Environmental Research 209 (2022) 112822

Contents lists available at ScienceDirect

Environmental Research
journal homepage: www.elsevier.com/locate/envres

Immobilization of Halomonas halodurans and Bacillus halodurans in packed

bed bioreactor for continuous removal of phenolic impurities in
waste water
N. Benit a, *, A. Amala Lourthuraj b, K. Barathikannan c, Ashraf Abdel-Fattah Mostafa d,
Hissah Abdulrahman Alodaini d, Mohamed Taha Yassin d, Ashraf Atef Hatamleh d
a
Department of Botany, Holycross College, Nagercoil, Kanyakumari District, Tamilnadu, India
b
Department of Biochemistry, Guru Nanak College (autonomous), Velachery, Chennai, 600042, Tamil nadu, India
c
Department of Food Science and Biotechnology, College of Agriculture and Life Science, Kangwon National University, Chuncheon, 24341, South Korea
d
Department of Botany and Microbiology, College of Science, King Saud University, P.O. Box 2455, Riyadh, 11451, Saudi Arabia

A R T I C L E I N F O A B S T R A C T

Keywords: Phenol is an organic contaminant widely distributed in wastewater. Biodegradation is one of the suitable
Wastewater methods used to remove phenol from the wastewater. In this study, the bacterial laccase and pectinase were
Phenol analyzed and phenol degradation potential was studied. A total of six bacterial strains were selected and their
Biocatalyst
phenol degrading potentials were studied. Laccase and pectinase producers were screened on substrate agar
Biodegradation
Immobilization
plates and several strains produced these enzymes in submerged fermentation. Among these enzyme producing
strains, strain PD8 and PD22 exhibited potent phenol degrading ability than other strains. These two bacterial
strains (Halomonas halodurans PD8 and Bacillus halodurans PD22) exhibited maximum growth in phenol-
supplemented culture medium. These two organisms grown well at wide pH values (pH 3.0 and 10.0), survive
well between 20 ◦ C and 50 ◦ C, and showed growth between 1 and 10% sodium chloride concentration. The
lyophilized enzyme from PD8 and PD22 were immobilized with alginate beads cross liked with divalent cations.
At 1% alginate, the binding efficiency was 40.2 ± 2.9% and it improved up to 2.0% concentration (67.5 ± 4.2%)
and further increase on alginate concentration affected binding efficiency. Phenol degradation was maximum
within 10 h of treatment in the immobilized packed bed column reactor (83.1 ± 3.2%) and colour removal
efficiency was maximum at 12 h treatment (82.1 ± 3.9%). After four successive experimental trials more than
40% efficiency was achieved.

1. Introduction industrial wastewater reduced environmental pollution, since many of

these phenolic compounds are highly toxic to the environment and
Phenol is one of the aromatic compounds and is a major pollutant in freshwater ecosystem. In recent years much more efforts have been
terrestrial and aquatic ecosystem. It is widely applied in the synthesis of taken to eliminate phenolic compound from the polluted wastewater.
various organic compounds, including, drugs, dyes, pesticides, and The removal of phenolic compounds from wastewater is achieved either
agricultural chemicals (Zhang et al., 2013). High concentration of by enzymatic activities of microorganisms, or by physico-chemical
phenol was detected in the environment (50–1500 mg/L) and also in treatment of which enzymatic methods have several advantages than
drinking water (0.21–1130 mg/L) (Jarvis et al., 1985). Phenols have other methods. Enzymatic methods target specific pollutants from the
toxic carcinogenic properties, and poses serious risk to environment and wastewater. These methods increase the potency of the biocatalyst used
public health. Phenolic compounds are determined in the effluents of for various uses, including soil and wastewater treatment (Vijayar­
various industries, including pulp and paper, textiles, dye industries, aghavan et al., 2012, 2021; Vijayaraghavan and Vincent, 2012).
metal coating, wood preservation, resin and plastics, petroleum refining Nanocomposites have various applications in waste management.
and coal conversion. Bioremediation of phenolic compounds from The fabricated electrochemical sensor based on ZnS/Au/f-MWCNT

* Corresponding author.
E-mail address: [email protected] (N. Benit).

https://doi.org/10.1016/j.envres.2022.112822
Received 28 September 2021; Received in revised form 12 December 2021; Accepted 19 January 2022
Available online 29 January 2022
0013-9351/© 2022 Elsevier Inc. All rights reserved.
N. Benit et al. Environmental Research 209 (2022) 112822

nanocomposites showed rapid detection of various pollutants from the morphologically distinct colonies were isolated and streaked on minimal
environment, including 4-nitrophenol (Naik et al., 2020, 2021a). Gold agar medium. Then it was incubated for 48 h at 30 ± 2 ◦ C and the iso­
nanoparticles have been used for the determination of lead from the lated bacterial strain was cultured in submerged fermentation with
aqueous media (Yu et al., 2021a,b). Photocatalytic method has been phenol substrate. Briefly, the isolated bacteria were inoculated in
suggested to degrade Rhodamine B. Naik et al. (2021b) have been used nutrient broth medium and incubated for 18 h and the cells were har­
metal and nonmetal-codoped ZnO for the degradation of Rhodamine B vested by centrifugation at 10,000 rpm for 10 min. Then it was re-
under solar light irradiation. The multidimensional interface of suspended in minimal medium containing increased concentration of
1D-Bi2S3@ 2D-GO/3D-BiOI ternary nanocomposites has been devel­ phenol (750 mg/L) and incubated for 48 h. Sample was withdrawn for
oped to treat tetracycline under visible-light (Arumugam et al., 2021). every 12 h and phenol degradation potential was determined. The cell
Photocatalytic degradation method has been considered as an effi­ fee supernatant was filtered using 0.22-μm filter and used as the crude
cient method for the degradation of various toxic substances in waste­ sample. The amount of phenol content in the sample was determined
water discharged from the industries. Moreover, the application of using HPLC equipped with C18 column. The mobile phase was adjusted
photocatalytic approach is associated with various limitations, such as to 0.5 mL/min and a diode array detector (DAD) was used to detect
the requirement of large amount of catalysts and lengthy procedures phenol concentration at 270 nm. External standard was used to correlate
(Theerthagiri et al., 2021; Yu et al., 2021a,b). Howeover, photocatalysis the peak area and phenol content was determined.
method combined with ultrafiltration, microwave, photothermal, elec­
trochemical and biological reactor has been suggested to treat pollutants 2.2. Screening and analysis of pectinase producing bacteria isolated from
from the environment (Lee et al., 2021; Naik et al., 2021). Microbial the coir retting wastewater
enzymes and reactors have been used to treat wastewater discharged
from the industries (Lekshmi et al., 2021; Vijayaraghavan et al., 2021). Bacterial strains were grown on pectin mineral salt medium for
Microorganisms, including bacteria and fungi synthesize laccases. initial screening of pectinase producing bacteria. The culture medium
Several bacteria degrade phenol using either aerobic or anaerobic composition consists of (g/L); pectin – 5, yeast extract – 1, (NH4)2SO4 –
metabolic process and utilize it as a sole source of carbon and energy 2, Na2HPO4 – 5, KH2PO4 – 1, MgSO4 – 0.2, CaCl2 – 0.001, MnSO4 -
(Gu, 2016; Wang et al., 2015). Bacteria such as, Arthrobacter sp. W1 0.00001, ZnSO4 - 0.00007, CuSO4 - 0.00005 and agar – 15. The pH of the
(Wang et al., 2015), Halomonas sp. (Haddadi and Shavandi, 2013), Ba­ culture medium was maintained at 7.2 ± 2 and the plates were incu­
cillus sp. (Iqbal et al., 2018), Acinetobacter sp. (Iqbal et al., 2018), Kocuria bated for 2 days at 30 ± 2 ◦ C. The bacteria were inoculated into mineral
sp. (Wu et al., 2018), Burkholderia sp. (Arora and Jain, 2012), and salt medium and incubated for 48 h at 150 rpm. Then the culture was
Pseudomonas sp. (Chen et al., 2017) degrade phenol. Laccases contain centrifuged and the supernatant was used for the determination of
copper ions at the catalytic centre and involved in the degradation of pectinase activity by well diffusion method. In the substrate agar plates,
lignin (Kiiskinen et al., 2002). These enzymes have various applications wells were made at 6 mm diameter. All wells were filled with crude
in the industries, as textile dye bleaching, bio-bleaching of coconut husk, enzyme and incubated for 30 ± 2 ◦ C for overnight. After 18 h incubation,
bio-pulping prevention of wine decolouration, enzymatic conversion of the plates were flooded with iodine solution. A clear zone formed around
various chemical intermediates, soil bioremediation, biodegradation of the wells was measured (mm).
various micropollutants, removal of phenolic compounds and endocrine
disruptors (Solé et al., 2012; Divya et al., 2013). Laccases from the 2.3. Pectinase assay
bacterial origin with high stability at extreme pH and temperature are
widely used for various industrial processes. Immobilization of enzymes Pectinase in the sample was determined by dinitrosalicylic acid
is one of the important techniques attracted much more attention in the (DNS) method. Quantification of enzyme assay was mainly based on the
treatment of wastewater. Immobilized enzymes have several advantages analysis of reducing sugars generated as a result of hydrolytic activity of
than free enzymes for various applications due to repeated usability and substrate by DNA reagent. Galactouronic acid was used as the standard
stability (Singh et al., 2021). Pollutant degrading enzymes are immo­ and for the determination of released reducing sugar. Pectinase activity
bilized on matrices. Chemical and physical methods are widely used to in the sample was finally expressed as enzyme units (U) per ml crude
entrap types of enzymes within various selected matrices. The oxidative sample. One enzyme unit was defined as the amount of enzyme which
enzymes such as, tyrosinases, laccases, and peroxidases are widely used releases 1 μmol of reducing sugar per min under standard assay
to remove phenols from the wastewater. These enzymes involved in the conditions.
oxidation of aromatic amines and phenolic compounds. Removal of
aromatic and phenolic compounds using fungal enzymes on various 2.4. Determination of laccases production
matrices has been reported previously. Moreover, laccases from fungal
strains are highly acidic in nature and are used to treat wastewater in The selected six bacterial strains were grown in 150 mL minimal
acidic nature and these failed to treat effluent from various industrial medium for the production of laccases. To the culture medium all six
wastewater. Hence, bacterial enzymes with high specificity and strains (PD2, PD9, PD8, PD21, PD22 and PD28) were inoculated and
maximum stability can be used to treat phenol from the industrial ef­ incubated at 30 ± 2 ◦ C and the pH of the medium was adjusted at 7.2 ±
fluents. The main aim of the study was to screen potent phenol 2. After two day the bacterial culture was centrifuged and the super­
degrading bacteria for the removal of phenolic compounds from the natant was applied for assaying the laccase activity. Laccase activity of
wastewater. the sample was assayed by dimethoxyphenol method. Enzyme activities
were finally expressed as μmol/min. One enzyme unit was defined as the
2. Materials and methods amount of sample that catalyzes the oxidation of dimethoxyphenol/min
at 30 ◦ C.
2.1. Isolation of phenol-degrading microorganisms from the coir retting
wastewater 2.5. Characterization of phenol-degrading bacterial strains

Enrichment method was used to isolate potent phenol-degrading Two maximum phenol-degrading bacterial strains were character­
bacteria from the natural environment. In order to isolate most potent ized by morphological, biochemical and molecular methods. These two
phenol-degrading bacteria, minimal medium was supplemented with bacterial strains (PD8 and PD22) were cultured in nutrient broth me­
500 mg/L phenol as the sole source of energy and carbon. The culture dium and incubated for 18 h at 32 ± 2 ◦ C. The cells were centrifuged at
plate was incubated for 48 h at 30 ◦ C. After incubation, the 10,000 rpm for 10 min. The pellet was washed several times with

2
N. Benit et al. Environmental Research 209 (2022) 112822

phosphate buffered saline solution and DNA was extracted using DNA 2.8. Immobilization of enzyme in a reactor
extraction kit as described by the manufacture. The 16S rDNA gene was
amplified using a PCR machine with primers (AAGGAGGTGATC­ Batch studies were performed in packed bed column bioreactor. The
CAGCCGCA and AGAGTTTGATCCTGGCT CAG), and Taq DNA poly­ immobilized active sodium alginate beads were loaded in the bioreactor
merase. The amplified DNA was purified and tested using a nanodrop (35 cm × 12.5 cm), and the headspace was 8 cm. The schematic pre­
machine. Then, the amplified 16S rDNA gene was sequenced and BLAST sentation of the used packed bed bioreactor system in the present study
search (ncbi.nlm.nih.gov/blast) was performed with other previously is described in Fig. 1. Phenol stock solution was prepared by mixing
published data bases submitted in GenBank. phenolic compounds such as catechol, pyrogallol, vanillic acid, tannic
acid, gallic acid, guaiacol, and resorcinol. These phenolic compounds
2.6. Optimized production of dual enzymes were common in coir retting industry. The prepared phenolic solution
was passed through the immobilized beads using a peristaltic pump for
The selection of the suitable variables for the production of laccase 10 h and the flow rate was 1 mL/min. The treated phenolic water sample
and pectinase were performed in 250-mL Erlenmeyer flask containing was collected from the sampling outlet hole and subjected for the
wastewater (100 mL) with various pH and temperatures were analyzed. determination of phenolic content from the water. Experiment was
The other factors such as carbon, nitrogen, surfactants and ions were performed in the laboratory scale reactor at 32 ± 2 ◦ C and the colour
also tested for their influence on pectinase and laccase production. removal efficiency was monitored using a UV–Visible
spectrophotometer.
2.7. Batch production and immobilization of partially purified enzyme
3. Results and discussion
The bacterial strains were cultured in optimized culture medium and
the cell free extract was precipitated with ammonium sulphate (30%– 3.1. Isolation of phenol-degrading bacteria
70% saturation). It was centrifuged and the precipitated enzyme was re-
suspended in phosphate buffer (0.1 M, pH 7.2) and it was dialyzed The strains PD2, PD8, PD21, PD22 and PD28 removed phenol within
against water (first change) and buffer (two changes). Laccase and 72 h incubation. Strain PD8 exhibited potent phenol degrading ability
pectinase activity were analyzed by standard method. The partially than other strains. Phenol degrading ability was only 16 ± 22% in strain
purified enzyme sample was mixed with sodium alginate solution PD9 (Fig. 2). Bacterial strains isolated from the soil and wastewater have
(2.5%) on an orbital shaker at 50 rpm for 60 min at ambient tempera­ the ability to degrade phenol from the medium at various concentrations
ture. Then the highly viscous alginate-enzyme was filled in a syringe and ranging from 120 to 1000 mg/L. Phenol is present in coir retting
the mixture was extruded into CaCl2 (0.5 M). The mixture was stored at wastewater at this concentration. So, the present study has significance
4 ◦ C for 3 h and the beads were filtered and washed with double distilled to remove phenol from the coir retting wastewater. Phenol-degrading
water. Total protein content was tested in the washed solution and bacteria have been isolated from various sources. Jayashree et al.
continued the procedure until there was no detectable level of protein in (2014) isolated phenol degrading organism from the coconut husk
the wash out solution. Pectinase and laccase activities were evaluated retting wastewater and bioelectricity generation have also been re­
from the wash out sample. The parameters such as, concentration of ported. Ereqat et al. (2018) characterized phenol degrading Bacillus
CaCl2, and sodium alginate were optimized and enzyme activities were thuringiensis J20 from olive waste. Phenol-degrading bacterial species
analyzed. have been isolated by Karthika et al. (2016) and reported the role of

Fig. 1. The schematic presentation of packed bed bioreactor developed using immobilized enzyme for the degradation of phenol from the wastewater.

3
N. Benit et al. Environmental Research 209 (2022) 112822

Fig. 2. Phenol removal by the selected six bacterial strains isolated from the coir retting pond. The strains were inoculated in to minimal medium containing 750 mg/
L at 32 ± 2 ◦ C.

bacteria on softness of coir fiber.

Table 2
Effect of pH on pectinase and laccase production by the strains PD8 and PD22.
3.2. Screening and analysis of pectinase and laccase producing bacteria
PD8 PD22

The bacterial strains from four different samples were grown on pH Pectinase Laccase Pectinase Laccase
activity (U/mL) activity (U/mL) activity (U/mL) activity (U/mL)
minimal medium containing pectin as the sole source of carbon. Table 1
reveals that all selected bacteria were able to utilize and grow on pectin 5.5 1.02 ± 0.03 1.3 ± 0.03 0.19 ± 0.76 0.41 ± 0.19
6 1.4 ± 0.92 1.51 ± 0.02 0.59 ± 0.48 1.19 ± 0.26
substrate. Among the bacterial strains PD8 showed improved production
65 2.9 ± 0.31 1.98 ± 0.05 2.75 ± 0.37 2.63 ± 0.17
of biomass and pectinase than other bacterial strains. The plate 7 4.87 ± 0.12 4.07 ± 0.24 3.86 ± 0.42 4.25 ± 0.28
screening method showed clear zone around the bacterial stains and 7.5 5.34 ± 0.08 5.3 ± 0.18 4.2 ± 0.15 4.9 ± 0.22
bacterial supernatant showed pectin degradation. A total of five bacte­ 8 5.08 ± 0.21 5.2 ± 0.76 4.01 ± 0.28 4.7 ± 0.36
rial strains showed potent pectinase production. Table 2 shows that 8.5 4.97 ± 0.02 4.8 ± 0.65 3.7 ± 0.22 4.4 ± 0.27
9 4.05 ± 0.17 4.65 ± 0.17 3.06 ± 0.16 2.05 ± 0.15
maximum pectinase production than other bacterial strains. The yield of
pectinase production varied widely among the bacterial species. Bacte­
ria have the ability to produce various hydrolytic enzymes, including degradation and finally two bacterial strains (PD8 and PD22) were
alkaline pectinases (Kavuthodi and Sebastian, 2018). The enzyme yield used for characterization studies. The analyzed bacterial strains showed
varied based on culture methods, for instance increased production of various degrees of resistance towards various concentrations of phenol.
pectinase has been obtained in solid state fermentation (Kaur and Gupta, Moreover, only two bacterial strains (PD8 and PD22) exhibited
2017). Chauhan et al. (2017) analyzed the significance of bacterial maximum growth in phenol-supplemented culture medium than other
laccases for various industrial processes and wastewater treatment. isolates. Hence, these two strains were selected for further studies. These
Biosurfactants have been reported as the inducers for the production of two organisms grown well at wide pH values (pH 3.0 and 10.0), survived
bacterial laccases and involved in phenol degradation (Zhou et al., well between 20 ◦ C and 50 ◦ C, and showed growth between 1 and 10%
2011). Kumar et al. (2018) have been characterized various microbial sodium chloride concentration, however maximum growth was
species for the production of laccases for the removal of phenol from the observed at 8% sodium chloride. The strain PD8 was Gram-positive,
wastewater. utilized-sorbitol, adonitol, inositol and arabinose, produced H2S and
phenylalanine deaminase. These two strains produced acid from D-
3.3. Identification of phenol-degrading bacteria fructose, D-ribose, D-arabinose, dextrose and D-galactose. The strain
PD22 was Gram-positive strain and hydrolyzed casein, and starch. It was
A total of six bacterial strains were selected based on phenol- tolerated up to 13% sodium chloride and showed growth under aerobic
condition. These two bacterial strains were identified as Halomonas
halodurans PD8 and Bacillus halodurans PD22. Regina et al. (2006) have
Table 1
Screening and production of pectinase and laccase by bacterial species isolated been reported hydrocarbon-utilizing bacterial strain isolated from crude
from the retting effluent. The zone of hydrolysis was measured and expressed as oil contaminated sites. Neumann et al. (2004) reported maximum sta­
mm and enzyme production was performed in submerged fermentation and bility of Pseudomonas strains up to 100 mg/L phenol concentration. Most
expressed as U/mL. of the organisms utilize phenol as the energy source and carbon. The
Pectinase activity Laccase
bacterial strain, B. thuringiensis have been reported to degrade various
activity environmental pollutants, including organophosphorus insecticides,
heavy metals, hydrocarbon from petroleum industry (Zeinat-Kamal
Bacterial Zone of Enzyme Zone of Enzyme
strain hydrolysis activity (U/ hydrolysis activity (U/ et al., 2008; Das et al., 2014; Das and Chandran, 2011). The bacterial
(mm) mL) (mm) mL) strain B. thuringiensis isolated from the mangrove sediment has the po­
PD2 14 ± 2 3.48 ± 0.31 17 ± 3 1.03 ± 0.09 tential to utilize phenol for carbon and energy source (Kaflizadeh and
PD7 10 ± 1 2.21 ± 0.29 10 ± 1 1.8 ± 0.7 Mokhtari, 2015). The screening of native bacterial species from local
PD8 21 ± 3 4.87 ± 0.17 19 ± 2 3.9 ± 0.42 contaminated wastewater has been reported to be highly efficient and
PD21 19 ± 1 4.4 ± 0.09 17 ± 2 2.7 ± 0.2 adaptive than non-native bacteria as biodegraders. Hence, the screening
PD22 20 ± 3 3.9 ± 0.21 14 ± 1 1.06 ± 0.04
of phenol-degrading indigenous bacterial strains has been
PD28 16 ± 2 4.1 ± 0.14 15 ± 2 6.6 ± 0.04

4
N. Benit et al. Environmental Research 209 (2022) 112822

recommended for the removal or degradation of phenol from various Table 4

environments (Silva et al., 2009). Effect of alginate concentration on laccase and pectinase binding efficiency. The
initial concentration of pectinase and laccase was 250 U/mL.
3.4. Enhanced production of laccase and pectinase Alginate (%) Binding efficiency (%) Ca2+ ions (M) Binding efficiency (%)

1 40.2 ± 2.9 0.02 37.8 ± 4.4

The variable such as, substrate concentration, incubation time (Patel 1.5 49.3 ± 4.8 0.04 49 ± 1.9
et al., 2009), bacterial cell suspension (Prasad et al., 2005), pH and 2 67.5 ± 4.2 0.06 62 ± 3.8
2.5 65.3 ± 1.2 0.08 71 ± 5.1
temperature (Nyanhongo et al., 2002), carbon source (Elshafei et al.,
3 60.7 ± 5.8 0.1 69 ± 2.75
2012) and minerals (Revankar and Lele, 2006) improved laccase pro­
duction. Laccase production was maximum at alkaline pH values. At pH
5.5, laccase production was 1.02 ± 0.03 U/mL and 0.41 ± 0.19 U/mL, chemical stability, high porosity and trapped enzymes in the suitable
respectively for the strain PD8 and PD22. The enzyme yield improved condition.
continuously up to pH 7.5 for both P8 and PD22 (5.3 ± 0.18 U/mL and
4.9 ± 0.22 U/mL). Pectinase biosynthesis also improved at pH 7.5 for
strain PD8 (5.34 ± 0.08 U/mL) and PD22 (4.2 ± 0.15), respectively. In 3.6. Improved degradation of phenol and reduction of colour in
PD8, pectinase production was improved at higher pHs. At 25 ◦ C, pec­ immobilized beads
tinase production was 4.9 ± 0.17 U/mL and at 35 ◦ C pectinase yield was
6.1 ± 0.12 U/mL (Table 2). Pectinase production was maximum at 35 ◦ C Batch experiments were used in the immobilized laboratory biore­
(5.7 ± 0.35 U/mL). Laccase production was high at 35 ◦ C for both actor containing Ca-alginate beads with optimum laccase and pectinase
bacterial strains (5.4 ± 0.19 U/mL and 4.98 ± 0.96 U/mL) (Table 3). activity. Table 5 shows the removal of colour and phenol degradation
Alkaline pH was optimum and induced the production of laccase and from the solution in an immobilized packed bed bioreactor run for 12 h.
pectinase. Other than these physical factors, trace elements, yeast In this experiment 49.2 ± 2.9% colour reduction was observed within 2
extract and MgSO4, improved laccase production in Streptomyces psam­ h of treatment and 52.1 ± 3.4% phenol was reduced at this incubation
moticus (Niladevi et al., 2009). time. Phenol degradation was maximum within 10 h of treatment in the
immobilized packed bed column reactor (83.1 ± 3.2%) and colour
3.5. Immobilization of enzymes removal efficiency was maximum after 12 h treatment (82.1 ± 3.9%).
Phenolic compounds in the sample aggregated with phenol oxidases
The bacteria were cultured in optimized culture medium and the and have been reported previously. This aggregation indicated
culture supernatant was subjected for enzyme assay. Then the partially improved laccase activity. These precipitated polymerized products
purified enzyme from PD8 and PD22 were immobilized with alginate were separated by filtration after the completion of catalytic cycle. At
beads cross liked with divalent cations. Alginate has been generally used the end of first catalytic cycle, 79.4 ± 3.2 colour removals were achieved
as the immobilizing agent and has been recognized as the natural and phenol reduction was 78.2 ± 3.5%. The stability of the immobilized
polymer and cross link with calcium chloride. Calcium chloride was laccase and pectinase was analyzed and the recovered immobilized
used for the preparation of Ca-alginate beads. Calcium alginate has been beads were further used for successive runs. A total of six runs were
used as a support material for the immobilization of enzymes. The performed and the results were described in Table 6. Enzyme leaching is
development and the structural and mechanical properties of alginate one of the major concerns with alginate beads. Moreover, after four
beads mainly based on concentration of the selected cations, nature of successive experimental trials more than 40% efficiency has been ach­
the cations and alginate concentrations. The influence of alginate con­ ieved. The present finding showed 22.7 ± 3.8% phenol degradation
centration on laccase and pectinase binding efficiency of calcium beads potential after sixth catalytic cycle. The decreased phenol degradation
was described in Table 4. The concentration of pectinase and laccase was was mainly due to immobilized enzyme leaching. In an initial operation,
250 IU/mL. At 1% alginate, the binding efficiency was 40.2 ± 2.9% and leaching percentage was high and leaching percentage decreased after
it improved up to 2.0% concentration (67.5 ± 4.2%) and further in­ four or five catalytic cycle. This is mainly due to the fact that enzymes
crease on alginate concentration affected binding efficiency. Likewise, coated at the surface of the beads leached out easily while the enzymes
addition of Ca2+ ions, improved binding efficiency up to 0.08 M con­ entrapped in the inner part of the beads retained activity for long
centration and binding efficiency was71 ± 5.1% and depleted binding duration. Laccases involved in the degradation of phenolic compounds
efficiency at 0.1 M concentration (69 ± 2.75%). The Table 4 showed the and have been reported previously (Lante et al., 2000). Moreover, the
influence of initial alginate concentration on the enzyme binding effi­ wastewater contains various phenolic compounds from different sour­
ciency of the immobilized beads. The formulated beads were highly ces, including phenol, steel, disinfectant, rubber proofing, plastic,
transparent at lower concentrations of alginate, and bead shape was pharmaceutical, coal conversion, petrochemical and petroleum in­
affected. In this study, the concentration of cations varied between 0.02 dustries (Nayak and Singh, 2007). Krastanov (2000) earlier reported the
and 0.1%. It has been earlier stated that Ca-alginate beads have high application of co-immobilization of tyrosinases and laccases for the
complete degradation of phenolics from industrial sources. The present
finding on phenol removal from the mixture of phenolic compounds
Table 3
indicated that co-immobilization of pectinase and laccase from two
Effect of temperature on pectinase and laccase production by the strains PD8 and
PD22.
bacterial sources revealed that these two enzymes are active at alkaline
pHs. Bacterial laccase and pectinase are active at alkaline pHs, moreover
Temperature PD8 PD22
(◦ C)
Table 5
Pectinase Laccase Pectinase Laccase
Degradation of phenol and colour removal by the immobilized biocatalyst.
activity (U/ activity (U/ activity (U/ activity (U/
mL) mL) mL) mL) Time (h) Colour removal (%) Phenol reduction (%)

25 4.9 ± 0.17 0.17 ± 0.21 0.19 ± 0.02 0.014 ± 2 49.2 ± 2.9 52.1 ± 3.4
0.01 4 57.3 ± 42 59.4 ± 3.1
30 6.2 ± 0.21 4.2 ± 0.29 1.84 ± 06 2.87 ± 0.27 6 67.5 ± 1.8 72.3 ± 2.7
35 6.1 ± 0.12 5.4 ± 0.19 5.7 ± 0.17 4.98 ± 0.96 8 72.9 ± 3.3 74.2 ± 1.5
40 4.1 ± 0.18 4.38 ± 0.14 5.2 ± 0.35 4.72 ± 0.22 10 79.4 ± 4.2 83.1 ± 3.2
45 3.9 ± 0.1 1.02 ± 0.28 4.4 ± 0.23 461 ± 0.49 12 82.1 ± 3.9 74.7 ± 1.9

5
N. Benit et al. Environmental Research 209 (2022) 112822

Table 6 Ereqat, S.I., Abdelkader, A.A., Nasereddin, A.F., Al-Jawabreh, A.O., Zaid, T.M., Letnik, I.,
Removal of colour and reduction of phenol in the wastewater treated with Abdeen, Z.A., 2018. Isolation and characterization of phenol degrading bacterium
strain Bacillus thuringiensis J20 from olive waste in Palestine. J. Env. Sci. Health, Part
immobilized enzyme. A. 53 (1), 39–45.
Catalytic cycle Colour removal (%) Phenol reduction (%) Gu, J.D., 2016. Biodegradation testing: so many tests but very little new innovation.
Appl. Environ. Biotechnol. 1, 92–95.
1 79.4 ± 3.2 78.2 ± 3.5 Haddadi, A., Shavandi, M., 2013. Biodegradation of phenol in hypersaline conditions by
2 74.2 ± 2.8 76.4 ± 1.8 Halomonas sp. strain PH2-2 isolated from saline soil. Int. Biodeterior. Biodegrad. 85,
3 69.8 ± 1.9 59.2 ± 2.9 29–34.
4 57.3 ± 3.4 45.8 ± 3.1 Iqbal, A., Arshad, M., Hashmi, I., Karthikeyan, R., Gentry, T.J., Schwab, A.P., 2018.
5 42.1 ± 1.5 39.2 ± 1.4 Biodegradation of phenol and benzene by endophytic bacterial strains isolated from
6 29.6 ± 2.3 22.7 ± 3.8 refinery wastewater-fed Cannabis sativa. Environ. Technol. 39, 1705–1714.
Jarvis, S., Straube, R., Williams, A., Bartlett, C., 1985. Illness associated with
contamination of drinking water supplies with phenol. Br. Med. J. 290, 1800–1802.
Jayashree, C., Arulazhagan, P., Kumar, S.A., Kaliappan, S., Yeom, I.T., Banu, J.R., 2014.
most of the laccases from fungal origin showed decreased activity Bioelectricity generation from coconut husk retting wastewater in fed batch
(Samanta, 2021; Muthuvelu et al., 2020). Hence, bacterial laccases are operating microbial fuel cell by phenol degrading microorganism. Biomass
suitable for the degradation phenol from the wastewater because of Bioenergy 69, 249–254.
Kaflizadeh, F., Mokhtari, S., 2015. Isolation and identification of phenol degrading
alkaline nature. bacteria from mangrove sediments in the Persian Gulf (Asaluyeh) and their growth
kinetics assay. Biomed. Pharmacol. J. 6 (2), 189–196.
4. Conclusions Karthika, S., Reshma, M.J., Wilson, P.A., Das, R.A., Sarma, U.S., Harikrishnan, K.,
Thomas, S., 2016. Characterization and evaluation of phenol degrading Bacillus spp.
for enhancing the softness of coir fiber. J. Nat. Fibers 13 (3), 253–260.
In this study analyzed the major advantages of co-immobilized lac­ Kaur, S.J., Gupta, V.K., 2017. Production of pectinolytic enzymes pectinase and pectin
cases and pectinase for the degradation of phenolic compounds from the lyase by Bacillus subtilis SAV-21 in solid state fermentation. Ann. Microbiol. 67 (4),
333–342.
wastewater. The degradation pattern of phenolic compounds showed
Kavuthodi, B., Sebastian, D., 2018. Review on bacterial production of alkaline pectinase
that Halomonas halodurans PD8 and Bacillus halodurans PD22 pectinase with special emphasis on Bacillus species. Biosci. Biotechnol. Res. Commun 11,
and laccases are able of removing tested phenolic compounds. The 18–30.
packed bed column reactor immobilized with dual biocatalyst improved Kiiskinen, L.L., Viikari, L., Kruus, K., 2002. Purification and characterisation of a novel
laccase from the ascomycete Melanocarpus albomyces. Appl. Microbiol. Biotechnol.
enzymes stability and phenol biodegradation. It can be concluded that 59 (2–3), 198–204.
immobilized H. halodurans PD8 and B. halodurans PD22 may be applied Krastanov, A., 2000. Removal of phenols from mixtures by co-immobilized laccase/
for effective removal of phenolic compounds from various natural tyrosinase and Polyclar adsorption. J. Ind. Microbiol. Biotechnol. 24 (6), 383–388.
Kumar, R., Raizner, Y., Kruh, L.I., Menashe, O., Azaizeh, H., Kapur, S., Kurzbaum, E.,
sources. 2018. Extracellular laccase production and phenolic degradation by an olive mill
wastewater isolate. Grasas Aceites 69 (1) e231-e231.
Credit to author statement Lante, A., Crapisi, A., Krastanov, A., Spettoli, P., 2000. Biodegradation of phenols by
laccase immobilised in a membrane reactor. Process Biochem. 36 (1–2), 51–58.
Lee, S.J., Theerthagiri, J., Nithyadharseni, P., Arunachalam, P., Balaji, D., Kumar, A.M.,
NB, AAL and BK designed the experiment. NB, AAL and BK per­ Madhavan, J., Mittal, V., Choi, M.Y., 2021. Heteroatom-doped graphene-based
formed the experiment. AAM, HAA, MTY, BK and NB performed vali­ materials for sustainable energy applications: a review. Renew. Sustain. Energy Rev.
143, 110849.
dated the experimental results and analyzed the data. NB and BK drafted Lekshmi, R., Nisha, S.A., Vasan, P.T., Kaleeswaran, B., 2021. A comprehensive review on
the manuscript and all the authors approved the manuscript. tannase: microbes associated production of tannase exploiting tannin rich agro-
industrial wastes with special reference to its potential environmental and industrial
applications. Environ. Res. 111625.
Declaration of competing interest
Muthuvelu, K.S., Rajarathinam, R., Selvaraj, R.N., Rajendren, V.B., 2020. A novel
method for improving laccase activity by immobilization onto copper ferrite
The authors declare that they have no known competing financial nanoparticles for lignin degradation. Int. J. Biol. Macromol. 152, 1098–1107.
Naik, S.S., Lee, S.J., Theerthagiri, J., Yu, Y., Choi, M.Y., 2021. Rapid and highly selective
interests or personal relationships that could have appeared to influence
electrochemical sensor based on ZnS/Au-decorated f-multi-walled carbon nanotube
the work reported in this paper. nanocomposites produced via pulsed laser technique for detection of toxic nitro
compounds. J. Hazard Mater. 126269.
Acknowledgement Naik, S.S., Lee, S.J., Begildayeva, T., Yu, Y., Lee, H., Choi, M.Y., 2020. Pulsed laser
synthesis of reduced graphene oxide supported ZnO/Au nanostructures in liquid
with enhanced solar light photocatalytic activity. Environ. Pollut. 266, 115247.
The authors extend their appreciation to the Researchers supporting Naik, S.S., Lee, S.J., Theerthagiri, J., Yu, Y., Choi, M.Y., 2021a. Rapid and highly
project number (RSP-2021/362) King Saud University, Riyadh, Saudi selective electrochemical sensor based on ZnS/Au-decorated f-multi-walled carbon
nanotube nanocomposites produced via pulsed laser technique for detection of toxic
Arabia. nitro compounds. J. Hazard Mater. 126269.
Naik, S.S., Lee, S.J., Yeon, S., Yu, Y., Choi, M.Y., 2021b. Pulsed laser-assisted synthesis of
References metal and nonmetal-codoped ZnO for efficient photocatalytic degradation of
Rhodamine B under solar light irradiation. Chemosphere 274, 129782.
Nayak, P.S., Singh, B.K., 2007. Instrumental characterization of clay by XRF, XRD and
Arora, P.K., Jain, R.K., 2012. Metabolism of 2-chloro-4-nitrophenol in a Gram negative
FTIR. Bull. Mater. Sci. 30 (3), 235–238.
bacterium, Burkholderia sp. RKJ 800. PLoS One 7, e38676.
Neumann, G., Teras, R., Monson, L., Kivisaar, M., Schauer, F., Heipieper, H.J., 2004.
Arumugam, M., Lee, S.J., Begildayeva, T., Naik, S.S., Yu, Y., Lee, H., Theerthagiri, J.,
Simultaneous degradation of atrazine and phenol by Pseudomonas sp. strain ADP:
Choi, M.Y., 2021. Enhanced photocatalytic activity at multidimensional interface of
effects of toxicity and adaptation. Appl. Environ. Microbiol. 70, 1907–1912.
1D-Bi2S3@ 2D-GO/3D-BiOI ternary nanocomposites for tetracycline degradation
Niladevi, K.N., Sukumaran, R.K., Jacob, N., Anisha, G.S., Prema, P., 2009. Optimization
under visible-light. J. Hazard Mater. 404, 123868.
of laccase production from a novel strain—Streptomyces psammoticus using response
Chauhan, P.S., Goradia, B., Saxena, A., 2017. Bacterial laccase: recent update on
surface methodology. Microbiol. Res. 164, 105–113.
production, properties and industrial applications. 3 Biotech 7 (5), 1–20.
Nyanhongo, G.S., Gomes, J., Gübitz, G., Zvauya, R., Read, J.S., Steiner, W., 2002.
Chen, J., Zhang, L., Jin, Q., Su, C., Zhao, L., Liu, X., et al., 2017. Bioremediation of phenol
Production of laccase by a newly isolated strain of Trametesmodesta. Bio Technol. 84
in soil through using a mobile plant-endophyte system. Chemosphere 182, 194–202.
(3), 259–263.
Das, N., Chandran, P., 2011. Microbial degradation of petroleum hydrocarbon
Patel, H., Gupte, A., Gupte, S., 2009. Effect of different culture conditions and inducers
contaminants: an overview. Biotechnol. Res. Int. 1–13, 2011.
on production of laccase by a basidiomycete fungal isolate Pleurotus ostreatus HP-1
Das, P., Sinha, S., Mukherjee, S.K., 2014. Nickel bioremediation potential of Bacillus
under solid state fermentation. Bioresources 4 (1), 268–284.
thuringiensis KUNi1 and some environmental factors in nickel removal. Ann. Finance
Prasad, K.K., Mohan, S.V., Rao, R.S., Pati, B.R., Sarma, P.N., 2005. Laccase production by
18, 169–177.
Pleurotus ostreatus 1804: optimization of submerged culture conditions by Taguchi
Divya, L.M., Prasanth, G.K., Sadasivan, C., 2013. Elimination of estrogenic activity of
DOE methodology. Biochem. Eng. J. 24 (1), 17–26.
thermal paper using laccase from Trichoderma sp NFCCI-2745. Appl. Biochem.
Regina, O.E., Emuobonuvie, I.F., Roseline, U.E., 2006. Growth responses of bacterial
Biotechnol. 169 (4), 1126–1133.
isolates on various concentrations of crude oil. Jounal Am. Sci. 2, 13–16.
Elshafei, A.M., Hassan, M.M., Haroun, B.M., Elsayed, M.A., Othman, A.M., 2012.
Revankar, M.S., Lele, S.S., 2006. Enhanced production of laccase using a new isolate of
Optimization of laccase production from Penicillium martensii NRC 345. Adv. Life Sci.
white rot fungus WR-1. Process Biochem. 41 (3), 581–588.
2 (1), 31–37.

6
 Chemosphere 291 (2022) 133057

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Integrating anaerobic digestion and microalgae cultivation for dairy

wastewater treatment and potential biochemicals production from the
harvested microalgal biomass
Adi Kusmayadi a, Po-Han Lu b, Chi-Yu Huang b, c, *, Yoong Kit Leong a, Hong-Wei Yen a,
Jo-Shu Chang a, c, d, **
a
Department of Chemical and Materials Engineering, Tunghai University, Taichung, Taiwan
b
Department of Environmental Science and Engineering, Tunghai University, Taichung, Taiwan
c
Research Center for Smart Sustainable Circular Economy, Tunghai University, Taichung, Taiwan
d
Department of Chemical Engineering, National Cheng Kung University, Tainan, Taiwan

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• C. sorokiniana SU-1 can treat dairy

wastewater with simultaneous biomass
production.
• High-strength dairy wastewater was
treated by combining AD and algal
cultivation.
• 4.25 g/L biomass concentration and
93% pollutants removal was obtained.
• Highest biomass productivity of 0.37 g/
L/d was obtained at 1 g/L inoculums
size.
• The harvested microalgal biomass had
12% lipids and 18% proteins content.

A R T I C L E I N F O A B S T R A C T

Handling Editor: Derek Muir Utilizing wastewaters as feedstock for microalgal cultivation has the dual benefits of water-saving and low
nutrient costs, with simultaneous remediation of pollutants and generation of value-added biochemical products.
Keywords: This study employed two different strategies to treat raw dairy wastewaters with moderate and high chemical
Dairy wastewater oxygen demand (COD) levels. For moderate-COD dairy wastewater, the wastewater was directly utilized as
Anaerobic digestion
feedstock for algal cultivation, in which the effects of wastewater dilution ratios and algal inoculum sizes were
Microalgae
investigated. The results show that the microalga strain used (Chlorella sorokiniana SU-1) was capable of
Nutrient removal
Lipids obtaining a high biomass concentration of 3.2 ± 0.1 g/L, accompanied by 86.8 ± 6%, 94.6 ± 3%, and 80.7 ± 1%,
removal of COD, total phosphorus (TP) and total nitrogen (TN), respectively. Meanwhile, the obtained micro­
algal biomass has lipids content of up to 12.0 ± 0.7% at a wastewater dilution ratio of 50% and an inoculum size
of 2 g/L. For high-COD dairy wastewater, an integrated process of anaerobic digestion and microalgal phycor­
emediation was employed, and the effect of inoculum sizes was also studied. The inoculum size of 2 g/L gave
highest biomass production of 4.25 ± 0.10 g/L with over 93.0 ± 2.0% removal of COD, TP, and TN. The har­
vested microalgal biomass has lipids and protein content of 12.5 ± 2.2% and 18.0 ± 2.2%, respectively. The
present study demonstrated potential microalgal phycoremediation strategies for the efficient COD removal and

* Corresponding author. Department of Environmental Science and Engineering, Tunghai University, Taichung, Taiwan.
** Co-corresponding author. Department of Chemical and Materials Engineering, Tunghai University, Taichung, Taiwan.
E-mail addresses: [email protected] (C.-Y. Huang), [email protected] (J.-S. Chang).

https://doi.org/10.1016/j.chemosphere.2021.133057
Received 25 September 2021; Received in revised form 12 November 2021; Accepted 22 November 2021
Available online 25 November 2021
0045-6535/© 2021 Elsevier Ltd. All rights reserved.
A. Kusmayadi et al. Chemosphere 291 (2022) 133057

nutrients recovery from dairy wastewater of different COD levels with simultaneous production of microalgal
biomass which contains valuable components, such as protein and lipids.

1. Introduction However, the excessively high dilution ratios could increase wastewater
volumetric flow, additional need for energy and chemicals, as well as the
Although there was a disruption to the food markets by the COVID- bigger size of the reactor (Khalid et al., 2016). Besides, if the algal-based
19 pandemic in early 2020, the dairy market has gradually recovered wastewater treatment is conducted in an open pond system, the
and became rather robust during 2021. The United States Department of contamination risk by other microbes will also greatly arise (Pacheco
Agriculture (USDA) has predicted that the production of dairy products et al., 2020). Therefore, for efficient nutrient removal, a rapid-growing
can reach 227.9 billion pounds in 2021 (Wankar et al., 2021). The in­ alga that can outgrow any other microorganisms is essential. The rapid
crease in global dairy production causes a rise in the number of dairy algal growth can increase the efficiency of wastewater treatment owing
farms, which in turn increases the amount of wastewater effluent dis­ to faster nutrient utilization by the algae (Khalid et al., 2016). In addi­
charged. The wastewaters discharged from dairy farms contains chem­ tion, higher inoculum size is considered as a crucial factor to promote
ical oxygen demand (COD) at a level of 80–95,000 mg/L with abundant the competitiveness of microalgae in the combined bacteria-algae
nutrients, such as total phosphate (TP) and total nitrogen (TN) at a level system.
of 9–280 mg/L and 14–830 mg/L as a result of different culturing and The objective of this study was to evaluate the effectiveness of
milking systems (Daneshvar et al., 2019). The high COD, phosphorus, integrating anaerobic digestion and microalgae cultivation for the
and nitrogen levels in dairy wastewater can cause severe eutrophication treatment of dairy wastewater with moderate and high COD levels,
and disrupt the ecosystem in the receiving water bodies (Nayak et al., along with biochemicals production. The study was performed in two
2016). Therefore, appropriate treatment of dairy wastewater before parts. In the first part, the raw dairy wastewater with a moderate COD
releasing it into natural water bodies is needed to diminish the risk of level was used directly as the culture medium for the cultivation of the
environmental deterioration. microalgae C. sorokiniana SU-1 taking dilution ratio and inoculum size
Regarded as an effective waste treatment technology, anaerobic as the investigated parameters. In the second part, the high-COD raw
digestion (AD) can be employed to recover organic compounds by dairy wastewater was first treated by AD, followed by microalgal
producing biogas or other high-value products (Xia et al., 2018). AD is cultivation with various inoculum sizes. This study demonstrates a
particularly suitable for wastewater or solid wastes of high organic sustainable approach to treating dairy wastewater and biomass pro­
contents. Nevertheless, the AD process also generates digestates, and the duction using integrated anaerobic digestion and microalgae-based
digested effluent contains a high concentration of phosphate and ni­ wastewater treatment.
trogen, which causes secondary pollution (Li et al., 2020). The con­
ventional aerobic process employed to treat digestate requires vigorous 2. Materials and methods
aeration and the addition of organic matters for nitrogen removal, while
the recovery of nutrients was not efficient (Li et al., 2020). Nowadays, 2.1. Microalgae strain and pre-culture condition
anaerobic digestate is frequently utilized as a bio-fertilizer in the agri­
cultural field (Ma et al., 2017). Nevertheless, when farmland is located The microalgae employed in this study was C. sorokiniana SU-1 iso­
far from the AD plant, the costs of transportation may be significant, and lated from southern Taiwan. The pre-culture medium utilized in this
demand for the digestate may be relatively restricted during the study was BG-11 medium that consisted of (g/L): NaNO3, 1.5; K2HPO4,
non-growing season (Li et al., 2021). Therefore, proper management of 0.04; MgSO4⋅7H2O, 0.075; Citric acid, 0.006; Na2CO3, 0.02; CaCl2,
AD digestate is needed to recover the nutrient compounds and fulfill the 0.036; Ferric (ammonium) citrate, 0.006; EDTA⋅2Na, 0.001; H3BO3, 2.8
discharge requirements effectively. × 10− 3; ZnSO4⋅7H2O, 2.2 × 10− 4; MnCl2⋅4H2O, 1.8 × 10− 3;
Microalgae as solar-driven cell factories have garnered great atten­ Na2MoO4⋅2H2O, 3.9 × 10− 4; CuSO4⋅5H2O, 7.9 × 10− 5; Co (NO3)2⋅6H2O,
tion as a potential feedstock for biofuel, with simultaneous production of 4.9 × 10− 5. The pre-culture of C. sorokiniana SU-1 was carried out in a
valuable metabolites, such as poly-unsaturated fatty acids, pigments, 200 ml photobioreactor illuminated at a light intensity of 140 μmol m− 2
and carbohydrates (Aravantinou et al., 2013; Chen et al., 2020). How­ s− 1 and aerated with 2% CO2 at a 200 ml/min flow rate according to our
ever, one of the key disadvantages of large-scale microalgae cultivation previous study (Chen et al., 2020).
is the immense water and nutrient consumption, which leads to exor­
bitant costs (Kuo et al., 2015). Microalgae-based phycoremediation 2.2. Characterization of wastewater composition and pre-treatment
technologies for simultaneous pollutant elimination and nutrient re­ condition
covery have been an emerging topic in recent years because of their
various benefits. Following the ‘waste-to-fuel’ approach, the The low-strength dairy wastewater used was taken from the dairy
phycoremediation-derived algal biomass can be integrated into the farm of Tunghai University, Taichung, Taiwan. The COD concentration
biorefinery framework to produce biodiesel and other high-value of this dairy wastewater fluctuated during different sampling times. The
products. Furthermore, utilizing algae for wastewater treatment could dairy wastewater had an initial pH of 6.8–7.5, and 100–130 mg/L sus­
minimize the use of chemicals, thus lowering the overall treatment cost pended solid (SS). The moderate-COD dairy wastewater contained
and enhancing the cost competitiveness of the algal biorefinery frame­ 1000–2500 mg/L COD, 20–150 mg/L TN, and 7.5–18.5 mg/L TP, while
work (Razzak et al., 2017). Containing a high level of nutrients and the high-COD dairy wastewater contained 4000 mg/L COD, 22 mg/L TN
organic matters, digestate can be considered a desirable feedstock for and 11.6 mg/L TP. The dairy wastewater was first filtered using 90 μm
algal cultivation (Xia et al., 2020). Nevertheless, high COD content and cellulose membranes to remove the SS and stored at − 20 ◦ C until uti­
extremely high turbidity of digestate might inhibit microalgae growth lization. As this study focused solely on the nutrient removal capability
for direct utilization (Tan et al., 2020; Wang et al., 2015). of microalgae, the dairy wastewater was then sterilized to eliminate the
Literature reported that the algal growth improved with the dilution possibility of nutrient assimilation by indigenous bacteria present in the
of various real wastewaters to overcome substrate inhibition and light wastewater.
limitation (Choi et al., 2018; Kobayashi et al., 2013). A 1:1 dilution ratio
of digested anaerobic effluent was effective for removing total nitrogen
up to 96% by extremophilic green microalgae (Pang et al., 2020).

2
A. Kusmayadi et al. Chemosphere 291 (2022) 133057

Fig. 1. (a) The schematic description of the process for the direct microalgae-based treatment of moderate-COD raw dairy wastewater (b) The schematic description
of the process for the integrated anaerobic digestion and microalgae cultivation for the treatment of high-COD raw dairy wastewater.

2.3. Experimental design achieve anaerobiosis. The anaerobic digester was incubated at 55 ± 1 ◦ C
with constant agitation of 120 rpm for 10 days (Chi-Yu Huang, personal
2.3.1. Direct treatment of moderate-COD raw dairy wastewater by communication). A homogenized mixed liquor of the anaerobic digester
microalgae cultivation was taken periodically to analyze COD, TP, and TN.
Microalgae were cultivated in a photobioreactor of 1 L working The anaerobic digestate was used as the medium for algal cultivation
volume with various dilution ratios of moderate-COD raw dairy waste­ with similar operating conditions as Section 2.3.1. The influence of
water (0%, 25%, 50%, 75%, and 100% dairy wastewater diluted with different inoculum sizes (1, 1.5, and 2 g/L) was studied. During algal
BG-11 medium) at 0.3 g/L inoculum size for 12 days. For the influence of cultivation, the sampling was performed at the desired time interval to
different inoculum sizes (0.1, 0,5 0.75, 1, and 1.5 g/L), the microalgae analyze the concentrations of algal biomass, TP, TN, and COD. At the
were cultivated for 10 days instead due to a faster growth rate at opti­ end of cultivation, the algal cell was examined for lipid and protein
mized dilution ratios. The microalgal cultivation conditions used in this content. Each experiment was carried out in two replicates. The sche­
study were similar to those reported previously (Chen et al., 2020; Kuo matic illustration of the integrated anaerobic digestion and microalgae
et al., 2015). The photobioreactor was continuously illuminated at a cultivation is shown in Fig. 1b.
light intensity of 140 μmol m− 2 s− 1. The operation was set up at 25 ±
1 ◦ C with a continuous supply of 0.2 vvm 2% CO2 and 200 rpm agitation. 2.4. Determination of algal growth and biomass concentration
During algal cultivation, the sampling was performed at the desired time
interval to analyze the concentrations of algal biomass, TP, TN, and The concentration of algal biomass was routinely analyzed by
COD. At the end of cultivation, the algal cell was examined for lipid measuring its optical density (OD) at 685 nm using a UV/Vis spectro­
content. All experiments were conducted in duplicates, and results are photometer (model U-5100, Hitachi, Tokyo, Japan) as previously
shown as means ± standard deviations. Fig. 1a depicts the schematic described (Chen et al., 2018). The biomass concentration (g/L) was
process of microalgae production with moderate-COD raw dairy obtained using the following equation:
wastewater.
Biomass concentration (g/L) = 0.4112 × OD685 R2 = 0.9988 (1)
2.3.2. Integrated anaerobic digestion and microalgae cultivation operation The overall biomass productivity could be determined using the
for the treatment of high-COD raw dairy wastewater following equation:
High-COD raw dairy wastewater was anaerobically digested to
remove the organic matter, and nutrients compound partially. The batch Biomass productivity (g/L/d) = (Wt – Wo)/Δt (2)
experiment was conducted using a 2 L round-bottomed flask with 1 L
Where Wo and Wt are the initial and final biomass concentration, and Δt
working volume. Ten grams of cow dung collected from Tunghai Uni­
is the cultivation period in days.
versity’s dairy farm was used as seed for the anaerobic digester. The
reactor was then flushed with nitrogen gas for 10 min and sealed to

3
A. Kusmayadi et al. Chemosphere 291 (2022) 133057

Fig. 2. Effect of the dairy wastewater dilution ratios on biomass and bio-compounds production and wastewater treatment performance. (a) Biomass production. (b)
Biomass productivity. (c) Lipid content. (d). Nutrients removal.

2.5. Determination of COD, TP, and TN concentrations in culture medium wastewater with BG-11 medium as control, 0%) on the performance of
direct treatment of moderate-COD raw dairy wastewater by microalgae.
For COD, TP, and TN analysis, centrifugation of sample was per­ As shown in Fig. 2a and b, C. sorokiniana SU-1 was able to grow in 100%
formed at 6000 rpm for 5 min. The supernatant was filtered through a undiluted moderate-COD dairy wastewater with biomass concentration
0.45 μm cellulose membrane to eliminate residual suspended solids and and productivity of 2.10 ± 0.14 g/L and 0.16 ± 0.007 g/L/d, though
algal cells. The filtrate was then determined for COD, TN, and TP con­ lower than other dilution ratios. This indicates that a high COD level of
centrations. COD was analyzed according to APHA standard methods 100% low-strength dairy wastewater still exhibits substrate inhibition
(Carranzo, 2012). TN was analyzed by using Nash-reagent spectropho­ and interferes with light penetration. The dilution ratio at 50% of dairy
tometric method (Chen et al., 2020). TP was determined by using wastewater produced the best results in terms of biomass production
molybdenum–antimony anti-spectrophotometric method (Olsen, 1954). (3.2 ± 0.1 g/L) and productivity (0.25 ± 0.01 g/L/d), even when
compared with the control. The abundant nutrients in BG-11 medium
were argued to be unfavorable for C. sorokiniana growth (Chen et al.,
2.6. Determination of lipids and proteins content
2020). Chen et al. reported that C. sorokiniana AK-1 demonstrated good
adaptability in piggery wastewater and generated maximum biomass
After harvesting by centrifugation at 9000 rpm for 10 min at room
production of 5.45 g/L when grown in 50% diluted high strength
temperature, the algal cells were washed and lyophilized for further
wastewater (Chen et al., 2020). In this study, based on biomass pro­
analysis. Lipid content was determined as fatty acid methyl esters
duction, a 50% dilution ratio of dairy wastewater was the optimal
(FAMEs) by the direct transesterification method (Chen et al., 2011).
cultivation medium for the growth of C. sorokiniana SU-1.
Quantification of the resulting FAMEs was determined by using gas
The algal cells were harvested after 12 days of cultivation to analyze
chromatography (Shimadzu GC-2014, Kyoto, Japan) with a flame
the lipid content. Fig. 2c showed that the algal cultivation utilizing
ionization detector (FID) (Chen et al., 2018). To determine protein
100% undiluted dairy wastewater obtained the highest lipid content of
content, the algae was first analyzed by the elemental analyzer (Ele­
22.0 ± 1.1%, followed by 75%, 0%, 50%, and 25% of dairy wastewater
mentar Vario EL III) for nitrogen content, and the microalgal crude
dilution ratios with lipid content of 18.5 ± 0.28%, 17.2 ± 0.77%, 15.0 ±
protein content was calculated based on the correlation given by the
0.77%, and 12.6 ± 1.4%, respectively. The results indicated that the use
literature (i.e. protein concentration = nitrogen content x 6.25) (Becker,
of dairy wastewater for C. sorokiniana SU-1 cultivation increased the
1994; Fuentes et al., 2000).
production of microalgal lipids. In natural environments, most wild
microalgae struggle to achieve high lipids content and productivity.
3. Results and discussion
Moreover, lipid-rich microalgae are usually correlated with low growth
rates (Yang et al., 2015). Nitrogen and phosphate concentrations are the
3.1. Direct treatment of moderate-COD raw dairy wastewater with
primary macronutrients for lipid production in the algal cells. As 100%
microalgae cultivation - effect of dilution ratios on wastewater treatment
dairy wastewater achieved the highest total nitrogen removal efficiency,
and algal biomass production
it was suggested that the microalgae had nitrogen deficiency during the
exponential growth phase. A previous study reported that the shortfall of
The effect of dilution ratios (25%, 50%, 75%, and 100% of

4
A. Kusmayadi et al. Chemosphere 291 (2022) 133057

Table 1
Effect of dilution ratios of wastewater on COD and nutrients removal by C. sorokiniana SU-1.
Dilution ratio of COD TP TN
wastewater (% w/w)
Initial Final Removal Initial Final Removal Initial Final Removal
(mg/L) (mg/L) Efficiency (%) (mg/L) (mg/L) Efficiency (%) (mg/L) (mg/L) Efficiency (%)

0 – – – 5.2 ± 0.2 0.49 ± 90.6 ± 4% 292.6 ± 98 ± 0.57 66.5 ± 2%

0.02 1.76
25 600 ± 7.1 132 ± 78 ± 6% 4.62 ± 0.52 ± 88.7 ± 5% 232.4 ± 68.2 ± 70.63 ± 4%
16.7 0.02 0.15 8.7 1.4
50 645 ± 3.5 85 ± 23.9 86.8 ± 6% 6.5 ± 0.3 0.35 ± 94.6 ± 3% 208.9 ± 40.33 ± 1 80.7 ± 1%
0.12 0.8
75 715 ± 197 ± 72.4 ± 9% 7.6 ± 0.5 0.7 ± 90.8 ± 4% 128.6 ± 37.6 ± 70.7 ± 1%
10.6 28.4 0.13 3.4 0.28
100 1350 ± 474 ± 64.9 ± 11% 12.1 ± 0.2 0.87 ± 92.8 ± 2% 119.6 ± 10.8 ± 91 ± 2%
35.5 77.7 0.21 7.5 2.39

Fig. 3. Effect of algal inoculum sizes on biomass and lipids production and wastewater treatment performance. (a) Biomass production. (b) Biomass productivity. (c)
Lipid content. (d). Nutrients removal.

nitrogen could divert the metabolic flux produced from photosynthesis includes direct NH4+ utilization by microalgae and NH3 stripping, which
to lipid production in Neochloris oleoabundans (Li et al., 2008). More­ usually occur at high pH value around or above 9 (Li et al., 2011). As the
over, the accumulation of lipids increases following the duration of ni­ final pH value of the culture medium was 9.1, thus, NH3 stripping may
trogen starvation (Widjaja et al., 2009). contribute to the removal of NH4+ in our study other than microalgal
The effect of dairy wastewater dilution ratios on COD and nutrients absorption. On the other hand, the mechanisms involved in phosphorous
removal by C. sorokiniana SU-1 is shown in Fig. 2d, and the data are removal include autotrophic assimilation, precipitation, and sedimen­
tabulated in Table 1. The results demonstrated that the 50% dilution tation (Cho et al., 2011). The precipitation of phosphorus normally takes
ratio of dairy wastewater provided the best results in terms of pollutant place between pH 8.9 and 9.5. Therefore, a high pH value of 9.1 in this
removal, with COD, TP, and TN removal of 86.8 ± 6%, 94.6 ± 3.0%, and study may facilitate the precipitation of phosphorus with cations such as
80.7 ± 1.0%, respectively. Literature showed that C. vulgaris have Ca2+ and Mg2+.
similar COD removal efficiency of 88% in the treatment of piggery
waste, though with a much lower initial COD concentration, 250 mg/L
(Travieso et al., 2006). Among various forms of inorganic nitrogen that 3.2. Direct treatment of moderate-COD raw dairy wastewater with
algae can assimilate (e.g., ammonium (NH4+), nitrate (NO3− ), and ni­ microalgae cultivation - effect of inoculum sizes on wastewater treatment
trite (NO2− )), NH4+ is the most preferred because it requires less energy and algal biomass production
to assimilate (Ruiz et al., 2011). The mechanisms of NH4+ removal
To study the influence of inoculum size on the performance of dairy

5
A. Kusmayadi et al. Chemosphere 291 (2022) 133057

Table 2
Effect of inoculum sizes on COD and nutrients removal by C. sorokiniana SU-1 with dairy wastewater of 50% dilution ratio.
Inoculum sizes COD TP TN
(g/L)
Initial (mg/ Final (mg/ Removal Efficiency Initial (mg/ Final (mg/ Removal Efficiency Initial (mg/ Final (mg/ Removal Efficiency
L) L) (%) L) L) (%) L) L) (%)

0.1 645 ± 3.5 240 ± 62.8 ± 8 65 ± 0.01 29 ± 1.7 55.4 ± 7 8 ± 0.7 2.3 ± 0.01 71.2 ± 6
20.3
0.5 645 ± 7 210 ± 9.7 67.4 ± 3 65 ± 0.7 24 ± 0.17 63.1 ± 2 8 ± 0.35 2 ± 0.01 75 ± 3
0.75 645 ± 10.6 183 ± 71.6 ± 8 65 ± 1.4 18 ± 3 72.3 ± 8 8 ± 0.21 1.6 ± 0.07 80 ± 3
27.8
1 645 ± 3.5 100 ± 3.9 84.5 ± 1 65 ± 0.7 13 ± 0.16 80 ± 1 8 ± 14 1.1 ± 0.06 86.2 ± 1
1.5 645 ± 3.5 75 ± 15.4 88.4 ± 4 65 ± 2.1 10 ± 1 84.6 ± 2 8 ± 0.21 0.8 ± 0.02 90 ± 1
2 645 ± 6.3 70 ± 35.9 89.1 ± 9 65 ± 2.8 7 ± 2.7 89.2 ± 7 8 ± 0.21 0.5 ± 0.18 93.7 ± 4

wastewater phycoremediation, microalgae cultivation with inoculum physiological activity affected by the initial algal biomass density.
sizes of 0.1, 0.5, 0.75, 1, 1.5, and 2 g/L at 50% dilution ratio of waste­ Inoculum size of 2 g/L was more successful in removing nutrients from
water were carried out in this study. As shown in Fig. 3a, the algal dairy wastewater. These findings can be explained by the fact that
biomass production increased with the increase of inoculum sizes. The appropriate inoculum size resulted in higher microalgal biomass,
best performance was achieved with a 2 g/L inoculum size where algal meaning that the microalgal cells remove more nutrients.
biomass concentration reached around 3.75 ± 0.15 g/L. Consistent with
the results of Lee and colleagues, increasing the inoculum size has a 3.3. Anaerobic digestion (AD) of high-COD raw dairy wastewater
positive influence on algae productivity (Lee et al., 2021). Large inoc­
ulum sizes provide a better chance of surviving for microalgae in Raw dairy wastewater sometimes contains a very high concentration
wastewater’s harsh and hazardous environment (Bohutskyi et al., 2016). of organic compounds (e.g., COD = 4000 mg/L) and nutrients, making it
As depicted in Fig. 3b, the biomass productivity obtained by the inoc­ unsuitable for microalgae cultivation. Moreover, the wastewater may
ulum size of 1 g/L was higher than the biomass productivity of 2 g/L also contain a significant amount of other microbes (such as other algae,
initial inoculum size. The results indicated that the biomass growth bacteria, fungi, etc.) that compete with or are antagonistic for the
performance decreased when the inoculum sizes rise beyond a threshold growth of microalgal cells (Chen et al., 2020, 2021). To overcome these
level. issues, this study proposed the combination of the AD process and algae
This study also investigated the impact of inoculum sizes on lipid cultivation to treat raw dairy wastewater with a high COD level.
content. Fig. 3c demonstrated that the highest lipid content of 12.0 ± Anaerobic digestion of raw dairy wastewater was conducted first to
0.7% was obtained with an inoculum size of 2 g/L. The inoculum size of significantly lower the concentration of organic compounds in the dairy
0.1 g/L gave the lowest lipid content of 7.8 ± 1.2%. These results sug­ wastewater. The performance of the integrated process is summarized in
gested that a larger inoculum size could lead to higher lipid content. The Table 3. The pH of the dairy wastewater increased from 6.8 to approx­
result is similar to Lu et al. which stated that inoculum size significantly imately 7.6 after the AD process. AD effectively reduced the concen­
influenced cell growth and lipid production (Lu et al., 2012). The high tration of COD from 4000 ± 71 mg/L to 978 ± 20 mg/L and TP from
inoculum size did change the metabolism of microalgae into lipid stor­ 11.6 ± 1.0 mg/L to 8.60 ± 0.42 mg/L. These would be advantageous for
age (Bohutskyi et al., 2016). However, further raising the inoculum size the digestate to be further utilized for algal cultivation since the inhib­
after a certain point did not substantially impact lipid content (Bohut­ itory effects of high COD and TP levels on microalgal growth were
skyi et al., 2016). According to Lu et al., high-density microalgal cultures reduced (Yang et al., 2015). In contrast, TN concentration rises from
derived from high inoculum size may face the issue of photo-limitation, 22.0 ± 1.4 mg/L to 51.0 ± 2.4 mg/L due to the transformation of organic
accumulating less energy storage products, such as lipids (Lu et al., N to NH4 by anaerobic bacteria (Crofcheck and Crocker, 2016). The
2012). Therefore, biomass and lipid production of microalga can be increase of TN concentration in anaerobic digestate can be an effective
effectively enhanced with appropriate inoculum size. nitrogen source for microalgae cultivation.
In terms of nutrient removal, larger inoculum sizes usually gave a
higher nutrient removal efficiency (Fig. 3d). For inoculum size of 2 g/L, 3.4. Effect of algal inoculum sizes on wastewater treatment and biomass
COD, TP, and TN removal efficiencies by C. sorokiniana SU-1 were 89.1 production for integrated anaerobic digestion of high-COD dairy
± 9.0%, 89.2 ± 7.0%, and 93.7 ± 4.0%, respectively (Table 2). A small wastewater and microalgae cultivation
inoculum size of 0.1 g/L resulted in the lowest removal efficiency of
62.8 ± 8.0% COD, 55.4 ± 7.0% TP and 71.2 ± 6.0% TN. These results In this section, C. sorokiniana SU-1 was cultured with an anaerobic
demonstrated that the wastewater-borne COD/nutrients removal effi­ digestate of dairy wastewater at different initial inoculum sizes of 1, 1.5,
ciency of C. sorokiniana SU-1 was related to the growth of algal cells and and 2 g/L. Fig. 4a showed that the biomass concentration of

Table 3
COD and nutrients removal of raw dairy wastewater by integrating anaerobic digestion and algal cultivation.
Inoculum COD TP TN
sizes (g/L)
Initial After After algal Removal Initial After After algal Removal Initial After After algal Removal
(mg/L) AD cultivation efficiency (mg/L) AD cultivation efficiency (mg/L) AD cultivation efficiency
(mg/ (mg/L) (%) (mg/ (mg/L) (%) (mg/ (mg/L) (%)
L) L) L)

1 4000 978 ± 130 ± 9.0 96.8 ± 4 11.6 ± 8.6 ± 4.5 ± 0.9 87.6 ± 6 22 ± 51 ± 0.4 ± 0.1 93 ± 4
± 70.7 19.7 1.0 0.42 1.4 2.47
1.5 4000 978 ± 100 ± 7.1 97.5 ± 6 11.6 ± 8.6 ± 4.1 ± 0.3 91.8 ± 4 22 ± 51 ± 0.35 ± 0.12 93.2 ± 3
± 77.7 20.1 1.2 0.32 0.7 3.1
2 4000 978 ± 80 ± 2.7 98.0 ± 2 11.6 ± 8.6 ± 3.9 ± 0.2 93.8 ± 2 22 ± 51 ± 0.3 ± 0.15 93.5 ± 5
± 84.8 18.2 0.49 0.12 1.06 1.7

6
A. Kusmayadi et al. Chemosphere 291 (2022) 133057

Fig. 4. Effect of algal inoculum sizes on the production of biomass, lipids, and proteins from anaerobic digestate as well as wastewater treatment performance. (a)
Biomass production. (b) Biomass productivity. (c) Lipid content. (d) Protein content. (e). Nutrients removal.

C. sorokiniana SU-1 increased as the initial inoculum sizes increased, auto-inhibitory aggregation are two possible explanations (Chen et al.,
reaching the highest biomass production of 4.25 ± 0.10 g/L at 2 g/L 2016). Light-limitation may occur due to higher inoculum sizes causing
inoculum size, though the biomass productivity showed a reverse trend. mutual shading which hinders light penetration into the liquid medium
The biomass productivity with 1 g/L inoculum size was higher than that (Wang and Han, 2013). Furthermore, excessive inoculum sizes could
obtained from using an inoculum size of 2 g/L, as illustrated in Fig. 4b. increase the viscosity of the culture medium and accelerate cell growth
According to the literature, the growth of C. vulgaris slowed down when in the initial growth stage, which then causes a lack of dissolved oxygen
the inoculum size increased above a certain level due to limited space for or nutrients and affects metabolite synthesis. On the other hand, low
cell growth with a high inoculum size (Chen et al., 2016). The threshold inoculum sizes could extend the growth period and increase the risk of
of inoculum size differs for different microalga species due to the microbial contamination to form mycelia and aggregate (Heidari et al.,
biochemical characteristics of each species. Self-shading and 2016). As a result, appropriate initial inoculum size is important way to

Table 4
Comparison of biomass production, COD/nutrient removal, lipids and proteins contents in wastewater treatment using different microalgae strains.
Wastewater Microalgae strain Biomass Lipid Protein COD TP removal TN removal References
concentration (g/L) content (%) content (%) removal (%) (%) (%)

Domestic Chlorella sp. + Scenedesmus sp. 1.78 34.83 – 83 95 94 Silambarasan et al.

wastewater (2021)
Municipal Chlorella vulgaris 1.28 – – 75 53 70 Amini et al. (2020)
wastewater
Urban Pseudochlorella pringsheimii — 1.27 21.5 – 83.2 97.2 95.2 Kumar et al.
wastewater Ind-Jiht-1 (2021)
Molasses Scenedesmus sp. Z-4 3.3 28.9 87.2 90.5 88.6 (C. Ma et al.,
wastewater 2017)
Dairy manure Chlorella sp. 1.71 – – 38.4 74.5 82.5 Wang et al.
(2010a)
Livestock waste C. vulgaris, Scenedesmus obliquus, – – – 63.1 63.2 76 Zhao et al. (2015)
and Neochloris oleoabundans
Dairy manure Chlorella vulgaris – – – 55.4 89.2 93 Wang et al.
(2010b)
Dairy Mixed microalgae 1.4 22 15 90 73 65.5 Hemalatha et al.
wastewater (2019)
Dairy Tetraselmis suecica 0.47 – – 40.16 42.18 44.92 Daneshvar et al.
wastewater (2018)
Dairy C. sorokiniana SU-1 3.75 12 – 89.1 93.7 89.2 This study
wastewater
Dairy C. sorokiniana SU-1 4.25 12 18 98 93.8 93.5 This study
wastewatera
a
Integration of anaerobic digestion with microalgae cultivation.

7
A. Kusmayadi et al. Chemosphere 291 (2022) 133057

ensure proper microalgae growth and metabolite production. cultivation strategy resulted in a high biomass concentration of 4.25 g/L
One of the most important algal metabolites is lipid as biomass with a coupled with the significant increase in COD, TP, and TN removal of
high lipid concentration is a promising feedstock for producing bio­ more than 93%. The harvested microalgal biomass contained 12% lipids
diesel. As shown in Fig. 4c, microalgae cultivation with 2 g/L initial and 18% proteins, suggesting a higher potential as an efficient solution
inoculum size resulted in the highest lipid content of 12.5 ± 2.2%. On for simultaneous wastewater phycoremediation and biomass yield with
the other hand, a lower inoculum size of 1 g/L gave lower lipid content valuable cellular components. The proposed microalgal phycor­
of 9 ± 1%. Hallenbeck et al. mentioned that the inoculum size, light emediation approaches demonstrated promising potential for a low-cost
intensity, and CO2 levels are all critical factors influencing the produc­ microalgal-based integrated algal biorefinery to treat wastewater and
tion of lipids in microalgae. An increased level of those parameters resource recovery from the wastewater for valuable bioproducts for­
seems to be favorable for lipids accumulation in microalgae (Hallenbeck mation. As large-scale wastewater treatment with microalgae is usually
et al., 2015). Moreover, enhancing the inoculum size by up to 4.16% led carried out outdoors, unsterilized wastewater or digestate will be the
to an increase in lipids and docosahexaenoic acid (DHA) accumulation sole source of nutrients for microorganisms. Thus, further investigation
in Thraustochytrium sp. (Wu et al., 2018). The results of this study sug­ is needed for integrated anaerobic digestion and microalgae cultivation
gest that a higher initial biomass concentration is more advantageous for for treating dairy wastewater and biochemicals production with un­
the accumulation of lipid. sterilized wastewater.
Another crucial component of microalgae biomass is its protein
content. The high protein content of microalgae makes them an excel­ Credit author statement
lent source for human food and animal feed (Kusmayadi et al., 2021).
With inoculum sizes of 1, 1.5, and 2 g/L, C. sorokiniana SU-1 culture Adi Kusmayadi: Investigation, Writing – original draft; Po-Han Lu:
achieved a high protein content of 23.0 ± 2.2%, 21.0 ± 1.0%, and 18.0 Investigation, Data curation; Chi-Yu Huang: Validation, Methodology;
± 2.2% (Fig. 4d). The protein concentration in the present study was Yoong Kit Leong: Writing- Reviewing and Editing; Hong-Wei Yen:
greater than the protein content of mixed microalgae biomass (15%) Funding acquisition, Resources; Jo-Shu Chang: Supervision, Concep­
cultivated in dairy wastewater, as reported by Huo et al. (2012). The tualization, Writing- Reviewing and Editing.
protein contents in this study are similar in all variations of inoculum
sizes, which may be related to the similar nitrogen removal efficiency.
Also, the low protein content of microalgae might be due to the con­ Declaration of competing interest
version of protein to other biochemical products. For example, nitrogen
depletion facilitates the conversion of pigments or proteins to lipids or The authors declare that they have no known competing financial
carbohydrates, as shown by the study of Hemalatha et al. (Hemalatha interests or personal relationships that could have appeared to influence
et al., 2019). The cell protein of C. sorokiniana SU-1 through the algal the work reported in this paper.
treatment of dairy wastewater, in particular, has the potential to be used
as a feed supplement to achieve the future goal of the circular economy Acknowledgments
of dairy farms.
The microalgae can effectively remove the COD and nutrients in the The authors thankfully acknowledge the financial support received
anaerobic digestate of dairy wastewater. Fig. 4e depicted the removal from the Ministry of Science and Technology (MOST), Taiwan, ROC
efficiency of COD and nutrient at different inoculum sizes. At inoculum under grant number 110-3116-F-006 -003, 110-2221-E-029 -004 -MY3,
sizes of 2 g/L, more than 93.0 ± 2.0% removal for COD, TP, and TN was 110-2621-M-029 -001, and 109-2622-E-110 -011.
achieved. The inoculum size of 1.5 g/L achieved 97.5 ± 6.0% COD, 91.8
± 4.0% TP, and 93.2 ± 3.0% TN removal. On the other hand, up to 96.8 References
± 4.0% COD, 87.6 ± 6.0% TP, and 93.0 ± 4.0% TN were removed from
anaerobic digestate of dairy wastewater with an inoculum size of 1 g/L. Amini, E., Babaei, A., Mehrnia, M.R., Shayegan, J., Safdari, M.-S., 2020. Municipal
wastewater treatment by semi-continuous and membrane algal-bacterial photo-
The results indicated that high inoculum size could also boost nutrient
bioreactors. J. Water Process Eng. 36, 101274–101282.
removal. These results can be related to the fact that large inoculum Aravantinou, A.F., Theodorakopoulos, M.A., Manariotis, I.D., 2013. Selection of
cultures generated increased microalgal biomass, implying that more microalgae for wastewater treatment and potential lipids production. Bioresour.
nutrients were assimilated by the microalgal cells (Tam and Wong, Technol. 147, 130–134. https://doi.org/10.1016/j.biortech.2013.08.024.
Becker, E.W., 1994. Microalgae: Biotechnology and Microbiology. Cambridge University
1989). The integration of anaerobic digestion with microalgae cultiva­ Press.
tion proved to be more effective for the remediation of high-COD raw Bohutskyi, P., Kligerman, D.C., Byers, N., 2016. Effects of inoculum size, light intensity,
dairy wastewater, achieving a high recovery of organic compounds and and dose of anaerobic digestion centrate on growth and productivity of Chlorella and
Scenedesmus microalgae and their poly-culture in primary and secondary
nutrients. Table 4 compared the biomass production, COD/nutrients wastewater. Algal Res 19, 278–290.
removal, lipids, and proteins content in wastewater treatment obtained Carranzo, I.V., 2012. Standard Methods for examination of water and wastewater. In:
in this study with those reported by literature. Various pure or mixed Anales De Hidrología Médica. Universidad Complutense de Madrid, p. 185.
Chen, C.-Y., Chang, Y.-H., Chang, H.-Y., 2016. Outdoor cultivation of Chlorella vulgaris
microalgae were employed for wastewater treatment with biochemical FSP-E in vertical tubular-type photobioreactors for microalgal protein production.
production. In comparison with other microalgae reported in the liter­ Algal Res 13, 264–270.
ature, the microalgal strain (C. Sorokiniana SU-1) used in the present Chen, C.-Y., Nagarajan, D., Chang, J.-S., 2020. Cultivating Chlorella sorokiniana AK-1
with swine wastewater for simultaneous wastewater treatment and algal biomass
work shows a lower lipid content but a higher protein content, along
production. Bioresour. Technol. 302, 1–10.
with higher pollutants (COD, TN, TP) removal efficiency. Chen, C.-Y., Nagarajan, D., Chang, J.-S., 2018. A highly efficient two-stage cultivation
strategy for lutein production using heterotrophic culture of Chlorella sorokiniana
MB-1-M12. Bioresour. Technol. 253, 141–147.
4. Conclusions Chen, C.-Y., Yeh, K.-L., Aisyah, R., Lee, D.-J., Chang, J.-S., 2011. Cultivation,
photobioreactor design and harvesting of microalgae for biodiesel production: a
The study proposed a flexible approach to treat raw dairy wastewater critical review. Bioresour. Technol. 102, 71–81.
Cho, S., Luong, T.T., Lee, D., Oh, Y.-K., Lee, T., 2011. Reuse of effluent water from a
of different COD levels. For moderate-COD raw dairy wastewater,
municipal wastewater treatment plant in microalgae cultivation for biofuel
phycoremediation by C. sorokiniana SU-1 under conditions of 50% production. Bioresour. Technol. 102, 8639–8645.
dilution ratio and 2 g/L initial inoculum size achieved COD, TP, and TN Choi, Y.-K., Jang, H.M., Kan, E., 2018. Microalgal biomass and lipid production on dairy
removal efficiencies of 89.1%, 93.7%, and 89.2% with 3.75 g/L biomass effluent using a novel microalga, Chlorella sp. isolated from dairy wastewater.
Biotechnol. Bioproc. Eng. 23, 333–340.
production with 12% lipid content in the algal biomass. For high-COD Crofcheck, C., Crocker, M., 2016. Application of recycled media and algae-based
raw dairy wastewater, integrated anaerobic digestion and microalgae anaerobic digestate in Scenedesmus cultivation. J. Renew. Sustain. Energy 8, 1–15.

8
 Bioresource Technology 332 (2021) 125108

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Efficient degradation of polyacrylate containing wastewater by combined

anaerobic–aerobic fluidized bed bioreactors
Qi Song a, Zheng Sun b, Yong Chang b, Weifeng Zhang c, Yingzhi Lv b, Jiayi Wang c,
Fenghao Sun b, Yanxue Ma d, Yuling Li d, Fengbo Wang a, Xiaoguang Chen a, *
a
College of Environmental Science and Engineering, State Environmental Protection Engineering Center for Pollution Treatment and Control in Textile Industry, Donghua
University, Shanghai 201620, China
b
Bashan Weaving Group Co., Ltd, Zibo 255000, China
c
China Filament Weaving Association, Beijing 100742, China
d
College of Textiles, Donghua University, Shanghai 201620, China

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Biological method rarely used to treat

polyacrylate containing wastewater
(PCW).
• Two special bioreactors were applicated
successfully in treating PCW.
• Co-substrate was used to reduce toxic
effect of PCW on microorganism.
• Monod equation was used to fit the cu­
mulative methane production curve.
• The diversity of microorganism was
detected during the treatment of PCW.

A R T I C L E I N F O A B S T R A C T

Keywords: Polyacrylate containing wastewater (PCW) is the typical sewage discharged by the textile industry. It has
Slurry wastewater extremely poor biodegradability, and chemical methods were used conventionally as the only way for treating
Polyacrylate containing wastewater PCW. This study is demonstrating a novel biological method. In batch experiment monod kinetics was applied to
Anaerobic treatment
the experimental data, which indicated that anaerobic treatment used for PCW is feasible. The pilot-scale
Biodegradability
experiment combined a Spiral Symmetry Stream Anaerobic Bioreactor (SSSAB) and an air-lift external circula­
Spiral symmetry stream anaerobic bioreactor
tion vortex enhancement nitrogen removal fluidized bed bioreactor (AFB). The COD and NH+ 4 -N removal reached
up to 95.2% and 96.6%, respectively, which were higher than the value obtained by other chemical methods.
High-throughput sequencing analysis indicated that the relative abundance of Proteobacteria, Firmicutes and
Bacteroidetes increased, which contribute to the degradation of PCW. Therefore, PCW can be degraded efficiently
by using a SSSAB-AFB technique and thus provides an alternative to the chemical methods.

* Corresponding author at: Room 4161, No.4 Academic Building, No. 2999, North Renmin Road, Songjiang District, Shanghai 201620, China.
E-mail address: [email protected] (X. Chen).

https://doi.org/10.1016/j.biortech.2021.125108
Received 10 March 2021; Received in revised form 25 March 2021; Accepted 27 March 2021
Available online 5 April 2021
0960-8524/© 2021 Published by Elsevier Ltd.
Q. Song et al. Bioresource Technology 332 (2021) 125108

1. Introduction enzyme and causing the bacteria proliferation supported by the biode­
gradable carbon resource (Friman et al., 2013). Glucose is known for a
Polyacrylate is widely used in sizing (Zhou et al., 2020), printing and good co-metabolic substrate for enhancing the degradation of the re­
finishing of the water jet weaving industry due to its good adhesion, fractory pollutants (Tian et al., 2020). In this study the enhancement of
aging resistance and fast production characteristics. Polyacrylate is glucose as co-metabolic substrate for PCW treatment was studied. While
formed by the polymerization of acrylic acids, and it has high molecular it is uneconomical to add glucose to wastewater, a substitute for glucose
weight and good thermal stability. The wide application of polyacrylate is need. Therefore, domestic sewage was introduced in a pilot-scale
makes the world more colorful, but it also caused an undesirable visible experiment and used as co-metabolism substrate since it is cost-free
pollution. Among the industries generating PCW, the textile industry is and nutritive (Guo et al., 2013).
prominent. It is traditional industry in China, with more than 36 thou­ The objective of this study was to provide a novel biological method
sand factories spread nationwide. According to the statistics China’s for the treatment of PCW to replace the chemical methods. The exper­
total filament manufacturing will account to 60 million tons by 2020, iment was divided into three steps. First, a batch experiment was carried
which accounts for more than half of the global production. Meanwhile, out to verify the feasibility of the anaerobic biological technique for
the sewage discharged by the textile industry will grow up to 2.05 billion treating PCW. Second, a lab-scale experiment using a SSSAB with
tons per year. PCW is one of the typical wastewaters of the sizing process glucose as co-substrates to improve the degradation. The performance of
in the textile industry. It contains a high concentration of organic the SSSAB is discussed. And third, a pilot-scale experiment using an AFB
compounds with extremely poor biodegradability. The conventional with the SSSAB for an improved treatment of PCW. Besides the chemical
treatment of PCW is by chemical methods, which have problems such as oxygen demand (COD), the high concentration of NH+ 4 -N was also
high cost and secondary pollution (Show et al., 2020). Therefore, it is removed. Additionally, the stability of the SSSAB and the AFB was
necessary to develop an efficient, economical and environmentally assessed by OLR shock experiment and statistical analysis.
friendly method for the treatment of PCW.
A combined anaerobic–aerobic biological treatment process has been 2. Materials and methods
regarded as the most logical concept for the treatment of refractory
pollutants (van der Zee & Villaverde, 2005). In this process high mo­ 2.1. Materials
lecular weight organic matter such as polyacrylate are anaerobically
biodegraded to low molecular weight organic matter, which could be 2.1.1. Inoculated sludge
further biodegraded aerobically (Liu et al., 2020). The step anaerobic For inoculation of the SSSAB, anaerobic granular sludge obtained
treatment is essential, because the refractory pollutants are recalcitrant from SSSAB in a soybean product processing plant in Jiangxi province,
to an aerobic bio-treatment (dos Santos et al., 2007). This holds for some China was used. The average particle size of the sludge was 2.25 mm, the
bacterial strains, for which refractory pollutants are recalcitrant under density was 1.17 g/cm3, and the VSS / SS value was 0.77.
aerobic but easily biodegradable under anaerobic conditions (Deng The sludge dosage accounted for about 35% of the effective volume
et al., 2008). of the SSSAB. The sludge used for the AFB was obtained from the
In order to provide a good and efficient anaerobic treatment, a high aeration tank of the municipal sewage treatment plant. The dosage
efficiency anaerobic reactor with own intellectual property rights, the accounted for 40% of the total volume of the AFB reactor.
Spiral Symmetry Stream Anaerobic Bioreactor (SSSAB) (Chen et al.,
2016) was developed. In the reaction zone of SSSAB three elliptic plates 2.1.2. Raw water
were spirally and symmetrically set. Then the reaction zone was divided The PCW originated from the sizing machine tank in the mixing
into three zones by the elliptic plates. Each zone was provided with a gas workshop of Zibo Bashan Weaving Co. Ltd. It contained a COD of
collection pipe to achieve respective biogas collection. The division was 810–3500 mg/L and a NH+ 4 -N of 19–92 mg/L. The co-substrate was used
based on the anaerobic three stage theory. The dominant bacteria for the in lab-scale and pilot-scale experiments (Tian et al., 2020). In lab-scale
different stages of the microbial anaerobic fermentation are enriched in experiment, glucose solution was applied as the co-substrate. In the
different zones of the reactor, which improves the reactors processing pilot-scale experiment glucose solution was used as the co-substrate at
efficiency. The superior operating performance of the SSSAB has been the beginning, and domestic sewage was introduced as the co-substrate
proven in previous studies (Chen et al., 2016, 2017; Dai et al., 2017, after a stable operation had been established. The domestic sewage
2016). Taking methanol-sodium acetate wastewater treatment for contained a COD of 244–311 mg/L and a NH+ 4 -N of 43–61 mg/L.
example, the super-high-rate OLR, volumetric removal rate (VRR) and
improved volumetric methane production (VMP) were 361.5 kg COD/ 2.2. Experiment
(m3⋅d), 268.3 kg COD/(m3⋅d) and 73.0 m3/(m3⋅d), respectively, which
have a leading position in literature. The reactor has been applied also 2.2.1. Batch experiment
for the treatment of other wasters, such as azo dye wastewater (Dai A high concentration of organic matter in the PCW can inhibit the
et al., 2018), soybean waste water and Chinese pharmaceutical waste­ activity of the anaerobic microorganism. In order to alleviate the
water (Chen et al., 2019). According to this study the treatment of PCW inhibitory effect, PCW was diluted to different concentrations (10, 20,
by SSSAB has been tested and a good performance was obtained. The air- 40, 60 and 80 times of COD) by tap water. The removal of the COD was
lift external circulation vortex enhancement nitrogen removal fluidized then measured.
bed bioreactor (AFB) (Hu et al., 2017) has also been developed, to use
for a post aerobic treatment. The AFB is divided into an aerobic space 2.2.2. Lab-scale experiment
and anoxic space. Thus, it can perform biological nitrification and A 5L lab-scale SSSAB, made of stainless steel, was used in lab-scale
denitrification on its own. According to the previous experiment, if AFB experiment. Influent was provided continuously to the SSSAB using a
is used to treat synthetic wastewater made of sucrose and ammonium peristaltic pump. The reactor was heated by hot water circulation and
chloride, the OLR and nitrogen loading rate (NLR) can reach 8.5 kg the temperature was kept at 38 ± 2 ℃ by a sponge layer. During the
COD/(m3⋅d) and 0.56 kg TN/(m3⋅d), respectively. This is obviously start-up period, glucose and polyacrylate were mixed to a certain COD of
higher than the values for a similar bioreactor (Bhuvanesh et al., 2013; around 820 mg/L. Then the OLR was gradually increased to make the
Feng et al., 2021; Omena et al., 2013). reactor run efficiently and stable, until the influent COD reached 3500
Co-metabolism is an effective approach to increase the degradation mg/L.
of refractory pollutants by anaerobic treatment. The improvement by co-
substrates addition can be attributed to the induction of oxidizing

2
Q. Song et al. Bioresource Technology 332 (2021) 125108

2.2.3. Pilot-scale experiment the colorimetric dish.

The SSSAB for the pilot-scale experiment, was also made of stainless Coefficient of variation (CV) and sensitivity (S) were used to assess
steel too. The reactor has a diameter of 0.8 m, a height of 3.8 m, an the operational stability of the AFB in the pilot-scale experiment. The CV
effective volume of 1 m3, and HRT of 41.7 h. The reactor is heated by hot was calculated according to formulas 2,3 and 4.
water circulation and the temperature 35 ± 3 ℃ by a layer of sponges.
1∑ n
The pilot-scale experiment additionally used the patented reactor AFB m= Xj (2)
coupled with SSSAB to treat the PCW. The AFB was used in the operation n j=1
mode of the SBR process: feeding for 0.75 h, aeration for 9 h, settling for √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
1 h, draining for 0.75 h and idling for 0.5 h. One cycle of the reactor √
√ 1 ∑ n

lasted 12 h and it operated at the room temperature (20-30℃). The ρ=√ (Xj − m)2 (3)
n − 1 j=1
apparatuses are shown in Fig. 1.
ρ
2.3. Analysis methods CV = (4)
m

The COD was determined according to the standard methods (APHA, where m is the average of sample parameters, n is the number of sam­
2005). The biochemical oxygen demand (BOD5) was determined by ples, j is the number of samples at different time, X is the sample pa­
iodimetry. The pH was measured by a PXSJ-216 electronic probe. VFAs rameters of the effluent, and ρ is the standard deviation of sample
were measured by gas chromatography (GC7890, Shanghai Aosong parameters.
Scientific Instrument Co., Ltd.). NH+
4 -N was measured by Nessler reagent The sensitivity (Se) of the reactor was calculated according to for­
colorimetry. mula (5).
Coenzyme F420 was determined by absorption spectrophotometry ⃒
⃒(Oi − Ob )/Ob ⃒
⃒
method (Reynolds & Colleran, 1987). The coenzyme F420 concentration Se = ⃒⃒ ⃒ (5)
(Ii − Ib )/Ib ⃒
was determined as follows:
Where Oi is the concentration of the effluent matrix at i d, Ob is the
(AS − A0 )f
Coenzyme F420 = (1) concentration of effluent at steady state condition, Ii is the OLR at i d,
εl
and Ib is the OLR at steady state condition.
where AS and A0 are the absorbancy of the sample and the blank; f is the
dilution factor; ε is the molar absorption coeffcient; and l the thickness of

Fig. 1. Batch, lab-scale and pilot-scale experiment setup and the diagram of device.

3
Q. Song et al. Bioresource Technology 332 (2021) 125108

2.4. Microbial community analysis

To identify how the PCW inhibits the activity of the microbial

community, high-throughput sequencing was used. Sludge samples
were collected from the inoculated and mature sludge for high-
throughput sequencing and metagenomic analysis. The bacterial and
archaeal 16S rRNA PCRs (polymerase chain reaction) were performed
using the primers 338F (ACTCCTACGGGAGGCAGCA)/806R (GGAC­
TACHVGGGTWTCTAAT) and 524F (TGYCAGCCGCCGCGGTAA)/958R
(YCCGGCGTTGAVTCCAATT) targeting the variable regions V3–V4 and
V4-V5, respectively.

3. Results and discussion

3.1. Batch experiment

3.1.1. Optimum influent concentration of COD

The biogas production relies strongly on the stability of the anaer­
obic microbial community growing in the bottle (Calusinska et al.,
2018). Fig. 2 depicts the methane production, COD removal and
methane production for different dilution ratios. As can be seen from
Fig. 2A, when the COD increased from 2040 mg/L to 16300 mg/L, the
methane production also increases. If the COD concentration of the PCW
is 178800 mg/L (a dilution ratio of 0), little methane is produced which
resulted from the anaerobic microorganism that was probably inhibited
by polyacrylate contained in PCW. Data in Fig. 2B indicated that with
the increase of dilution ratio, the methane production first increased
from 37.3 to 119.0 ml/g COD and later decrease to 107.8 ml/g COD.
Similar to the methane production rate (see Fig. 2B), with the increase of
dilution ratio the COD removal rate also increased from 7.8% to 37.9%
and then decreased to 24.4%. The maxima of methane production and
COD removal were obtained for a dilution ratio of around 40x. At this
value the methanogenic activity was best.

3.1.2. Reaction kinetics

The Monod equation was used to fit the cumulative methane pro­
duction curve values (Zhang et al., 2019). Since there is only poly­
acrylate slurry as the limiting factor for microbial growth in the matrix
of the reaction process, Eq. (6) was selected as the basic model.
Vmax S
V= (6)
Ks + S

where V is the specific growth rate of the microorganisms, Vmax the

maximum specific growth rate of microorganisms, KS is a constant, its
value is the matrix concentration, if V = Vmax/2, and S the matrix
concentration.
On the basis of Eq. (6), Eq. (7) can be derived.
1 KS 1 1 Fig. 2. Anaerobic biodegradation of PCW by batch experiment. A) Methane
= ⋅ + (7) production values. B) COD removal and methane production for different COD.
V Vmax S Vmax
C) Linear regression curve of reaction kinetics with methane production and
where V can be expressed by the specific methanogenic activity (Gon­ substrate degradation.
zalez-Estrella et al., 2017), the unit is ml CH4/(gVSS⋅d). If the 1/V
(specific matrix degradation rate) and 1/S (methanogenic activity) 0.106 d-1 and KS = 122 mg/L. It is obvious that the effect of the
under different COD values are analyzed, a fitting straight line is per­ degradation of PCW by an anaerobic process yield better results than
formed, as shown in Fig. 2C. From the data, the relevant parameters cited ones.
Vmax and KS of the Monod equation can be calculated.
According to the linear regression curve in Fig. 2C, the equation is y
3.2. Lab-scale experiment of SSSAB
= 0.20523x + 0.06524. The correlation coefficient is R2 = 0.96742,
Vmax = 15.33d-1 and KS = 3.15 mg/L. According to the characteristics of
3.2.1. Removal of COD
anaerobic degradation, if Vmax is large and KS is small, it means that the
According to the results of the batch experiment, the diluted PCW
effect of the degradation is great. For comparation the data obtained by
was used as influent for the SSSAB in a lab-scale experiment. Its dilution
Wang (Wang et al., 2014), for the treatment of coking wastewater using
ratio was gradually adjusted, until the COD reached the optimum con­
an anaerobic expanded granular sludge bed reactor are given: Vmax =
centration (COD 4000 mg/L). During the start-up period of the reactor,
0.576 d-1 and KS = 44.07 mg/L. Lin Shaoning (Lin et al., 1991) and
the glucose was used as a co-substrate and mixed with the PCW. The OLR
others achieved in the study of PVA slurry wastewater treatment Vmax =
of the reactor was gradually increased from 0.27 kg COD/m3∙d to 1.17

4
Q. Song et al. Bioresource Technology 332 (2021) 125108

kg COD/m3∙d with an HRT maintained at 72 h. removed in subsequent process (Yu et al., 2019a).
Data for the performance of the SSSAB in a lab-scale experiment are
shown in Fig. 3. Phase one (days 1–15) is the initial start-up period. The 3.3. Pilot-scale experiment
influent was mixed wastewater (glucose and PCW) with an average COD
of 853.6 mg/L, and the COD removal account of 48%. The probable 3.3.1. Performance of SSSAB
reason for the low efficiency is that the anaerobic sludge had just been
inoculated. It was not adapted to PCW, and some of the microorganism (1) COD removal
would be inhibit by the new external environment, and the sludge might
bring in oxygen during inoculation. In phase two (days 16–30) the Domestic sewage was used in the pilot-scale experiment as co-
dosages of PCW and glucose were increased causing the influent COD substrate mixed with PCW and diluted to a concentration of COD
increase to 1600 mg/L. During this period the reactor operated stably. around 1500 mg/L. The SSSAB was started by gradually increasing the
The COD removal reached 75% and VFA was not detected in the proportion of PCW in the influent.
effluent. In phase three (days 30–45) the proportion of PCW in the mixed The pilot-scale experiment was divided into four phases as shown in
influent was increased to 2/3 (PCW in the mixed wastewater of PCW and Fig. 4A. In Phase I (days 1–4) the wastewater (glucose solution without
co-substrate). Then the COD in the influent reached 2400 mg/L. At this PCW) was used as influent for adaptation of the anaerobic granular
time the COD removal efficiency accounted to above 65%. In phase four sludge to the reactor conditions to improve its activity. The COD
(days 45–61) no co-substrate, but the COD of the influent was kept at removal reached up to 85% in these four days. In phase II (days 4–24)
2400 mg/L with an increase of the PCW dose. In the following the COD the domestic sewage was introduced as co-substrate (75% domestic
removal efficiency rapidly decreased to 43.16%, while after the sewage mixed with 25% PCW). Then the COD removal dropped to
following days, it rebounded to around 50%. The drop of COD removal 46.3%. Probably the polyacrylate in the PCW exerted a toxic effect on
indicates the importance of co-substrate. Throughout phase four the pH microorganisms. During the next days the microorganisms gradually
in the reactor was stable at about 8.42 (range from 8.3 to 8.6). Also, the adapted to the influent. Thus, the COD removal increase. In phase III
sludge has stable properties indicating that the reactor operated stably. (days 25–64) the proportion of PCW in influent was increased to 50%.
Finally, the reactor entered the steady operation phase (days 61–78). After the reactor operated for one week. The average COD removal still
The COD increased to 3510 mg/L with average COD removal at this reached 75.25% and stabilized. In phase IV (days 65–100) while the
phase of 51.30%. This value is significantly improved compared to the proportion of PCW was increased further to 75%. This caused the COD in
37.89% obtained in the batch experiment. The data show that the effluent increase significantly, while the corresponding COD removal
addition of co-substrate makes the anaerobic granular sludge more dropped to 40.7%. After ten days of steady operation the average COD
active in the anaerobic treatment of PCW. concentration in the effluent was 921.6 mg/L with a corresponding
Fig. 3B depicts the changes of ammonia nitrogen in the steady state. average COD removal of 44.28%. This value is significantly lower than
The concentration of NH+ 4 -N has increased significantly from 266% to those obtained in lab-scale experiment (phase three in lab-scale exper­
426%. The reason may be that the PCW contains a large amount of iment, 65%). Accordingly, a conclusion can be drawn indicating that
organic nitrogen compounds, which will be converted to NH+ 4 -N in domestic wastewater can be used as co-substrate for treating PCW, but
anaerobic condition by ammonifiers. The good activity of the ammo­ its induction of oxidase and the support of bacterial proliferation were
nifiers also proves the biodegradability of PCW has been effectively not as effective as glucose solution.
improved. For this high NH+ 4 -N, the COD removal fluctuated in the range
of 40% − 50%. Thus, the SSSAB has shown a strong resistance to high (2) OLR shock experiment
NH+ 4 -N concentration. However, the increased NH4 -N needs to be
+

Fig. 3. Performance of the SSSAB treating PCW in a lab-scale experiment. A) Changes of COD. B) Growth of NH+
4 -N in the steady running phase.

5
Q. Song et al. Bioresource Technology 332 (2021) 125108

Fig. 4. Performance of the SSSAB treating PCW and characteristics of anaerobic granular sludge (AGS) in SSSAB in a pilot-scale experiment. A) Changes of COD. B)
SSSAB under shock of organic loading rate (OLR). C) Changes of B/C (biochemical oxygen demand/chemical oxygen demand). D) Particle size distribution of AGS E)
Coenzyme F420 concentration of AGS.

To assess the stability of the SSSAB operation in the steady operation The BOD/COD (B/C) values of the samples at different stages are
stage, the reactor was operated under shock OLRs. In Fig. 4B is shown, shown in Fig. 4C. The principle of co-metabolism was used with do­
the COD removal decreased from 41.1% to 36.8%, 34.2%, 32.1% mestic sewage as the co-substrate. The B/C values increased from 0.031
respectively, when the OLR of SSSAB increased from 0.91 to 1.15, 1.73, to 0.156. After the treatment in the SSSAB, the B/C values rose from
2.30 kg COD/m3∙d. It is in inevitable that the strong and rapid increase 0.156 to 0.287. The results indicate that the biodegradability of PCW
of the COD caused a decrease of the COD removal, but the impact and was improved gradually (Camacho-Munoz et al., 2020). The improve­
recovery time were all within acceptable ranges. ment of biodegradability of PCW provided a good condition for the
subsequent aerobic treatment.
(3) improvement of biodegradability
(4) anaerobic granular sludge characteristics (AGS) in SSSAB

6
Q. Song et al. Bioresource Technology 332 (2021) 125108

Fig. 4D shows the change of the particle size distribution of the seed which then converted to ammonia nitrogen. Also, the microorganisms in
and the adapted AGS. The mean diameter of the AGS increased from the SSSAB were active. The concentration of NH+ 4 -N in the influent of the
2.62 mm to 3.46 mm. The wet density and the average setting velocity AFB fluctuated in the range of 84 mg/L to 163 mg/L. However, the
changed from 1.13 g/ml and 72.47 m/h to 1.22 g/ml to 107.39 m/h, concentration of NH+ 4 -N remained at low level (around 1.8 mg/L) after
respectively. The changes of the AGS characteristics indicate that it was the treatment of AFB. This concentration meets the discharge standards
adapted and granulated in the SSSAB. The larger size and relatively high for the textile dyeing industry. Therefore, the AFB perfectly meets the
density of individual granules have better settleability. And the good demands for the removal of NH+ 4 -N.
settleability of the granules would simplify the separation of purified
effluent from the granular sludge(Liu & Tay, 2002; Tan et al., 2020). (3) Stability of the AFB operation
Coenzyme F420 as a low-potential electron carrier, participates in
catabolism and it is a coenzymes unique to methanogens (Reynolds & The values for CV and S of the were shown in Table 1. The CV of the
Colleran, 1987). Fig. 4E depicts the concentration of coenzyme F420 at AFB together with the COD and the NH+ 4 -N decreased from 32.79% and
each stage. The concentration of coenzyme F420 in seed and adapted 58.64% to 5.05% and 10.47%, respectively. This indicates that the
sludge were 0.152 μmol/g and 0.189 μmol/g, respectively. This increase stability of the AFB in steady-running phase was better than in the start-
shows that the number of methanogenic archaea had increased during up phase. Compared with the data obtained by Hu et al. (2017) for
the start-up process. Consequently, the methanogenic activity in the treating synthetic wastewater by AFB, CV = 11% and 23% (associate
sludge is significantly improved, also for the degradation of PCW. with COD and NH+ 4 -N, respectively), the stability of AFB in treating PCW
has obvious advantages. S (Sensitivity) refers to the degree of change in
3.3.2. Performance of the AFB response to the change of the substance concentration to be measured.
With the COD concentration and its removal was increased, S also
(1) Removal of COD increased a little, but still remained at a low level. The fluctuation of the
water quality within a certain range will not cause significant changes in
Fig. 5A depicts the COD removal by the AFB in the pilot-scale the effluent. This statistical analysis has shown that the AFB was oper­
experiment. The influent COD varies greatly during the stable opera­ ating steadily.
tion, but the effluent COD remains at 50–70 mg/L. Thus, the corre­
sponding COD removal accounts to around 85%. During the day 34th-
3.4. Inhibition of anaerobic microorganisms
37th, COD removal suddenly rose up to over 90%. The reason could
be that some new sludge had been added to the reactor on the 33rd day.
The microbial communities in the inoculated and the mature sludge
The adsorption of the anaerobic granular sludge in a short period of time
have been analyzed by 16S rRNA high-throughput sequencing. The re­
resulted in the “false appearance” of the decrease of the content the
sults are presented in Fig. 6. The rarefaction curves shown in Fig. 6A and
organic. However, in the following few days the effluent COD and COD
6B gradually flatten, which is indicating that the sequencing depth was
removal were not significantly different from those before. Therefore,
the performance of the AFB indicates that it has a good efficiency, if
coupled with a SSSAB for treating PCW. Table 1
Statistical analysis of the operational stability of the AFB.
(2) Removal of NH+
4 -N COD NH+
4 -N

CV (%) Se CV (%) Se
4 -N increased
The data in Fig. 5B show that the concentration of NH+
Start-up 32.79 0.08 58.64 3.25
from 54 mg/L to 116 mg/L after the treatment in the SSSAB. Obviously Steady-running 5.05 6.02 10.74 12.61
the ammonolysis reaction have resulted in a lot of organic nitrides,

Fig. 5. Performance of the AFB treating PCW during stable operation. A) Changes of the COD. B) Changes of the NH+
4 -N in combined process of the SSSAB with
the AFB.

7
Q. Song et al. Bioresource Technology 332 (2021) 125108

Fig. 6. Microbial community analysis of samples from the inoculated sludge and the mature sludge: A) Rarefaction curve for the Chao 1; B) Rarefaction curve on
shanon. C) Community composition of bacteria in inoculated sludge. D) Community composition of bacteria in mature sludge. E) Relative abundance at the genus
level of the two samples.

reasonable. Fig. 6A and B displays the results of the alpha diversity granules were identified using MiSeq analysis of the 16S rRNA sequence.
analysis, and the Chao and Shannon indices of the microbial community Fig. 6C and D outer shows the phylum level distributions of the micro­
in the mature sludge are 1795.45 and 7.80, respectively. They are lower organisms in the two samples. The main phyla in the inoculated and the
than those of the inoculated sludge (2013.77 and 7.95). This shows a mature sludge include Chloroflexi, Proteobacteria, Firmicutes, Bacter­
decrease in both the community richness and diversity after acclimati­ oidetes, Actinobacteria, Synergistetes, Acidobacteria. Those groups totally
zation, which indicates that the PCW significantly inhibits the account for 86.58% and 88.39% of the microbial population, respec­
microorganism. tively. Especially, the abundances of the Chloroflexi decreased by
Microbial communities of the inoculated sediment and the mature 13.08%, while Proteobacteria, Firmicutes, Bacteroidetes significantly

8
Q. Song et al. Bioresource Technology 332 (2021) 125108

increased by 4.44%, 5.17%, and 4.98%, compared to those in the control Yuling Li: Resources. Fengbo Wang: Data curation, Investigation.
group (Fig. 6C). These phyla are common in anaerobic fermentation Xiaoguang Chen: Formal analysis, Funding acquisition.
reactors, since they are responsible for the degradation of organic
matter. Chloroflexi is the most abundant phylum in this study. It is Declaration of Competing Interest
known to be involved in the hydrolysis and fermentation of sugars in
anaerobic digesters and may play a role in the degradation of a variety of The authors declare that they have no known competing financial
carbohydrates (Lei et al., 2020). Moreover, Synergistetes are reported to interests or personal relationships that could have appeared to influence
be able to utilize amino acids from proteins and peptides which have the work reported in this paper.
been degraded by other microbes to provide short-chain fatty acids and
sulfates for sulfate-reducing bacteria. In this way it has a good resistance Acknowledgements
to PCW degradation products (Xue et al., 2019). Further studies of
genus-level (Fig. 6C and D inner) indicate that the abundance of the This work was sponsored by the Bashan Weaving Group Co., Ltd,
Aminicenantales who are involved in hydrolysis (Candry et al., 2020) in Shanghai Pujiang Program (No: 2020PJD001), Key Laboratory of Water
the mature sludge account to 9.00%. This genus significantly decreased Pollution Control and Environmental Safety of Zhejiang Province,
to 5.69% compared to the inoculated sludge. The abundance of syntro­ China, and Graduate Student Innovation Fund of Donghua University
phobacter, who are involved in the degradation of propionic acid (No: GSIF-DH-M-2021009).
increased from 2.60% to 5.48%. The results reveal that PCW has inhibits
some of the microorganisms in anaerobic sludge, while some others References
specialized in the degradation of polyacrylate became more dominant
(Zhang et al., 2020). Bhuvanesh, S., Maneesh, N., Sreekrishnan, T.R., 2013. Start-up and performance of a
hybrid anoxic reactor for biological denitrification. Bioresour. Technol. 129, 78–84.
In Fig. 6E data are shown for the archaeal communities. At the genus Borowski, S., Kubacki, P., 2015. Co-digestion of pig slaughterhouse waste with sewage
level the dominant genera included Methanosaeta (62.83%), Meth­ sludge. Waste Manage. 40, 119–126.
anobacterium (30.07%), Candidatus_Methanofastidiosum (2.63%), and Calusinska, M., Goux, X., Fossépré, M., Muller, E.E.L., Wilmes, P., Delfosse, P., 2018.
A year of monitoring 20 mesophilic full-scale bioreactors reveals the existence of
Bathyarchaeia (1.93%). A significant decrease of members of Candida­ stable but different core microbiomes in bio-waste and wastewater anaerobic
tus_Methanofastidiosum occurred in the mature sludge compared with the digestion systems. Biotechnol. Biofuels 11 (1). https://doi.org/10.1186/s13068-018-
seed sludge. This led to a reduction of the relative abundance of the 1195-8.
Camacho-Munoz, R., Villada-Castillo, H.S., Solanilla-Duque, J.F., 2020. Anaerobic
archaea. Interestingly, the relative abundance of the Methanosaeta biodegradation under slurry thermophilic conditions of poly(lactic acid)/starch
increased by 44.30%. This genus is known to be able to convert acetate blend compatibilized by maleic anhydride. Int. J. Biol. Macromol. 163, 1859–1865.
to methane(Lu et al., 2020). Thus, the data on the relative abundance of Candry, P., Radic, L., Favere, J., Carvajal-Arroyo, J.M., Rabaey, K., Ganigue, R., 2020.
Mildly acidic pH selects for chain elongation to caproic acid over alternative
the archaea reveal that the methane producing activity in the sludge in
pathways during lactic acid fermentation. Water Res 186, 116396.
the SSSAB was reduced but still acceptable. This can be attributed to the Chen, X., Dai, R., Xiang, X., Li, G., Xu, Z., Hu, T., Abdelgadir, A., 2016. Shock resistance
structure of the reactor and the application of the co-substrate. characteristic of a spiral symmetry stream anaerobic bio-reactor. Water Sci. Technol.
73 (4), 916–923.
Chen, X., Li, G., Lin, H., Li, Y., Ma, Y., Dai, R., Zhang, J., 2017. Operation performance
3.5. Comparison with other treatment methods. and membrane fouling of a spiral symmetry stream anaerobic membrane bioreactor
supplemented with biogas aeration. J. Membr. Sci. 539, 206–212.
Yu et al used fenton-like to clean PCW. For an influent COD of 5200 Chen, X., Wang, Y.u., Wang, Z., Liu, S., 2019. Efficient treatment of traditional Chinese
pharmaceutical wastewater using a pilot-scale spiral symmetry stream anaerobic
mg/L the COD removal reached 92.7% (Yu et al., 2019b). Xu et al used bioreactor compared with internal circulation reactor. Chemosphere 228, 437–443.
an ABR to clean a similar textile wastewater (PVA - polyvinyl alcohol Dai, R., Chen, X., Ma, P., Chu, X., Xiang, X., Luo, Y., Ni, S., Li, G., 2017. Modelling and
wastewater). If the influent COD was 1200 mg/L, the COD removal predicting transport and concentration distribution of sludge in a high performance
anaerobic bioreactor. Biochem. Eng. J. 120, 17–24.
reached 80% (Xu et al., 2005). Therefore, the treatment of PCW by a Dai, R., Chen, X., Xiang, X., Huang, D., Lin, H., Xu, M., 2016. Dispersion characteristics of
combined system of a SSSAB and an AFB is an advanced technology. a spiral symmetry stream anaerobic bio-reactor. Biochem. Eng. J. 110, 27–34.
Based on the results, a SSSAB-AFB process can be used to replace Dai, R., Chen, X., Xiang, X., Wang, Y., Wang, F., 2018. Understanding azo dye anaerobic
bio-decolorization with artificial redox mediator supplement: Considering the
chemical methods successfully. The advantages are obvious: 1) no sec­ methane production. Bioresour. Technol. 249, 799–808.
ondary pollution, 2) low costs (Borowski & Kubacki, 2015), 3) increased Deng, D., Guo, J., Zeng, G., Sun, G., 2008. Decolorization of anthraquinone,
treatment efficiency, 4) generation of renewable energy in the form of triphenylmethane and azo dyes by a new isolated Bacillus cereus strain DC11. Int.
Biodeterior. Biodegrad. 62 (3), 263–269.
biogas (Madigou et al., 2019a, 2019b, 2019c, 2019d, 2019e). dos Santos, A.B., Cervantes, F.J., van Lier, J.B., 2007. Review paper on current
technologies for decolourisation of textile wastewaters: Perspectives for anaerobic
4. Conclusion biotechnology. Bioresour. Technol. 98 (12), 2369–2385.
Feng, F., Tang, X., Qu, C., Lu, X., Liu, Z., Tang, J., Tang, C.-J., Chai, L. 2021.
Hydroxylamine addition enhances fast recovery of anammox activity suffering Cr
The application of co-substrate theory and the use of a SSSAB alle­ (VI) inhibition. Bioresour. Technol., 329, 124920-124920.
viated the inhibitory effect of polyacrylate and effectively improved the Friman, H., Schechter, A., Ioffe, Y., Nitzan, Y., Cahan, R., 2013. Current production in a
biodegradability. The SSSAB-AFB process is efficient particularly for the microbial fuel cell using a pure culture of Cupriavidus basilensis growing in acetate
or phenol as a carbon source. Microb. Biotechnol. 6 (4), 425–434.
treatment of PCW. The COD removal reaches up to 95.2%, which is Gonzalez-Estrella, J., Asato, C.M., Jerke, A.C., Stone, J.J., Gilcrease, P.C., 2017. Effect of
higher than the value obtained by other chemical methods (the highest structural carbohydrates and lignin content on the anaerobic digestion of paper and
COD removal is 92.7%) (Yu et al., 2019b). High-throughput sequencing paper board materials by anaerobic granular sludge. Biotechnol. Bioeng. 114 (5),
951–960.
analysis indicated the relative abundance of Proteobacteria, Firmicutes Guo, J., Chen, W., Wang, L., Ma, F. 2013. SBR co-metabolism process for the advanced
and Bacteroidetes increased, which contribute to the degradation of treatment of petrochemical wastewater. Chem. Indust. Eng. Progr. , 32(11), 2768-
PCW. This study provides a new way for the degradation of PCW as an 2772,2777.
Hu, T., Chen, X., Wang, Y., Zhou, W., Xu, Y., Lin, H., Huang, D., 2017. Operation
alternative to the existing chemical method. characteristics of air-lift external circulation vortex enhancement nitrogen removal
bio-reactor. Chin. J. Environ. Eng. 11 (5), 2723–2730.
CRediT authorship contribution statement Lei, Z., Yang, S., Wang, L., Huang, X., Wang, X.C., Li, Y.-Y., Li, Q., Zhao, Y., Chen, R.,
2020. Achieving successive methanation and low-carbon denitrogenation by a novel
three-stage process for energy-efficient wastewater treatment. J. Cleaner Prod. 276.
Qi Song: Methodology, Writing - review & editing. Zheng Sun: Lin, S., Yu, Z., Jiang, X., 1991. Research on biochemical dynamics of dyeing wastewater
Conceptualization. Yong Chang: Project administration. Weifeng containing PVA and its application (Chinese). Shanghai Environ. Sci. 12, 5–9.
Zhang: Software. Yingzhi Lv: Writing - review & editing. Jiayi Wang:
Visualization. Fenghao Sun: Validation. Yanxue Ma: Supervision.

9
 Journal of Hazardous Materials 403 (2021) 123986

Contents lists available at ScienceDirect

Journal of Hazardous Materials

journal homepage: www.elsevier.com/locate/jhazmat

Effective abatement of 29 pesticides in full-scale advanced treatment

processes of drinking water: From concentration to human exposure risk
Huiyu Dong a, *, Lei Xu a, b, Yuanxiang Mao a, c, Yan Wang a, Shule Duan a, c, Junfeng Lian d,
Jin Li b, Jianwei Yu a, Zhimin Qiang a, *
a
Key Laboratory of Drinking Water Science and Technology, Research Center for Eco-Environmental Sciences, University of Chinese Academy of Sciences, Chinese
Academy of Sciences, Beijing 100085, China
b
College of Environmental Science and Engineering, Qingdao University, Qingdao 266071, China
c
School of Chemical & Environmental Engineering, China University of Mining and Technology-Beijing, Beijing 100083, China
d
School of Architectural, Surveying and Mapping Engineering, Jiangxi University of Science and Technology, 86 Hong-qi Road, Ganzhou 341000, China

A R T I C L E I N F O A B S T R A C T

Keywords: This study investigated the occurrence and removal of 29 pesticides in 4 drinking water treatment plants
Drinking water (DWTPs) with conventional and advanced treatment processes (i.e., ozonation + biological activated carbon, and
Pesticide ultrafiltration) in Shanghai, China from 2018 to 2019. The concentration levels of target pesticides in raw waters
Advanced treatment process
ranged from below the limit of quantification (<LOQ) to 2391.3 ng L− 1, whereas in finished waters significantly
Human exposure risk
decreased to a range of <LOQ to 269 ng L− 1. The application of the advanced treatment process could effectively
remove target pesticides. For example, the highest removal of isocarbophos after the advanced treatment process
reached 94.6%, whereas its removal after the conventional treatment process was just 42.6%. The estimated
daily intake level of target pesticides in finished water, which were pretty lower than acceptable daily intake
levels in finished water, could be reduced by 67% if combining the conventional and advanced treatment pro­
cesses. Although the multiple pesticides in raw waters were effectively removed by advanced treatment process,
their synergistic toxic effects induced by the residual pesticides and their degradation by-products on human
health deserve further attentions.

1. Introduction has implemented a rigorous Drinking Water Directive (Karabelas et al.,

2009), which sets a limit of 100 ng L− 1 for a single pesticide and 500 ng
In recent years, with rapid developments of agricultural application, L− 1 for the sum of all pesticides by monitoring and quantifying active
the consumption of pesticides is ever-increasing (Rahman, 2013). The ingredients (98/83/EC).
effectiveness of pesticide in preventing or eliminating pests and weeds Nowadays, the source of drinking water is mostly surface water
caused the wide use in agricultural and non-agricultural fields (Aspelin, around the cities or upstream of the agricultural watershed, often con­
1994; Barbash and Resek, 1996), resulting in the detection of pesticide taining multiple pesticides. The occurrence of micro-pollutants in the
residues in aquatic environment. For example, Ccanccapa et al. and source water induces a severe challenge to the conventional drinking
Papadakis et al. detected 50 and 68 pesticide residues in the Turia and water treatment processes (i. e., coagulation-flocculation, sedimenta­
Júcar Rivers and the basin of Lake Vistonis, respectively (Ccanccapa tion, filtration, and disinfection) (Valcárcel et al., 2013). Combining
et al., 2016; Papadakis et al., 2015). Multiple pesticides were also conventional and advanced water treatment processes has become a
identified from the Songhua River and Jiulong River, China (Cai et al., feasible option to enhance the removal of target micro-pollutants (Yang
2014; Zheng et al., 2016). These pesticide residues enter surface water et al., 2017).
or groundwater (Plakas and Karabelas, 2012a) via spray drift, runoff At present, ozonation, activated carbon adsorption, and membrane
water, and drainage water (Ikehata and Gamal El-Din, 2005), posing a filtration are popular advanced treatment processes applied in
threat to the quality of source waters and human health. Considering the drinking water treatment plant (DWTP) (Plakas and Karabelas,
possibility of human exposure to pesticide residues, the European Union 2012b; Wen et al., 2018; Zularisam et al., 2006). Combination of

* Corresponding authors.
E-mail addresses: [email protected] (H. Dong), [email protected] (Z. Qiang).

https://doi.org/10.1016/j.jhazmat.2020.123986
Received 24 February 2020; Received in revised form 5 September 2020; Accepted 12 September 2020
Available online 16 September 2020
0304-3894/© 2020 Elsevier B.V. All rights reserved.
H. Dong et al. Journal of Hazardous Materials 403 (2021) 123986

ozonation and biological activated carbon (BAC) processes can pro­ 2. Materials and methods
long the service life of granular activated carbon and curb the for­
mation of disinfection by-products (DBPs) (Reungoat et al., 2012). 2.1. Process and sampling details
BAC process is one of the most promising, environmentally friendly,
and economically viable advanced treatment processes (Kor­ The samplings of A, B, C, and D DWTPs in eastern China were con­
otta-Gamage and Sathasivan, 2017). As the ozonation process can ducted once a month for a total of 12 months (April 2018–March 2019).
increase the biodegradability of organic pollutants, the combination A schematic diagram of the 4 DWTPs is shown in Fig. 1. The advanced
of ozonation and BAC processes can effectively remove natural treatment processes of A, B, and C DWTPs include ozonation +BAC
organic matter (NOM), ozone transformation products, DBP pre­ processes, where water samples were collected at six sites (i.e., raw
cursors, and taste and odor compounds (Simpson, 2008). Ultrafiltra­ water, sedimentation effluent, sand filtration effluent, ozonation
tion (UF) process, using hydrostatic pressure to force water through a effluent, BAC effluent, and finished water). The advanced treatment
semi-permeable membrane, is usually set up after the BAC process to process of D DWTP includes ozonation, BAC, and UF units, where water
further polish micro-pollutants and improve water quality (Benitez samples were also collected at 6 points (i.e., raw water, sedimentation
et al., 2009). The UF process usually has the following advantages: effluent, ozonation effluent, BAC effluent, UF effluent, and finished
better quality percolate, moderate operating temperature, low energy water). The detailed information about the Huangpu River could be
consumption, and no chemical substances. However, whether the found in Text S1. The raw water quality and advanced treatment process
advanced treatment processes can effectively remove various pesti­ parameters are listed in Table S1 and Table S2. The 12 samplings were
cides is not yet clear. Due to the increased application of advanced carried out in 2018–2019 and 2 L water sample per site was taken for
treatment processes, it is necessary to investigate the removal of subsequent analysis. All water samples were stored in pre-cleaned amber
micro-pollutants in full-scale advanced treatment processes for future glass bottles and pretreated within 2 days.
optimization and improvement.
In this study, the occurrence and removal of 29 pesticides in 4 2.2. Chemicals
DWTPs (A, B, C, and D) in the Huangpu River Basin were investigated
once a month in 2018–2019, covering spring, summer, autumn, and All 29 pesticides were purchased from Ark Pharm, Inc. (Libertyville,
winter. The 29 target pesticides were selected due to their wide appli­ IL, US), with high purity (>97%) standards. The CAS, molecular for­
cations and frequent detections in China (Pan et al., 2018; Tang et al., mula, molecular mass, and molecular structure of each pesticide were
2019; Zheng et al., 2016). The objectives of this study are to: (1) listed in Table S3. Other chemicals used in sample preparation and
investigate the temporal variation of 29 pesticides in raw water; (2) analysis were HPLC grade or analytical grade.
assess the removal efficiency of the target pesticides by conventional and
advanced treatment processes; (3) evaluate exposure risk of each 2.3. Sample analysis
pesticide to humans.
Solid-phase extraction (SPE) using the Oasis HLB cartridges (500 mg,
6 mL, Waters, Millford, MA, USA) was adopted to enrich water samples.
The samples were filtered through 1 µm glass microfiber filters (GF/B,

Fig. 1. Schematic diagram of treatment processes of 4 drinking water treatment plants investigated.

2
H. Dong et al. Journal of Hazardous Materials 403 (2021) 123986

Whatman, Mainstone, England), and then acidified to pH 3.0 with 40% buprofezin (33%) (Table 1). In the finished water samples, the detection
H2SO4 (v/v), followed by adding 0.5 g L− 1 Na2EDTA as the chelating frequencies of 12 pesticides reached 100% (one benzimidazoles, five
agent before extraction. 1 L of each water sample was passed through nicotine and amides, and six triazoles) and the lowest were found with
the preconditioned SPE cartridge. Methanol (4 mL × 4) was employed thiabendazole and pyrimethanil (around 15%), indicating that the re­
to eluted samples, which stripped off using high purity nitrogen. Prior to movals of benzimidazoles, triazoles, nicotine and amides were not
sample injection, the sample was dissolved with 400 μL methanol and complete.
600 μL ultrapure water to a final volume of 1 mL. The detected concentration of pesticides in the raw water samples
The extracted water samples were analyzed by an ultra-performance ranged from <LOQ to 2391.3 ng L− 1, whereas decreased to
liquid chromatography-triple quadrupole mass spectrometer (with <LOQ–269 ng L− 1 in the finished water samples. Notably, the highest
electrospray ion source) (Waters ACQUITY UPLC/MS/MS Xevo TQ, concentrations of pesticides in raw water and finished water were both
Waters, USA); An ACQUITY BEH C18 column (2.1 mm × 100 mm, observed with carbendazim, which is widely used to control various
1.7 µm, Waters, USA) was used for the chromatographic separation of fungi affecting crops such as fruits, nuts, vegetables, etc. (Panadés et al.,
the target compounds. Table S4 and Table S5 show the specific opera­ 2000). After treatment of DWTPs, the maximum concentration of each
tion parameter, retention time, precursor ion, and product ion. pesticide in finished water was reduced nearly by 57% and the average
Pesticides were identified based on the corresponding parent ion and concentration was reduced nearly by 64%. In general, the combination
two product ions, together with the retention time, and quantified based of conventional treatment process with the advanced treatment process
on the calibration standard curves constructed from the SPE and analysis achieved significant removals of target pesticides. Except for carben­
of spiked water samples. A series of spiked samples were prepared at dazim, the maximum concentrations of other pesticides were
eight different concentrations (0.1, 0.5, 1, 10, 50, 100, 500, <100 ng L− 1, below the limit established for individual pesticide in
1000 ng L− 1). They were extracted and analyzed according to the same drinking water according to the Drinking Water Directive (98/83/EC).
procedures as that used for the surface water. The calibration curves Out of 29 pesticides in raw water, over 22 were detected in each
were constructed by the peak area ratios of analysis to the internal season, with the detection frequency of buprofezin being the lowest
standard. These calibration standard curves compensated for losses in (33.3%, in spring; 33.3%, in summer; 58%, in autumn; not detected, in
both SPE process and instrument variations in UPLC/MS/MS. winter) (Fig. S1). As shown in Fig. S1, the detection frequency of tri­
azoles is relatively high. Except for diniconazole, the other 8 pesticides
2.4. Quality control were all detected in 4 seasons. The detection frequencies of diniconazole
in spring and winter were 66.7% and 83.3%, respectively, and achieved
By measuring pesticide recovery, the limit of detection (LODs) and 100% in summer and autumn, indicating the wide application of tri­
quantification (LOQs) and the regression coefficient prior to sample azoles in this basin.
analysis, the method validation was performed. The recoveries, LODs, To analyze the temporal variation, 2 pesticides with relatively high
and LOQs of target pesticides are shown in Table S6. The recoveries of concentrations from each group were selected. Out of 12 representative
29 target pesticides (500 ng L− 1) spiked to the filtered surface water pesticides, 8 pesticides (isoprocarb, fenobucarb, carbendazim, metola­
(n = 3) were between 68.5% and 119.2%. The LOD and the LOQ for chlor, dimethoate, paclobutrazol, azoxystrobin, and isoprothiolane)
each pesticide were defined as the concentrations corresponding to the were detected at their lowest levels in autumn and 6 pesticides (thia­
signal/noise (S/N) of 3 and 10, respectively. HLB cartridges and water bendazole, carbendazim, metolachlor, dimethoate, isocarbophos, and
samples were stored in a cold box (<4 ◦ C) to prevent degradation until isoprothiolane) were present at their highest levels in spring (Fig. 2).
extraction. The seasonality of target pesticide occurrence may be attributed to the
precipitation dilution. The precipitation in June, 2009 was 158.9 mm,
2.5. Risk assessment while the precipitation was only 48.8 mm in December (Jiang et al.,
2011). High-flow flux in early autumn would dilute the concentration of
Human exposure risk assessment was based on the estimated daily pesticides in the Huangpu River.
intake (EDI), which was calculated from the concentration of pesticides
in water and the water consumption, which was further compared with 3.2. Variations of pesticide levels in different water treatment processes
the acceptable daily intake (ADI) for the three groups of people with
different ages (i.e., adults, juveniles, and infants). The ADI is an estimate Then the 12 pesticides with high detection frequency were selected
of the maximum daily intake, which is expressed on a weight basis, and to analyze the concentration variations in different units of 4 DWTPs. As
were taken from the pesticide toxicity database (University of Hert­ shown in Fig. 3, the pesticide concentrations show an obvious down­
fordshire, PPDB: Pesticide Properties DataBase), while the EDI is ward trend in DWTPs, which decrease significantly after the advanced
calculated by the following equation (Eq. 1): treatment processes. The removal efficiencies of pesticides by conven­
tional processes (i.e., coagulation, sedimentation, and sand filtration)
Cw ∗ Iw
EDI = (1) were limited, with removal ratios of only five pesticides reaching 50%
Bw (55.1% for fenobucarb, 70.7% for thiabendazole, 65.3% for carbenda­
Cw is the concentration of pesticide detected in the water sample; Iw zim, 55.2% for dimethoate, and 68.0% for azoxystrobin). The removal of
is the daily water intake (2 L for adults, 1 L for juveniles, 0.75 L for benzimidazoles by conventional processes was higher than those of
infants); Bw is the body weight (60 kg for adults, 10 kg for juveniles, nicotine and amides (19.3% for imidacloprid, and 31.6% for metola­
5 kg for infants) (Younes and Galal-Gorchev, 2000). chlor), which was consistent with the previous study (Griffini et al.,
1999). These results were in line with our anticipation because con­
3. Results and discussions ventional processes are an effective way to remove small suspended
solids and colloidal particulate matter (Crittenden et al., 2012), while
3.1. Occurrences of target pesticides in raw water and finished water deficient in removing soluble micro-pollutants. For example, low re­
movals of micro-pollutants were also observed in conventional drinking
In the raw water samples, out of 29 pesticides (six carbamates, two water treatment processes (<30%) (Diemert et al., 2013; Huerta-Fontela
benzimidazoles, six nicotine and amides, two organophosphates, nine et al., 2011; Liu et al., 2016). Micro-pollutants associated with surfaces
triazoles, and six others), the detection frequencies of 20 pesticides were of suspended particles and flocculants can be removed during sedi­
100% (two carbamates, one benzimidazoles, four nicotine and amides, mentation, which is a major contributor to the removal of
eight triazoles, and five others) and the lowest was found with micro-pollutants during conventional treatment process. Due to strong

3
H. Dong et al. Journal of Hazardous Materials 403 (2021) 123986

Table 1
Detection concentration and detection frequency of six major families of 29 pesticides in the raw water and finished water samples.
Family Pesticide LOQa (ng L− 1) Raw water (ng L− 1) Finished water (ng L− 1)
b c d e
Max. Min. Mean Freq. (%) Max. Min. Mean Freq. (%)

Carbamates Isoprocarb 6.25 4.7 0.2 1.62 100 2.1 0.1 0.73 90
Fenobucarb 2.96 1.2 0.16 0.49 100 0.5 <LOQ 0.21 73
Pirimicarb 1.37 0.2 <LOQ 0.07 67 0.02 <LOQ 0.01 27
Carbofuran 1.7 6 <LOQ 2.45 81 3.5 <LOQ 0.79 81
Benzimidazoles Thiabendazole 54.46 11.9 0.06 0.54 77 0.1 <LOQ 0.03 15
Carbendazim 26.9 2391.3 32 121.82 100 269 0.1 14.35 100
Nicotine and amides Metalaxyl 2.07 18.2 2.6 8.09 100 10.2 0.5 4.68 100
Imidacloprid 1.88 50.4 8.64 21.26 100 26.6 1.5 10.97 100
Acetamiprid 1.3 21.1 1.26 10.35 100 13.2 0.78 5.49 100
Boscalid 2.28 23.52 <LOQ 3.86 73 27.52 <LOQ 2.08 65
Metolachlor 16.04 27.6 2.7 12.16 100 15.9 1.3 5.55 100
Thiamethoxam 7.08 65.9 1.02 13.19 92 31.8 <LOQ 9.57 100
Organophosphates Dimethoate 3.36 11 <LOQ 1.98 92 2.76 <LOQ. 0.57 42
Isocarbophos 3.4 80.5 15 35.96 98 10.7 <LOQ 1.94 85
Triazoles Diniconazole 8.98 1.5 <LOQ 0.37 88 0.6 <LOQ 0.15 33
Difenoconazole 2.67 7.1 0.1 1.32 100 2.1 <LOQ 0.43 96
Myclobutanil 3.78 5.5 0.28 2.14 100 4.1 0.1 1.61 100
Epoxiconazole 5.97 5.7 0.2 1.17 100 1.6 <LOQ 0.34 81
Triadimenol 6.48 41.6 2.22 19.81 100 29.7 1.24 10.59 100
Propiconazole 3.82 27.1 3.24 12.03 100 19.8 0.42 7.53 100
Hexaconazole 1.44 36.1 3.72 18.05 100 34.5 0.68 11.87 100
Tebuconazole 3.37 82 5.48 28.85 100 29.8 0.28 9.62 100
Paclobutrazol 1.64 36.3 8.5 17.77 100 17 1.9 8.89 100
Others Buprofezin 10 0.16 <LOQ 0.03 33 0.1 <LOQ 0.01 33
Ametryn 1.16 20.6 4.2 10.60 100 2.02 <LOQ 0.40 88
Azoxystrobin 1.89 87.5 8.4 23.06 100 19.2 <LOQ 2.21 88
Dmethomorph 2.19 16.7 0.4 4.89 100 4.3 <LOQ 1.16 90
Pyrimethanil 45.12 3.5 0.1 1.19 100 0.4 <LOQ 0.17 15
Isoprothiolane 1.73 91.6 15.4 40.25 100 9.9 0.1 2.26 88
a
Limit of quantification;
b
Maximum;
c
Minimal;
d
Mean;
e
Frequency of detection of each pesticide in all water samples.

60
5 Isoprocarb 75% Max. (a) Thiabendazole (b) Imidacloprid (c)
Fenobucarb 50% Mean 1000 Carbendazim 50 Metolachlor
Min.
Concentration (ng L-1)

Concentration (ng L-1)

Concentration (ng L-1)

4 25%
100 40
3
30
10
2
20
1
1 10
0.1
0 0

Spring Summer Autumn Winter Spring Summer Autumn Winter Spring Summer Autumn Winter
100 100
Dimethoate (d) Tebuconazole (e) Azoxystrobin (f)
80 Isocarbophos Paclobutrazol 100 Isoprothiolane
80
Concentration (ng L-1)

Concentration (ng L-1)

Concentration (ng L-1)

60
60

40 50
40

20
20
0 0
0
-20
Spring Summer Autumn Winter Spring Summer Autumn Winter Spring Summer Autumn Winter

Fig. 2. Variations in concentration of representative pesticides in the raw waters in different seasons.

hydrophobicity, large molecular mass, or the tendency to form hydrogen After advanced treatment processes, the removals of pesticides were
bonds with other substances, fenobucarb, thiabendazole, carbendazim, further polished, among which the removal of isocarbophos was the
and azoxystrobin may be prone to adsorbed by suspended particles or highest (94.6%), followed by thiabendazole and isoprothiolane (both
flocculants in water, resulting in high removals in the conventional 94.4%). The removals of benzimidazoles, organophosphates, and others
treatment processes. were also > 70%, while those of nicotine and amides were relatively low

4
H. Dong et al. Journal of Hazardous Materials 403 (2021) 123986

6 10000 60
Isoprocarb 75% Thiabendazole Imidacloprid
Max.
Fenobucarb 1000 Carbendazim 50 Metolachlor
50% Mean
Concentration (ng L-1)

Concentration (ng L-1)

Concentration (ng L-1)

Min.
4 100
25% 40

10
30
2 1
20
0.1
10
0 0.01
0
0.001
RW CS SF O3 BAC FW RW CS SF O3 BAC FW RW CS SF O3 BAC FW
100 100
Dimethoate Tebuconazole 100 Azoxystrobin
Isocarbophos Paclobutrazol Isoprothiolane
80 80

Concentration ( ng L-1)
Concentration (ng L-1)

Concentration (ng L-1)

80
60 60
60

40 40
40

20 20 20

0 0 0

-20 -20 -20

RW CS SF O3 BAC FW RW CS SF O3 BAC FW RW CS SF O3 BAC FW

Fig. 3. Variations in concentration of representative pesticides in different processes.

(48.4% for imidacloprid, 54.4% for metolachlor). In general, the ozon­

ation process can effectively remove the target pesticides. Ozonation
150
O3 influent 75%
process promotes the formation of molecules with a lower average Max. 6
molecular weight, which facilitates subsequent adsorption of these O3 effluent 50% Mean
molecules onto activated carbon. The removal of organic matter in BAC 25% Min.
Concentration (ng L-1)

involves three ways including oxidation, adsorption, and biodegrada­ 100 5

tion (Lambert and Graham, 1995; Li et al., 2006; Takeuchi et al., 1997).

TOC (mg L-1)

Ormad et al. (2008) found that combined ozonation +BAC processes can
significantly improve the removal efficiency of pesticides with the re­ 4
movals of 44 target pesticides being > 60% . Moreover, the ozonation 50
+BAC processes can remove the precursors of DBPs, and help to
decrease the dose of chlorine in the disinfection process, thereby 3
reducing the formation of chlorinated DBPs (Takeuchi et al., 1997).
0
3.3. Contribution of each advanced treatment process
2

To evaluate the removal contribution of each advanced treatment Carbendazim Isoprothiolane Tebuconazole Isocarbophos TOC

process, four pesticides (i.e., carbendazim, isoprothiolane, tebucona­ Fig. 4. Variations in concentration of representative pesticides and TOC in the
zole, and isocarbophos) with higher concentrations were selected as ozonation process.
target pesticides in this section to compare the influent and effluent
concentration of pesticides. Mazellier et al. (2002) reported that the k(HO• + carbendazim) was as high as
(2.2 ± 0.3) × 109 M− 1 s− 1, which can interpret the high removal effi­
3.3.1. Ozonation process ciency of carbendazim. Previous studies verified that ozone molecules
As shown in Fig. 4, the ozonation process can further abate the target tend to attack deprotonated amine organic compounds that contain
pesticides after conventional treatment processes, whereas TOC exhibits double bonds and aromatic rings (Hollender et al., 2009; Wang et al.,
a slight increase (2.9%). Compared with the influent of ozonation pro­ 2020). In addition, the CO double bond has a high polarity and easily
cess, the concentrations of four pesticides in effluents further decreased reacts with HO•, resulting in double bond breakage (Xu et al., 2019). The
by 18.4–33.6% (33.6% of carbendazim, 32% of isoprothiolane, 18.4% of four target pesticides selected in this study all contain aniline groups or
tebuconazole, and 30.5% of isocarbophos). Ozone is widely used in double bonds (like CC and CO), which can be easily attacked by the
disinfection, odor control, and color removal in water treatment (Ike­ ozonation process. For example, Li et al. (2019) reported that the CC
hata and Gamal El-Din, 2005). There are two mechanisms for ozone double bond and thioether in isoprothiolane were the reactive sites
reaction: (1) ozone molecules directly oxidize with pollutants; (2) ozone during ozonation process, triggering the formation of some trans­
molecules generate hydroxyl radical (HO•) through radical-type chained formation products including isoprothiolane sulfoxide, isoprothiolane
reactions, thereby efficiently removing contaminants (Am Water Works disulfoxide, isoprothiolane sulfone, etc. Xu et al. (2019) found that the
Res et al., 1991). During the direct oxidation process, ozone selectively C˭O double bond in carbendazim can be easily broken by HO•, yielding
reacts with specific functional group, such as olefins, activated aro­ an intermediate, TP237. In addition, Rokbani et al. reported that during
matics, and tertiary amines, the reaction rate of which may differ by the ozonation process, the reaction between tebuconazole and HO• is
several orders of magnitude (Wang et al., 2020). For the indirect expected to proceed through several routes, including the addition of
oxidation, it is not selective and the reaction rate is relatively high, aromatic C6-cycle or C5 aromatic N-heterocycle and the H-abstraction
usually greater than 109 M1 s1 (Von Gunten, 2003). For example,

5
H. Dong et al. Journal of Hazardous Materials 403 (2021) 123986

process from different sites (Rokbani et al., 2019). Although the ozon­ occur at low temperatures, and the amount of extracellular enzymes,
ation process can effectively abate pesticides, it should be noted that such as β-glycosidase, is reduced, which is necessary for the pollutant
ozone may react with bromide that present in water to form bromate abatement (Le Bihan and Lessard, 2000). Overall, the contribution rate
(Ruffino et al., 2020; Wu et al., 2019; Yang et al., 2019), which is defined of the ozonation +BAC process is much higher than that of the con­
as a highly carcinogenic substance by the World Health Organization ventional treatment process (Fig. S2). The contribution rate of advanced
and is prescribed to have a maximum concentration of 10 μg L− 1 in treatment processes ranged from 63.1% (tebuconazole, in winter) to
drinking water (Yang et al., 2019; Yao et al., 2020). Although the con­ 95.6% (carbendazim, in summer), further indicating that the advanced
centration of bromate in the effluent did not reach the guideline in this treatment process is more efficient than the conventional treatment
study, the dosage of ozone should also be strictly controlled. process.

3.3.2. Biological activated carbon process 3.3.3. Ultrafiltration process

As shown in Fig. 5, the BAC process further polishes the four target To further improve the drinking water quality, D DWTP was equip­
pesticides in the ozonation effluents. The removal ratios ranged from ped with UF process after the ozonation +BAC process. From Fig. 6, it
23.6% to 58.1%, with 58.1% of carbendazim, 26.8% of isoprothiolane, can be seen that the UF process can further abate the concentration of
36.7% of tebuconazole, and 23.6% of isocarbophos. Moreover, the TOC target pesticides. The removal efficiencies ranged from 20.7% to 46.4%
values of water samples also showed an obvious decreasing trend, (46.4% of carbendazim, 28% of isoprothiolane, 20.7% of tebuconazole,
approximately 11.1%. The chemical structure, functional groups, and 27.5% of isocarbophos). In addition, the removal of TOC was 9.7%.
morphology and the competitive effects of the compounds all have an After the UF process, the average concentrations of the four pesticides all
impact on the removal of different pesticides. Carbendazim has a lower decreased to < 10 ng L− 1, which is much lower than the limit
molecular mass and a smaller volume (Table S3), which makes it easy to (100 ng L− 1 per pesticide) of Drinking Water Directive (98/83/EC). The
enter the interior of activated carbon, causing a relatively high removal removal of pollutants by the UF process is mainly through the me­
efficiency. In addition, the increase of the adsorption capacity is also chanical screening of the UF membrane and the adsorption of the
related to the enhancement of the binding capacity of the hydrogen membrane surface (Chu et al., 2017; Sheng et al., 2016; Yoon et al.,
bond and π‒π bond between the heterocyclic atom of the compound and 2006). The molecular weight cut-off range of UF membranes is in the
the adsorbent (Ji et al., 2009). Compared with the other 3 pesticides, range of 500–300,000 Da, while the molecular masses of the four target
there is only one methyl group in the molecular structure of carbenda­ pesticides are all less than 500 Da, indicating that the adsorption of UF
zim (other pesticides have three or four methyl groups), which leads to membranes may be dominant during the removal of these pesticides.
smaller steric hindrance and higher removal rate. Benitez et al. (2009) reported that the removal of phenyl urea herbicides
Ambient temperature may impact the removal of pesticides in the was mainly due to the adsorption of the UF membrane surface. However,
BAC process. Low temperatures in winter may inhibit the metabolism of it should be noted that the UF process would inevitably induce the
microorganisms in the BAC process. To investigate the effect of tem­ membrane fouling, which leads to a reduction in membrane flux. A
perature on the BAC performance, the removals of carbendazim and previous study showed that the ozonation process can increase the hy­
tebuconazole in A DWTP were compared in summer (28 ◦ C) and winter drophilicity of organic pollutants, thereby optimizing the interaction
(13 ◦ C). As shown in Fig. S2, the contribution of BAC in winter decreased between them and UF membranes, and reducing membrane fouling (Li,
obviously compared with summer (carbendazim decreased by 40.9% 2019). The removal difference between the D DWTP and the other three
and tebuconazole decreased by 20.4%). Previous study reported that the DWTPs were further compared in Fig. S3. Overall, higher removal ratios
temperature may be the most important abiotic factor affecting the of the target pesticides were observed in D DWTP. The average removal
growth and activity of microbial populations (Kim et al., 2014). The ratio of pesticides in the D DWTP was 70%, while 54% for other 3
performance of BAC filtration is positively correlated with water tem­ DWTPs. The application of UF process can polish the removal of residual
perature and low temperature has an adverse effect on the performance pesticides from ozonation +BAC effluent.
of BAC systems (Dong et al., 2019; Laurent et al., 2003; Van der Aa et al.,
2011). Lower temperatures may cause slower metabolism, which 3.4. Exposure assessment for human health
reduced nutrient utilization and absorption of biofilms and inhibited
biodegradation processes. The temperature can also affect the microbial To evaluate the human exposure risk induced by pesticides in
community structure composition (Chen et al., 2017; Gonzalez-Martinez drinking water, risk assessment was conducted for 3 groups with
et al., 2018; Kim et al., 2014). A large number of ciliated protozoans different ages (i.e., adults, juveniles, and infants). The ADI and EDI of 29

30
120 BAC influent 75% Max. 6 UF influent 75% Max. 5.0
BAC effluent 50% Mean UF effluent
Min. 50% Mean
25%
Concentration (ng L-1)

25% Min. 4.5

Concentration (ng L-1)

5 20
TOC (mg L-1)

80
TOC (mg L-1)

4.0

4 3.5
10
40 3.0
3
2.5
0
2
0 2.0

Carbendazim Isoprothiolane Tebuconazole Isocarbophos TOC Carbendazim Isoprothiolane Tebuconazole Isocarbophos TOC

Fig. 5. Variations in concentration of representative pesticides and TOC in the Fig. 6. Variations in concentration of representative pesticides and TOC in the
BAC process. UF process.

6
H. Dong et al. Journal of Hazardous Materials 403 (2021) 123986

pesticides are plotted in Fig. 7 and Fig. S4. Due to the effectiveness of the drinking water might induce synergistic toxic effects on human health,
conventional and advanced treatment processes, the average EDI level which warrants further attention.
in finished water was 67% lower than the raw water. Each EDI level for
29 pesticides was 3–9 orders of magnitude lower than the ADI level, 3.5. Environmental indications
indicating that there was almost no chronic exposure risk of these target
pesticides. The highest EDI level was found from carbendazim in raw The application of conventional treatment process +advanced
water and hexaconazole in finished water, with the order of EDI levels treatment process in 4 DWTPs effectively removed the multiple pesti­
for the three groups being infants > juveniles > adults. Although this cides in the source water. Based on these results, the following sugges­
risk assessment was a meaningful tool to conduct risk assessment, it may tions are concluded: (1) strict protection measures should be taken in the
be simplified and these results should be treated with caution since the source water area and buffer zones should be established to minimize
risk data was obtained from the Pesticide Properties DataBase. Due to source pollution; (2) regularly monitor variations in water quality and
the presence of multiple pesticides and degradation products in the make corresponding adjustments to each process section; (3) ozone
water sample, the single compound exposure scenario is simplified. dosage can be increased appropriately to make up for the loss of BAC
Long-term exposure to multiple pesticides and degradation products in removal performance in winter; (4) indicative pollutants can be

Raw Water Finished Water

0.006 adults 75% (a) 0.006 (d)
Max.
juveniles
50% Mean
infants
Min.
0.004 25% 0.004

0.002 0.002

0.000 0.000

(b) (e)
0.012 0.012
EDI (μg/kg bw/day)
EDI (μg/kg bw/day)

0.008 0.008

0.004 0.004

0.000 0.000

0.020 (c) 0.020 (f)

0.015 0.015

0.010 0.010

0.005 0.005

0.000 0.000
Metalaxyl

Metolachlor

Triadimenol

Hexaconazole

Paclobutrazol

Metalaxyl

Metolachlor

Triadimenol
Dimethoate

Propiconazole
Hexaconazole

Paclobutrazol
Diniconazole
Difenoconazole
Myclobutanil

Propiconazole

Pyrimethanil

Dimethoate

Diniconazole
Epoxiconazole

Tebuconazole

Difenoconazole
Myclobutanil
Epoxiconazole

Tebuconazole

Pyrimethanil
Thiabendazole

Isoprothiolane

Thiabendazole

Isoprothiolane
Pirimicarb

Thiamethoxam

Pirimicarb
Carbendazim

Boscalid

Thiamethoxam
Carbendazim
Buprofezin
Ametryn

Acetamiprid
Boscalid
Fenobucarb

Imidacloprid
Acetamiprid

Imidacloprid

Buprofezin
Ametryn
Isoprocarb

Carbofuran

Azoxystrobin
Dimethomorph

Isoprocarb
Fenobucarb

Carbofuran

Azoxystrobin
Dimethomorph
Isocarbophos

Isocarbophos

Fig. 7. Comparison of EDI levels for adults, juveniles and infants in the raw waters and finished waters.

7
H. Dong et al. Journal of Hazardous Materials 403 (2021) 123986

screened for monitoring (e.g., carbendazim with higher detection con­ Ccanccapa, A., Masiá, A., Andreu, V., Picó, Y., 2016. Spatio-temporal patterns of
pesticide residues in the Turia and Júcar Rivers (Spain). Sci. Total Environ. 540,
centration and detection frequency in this study).
200–210.
Chen, Y., Lan, S., Wang, L., Dong, S., Zhou, H., Tan, Z., Li, X., 2017. A review: driving
4. Conclusions factors and regulation strategies of microbial community structure and dynamics in
wastewater treatment systems. Chemosphere 174, 173–182.
Chu, K.H., Fathizadeh, M., Yu, M., Flora, J.R., Jang, A., Jang, M., Park, C.M., Yoo, S.S.,
In light of the fact that there is little field research on full-scale Her, N., Yoon, Y., 2017. Evaluation of removal mechanisms in a graphene oxide-
conventional treatment process combined with advanced treatment coated ceramic ultrafiltration membrane for retention of natural organic matter,
pharmaceuticals, and inorganic salts. ACS Appl. Mater. Interfaces 9, 40369–40377.
process to remove pesticides, this study investigated the occurrence and
Crittenden, J.C., Trussell, R.R., Hand, D.W., Howe, K.J., Tchobanoglous, G., 2012.
removal of 29 pesticides in 4 DWTPs in eastern China and evaluated the MWH’s Water Treatment: Principles and Design. John Wiley & Sons,, USA.
exposure risk of different human groups to residual pesticides. Based on Diemert, S., Wang, W., Andrews, R.C., Li, X.-F., 2013. Removal of halo-benzoquinone
the above, the following conclusions can be drawn: (emerging disinfection by-product) precursor material from three surface waters
using coagulation. Water Res. 47, 1773–1782.
Directive 98/83/EC, 1998. Council directive of 3 November 1998 on the quality of water
• The occurrence of 29 pesticides in source water exhibited an obvious intended for human consumption. Official J. Eur. Communities. L/330/32-54.
seasonal characteristic. Dong, L., Liu, W., Yu, Y., Gu, P., Chen, G., 2019. Preparation, characterization, and
application of macroporous activated carbon (MAC) suitable for the BAC water
• The removal efficiency of pesticides by conventional process is treatment process. Sci. Total Environ. 647, 1359–1367.
limited, while the combination of conventional with advanced Gonzalez-Martinez, A., Muñoz-Palazon, B., Maza-Márquez, P., Rodriguez-Sanchez, A.,
treatment process can effectively enhance the removal of pesticides, Gonzalez-Lopez, J., Vahala, R., 2018. Performance and microbial community
structure of a polar Arctic Circle aerobic granular sludge system operating at low
with the average concentration of each pesticide being lower than temperature. Bioresour. Technol. 256, 22–29.
those of EU standards in finished water. Griffini, O., Bao, M., Burrini, D., Santianni, D., Barbieri, C., Pantani, F., 1999. Removal of
• Although the mineralization of organic matter by ozone is negligible, pesticides during the drinking water treatment process at Florence water supply,
Italy. J. Water Supply Res. T 48, 177–185.
ozonation process plays an important role in the abatement of target Hollender, J., Zimmermann, S.G., Koepke, S., Krauss, M., McArdell, C.S., Ort, C.,
pesticides. Singer, H., von Gunten, U., Siegrist, H., 2009. Elimination of organic micropollutants
• The EDI values of target pesticides in the finished water, significantly in a municipal wastewater treatment plant upgraded with a full-scale post-ozonation
followed by sand filtration. Environ. Sci. Technol. 43, 7862–7869.
lower than of ADI, decreased simultaneously with the pesticide
Huerta-Fontela, M., Galceran, M.T., Ventura, F., 2011. Occurrence and removal of
concentration along the treatment process. pharmaceuticals and hormones through drinking water treatment. Water Res. 45,
1432–1442.
Ikehata, K., Gamal El-Din, M., 2005. Aqueous pesticide degradation by ozonation and
CRediT authorship contribution statement ozone-based advanced oxidation processes: a review (Part II). Ozone-Sci. Eng. 27,
173–202.
Huiyu Dong: Sampling, Analysis, Writing - review & editing. Lei Xu: Ji, L., Chen, W., Zheng, S., Xu, Z., Zhu, D., 2009. Adsorption of sulfonamide antibiotics to
multiwalled carbon nanotubes. Langmuir 25, 11608–11613.
Sampling, Analysis, Writing. Yuanxiang Mao: Sampling, Analysis. Yan
Jiang, L., Hu, X., Yin, D., Zhang, H., Yu, Z., 2011. Occurrence, distribution and seasonal
Wang: Sampling, Analysis. Shule Duan: Sampling, Analysis. Junfeng variation of antibiotics in the Huangpu River, Shanghai, China. Chemosphere 82,
Lian: Reviewing, Editing. Jin Li: Supervision, Writing - review & edit­ 822–828.
ing. Jianwei Yu: Supervision, Writing - review & editing. Zhimin Karabelas, A., Plakas, K., Solomou, E., Drossou, V., Sarigiannis, D., 2009. Impact of
European legislation on marketed pesticides—a view from the standpoint of health
Qiang: Conceptualization, Supervision, Writing - review & editing, impact assessment studies. Environ. Int. 35, 1096–1107.
Resources. Kim, T.G., Yun, J., Hong, S.-H., Cho, K.-S., 2014. Effects of water temperature and
backwashing on bacterial population and community in a biological activated
carbon process at a water treatment plant. Appl. Microbiol. Biot. 98, 1417–1427.
Declaration of Competing Interest Korotta-Gamage, S.M., Sathasivan, A., 2017. A review: Potential and challenges of
biologically activated carbon to remove natural organic matter in drinking water
purification process. Chemosphere 167, 120–138.
The authors declare that they have no known competing financial Lambert, S., Graham, N., 1995. Removal of non-specific dissolved organic matter from
interests or personal relationships that could have appeared to influence upland potable water supplies—II. Ozonation and adsorption. Water Res 29,
2427–2433.
the work reported in this paper.
Laurent, P., Kihn, A., Andersson, A., Servais, P., 2003. Impact of backwashing on
nitrification in the biological activated carbon filters used in drinking water
Acknowledgment treatment. Environ. Technol. 24, 277–287.
Le Bihan, Y., Lessard, P., 2000. Monitoring biofilter clogging: biochemical characteristics
of the biomass. Water Res 34, 4284–4294.
This work was financially supported by the National Natural Science Li, L., Zhu, W., Zhang, P., Zhang, Q., Zhang, Z., 2006. AC/O3-BAC processes for removing
Foundation of China (51525806, 51878648, 52070184), Chinese refractory and hazardous pollutants in raw water. J. Hazard. Mater. 135, 129–133.
Li, M., 2019. Study on reduction of orrganic matters and bio-toxicity from wastewater
Academy of Sciences (QYZDY-SSW-DQC004) and Ministry of Housing
treatment plant effluent by oxidation-ultrafiltration. Harbin Inst. of Technol.
and Urban-Rural Development of China (2017ZX07207–004). Li, Y., Wang, X., Yang, H., Wang, X., Xie, Y., 2019. Oxidation of isoprothiolane by ozone
and chlorine: reaction kinetics and mechanism. Chemosphere 232, 516–525.
Liu, J., Sun, Q., Zhang, C., Li, H., Song, W., Zhang, N., Jia, X., 2016. Removal of typical
Appendix A. Supporting information antibiotics in the advanced treatment process of productive drinking water. Desalin.
Water Treat. 57, 11386–11391.
Supplementary data associated with this article can be found in the Mazellier, P., Leroy, E., Legube, B., 2002. Photochemical behavior of the fungicide
carbendazim in dilute aqueous solution. J. Photoch. Photobio. A 153, 221–227.
online version at doi:10.1016/j.jhazmat.2020.123986.
Ormad, M., Miguel, N., Claver, A., Matesanz, J., Ovelleiro, J., 2008. Pesticides removal in
the process of drinking water production. Chemosphere 71, 97–106.
References Pan, L., Sun, J., Li, Z., Zhan, Y., Xu, S., Zhu, L., 2018. Organophosphate pesticide in
agricultural soils from the Yangtze River Delta of China: concentration, distribution,
and risk assessment. Environ. Sci. Pollut. R 25, 4–11.
Am Water Works Res, F., Langlais, B., Reckhow, D.A., Brink, D.R., 1991. Ozone in Water
Panadés, R., Ibarz, A., Esplugas, S., 2000. Photodecomposition of carbendazim in
Treatment: Application and Engineering. CRC press,, USA.
aqueous solutions. Water Res. 34, 2951–2954.
Aspelin, A.L., 1994. Pesticides industry sales and usage: 1992 and 1993 market
Papadakis, E.-N., Tsaboula, A., Kotopoulou, A., Kintzikoglou, K., Vryzas, Z.,
estimates. Biological and economic analysis division, office of prevention. Pestic.
Papadopoulou-Mourkidou, E., 2015. Pesticides in the surface waters of Lake Vistonis
Barbash, J.E., Resek, E.A., 1996. Pesticides in Ground Water: Distribution,Trends, and
Basin, Greece: occurrence and environmental risk assessment. Sci. Total Environ.
Governing Factors. Ann Arbor Press, USA.
536, 793–802.
Benitez, F.J., Acero, J.L., Real, F.J., Garcia, C., 2009. Removal of phenyl-urea herbicides
Plakas, K.V., Karabelas, A.J., 2012. Removal of pesticides from water by NF and RO
in ultrapure water by ultrafiltration and nanofiltration processes. Water Res 43,
membranes—a review. Desalination 287, 255–265.
267–276.
Rahman, S., 2013. Pesticide consumption and productivity and the potential of IPM in
Cai, S., Sun, K., Dong, S., Wang, Y.M., Wang, S., Jia, L., 2014. Assessment of
Bangladesh. Sci. Total Environ. 445, 48–56.
organochlorine pesticide residues in water, sediment, and fish of the Songhua River,
China. Environ. Forensics 15, 352–357.

8
 Journal of Water Process Engineering 55 (2023) 104165

Contents lists available at ScienceDirect

Journal of Water Process Engineering

journal homepage: www.elsevier.com/locate/jwpe

A combined treatment of aerobic activated sludge and powdered activated

carbon: Pilot-scale study of per/polyfluoroalkyls (PFASs), organic matters,
chromium, and color removal from tannery wastewaters
Marco Gottardo *, Giulia Adele Tuci , Aditi Chitharanjan Parmar , Paolo Pavan ,
Francesco Valentino
Department of Environmental Sciences, Informatics and Statistics, Ca’ Foscari University of Venice, Via Torino 155, 30172 Mestre-Venice, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: A pilot-scale system of three parallel plug flow reactors was utilized for the treatment of tannery wastewaters
Tannery wastewater with powdered activated carbon (PAC) at different dosage in activated sludge process. The SRT and the carbon
Per- and polyfluorinated alkylated substances replacement (XCI) were set at 35 d and 0 mg PAC/L (R1), 25 d and 100 mg PAC/L (R2), 20 d and 300 mg PAC/L
(PFASs)
(R3), at fixed HRT (4 days) and temperature (20–22 ◦ C). The removal of per- and polyfluoroalkyls substances
Chromium
Powdered activated carbon (PAC)
(PFASs), organics, chromium and color were assessed. Due to competition phenomena for adsorption between
the organics and the PFASs, the highest PAC dosage (R3) was the only one able to decrease the concentration of
total PFASs below the limit of 500 ng/L. Soluble COD and chromium removal efficiency up to 92.8 % and 68 %
was respectively achieved; in terms of color abatement, the treated wastewaters showed a reduced absorbance up
to 56 % compared to condition R1. The mass balance assessment revealed a maximum cost's increase of 0.923
€/m3 of treated water associated to the use of the PAC (R3), highlighting the necessity to adopt sustainable
strategies to valorize the generated tannery sludge instead of landfilling, which is still the only disposal practice
of this waste.

1. Introduction also bioaccumulate in humans, animals, and plants and, among the few
that are well-studied, most are considered toxic [3]. Some PFASs are
The leather tanning is a manufacturing process that requires a high classified as Persistent, Bioaccumulative and Toxic (PBT) and very
use of water and produces wastewater rich in pollutants [1]; therefore, it Persistent and very Bioaccumulative (vPvB) under the regulation of the
must be properly treated before being discharged into the receiving Registration, Evaluation, Authorisation of Chemicals (REACH legisla­
water bodies. Although conventional purification techniques are very tion [4]). The European Environment Agency lists the main effects of
effective for most of the pollutants present in tannery wastewater, for PFASs on human health, which include thyroid disease, increased
some classes of pollutants they are not very efficient. The main con­ cholesterol levels, effects on reproduction and fertility, immunotoxicity,
taminants belonging to these classes are poly- and perfluoroalkyl com­ liver damage, kidney, and testicular cancer.
pounds (PFASs), complexed trivalent chromium, and organic dyes. The Directive 98/83/EC was recently reviewed and standards for
The PFAS compounds represent a class of chemicals of industrial PFASs in drinking water were introduced; the new Directive 2020/2184
origin characterized by a linear or branched carbon chain, partially or includes in Annex I the limit value of 100 ng/L for a sum of 20 individual
completely fluorinated. These compounds have found use in many in­ PFASs listed in Annex III, as well as the limit value of 500 ng/L for total
dustrial sectors including textiles and tanneries [2]. Several PFASs can PFASs concentration [5]. As highlighted by several authors [6–8],

Abbreviations: Activated Carbon, (AC); analysis of variance, (ANOVA); biochemical oxygen demand, (BOD); chemical oxygen demand, (COD); color unit, (CU);
European Union, (EU); extracellular polymeric substances, (EPS); hydraulic retention time, (HRT); membrane bioreactor, (MBR); micropollutants, (MPs); mixed
liquor suspended solids, (MLSS); mixed liquor volatile suspended solids, (MLVSS); natural gas, (NG); per- and polyfluorinated alkylated substances, (PFASs); per­
fluorobutanesulfonate, (PFBS); perfluorobutyrate, (PFBA); perfluorooctanesulfonate, (PFOS); perfluorooctanoate, (PFOA); powdered Activated Carbon, (PAC); sludge
retention time, (SRT); soluble COD, (CODSOL); total nitrogen, (TN); total solids, (TS); volatile solids, (VS); wastewater treatment plant, (WWTP).
* Corresponding author.
E-mail address: [email protected] (M. Gottardo).

https://doi.org/10.1016/j.jwpe.2023.104165
Received 15 May 2023; Received in revised form 27 July 2023; Accepted 11 August 2023
Available online 17 August 2023
2214-7144/© 2023 Elsevier Ltd. All rights reserved.
M. Gottardo et al. Journal of Water Process Engineering 55 (2023) 104165

conventional purification treatments are completely ineffective in studies highlighted how the PAC utilization can strongly support the
removing PFASs from wastewater; moreover, PFASs compounds may be biological processes in the removal of slowly biodegradable (or even
formed through the metabolization of polyfluoroalkyl compounds (e.g., recalcitrant) microcontaminants, and in the protection of the biomass
fluorotelomer alcohols; FTOHs) and perfluoroalkane sulfonamides from potentially toxic compounds, favoring the flock's formation and
(FASAs) in the aerobic biological process [9]. improving the sludge settling properties [22–24]. The pilot-scale pro­
Among the unconventional treatments, the process that has shown posed in this study is aimed at demonstrating the technical feasibility of
greater feasibility for the treatment of PFASs is the adsorption on acti­ the combination of biologically activated sludge process integrated with
vated carbon (AC). As demonstrated in several studies, AC can adsorb a the use of powdered activated carbon (PAC) for the treatment of real
wide range of organic contaminants of synthetic and natural origin, tannery wastewater coming from different tannery industries to remove
including PFASs compounds [10]. However, these studies were mainly the PFAS compounds, complexed chromium as [Cr(III)], and organic
based on the purification treatment of surface or groundwater using AC matter quantified as chemical oxygen demand (COD). The color abate­
filters and highlighted how the lifetime of the AC did not depend ment has been also evaluated as well as the economic outcomes of the
exclusively on the concentration of the PFASs in the influent but also on approach.
the concentration of organic material in the treated water. Then, PFASs
compounds adsorbed on AC carbon can be destroyed by combustion at 2. Materials and methods
temperatures above 600 ◦ C [11,12]. Recently, an efficient mineraliza­
tion to fluoride ions (higher than 80 %) of some PFASs such as per­ 2.1. Pilot units' description and operating conditions
fluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) has been
obtained on granular AC at temperature ≥ 700 ◦ C, with an almost Three pilot lines operating in parallel, each consisting of one plug
complete decomposition (>99.9 %) for both compounds [13]. On the flow reactor (PFR) of 2.0 m3 working volume, were aerobically main­
other hand, the use of advanced oxidative processes has proved to be tained, and connected to a settler for the biomass recirculation (Fig. 1).
ineffective in destroying perfluoroalkyl compounds [14,15]. The dissolved oxygen (DO) was maintained in the range of 3.1–3.7 mg
Tannery industry faces also with the environmental concerns related O2/L. The three lines were inoculated with the same biological sludge,
to the use of chromium. At present times, over 90 % of leather products taken from the wastewater treatment plant (WWTP) of Montebello
are tanned using trivalent chromium [Cr(III)] salts, due to their excellent Vicentino (northeast Italy). This WWTP treats around 10,000 m3/d of
hydrothermal stability, physical and mechanical properties. The industrial wastewaters produced by >20 tannery plants. The real
chromium-related tanning mechanism is mainly based on coordination wastewater taken from the above cited WWTP was used as substrate for
complexation between [Cr(III)] and deprotonated carboxyl groups of the pilot units. The duration of the tests was 90 days, and, for the entire
skin collagen, under acidic (pH ~ 4.0) conditions, which generate the period, the temperature of the pilot units was in the range of 20–22 ◦ C.
chemical cross-linking of collagen fibers and the conversion of rawhide The following Table 1 shows the employed operating conditions in the
to stable leather. In addition, the masking agents (among others citrate three pilot scale reactors (R1, R2 and R3) as main component of the
and acetate) added during tanning to favor the permeation of [Cr(III)] water line required for the tannery wastewater treatment. All the aera­
into collagen fibers, can also form complexes with [Cr(III)]. Generally, tion tanks were maintained with a hydraulic retention time (HRT) of 4
these structures are difficult to be removed because of their high days and with a sludge retention time (SRT) in the range 20–35 days. In
complexation affinity and high stability over a wide pH range [16]. In the three reactors, the food-to-microorganisms (F/M) ratio was equally
the conventional chromium tanning process, only 70 % of the total maintained at 0.09 kg COD/(kg VS d) (volatile solids) and 0.02 kg TN/
chromium agent (approximately) is absorbed by the leather, leading to (kg VS d) as total nitrogen. The XPAC and XCI were set in the range
the accumulation of [Cr(III)] in tannery wastewater to concentrations 0–1.50 g PAC/L and 0–300 mg PAC/L respectively, where XPAC was the
higher than 3000 ppm [17]. The main chromium removal process in
wastewater treatment is precipitation, also chemically supported, as a
simple and economically inexpensive process. However, it is commonly
observed that in the treatment of tannery wastewater, at the end of the
chromium sedimentation, some [Cr(III)] is still present in the solution as
it is bound to organic material ([Cr(III)]-complexes), such as phenolic-
type compounds [18,19].
Besides the [Cr(III)] pollutant in the chromium tanning process,
various cationic and anionic organic dyes that can combine with skin
collagen through electrostatic interactions are also widely used in the
subsequent dyeing process during leather-making. In addition, most of
these organic dyes are aromatic compounds (harmful to humans but also
toxic to microorganisms), another source of pollution in the tannery
wastewaters. Moreover, the presence of such components (both of
organic and inorganic nature) gives to the tannery wastewater an
aesthetically displeasing color, which can heavily damage the receiving
water bodies [20]. Indeed, tannery wastewaters containing [Cr(III)] and
organic dyes have become a serious environmental pollution issue,
which has severely affected the sustainable development of the leather
industry [17].
A recent review reported some examples in which the adsorption
process with AC can be an effective technique for the removal of organic
dyes from aqueous solution, even though no information about the
competitive adsorption between the organic dyes and the natural
organic matter of a real wastewater are available [16]. In addition,
studies on the adsorption of [Cr(III)]-complexes with AC in tannery
wastewater are very limited [21]. Fig. 1. Simplified scheme of the pilot units installed at the Montebello Vice­
Regarding the use of the powdered activated carbon (PAC), several ntino WWTP.

2
M. Gottardo et al. Journal of Water Process Engineering 55 (2023) 104165

Table 1 “Color Unit” (CU). The analysis of the soluble COD (CODSOL) was carried
Operating conditions of three pilot-scale water lines adopted for the tannery out on filtered samples by using a 0.2 μm porosity filters. The analysis of
wastewater treatment. the complexed [Cr(III)] was performed on a sample previously basified,
Parameter Units PFR centrifuged, and filtered with a 0.2 μm porosity filters. The PFASs were
R1 R2 R3
analyzed according to the ASTM 7979-20 method (taking into account
the 50 % dilution with methanol) with an ultrahigh performance liquid
HRT d 4 4 4
chromatography (UHPLC, Ultimate 3000 Thermo Fisher Scientific)
SRT d 35 25 20
F/M kg COD/(kg VS d) 0.09 0.09 0.09 system coupled to a high-resolution mass spectrometer (Q-Orbitrap,
kg TN/(kg VS d) 0.02 0.02 0.02 Thermo Fisher Scientific). Each sample has been taken in polypropylene
DO mg/L 3.1–3.7 3.1–3.7 3.1–3.7 vessels, avoiding any contacts (from the sampling to the final analysis)
XPAC g PAC/L 0 0.63 1.50 with glass and/or polytetrafluoroethylene materials. The entire class of
XCI mg PAC/L 0 100 300
the measured PFAS compounds is shown in Table 3.

PAC concentration in the tank (g/L) as defined in the Eq. (3), and XCI was 2.4. Calculation and statistical analysis
the PAC concentration dosage [25]. The SRT values have been chosen to
prevent the uncontrolled increase of mixed liquor suspended solids The soluble organic dyes together with other undefined organic
(MLSS) concentration in the aerobic tanks, avoiding difficulties in the substances were quantified as soluble COD (CODSOL). The CODSOL
operation of the pilot plant and also inapplicable conditions for a scale- removal efficiency was quantified as percentage, by considering the
up perspective. mass flow rates (kg COD/d) of the influent (FIN)COD and the effluent
(FOUT)COD, according to the following formula:
2.2. Tannery wastewaters and PAC characteristics
(FIN)COD − (FOUT)COD
COD removal (%) = (1)
(FIN)COD
Tannery wastewaters came from different tannery industries located
close to the WWTP of Montebello Vicentino (northeast Italy). In the full- The complexed [Cr(III)] removal efficiency was quantified as per­
scale plant, the wastewaters are firstly subjected to physicochemical centage, by considering the mass flow rates (kg COD/d) of the influent
primary treatment (a settling supported by a coagulation-flocculation (FIN)Cr and the effluent (FOUT)Cr, according to the following formula:
process with aluminum polychloride); a fraction of the effluent (40 %
(FIN)Cr − (FOUT)Cr
v/v) is pre-oxidized before the secondary biological treatment (anoxic/ [Cr(III) ] removal (%) = (2)
(FIN)Cr
aerobic) for the biochemical oxygen demand (BOD) and nitrogen
removal [26]. For the pre-oxidation step, a PFR is utilized under the The PAC load was calculated according to the formula reported
following operating conditions: HRT 1 d, SRT 12–15 d, F/M 0.20–0.24 elsewhere [25], as it follows:
kg COD/(kg VS d), DO <2.0 mg O2/L. The remaining 60 % of the
XCI*SRT
physicochemical primary treatment effluent is mixed with the pre- XPAC = (3)
HRT
oxidizing PFR effluent; this effluents' mixture has been utilized for the
scopes of this work and it represents the real wastewaters usually sent to where XPAC (mg PAC/L) and XCI (mg PAC/L) have been defined above.
the full-scale biological treatment. It is important to highlight that the For statistics, the analysis of variance (ANOVA) and Tukey HSD post
pre-oxidation step has been conceived to partially remove biodegrad­ hoc tests were conducted to detect the significance of the results.
able COD and, in turn, to reduce the OLR in the full-scale biological Shapiro-Wilk test was carried out to verify if the considered data were
water line. characterized by a normal distribution. Bartlett test was used to verify
For the pilot scale study, the tannery wastewater was collected after the homoscedasticity between the different groups of considered data.
the physicochemical primary and the pre-oxidation treatment in a The null hypothesis (H0) was rejected when the calculated p-values were
dedicated tank equipped with a mechanical stirrer, and characterized in equal to or less than the level of significance (α). Statistical analyses
terms of pH, MLSS, soluble chemical oxygen demand (CODSOL), con­ were performed using the open-source program, R (The R Foundation
ductivity, chlorides, sulphates, PFAS compounds and complexed [Cr for Statistical Computing, version 4.2.2).
(III)]. The PAC used in the experiment (ACTISORB M.J.P. 120) was a
thermally activated carbon of mineral origin; its physical and chemical 2.5. Energy-mass balance and economic considerations
features are summarized in Table 2.
A hypothetical full-scale plant with a volumetric flow rate (F) of
2.3. Analytical methods 1000 m3/d of tannery wastewaters was considered. The three conditions

The determination of all parameters was performed according to the Table 3

Standard methods [27], with the exception of the color parameter, for PFAS compounds analyzed in the tannery wastewaters.
which the spectrophotometric analysis was performed following the EN
Compound CAS Compound CAS
ISO 7887 method (wavelength 410 nm) and the value was expressed as a
PFOS 1763-23-1 PF 3,7 DMOA 172155-07-6
PFOA 335-67-1 4:2 FTS 757124-72-4
Table 2 PFBA 375-22-4 6:2 FTS 27619-97-2
Physical and chemical characteristics of the PAC utilized in this study (ACTI­ PFBS 375-73-5 8:2 FTS 39108-34-4
SORB M.J.P. 120). PFPeA 2706-90-3 4:2 FTOH 2043-47-2
PFHxA 307-24-4 6:2 FTOH 647-42-7
Parameters Units Value
PFHxS 355-46-4 8:2 FTOH 678-39-7
Humidity % <5 % PFHpA 375-85-9 PFOSA 754-91-6
Ashes % 12 % PFNA 375-95-1 N-MeFOSA 31506-32-8
Apparent density kg/m3 400 ± 20 PFDeA 335-76-2 N-EtFOSA 4151-50-2
Specific surface m2/g 1100 PFUnA 2058-94-8 N-MeFOSE 24448-09-7
Iodine number mg/g 1000 PFDoA 307-55-1 N-EtFOSE 1691-99-2
Methylene blue mg/g 250 PFHpS 375-92-8 Me-FOSAA 2355-31-9
pH – Alkaline HFPO – DA (GenX) 13252-13-6 Et-FOSAA 2991-50-6

3
M. Gottardo et al. Journal of Water Process Engineering 55 (2023) 104165

have been compared in terms of produced sludge and required energy PFAS compounds (e.g., PFOS and PFOA), some possible alternatives
for its drying, also according to the specific PAC dosage in R2 and R3. have been identified, including the use of short-chain PFASs and, in
The total solids (TS) level of the produced sludge was assumed to be 25 particular, the use of chemical compounds based on PFBS as a replace­
% and 80 % w/w as inlet and outlet of the drying unit. Natural gas (NG) ment for PFOS-based chemicals in the textile and tanning industry [31].
was considered as energy source of the drying unit: for its quantification This fact can be tricky if the decontamination efficiency in terms of
(as standard cubic meter, Sm3), the lower heating value and the thermal PFASs removal from wastewaters is considered; in fact, several studies
energy yield were assumed to be equal to 35 MJ/Sm3 and 0.9 respec­ reported in the review of Rahman et al. (2014) [31] have shown that the
tively. The energy required for water evaporation was considered equal adsorption potential on activated carbon (both granular and powdered)
to 3.35 MJ per kg of water (MJ/kgwater). For the economic consider­ of PFAS compounds decreased with the decreasing of the carbon chain
ations, a cost of 1.90 €/kg of the PAC was considered [28]. In addition, length, due to the negative effect of decreasing hydrophobicity with
the cost associated to the use of NG and to the sludge landfilling, as the decreased carbon chain length on adsorption.
only current disposal practice (according to Italian legislation, tannery The following Fig. 2A shows the PFASs concentration (ng/L) as sum
sludge is classified as “special non-hazardous residue” and its destina­ of each PFAS quantified in the effluent of the three reactors (R1, R2 and
tion is to a second-class type B controlled landfill; [26]) have been R3) in the 90 days of experimentation, compared with the two limit
considered: the price of the NG and the disposal cost of the dried sludge values of 500 and 100 ng/L, according to the drinking water directive by
(80 % TS level) were assumed to be 0.90 €/m3 [29] and 230 €/ton [1], European Commission [5]. In the first condition with no PAC addition
respectively. (R1), the total PFASs concentration did not show remarkable fluctua­
tions, being most of the data in the range 600–700 ng/L (667 ± 45 ng/L,
3. Results and discussions on average). These values were similar to the total PFASs concentration
in the influent (675 ± 42 ng/L, on average); it is reasonable to assume
3.1. Tannery wastewater characteristics and PFASs removal that such similarity was a consequence of the inability of the aerobic
biomass for PFASs degradation. In fact, no data of PFASs concentration
Table 4 summarizes the main chemical features of the tannery out of R1 were below the higher limit of 500 ng/L. It is known from the
wastewaters utilized as inlet, after the physicochemical primary treat­ literature that the reductive defluorination can be thermodynamically
ment and sedimentation full-scale units; effluents' characteristics, 30 favorable and the released energy may allow the microbial community
min sludge volume index (SVI) and related thickened total solids' con­ to thrive; however, the PFAS compounds do not seem to be commonly
centration XR (g/L) are also furnished. These wastewaters were char­ used as a carbon source [33]. A certain removal of both PFOS and PFOA
acterized for a wide range of time, covering all the period of the pilot from wastewater in a lab-scale closed-loop anaerobic bioreactor was
scale activities. Regarding the PFAS compounds, the most abundant observed; however, no biodegradation phenomena were detected under
were represented by the PFOS and PFOA, followed by perfluorobutyrate aerobic conditions [34]. The PAC dosage corresponding to 0.63 g/L as
(PFBA) and perfluorobutanesulfonate (PFBS) (Table 4). The total con­ XPAC (R2) did not substantially improve the performance for PFASs
centration of the PFAS compounds was equal to 675 ± 45 ng/L; among removal; despite of some short fluctuations, the total PFASs concentra­
others, PFBS represented the 75 % (roughly) of the total PFASs, with a tion was in the same range observed in R1 (the average value was 664 ±
concentration equal to 509 ± 43 ng/L. 44 ng/L). A remarkable decrease of the total PFASs concentration was
This high difference in concentration between the PFBS and all the instead observed in the condition R3, performed at the highest PAC
other PFASs can be explained as an effect of the introduction of new dosage (1.5 g/L as XPAC). The PFASs trend was quite stable, exhibiting
PFASs congener among the compounds listed in Annex B of the Stock­ concentration values often in the range 400–500 ng/L (415 ± 40 ng/L
holm Convention (2001) [30], which aimed at the gradual decrease and on average).
progressive elimination of the use of some substances harmful to human In general, none of the three tested conditions involved in a reduc­
health and the environment. In fact, as an alternative to the long-chain tion of the total PFASs concentration sufficiently high to achieve a level

Table 4
Main characteristics (with average data and standard deviations) of the tannery wastewaters utilized in the pilot scale facilities after the primary treatment and pre-
oxidation stage (inlet) and after the system PFR + Settler (outlet).
Parameter Units n. of analysesa Inlet Outlet

R1 R2 R3

pH – 35 8.0 ± 0.4 7.8 ± 0.3 7.9 ± 0.5 7.9 ± 0.1

Mixed Liquor Suspended Solids (MLSS) mg/L 20 454 ± 71 73 ± 16 44 ± 14 38 ± 14
Soluble COD (CODSOL) mg/L 88 1048 ± 90 186 ± 18 159 ± 11 75 ± 6
Total Nitrogen (TN) mg/L 25 196 ± 14 150 ± 13 149 ± 13 147 ± 12
Conductivity μS/cm 25 9606 ± 880 8475 ± 742 8537 ± 664 8721 ± 721
Chlorides mg/L 25 1538 ± 282 1488 ± 173 1503 ± 221 1435 ± 192
Sulphates mg/L 25 1245 ± 232 1232 ± 195 1216 ± 184 1239 ± 177
Total per- and polyfluoroalkyl compounds (PFASs) ng/L 88 675 ± 45 664 ± 47 658 ± 45 412 ± 39
Perfluorooctanesulfonates (PFOS) ng/L 88 17 ± 4 16 ± 4 16 ± 3 4±2
Perfluorooctanoates (PFOA) ng/L 88 25 ± 7 25 ± 5 23 ± 6 9±5
Perfluorobutyrates (PFBA) ng/L 88 18 ± 2 17 ± 4 19 ± 4 19 ± 3
Perfluorobutanesulfonates (PFBS) ng/L 88 509 ± 43 507 ± 40 495 ± 43 340 ± 37
Other PFASs ng/L 88 106 ± 12 99 ± 17 105 ± 21 41 ± 20
Complexed [Cr(III)] mg/L 88 0.94 ± 0.14 0.49 ± 0.15 0.44 ± 0.13 0.30 ± 0.09
Color (410 nm) CU 35 – 292 ± 39 250 ± 34 129 ± 26
Mixed Liquor Suspended Solids (MLSS)b mg/L 20 – 4002 ± 146 4602 ± 177 5500 ± 152
Mixed Liquor Volatile Suspended Solids (MLVSS)b 20 – 3082 ± 120 3684 ± 148 4561 ± 127
SVI30c mL/g 15 – 160 ± 5 119 ± 4 86 ± 5
(TS)Rc g/L 20 – 6.3 ± 0.4 8.4 ± 0.1 11.2 ± 0.3
a
Analyses related to the inlet and to the effluent in each operating condition;
b
Data related to samples collected in the aerobic tanks;
c
Data related to the settled sludge stream, recirculated in the aerobic tanks.

4
M. Gottardo et al. Journal of Water Process Engineering 55 (2023) 104165

Fig. 2. (A) PFASs concentration values (as sum of each PFASs) of the effluent from the three reactors (R1, R2, R3), for the whole experimentation period (days 1–50).
The two limits (100 and 500 ng/L) are also indicated. (B) Average values (with standard deviation) of the daily flow rates (g PFAS/d) of the total PFAS in the influent
(PFASin), effluents (PFASout) and related equilibrium PAC content (XPAC).

below 100 ng/L; however, with the PAC dosage corresponding to the respectively (Fig. 3A and B). As regards the short-chain congeners,
XPAC of 1.5 g/L (R3), the concentration of total PFASs in the effluent was namely PFBS and PFBA, a considerable removal (32 %) was only
below 500 ng/L for the whole experimentation period. observed for the PFBS congener (Fig. 3C and D). Regarding the PFBA, no
Fig. 2B shows the average value (with standard deviation) of the significance difference was found between the inlet and outlet flow
daily flow rates (g PFAS/d) of the total PFASs detected in the influent values (ANOVA test, Tukey HSD post hoc test, α = 0.05). These data
(equal for the three reactors) and in the effluents of the three reactors. were in line with data reported in the literature, where a lower efficiency
From the statistical analysis, a significant abatement efficiency is found of short chain perfluorinated compounds adsorption on AC (in particular
exclusively in R3 (1.5 g/L as XPAC) while both R1 and R2 effluents (0 and the PFBA congener) was usually reported [32,35]. In addition, the major
0.6 g/L as XPAC respectively) the calculated inlet and outlet flow values removal efficiency observed for the PFBS congener, compared to PFBA,
were significantly similar (ANOVA test, Tukey HSD post hoc test, α = could be a consequence of the different hydrophobic grade of the two
0.05). compounds; in fact, PFBA has a lower partition coefficient (logKOW)
Analyzing the reductions for individual congeners, the main conge­ compared to PFBS (− 0.66 and 0.03 respectively; [36]). This character­
ners removed in the reactor R3 were the two longer-chain compounds istic could be the basis of the greater affinity between the PAC and the
PFOS and PFOA, with an average removal efficiency of 75 % and 68 % PFBS congener, and its higher adsorption compared to the PFBA.

5
M. Gottardo et al. Journal of Water Process Engineering 55 (2023) 104165

Fig. 3. Average values (with standard deviation) of the daily flow rates (g/d) of the PFOS (A), PFOA (B), PFBS (C), PFBA (D) congeners and other PFASs (E) in the
influent (in) and in the effluents (out) with related equilibrium PAC content (XPAC).

The input and output flow for each congener were significantly 3.2. Removal of the organic matter as COD
similar in the case of zero (R1) and intermediate (R2; XPAC 0.6 g/L) PAC
dosage. This fact was in agreement with the well-documented inability As confirmation of the competition phenomena between the organics
to remove the PFAS compounds by employing a traditional activated and the PFAS compounds for the PAC adsorption, the following Fig. 4A
sludge biological process compared to the PFAS removal efficiency from shows the trends of the flow rate of the COD influent and the flow rates
water samples (sorption and stabilization, thermal destruction, of the effluent soluble COD (from the three reactors). A substantial
advanced oxidation [37,38]). abatement was obtained even without the use of PAC (R1), from 0.52 ±
A significant abatement was still not detected in the second case 0.05 to 0.093 ± 0.008 kg CODSOL/d. This was due to the utilization of a
(R2). This suggested that the PAC dosage (corresponding to XPAC 0.6 g/ culture already acclimated to the same tannery wastewater utilized in
L) was excessively low, considering that the tannery wastewater con­ this study. The increase of the PAC dosage improved the reactors' per­
tained, in addition to the PFAS compounds, other components which formances since the mass flow rate was further reduced to 0.078 ±
were most likely competitive in the PAC adsorption process with PFAS 0.008 and 0.037 ± 0.005 kg CODSOL/d in R2 and R3 respectively.
compounds (mainly organics, [32,39]). Hence, the use of PAC significantly increased the COD removal effi­
ciency compared to the performance obtained from the biological pro­
cess alone, by roughly 3 % with an XPAC of 0.6 g/L (R2) and 10 % with an

6
M. Gottardo et al. Journal of Water Process Engineering 55 (2023) 104165

phenolic compounds, sulfur, solvents, and chloride. The authors ob­

tained COD and phenol removal efficiencies of 98 % and 99 % respec­
tively, with a process characterized by a higher stability in the
performances compared to the control reactor process (activated sludge
without PAC). To the best authors' knowledge, this is the first example
describing the treatment of tannery wastewater with a simultaneous
approach as depicted in Campos et al. [42]. As described for the oil-
refinery wastewaters, also tannery wastewaters are generated by com­
plex processes which in turn generate a diluted mixture of different
organic compounds. Hence, the PAC addition in tannery wastewaters
may conferred several advantages to the process, among others, stability
in COD removal during any eventual load shocks, reduction or elimi­
nation of biological inhibition phenomena, improvement in sludge
settling, removal of color and chromium as discussed below.
These positive effects are due to the additional organic biodegrada­
tion because of the adsorption phenomena of the slowly biodegradable
substances onto PAC (a mechanism which causes a decrease in toxicity
due to the possible presence of recalcitrant compounds) [21]. In other
words, the presence of PAC increased the time of exposure between the
adsorbed substances (which remained in the system according to the
chosen SRT) and the activated biomass. In the absence of PAC, these
substances remained in the system for the chosen hydraulic retention
time of 4 days only. For these reasons, the COD removal efficiency
increased from 85.1 ± 0.5 % to 93 ± 1 %, in R1 and R3 effluents
respectively.

3.3. [Cr(III)] complexes removal and color abatement

The dosage of PAC showed an improvement in efficiency also for the

treatment of complexed chromium [Cr(III)] and the reduction of color,
compared to the results obtainable exclusively from the activated sludge
biological process. Regarding the fate of the complexed [Cr(III)], Fig. 5A
shows the mass flow rate of the [Cr(III)] influent and the mass flow rates
Fig. 4. Trend of the mass flow rates of the influent COD, and of the effluent of the [Cr(III)] effluents from R1, R2 and R3. The high variability as well
COD from the three reactors (A); average values of the COD removal efficiency
as the values of the influent mass flow rate (0.25–0.69 kg [Cr(III)]/d)
with related equilibrium PAC content (XPAC) (B).
were substantially reduced in the aerobic system, also when the PAC was
not dosed. In addition, the obtained data showed that by operating with
XPAC of 1.5 g/L (Fig. 4B; ANOVA test, Tukey HSD post hoc test, α = the commonly used activated sludge process, a removal efficiency of the
0.05). At the highest PAC dosage (R3), the soluble COD removal effi­ complexed [Cr(III)] of 48.4 ± 0.7 % was achieved (Fig. 5B). The [Cr(III)]
ciency was equal to 93 ± 1 %. removal in R1 could be due to a two-steps mechanism which was
Despite of the benefits of PAC utilization in the activated sludge characterized by the first [Cr(III)] release because of the degradation of
wastewater treatment (high quality of the treated wastewater, totally the [Cr(III)]-linked organic material, followed by the [Cr(III)] precipi­
reusable and with stable characteristics [40]), in most cases, its addition tation in form of hydroxide or linked with the extracellular polymeric
has been limited to the tertiary treatments and few information are substances (EPS) [16]. After precipitation or settling, the [Cr(III)] can be
currently available on simultaneous biological oxidation and physical removed from the water line with the excess sludge. A little improve­
adsorption processes. On the contrary, much information are available ment in the reduction of the [Cr(III)] flow rate values was obtained by
on the removal efficiencies of a wide spectrum of micropollutants (MPs) using the PAC at the lower dosage tested (R2); The related removal ef­
in biological membrane bioreactor (MBR) coupled with PAC (directly ficiency was equal to 53.6 ± 0.2 % (Fig. 5B).
added in the bioreactor or used in a separate unit as post treatment). A On the other hand, the abatement efficiency of complexed [Cr(III)]
recent review showed the effective contribution of PAC in advanced increased significantly with the highest dosage of PAC in the oxidation
biological wastewater treatment in enhancing the MPs removal and the tank (R3) (ANOVA test, Tukey HSD post hoc test, α = 0.05); the average
MBR performances (mitigation in membrane fouling and improvement mass flow rate was 0.15 ± 0.04 kg [Cr(III)]/d, and the related removal
in sludge characteristics among others) [41]. The MPs nature, as well as efficiency was equal to 68 ± 2 % on average (with maximum peak of 72
the PAC characteristics and the presence of dissolved organic matter in %).
the wastewater were the main aspects for the definition of MPs removal The literature is not well furnished of examples on the adsorption [Cr
efficiencies. In addition, most of the available data showed a higher (III)] complexes; apart from the physical adsorption phenomena, the
variability in MPs removal when PAC was utilized in post-treatment only more often reported adsorption mechanisms involve the active sites of
(in comparison to when it was added in the MBR). Apart from the chosen adsorbents or the ion exchange of both free organic ligands and/or [Cr
configuration, a loss in PAC potential adsorption capacity occurred with (III)] species [43,44]. Complexes concentration and pH heavily affected
its progressive saturation and its replacement was always necessary to the adsorption efficiency [16]; for the case investigated in this study, the
maintain the desired MPs removal capacity. pH was similar in all the three reactors and close to the average value of
Apart from MBR configuration, a representative example was made the influent wastewaters (8.0 ± 0.4). Reasonably, the increase in PAC
by Campos et al. [42]; the author described the use of PAC in the acti­ dosage increased the concentration of the adsorption sites, leading to a
vated sludge process for the treatment of oil-refinery wastewater, which higher [Cr(III)] removal (in addition to PFASs and COD abatement as
contained different types of complex and potentially recalcitrant com­ described above). To the best authors' knowledge, there are no examples
pounds such as aliphatic and aromatic hydrocarbons, mercaptans, in the literature describing the use of PAC for the treatment of [Cr(III)]

7
M. Gottardo et al. Journal of Water Process Engineering 55 (2023) 104165

Fig. 6. Average values of treated tannery wastewater color (in CU) with related
equilibrium PAC content (XPAC).

cost associated to the treatment of tannery wastewaters: the two main

factors that have the major contribution are the cost of the virgin PAC
and the increased disposal cost deriving from the greater production of
sludge. The importance of the second factor depends on the dosage of
PAC as well as the TS level achievable in the final sludge to be disposed
of. This level can be increased up to 80 % w/w (roughly), as depicted in
this example with the operation of a drying unit. The disposal costs of
the tannery sludge are generally high, even in comparison with
municipal sewage sludge; currently, due to the presence of chromium,
the only disposal practice is exclusively the landfill [26]. It has to be
underlined that the sludge settling has been strongly improved with the
increasing PAC dosage, probably reducing the concentration of fila­
mentous organisms in the sludge flocks [42]: the 30 min sludge volume
index (SVI30) was 160 ± 5, 119 ± 4 and 86 ± 5 at the corresponding
Fig. 5. Trend of the mass flow rates of the influent [Cr(III)], and of the effluent XPAC value of 0.0, 0.63 and 1.50 g/L respectively (Fig. 7).
[Cr(III)] from the three reactors (A); average values of the [Cr(III)] removal This led to an increase of the TS level in the settled sludge up to 11.2
efficiency with related equilibrium PAC content (XPAC) (B).
g TSR/L (R3) and, in turn, to a substantial reduction of both recirculated
flowrate (L/d) and required energy to maintain the recirculation typi­
complexes in tannery wastewaters, as well as the PAC addition for the cally adopted in the water line of a suspended biomass based WWTP.
simultaneous biological treatment and [Cr(III)] adsorption processes. The quantification of the economic benefit due to this improvement was
Only one previous example investigated the removal of dyes-[Cr(III)] out of the scope of this work.
complexes from aqueous solutions using AC and Macronet MN200, a Considering a PAC cost of 1.90 €/kg and the operating conditions
hyper-crosslinked neutral polymer. The removal efficiency of the com­ adopted (Table 1), the cost of PAC per m3 of treated water is equal to
plexes was higher in the acidic solution, due to the positive effect of the 0.190 €/m3 and 0.570 €/m3, respectively in R2 and R3 configurations.
pH value on the surface charge of the adsorbent; in addition, the process The cost of mechanical sludge dewatering was not considered, being
removal was not effective at higher dyes-[Cr(III)] complex concentra­ negligible compared to the cost of the drying process. The inlet of the
tion due to the limited available adsorption sites [45]. dewatering unit (FW-TSR; Fig. 8) consisted in a settled sludge at different
Regarding color, it is known from the literature that this parameter is TS level (0.6–1.1 % w/w) according to the XPAC value. The sludge
one of the most important indicators of the wastewater's pollution and,
for tannery wastewater, it is usually not enough reduced through the
conventional activated sludge treatment technologies [20]. The PAC
dosage showed a significant increase in the transmittance of the treated
wastewaters (effluents) from R2 and R3 reactors, compared to the
effluent obtained in R1 reactor (without the PAC dosage). Fig. 6 shows
the average absorbance values from the three treated tannery effluents.
Therefore, the efficiency in color removal strongly increased with the
increase in PAC dosage (R3), due to the adsorption phenomena and the
consequent organics and [Cr(III)]-complexes removal. Hence, the use of
the PAC at the two chosen dosages showed a significant improvement in
the quality of the wastewater, the average CU values detected in the R2
and R3 effluents (250 ± 34 and 129 ± 26 respectively) were signifi­
cantly lower than average value of R1 effluent (295 ± 39, ANOVA test,
Tukey HSD post hoc test, α = 0.05).

3.4. Economic aspects in full-scale application

It is out of any doubt that the use of PAC has a certain impact on the Fig. 7. Trend of the SVI30 as a function of XPAC.

8
M. Gottardo et al. Journal of Water Process Engineering 55 (2023) 104165

Fig. 8. Mass-energy balance and related costs of the proposed activated sludge-PAC technology for tannery wastewaters treatment.

dewatering unit allows to obtain a concentrated stream (FW in, Fig. 8) at sludge can strongly improve the methane yield or bio-methanation
25 % w/w of TS content. Based on the purged sludge and the cost process if the anaerobic digestion technology is applied [46].
deriving from the drying of dewatered sludge up to a TS content of 80 % Following this route, the amount of sludge to be disposed in landfilling
w/w (FW out, Fig. 8), the estimated drying costs per m3 of treated water can be remarkably reduced as well as the energy cost, since the produced
were equal to 0.120 €/m3, 0.194 €/m3 and 0.288 €/m3 respectively for biomethane could be utilized as energy source for the drying unit.
R1, R2 and R3 configuration, assuming 1.05 €/m3 as the representative
cost of NG. These data were obtained by a drawback calculation from 4. Conclusions
the evaporated water (based on the TS level in the inlet and outlet of the
drying unit), the energy required to obtain such evaporation (3.35 MJ/ The pollution related to the tannery wastewaters containing high
kgwater), and the NG to be utilized to obtain the required energy concentrations of PFASs compounds, trivalent chromium ([Cr(III)]-
(assuming a lower heating value and the thermal energy yield of 35 MJ/ complexes) and organic dyes introduces many limits to the sustainable
Sm3 and 0.9 respectively). development of the leather industry. This study highlighted the effec­
The disposal costs were estimated at 0.131 €/m3 of treated water, tiveness of the combined activated sludge-PAC processes in a pilot scale
0.213 €/m3 and 0.316 €/m3 respectively for R1, R2 and R3 configura­ PFR for PFASs, Cr(III), organic matter removal and color abatement
tion, assuming the landfilling as the only currently available practice from tannery wastewaters, operating at an HRT and SRT of 4 and 20
and a disposal cost of 230 €/ton. days respectively (R3 condition). The highest applied PAC dosage (1.50
Ultimately, based on the three contributions reported, the total costs g/L as XPAC) allowed reducing the concentration of total PFASs below
estimated for the three operating conditions, namely XPAC of 0 g/L (R1), the limit of 500 ng/L, as standard for drinking water. High removal ef­
0.63 g/L (R2) and 1.50 g/L (R3), were respectively 0.251 €/m3 of water ficiency was observed for the two longer-chain compounds PFOS and
treated, 0.597 €/m3 and 1.174 €/m3. In other words, the PAC dosage PFOA (75 % and 68 % respectively); on the contrary, for the short-chain
caused a minimum increase in the treatment cost of 0.345 €/m3 and congeners, only PFBS was removed at 32 % and no removal was
0.923 €/m3 of treated water respectively for R2 and R3 condition observed for the PFBA. A substantial removal of CODSOL (92.8 %) and
compared to the conventional activated sludge process (R1). The utili­ [Cr(III)] (68 %) was also achieved, in addition to the 56 % of absorbance
zation of the anaerobic approach in conjunction with PAC was out of the reduction as indication of the color abatement in the wastewaters,
scope and it has not been evaluated in this study. However, given the compared to the biological treatment only. On the other hand, the
average size of the full-scale wastewater treatment plant that hosted the proposed technology increased the global cost for the wastewater
pilot facilities (450,000 PE; Person Equivalent), the anaerobic technol­ treatment by 0.923 €/m3 of treated water compared to the conventional
ogy might have needed HRT longer than 4.0 days, with consequent activated sludge process. Hence, the benefits achieved with the use of
unsustainable reactor's volume. the PAC can be more valorized if conscientious practices addressed to
On the other hand, the scenario hypothesized in this study has a large the reduction of tannery sludge landfilling, towards some recovering
potential to be improved since the landfilling of tannery sludge could be strategies which can include renewable energy or the synthesis of added-
progressively reduced along with a desirable development of the cir­ value products.
cular economy principles and biorefinery-based platform chemical
production [26]. Moreover, the presence of PAC in the dried tannery

9
M. Gottardo et al. Journal of Water Process Engineering 55 (2023) 104165

CRediT authorship contribution statement granular activated carbon, Environ. Sci. Technol. Lett. 7 (2020) 343–350, https://
doi.org/10.1021/acs.estlett.0c00114.
[14] K.H. Kucharzyk, R. Darlington, M. Benotti, R. Deeb, E. Hawley, Novel treatment
Marco Gottardo: Conceptualization, Methodology, Investigation, technologies for PFAS compounds: a critical review, J. Environ. Manag. 204 (2017)
Formal analysis, Visualization, Writing – original draft, Writing – review 757–764, https://doi.org/10.1016/j.jenvman.2017.08.016.
& editing. Giulia Adele Tuci: Investigation, Formal analysis, Writing – [15] D. Lu, S. Sha, J. Luo, Z. Huang, X. Zhang Jackie, Treatment train approaches for the
remediation of per- and polyfluoroalkyl substances (PFAS): a critical review,
review & editing. Aditi Chitharanjan Parmar: Formal Analysis, Visu­ J. Hazard. Mater. 386 (2020), 121963, https://doi.org/10.1016/j.
alization. Paolo Pavan: Supervision, Conceptualization, Visualization, jhazmat.2019.121963.
Methodology. Francesco Valentino: Writing – review & editing, [16] Y. Hao, H. Ma, Q. Wang, C. Zhu, A. He, Complexation behaviour and removal of
organic-Cr(III) complexes from the environment: a review, Ecotoxicol.
Funding acquisition, Project administration, Supervision. Environmental Safety 240 (2022), 113676, https://doi.org/10.1016/j.
ecoenv.2022.113676.
[17] K. Li, R. Yu, R. Zhu, R. Liang, G. Liu, B. Peng, pH-sensitive and chromium-loaded
Declaration of competing interest mineralized nanoparticles as a tanning agent for cleaner leather production, ACS
Sustain. Chem. Eng. 7 (2019) 8660–8669, https://doi.org/10.1021/
acssuschemeng.9b00482.
The authors declare that they have no known competing financial [18] D. Wang, S. He, Y. Ye, H. Ma, X. Zhang, W. Zhang, et al., Chromium speciation in
interests or personal relationships that could have appeared to influence tannery effluent after alkaline precipitation: isolation and characterization,
the work reported in this paper. J. Hazard. Mat. 316 (2016) 169–177, https://doi.org/10.1016/j.
jhazmat.2016.05.021.
[19] D. Wang, Y. Ye, H. Liu, H. Ma, W. Zangh, Effect of alkaline precipitation on Cr
Data availability species of Cr(III)-bearing complexes typically used in the tannery industry,
Chemosphere 193 (2018) 42–49, https://doi.org/10.1016/j.
chemosphere.2017.11.006.
Data will be made available on request.
[20] J.S. Piccin, C.S. Gomes, B. Mella, M. Gutterres, Color removal from real leather
dyeing effluent using tannery waste as an adsorbent, J. Env. Chem. Eng. 4 (2016)
Acknowledgements 1061–1067, https://doi.org/10.1016/j.jece.2016.01.010.
[21] G. Liu, R. Yu, T. Lan, Z. Liu, P. Zhang, R. Liang, Gallic acid-functionalized graphene
hydrogel as adsorbent for removal of chromium (III) and organic dye pollutants
This research was financially supported by the European Union - from tannery wastewater, RSC Adv. 9 (2019) 27060, https://doi.org/10.1039/
Next Generation EU: PON “Ricerca e Innovazione”. This study was also c9ra05664e.
[22] J. Altmann, A.S. Ruhl, F. Zietzschmann, M. Jekel, Direct comparison of ozonation
carried out within the “Interconnected Nord-Est Innovation Ecosystem and adsorption onto powdered activated carbon for micropollutant removal in
(iNEST)” project and received funding from the European Union Next- advanced wastewater treatment, Wat. Res. 55 (2014) 185–193, https://doi.org/
GenerationEU - National Recovery and Resilience Plan (NRRP) – 10.1016/j.watres.2014.02.025.
[23] M. Boehler, B. Zwickenpflug, J. Hollender, T.A. Ternes, A. Joss, H. Siegrist,
MISSION 4 COMPONENT 2, INVESTIMENT N. ECS00000043 – CUP N.
Removal of micropollutants in municipal wastewater treatment plants by powder-
H43C22000540006. This manuscript reflects only the authors' views activated carbon, Wat. Sci. Technol. 66 (2012) 2115–2121, https://doi.org/
and opinions, neither the European Union nor the European Commission 10.2166/wst.2012.353.
[24] F. Meinel, A.S. Ruhl, A. Sperlich, F. Zietzschmann, M. Jekel, Pilot-scale
can be considered responsible for them.
investigation of micropollutant removal with granular and powdered activated
carbon, Wat. Air Soil Pollut. 226 (2015) 1, https://doi.org/10.1007/s11270-014-
References 2260-y.
[25] Metcalf & Eddy, Inc, Wastewater Engineering Treatment and Reuse, fourth ed.,
McGraw-Hill Higher Education, New York, 2003.
[1] A.B. Mpofu, O.O. Oyekola, P.J. Welz, Anaerobic treatment of tannery wastewater
[26] G.A. Tuci, F. Valentino, E. Bonato, P. Pavan, M. Gottardo, Microwave-hydrogen
in the context of a circular bioeconomy for developing countries, J. Cleaner
peroxide assisted anaerobic treatment as an effective method for short-chain fatty
Production 296 (2021), 126490, https://doi.org/10.1016/j.jclepro.2021.126490.
acids production from tannery sludge, Processes 10 (2022) 2167, https://doi.org/
[2] V.A. Espana, M. Mallavarapu, R. Naidu, Treatment technologies for aqueous
10.3390/pr10112167.
perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA): a critical review
[27] APHA, AWWA, WEF, Standard Methods for the Examinations of Water and
with an emphasis on field testing, Env. Technol. Innovation 4 (2015) 168–181,
Wastewater, twentieth ed., American Public Health Association, 1015 Fifteenth
https://doi.org/10.1016/j.eti.2015.06.001.
Street, NW, Washington, DC, 1998, pp. 20005–22605.
[3] I.T. Cousins, Per- and polyfluoroalkyl substances in materials, humans and the
[28] R.M.C. Viegas, E. Mesquita, M. Campinas, M.J. Rosa, Pilot studies and cost analysis
environment, Chemosphere 129 (2015) 1–3, https://doi.org/10.1016/j.
of hybrid powdered activated carbon/ceramic microfiltration for controlling
chemosphere.2014.08.036.
pharmaceutical compounds and organic matter in water reclamation, Water 12
[4] REACH legislation, https://echa.europa.eu/regulations/reach/legislation.
(2020) 33, https://doi.org/10.3390/w12010033.
[5] EU, The drinking water directive 98/83/EC. http://data.europa.eu/eli/dir/2020/2
[29] B. Akcora, O.K. Kocaaslan, Price bubbles in the European natural gas market
184/oj, 2020.
between 2011 and 2020, Resources Policy 80 (2023), 103186, https://doi.org/
[6] C. Flores, F. Ventura, J. Martin-Alonso, J. Caixach, Occurrence of perfluorooctane
10.1016/j.resourpol.2022.103186.
sulfonate (PFOS) and perfluorooctanoate (PFOA) in N.E. Spanish surface waters
[30] Stockholm Convention on Persistent Organic Pollutants (POPS). http://chm.pops.
and their removal in a drinking water treatment plant that combines conventional
int/theconvention/overview/textoftheconvention/tabid/2232/default.aspx, 2001.
and advanced treatments in parallel lines, Sci. Total Env. 461-462 (2013) 618–626,
[31] M. Ateia, A. Maroli, N. Tharayil, T. Karanfil, The overlooked short- and ultrashort-
https://doi.org/10.1016/j.scitotenv.2013.05.026.
chain poly- and perfluorinated substances: a review, Chemosphere 220 (2019)
[7] B. Shivakoti, S. Fujii, M. Nozoe, S. Tanaka, C. Kunacheva, Perfluorinated chemicals
866–882, https://doi.org/10.1016/j.chemosphere.2018.12.186.
(PFCs) in water purification plants (WPPs) with advanced treatment processes,
[32] M. Rahman, S. Peldszus, W. Anderson, Behaviour and fate of perfluoroalkyl and
Water Sci. Technol. Water Supply 10 (2010) 87–95, https://doi.org/10.2166/
polyfluoroalkyl substances (PFASs) in drinking water treatment: a review, Wat.
ws.2010.707.
Res. 50 (2014) 318–340, https://doi.org/10.1016/j.watres.2013.10.045.
[8] S. Takagi, F. Adachi, K. Miyano, Y. Koizumi, H. Tanaka, I. Watanabe, et al., Fate of
[33] A. Chiavola, C. Di Marcantonio, M.R. Boni, S. Biagioli, A. Frugis, G. Cecchini, PFOA
perfluorooctanesulfonate and perfluorooctanoate in drinking water treatment
and PFOS removal processes in activated sludge reactor at laboratory scale,
processes, Wat. Res. 45 (2011) 3925–3932, https://doi.org/10.1016/j.
Advances Sci. Technol. Innovation (2020) 375–377, https://doi.org/10.1007/978-
watres.2011.04.052.
3-030-13068-8_94.
[9] US Environmental Protection Agency, Accelerated Biodegradation of 8-2 Telomer B
[34] Z. Zhang, D. Sarkar, J. Kumar Biswas, R. Datta, Biodegradation of per- and
Alcohol 45, Office of Pollution Prevention and Toxics, Docket AR226-1264, ed.
polyfluoroalkyl substances (PFAS): a review, biores, Technol. 344 (2022), 126223,
Washington DC, 2003.
https://doi.org/10.1016/j.biortech.2021.126223.
[10] Z. Du, S. Deng, Y. Bei, Q. Huang, B. Wang, J. Huang, et al., Adsorption behavior
[35] F. Li, J. Duan, S. Tian, H. Ji, Y. Zhu, Z. Wei, et al., Short-chain per-and
and mechanism of perfluorinated compounds on various adsorbents - a review,
polyfluoroalkyl substances in aquatic systems: occurrence, impacts and treatment,
J. Hazardous Mat. 274 (2014) 443–454, https://doi.org/10.1016/j.
Chem. Eng. J. 380 (2020), 122506, https://doi.org/10.1016/j.cej.2019.122506.
jhazmat.2014.04.038.
[36] A. Maimaiti, S. Deng, P. Meng, W. Wang, B. Wang, J. Huang, et al., Competitive
[11] US Environmental Protection Agency, Laboratory-Scale Thermal Degradation of
adsorption of perfluoroalkyl substances on anion exchange resins in simulated
Perfluorooctanyl Sulfonate and Related Substances 13, Office of Pollution
AFFF-impacted groundwater, Chem. Eng. J. 348 (2018) 494–502, https://doi.org/
Prevention & Toxics, Docket AR226-1366, ed. Washington DC, 2003.
10.1016/j.cej.2018.05.006.
[12] US Environmental Protection Agency, Laboratory-Scale Thermal Degradation of
[37] R.K.M. Bulusu, R.J. Wandell, Z. Zhang, M. Farahani, Y. Tang, B.R. Locke,
Perfluoro- Octanyl Sulfonate and Related Substances 142, Office of Pollution
Degradation of PFOA with a nanosecond-pulsed plasma gas-liquid flowing film
Prevention & Toxics, Docket AR226-1367, ed. Final Report: Washington DC, 2003.
reactor, Plasma Process. Polym. 17 (2020) 2000074, https://doi.org/10.1002/
[13] F. Xiao, P.C. Sasi, B. Yao, A. Kubátová, S.A. Golovko, M.Y. Golovko, D. Soli,
ppap.202000074.
Thermal stability and decomposition of perfluoroalkyl substances on spent

10
 Water Research 170 (2020) 115333

Contents lists available at ScienceDirect

Water Research
journal homepage: www.elsevier.com/locate/watres

Kinetics of biological and chemical processes in anoxic-aerobic

digestion of phosphorus rich waste activated sludge
D.S. Ikumi*, T.H. Harding
Future Water Institute, New Engineering Building, University of Cape Town, Rondebosch, 7701, Cape Town, South Africa

a r t i c l e i n f o a b s t r a c t

Article history: Because the functions of these water and resource recovery facilities (WRRFs) stretches beyond simply
Received 14 July 2019 meeting effluent requirements (i.e., also includes optimisation of products to be generated as recovered
Received in revised form resources), a high level of accuracy is required in using mathematical models that virtually replicate
20 November 2019
(hence predict) WRRF system responses to dynamic conditions. The currently developed mathematical
Accepted 22 November 2019
models embrace the majority of advances made towards tracking nitrogen (N) and phosphorus (P)
Available online 25 November 2019
through the entire WRRF, and significant effort has been made towards calibrating them to predict
realistic outcomes. This paper presents the stepwise calibration of the PWMSA model (Ikumi et al., 2015)
Keywords:
Phosphorus removal
for aerobic (AerD) and anoxic-aerobic digestion (AAD) processes, through predictions of (i) mineral
Activated sludge precipitation potential, in isolation to biological reactions (ii) AerD bioprocesses (including nitrification,
Anoxic-aerobic digestion orthophosphate (OP) release, and endogenous respiration), in isolation to mineral precipitation (iii)
Plant-wide modelling predicted interaction of the mineral precipitation and the biological processes of organic removal and
nitrification, excluding P accumulating organisms (PAOs) and polyphosphate (PP) release during AerD,
(iv) replicated interaction of mineral precipitation and bioprocesses of P release and nitrification kinetics
(v) predicted PAO behavioural kinetics of anaerobic OP release with acetate uptake and aerobic PP up-
take, in isolation to nitrification and (vi) predicted nitrate denitrification and anoxic OP release. The
calibrated kinetic parameters allowed for the model capability of reproducing the data from the key
biological, physical and chemical processes occurring in the various environments of sludge treatment
(aerobic, anoxic and anaerobic) within satisfactory level of accuracy.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction anoxic zones of the activated sludge (AS) systems for effective
nutrient removal (Ekama, 2017). To assess the impact of operation
For conventional methods of tertiary (sludge) treatment pro- strategies and to optimise nutrient removal in these systems re-
cesses, wastewater treatment systems could adopt either anaerobic quires the utilisation of predictive mathematical models that can
digestion (AD) or anoxic-aerobic digestion (AAD) technology (Shao simulate the sludge treatment processes with reasonable accuracy
et al., 2013; Banjade, 2008). The AD systems offer energy savings (Solon et al., 2019).
but are less robust and result in high nutrient concentrations Similar to ADs, the AAD of phosphorus (P) - rich waste activated
within the dewatering liquor (Ekama, 2017; Christensen et al. sludge (WAS) raises research questions, regarding the rate of
1984). Ideally, these nutrients could be utilised towards produc- release of P from the biomass P and polyphosphate (PP) on the 3
tion of fertiliser (e.g. via processes such as struvite formation) or get phase (aqueous-gas-solid) mixed weak acid/base chemistry of AAD
transferred to delicately poised biological systems like Sharon- systems - i.e., (i) are the P and metal (Mg2þ, Kþ, Ca2þ) content of PP
Ammanox (Macura et al., 2019; Kominko et al., 2018; Jamil et al., released rapidly from P accumulating organisms (PAOs) as noticed
2006). However, these nutrients often they get recycled upstream in AD or gradually with PAO endogenous respiration and (ii) are
to the parent activated sludge unit operations - this necessitates their concentrations, once released in the aqueous medium, suffi-
dosage of readily biodegradable organics to the anaerobic and cient for formation of mineral precipitates? This paper presents the
results from a three phase (aqueous/solid/gas) mixed weak acid/
base kinetic model developed by Ikumi et al. (2015), when applied
* Corresponding author. to simulate the biological, physical and chemical processes
E-mail address: [email protected] (D.S. Ikumi).

https://doi.org/10.1016/j.watres.2019.115333
0043-1354/© 2019 Elsevier Ltd. All rights reserved.
2 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333

occurring in dynamic batch aerobic, anoxic and anaerobic sludge 0.45 mm membrane filtered, the solids tested for concentrations of
treatment conditions. A model calibration procedure is applied, volatile suspended solids (VSS), total suspended solids (TSS), and
whereby the data generated from various experimental in- the filtrate tested for free and saline ammonia (FSA), nitrates (NO3)
vestigations are matched against the model’s capability to predict and OP. The batch reactors maintained a monitored pH and volume
the main biological and chemical processes in this sludge treat- control and sludge mass loss due to sampling was taken into
ment, including the endogenous respiration of the ordinary het- account.
erotrophic organism (OHO) and PAO biomass, organic utilisation, Table 1, below provides a summary of the experimental systems,
nitrification, PP release and mineral precipitation. used towards calibrating this model. Also included is the reference
to the processes, described in Table 3, that are catered for (i.e., in
2. Materials and methods calibration of the model) by the experimental systems.

The selection of batch tests towards the model calibration ex- 2.2. Model description
ercise is performed by stepwise isolation of the relevant processes
(i.e., physical and chemical processes of gas stripping and precipi- The UCT three-phase plant-wide model (PWM_SA) was devel-
tation respectively followed by biological processes of endogenous oped for simulating the biological processes to track and predict the
respiration, nitrification and polyphosphate release). The calibra- output of materials (chemical oxygen demand (COD), carbon (C),
tion exercise is done this way, towards accurate determination of hydrogen (H), oxygen (O), nitrogen (N), phosphorus (P), magne-
the kinetic parameters of importance. sium (Mg), potassium (K) and calcium (Ca)) along the unit pro-
cesses of a wastewater treatment plant (WWTP). It comprises three
2.1. Experimental set-up and data sub-models, integrated for simulation of the entire WWTP under
various configurations (e.g., NDBEPR AS system linked to an AD or
The dynamic results of the following batch tests, operated under an anoxic-aerobic digestion (AAD) for WAS stabilisation). These
varying experimental conditions, during four main experimental sub-models include:
periods (see results section below for further details) at the Water
Research Laboratory of the University of Cape Town (UCT), are used 1. The ionic speciation model (Brouckaert et al., 2010). This model
in this process: (i) Operated by Musvoto (1998) generating data includes pairing of ionic components and inter-phase transfers
useful towards (a) validation of models for chemical and physical of component species.
processes of precipitation and gas stripping (anaerobic digester 2. The ASM2-3P model: This is the activated sludge model No. 2
supernatant, ADS) in isolation to bioprocesses (the data obtained (ASM2, Henze et al., 1995), modified to include the ionic speci-
from batch test 5; i.e., AerD BT5) and (b) validation of the model for ation model (Brouckaert et al., 2010), the Inorganic Settleable
chemical and physical processes, with the inclusion of and/or Solids (ISS) model of Ekama and Wentzel (2004) and including
interaction with nitrification and ordinary heterotrophic organism multiple mineral precipitation according to Musvoto et al.
(OHO) biomass biological processes (data obtained from aerobic (2000a,b).
batch digestion test ART1; i.e. AerD BT ART1), with negligible PAOs 3. The SDM3P Model: This is the UCT sludge digestion model
present (hence no PP disintegration expected), (ii) data from aer- (UCTSDM; So €temann et al., 2005a,b,c), modified to include the
obic batch digestion test 8 (AerD BT8) of Mebrahtu et al. (2007) to hydrolysis of multiple organic sludge types (primary sludge (PS),
validate orthophosphate (OP) release and nitrification (with min- NDBEPR WAS and PS-WAS blends), the Ekama and Wentzel
eral precipitation excluded through dilution of the WAS for main- (2004) inorganic suspended solids (ISS) model, multiple min-
tenance of a low concentrations) during aerobic endogenous eral precipitation processes according to Musvoto et al. (2000a,
breakdown of mixed cultures of OHO and PAO biomass (iii) Results b) and the Brouckaert et al. (2010) aqueous speciation model
from Vogts (2015), used for validation of (a) integrated biological, which facilitates ionic speciation (Ikumi, 2011).
physical and chemical processes (i.e., inclusion of the endogenous
respiration, OP release and nitrification) during aerobic digestion of For their compatibility, the ASM2-3P and SDM3P models have
nitrification-denitrification biological excess P removal (NDBEPR) the same comprehensive set of model components (supermodel
WAS (data obtained from AerD BT2 and AerD BT4 of the study) and approach, Volcke et al., 2006), including parameterized stoichi-
(b) To observe OP release with denitrification, using the anoxic ometry for the bioprocesses and sharing the same ionic speciation
digestion batch test 7 (AxD BT7) and (c) To validate the kinetics of subroutine model (1).
anaerobic OP release with PAO biomass anaerobic acetate uptake
and aerobic OP uptake for PP formation using data from the 2.3. Model evaluation process
anaerobic-aerobic batch digester 8 (AnAerD BT8) (iv) results pre-
sented by Harding et al. (2010) - batch AD tests on NDBEPR WAS - to The model evaluation process is presented by Ikumi et al. (2014),
calibrate the AD model for the kinetics of AD OP release (including which was based on the adoption of BIOMATH protocol developed
that from PP breakdown), and mineral precipitation, with the in- by Vanrolleghem et al. (2003). This process involved (i) the sys-
clusion of AD bioprocesses (AD BT1). tematic model verification method proposed by Hauduc et al.
Detailed descriptions of these experimental systems, used to- (2010) to check material (COD, C, H, O N, P, Mg K and Ca) mass
wards calibrating this model are provided prior to presentation of balances; (ii) preparation of initial parameter values (determined
the results, in the respective sections below. For the batch tests, experimentally or obtained from literature) and given their typical
completely mixed reactors were set up (with stirrers in the form of range using the method proposed by Brun et al. (2002); (iii) eval-
paddles driven by electric motors mounted on top of the reactor uation of parameters using (a) the Standardised Regression Co-
lids to mixed the sludge continuously). For aerobic batch tests, the efficients (SRC) and (b) Morris Screening methods. This resulted in
batch reactors were also fitted with dissolved oxygen (DO) probes, prioritised model parameters (the ones with great effects), the
valves, oxygen utilisation rate (OUR) meters and humidified air determination of non-influential parameters (those that can be
supply. Fig. 1 shows a typical set up for these batch reactors. All fixed at any range, without influencing model variable outputs) and
batch test samples were 50m[ unless otherwise stated. These quantified degrees of interaction between parameters that intri-
samples were centrifuged at 2500 rpm for 10min, the supernatant cately worked together to influence model output results
 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333 3

Fig. 1. Experimental set-up of typical batch reactor used to carry out research investigations.

Table 1
The selection of batch tests contributing towards the model calibration exercise.

Reference Purpose Reactor Feed and Operation

1 Musvoto Mineral precipitation and gas stripping in isolation to 5 L Aerobic Batch Fed anaerobic digester supernatant (ADS; from an AD fed blended
(1998) bioprocesses (i.e., II and III of Table 3) Digestion (AerD BT5) PS and WAS and operated at 20 d R); 100 mg/l of Ca, Mg and P (as
Ca, Mg and P from CaCl2, MgCl2 and K2HPO4 respectively) were
added to the supernatant prior to its aeration
2 Musvoto Chemical precipitation, integrated with nitrification (by ANOs) 5 L Aerobic batch Fed mixture of raw sewage and MLE WAS (containing OHOs and
(1998) and OHO biomass biological processes with negligible PAOs digester (AerD BT negligible PAOs) with addition of Ca, Mg, P and FSA
present (i.e., I to IV of Table 3) ART1)
3 Mebrahtu Validate orthophosphate (OP) release and nitrification (with 3 L Aerobic batch Fed diluted (from 20gTSS to 4gTSS/l; i.e.,  5) MBR NDBEPR WAS
et al. mineral precipitation excluded) during aerobic endogenous digester (AerD BT8) (containing mixed cultures of OHO and PAO biomass) from AS
(2007) breakdown of mixed cultures of OHO and PAO biomass (i.e., I, III, system operated at 20 d R
IV, VI and VII of Table 3)
4 Vogts Validate integrated biological, physical and chemical processes 2 L Aerobic batch Fed dilluted (from 10gTSS/l to 3 gTSS/l) MBR NDBEPR WAS
(2015) (including endogenous respiration, OP release and nitrification) digester (AerD BT2) (containing mixed cultures of OHO and PAO biomass) from AS
during aerobic digestion of nitrification-denitrification biological system operated at 10 d R; 20d experimental period, NAOH not
excess P removal (NDBEPR) WAS (i.e., I, II, III, IV and VI of Table 3). added for pH control
5 Vogts Same as for (4) above. 2.5 L Aerobic batch Fed dilluted (from 10gTSS/l to 2 gTSS/l) MBR NDBEPR WAS
(2015) digester (AerD BT4) (containing mixed cultures of OHO and PAO biomass) from AS
system operated at 10 d R; 20d experimental period, with pH
controlled at 7.3 to 7.6 with measured addition of a 19.05 g/[ NaOH
solution
5 Vogts Validate OP release with denitrification (i.e., process V of Table 3). 2 L Anoxic batch Fed a nitrate rich and OP deficient mixture of 1 L of effluent from
(2015) digester (AxD BT7) continuous AAD that was fed thickened (from 10gTSS/l to 20gTSS/
l) MBR NDEBPR WAS mixed with 700m[ of the NDBEPR AS system
effluent and 300m[ of supernatant from a continuous dilluted AAD
(from the 10gTSS/l to 3gTSS/l); stirring without aeration for 24 h
6 Vogts Validate the kinetics of anaerobic OP release with PAO biomass 2 L Anaerobic- Fed 1 L of effluent from continuous AAD that was fed thickened
(2015) anaerobic acetate uptake and aerobic OP uptake for PP formation, Aerobic batch (from 10gTSS/l to 20gTSS/l) MBR NDEBPR WAS mixed with 1[ of
while excluding aerobic nitrification processes (i.e. process I, VI digester (AnAerD distilled water with a measured dose of sodium acetate (5
and VII of Table 3) BT8) gCH3COONa) dissolved in it; operated at 6hr test of 3hr with
aeration off and 3hr aeration on; pH was monitored and
maintained between 7.3 and 7.7, by adding measured doses of 1N
NaOH and HC[
7 Harding Kinetics of AD OP release with PP breakdown, mineral 5 L Mesophilic (35  C) Fed 2.5 L from 10gTSS/l MBR NDBEPR WAS (containing mixed
et al. precipitation, with the inclusion of all AD bioprocesses. Anaerobic batch cultures of OHO and PAO biomass) from AS system operated at
(2010) digester (AD BT1) 10 d R mixed with 2.5 L of effluent from a continuous AD (operated
at 18 d R) fed the NDBEPR WAS; operated for 10 days with no
chemical additives included for pH regulation

(Neumann, 2012). In the final step (iv) of model calibration and of parameters. For the sampled parameter sets, simulations were
validation non-influential parameters were set at their default conducted and predicted model outputs were compared with
values and random samples were drawn from the remaining subset observed outputs. During this calibration phase, a consistent set of
4 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333

Table 2a
Stoichiometric Parameters used in Simulating the ASM2-3P Model.

No. Parameter Description Value Units Reference

1 f_S_I Fraction of inert COD in particulate substrate 0.003 ratio Henze et al. (1995)
2 f_XU_Bio_lysis Fraction of inert COD generated in biomass lysis 0.08 gCOD/gCOD Dold et al. (1980)
3 ISS_BM ISS to biomass for OHO and PAO 0.15 g/gCOD Ekama and Wentzel (2004)
4 i6_Ca_PP_mol_perP Ca/P: Polyphosphate 0.053 ratio Ikumi (2011)
5 i6_K_PP_mol_perP K/P:Polyphosphate 0.312 ratio Ikumi (2011)
6 i6_Mg_PP_mol_perP Mg/P:Polyphosphate 0.297 ratio Ikumi (2011)
7 i_H_XBInf_mol_perC H/C: PS biodegradable particulate organics (BPO,PS) 2.469 ratio Ikumi (2011)
8 i_H_XUOrg_mol_perC H/C: endogenous residue organics (ER) 1.567 ratio Ikumi (2011)
9 i_H_SF_mol_perC H/C: fermentable biodegradable soluble organics (FBSO) 2.004 ratio Ikumi (2011)
10 i4_H_Org_mol_perC H/C: organisms 1.454 ratio Ikumi (2011)
11 i_H_XUInf_mol_perC H/C: unbiodegradable particulate organics (UPO) 1.567 dUnit/dUnit Ikumi (2011)
12 i_H_SU_mol_perC H/C: unbiodegradable soluble organics (USO) 1.648 dUnit/dUnit Ikumi (2011)
13 i_O_XBInf_mol_perC O/C: PS biodegradable particulate organics (BPO,PS) 0.848 ratio Ikumi (2011)
14 i_O_XUOrg_mol_perC O/C: endogenous residue organics (ER) 0.565 ratio Ikumi (2011)
15 i_O_SF_mol_perC O/C: fermentable biodegradable soluble organics (FBSO) 0.66 ratio Ikumi (2011)
16 i_O_Org_mol_perC O/C: organisms 0.357 ratio Ikumi (2011)
17 i_O_XUInf_mol_perC O/C: unbiodegradable particulate organics (UPO) 0.565 ratio Ikumi (2011)
18 i_O_SU_mol_perC O/C: unbiodegradable soluble organics (USO) 0.511 ratio Ikumi (2011)
19 i_N_XBInf_mol_perC N/C: PS biodegradable particulate organics 0.033 ratio Ikumi (2011)
20 i_N_XUOrg_mol_perC N/C: endogenous residue organics (ER) 0.062 ratio Ikumi (2011)
21 i_N_SF_mol_perC N/C: fermentable biodegradable soluble organics (FBSO) 0.058 ratio Ikumi (2011)
22 i2,4
_N_Org_mol_perC N/C: organisms 0.227 ratio Ikumi (2011)
23 i_N_XUInf_mol_perC N/C: unbiodegradable particulate organics (UPO) 0.062 ratio Ikumi (2011)
24 i_N_SU_mol_perC N/C: unbiodegradable soluble organics (USO) 0.135 ratio Ikumi (2011)
25 i_P_XBInf_mol_perC P/C: PS biodegradable particulate organics (BPO,PS) 0.013 ratio Ikumi (2011)
26 i_P_XUOrg_mol_perC P/C: endogenous residue organics (ER) 0.012 ratio Ikumi (2011)
27 i_P_SF_mol_perC P/C: fermentable biodegradable soluble organics 0.005 ratio Ikumi (2011)
28 i_P_Org_mol_perC P/C: organisms 0.031 ratio Ikumi (2011)
29 i_P_XUInf_mol_perC P/C: unbiodegradable particulate organics (UPO) 0.012 ratio Ikumi (2011)
30 i_P_SU_mol_perC P/C: unbiodegradable soluble organics (USO) 0.03 ratio Ikumi (2011)
31 Y_ANO Yield for Autotrophic Biomass 0.24 gCOD/gN Henze et al. (2000)
32 Y1_OHO Yield for Heterotrophic Biomass 0.67 gCOD/gCOD Dold and Marais (1986)
33 Y1_PAO Yield coefficient (biomass/PHA) 0.67 gCOD/gCOD Wentzel et al. (1990)
34 Y_Stor_PP_PAO PHA requirement for PP storage 0.2 gCOD/gP Smolders et al. (1995)
35 Y_PP_Stor_PAO PP requirement (S_PO4 release) per PHA stored 0.4 gP/gCOD Wentzel et al. (1990)

Nb: The numbers1 to 6 indicate parameters that are most sensitive to prediction of organic removal (1), nitrification (2), OP release (3), PP uptake (4), denitrification (5) and
mineral precipitation (6) respectively.
All stoichiometric parameters were set at default from literature references given and none of them were newly calibrated.

parameter values was used to simulate all experimental systems/ These results include volatile suspended solids (VSS; to match the
periods, and detailed explanations of observed discrepancies (if sum of all particulate organics), total phosphates (TP; to match the
any) were reported by Ikumi et al. (2015). The sensitivity analysis sum of all components containing P, including PP and OrgP and OP),
together with intuitive observation on a steady state stoichiometric ISS (to match the mass concentration of PP and reactor accumu-
model played a significant role in selecting the ‘best’ set of pa- lated influent ISS), etc.
rameters. However, using this selected set of parameters to model
the different unit processes (in this case, both NDBEPR and AAD 3.1. Mineral precipitation - calibration against Musvoto (1998)
systems) required versatility in the ASM2-3P model kinetics to aerobic batch test 5 data
simulate the variation in unit processes.
A batch aerobic test, amongst the ones operated by Musvoto
3. Results and discussions (1998), provides useful data towards validation of models for
chemical and physical processes of precipitation and gas stripping.
The PWM_SA model (Ikumi et al., 2015) is used to simulate both The aerobic batch experiment was conducted on the anaerobic
the NDBEPR parent system and the batch digestion tests. digestion supernatant (ADS) from an AD (operated at 20 d sludge
The model calibration process was done via systematic isolation age) in Cape Flats wastewater treatment works (CF WWTW, in
of relevant processes according to the emphasis placed on different South Africa). The AD was fed a blend of PS and WAS from a 5 stage
variables for each experimental data set. The model predicted re- bardenpho AS system treating municipal sewage). Its effluent
sults are compared to experimental measurements, firstly for sludge was sampled directly into sealed pressurised containers
mineral precipitation excluding bioprocesses, then bioprocesses (with care taken to ensure no CO2 loss) which were stored at 20  C.
excluding the precipitation and finally for the combination of bio- Five litres of the ADS, decanted from the containers, were placed in
processes and mineral precipitation (i.e., the biological, chemical batch reactors and 100 mg/l of Ca, Mg and P (as Ca, Mg and P from
and physical processes). CaCl2, MgCl2 and K2HPO4 respectively) were added to the super-
To simulate the given experimental scenarios with the ASM2-3P natant prior to its aeration. To ease CO2 loss concerns, the ADS was
model, a pre-processor (calculations done in Ms Excel to reconcile sparged with CO2/CH4 (60/40) gas mixture typical of AD. This was
experimentally measured data towards determination of input done until the pH dropped to a constant value (after about
model component and parameter values) was set up that could 5e10 min) and the initial sample taken immediately after. The su-
calculate the initial conditions (day zero) of the batch reactor, by pernatant was then aerated (using compressed air passed through a
wastewater and sludge characterisation procedure of Ikumi (2011). diffuser stone in the batch reactor) for above 24 h and its pH
 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333 5

Table 2b
Kinetic Parameters used in Simulating the ASM2-3P Model.

No. Parameter Description Value (literature) Value (calibrated) Units Reference

36 b_ANO Decay rate 0.15 1/d Henze et al. (2000)

37 b1_OHO Rate constant for lysis and decay 0.62 1/d Dold et al. (1991)
38 b1_PAO Rate constant for lysis of PAO 0.04 1/d Wentzel et al. (1990)
39 b_PHA Rate constant for lysis of PHA 0.04 1/d Henze et al. (2000)
40 b4_PP Rate constant for lysis of PP 0.2 0.03 1/d Henze et al. (2000)
41 k’r6_cal Dissolution/precipitation of calcite 0.5 1/d Musvoto et al. (2000a,b)
42 k’r6_cap Dissolution/precipitation of calcium phosphate 350 0.15 1/d Musvoto et al. (2000a,b)
43 k’r6_mag Dissolution/precipitation of magnesite 50 1/d Musvoto et al. (2000a,b)
44 k’r6_mgkp Dissolution/precipitation of K-struvite 100 1/d Musvoto et al. (2000a,b)
45 k’r6_newb Dissolution/precipitation of newberyite 0.05 1/d Musvoto et al. (2000a,b)
46 k’r6_stru Dissolution/precipitation of struvite 3000 stays same 1/d Musvoto et al. (2000a,b)
48 K_S_VFA saturation coeff for S_A (acetate) 4 gCOD/m3 Henze et al. (2000)
49 K_S_ALK saturation coeff for alkalinity (HCO3-) 0.1 mol HCO3/m3 Henze et al. (2000)
50 K_S_ALK_ANO saturation coeff of autotrophs for alkalinity 0.5 0.1 mol HCO3/m3 Henze et al. (2000)
51 K_F_OHO saturation/inhibition coeff for growth on S_F 4 gCOD/m3 Henze et al. (2000)
52 K_I_PP_PAO Inhibition coeff for X_PP storage 0.02 gPP/gPAO Henze et al. (2000)
53 K4_MAX_fPP_PAO Maximum ratio of X_PP/X_PAO 0.34 1 gPP/gPAO Wentzel et al. (1990) Henze et al. (2000)
54 K_S_NHx saturation coeff for ammonium (nutrient) 0.05 gN/m3 Henze et al. (2000)
55 K2_S_NHx_ANO saturation coeff of autotrophs for ammonium 1 0.3 gN/m3 Henze et al. (2000)
56 K_NOx_OHO saturation/inhibition coeff for nitrate 0.5 gN/m3 Henze et al. (2000)
57 K_O2 saturation/inhibition coeff for oxygen 0.2 gO/m3 Henze et al. (2000)
58 K_O2_ANO saturation/inhibition coeff of ANOs for oxygen 0.5 stays same gO/m3 Henze et al. (2000)
59 K_S_PO4 saturation coeff for phosphorus (nutrient) 0.01 gP/m3 Henze et al. (2000)
60 K_S_PHA_PAO saturation coeff for PHA 0.01 gPHA/gPAO Henze et al. (2000)
61 KS_fPP_PAO_PHA saturation coeff for poly-phosphate 0.01 gPP/gPAO Henze et al. (2000)
62 K_S_PO4_PAO_PP saturation coeff for phosphorus in PP storage 0.2 gP/m3 Henze et al. (2000)
63 K1_S_BInf_OHO_hyd saturation coeff for particulate COD 0.1 gBPO/gOHO Henze et al. (2000)
64 K4_S_F_OHO_ferm saturation coeff for fermentation on S_F 20 stays same gCOD/m3 Henze et al. (2000)
65 k_M_BInf_OHO_hyd Hydrolysis rate constant 3 gCOD/(gCOD.d) Henze et al. (2000)
66 mu_2ANO Maximum growth rate of ANOs 1 stays same 1/d Dold et al. (1991)
67 mu_OHO Maximum growth rate on substrate 6 1/d Dold et al. (1991)
68 mu_PAO Maximum growth rate 1 1/d Dold et al. (1991)
69 n5_NO_Het Reduction factor for de-nitrification 0.6 0.3 e Wentzel et al. (1992)
70 n_OHO_BInf_hyd Anoxic hydrolysis reduction factor 0.6 e Henze et al. (2000)
71 n_OHO_BInf_ferm Anaerobic hydrolysis reduction factor 0.1 e Henze et al. (2000)
72 Q_PAO_PP_Stor Rate constant for storage of PHA (base: X_PP) 3 stays same 1/d Henze et al. (2000)
73 Q4_PAO_PO4_PP Rate constant for storage of PP 1.5 2 1/d Henze et al. (2000)
74 Q4_OHO_F_VFA Maximum rate for fermentation 3 20 1/d Henze et al. (2000)
75 S_O2_Sat Oxygen saturation concentration 8.9 g/m3 € temann et al., (2005a,b,c)
So

Nb: The numbers1 to 6 indicate parameters that are most sensitive to prediction of organic removal (1), nitrification (2), OP release (3), PP uptake (4), denitrification (5) and
mineral precipitation (6) respectively (Ikumi et al., 2011). Since the ASM2-3P model is an extension of ASM2 kinetic model of Henze et al. (1995, 2000), most parameter values
are obtained from Henze et al. (2000). However, some new values were replaced with those observed in other literature, for new processes (e.g., mineral precipitation, where
values used by Musvoto et al. (2000a,b) were adopted) and where the predictions were significantly improved by utilisation of the other literature values (e.g., bOHO value of
0.62 from Dold et al. (1991) is used instead of the original value of 0.4 from Henze et al. (2000)). The newly calibrated values are adjusted where further improvements were
noted in the model predictions according to processes noted in Table 3 below.

Table 3
Parameters determined from parameter estimation (manually done during calibration process) compared to the parameters of importance in sensitivity analysis.

Process Stoichiometric processes involved Noted from sensitivity analysis Parameters Determined

Stoichiometric Kinetic

(I.) Ordinary heterotrophic Biomass BPO hydrolysis2; OHO growth on FSO3; Y_OHO b_OHO All set at default
organism (OHO) endogenous OHO growth on Ac4; Fermentation of FSO7; OHO
respiration (ER) Lysis8
(II.) Mineral precipitation Struvite21; ACP22; MgKP23; Calcite24; Magnesite25; i_Ca_PP_mol_perP; k’r_cal; k’r_cap; k’r_mag; k’r_mgkp; k’r_cap; k’r_stru
Newberryte26 i_K_PP_mol_perP; k’r_newb; k’r_stru
i_Mg_PP_mol_perP
19 20
(III.) Aeration and CO2 active gas Active CO2 gas exchange ; Aeration y_S y_S; KLa_CO2
exchange
(IV.) Autotrophic nitrifying ANO growth17; ANO Lysis18 i_N_Org_mol_perC mu_ANO; K_S_NHx_ANO mu_ANO; K_S_NHx_ANO;
organisms (ANO) bioprocesses K_O2_ANO
(V.) Denitrification by OHOs OHO growth on FSO5; OHO growth on Ac6; Y_OHO n_NO_Het; b_OHO All set at default
Biomass BPO hydrolysis2; Fermentation of FSO7;
8
OHO Lysis
(VI.) Phosphorus accumulating Glycogen Storage9; PHB Storage10; PAO Lysis13; PP Y_PAO; K_S_F_OHO_ferm; Q_PAO_PO4_PP; K_S_F_OHO_ferm;
organism (PAO) aerobic Lysis14; PHB Lysis15; Glycogen hydrolysis16 i_N_Org_mol_perC; Q_OHO_F_VFA; K_MAX_fPP_PAO; Q_PAO_PO4_PP; Q_OHO_F_VFA;
bioprocesses i_H_Org_mol_perC b_PAO; b_PP K_MAX_fPP_PAO
(VII.) Phosphorus accumulating PP Storage11; PAO growth on PHB12; PAO Lysis13; i_N_Org_mol_perC; K_S_F_OHO_ferm; Q_OHO_F_VFA; b_PP; Q_PAO_PP_Stor
organism (PAO) anaerobic PP Lysis14; PHB Lysis15; Glycogen hydrolysis16 i_H_Org_mol_perC b_PAO; b_PP
bioprocesses
2 26
Nb: the numbers to indicate the process number in the Gujer matrix of theASM2-3P model.
6 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333

continuously monitored. This experimental set up was simulated As notable from Fig. 2, most PWM_SA predicted results (OP, Mg,
using the PWM_SA model to calibrate the model against the pro- Ca, pH and H2CO3 Alk) are closely matched to those measured by
vided data for chemical precipitation and gas stripping processes. Musvoto (1998). The poor correlation between measured and
Table 2 (a and b) provides the determined set of significant pa- simulated FSA (Fig. 2f) is attributed to testing for FSA after vacuum
rameters and their values that allowed the best match between filtration. Musvoto (1998) noted that this vacuum filtration resulted
experimental and simulated variables (See Fig. 2 for the compari- in the stripping of ammonia, as would occur with reactor aeration.
son of measured and simulated outputs). This was not captured in the predictions because the ammonia was

Fig. 2. (a to f). Simulation of chemical precipitation and gas stripping processes occurring in the batch aerobic digestion of anaerobic digestion supernatant (ADS), as operated by
Musvoto (1998).
 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333 7

stripped outside of the confines of the experimental environment Moreover, the values did not reflect the dilution of 20gTSS to a fifth
virtually replicated by the reactor and better predictions would be of the concentration. Therefore, due to the spurious TP data, the ISS
obtainable with the sample analysed before vacuum filtration. The was alternatively used to determine the concentration of influent
poor correlation between the measured and simulated HCO 3 is PP (this was done as shown in data reconciliation methods by
linked to the ammonia loss during the filtration process. The con- Ikumi, 2011) as input component for the PWM_SA model simula-
version of ammonia to nitrates, which is deemed to have occurred tions. This resulted in the ASM2-3P having a better match of
during the filtration process results in pH and alkalinity reduction, effluent OP and similar effluent ISS concentrations to those exper-

i.e.: NHþ4 þ 2O2 /NO3 þ H2 O þ 2H
þ
where the carbonate weak imentally measured.
acid/base system plays a significant role in the availability of Hþ To match the predicted P release in the diluted BT8 (in which no
ions for system pH adjustments i.e., 2HCO þ
3 þ 2H /2H2 CO3 . precipitation took place) the polyphosphate lysis rate (bPP) is

Hence the HCO3 which contributes significantly to the system reduced from the value of 0.2/d (from Henze et al., 2000) to a value
H2CO3 Alkalinity is lower measured (where ammonia to nitrate of 0.03/d. Therefore, it is notable from these low PP release kinetic
conversion took place during sample filtration) than the predicted constants, that in aerobic digestion the PAOs can be modelled to
value (where the virtually replicated experimental environment remain alive, still withhold their stored PP and release it at a
did not cater for the aeration during sample filtration. similarly slow rate as associated with PAO endogenous respiration.
Moreover, it was noted that the initial sparging of the ADS with This release behaviour is completely different to that in AD, where
CO2/CH4 did not affect the H2CO3 alkalinity, but instead allowed for all the PP was rapidly released in about 5e7 days (Harding et al.,
struvite precipitation at low pH and solubilisation of a calcium 2010; see Section below).
containing compound (possibly calcium carbonate, CaCO3). This The kinetic constants used for the endogenous respiration
decreased Mg but increased Ca concentrations respectively and was process of OHOs and PAO biomass allowed for simulation of VSS
taken into consideration during the model calibration process. concentrations that are reasonably well matched to those
Simulating this AerD of ADS process is taken as a step in the model measured for the organic degradation during the AerD BT8.
calibration exercise, whereby the focus is distinctly towards accu- Consequentially, there is also a close match in declining trends of
rate prediction of mineral precipitation potential, without the total kjeldahl nitrogen (TKN, of Fig. 3d), via reduction of the organic
interference of biological reactions. The closely matched results VSS bound N (OrgN). During this AerD BT8, nitrification is only the
provide reasonable evidence that the model is capable of repro- bioprocess simulated to utilise the ammonia (i.e., convert FSA to
ducing these results for similar reactor conditions, using the nitrates) that was made available from breakdown of OrgN. The
selected kinetic and stoichiometric constants. ammonia released (to aqueous phase) could not have been taken
up to precipitate struvite (MgNH4PO4, part of the solid phase TKN)
3.2. Bioprocesses calibration against Mebrahtu (2007) aerobic because if this had occurred, the measured TKN and ISS would be
batch test 8 data higher (i.e., would not match our simulated model predictions that
excluded occurrence of precipitation) and nitrates (which is not
Mebrahtu et al. (2007) carried out eight aerobic batch digestion measured in TKN) would be lower (See Fig. 3). Eliminating mineral
tests on WAS from the laboratory scale membrane (MBR) UCT precipitation potential in this experiment also assists towards cal-
NDBEPR, at a controlled pH of 7.2 and in the temperature regulated ibrating the AAD effluent OP prediction. However, further in-
(20  C) Water Research Laboratory of the University of Cape Town tricacies are brought about by the significant influence of aerobic PP
(UCT). The MBR UCT NDBEPR system that provided the WAS was disintegration on the AAD effluent OP (PP, which also contributes to
operated at steady state by Du Toit et al. (2010) at a sludge age of 20 the ISS, generates significantly more OP than the organically bound
days and a high aerobic reactor TSS concentration of about 20gTSS/ biomass P, which is released with the disintegration of biomass
l. Six of the eight batch tests were performed in a 5 l aerobic organic material) and effluent alkalinity. The aerobic disintegration
digester, while the last two batch tests were carried out in parallel of PP results in alkalinity reduction, which causes a significant
using two separate 3 l aerobic digesters, one with concentrated decline in the rate of nitrification. Mebrahtu (2007) added some
WAS (~20gTSS/l) and the other with WAS diluted 5 times (~4gTSS/ sodium bicarbonate to the aerobic batch digestion to ensure a pH
l). The influent COD concentration fed to the parent NDBEPR system value of around 7.2 is maintained. Without the knowledge of
was 800 mgCOD/l real wastewater, to which 200 mgCOD/l of ace- alkalinity added to the system, it would be difficult to calibrate the
tate was added to increase PAO growth. To maintain the 20-day PP breakdown rate without compromising the prediction of ni-
sludge age, 2.85 l/day of mixed liquor WAS was harvested from trates generated and vice versa. The simulated results indicate that
the aerobic reactor. Mebrahtu et al. (2007) noted that struvite a reasonable compromise has been made between the nitrates
precipitation took place in the concentrated batch tests (these were generated and OP released. The determined set of suitable pa-
the 1st 3rd and 7th of eight batch tests performed) and not in the rameters and their values that allowed the best match between
diluted batch test (BT 8). To calibrate the PWM_SA model towards experimental and simulated variables are given in Tables 2a and 2b
its capability of predicting the AerD bioprocesses (including nitri- Table 3 further describes the parameters that are noted to be sig-
fication, OP release, and endogenous respiration), without the in- nificant during the parameter estimation exercise as compared to
fluence of mineral precipitation, it is simulated against the the ones observed to be significant during the sensitivity analysis
conditions of this diluted (4gTSS) batch test (BT8). The results from process.
this calibration exercise are presented in Fig. 3 below.
The comparisons between the results simulated by the PWM_SA 3.3. Chemical, physical and biological processes calibrated against
dynamic model and measured data from the batch test (fed 20gTSS Musvoto (1998) and Vogts (2015) aerobic batch test data
NDBEPR WAS, but diluted 5 times to 4gTSS) exhibit close correla-
tions between the measured and predicted data. For model validation against chemical and physical processes, as
Prior to simulation, during the pre-processing of input compo- well as aerobic biological processes, Musvoto (1998), conducted
nents from the measured digester starting point data, it was noted further aerobic batch tests on a mixture of raw sewage and WAS
that, with PP present, the measured TP values seemed to indicate a (from a laboratory scale modified Ludzack Ettinger (MLE) system
lower than expected corresponding ISS concentration (PP contrib- ei.e., the PAO biomass and PP assumed to be at negligible quantity)
uting significantly to the ISS; Ekama and Wentzel., 2004). with addition of Ca, Mg, P and FSA. This mixture was aerated, but
8 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333

Fig. 3. (a to f): Simulation of orthophosphate (OP) release and nitrification during batch AerD of WAS from the laboratory scale membrane (MBR) UCT NDBEPR containing mixed
cultures of OHO and PAO biomass (with mineral precipitation excluded through dilution of the WAS for maintenance of low concentrations). The AerD batch test was operated by
Mebrahtu et al. (2007).

because the concentration of carbonate species in raw sewage is was added to one filtered sample to reduce the pH to below 2 and
very low, there was no increase in pH (it remained in the region of prevent further precipitation. This acidified sample was analysed
7.2). This allowed for the interaction of the precipitation of minerals for Ca, Mg and iron (Fe) by emission spectroscopy (using the
and the biological processes to be investigated. The OUR was inductively coupled plasma method) and total P (using persulphate
automatically measured with the OUR meter of Randall et al. (1991) digestion and molybdate/vanadate colour reaction; standard
because continuous aeration was not required to raise the pH. methods, 1985). The Fe was found to be low enough to be neglected.
During its operation, 100 ml and 10 ml samples were drawn at The other filtered sample was not acidified and was tested for total
intervals from the batch reactor for analysis. The analysis of 10 ml carbonate and short chain fatty acids (determined with measure-
sample included FSA before filtration e experience had shown that ment of H2CO3 alkalinity, volatile fatty acid (VFA) and pH with the
FSA loss occurs during filtration at high pH. The 100 ml sample was 5-point titration method of Moosbrugger et al., 1993). The PWM_SA
filtered immediately (through 0.45 mm filters) and the filtrate model simulations of this experimetal set up was matched against
divided into two. About 1.5 ml of concentrated nitric acid (HNO3) the experimental results of Musvoto (1998) to calibrate the model
 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333 9

Fig. 4. (a to f): Simulation of integrated biological (including endogenous respiration and nitrification) and chemical processes (including mineral precipitation) during batch AerD
on a mixture of raw sewage and WAS (from a laboratory scale MLE system ei.e., the PAO biomass and PP assumed to be at negligible quantity) with addition of Ca, Mg, P and FSA.
The measured data was obtained the study by Musvoto (1998).

processes against conditions where both the aerobic biological (in different concentration of this WAS - one fed concentrated (high
this case, excluding PAOs and PP but including organic removal and TSS) NDBEPR WAS (the 10 gTSS/l was thickened to 20 g/[ as feed to
nitrification) and chemical reactions (including mineral precipita- the high TSS AAD), the other fed dilute (low TSS) NDBEPR WAS (the
tion) were taking place. These results (Fig. 4) are discussed further low TSS AAD feed of 3 g/l was made by adding the NDBEPR AS
below together with the results from the similar set up by Vogts effluent to dilute the WAS to about 1/3rd). The two AADS were
(2015). operated continuously in order to observe the effect of intermittent
Vogts (2015) set up two anoxic-aerobic digesters (AADs), fed aeration cycle (3h aeration on, 3h aeration off) and P precipitation
WAS from the same NDBEPR UCT AS system as for Du Toit et al. on N and P removal in anoxic-aerobic digestion of P rich (including
(2010), but operated at a sludge age of 10 days and an aerobic PAOs and PP) WAS (Vogts, 2015).
reactor TSS concentration of about 10gTSS/l. The two AADs were fed Various batch tests were performed by Vogts (2015) on the
10 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333

influent to the AAD (i.e., the UCT WAS). To calibrate the model shows that rate of degradation via endogenous respiration is
further for replication of nitrification and P release kinetics, the accurately predicted. A further match with the oxygen utilisation
data from aerobic batch tests 2 and 4 (AerD BT2 and AerD BT4) from results (Fig. 4f) validates the model further since the total oxygen
Vogts (2015) was used. The AerD BT 2, with 2 L volume of ~3gTSS, demand is the output of three significant bioprocesses, i.e., sub-
was made from adding 700 ml of the 10gTSS WAS to 1.4 L of effluent strate utilisation, nitrification and endogenous respiration. Further,
and AerD BT4, with 2.5 L volume of ~2gTSS, was made from adding even with the inclusion of struvite precipitation (a process that also
500 ml of the 10gTSS WAS to 2 L of effluent. Both BT2 and BT4 were takes up ammonia), the measured nitrates concentration follows
kept aerobic (aerated) for their experimental period of 20 days and the same trend as the simulated values. In the results from Fig. 4, it
had 50m[ samples taken daily (distilled water was used to replace is noted that although nitrification and struvite formation both
the volume lost). The pH for BT4 was controlled at 7.3 to 7.6 with require ammonia, they both occurred to a given extent, according
measured addition of a 19.05 g/[ NaOH solution, while that of BT2 to the calibrated kinetics of the two processes. However as noted in
was not controlled with alkalinity addition and eventually dropped Fig. 4a and b, the magnesium and phosphorus precipitation
below 7. occurred rapidly at the beginning of the batch test and then
The same kinetic and stoichiometric constants (Tables 2 and 3) redissolution occurred allowing the Mg and OP to be released back
used for the calibration of the chemical processes (see Fig. 2) and in the aqueous phase (hence the decline and increase in their
the biological processes (see Fig. 3) were utilised in the simulation concentrations). However, the ammonia continuously declined,
of both processes (chemical and biological) occurring interactively and nitrate formation increased with the increased retention time.
to assess the extent to which potential non-linearity would affect This allows for some confidence towards the model replication of
the correlation to measured data (non-linearity is the result of in- the system performance in tracking N - i.e., according to the process
tricacies in output variables, due to parameter inter-relations kinetics, the right portions of ammonia released during organic
within the model processes). degradation goes towards struvite precipitation and nitrification
From the correlations exhibited in Figs. 4 and 5 it is noted that processes. Initially, it was expected that K-struvite (MgKPO4) rather
these model parameters continue to work well under the reactor than MgNH4PO4 would precipitate (with most of the NH4 used in
with integrated biological and chemical processes (i.e., including nitrification). However, measurements of the metals that would be
mineral precipitation). A good correlation of VSS (Fig. 5a) results made available with PP release (i.e., Mg, K and Ca) showed that only

Fig. 5. (a to d): Simulation of integrated biological (including endogenous respiration, OP release by PAOs and nitrification), physical and chemical processes during aerobic
digestion of nitrification-denitrification biological excess P removal (NDBEPR) WAS. The measured data was obtained from AerD BT2 and AerD BT4 of the study by Vogts (2015).
 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333 11

Mg and not K participated in the mineral precipitation, i.e., was released with uptake of VFA for polyhydroxy-butyrate formation;
transformed back from the aqueous to the solid phase. The pH Smolders et al., 1995) and then decreases afterwards in the aerobic
prediction, post precipitation, is very close to that measured phase (i.e., in the last 3 h, which indicate OP uptake to form PP;
because the model is safeguarded against decline in pH that would Smolders et al., 1995; Wentzel et al., 1990).
occur due to nitrification. Vogts (2015) performed further batch tests on the effluent to the
20 d R AAD to observe the denitrification performance and OP
3.4. Chemical, physical and biological processes calibrated against concentrations after 20d of aerobic digestion. This was the batch
Vogts (2015) anaerobic- aerobic and anoxic batch test data test 7 of Vogts (2015) that involved the setup of an anoxic batch
reactor (i.e., AxD BT7) comprising of diluted high TSS AAD effluent
The experimental data from Vogts (2015) was used further to (1 L of the AAD sludge effluent diluted with 1 L of “dewatering li-
calibrate the PWM_SA model against conditions of anaerobic - quor”, made up of 700m[ UCT effluent and 300m[ low TSS AAD (i.e.,
aerobic digestion and anoxic digestion of WAS. This was to ensure the one fed 3gTSS/l) supernatant, to create the 2l batch reactor
that accurate predictions could be achieved in cases where very volume). The mixture was high in nitrates and low in OP and the
little or no nitrification is occurring in the aerobic phase of the AAD. reactor was stirred without aeration (anoxic conditions due to ni-
According to known PAO metabolism reported by Wentzel et al. trates present) for 24 h, with samples taken at increasing time gaps,
(1990), the presence of high OP and acetate in alternating anaer- i.e. at 5, 10, 20, 30, 60min, 1.5h, 2.5h, 3.5h, 5.5h. The results in Fig. 7
obic eaerobic conditions could result in the growth PAO biomass, below show that the model was also able to replicate same trend of
rather than the exclusive OHO pre-domination. This meant that nitrate denitrification and OP as the system operated by Vogts
simulating AADs containing mixed cultures of OHO and PAO (2015). It can be noted that the nitrates decrease (due to denitrifi-
biomass, requires careful monitoring of the fermentation rate cation is replicated well) together with the OP concentration (none
(Q_OHO_F_VFA; see Tables 2 and 3) to control the quantity of acetate was released or taken up in this experiment). The consistency in OP
produced, hence direct PAO growth (and the consequential PP concentrations may be due to significant quantities having already
storage). For further investigations on this aspect, Vogts (2015) set been released during the AAD process and the short retention time
up anaerobic-aerobic batch reactor, to calibrate anaerobic P release (7 h) available during the experiment to observe any release (rapid
with acetate uptake and aerobic P uptake after 20d aerobic diges- OP releases usually occur in anaerobic zones with availability of
tion (i.e., the data is obtained from the anaerobic-aerobic batch VFAs, while the OP release with breakdown of PAOs is expected to
digester 8; AnAerD BT8). This reactor was made up of 1[ of the high be slow).
TSS (i.e., fed the NDBEPR WAS thickened form 10 to 20gTSS/l) AAD
effluent mixed 1[ of distilled water with a measured dose of sodium 3.5. Review of observed P release during batch-anaerobic digestion
acetate (5 gCH3COONa - this was 1952 mgCOD/[ in the 2[ final (AD) of NDBEPR WAS
batch test volume) dissolved in it. The batch reactor was mixed
without aeration (anaerobic conditions - no nitrates present) for 3h To observe the rate of phosphorus release in AD, Harding et al.
and then aeration was switched on for 3h (i.e., the test was carried (2010) conducted experimental AD batch tests on the NDBEPR
out over 6h). During the 6h batch test, hourly samples were taken WAS. These batch tests were performed using a 5 L continuously
and the reactor pH was monitored and maintained between 7.3 and stirred mesophilic (35  C) AD reactor. This reactor was seeded with
7.7, by adding measured doses of 1N NaOH and HC[. The results a 50:50 ratio mixture of NDBEPR WAS to effluent waste sludge from
(VFA, which is taken up, and OP, which is expended to be released) the continuous AD system described above, for the 18-day Rs
from simulation of this experimental set up are shown in Fig. 6 (it is operation. One batch test was conducted on undiluted NDBEPR
noticeable here that the model can reproduce the data within WAS (~10gTSS/l), the other was diluted five times. Fig. 1 below
reasonable range). shows a comparison between the results measured in the undi-
From Fig. 6 above, it is notable that the VFA is being utilised luted batch test and results simulated with the three phase AD
(decreasing over time) in both the anaerobic and aerobic phases, model, with and without struvite precipitation.
while the OP increases during the anaerobic phase (first 3 h; OP Fig. 8 a to c show that a closer match between experimental and

Fig. 6. 6a and 6b: Simulation of anaerobic OP release with PAO biomass anaerobic acetate uptake and aerobic OP uptake for PP formation using data from the anaerobic-aerobic
batch digester 8 (AnAerD BT8) of Vogts (2015).
12 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333

Fig. 7. Simulation of OP (observed to stay consistent) with denitrification, using the anoxic digestion batch test 7 (AxD BT7) of Vogts (2015).

Fig. 8. (a to c): Simulated results for a batch AD fed P-rich WAS (from a UCT NDEBPR AS system) blended with mixed liquor from a continuous AD system (also treating NDEPR WAS)
in 50:50 ratio.

simulated results is made, when including struvite precipitation (multiplied by the dilution factor), while the OP simulated with
than without precipitation. Most of the P is released (as ortho- inclusion of struvite precipitation matches the OP measured in the
phosphate, OP) within 4 days, after which the P release, if any undiluted batch test. This validates the capability of the model to
was very slow or gradual. This conforms to the observation by predict mineral precipitation during AD. The other graphs which
Jardin and Popel (1994), confirming that stored polyphosphate in show increased TSS (Fig. 1b, since struvite contributes to inorganic
PAOs (which forms a large portion of the TP content) is released solids concentration in the AD) provide further support of the
within seven days. The OP simulated for infinite solubility of stru- reduced simulated OP being due to struvite precipitation. The re-
vite matches well with that measured in the diluted batch test sults match well when the kinetics of PP release are modelled such
 D.S. Ikumi, T.H. Harding / Water Research 170 (2020) 115333 13

that most of the PP disintegrates rapidly, with PHA uptake, as would Town, and is published with their permission.
occur in the anaerobic phase of the AS system. However, further
work is still required to compare whether the rates for PP release in References
anaerobic phase AS is actually equal to that in the AD and also to
check the differences in alkalinity and pH brought about by the Banjade, Sarita, 2008. Anaerobic/Aerobic Digestion for Enhanced Solids and Nitro-
gen Removal. Master’s Thesis. Virginia Polytechnic Institute and State
different reactor conditions (Ikumi and Ekama, 2019). In the AD, the University.
OHO and PAO biomass were allowed have rapid death rates and the Brouckaert, C.J., Ikumi, D.S., Ekama, G.A., 2010. A 3-phase anaerobic digestion
resulting biodegradable organic particulates (BPO) are hydrolysed model. In: Proceedings. 12th IWA Anaerobic Digestion Conference (AD12).
Guadalajara, Mexico, 1-4 Nov, 2010.
much slower. The kinetics of this biomass BPO hydrolysis has been Brun, R., Kühni, M., Siegrist, H., Gujer, W., Reichert, P., 2002. Practical identifiability
calibrated (Ikumi et al., 2014). However, further validation towards of ASM2d parametersdsystematic selection and tuning of parameter subsets.
the kinetics of PAO biomass death in AD and the resulting endog- Water Res. 36, 4113e4127.
Christensen, D.R., Gerick, J.A., Eblen, J.E., 1984. Design and operation of an upflow
enous mass fraction is required. To observe this endogenous death
anaerobic sludge blanket reactor. J. Water Pollut. Control Fed. 56 (9),
rate and unbiodegradable endogenous fraction could be achieved 1059e1106.
by further experimental observation of enhanced PAO cultures, Dold, P.L., Ekama, G.A., Marais, G.vR., 1980. A general model for the activated sludge
when fed to the AD environments. process. Prog. Water Technol. 12 (Tor), 47e77.
Dold, P.L., Marais, G., 1986. Evaluation of the general activated sludge model pro-
posed by the IAWPRC Task Group. Water Sci. Technol. 18, 63e89.
4. Conclusion Dold, P.L., Wentzel, M.C., Billing, A.E., Ekama, G.A., Marais, G.vR., 1991. Activated
Sludge Simulation Programs. Water Research Commission, Private Bag X03.
Gezina, 0031, South Africa.
The experiment al data obtained from literature has assisted in Du Toit, G.J.G., Parco, V., Ramphao, M., Wentzel, M.C., Lakay, M.T., Mafungwa, H.,
the stepwise calibration of the PWMSA model for aerobic (AerD) Ekama, G.A., 2010. The Performance and Kinetics of Biological Nitrogen and
and anoxic-aerobic digestion (AAD) processes. The calibrated ki- Phosphorus Removal with Ultra-filtration Membranes for Solid-Liquid Separa-
tion. WRC Report No. 1537/1/09. Dept. of Civil Eng., University of Cape Town,
netic parameters allowed for the model capability of reproducing Rondebosch, Cape Town, South Africa, p. 7701.
the data from the biological processes occurring in the various Ekama, G., 2017. Optimizing water and resource recovery facilities (WRRF) for en-
environments of sludge treatment (aerobic, anoxic and anaerobic) ergy generation without compromising effluent quality. In: IWA 2017
Conference.
within satisfactory level of accuracy. This includes. Ekama, G.A., Wentzel, M.C., 2004. A predictive model for the reactor inorganic
suspended solids concentration in activated sludge systems. Water Res. 38,
 Accurately predicting relevant outputs of VSS removal and ox- 4093e4106.
Harding, T.H., Ikumi, D.S., Ekama, G.A., 2010. In: A Steady State Stoichiometric Model
ygen utilisation, which are intricately related to the significant
Describing the Anaerobic Digestion of Biological Excess Phosphorus Removal
bioprocesses of substrate utilisation, nitrification and endoge- Waste Activated Sludge. Research Report W132 (MSc Thesis), vol. 7701. Dept. of
nous respiration occurring during the AAD. Civil Eng., Univ of Cape Town, Rondebosch, Cape Town, South Africa.
 Polyphosphate (PP) lysis kinetics e it was notable that during Hauduc, H., Rieger, L., Tak duit, A., Vanrolleghem, P.A., Gillot, S., 2010.
acs, I., He
A systematic approach for model verification: application on seven published
AerD, the PP can be modelled to remain withheld by the PAOs activated sludge models. Water Sci. Technol. 61 (4), 825e839.
while alive and released with the PAO lysis at a similarly slow Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M.C., Marais, G.v.R., 1995.
rate (bpp ¼ 0.03/d) to the PAO endogenous respiration (0.04/d). Activated Sludge Model No.2 (ASM2). IWA Scientific and Technical Report No.3.
IWA Publishing, London, U.K.
This release behaviour is completely different to that in AD, Henze, M., Gujer, W., Mino, T., van Loosdrecht, M.C.M., 2000. Activated Sludge
where almost all the PP was released in about 5e7 days Models: ASM1, ASM2, ASM2d and ASM3. Scientific and Technical Report No. 9.
(Harding et al., 2010). IWA Publishing, London. U.K.
Ikumi, D.S., Brouckaert, C.J., Ekama, G.A., 2011. A 3 phase anaerobic digestion model.
 Model replication of the system performance in tracking N - i.e., In: Proceedings 8th IWA Watermatex Conference, pp. 20e22. San Sebastian,
according to the process kinetics, the right portions of ammonia Spain June, 2011.
released during organic degradation goes towards struvite Ikumi, D.S., 2011. The Development of a Three Phase Plant-wide Mathematical
Model for Sewage Treatment. Water Research Group (WRG). Department of
(MgNH4PO4, part of the solid phase TKN) precipitation and
Civil Engineering. University of Cape Town, South-Africa.
nitrification (nitrates formed not being part of TKN) processes. Ikumi, D., Vanrolleghem, P.A., Bouckaert, V., Neumann, M., Ekama, G., 2014. Towards
Although in AerD and AAD some of the NH4 is used in nitrifi- calibration of phosphorus (P) removal plant-wide models. In: Proc. 4th IWA/
WEF Wastewater Treatment Modelling Seminar (WWTmod2014). Spa, Belgium,
cation, N-struvite (MgNH4PO4) is still likely to precipitate, with
pp. 197e206. March 30 - April 2 2014.
K-struvite (MgKPO4) precipitation not observed experimentally. Ikumi, D.S., Harding, T.H., Vogts, M., Lakay, M.T., Mafungwa, H.Z., Brouckaert, C.J.,
 Accurate prediction of the nitrate denitrification and OP release Ekama, G.A., 2015. Mass Balances Modelling over Wastewater Treatment Plants III.
in anoxic phase of AAD. WRC Report No. 1822/1/14. Water Research Commission, Pretoria.
Ikumi, D.S., Harding, T.H., Ekama, G.A., 2014. Plant-wide wastewater treatment
 Accurate prediction of the OP release in anaerobic periods. modelling (1) e biodegradability of wastewater and activated sludge organics
However, it is noted that further work may be required to vali- in anaerobic digestion. Water Res. 56 (1), 267e279.
date PP release in AD systems e including observation of the Ikumi, D.S., Ekama, G.A., 2019. Plant-wide wastewater treatment modelling (1) e
biodegradability of wastewater and activated sludge organics in anaerobic
impact of the process on system pH and selecting the correct digestion. Water S.A. 31, 545e568.
stoichiometric path and kinetics for accurate prediction of sys- Jamil, M., Qasim, M., Umar, M., 2006. Utilization of sewage sludge on organic fer-
tem performance. Experimentally observing the AD of PAO tilizer in sustainable agriculture. J. Appl. Sci. 6 (3), 531e535. https://doi.org/
10.3923/jas.2006.531.535.
biomass enhanced cultures may be able to address this aspect. €pel, H.J., 1994. Phosphate release of sludges from enhanced biological
Jardin, N., Po
P-removal during digestion. Water Sci. Technol. 30 (6), 281e292.
Declaration of competing interest Kominko, H., Gorazda, K., Wzorek, Z., Wojtas, K., 2018. Sustainable management of
sewage sludge for the production of organo-mineral fertilizers. Waste Biomass
Valorization 9 (10), 1817e1826. https://doi.org/10.1007/s12649-017-9942-9.
The authors declare that they have no known competing Macura, B., Johannesdottir, S.L., Piniewski, M., Haddaway, N.R., Kvarnstro €m, E., 2019.
financial interests or personal relationships that could have Effectiveness of ecotechnologies for recovery of nitrogen and phosphorus from
anaerobic digestate and effectiveness of the recovery products as fertilisers: a
appeared to influence the work reported in this paper.
systematic review protocol. Environ. Evid. 8 (1), 29. https://doi.org/10.1186/
s13750-019-0173-3.
Acknowledgements Mebrahtu, M., 2007. Aerobic Digestion of Waste Activated Sludge from Biological
Nutrient Removal Activated Sludge Systems. Water Research Group (WRG).
Department of Civil Engineering. University of Cape, Town, South-Africa.
This research was supported by the Water Research Commis- Mebrahtu, M.K., Wentzel, M.C., Ekama, G.A., 2007. Aerobic Digestion of Waste Acti-
sion, the National Research Foundation and the University of Cape vated Sludge from Biological N Ad P Removal Systems. Research Report No. W126.
 Chemosphere 252 (2020) 126513

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

In situ textile wastewater treatment in high rate transpiration system

furrows planted with aquatic macrophytes and floating phytobeds
Vishal Chandanshive a, Suhas Kadam a, Niraj Rane b, Byong-Hun Jeon c, **, Jyoti Jadhav d,
Sanjay Govindwar a, c, *
a
Department of Biochemistry, Shivaji University, Kolhapur, 416004, India
b
Savitribai Phule Pune University, Pune, 411007, India
c
Department of Earth Resources and Environmental Engineering, Hanyang University, Seoul, 04763, South Korea
d
Department of Biotechnology, Shivaji University, Kolhapur, 416004, India

h i g h l i g h t s g r a p h i c a l a b s t r a c t

Remazol Red, dye mixture and textile

effluent treated efficiently by
V. zizanioides.

Histological analysis confirmed entry
and degradation of dyes in roots.

Floating phyto-bed reactor treated
textile wastewater effectively.

Textile wastewater was noteworthily
treated in plant cultivated furrows of
HRTS.

Plant consortium enhanced the po-
tential of textile dye removal.

a r t i c l e i n f o a b s t r a c t

Article history: Plants are known to remediate dyes, metals and emerging contaminants from wastewaters. Vetiveria
Received 5 November 2019 zizanioides, a perennial bunchgrass showed removal of Remazol Red (RR, 100 mg/L) up to 93% within
Received in revised form 40 h. Root and shoot tissues of V. zizanioides revealed induction in dye degrading enzymes viz. lignin
8 March 2020
peroxidase by 2.28 and 1.43, veratryl alcohol oxidase 2.72 and 1.60, laccase 6.15 and 3.55, and azo
Accepted 14 March 2020
reductase 2.17 and 2.65-fold, respectively, during RR decolorization. Substantial increase was observed in
Available online 16 March 2020
the contents of chlorophyll a, chlorophyll b, and carotenoids in the plant leaves during treatment.
Handling Editor: T Cutright Anatomical studies of roots, HPLC and GC-MS analysis of metabolites, and phytotoxicity assessment
confirmed phytotransformation of RR into nontoxic metabolites. Floating phytobed with V. zizanioides
Keywords: treated textile wastewater (400 L) effectively and reduced ADMI, COD, BOD, TDS, and TSS by 74, 74, 81, 66
Phytoremediation and 47%, respectively within 72 h. In-situ treatment of textile wastewater for 5 days in constructed
Constructed furrows furrows planted with semiaquatic plants, V. zizanioides, Ipomoea aquatica and its consortium-VI
Vetiveria zizanioides decreased ADMI by 68, 61 and 76%, COD by 75, 74 and 79%, BOD by 73, 71 and 84%, TDS by 77, 75
Oxido-Reductive enzymes
and 83%, and TSS by 34, 31 and 51%, respectively. This treatment was also useful to remove arsenic,
Textile wastewater
cadmium, chromium and lead from wastewater. Overall observation suggests wise strategy to use this
plantation in the furrows of high rate transpiration system and phytobeds in deep water for textile
wastewater treatment.
© 2020 Elsevier Ltd. All rights reserved.

* Corresponding author. Department of Earth Resources and Environmental Engineering, Hanyang University, Seoul, 04763, South Korea.
** Corresponding author.
E-mail addresses: [email protected] (B.-H. Jeon), [email protected] (S. Govindwar).

https://doi.org/10.1016/j.chemosphere.2020.126513
0045-6535/© 2020 Elsevier Ltd. All rights reserved.
2 V. Chandanshive et al. / Chemosphere 252 (2020) 126513

1. Introduction Ambewadi, Kolhapur, Maharashtra. Riboflavin and 2, 2-azino-bis

(3-ethylbenzothiazoline)-6-sulphonic acid (ABTS) were purchased
The dyes are extensively used in a variety of industries like from Sigma Aldrich (St Louis, MO, USA). Veratryl alcohol, nicotin-
plastic, textile, pulp, rubber and paper industries. Among all in- amide adenine dinucleotide (di-sodium salt), n-propanol, 2, 6-
dustries, the textile is at first rank in the usage of dyes for coloration dichlorophenol indophenol (DCIP) and catechol were procured
of fiber. The effluent containing textile dyes are the largest from Sisco Research Laboratories, Mumbai, India. The textile dyes
pollutant source of valuable water bodies due to its carcinogenic, Remazol Red, Scarlet RR, Rubine GFL, Methyl Orange, Disperse
allergic, cytotoxic and mutagenic nature. The present physico- EMGR, Navy Blue Ex and Scarlet GDR were acquired from Mahesh
chemical methods such as, flocculation, adsorption, filtration, dye processors, Ichalkaranji. The textile wastewater was acquired
coagulation, photo-degradation, membrane processes, reverse from Common Effluent Treatment Plant (CETP), Kagal, India. An
osmosis and chemical oxidation are efficient in color removal, but analytical grade and highly purified chemicals were used for
they have some limitations like high cost, limited applications and experiments.
secondary pollution problems such as sludge formation, toxic gases
etc. To comply with environmental laws, it is necessary to look for 2.2. Textile dye decolorization by V. zizanioides
less sludge producing, eco-friendly, in situ applicable and low-cost
methods for dye wastewater treatment. Wild plants of V. zizanioides were used after removing soil from
Biological mode of the treatment mineralizes toxic compounds the roots for decolorization experiments. Two plants were sub-
into an organic form which are nontoxic to life (Tahir et al., 2016). jected to Remazol Red, Scarlet RR, Rubine GFL, Methyl Orange,
Presence of high amount of toxic dyes has become environmental Disperse EMGR, Navy Blue Ex and Scarlet GDR solutions (100 mg/L)
concern that affect agriculture and impacting on accumulation and in Erlenmeyer flasks. Samples (1 mL) were collected at every 12 h.
magnification of the components in food chain (Dipu et al., 2011). The samples were centrifuged for 10 min at 4561g. Remazol Red
Phytoremediation used as a novel approach in dye degradation concentration was assessed at 530 nm using UVevisible spectro-
since last few years; it deals with the use of plant and rhizospheric photometer. Remazol Red was used for further experiments,
microorganisms to exclude toxic compounds from the contami- because it took less time for the decolorization than other dyes.
nated sites. Phytoremediation can be used for treatment of different In another experiment, V. zizanioides was used for the decolor-
kinds of contaminant, such as landfill leachates, explosives, chlo- ization of textile wastewater. Tristimulus filter method was used to
rinated solvents, radionuclides, polyaromatic hydrocarbons, heavy measure the color before and after treatment of textile wastewater.
metals, and pesticides (Khandare and Govindwar, 2015). Although, Percent decolorization was computed using ADMI color removal.
phytoremediation potential of many plants was studied in textile Textile wastewater and Remazol Red solution were characterized
wastewater treatment, plenty of them remained at lab scale. Hy- before and after treatment using parameters like chemical oxygen
droponic phyto-tunnel (HPT) system has shown to be an efficient demand (COD), biological oxygen demand (BOD), color value
textile wastewater treatment (Khandare et al., 2013). Laboratory (American dye manufacturers institute, ADMI) (Chandanshive
scale plant-bacterial consortia reactors such as vertical (VSbF) and et al., 2017), total suspended solid (TSS) and total dissolved solid
horizontal subsurface flow, (HSbF) were built for textile wastewater (TDS) (APHA, 1998). Atomic absorption spectrophotometer (AAS)
treatment (Khandare and Govindwar, 2015). Some pilot scale sys- was used to estimate the concentration of heavy metals
tems with aquatic plants (Chrysopogon zizanioides, Typha domi- (Chandanshive et al., 2017). The average mean value of three ex-
ngensis and Phragmites australis) were reported to treat textile periments was considered for inference. Abiotic controls of dye and
wastewater (Ong et al., 2010; Shehzadi et al., 2014). However, only a textile wastewater was devoid of the plant.
few studies have been recorded for actual dye contaminated site
remediation. Constructed wetlands with individual plants (Salvinia 2.3. Analysis of photosynthetic pigments and plant histology
molesta, Ipomoea aquatica, Typha angustifolia and Fimbristylis
dichotoma) were proposed for dye wastewater management (Rane Analysis of chlorophylls and carotenoids is important to un-
et al., 2016; Chandanshive et al., 2017; Kadam et al., 2017). Co- derstand the health and energy requirement of the plants. Two
plantation was advantageous in waste wastewater remediation gram leaves V. zizanioides plant (Control and treated) were ho-
because of synergistic effect of oxido-reductases from the rhizo- mogenized using mortar and pestle, separately. Leaves were
spheric and individual plants. crushed in 50 mL of cold acetone (80%) with bit amount of MgCO3.
This work describes on field application of Vetiveria zizanioides, I. The obtained homogenate was filtered and centrifuged (2000g)
aquatica and its consortium-VI when planted in constructed fur- for 10 min. The chlorophylls were quantified spectrophotometri-
rows of high rate transpiration system (HRTS), MIDC, Kagal, India cally at 663 nm and 645 nm; however, carotenoids were measured
for textile wastewater treatment. Both the plants are annual herb at 470 nm using acetone as blank (Chandanshive et al., 2017). Root
and able to sustain the presence of waterlogged condition. Textile transverse section mounted in glycerine was examined for accu-
wastewater was also treated with phyto-beds possessing these mulation of dye at 100X magnification using a Trinocular Micro-
plants, which can further be used in furrows, where plantation is scope (Zeiss Axio Imager 2).
not possible.
2.4. Preparation of enzyme extracts
2. Materials and methods
Roots and shoots of V. zizanioides from control and after decol-
2.1. Plant material and chemicals orization of the textile wastewater were taken to prepare crude
enzyme extract (Chandanshive et al., 2017). Roots and shoots
Aestivum sativum and Phaseolus mungo seeds were bought from (2 gm) were chopped and grounded using morter and pestle in
a local grain market. V. zizanioides was recovered from Botany Lake, 50 mM potassium phosphate buffer at pH 7.4. Obtained sample was
Shivaji University, Kolhapur; while I. aquatica was collected from centrifuged at 9000g for 20 min in a cold centrifuge at 4  C.
 V. Chandanshive et al. / Chemosphere 252 (2020) 126513 3

Enzyme assayes of dye degrading enzymes were performed at special nutrients. The samples were collected for 5 d at the interval
room temperature using spectrophotometer. Laccase activity was of 24 h.
assessed at 420 nm using ABTS as a substrate (Hatvani and Mecs, Samples (1 mL) of textile wastewater were withdrawn at 0 time
2001). Veratryl alcohol oxidase was assessed using a substrate, and after phytoremediation. Bacterial count was measured using
veratryl alcohol at 300 nm (Jadhav et al., 2009). Activity of lignin serial dilution (with 0.9% NaCl) technique. The obtained samples
peroxidase was assessed at 300 nm by measuring propionaldehyde were grown independently in Petri dishes containing nutrient agar
formation due to n-propanol oxidation (Shanmugam et al., 1999). medium at 37  C. The colony forming units (CFUs) were measured
Azo, riboflavin and DCIP reductase were estimated measuring at after 24 h.
430, 340 and 600 nm, respectively (Chandanshive et al., 2017).
Triplicate experments were performed for all enzymes. Lowry 2.9. Statistical analysis
method was used to estimate protein concentration in the enzyme
extract (Lowry et al., 1951). Data were analysed using (ANOVA) one way analysis of variance
and Tukey-Kramer comparisons test, considering significant level
2.5. Extraction and analyses of phytodegraded products P < 0.05.

Phytodegraded product were extracted taking equal amount of 3. Results and discussion
ethyl acetate. The extracted samples were dried by evaporating at
room temperature and dissolved in methanol (HPLC grade) 3.1. Textile dyes decolorization using V. zizanioides
(Chandanshive et al., 2016). Metabolites separation and identifica-
tion were done using HPLC, GC-MS, and FTIR (Kagalkar et al., 2015). Azo dyes such as Scarlett RR, Green HE4B, Navy Blue HER and
NIST library and mass spectram (m/z) was used to idetify the Rubine GFL decolorized by 45, 57, 39 and 62%, respectively; due to
metabolites. the treatment of long rooted wild plant of V. zizanioides within 40 h.
However, Remazol Red was decolorized up to 93% within 40 h (data
2.6. Toxicity analysis not shown). Variation in percent degradation of these dyes might
be due to structure of the dyes. Wild plants such as S. molesta,
Toxicity of Remazol Red and its products produced by Ipomoea hederifolia, F. dichtomas and Asparagus densiflorus were
V. zizanioides treatment were independently tested on A. sativum found to decolorize dyes Rubine GFL (97%), Scarlett RR (96%),
and P. mungo seeds. Fifty seeds were watered for 7 days with 5 mL Methyl Orange (91%) and Rubine GFL (91%) within 72, 60, 60 and
of Remazol Red (100 mg/L) and metabolite solution separately. 48 h, respectively (Kadam et al., 2017; Watharkar et al., 2018).
Control set of seeds was watered with distilled water. The root and Bouteloua dactyloides and T. angustifolia were found to efficiently
shoot lengths were measured on 7th days. decolorize the Reactive Blue 19 and textile effluent, respectively
(Mahmood et al., 2005; Vijayalakshmidevi and Muthukumar, 2015).
2.7. Construction of floating phyto-bed reactor of V. zizanioides V. zizanioides also promote the biodegradation of organic com-
pounds such as phenol, 2,4,6-trinitroluene, atrazine and benzo[a]
Green-remediation process of textile wastewater was per- pyrene (Danh et al., 2009). A. philoxeroides also displayed a decol-
formed using floating phyto-bed in rectangle metal tank (400 L, ourization potential of Reactive Green up to 3000 mg/L concen-
dimensions: 1.2 m  0.61 m x 0.61 m). The floating phyto-beds tration within 60 days treatment, whereas, association of Klebsiella
(1.0  0.58 m) were prepared using PVC pipes, elbows, thermacol sp. VITAJ23 was found more efficient (Sinha et al., 2019). Cerato-
sheet and aluminium metal wire gauze. The aluminium wire guaze phylum demersum showed phytoremediation of Reactive Black,
holds the plants and allow roots to enter in the wastewater tank. Reactive Red and Reactive Brown to the extent of 76e84% at 50 ppm
Twenty-four V. zizanioides plants were planted in a PVC reducer concentration (Priyanka and Krishnaswamy, 2019). Lemna minor
along with 100 g of soil making appropriate holes on the thermacol decolorized triphenylmethane dyes (Crystal Violet and Malachite
sheet. Initially, floating phyto-bed was kept floating in tap water for Green) significantly during phytoremediation process (To €ro
€;k et al.,
3 months to develop the root system. Then, developed floating beds 2015). Bacopa monnieri (L.) Pennell showed potential to degrade
were exposed to textile wastewater. The samples were withdrawn fourteen azo dyes at 40 mg/L concentration by 90e100% within 2
at 24 h interval and analysed for environmental parameters (APHA, weeks (Shanmugam et al., 2020).
1998). Treatment of simulated dye mixture and textile wastewater by
V. zizanioides decreased ADMI color value up to 74 and 73%,
2.8. Textile wastewater treatment in planted furrows respectively within 40 h. COD, BOD, TSS, TDS and turbidity were
reduced by 73 and 73%, 75 and 78%, 58 and 68%, 55 and 59%, 57 and
Furrows (91.4 m  1.2 m  0.6 m) were constructed in HRTS to 60% after 40 h treatment of simulated dye mixture and textile
treat textile wastewater. This furrow was mulched correctly using wastewater, respectively with V. zizanioides. The pH of the solution
mulching paper to reduce seepage of textile wastewater. Three was decreased to 7.7 and 7.3 from 10.6 to 9.8, respectively, after the
furrows were planted independently with V. zizanioides, I. aquatica treatment of textile wastewater and simulated dye mixture by
and in consortium. V. zizanioides was planted 15 cm away from each V. zizanioides. The electric conductivity of textile wastewater and
other at the bottom of furrow (49 plants/m2), because it is a sub- simulated dye mixture was found to increase by 3.48 and 5.52-fold,
merged rooted plant. Whilst, I. aquatica, a free-floating aquatic respectively by the exposure of V. zizanioides (Table 1). Plants like
plant was planted on the upper edges of furrow which further S. molesta (in lagoon (52,500 L) treatment), T. angustifolia (planted
covered surface of water. Initially, plants were grown watering tap in furrows of HRTS) and I. hederifolia (in flask studies) were found to
water for three months to develop and sustain in the environment. significantly reduce ADMI, BOD, COD, TSS, TDS and electric con-
The planted furrow was then independently supplied with textile ductivity after treatment of dye mixture and textile effluent (Rane
wastewater in a stagnant state. The experiments were carried out in et al., 2016; Chandanshive et al., 2016). Textile wastewater treat-
winter season (September to January) at atmospheric conditions ment with Gaillardia grandiflora, Tagetes patula, Aster amellus, and
such as, temperature range- 15e25  C, light- 11/13 h (light/dark Portulaca grandiflora showed decolorization of dye after 30 d of
cycle), CETP textile effluent pH range 7e10, and without adding treatment at HRTS (Patil and Jadhav, 2013; Chandanshive et al.,
4 V. Chandanshive et al. / Chemosphere 252 (2020) 126513

2018). C. zizanioides L. treated textile industry wastewater and Several reports showed an increase in pigments when exposed to
showed removal of BOD (98.47%) and COD (89.05%) (Tambunan textile effluent in phytoreactors/lagoons (Chandanshive et al.,
et al., 2018). Eichhornia crassipes, Pistia stratiotes and S. molesta 2018).
were found effective in removing pollutants, COD, TDS, nutrients Histological study of root tissues of V. zizanioides was performed
and metals (Cd, Ni and Zn) from textile wastewater at 10 and 50 h for understanding the association of dye Remazol
(Wickramasinghe and Jayawardana, 2018). Red in plant tissue and to locate phytodegradation site. Accumu-
lation of dye was found at adjacent cortical cells with an extended
3.2. Enzyme induction in roots and stems of V. zizanioides time period. The plant of V. zizanioides revealed accretion and
transformation of dye in plant tissues. Cells of control (untreated)
Phytoremediation possesses accumulation, degradation, plant root were normal in size and shapes without having any color
adsorption and biotransformation of contaminant through meta- (Suppl. Fig. 1a). Remazol Red was accumulated in epidermal cells of
bolic processes of the plant. Root and shoot tissues were separately plant roots after 10 h of dye treatment (Suppl. Fig. 1b) and further
studied for a variety of oxido-reductive enzymes involved in the exposure affected to the cortical region at 20 h (Suppl. Fig. 1c). Dye
dye phytodegradation process. The activities of the degrading en- was accumulated at both the cortical and epidermal regions at 30 h
zymes such as veratryl alcohol oxidase, lignin peroxidase, laccase (Suppl. Fig. 1d). Both cortical and epidermal cells showed degra-
and azo reductase were increased by 2.72, 2.18, 6.15 and 2.17-fold in dation of dye at 40 h (Suppl. Fig. 1e); while phloem and xylem was
root tissues; 1.6, 1.43, 3.55 and 2.65-fold in stem tissues of containing some traces of dye. These results signify the phytode-
V. zizanioides, respectively, due to exposure of Remazol Red. Root gradation of dye in the cortical and epidermal region. Further
and stem tissues demonstrated a reduction in the activity of ribo- exposure of plants to the tap water for 10 h, histology of root did not
flavin reductase and NADH-DCIP reductase (Table 2). The contri- find color in epidermal cells and found some distorted cells in
bution of these enzymes is proposed in I. aquatica (Rane et al., cortical containing a minute traces of dye (Suppl. Fig. 1f).
2016). Involvement of dye degrading enzyme was the key cause T. angustifolia and Paspalum scrobiculatum root epidermis as well as
behind better dye degradation and removal. Additionally, it always cortical tissue showed the presence of Congo Red after 24 h and
showed induction in the enzyme activities of lignin peroxidase, subsequently phytotransformed at 48 h (Chandanshive et al., 2018).
veratryl alcohol oxidase, laccase and azo reductase during the
decolorization process. These oxido-reductive enzyme plays an 3.4. Products analysis
important role in biotransformation of textile dyes. Because of
these reasons, V. zizanioides might have shown a better dye removal Remazol Red indicated peaks at retention time of 1.565, 1.657,
potential from the textile wastewater. The participation of these 1.846, 2.462, 3.426, 3.864, 4.556 and 5.554 min (Fig. 1a) in HPLC
enzymes has been reported earlier in the degradation of several spectrum, while new peaks were observed at 1.498, 1.653, 1.748 and
textile dyes like Scarlett RR, Rubine GFL, Methyl Orange and Congo 6.636 min after treatment (Fig. 1b). Simulated dye mixture indi-
Red by plants I. hederifolia, S. molesta, F. dichotomas and cated peaks at 1.732, 2.118, 2.564, 4.018 and 4.452 min (Fig. 1c).
T. angustifolia, respectively (Rane et al., 2016; Kadam et al., 2017; Some distinct peaks were appeared at 1.552, 1.792, 2.532 and
Chandanshive et al., 2017). Similarly, Ipomea palmata and Phrag- 3.428 min after the treatment of simulated dye mixture (Fig. 1d).
mites australis induced peroxidase enzyme in decolorization pro- Untreated textile wastewater indicated peaks at retention time of
cess of textile dyes (Shaffiqu et al., 2002; Caries et al., 2007). 1.741, 2.446, 3.087, 3.826, 4.475, 4.922 and 11.784 min (Fig. 1e).
However, treated textile wastewater showed some distinct peaks at
3.3. Photosynthetic pigment and anatomical analyses of plants 1.568, 1.745, 1.921, 2.044, 3.602, 3.925, 4.933, 5.657 and 11.780 min
(Fig. 1f). The HPLC peak pattern of untreated and treated dye,
Photosynthetic pigments (Chlorophyll a, Chlorophyll b and simulated dye mixture and textile wastewater demonstrated the
carotenoid) are involved in the cell energy metabolism, hence, it is variation in retention time suggesting dye degradation.
essential to know the effect of dyes on the level of these pigments. GC-MS analysis of treated Remazol Red was used to propose
Plants V. zizanioides showed elevated levels by 41% in chlorophyll a; chemical nature and structure of extracted metabolites. The
24% in chlorophyll b, and 36% in carotenoids, as compared to un- biotransformation of Remazol Red was projected based on the
treated plants (control) (Suppl. Table 1). V. zizanioides is a long formed metabolites and enhanced enzyme activities in the root and
rooted macrophyte, might synthesize additional amount of chlo- stem tissues of V. zizanioides. Asymmetric break down of Remazol
rophylls and carotenoids to complete the energy requirement and Red through the action of laccase or lignin peroxidase formed in-
metabolic activities during phytoremediation of Remazol Red. termediate I and 2-[(3-diazenylphenyl) sulfonyl] ethanesulfonate

Table 1
Characterization of dye mixture and textile wastewater before and after treatment by Vetiveria zizanioides at laboratory scale (40 h), and using floating phyto-bed (72 h).

Treatment/Parameters Laboratory scale (40 h) Floating phyto-bed (72 h)

Dye mixture Textile wastewater Textile wastewater

Control Test Control Test Control Test

ADMI 821 ± 12 215 ± 3*** 1424 ± 8 380 ± 17*** 1142 ± 21 295 ± 12**
pH 9.8 7.3 10.6 7.7 9.5 7.8
COD (mg/L) 1189 ± 22 325 ± 15*** 1645 ± 22 491 ± 10*** 1495 ± 20 395 ± 18**
BOD (mg/L) 832 ± 9 207 ± 3*** 1309 ± 11 289 ± 12*** 1135 ± 20 218 ± 7**
TDS (mg/L) 21 ± 2.3 9 ± 1.2*** 4280 ± 5.7 1364 ± 6.4*** 4706 ± 28 1580 ± 10**
TSS (mg/L) 35 ± 1.1 15 ± 0.2* 973 ± 14.4 392 ± 12.2*** 734 ± 8 392 ± 7**
Electric conductivity 0.54 ± 0.02 2.98 ± 0.05*** 0.89 ± 0.05 3.10 ± 0.06*** 0.72 ± 0.07 4.18 ± 0.05*
Turbidity (NTUs) 42 ± 0.8 19 ± 2.2** 291 ± 3.5 118 ± 3.6*** 305 ± 5.0 125 ± 2.9***

Values are a mean of three experiments ± SEM.

Significantly different from control (0 h) at *P < 0.05, **P < 0.01 and ***P < 0.001. by one-way ANOVA with Tukey-Kramer comparison test.
 V. Chandanshive et al. / Chemosphere 252 (2020) 126513 5

Table 2
Enzyme activities of root and shoot of Vetiveria zizanioides at 0 h and 40 h for textile wastewater.

Enzyme Root Stem

Control Test Control Test

a
Lignin peroxidase 0.21 ± 0.02 0.48 ± 0.10* 0.07 ± 0.01 0.10 ± 0.02
Laccasea 29.6 ± 1.8 182.2 ± 7.0*** 14.7 ± 1.9 52.2 ± 4.0**
Veratryl alcohol oxidasea 198.2 ± 6.8 539.1 ± 15.6*** 78.5 ± 5.4 125.5 ± 2.2*
Azo reductaseb 5.87 ± 0.47 12.75 ± 1.43*** 1.16 ± 0.09 3.08 ± 0.07*
NADH-DCIP reductasec 24.15 ± 1.24 23.09 ± 0.95 0.28 ± 0.01 0.10 ± 0.01*
Riboflavin reductased 0.77 ± 0.02 0.03 ± 0.01*** 0.75 ± 0.03 0.07 ± 0.01***

Values are a mean of three experiments ±SEM.

Significantly different from control (0 h) at *P < 0.05, **P < 0.01 and ***P < 0.001.by one-way ANOVA with Tukey-Kramer comparison test.
a
Enzyme activity unit in min1 mg1.
b
mg of azo dye reduced min1 mg protein1.
c
mg of DCIP reduced min1 mg protein1.
d
mg of riboflavin reduced min1 mg protein1.

Fig. 1. HPLC analysis of a) untreated Remazol Red, b) untreated dye mixture, c) untreated textile wastewater, d) degraded product of Remazol Red, e) dye mixture after treatment, f)
treated textile wastewater.

(mw ¼ 277, m/z ¼ 277). Further, intermediate I undergoes desul- Remazol red and textile wastewater after treatment by
fonation to yield 8-[(4-chloro-1,3,5-triazin-2-yl)amino]naph- V. zizanioides were also increased when compared to plants
thalen-1-ol (mw ¼ 272, m/z ¼ 271) and intermediate II. germinated with normal water. While notable decrease in the
Intermediate II underwent denitrification to form 8-[(4-chloro- length of shoot and root of the seedlings was observed when
1,3,5-triazin-2-yl)amino]naphthalen-1-ol (mw ¼ 142, m/z ¼ 143) watered with untreated Remazol Red and textile wastewater
(Fig. 2). Oxidative and asymmetrical breakdown revealed to be an (Table 3). V. zizanioides treatment confirmed significant reduced
important mechanism for the cleavage of dye structure because of toxicity of produced metabolites.
laccase, lignin peroxidase and veratryl alcohol oxidase enzymes
action (Khandare and Govindwar, 2015). 3.6. Treatment of textile wastewater using floating phytobed

3.5. Toxicity analysis of dye products Treatment of textile wastewater by chemical precipitation and
microbial treatment fails to remove color and reduce TDS. The
Treated textile wastewater is being used to irrigate agricultural irrigation with high TDS wastewater enhanced soil salinity, and
field. Hence, it is imperative to study the harmful impact of treated increased TDS of surface and ground water. Hence, phytobeds were
wastewater on the plantlets of routinely grown crops. Remazol Red used to improve earlier treatment practices. Wild macrophytes
and textile wastewater demonstrated inhibition of A. sativum and V. zizanioides were grown with long, intense and strong root (45 cm,
P. mungo seed sprouting. The maximum germination was 50 and 180 g biomass) arrangement in normal tap water for two months
55%; 40 and 50%, respectively. However, V. zizanioides treated dye on floating phytobed. V. zizanioides could grow healthily in an
Remazol Red and textile wastewater showed 85 and 90%; 80 and extensive range of pH 4e10. Textile wastewater treatment with
80% seed germination of A. sativum and P. mungo, respectively. The fully grown macrophytes on floating phytobed in a reactor (400 L)
lengths of root and shoot watered with formed metabolites of for 72 h notably reduced the parameters like ADMI values, BOD,
6 V. Chandanshive et al. / Chemosphere 252 (2020) 126513

Fig. 2. Proposed pathways for metabolites of Remazol Red by V. zizanioides.

COD, TSS and TDS by 74, 81, 73, 47 and 66%, respectively (Table 1). textile wastewater and achieved the BOD, COD, turbidity, TSS, TDS
Static reactor planted with macrophyte Pogatherum orinithum also and TOC up to 38, 59, 41, 60, 71 and 37%, respectively (Khandare
revealed noteworthy performance and reduced ADMI, COD, BOD, et al., 2013). Ornamental plant, Glandularia pulchella in static phy-
pH, TSS and TDS (Watharkar et al., 2015). Pilot scale phytoreactor toreactor demonstrated significant reduction of BOD (70%), COD
along with P. grandiflora as well showed efficient treatment of real (70%) and TOC (74%) within 60 h (Kabra et al., 2013). It was also
 V. Chandanshive et al. / Chemosphere 252 (2020) 126513 7

Table 3
Phytotoxicity of Remazol Red and textile wastewater before and after (40 h) on Aestivum sativum and Phaseolus mungo.

Water Control dye Treated dye Untreated textile wastewater Treated textile wastewater

Aestivum sativum
Germination % 100 50 85 40 80
Root length 6.15 ± 0.37 2.75 ± 0.15* 5.45 ± 0.12$ 1.85 ± 0.16* 4.92 ± 0.51$$
Shoot length 10.50 ± 0.32 4.89 ± 0.06* 7.64 ± 0.29$$ 3.40 ± 0.19* 8.78 ± 0.31$$
Phaseolus mungo
Germination % 100 55 90 50 80
Root length 7.25 ± 0.24 2.10 ± 0.07* 5.96 ± 0.14$$ 2.92 ± 0.26* 6.00 ± 0.48$$
Shoot length 12.50 ± 0.43 3.58 ± 0.19* 10.45 ± 1.01$$ 4.21 ± 0.37* 11.20 ± 0.21$$

Values are mean of three experiment ± SEM. Shoot and root lengths of plants grown in dye Remazol Red and untreated textile wastewater are significantly different from that
of plant grown in distilled water by *P < 0.001.
Shoot and root lengths of fifty plants grown in the treated dye Remazol Red and textile wastewater are also significantly different from that of plants grown in untreated
Remazol Red and textile wastewater, respectively by $P < 0.05 and $$P < 0.01.

noted the decrease the pH 7.8 from initial pH 9.5 due to the treat- from 10.7, within 4 d treatment) was also observed when textile
ment of wastewater. It was also observed in the phyto-treatment effluent treated with P. scrobiculatum, T. angustifolia and
(with I. hederifolia) of real textile wastewater and simulated dye consortium-TP planted on ridges of HRTS (Chandanshive et al.,
mixture, that reduced pH to 7.5 and 8.1 from 7.9 to 10.2, respectively 2018). Wastewater exposed to V. zizanioides, I. aquatica and
(Rane et al., 2016). While, Alternanthera philoxeroides exposed to consortium-VI planted furrows were found to enhance bacterial
effluent decreased pH from acidic to neutral (Rane et al., 2015). counts to 1.89, 1.87 and 1.93-fold, respectively within 120 h. This
Electric conductivity of wastewater was found to enhance after might be due to decrease in dye toxicity (Table 4). Lower bacterial
treatment of V. zizanioides by 5.8-fold within 72 h (Table 1). Floating population in untreated/unplanted textile effluent might be due to
beds with P. australis and T. domingensis augmented with a bacterial toxicity of dyes present in the effluents. Increased bacterial popu-
consortium (Acinetobacter junii, Pseudomonas indoloxydans and lation by the plantation of V. zizanioides, I. aquatica and consortium-
Rhodococcus sp.) also resulted in the removal of colour, organic VI could be the result of release of exudates and carbon sources
matter, toxicity, and heavy metals from textile wastewater (Tara from dye metabolism, which ultimately supported the growth of
et al., 2019). bacteria. Eventually, increased bacterial population might have
contributed for the degradation of textile dyes as well as decreased
3.7. On field treatment of textile wastewater in a constructed toxicity of effluent (Kabra et al., 2013; Hussain et al., 2018).
furrow planted with V. zizanioides, I. aquatica and their consortium T. angustifolia, P. grandiflora, P. scrobiculatum, F. dichotoma,
Ammannia baccifera, G. grandiflora, A. amellus, and T. patula exposed
Individual furrow cultivated with V. Zizanioides, I. aquatica and to the textile effluent in a laboratory scale reactor also increased
consortium-VI in modified high rate transpiration system con- bacterial count after phytoremediation (Kadam et al., 2017;
structed at Kagal 5 star MIDC, Maharashtra, India when used to Chandanshive et al., 2018). Textile wastewater treated by
treat textile wastewater showed significant decrease in ADMI (67, V. zizanioides planted in furrow showed 18, 47, 64 and 56% reduc-
61 and 76%), BOD (73, 70 and 83%), COD (75, 74 and 79%), TDS (77, tion in cadmium (Cd), lead (Pb), arsenic (As) and chromium (Cr),
75 and 83%) and TSS (34, 31 and 51%), respectively after 5 d respectively. I. aquatica grown in furrow also decreased in Cd, Pb, As
(Table 4). Constructed lagoon (52,500 L) with floating plants like and Cr up to 18, 50, 46 and 52%, respectively, within 120 h.
S. molesta reduced BOD, ADMI and COD of textile wastewater Consortium-VI showed significant reduction of Cd (63%), Pb (74%),
considerably by 82, 81 and 76% within 8 d treatment, respectively As (66%) and Cr (70%) after treatment of 120 h. A healthy root zone
(Chandanshive et al., 2016). Developed vertical and horizontal microflora is always advantageous in phytoremediation trials.
constructed wetland planted with P. australis reduced color value, Metals from textile effluent could have been absorbed/bio-
BOD, COD, TSS up to 90, 66, 84 and 93%, respectively of textile transformed by soil microbial community and further make avail-
effluent (Bulc and Ojstrsek, 2008). Macrophytes V. zizanioides, able to the plants for easy uptake and accumulation, establishing
I. aquatica and consortium-VI were observed to reduce pH of textile synergistic removal of metals. Plants such as, T. angustifolia, P.
wastewater to 8.1, 7.9 and 7.4, respectively from pH 10.9 within scrobiculatum and consortium-TP also showed notably decreases of
120 h (Table 4). Similar decrease in pH (8.8, 8.5 and 7.5, respectively Pb (45, 50 and 69%), Cd (28, 28 and 71%), Cr (59, 63 and 77%) and As

Table 4
Characterization of untreated and treated textile wastewater (5 days) in furrows cultivated Vetiveria zizanioides, Ipomoea aquatica and their consortium.

Parameters Textile wastewater Vetiveria zizanioides Ipomoea aquatica Consortia VI

ADMI 1308 ± 11 426 ± 8*** 512 ± 5*** 319 ± 4***

pH 10.9 8.1 7.9 7.4
COD (mg/L) 1794 ± 7 452 ± 3*** 471 ± 5*** 382 ± 7***
BOD (mg/L) 1350 ± 13 368 ± 4*** 397 ± 8*** 221 ± 7***
TDS (mg/L) 5143 ± 21 1185 ± 9*** 1268 ± 8*** 859 ± 5**
TSS (mg/L) 1900 ± 15 1249 ± 12*** 1310 ± 10*** 928 ± 5**
Bacterial count (CFUs) 07 ± 0.5  107 62 ± 1.7  107** 55 ± 2.3  107** 97 ± 2.9  107***
Cadmium (mg/L) 0.11 ± 0.01 0.09 ± 0.01 0.09 ± 0.01** 0.04 ± 0.01***
Lead (mg/L) 0.80 ± 0.09 0.42 ± 0.07* 0.38 ± 0.09* 0.21 ± 0.01**
Arsenic (mg/L) 2.89 ± 0.03 1.05 ± 0.05*** 1.57 ± 0.10*** 0.98 ± 0.04***
Chromium (mg/L) 1.93 ± 0.28 0.85 ± 0.07** 0.92 ± 0.09** 0.58 ± 0.01***

Values are a mean of three experiments ± SEM.

Significantly different from control (untreated wastewater) at *P < 0.05, **P < 0.01 and. P < 0.001 by one-way ANOVA with Tukey-Kramer comparison test.
8 V. Chandanshive et al. / Chemosphere 252 (2020) 126513

(60, 54 and 72%) within 96 h of textile effluent treatment Appendix A. Supplementary data
(Chandanshive et al., 2017). Several aquatic and semi-aquatic plants
have potential to eliminate heavy metals from textile wastewater Supplementary data to this article can be found online at
(Miretzky et al., 2004; Alvarado et al., 2008; Rezania et al., 2015). https://doi.org/10.1016/j.chemosphere.2020.126513.
Three V. zizanioides ecotypes also showed removal of Fe, Mn, Zn, Cu,
Cd and Pb from wastewater of milk factory, an electric lamp plant, a References
battery manufacturing plant and ink manufacturing facility
(Roongtanakiat et al., 2007; Kafil et al., 2019). E. crassipes, a mac- Alvarado, S., Gue dez, M., Lue -Merú, M., Nelson, G., Alvaro, A., Jesús, A., Gyula, Z.,
2008. Arsenic removal from waters by bioremediation with the aquatic plants
rophytes having extensive root system showed effective removal of water hyacinth (Eichhornia crassipes) and lesser duckweed (Lemna minor).
heavy metal and color (85%) due to adsorption from wastewater Bioresour. Technol. 99, 8436e8440.
(Roy et al., 2018; Tabinda et al., 2019). It also has a potential to APHA, 1998. Standard Methods for the Examination of Water and Wastewater,
twentieth ed. American Public Health Association.
decrease COD, TSS and color from batik textile effluents (Safauldeen
Bulc, T., Ojstrsek, A., 2008. The use of constructed wetland for dye-rich textile
et al., 2019). Simulated shallow pond system planted with L. minor wastewater treatment. J. Hazard Mater. 155, 76e82.
was able to remove Basic Red 46 at low concentrations (Yaseen and Carias, C.C., Novais, J.M., Martins-Dias, S., 2007. Phragmites australis peroxidases role
Scholz, 2016; Ekperusi et al., 2019). A pilot-scale vertical flow in the degradation of an azo dye. Water Sci. Technol. 56, 263e269.
Chandanshive, V., Rane, N., Tamboli, A., Gholave, A., Khandare, R., Govindwar, S.,
constructed wetland with Brachiaria mutica demonstrated suc- 2017. Co-plantation of aquatic macrophytes Typha angustifolia and Paspalum
cessful plant-dye degrading endophytic bacteria partnership scrobiculatum for effective treatment of textile industry effluent. J. Hazard
showing efficient degradation of textile effluent and reduction in Mater. 338, 47e56.
Chandanshive, V.V., Kadam, S.K., Khandare, R.V., Kurade, M.B., Jeon, B.-H.,
toxicity. Bacterial augmentation enhanced the remediation effi- Jadhav, J.P., Govindwar, S.P., 2018. In situ phytoremediation of dyes from textile
ciency which decreased COD (81%), BOD (81%), TDS (32%), and color wastewater using garden ornamental plants, effect on soil quality and plant
(74%) along with reduction of nitrogen (84%), phosphorous (79%), growth. Chemosphere 210, 968e976.
Chandanshive, V.V., Rane, N.R., Gholave, A.R., Patil, S.M., Jeon, B.-H., Govindwar, S.P.,
and heavy metals viz. Cr, Fe, Ni, and Cd by 72e97% (Hussain et al., 2016. Efficient decolorization and detoxification of textile industry effluent by
2018). Salvinia molesta in lagoon treatment. Environ. Res. 150, 88e96.
Danh, L.T., Truong, P., Mammucari, R., Tran, T., Foster, N., 2009. Vetiver grass, Veti-
veria zizanioides: a choice plant for phytoremediation of heavy metals and
4. Conclusion organic wastes. Int. J. Phytoremediation 11, 664e691.
Dipu, S., Kumar, A.A., Thanga, V.S.G., 2011. Phytoremediation of dairy effluent by
constructed wetland technology. Environmentalist 31, 263e278.
Significant decolorization and removal of Remazol Red, removal Ekperusi, A.O., Sikoki, F.D., Nwachukwu, E.O., 2019. Application of common duck-
of dyes from textile wastewater and dye mixture, induction of weed (Lemna minor) in phytoremediation of chemicals in the environment:
oxido-reductive enzymes, significant reduction in environmental state and future perspective. Chemosphere 223, 285e309.
Hatvani, N., Me cs, I., 2001. Production of laccase and manganese peroxidase by
parameters such as COD, BOD, TDS, TSS, and ability to survive in Lentinus edodes on malt-containing by-product of the brewing process. Process
unfavourable conditions suggests the potential of V. zizanioides for Biochem. 37, 491e496.
phytotransformation. Floating phytobeds of V. zizanioides could be Hussain, Z., Arslan, M., Malik, M.H., Mohsin, M., Iqbal, S., Afzal, M., 2018. Treatment
of the textile industry effluent in a pilot-scale vertical flow constructed wetland
used in a deep textile wastewater of the reservoirs. The combina-
system augmented with bacterial endophytes. Sci. Total Environ. 645, 966e973.
torial plantation of V. zizanioides and I. aquatica in furrows of Jadhav, U.U., Dawkar, V.V., Tamboli, D.P., Govindwar, S.P., 2009. Purification and
existing HRTS could effectively treat textile wastewater than with characterization of veratryl alcohol oxidase from Comamonas sp. UVS and its
individual plants. Plantation of aquatic plants consortium in fur- role in decolorization of textile dyes. Biotechnol. Bioproc. Eng. 14, 369e376.
Kabra, A.N., Khandare, R.V., Govindwar, S.P., 2013. Development of a bioreactor for
rows of HRTS will be intelligent approach to clean textile waste- remediation of textile effluent and dye mixture: a plant-bacterial synergistic
water and management strategies. strategy. Water Res. 47, 1035e1048.
Kadam, S.K., Chandanshive, V.V., Rane, N.R., Patil, S.M., Gholave, A.R., Khandare, R.V.,
Bhosale, A.R., Jeaon, B.-H., Govindwar, S.P., 2017. Phytobeds with Fimbristylis
Declarations of competing interest dichotoma and Ammannia baccifera for treatment of real textile effluent: an in-
situ treatment, anatomical studies and toxicity evaluation. Environ. Res. 160,
None. 1e11.
Kafil, M., Boroomand, N.S., Moazed, H., Bhatnagar, A., 2019. Phytoremediation po-
tential of vetiver grass irrigated with wastewater for treatment of metal
CRediT authorship contribution statement contaminated soil. Int. J. Phytoremediation 21, 92e100.
Kagalkar, A.N., Khandare, R.V., Govindwar, S.P., 2015. Textile dye degradation po-
tential of plant laccase significantly enhances upon augmentation with redox
Vishal Chandanshive: Data curation, Formal analysis, Investi- mediators. RSC Adv. 5, 80505e80517.
gation, Methodology, Resources, Software, Validation, Visualiza- Khandare, R.V., Govindwar, S.P., 2015. Phytoremediation of textile dyes and efflu-
ents: current scenario and future prospects. Biotechnol. Adv. 33, 1697e1714.
tion, Writing - original draft, Writing - review & editing. Suhas Khandare, R.V., Kabra, A.N., Awate, A.V., Govindwar, S.P., 2013. Synergistic degra-
Kadam: Data curation, Formal analysis, Investigation, Methodol- dation of diazo dye Direct Red 5B by Portulaca grandiflora and Pseudomonas
ogy, Resources, Software, Validation, Visualization. Niraj Rane: putida. Int. J. Environ. Sci. Technol. 10, 1039e1050.
Lowry, O.H., Rosebrough, N.J., Farr, L., Randall, R., 1951. Protein measurement with
Data curation, Formal analysis, Investigation, Methodology, Re-
the folin phenol reagent. J. Biol. Chem. 193, 265e275.
sources, Software, Validation, Visualization. Byong-Hun Jeon: Mahmood, Q., Zheng, P., Islam, E., Hayat, Y., Hassan, M.J., Jilani, G., Jin, R.C., 2005. Lab
Conceptualization, Funding acquisition. scale studies on water hyacinth (Eichhornia crassipes marts solms) for bio-
treatment of textile wastewater. Caspian J. Environ. Sci. 3, 83e88.
Miretzky, P., Saralegui, A., Cirelli, A., 2004. Aquatic macrophytes potential for the
Acknowledgements simultaneous removal of heavy metals (Buenos Aires, Argentina). Chemosphere
5, 997e1005.
Ong, S., Uchiyama, K., Inadama, D., Ishida, Y., Yamagiwa, K., 2010. Treatment of azo
Vishal Chandanshive and Sanjay Govindwar thanks DBT (Grant dye Acid Orange 7 containing wastewater using up-flow constructed wetland
No. BT/PR18965/BCE/8/1401/2016) for research support. Sanjay P with and without supplementary aeration. Bioresour. Technol. 101, 9049e9057.
Patil, A.V., Jadhav, J.P., 2013. Evaluation of phytoremediation potential of Tagetes
Govindwar also thanks National Research Foundation, South Korea
patula L. for the degradation of textile dye Reactive blue 160 and assessment of
for financial support (Grant no. NRF 201900000002057). Prof. Jeon the toxicity of degraded metabolites by cytogenotoxicity. Chemosphere 92,
thanks to Ministry of Education, Science, and Technology (MEST) 225e232.
for the grant (NRF-2017R1A2B2004143). Authors are thankful to Priyanka, J.V., Krishnaswamy, V.G., 2019. Phytoremediation of mixed reactive Azo
dyes in contaminated water by Ceratophylum demersum and its toxicity analysis
Anil N. Kamble and Dipak S. Patil for monitoring and maintenance on other life forms. SPC J. Environ. Sci. 1, 16e25.
of plants at HRTS. Rane, N.R., Chandanshive, V.V., Watharkar, A.D., Khandare, R.V., Patil, T.S.,
 Fuel 329 (2022) 125524

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Kinetic modelling of anaerobic co-digestion of sewage sludge and

Sherry-wine distillery wastewater: Effect of substrate composition in
batch bioreactor
Vanessa Ripoll a, b, Rosario Solera a, *, Montserrat Perez a
a
Department of Environmental Technologies, University of Cadiz, Campus de Puerto Real, 11500 Puerto Real, Cadiz, Spain
b
Facultad de Ciencias Experimentales, Universidad Francisco de Vitoria (UFV), Ctra. Pozuelo‑Majadahonda km 1.800, 28223 Pozuelo de Alarcón, Madrid, Spain

A R T I C L E I N F O A B S T R A C T

Keywords: Batch thermophilic anaerobic co-digestion of sewage sludge (SS) and Sherry-wine distillery wastewater (SW-DW)
Anaerobic co-digestion was investigated through biochemical methane potential tests (BMP). The results pointed out that biodegrad­
Sherry-wine distillery wastewater ability and biomethane potential were enhanced proportionally to the percentage of SW-DW of the feedstock,
Sewage sludge
whose soluble biodegradability fraction is 10-fold higher than SS. Specifically, organic matter removal increases
Kinetic modelling
from 37 % (employing sole SS as feedstock) to 60 % (employing sole SW-DW as feedstock). SW-DW almost
Batch bioreactor
doubles methane yield in comparison to SS (302 ± 15 and 175 ± 9 NL/kg, respectively).
A structured kinetic model was developed considering hydrolysis, acidogenic and methanogenic steps of
anaerobic digestion. A non-linear multiple-response regression was employed to estimate the kinetic parameters
for each feedstock (0%(v/v) SW-DW, 25%(v/v) SW-DW, 50%(v/v) SW-DW, 75%(v/v) SW-DW, and 100%(v/v)
SW-DW). The first-order kinetic parameter estimated of the hydrolysis step varies inversely proportional to the
percentage of SW-DW content in the feedstock. Whereas, there is no significant influence of feedstock compo­
sition on kinetic parameters value regarding acidogenesis and methanogenesis. These results showed that
rate–limiting step switched during the fermentation and the addition of SW–DW favours acidogenic and meth­
anogenic steps. In summary, the proposed kinetic model was able to predict batch experimental data, supporting
the application of biogas production from anaerobic co-digestion of SS and DW-DW.

1. Introduction The application of biological aerobic and anaerobic treatments is a

promising technology to reduce the organic matter content in SW-DW
Wine production is part of the economic engine in the Mediterranean [2,4–6]. Comparing to the aerobic processes, the anaerobic digestion
societies. Specifically, Sherry-wine is the most important wine produced (AD) treatment presents great advantages, such as low operating costs, a
in Cadiz region (South of Spain) having an under cultivation area higher well-known understanding of the available technology, and biogas
than 7100 ha and more than 36 million of litres produced in 2020 [1]. production during the treatment, which could eventually cover part of
However, associated to the industrial production process, Sherry-wine the energy requirements in the plant. Recently, the use of SW-DW as co-
distillery wastewater (SW-DW) is generated as a by-product composed substrate in AD of municipal sewage sludge (SS) has been already started
by a mixture of liquid waste from pressing and first fermentation phases studying [3,7–9]. And yet, the addition of co-substrates derived from
of wine processing, the bottom of the distillery unit, grape juice spillage meat industry, such as chicken manure and poultry manure, are
and chemical cleaning products from equipment and tanks. SW-DW is emerging researches [10,11]. Anaerobic co-digestion (ACoD) presents
considered one of the most important environmental concerns in viti­ many advantages, including shared treatment facilities, reducing initial
culture agro-industries due to its strongly acid pH, the high organic load investment and operating costs, buffering of the variations in the
and its cytotoxic components, such as polyphenolic derivatives [2]. In composition of the waste over time and dilution of toxic compounds
addition, the seasonality of the wine production also implies an issue for [12].
the waste management at full-scale [3]. It is well-known that, AD takes places by means of a complex reaction

* Corresponding author.
E-mail addresses: [email protected] (V. Ripoll), [email protected] (R. Solera), [email protected] (M. Perez).

https://doi.org/10.1016/j.fuel.2022.125524
Received 28 June 2022; Received in revised form 27 July 2022; Accepted 3 August 2022
Available online 9 August 2022
0016-2361/© 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-
nc/4.0/).
V. Ripoll et al. Fuel 329 (2022) 125524

network involving several steps and a large number of microbial species. Table 1
Overall, AD is lumped in four serial steps: hydrolysis, acidogenesis, Inoculum and raw substrates characteristics.
acetogenesis, and methanogenesis. However, there are many other steps Parameter Inoculum SS SW-DW
that occur in series and parallel reactions, such as butyrate-degrading
pH 7.8 ± 0.2 6.9 ± 0.2 6.4 ± 0.2
acetogenesis, acetoclastic, hydrogenotrophic methanogenesis, propio­ CODt (kg/m3) 19.9 ± 0.4 53.9 ± 1.2 24.6 ± 2.2
nate degradation, and inhibition by pH and high concentration of acid CODs (kg/m3) 9.7 ± 0.3 19.0 ± 0.3 20.7 ± 0.6
components [13]. The overall rate of AD will be controlled by the TS (kg/m3) 2.09 ± 0.03 3.67 ± 0.01 1.47 ± 0.11
slowest stage of the process, known as the rate-limiting step. Although, vS (kg/m3) 1.21 ± 0.01 2.69 ± 0.03 1.06 ± 0.09
Alkalinity (kg/m3) 5.81 3.53 0.02
some authors describe that the hydrolysis of wastes with a high content VFA (kg/ m3) 0.41 2.85 0.75
of solid organic matter controls the AD due to the mass transfer limi­ TN (kg/ m3) 2.15 14.8 1.09
tations, it is also accepted that acetogenesis or methanogenesis might be C/N 9.2 5.2 17.5
the rate-controlling step in AD of complex wastes [14,15].
Kinetic model of anaerobic treatment systems plays a key role to
inoculum was stored at 55 ◦ C for 7 days before employing in batch tests
develop and controlling more stable processes [16]. It is also a powerful
in order to deplete the residual biodegradable organic material present
tool in the scale-up of bioprocesses. To date, a large number of published
in it. This step is highly recommended to minimize endogenous methane
papers aim to describe biogas production of AD, considering only the
production of the inoculum and assuring the reproducibility of the
last step of the process. Through simple kinetic equations, such as first-
assays.
order, logistic, Cone, and Gompertz model, cumulative methane pro­
duction is successfully fitted and predicted for batch experiments
[17–21]. However, this mathematical approach limits to describe biogas 2.2. Biomethane potential tests
evolution over time without considering the connection between the
generation of the final product and the organic matter consumption. In BMP tests were carried out in order to evaluate biodegradability and
order to provide a more comprehensive AD model, a further kinetic biogas production that can be potentially reached by ACoD treatment of
description focusing on the involving metabolic pathways is required. In each substrate.
this sense, only few researchers have developed complex kinetic Batch reactors employed in BMP tests were 250 mL serum bottles
modelling of AD considering the relationship between enzymatic hy­ with a working volume of 130 mL. The initial broth composition was 60
drolysis, acidogenesis, acetogenesis and methane production. Most of % (v/v) of substrate, and 40 % (v/v) of inoculum. All bottles were
them included first-order kinetic and Monod-type equations to describe subsequently purged with 100 % N2 for 3–4 min in order to achieve
the degradation of different wastes, such as livestock manure [22,23], anaerobic atmosphere for the run. BMP tests took place in an orbital
olive pomace [13], orange rind [24], organic fraction of municipal solid shaker at 85 rpm under thermophilic conditions (55 ± 1 ◦ C). Control
waste [25], microalgae [14], cow manure [26] and SS [27]. run, containing only anaerobic inoculum and distillate water, was per­
The present work aimed to develop a kinetic model to describe the formed to determine background gas production. For each mixture,
anaerobic co-digestion of SS and SW-DW at thermophilic temperature eight runs were conducted initially. Every 3–4 days, a duplicate was
(55 ◦ C) in batch bioreactor. In order to reach this goal, the influence of used to analyse medium composition in terms of total and soluble
feedstock composition on biodegradability and biomethane production chemical oxygen demand (CODt, and CODs), total and volatile solids
was studied in biochemical methane potential (BMP) tests. Based on the (TS, and VS), and volatile fatty acids (VFA) content.
experimental tendencies over the time, a kinetic model including hy­ BMP tests were performed until daily methane production meant less
drolysis, acidogenic, and methanogenic steps was proposed. Finally, a than 1 % of total (25 days). Biomethane production data relating to
comparison of the estimated kinetic parameters was conducted to study produced volume and biogas composition showed in this paper corre­
the influence of feedstock composition on the rate-limiting step. This is spond to the average daily values of each bottle.
the first time that kinetic modelling of ACoD of SS and SW-DW was
developed. 2.3. Analytical methods

2. Material and methods pH, TS, vS CODt, and CODs were determined according to Standard
Methods [30]. pH determination was taken by pHmeter type CRISON
2.1. Inoculum and substrates MICROPH 2001 with a temperature probe. For TS, vS and FTS, samples
were weighed in ceramic boats in a laboratory balance Cobos type and
The wastes used as substrates in the present study came directly from drying in oven type ELF14 de CARBOLITE.
industrial features located in Jerez de la Frontera (Cadiz, Spain). SS were VFA (acetic, propionic, isobutyric, butyric, isovaleric, valeric, iso­
collected from a secondary treatment floatation unit from Guadalete caproic, caproic and heptanoic acid) were determined by gas chroma­
wastewater treatment plant. SW-DW was obtained from Gonzalez-Byass, tography (GC-2010 Plus Shimadzu). Total acidity was calculated by
an ethanol producing wine-distillery plant. Both substrates were kept summing up the individual fatty acids in terms of equivalent mass of
under refrigeration at 4 ◦ C until use to avoid biodegradation. acetic acid.
Different mixtures of SS and SW-DW, as well as sole SS and sole SW- Gas composition was determined employing a gas chromatography
DW, were used as substrates to evaluate the biomethane potential of technique (GC-2010 Shimadzu). The analysed gases (H2, CH4, CO2, O2
each waste. Five mixtures were tested as substrate in BMP test, being its and N2) were measured by means of a thermal conductivity detector
proportion of SW-DW 0, 25, 50, 75, and 100 % (v/v). The pH values of (TCD) at 250 ◦ C using a Supelco Carboxen 1010 Plot column. The oven
co-digestion mixtures were in a range of 6.0–7.0. Before it used, pH was temperature was programmed between 35 and 200 ◦ C. Manual injection
adjusted to 7.8–8.2 using sodium hydroxide solution (2 M). The char­ was carried out employing a sample volume of 250 mL. The carrier gas
acteristics of each raw substrate and inoculum are summarised in was helium at 35 kPa of pressure.
Table 1.
The effluent of 5 L laboratory-scale mesophilic anaerobic digester 2.4. Mathematical methods
was used as inoculum in the BMP test. The reactor was operated under
semi-continuous mode setting with a hydraulic retention time of 20 Aspen Custom Modeler v10 software (AspenTech, USA) was
d and feeding with activated SS. According to the established consid­ employed to fit the kinetic model to experimental data in order to es­
erations in literature to standardize protocol for BMP assays [28,29], timate kinetic parameters. The set of differential equations forming the

2
V. Ripoll et al. Fuel 329 (2022) 125524

kinetic model were integrated using the numerical implicit Euler

method. A multiple-response method to fit the model to the data was set.
The kinetic parameters were estimated by minimizing the difference
between experimental observations and model simulations according to
“least squares method” by an adaptive non-linear least-squares algo­
rithm (NL2SOL).
For the porpoise of the validation regarding the proposed model,
physicochemical and statistical criteria were performed, including:

• F-Fischer’s value (F), which should be higher than its tabulated value
at 95 % confidence (F95) to fulfil the null hypothesis.
• Sum of squared residuals (SSR), which should be as close to zero as
possible.
• Residual mean square error (RMSE), which should be as close to zero
as possible.
• Percentage of explained variance (VE), which should be equal or
near 100 % to indicate an identical change of experimental and
simulated data with time.

3. Results and discussion

3.1. Substrate biodegradability and methane production in BMP tests

Substrate biodegradability was evaluated based on initial and final

characteristics in BPM test. Physicochemical characteristics measured at
the beginning and at the end in serum bottles for each experiment are
shown in Table 2.
In general, organic matter is reduced in terms of CODt, TS and vS
though the batch AD process for all the employed substrates. The
removal percentage of each parameter has been calculated in order to
discuss the influence of the substrate composition. As it is shown in the
Fig. 1, the presence of SW-DW in the substrate has a positive effect on the
biodegradability of the feedstock: the higher SW-DW content, the higher
CODt (Fig. 1A), TS (Fig. 1C) and vS removal (Fig. 1D). While in contrast,
only around 37 % of CODt and vS are reduced when SS is employed as
carbon source, 50 % of the organic matter is degraded in the mixture 50
% SS – 50 % SW-DW and values up to 60 % are reached employing raw
SW-DW. These results showed that SW-DW contains more soluble and
easily metabolized organic matter, which promotes ACoD in comparison
to sole SS AD. Similar improvement of SS biodegradability by adding
SW–DW as co-substrate in AD has been previously reported in meso­
philic temperature range [7].
Moreover, operational temperature has a perceptible effect on CODs
removal. Whereas mesophilic temperature range an effective CODs

Table 2
Initial and final characteristics of substrates in serum bottles.
Parameters SW-DW (%)
(kg/m3)
0 25 50 75 100

CODt_initial 50.1 ± 46.8 ± 38.4 ± 36.0 ± 31.5 ±

0.3 0.4 0.9 1.4 1.6
CODt_final 31.0 ± 25.0 ± 19.5 ± 17.8 ± 12.4 ±
0.5 1.1 1.5 1.1 0.1
CODs_initial 11.0 ± 11.9 ± 15.4 ± 17.8 ± 21.6 ±
1.7 1.1 0.6 0.8 0.5
CODs_final 14.8 ± 14.0 ± 10.0 ± 10.5 ± 8.7 ± 0.9
1.0 0.6 0.2 0.8
TS_initial 36.9 ± 35.0 ± 31.7 ± 28.3 ± 24.7 ±
0.1 0.1 0.1 0.2 0.7
TS_final 28.5 ± 25.4 ± 22.2 ± 19.3 ± 15.1 ±
0.2 0.3 0.2 0.2 0.3
Fig. 1. Influence of waste composition on CODt removal, CODs removal, TS
VS_initial 26.1 ± 24.5 ± 21.7 ± 19.3 ± 15.7 ±
removal, and vS removal in BMP test.
0.3 0.2 0.4 0.2 0.1
VS_final 16.9 ± 14.2 ± 11.3 ± 8.7 ± 0.2 4.7 ± 0.6
0.4 0.4 0.4
VFA_initial 1.1 ± 0.1 1.5 ± 0.1 0.8 ± 0.1 1.4 ± 0.1 1.1 ± 0.1
VFA_final 2.7 ± 0.1 2.5 ± 0.1 2.8 ± 0.1 2.5 ± 0.1 2.0 ± 0.1

3
V. Ripoll et al. Fuel 329 (2022) 125524

removal was showed independently the proportion of SW-DW added - Since hydrogen production was negligible during the experiments,
[7], in thermophilic temperature range lower SW-DW proportion than hydrogenotrophic methanogenesis pathway was not considered in
30 % in the feedstock leads to an accumulation of soluble organic matter the kinetic description of the process.
in the final effluent, as it is showed in Fig. 1B on the negative values of - All concentrations, except methane production, were expressed as
CODs removal. For higher values of SW-DW than 30 %, an effective COD. Therefore, it is necessary to define the yield methane from COD
consumption of soluble organic matter is achieved, reaching CODs (YCH4/COD), a pseudo-stoichiometric coefficient in macroscopic
degradation between 40 and 60 %. However, in all those cases VFA were terms.
accumulated in the effluent respect to the initial VFA content (Table 2). - Employed substrates, SS and SW-DW, were partially biodegradable,
As it will be seen in the biomethane production results, VFA production as it had been previously suggested for other organic wastes [24].
did not inhibit biogas generation. Therefore, the fraction of non-biodegradable particulate substrate
Regarding the methane production, Fig. 2 shows the influence of (CODp_nb) and the fraction of non-biodegradable soluble substrate
feedstock composition on methane yield, which was calculated as the (CODs_nb) have been considered in the kinetic model.
quotient of the cumulative volume of produced methane at the end of - A first-kinetic order kinetic equation was proposed for each step.
the experiment over the initial organic matter content in terms of vol­ - Deviations in conservation of mass derived from water evaporation
atile solids. It is noteworthy that the methane yield increases when the and biomass adhesion to the bioreactoŕs surface was assume as
amount of SW-DW increases in co-digestion even though, the initial vS negligible, as other authors also previously assumed [13,24]
content is lower than vS presenting in SS. In spite of the above-
mentioned differences relating to solubilisation of organic matter, the Taking this information into consideration, the following kinetic
obtained results of methane production in thermophilic temperature model was developed. The simplified reaction network is composed by 3
range are quite similar to the previously published in batch runs at steps: hydrolysis, acetogenesis and methanogenesis, as is showed in Eq.
mesophilic temperature range [7]. In short, these results support the (1)–(3). For each step, a first-kinetic order reaction rate was defined, as
organic matter containing in the waste derived from distillery industry is is showed in Eq. (4)–(6). Based on this information, a differential
highly biodegradable and its addition as co-substrate in SS digestion equation regarding the production rate of each component involved was
improves methane production. established, as is showed in Eq. (7)–(10).
Hydrolysis step.

3.2. Kinetic modelling of anaerobic co-digestion r1

CODp → CODs (1)

The kinetic modelling of AcoD has been based on the evolution of Acetogenic step.
particulate chemical oxygen demand (CODp) (defined as the difference
(2)
r2
CODs → AceticAcid
between CODt and CODs), CODs, acetic acid (AA), and methane (CH4).
Based on the experimental results, a structured kinetic model was pro­ Methanogenic step.
posed to describe the macroscopic observed tendencies related to CODp,
(3)
r3
CODs, AA, and methane production. In order to develop a kinetic model AceticAcid → YCH4/AA Â⋅CH 4
to describe AD experiments, the following considerations were taken ( )
into consideration: r1 kgCOD /m3 Â⋅d = k1 Â⋅[[CODp] − CODp nb ] (4)
( )
- Batch bottles were considered as completely mixed bioreactors. r2 kgCOD /m3 Â⋅d = k2 Â⋅[[CODs] − CODs nb ] (5)
- At first, particulate substrate (CODp) must have been hydrolysed to ( )
obtain an accessible soluble substrate (CODs). r3 kgCOD /m3 Â⋅d = k3 Â⋅[AA] (6)
- The high-molecular-weight compounds containing in the accessible
soluble substrate were transformed into acetic acid, the final and d[CODp] ( )
kgCOD /m3 Â⋅d = − k1 Â⋅[[CODp] − CODp nb ] (7)
simplest volatile fatty acid in acetogenesis step. dt
- Finally, AA is transformed into methane by means of aceticlastic
methanogenesis pathway. d[CODs] ( )
kgCOD /m3 Â⋅d =k1 Â⋅[[CODp] − CODp nb ] − k2 Â⋅[[CODs]
dt (8)
− CODs nb ]

d[AA] ( )
kgCOD /m3 Â⋅d = k2 Â⋅[[CODs] − CODs nb ] − k3 Â⋅[AA] (9)
dt

d[CH 4 ] ( )
NLCH4 /m3 Â⋅d = YCH4/AA Â⋅k3 Â⋅[AA] (10)
dt
The fitting of the proposed kinetic equation to experimental data
(CODp, CODs, AA and methane) for each BMP run was carried out to
estimate the value of the kinetic parameter involving the kinetic model.
Fig. 3 shows the fit of the values predicted by the model to the experi­
mental data over the time for each mixture SS:SW-DW evaluated as
substrate.
In general, particulate substrate content decreases over the time as a
consequence of the hydrolysis of solid molecules such as carbohydrates,
proteins and lipids is taking place. The higher particulate organic matter
content linked to the higher SS proportion leads to a faster consumption
at the beginning of the fermentation. The evolution of soluble organic
matter is the net result of two opposing steps: hydrolysis and aceto­
Fig. 2. Influence of feedstock composition on methane yield reached at the end genesis, which causes the production and the consumption of CODs,
of BMP tests.

4
V. Ripoll et al. Fuel 329 (2022) 125524

respectively. Consequently, a peak in CODs content is observed. How­

ever, SW-DW content in the feedstock increases the initial accessible
soluble substrate that leads to slighter maximum tendencies. Even more,
evolution of soluble organic matter is always decreasing over the time
when sole SW-DW is used as sole carbon source (Fig. 3B), which means
that acetogenesis rate takes place faster than hydrolysis during the
whole experiment. The evolution of acetic acid also follows the typically
tendency of an intermediate compound, increasing at the beginning of
the fermentation due to the acidogenic step and decreasing afterwards
due to its consumption in methanogenic step. However, the maximum
value is achieved at shorter periods of time with the increasing of SW-
DW in the feedstock. AA concentration reached the maximum at 10 h
when sole SS was employed, whereas the peak is observed around 7 h
ahead (at 3 h) when sole SW-DW was used (Fig. 3C). Regarding methane
production, its concentration increases over the time as final product in
the reaction network. Accordingly to AA exhaustion, methane produc­
tion reaches a plateau at the end of the batch fermentation. It should be
pointed out that the faster methane production is observed in co-
digestion runs (25 %, 50 % and 75 % SW-WD) due to the combination
of high total and soluble organic matter presenting in theses feedstock.
Table 3 shows the estimated kinetic parameters, as well as the sta­
tistical parameters that provide the information on the goodness of fit.
The proposed model fits reasonably to all relevant data. Goodness-of-fit
statistical parameters indicate a high value for Fisheŕs F, very much over
the limiting value (the value of the F tabulated at 95 % was 8.6), and a
low value for the RMSE and the SSR. Moreover, the experimental ten­
dencies and the predicted from the model are quite similar, as the VE
percentages higher than 90 % showed.
The influence of feedstock composition on the value of the kinetic
parameters is shown in Fig. 4. The fractions of non-biodegradable par­
ticulate and soluble substrates follow a linear tendency with the
composition of the employed feedstock. Assuming that non-
biodegradable content was presented initially in the wastes, SS has
41.4 % of non-biodegradable particulate organic matter and 93.6 % of
non-biodegradable soluble organic matter. In the case of SW-DW, the
non-biodegradable fraction is around a third: only 38.4 % of the par­
ticulate matter and 31.0 % of the soluble matter are non-metabolizable.
Regarding the first-order kinetic parameter of hydrolysis step, k1 also
depends directly on the composition of the employed substrate, being
proportional to the ratio of the mixture. More specifically, the faster
hydrolysis takes places when SS is used as substrate probably due to its
higher content in solid particles. On the other hand, first-order kinetic
parameter of acidogenic and methanogenic steps, k2 and k3, are around
the same value independently the feedstock composition. A similar
tendency is observed for the pseudo-stoichiometric coefficient YCH4/COD.
Based on these results, feedstock composition affects basically hydro­
lysis kinetics in anaerobic co-digestion of SS and SW-WD, which address
the resulting differences in waste degradation and biogas production.
In order to determine the rate-limiting step, the evolution of the
reaction rate of each step in BMP tests was calculated for each mixture
SS:SW-DW employing estimated kinetic parameters values and pre­
dicted concentration profiles, as it is shown in Fig. 5. In general, reaction
rate of hydrolysis step (r1) decreases over the time due to the ongoing
consumption of particulate organic matter. In all the cases, r1 eventually
tends to zero, which is relating to the total exhaustion of the biode­
gradable particulate content of the wastes. The higher initial particulate
content leads to a faster initial rate and a longer required time to be
entirely consumed. Except for 100 % of SW-WD, the reaction rate of
acidogenesis step (r2) reaches a peak around the crossing point of hy­
drolysis with acidogenesis rates, which regards to the maximum amount
Fig. 3. Experimental evolution (points) and kinetic model prediction (dashed
of soluble COD in the system. In AD of sole SW-WD (Fig. 5E), r2 is
lines) of (A) particulate, (B) soluble organic matter, (C) acetic acid concentra­ maximal at the beginning of the experiment just as the soluble organic
tion, and (D) methane production in the BMP tests. Key: (SW-DW percentage in matter content (Fig. 3B). The reaction rate of methanogenic step (r3)
feedstock composition) black: 0%; blue: 25%; red: 50%; orange: 75%; grey: follows a similar tendency than acidogenesis rate, showing similar
100%. (For interpretation of the references to colour in this figure legend, the values in the second part of the experiments (since around 5 h of
reader is referred to the web version of this article.) fermentation). Methanogenic step is completely limited by acetic acid

5
V. Ripoll et al. Fuel 329 (2022) 125524

Table 3
Kinetic and statistical parameter values calculated by fitting the kinetic model to
experimental data of BMP experiments for each substrate.
0 % SW-DW

Kinetic parameters Statistical parameters

CODp_nb (kgCOD/m3) 16.2 ± 0.4 F / F95 2390

CODs_nb (kgCOD/m3) 10.3 ± 1.5 RMSE 0.55
k1 (d-1) 0.23 ± 0.01 SSR 11.3
k2 (d-1) 0.09 ± 0.01 VE (%) 98.0
k3 (d-1) 0.29 ± 0.03 – –
YCH4/COD (NLCH4/kgCOD) 156 ± 19 – –

25 % SW-DW

Kinetic parameters Statistical parameters

CODp_nb (kgCOD/m3) 11.6 ± 0.6 F / F95 1020

CODs_nb (kgCOD/m3) 9.1 ± 0.9 RMSE 0.79
k1 (d-1) 0.21 ± 0.01 SSR 23.9
k2 (d-1) 0.08 ± 0.01 VE (%) 96.6
k3 (d-1) 0.57 ± 0.06 – –
YCH4/COD (NLCH4/kgCOD) 186 ± 17 – –

50 % SW-DW

Kinetic parameters Statistical parameters

CODp_nb (kgCOD/m3) 9.9 ± 0.8 F / F95 1686

CODs_nb (kgCOD/m3) 8.6 ± 1.2 RMSE 0.75
k1 (d-1) 0.18 ± 0.02 SSR 21.5
k2 (d-1) 0.11 ± 0.02 VE (%) 93.3
k3 (d-1) 0.44 ± 0.05 – –
YCH4/COD (NLCH4/kgCOD) 202 ± 24 – –

75 % SW-DW

Kinetic parameters Statistical parameters

CODp_nb (kgCOD/m3) 8.1 ± 0.4 F / F95 1120

CODs_nb (kgCOD/m3) 7.8 ± 0.9 RMSE 0.54
k1 (d-1) 0.17 ± 0.02 SSR 11.4
k2 (d-1) 0.11 ± 0.02 VE (%) 95.2
k3 (d-1) 0.54 ± 0.05 – –
YCH4/COD (NLCH4/kgCOD) 203 ± 14 – –

100 % SW-DW

Kinetic parameters Statistical parameters

3
CODp_nb (kgCOD/m ) 3.8 ± 0.3 F / F95 2840
CODs_nb (kgCOD/m3) 6.7 ± 0.6 RMSE 0.29
k1 (d-1) 0.15 ± 0.02 SSR 3.5
k2 (d-1) 0.12 ± 0.01 VE (%) 98.5
k3 (d-1) 0.51 ± 0.04 – –
YCH4/COD (NLCH4/kgCOD) 183 ± 9 – –

formation, which can also be seen in the comparable trend of AA with

CODs showed in Fig. 3C. The lower values of AA concentration in
comparison to CODs ones favours that r3 is limited by the previous step
in the reaction network, r2.
The slowest rate performs as the rate-limiting step and it controls the
overall rate. In BMP tests the rate-limiting step changes along the
fermentation, being methanogenic reaction at shorter periods of time
and hydrolysis reaction at longer periods of times. For 100 % of SW–WD,
hydrolysis step is always de controlling phenomena due to its higher
initial soluble organic matter. In conclusion, the addition of SW-WD as
co-substrate in SS reduces the role of solubilisation of organic matter
step in AD and, as a result, biomethane production is favoured as it was
shown in Fig. 2. This fact must be taking into account for the optimi­ Fig. 4. Influence of on SW-DW percentage in feedstock composition on the
zation of the hydraulic retention time of the process under continuous kinetic parameter value: (A) fraction of non-biodegradable particulate substrate
operational mode. (CODp_nb), (B) fraction of non-biodegradable soluble substrate (CODs_nb), (C)
first-order kinetic parameter of hydrolysis step (k1), (D) first-order kinetic
parameter of acidogenic step (k2), (E) first-order kinetic parameter of meth­
anogenic step (k3), and (F) pseudo-stoichiometric coefficient yield methane
from COD (YCH4/COD).

6
V. Ripoll et al. Fuel 329 (2022) 125524

4. Conclusions

Revalorization of agro-industrial wastes through anaerobic co-

digestion contributes to develop the concept of biorefinery in the
framework of circular economy proposed by H2020 EU program. In this
sense, kinetic modelling is a powerful tool to provide useful information
on the overall performance of the process for the scaling up.
The addition of SW-DW, as a co-substrate in anaerobic digestion of
SS, enhances feedstocḱ s biodegradability as well as biomethane pro­
duction. These improvements are linked to the higher particulate and
soluble biodegradable fractions of SW-DW (61.6 % and 69.0 %,
respectively) in comparison to SS (58.6 % and 6.4 %, respectively).
A kinetic model considering hydrolysis, acidogenic, and methano­
genic steps was developed that successfully reproduces experimental
data supporting by statistical criteria (F–Fischer’s value, sum of squared
residuals, residual mean square error, and percentage of explained
variance). The percentage of SW-DW of feedstock has an inverse effect
on hydrolysis kinetic parameter and, therefore, on hydrolysis step rate.
Regarding acidogenesis and methanogenesis, substrate composition
does not influence the value of kinetic parameters involving in each step.
Analysing the rate-limiting step along the fermentation, methano­
genic step limits the overall rate at shorter periods of time whereas
hydrolysis step do it at longer times. In addition, SW-DW plays an
important role whithin switching the mechanism of methane produc­
tion, where acidogenesis and methanogenesis steps are enhanced since
the beginning due its higher initial soluble organic matter. This
consideration might be decisive to design and control the process in
continuous operational mode.

CRediT authorship contribution statement

Vanessa Ripoll: Conceptualization, Methodology, Validation,

Formal analysis, Writing – original draft. Rosario Solera: Conceptual­
ization, Resources, Writing – review & editing, Project administration,
Funding acquisition. Montserrat Perez: Conceptualization, Resources,
Writing – review & editing, Project administration, Funding acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence
the work reported in this paper.

Data availability

The authors do not have permission to share data.

Acknowledgement

This research work was supported by Andalusian Research, Devel­

opment and Innovation Plan (PAIDI 2020) under contract P18-RT-1348.

References

[1] Regulatory council of denomination of origin “Jerez-Xeres-Sherry” -“Manzanilla-

Sanlucar de Barrameda” - “Vinagre de Jerez”. Activity Memories Wine Sector
Technology Platform., https://www.sherry.wine/sites/default/files/memoria_
2020.pdf, 2020. [accessed 10 April 2022].
Fig. 5. Evolution of predicted reaction rate of each step (hydrolysis, acido­ [2] Borja R, Martin A, Maestro R, Luque M, Duran MM. Enhancement of the anaerobic
genesis and methanogenesis) in BMP test employing different SW–DW pro­ digestion of wine distillery wastewater by the removal of phenolic inhibitors.
portion in feedstock: (A) 0%, (B) 25%, (C) 50%, (D) 75%, and (E) 100%. Key: Bioresour Technol 1993;45:99–104. https://doi.org/10.1016/0960-8524(93)
black: hydrolysis step, red: acidogenic step, blue: methanogenic step. Solid 90097-U.
[3] Agabo-Garcia C, Perez M, Solera. Adaptation of thermophilic sludge-inoculum to
lines: model prediction. Dashed-lines: model prediction considering error range co-digestion with Sherry-wine distillery wastewater. Biomass Bioenergy 2020;139:
of estimated parameters. (For interpretation of the references to colour in this 105628. https://doi.org/10.1016/j.biombioe.2020.105628.
figure legend, the reader is referred to the web version of this article.) [4] Petta L, de Gisi S, Casella P, Farina R, Notarnicola M. Evaluation of the treatability
of a winery distillery (vinasse) wastewater by UASB, anoxic-aerobic UF-MBR and
chemical precipitation/adsorption. J Environ Manage 2017;201:177–89. https://
doi.org/10.1016/j.jenvman.2017.06.042.

7
V. Ripoll et al. Fuel 329 (2022) 125524

[5] Perez-Garcia M, Romero-Garcia LI, Rodríguez-Cano R, Sales-Marquez D. Effect of [18] Khadka A, Parajuli A, Dangol S, Thapa B, Sapkota L, Carmona-Martinez AA, et al.
the pH influent conditions in fixed-film reactors for anaerobic thermophilic Effect of the substrate to inoculum ratios on the kinetics of biogas production
treatment of wine-distillery wastewater. Water Sci Technol 2005;51(1):183–9. during the mesophilic anaerobic digestion of food waste. Energies 2022;15(3):834.
https://doi.org/10.1177/0734242X07079060. https://doi.org/10.3390/en15030834.
[6] Benitez FJ, Beltran-Heredia J, Real FJ, Gonzalez T. Aerobic and anaerobic [19] De Oliveira LRG, dos Santos DA, Fraga TJM, Juca JFT, Sobrinho MAD. Kinetics
purification of wine distillery wastewater in batch reactors. Chem Eng Technol assessment and modelling of biogas production by anaerobic digestion of food
1999;22:165–72. https://doi.org/10.1002/(SICI)1521-4125(199902)22:2<165:: wastes and acclimated sewage sludge. J Mater Cycles Waste Manag 2021;23(4):
AID-CEAT165>3.0.CO;2-T. 1646–56. https://doi.org/10.1007/s10163-021-01248-x.
[7] Ripoll V, Agabo-Garcia C, Perez M, Solera R. Improvement of biomethane potential [20] Abid M, Wu J, Seyedsalehi M, Hu Y, Tian G. Novel insights of impacts of solid
of sewage sludge anaerobic co-digestion by addition of “sherry-wine” distillery content on high solid anaerobic digestion of cow manure: kinetics and microbial
wastewater. J Clean Prod 2020;251:119667. https://doi.org/10.1016/j. community dynamics. Bioresour Technol 2021;333:125205. https://doi.org/
jclepro.2019.119667. 10.1016/j.biortech.2021.125205.
[8] Ripoll V, Agabo-Garcia C, Perez M, Solera R. Modelling of the anaerobic semi- [21] Zhang W, Xiao B, Zhang KE, Chen H, Guo X. Effects of mixing ratios on anaerobic
continuous codigestion of sewage sludge and wine distillery wastewater. Environ co-digestion of swine manure and rice straw: methane production and kinetics.
Sci Water Res Technol 2020;6:1880. https://doi.org/10.1039/D0EW00275E. Biomass Convers 2021. https://doi.org/10.1007/s13399-020-01211-0.
[9] Tena M, Luque B, Perez M, Solera R. Enhanced hydrogen production from sewage [22] Garcia-Ochoa F, Santos VE, Naval L, Guardiola E, Lopez B. Kinetic model for
sludge by cofermentation with wine vinasse. Int J Hydrog Energy 2020;45(32): anaerobic of livestock manure. Enzyme Microb Technol 1999;25:55–60. https://
15977–84. https://doi.org/10.1016/j.ijhydene.2020.04.075. doi.org/10.1016/S0141-0229(99)00014-9.
[10] Sillero L, Solera R, Perez M. Anaerobic co-digestion of sewage sludge, wine vinasse [23] Angelidaki I, Ellegaard L, Ahring BK. A mathematical model for dynamic
and poultry manure for bio-hydrogen production. Int J Hydrog Energy 2022;47(6): simulation of anaerobic digestion of complex substrates: focusing on ammonia
3667–78. https://doi.org/10.1016/j.ijhydene.2021.11.032. inhibition. Biotechnol Bioeng 1993;42:159–63. https://doi.org/10.1002/
[11] Zahedi S, Martin C, Solera R, Perez M. Evaluating the Effectiveness of Adding bit.260420203.
Chicken Manure in the Anaerobic Mesophilic Codigestion of Sewage Sludge and [24] Siles JA, Martin MA, Chica A, Borja R. Kinetic modelling of the anaerobic digestion
Wine Distillery Wastewater: Kinetic Modeling and Economic Approach. Biofuels of wastewater derived from the pressing of orange rind produced in orange juice
Biomass 2020;34(10):12626–33. https://doi.org/10.1021/acs. manufacturing. Chem Eng J 2008;140:145–56. https://doi.org/10.1016/j.
energyfuels.0c01852. cej.2007.09.026.
[12] Mata-Alvarez J, Dosta J, Romero-Güiza MS, Fonoll X, Peces M, Astals SA. Critical [25] Bollon J, Le-Hyaric R, Benbelkacem H, Buffiere P. Development of a kinetic model
review on anaerobic codigestion chievements between 2010 and 2013. Renew for anaerobic dry digestion processes: focus on acetate degradation and moisture
Sustain Energy Rev 2014;36:412–27. https://doi.org/10.1016/j.rser.2014.04.039. content. Biochem Eng J 2011;56:212–8. https://doi.org/10.1016/j.
[13] Borja R, Martin A, Sanchez E, Rincon B, Raposo F. Kinetic modelling of the bej.2011.06.011.
hydrolysis, acetogenic and methanogenic steps in the anaerobic digestion of two- [26] Lopez I, Benzo M, Passeggi M, Borzacconi L. A simple kinetic model applied to
phase olive pomace (TPOP). Process Biochem 2005;40:1841–7. https://doi.org/ anaerobic digestion of cow manure. Environ Technol 2020;42(22):3451–62.
10.1016/j.procbio.2004.06.026. https://doi.org/10.1080/09593330.2020.1732473.
[14] Lee E, Cumberbatch J, Wang M, Zhang Q. Kinetic parameter estimation model for [27] Lafratta M, Thorpe RB, Ouki SK, Shana A, Germain E, Willcocks M, et al.
anaerobic co-digestion of waste activated sludge and microalgae. Bioresour Development and validation of a dynamic first order kinetics model of a
Technol 2017;228:9–17. https://doi.org/10.1016/j.biortech.2016.12.072. periodically operated well-mixed vessel for anaerobic digestion. Chem Eng J 2021;
[15] Valvilin VA, Fernandez B, Palatsi J, Flotats X. Hydrolysis kinetics in anaerobic 426:131732. https://doi.org/10.1016/j.cej.2021.131732.
degradation of particulate organic material: An overview. Waste Manage 2008;28: [28] Angelidaki I, Alves M, Bolzonella D, Borzacconi L, Campos JL, Guwy AJ, et al.
939–51. https://doi.org/10.1016/j.wasman.2007.03.028. Defining the biomethane potential (BMP) of solid organic wastes and energy crops:
[16] Pavlostathis SG, Giraldo-Gomez E, 1991. Kinetics of Anaerobic Treatment: a a proposed protocol for batch assays. Water Sci Technol 2009;59(5):927–34.
critical review. Crit Rev Env Contr 1991; 21(5-6); 411-490. https://doi.org/ https://doi.org/10.2166/wst.2009.040.
10.1080/10643389109388424. [29] Hollinger C, Alves M, Andrade D, Angelidaki I, Astals S. Towards a standardization
[17] Ketsub N, Whatmore P, Abbasabadi M, Doherty WOS, Kaparaju P, O’Hara IM, et al. of biomethane potential tests. Water Sci Technol 2016;74(11):2515–22. https://
Effects of pretreatment methods on biomethane production kinetics and microbial doi.org/10.2166/wst.2016.336.
community by solid state anaerobic digestion of sugarcane trash. Bioresour [30] American Public Health association. Standard Methods for Examination of Water
Technol 2022;352:127112. https://doi.org/10.1016/j.biortech.2022.127112. and Wastewater. 21st ed. APHA-AWWA-WPCF, New York, USA; 2005.

8
 Journal of Environmental Management 241 (2019) 32–43

Contents lists available at ScienceDirect

Journal of Environmental Management

journal homepage: www.elsevier.com/locate/jenvman

Research article

Technical and economic investigation of chemical scrubber and bio- T

filtration in removal of H2S and NH3 from wastewater treatment plant
Ebrahim Alinezhada, Mohsen Haghighib, Farhad Rahmanic, Habib Keshizadeha, Mahyar Abdia,
Kazem Naddafia,∗,1
a
Department of Environmental Health Engineering, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran
b
Department of Environmental Health Engineering, Faculty of Health, Kashan University of Medical Science, Kashan, Iran
c
Department of Chemical Engineering, Faculty of Engineering, University of Kurdistan, Sanandaj 66177, Iran

ARTICLE INFO ABSTRACT

Keywords: A detailed techno-economic comparison of a chemical scrubber (CS) and a bio-filter (BF) was conducted over a
Ammonia 45-day time period at a municipal wastewater treatment plant (WWTP), Yazd city. The assessment of emissions
Hydrogen sulphide quantity indicated that odor emissions from the Yazd WWPT mainly consist of hydrogen sulfide (H2S) and
Biofilter ammonia (NH3). It was also found that odor gaseous loading changes corresponding to water consumption
Chemical scrubber
pattern in society (R2 = 0.922) for H2S and (R2 = 0.978) for NH3. The highest level of 25 and 3 ppm for H2S and
Techno-economic evaluation
NH3, respectively were detected at specific times during the day. The BF system was continuously supplied with
Yazd WWPT's off-gas treatment while the CS was only examined at the times during the day when the gas
emissions are at the highest level. The removal efficiency of NH3 and H2S were found to be affected by their
respective loading rate. Additionally, among the various oxidants examined in the CS, the NaOCl solution
showed the best results in terms of removal efficiency and compatibility. The experiment revealed almost
complete removal of NH3 while the H2S removal efficiency remained above 95% for both systems regardless of
the operating conditions. This study clearly demonstrates the effectiveness of both systems in treating actual
waste gases containing H2S and NH3. By comparing the gas loading rate of both systems and considering lim-
itations of the BF system, the CS seems to be more efficient applicable odor control technology from a technical
viewpoint. From the economic viewpoint, comparisons revealed that chemical usage and operating expenses
were costly parts of the CS and the BF, respectively. The economic indexes of 1.58 €.m−3. h−1 and 2.57 €.m−3.
h−1 were obtained for the BF and CS, respectively, reflecting cost-effectiveness of the BF system.

1. Introduction hydrogen sulfide (H2S) (Alfonsín et al., 2015; Nisola et al., 2009). H2S
and NH3 are highly toxic, colorless, corrosive and irritating malodorous
The development and infringement of urban residential areas on nuisances with very low odor thresholds. These compounds not only are
potential malodor sources specifically wastewater treatment plants a direct threat to human health but also represent a remarkable con-
(WWTP) has led to an increasing number of public grievances against tribution to photochemical smog formation and particulate secondary
these malodorous compounds (Lebrero et al., 2011). Frequent ex- emission (Gao et al., 2001). Hence, it is essential to control the emission
posures to these malodors have a direct threat to people's health and of these gases not only for public health and safety but also for en-
welfare (Chen et al., 2001; Cheng and Hsieh, 2010). This factor, along vironmental protection using a cost-effective and eco-friendly odor
with environmental sticking legislations about air pollution control in abatement system (Van der Heyden et al., 2015). Up to now, various
connection with industries and odor resources, has led to the need for technologies have been proposed and applied for the treatment of
odor management (Sucker et al., 2009). Odorous compounds are pri- malodorous components from contaminated air which can be classified
marily produced during biological metabolisms in wastewater collec- into physical/chemical (chemical scrubbers, incinerators, adsorption
tion and treatment systems (Jaber et al., 2014). Generally, major systems, and so forth), biological systems (biofilter, bio-trickling filter
sources of smell encountered in WWTP are ammonia (NH3) and (Arellano-Garcia et al., 2012; Gabriel et al., 2013; Qiu and Deshusses,

∗
Corresponding author. Department of Environmental Health Engineering, Tehran University of Medical Sciences, P.O.Box 6446-14155, Tehran, Iran.
E-mail address: [email protected] (K. Naddafi).
1
web: https://www.tums.ac.ir.

https://doi.org/10.1016/j.jenvman.2019.04.003
Received 30 April 2018; Received in revised form 16 February 2019; Accepted 2 April 2019
Available online 10 April 2019
0301-4797/ © 2019 Elsevier Ltd. All rights reserved.
E. Alinezhad, et al. Journal of Environmental Management 241 (2019) 32–43

2017), bio-scrubber, and activated sludge diffusion reactors) or a on simultaneous H2S and NH3 removal were also investigated.
combination of them. Chemical scrubbing (CS) and bio-filter (BF) are
among the most commonly employed treatment techniques in WWTPs 2. Materials and methods
(Kumar et al., 2013; Lebrero et al., 2011).
The CS is a well-known, widely established, and reliable technology. 2.1. Site description of biofilter and chemical scrubber
This mainly arises from short gas retention time (as low as 1–2.5), ex-
tensive experience in design and operation, high robustness when Two odor control systems were designed and installed at the Yazd
properly operated, and rapid start-up of the wet scrubber (Moussavi wastewater treatment plant in Yazd, Iran. The treatment plant is an SBR
et al., 2008; Vega et al., 2014). Removal efficiencies of up to 98% for activated sludge treatment facility, which treats municipal wastewater
H2S and 99.8% for NH3 are commonly achieved under good operation of Yazd city. This study was conducted over the months of June and
practices (Lebrero et al., 2011). However, this technique is relatively July 2017 in which odor emissions were at a high level during the year.
costly and produces secondary pollutants. Both control systems were designed to control off-gases and odors from
In contrast to physical/chemical treatments, biological process is the pumping station because this section had a great portion of waste
often considered a cost-effective and eco-friendly route for the pur- gas emission from the plant.
ification of gas streams (Van Groenestijn and Kraakman, 2005). The
main environmental impact of biological techniques was caused by 2.2. Chemicals
high water consumption to maintain biological activity. Among mi-
crobiological processes, the BF has proved to be one of the most pro- The material used were: N, N-dimethyl-p-Phenylenediamine, H2SO4
mising methods in WWTPs’ odor treatment (Lewkowska et al., 2016). 50%, FeCl3, and Na2S.9H2O that was applied for the determination of
The removal efficiency of an odor gas by a BF is generally satisfactory H2S ranges in samples, and Na2 [Fe (CN) 5NO], NaOH 6.75 g mol−1,
under appropriate sets of conditions. However, it reduced significantly NaOCl 0.1 N, C6H5OH 45%, Na3PO4.12H2O, CHCl3, and (NH4)2SO4,
when applied to undiluted and hardly biodegradable waste gases at were used for preparation of solutions which were utilized for detection
large quantities of the contaminated gas stream (Lebrero et al., 2011). of NH3. Besides, CdSO4.H2O used as a precursor to making alkaline
The BF is also criticized for high land and material requirements; and suspension of cadmium hydroxide as an absorption solution for cap-
their long period of acclimation required for microbial population turing H2S from polluted air. In addition, H2SO4 (0.1 N) was used as
(weeks or even months) (Lebrero et al., 2011). In addition, treatment of NH3 absorption solution. The oxidant compounds (NaOCl and H2O2)
low solubility gases such as NO in conventional biofilters has typically purchased in the commercial grade with purities of 12% and 35%, re-
required long contact times (Kumar et al., 2010). Recently, the mem- spectively, but KMnO4 was 99% purity in analytic grade. All the che-
brane bioreactor (MBR) technology for air pollution control has been mical compounds in this study except NaOCl and H2O2 were analytic
proposed (Kumar et al., 2008, 2009). In spite of operational advantages grade (Merck-Germany) and used without further purification.
compared with the biofiltration, membrane bioreactors for air pollution
treatment have yet to be tested at full scale and their long-term op- 2.3. Biofilter odor control system
erational stability still has to be demonstrated. Furthermore, an MBR
has a high construction cost disadvantage (Luvsanjamba et al., 2008). Fig. 1 indicates the experimental setup of BF at Yazd wastewater
Both of these technologies (BF and CS) have been widely reviewed treatment plant. It was made up of Plexiglas with a height of 95 cm and
in the literature, and their optimal range of application and perfor- an internal diameter of 14 cm. BF consists of two sections and a 2 L
mance for odorous compounds treatment has been clearly established bottle at the bottom for the collection of the solution. Each section is
on the basis of laboratory and field experiences. Despite numerous filled with equal amount of bed material to a height of 40 cm giving a
studies conducted on the BF and wet scrubbers, only a few ones have total working volume of 12.03 L. The bed materials (Fig. 2) used for the
been focused on the comparative evaluation of these odor-treatment treatment of real gaseous mixture were activated sludge with con-
technologies (Estrada et al., 2010; Gao et al., 2001). In this regard, Gao centration of 4–6% containing ammonium-oxidizing bacteria (AOB) to
et al. compared a BF and a sodium hypochlorite/sodium hydroxide wet improve nitrification, which mixed with Poly Vinyl Chloride (PVCs),
scrubber in order to control H2S in the Mill Greek Wastewater Treat- which were first chopped to pieces with 2 cm dimension and afterward,
ment Plant in Cincinnati, Ohio economically and technically (Gao et al., these pieces were cut into three equal parts with using longitudinal
2001). In another study, various odor control technologies including incision (approximately 6.6 mm) and a compost (density of
BF, activated sludge diffusion, bio-trickling filtration, CS, activated 500 kg m−3). The bed materials in each section were supported by
carbon adsorption, regenerative incineration, and a hybrid technology plates made of Plexiglas®. The experimental air for the operation is
have been comparatively evaluated in terms of environmental perfor- supplied using an air compressor, which was prepared by sucking in the
mance, process economics, and social impact by using the IChemE ambient air. The Flow of air into the BF was adjusted using a flow
Sustainability Metrics in the context of odor treatment from wastewater control valve attached to a flow meter (Brooks Instrument Company).
treatment plant (WWTP). The odor pollution from a WWTP located at
Stuttgart University was selected as a model of malodorous emission for 2.3.1. Moisture supply methods in BF system
the environmental evaluation. Methyl mercaptan, hydrogen sulfide and In fact, we applied two methods for creating moisture for micro-
mainly, VOC components were included in the above-mentioned model organism activity in BF: (1) a water bath, and (2) water trickling from
emission (Estrada et al., 2010). According to this simple literature the top of the BF. In the first method, input air was passed through the
survey and to the best of our knowledge, no research has dealt with the water bath (diameter of 15 cm and height of 80 cm) before entering the
detailed comparison of the wet scrubber and BF in order to simulta- bed. It caused input air loads water vapor droplets, and then these
neously remove NH3 and H2S in an air stream. On this basis, a detailed vapor droplets could be dispersed uniformly through the whole of the
techno-economic comparison of these two odor-abatement technologies bed. In the second method, we added 200 mL of the deionized water
was conducted over a 45-day time period at a municipal WWTP (lo- (pH ≈ 7) into the bed about every three days. In low airflow rates, the
cated at Yazd, Iran) as model odor source to provide a basic guideline water bath used to provide moisture, and in high input airflow rates,
for technology selection in odor management in Yazd WWTP. In this both methods (1 and 2) were applied to provide moisture in the bed. As
regard, “removal efficiency” and “total annualized cost” are taken as regards, in the high input airflow rates, there is a high microorganism
technical and economic indexes, respectively. In line with technical activity in the BF bed , the temperature of the system increases further,
evaluation, the effects of the gas loading rate and EBRT for both systems and conversely, the moisture content decreases. Thus, the supply of
and also, oxidant nature and scrubbing flow rate for the wet scrubber, moisture in high airflow rates is more difficult.

33
E. Alinezhad, et al. Journal of Environmental Management 241 (2019) 32–43

Fig. 1. Schematic flow chart and real pilot of systems, (A) Chemical scrubber, (B) Biofilter.

34
E. Alinezhad, et al. Journal of Environmental Management 241 (2019) 32–43

Fig. 2. Biofilter bed components and preparation.

2.4. Chemical scrubber contains 20 mL adsorbent solution of sulfuric acid. Impinger solutions
were analyzed by DR 5000 at wavelength λ = 630 nm. The NH3 con-
The schematic of the CS system is depicted in Fig. 1, the pilot-scale centration (ppm) in the air calculated by the following equation (Lodge,
packing bed scrubber was constructed using Plexiglas with an internal 1988).
diameter of 14 cm and a total height of 120 cm. The bed height and
NH3 (mg.m 3 ) = ((17.03×NH3 ppm))/24.45 (2)
volume were 70 cm and 10.77 L, respectively. The bed material made of
Polyethylene (PE) with the porosity of 80% and specific surface area In here 17.03 is NH3 molar mass and 24.45 is the molar volume of
369.6 m2 m−3. In this system, the polluted air is supplied by a suction an ideal gas in 25 °C and 101.3 kPa.
pump, the inlet airflow rate to the CS was adjusted and monitored by a
flow meter and provided valves. Three plates were placed across the 2.6. Measurement of moisture, pH, and temperature in biofilter
reactor for holding the packing material and helped to diminish the
channeling of the fluid through the bed. The area below the bed served In order to maintain a suitable level of moisture, every three days
as a liquid pool to save the scrubbing solution for recirculation. A dif- roughly 200 mL of deionized water was trickled from top of the BF.
fusing pump circulated this liquid from the pool and sprayed on the top Therefore, everyday sampling from two sections of the bed was done
of the bed using a shower. Three chemical oxidants consisting of NaOCl, and after analysis in the laboratory, the moisture in each bed of the BF
H2O2, and KMnO4 were utilized for oxidizing odor mixture compounds. calculated by Mn = Mw/Mo
The odor treatment formula in this system was the reactions that Where Mn is the moisture content, Mw is the mass of medium with
happen between circulating liquid and malodorous air in a reverse water, Mo is the mass of medium without the water. For measurement
current in the CS bed. The oxidants, which were made in a basin, had pH, it took 5 g of each bed substrate, from sampling ports (6) in Fig. 1
been injected (4–10 L min−1) into the CS bed by a centrifuge pump and and mixed it with 50 mL water, and afterward, this was set up for the
a shower. time of 30 min on the mixer with a speed of 350 rpm. Finally, the pH of
this composite was determined by using an Ecoscan pH meter (Eutech
2.5. H2S and NH3 analysis instruments). The temperature of the bed was also measured in three
periods of the day (morning, afternoon and evening) by placing a
In order to assess the efficiency of each system, we need to measure thermometer in each bed of BF.
the concentrations of malodorous chemicals in the inlet and outlet of
the gas. 2.7. The EBRT, airflow rate (Q), pressure (ΔP), and gas velocity
H2S: It was used per the Jacob and et al. the method for measuring measurement in systems
H2S gas. The H2S sampling set up consisted of two impingers and an air
vacuum pump and tubes. A vacuum pump aspirated the air with The EBRT and airflow rate (Q) in systems are calculated by the
Q = 1 L.min−1 and T = 5–30 min into the impinger that contains following equations (Eq. (4));
45 mL adsorption solution of cadmium hydroxide. To protect samples
EBRT = Q/V and Q = V/t (3)
from sunlight, impingers were covered entirely by aluminum foil.
Impinger solutions were analyzed using DR 5000 at wavelength −1
Where Q is the airflow rate (L.min ), V is reactors bed volume
λ = 670 nm. The H2S concentration (ppm) in the air was calculated by (m3), and t is gas contact time (s). The pressure drop in either the CS or
the following equation (Lodge Jr, 1988). the BF bed was determined by a water manometer, which was installed
in two head of the beds and the water level fluctuates in this device
H2 S(ppm) = ((A×0.719))/Vs (1)
before and after sampling for regarding EBRT, calculated ΔP in systems.
Where A is H2S (μg) in solution, Vs (L) is the volume of air that passes The gas velocity in systems was estimated by this equation:
through the adsorbent solution in standard circumstance, and 0.719 is
USG = QG /A (4)
the volume of 1 mg H2S gas in stoichiometric conditions.
NH3: The NH3 sampling method resembles H2S but the Q of the Where USG is gas velocity m.s and QG is inlet gas flow rate m . s−1
−1 3

vacuum pump is 1–3 L min−1 and T = 30 min into the impinger that and A is the surface area of the bed m2.

35
E. Alinezhad, et al. Journal of Environmental Management 241 (2019) 32–43

Fig. 3. (a) H2S and NH3 daily emissions and those peaks (for 23rd day); (b): Relevance between inlet wastewater to Yazd WWTP with H2S and NH3 emission ranges
during a day.

Table 1 2.8. Economical evaluations

H2S efficiency & removal rate for BF.
The economic analysis must take into account equipment, con-
Time (day) EBRT (s) Q (L/min) C inlet (ppm) C out (ppm) Efficiency (%)
struction costs, and operating costs (raw materials, utilities, and op-
air

1–9 60 9.2 17.75 2.8 84 erator wages). Then, costs including capital and expense of the various
12–16 60 9.2 24.02 0.29 99 options should be compared using economic parameters as net present-
16–26 30 18.4 20.13 0.54 97 worth or annualized cost. In this study, the annualized cost method has
26–38 20 27.6 23.36 0.56 98
38–45 15 36 22.23 1.7 92
been chosen and applied. In the Supplementary Data, details of the cost
estimation procedure are depicted (Gao et al., 2001).

36
E. Alinezhad, et al. Journal of Environmental Management 241 (2019) 32–43

Table 2 start-up it in times when there are high concentrations of them. Fig. 3a
NH3 efficiency & removal rate for BF. shows daily H2S and NH3 emissions at Yazd WWTP in Iran. As can be
Time (day) EBRT (s) Q air (L/ C inlet (ppm) C out (ppm) Efficiency (%) seen, odor emission ranges are low, except for peak periods. The highest
min) level of the odorant gases observed in time ranges of 9:00–11:00 a.m.,
3:00–5:00 pm, and 11:00–12:00 p.m. during the day with the main
1–6 60 9.2 2.1 0.26 88
peak concentration of about 25 and 3 ppm for H2S and NH3, respec-
12–16 60 9.2 2.71 0.018 97
16–26 30 18.4 2.14 0.037 98
tively. According to the results, it can be seen that there is a direct
26–38 20 27.6 1.88 0.036 98 relationship between water consumption pattern in the community and
38–45 15 36 1.97 0.007 99.6 odor emissions. It is clear that the highest level of odor emissions occur
during the peak period of water consumption. With increasing water
consumption the QInlet wastewater increases, and this phenomenon leads
3. Results and discussions to an increase in the odor emissions at WWTP. Meanwhile, with this
scrutiny, it can be seen that there is a relationship between water
3.1. H2S and NH3 emission ranges consumption and production of NH3 and H2S gases. In general, several
factors have a role in the occurrence of sulfur cycle in wastewater
Assessment of the odor compounds level and quantity of emissions networks such as microorganisms in wastewater, wastewater volume,
is essential for the better performance of odor control systems. This data the amount of sediments in the sewer, and wastewater contact area
will also help operators to shut the odor control system down when with air (Lens and Kuenen, 2001). The sulfur cycle is completed in four
marginal levels of malodorous compounds are in the environment or main processes including production of sulfide, transition sulfide from

Fig. 4. H2S and NH3 inlet, outlet, and removal efficiencies in BF. (a) H2S and (b): NH3.

37
E. Alinezhad, et al. Journal of Environmental Management 241 (2019) 32–43

Fig. 5. H2S and NH3 removal efficiencies by CS in present of three oxidants. (a) H2S and (b): NH3.

liquid to gas phase, chemical and biological oxidation of sulfur, and the QInlet wastewater has been shown in Fig. 3b with the correlation coeffi-
reaction of sulfide ion-containing compounds with iron and chemical cient R2 = 0.990 and R2 = 0.978 for H2S and NH3, respectively. This
sequestration, sulfide created by bacteria, which converted sulfate into relation is important in the CS start-up times more than other tech-
the sulfide (Parande et al., 2006). This bacteria activity occurred in the nologies.
anaerobic condition in wastewater treatment collection networks and is
seen in areas where the wastewater flow moves slowly and is poorly 3.2. Biofilter efficiency and removal rate study
aerated (Hvitved-Jacobsen et al., 2000). Ammonia is another substance
that produces odor in the sewer by bacterial decomposition of urea in 3.2.1. Effect of gas flow rate and EBRTs in simultaneous removal of H2S
wastewater networks. Nitrogen organic compounds and urea are con- and NH3
verted into ammonia through the ammonification aerobic process in The efficient and simultaneous removal of H2S and NH3 by the bio-
wastewater collection networks. On the other hand, NH3 has a low filtration process as evaluated and the results are given in Tables 1 and
evaporation temperature, and it can quickly be evaporated and released 2. The H2S and NH3 removal process can be classified into four stages in
into the environment (Talaiekhozani et al., 2016). Therefore, H2S and the BF system in terms of inlet air (9.2, 18.4, 27.6 and 36 L. min−1) and
NH3 are produced when the water consumption rate is low and was- EBRT (15, 20, 30 and 60 s) Fig. 5. The incubation time is one of the
tewater flow speed is slow in the sewer, but when this flow goes up in factors determined when BF has reached a stable state and is robust in
wastewater pipes (high water consumption) these odorants are released the reduction of malodorous compounds. In fact, after this time biolo-
into the air. This marked relevance between odor emission rate and gical activity starts in the BF bed. In the start-up, these incubation

38
E. Alinezhad, et al. Journal of Environmental Management 241 (2019) 32–43

is loaded in the system increased (maximum 35 ppm) and caused a

reduced removal efficiency from 98 to 92%. These factors (ΔP and mass
loading) did not have any noticeable effect on the BF performance for
the removal of NH3. This is attributed to this fact that the mass loading
rate (maximum 3 g m−3. h−1) entering to the system is not so high and
the BF is able to degrade approximately all of NH3 gas (99.6% EBRT of
15s) Fig. 4b. The obtained results are consistent with the results of
Rehman et al. (2009) when after a power failure, the maximum removal
efficiency of H2S diminished from 100 to 90% (Rehman et al., 2009).
Also, similar observations were reported previously by Rabbani et al.
(2016). In conclusion, it is obvious that QInlet air and EBRT are crucial
factors in BF design and performance. Due to the high sensitivity of this
system, it needs more consideration to select these parameters in order
to facilitate the reactor construction.

3.3. Chemical scrubber efficiency and removal rate study

3.3.1. Effect of gas flow rate and EBRTs in simultaneous removal of H2S
and NH3 in the presence of oxidants
Fig. 6. ΔP fluctuates corresponding to gas velocity (airflow rate) changes in CS. The performance of the CS was evaluated for efficient simultaneous
removal of malodorous compounds including H2S and NH3 in terms of
gas flow rate (100, 150, 200 and 250 L min−1), EBRTs (6.46, 4.3, 3.23
Table 3
and 2.58 s) and scrubbing flow rate (4–10 L min−1) in the presence of
Oxidants fractions in CS.
three oxidants (NaOCl, H2O2 and KMnO4).
Oxidant Consumed oxidant mg/Oxidized malodorous mg Scrubbing liquid pH Adsorption process coupled with chemical oxidation is responsible
for the removal of H2S and NH3 in the CS. Therefore, the inlet flow rate,
Cl2 6.5–7.1 8–9
H2O2 1.5–2.2 7.8–8.1 gas contact time and scrubbing liquid are crucial factors affecting H2S
KMnO4 9.5–10.4 7.5–8.1 and NH3 removal efficiency via mass transfer from the gas phase to the
liquid phase. In this regard, the effect of air flow rate (100, 150, 200
and 250 L min−1 which considered areactor bed volume of
periods were 9 and 6 days for H2S and NH3, respectively. These in- 10.77 L min−1) and gas contact time (6.46, 4.3, 3.23 and 2.58 s) on H2S
cubation figures were partly higher in comparison with other studies and NH3 removal efficiency was investigated in the presence of NaOCl,
mentioned in the literature (Chen et al., 2018; Chung et al., 2000; KMnO4 and H2O2 oxidants as scrubbing liquid and depicted in Fig. 5. As
Galera et al., 2008; Hou et al., 2016; Jones et al., 2004; Kim et al., 2002; can be seen in Fig. 5a, with increasing the air flow rate from
Rabbani et al., 2016). Forasmuch as prepared biomass (containing Qair = 150–200 L min−1 and subsequently, decreasing bed contact time
adapted bacteria), was not used in the BF system, so it needed more from 4.3 to 2.58 s, the H2S removal efficiency decreases from 100 to
time to start up biological activity (adaptation) and reach stability in 95% for NaOCl and KMnO4 and also drops from 100 to 45% for H2O2.
the system. On the other hand, in aerobic conditions, H2S is converted According to Fig. 5b, NH3 removal efficiency does not vary too much
to sulfate (SO4)2- by sulfur-oxidizing bacteria (SOB), and the pH ranges because of low NH3 concentrations in the inlet gas to the CS. The re-
for optimal growth of SOB are variable (1.8–7.4), which depends on the moval efficiencies of 100, 98.5 and 96.3% by employing NaOCl and
type of bacteria. The pH value is 7.4 and below 1 for Thiobacillus de- 100, 97.5 and 95% using KMnO4 were obtained in the gas contact times
nitrificans and Acidithiobacillus thiooxidans, respectively. In the same of 4.3, 3.23, and 2.58 s, respectively. Statistic appraisal of the results,
environment, AOB (ammonia oxidizing bacteria) and NOB (nitrite presented in Fig. 5, indicates that increasing superficial gas velocity
oxidizing bacteria) convert NH3 to NO3− and the optimum range of pH (airflow rate) and subsequent reduction of gas contact time; the re-
for these kinds of bacteria are 6.0–9.0 for Nitrosomonas and 7.3–7.5 for moval efficiency lessens regardless of oxidants used.
Nitrobacter (Lors et al., 2009; Solcia et al., 2014). According to pH Increasing QInlet air up to 250 L min−1 makes an increase in pressure
graphs in two beds (Fig. S1b), pH in initial days (around 7.9–7.8) was drop (ΔP) from 4 to 7.1 cm H2O in the bed (Fig. 6), and hinders the
favorable to NH3 converting bacteria rather than H2S ones, so NH3 mass transfer from gas to liquid phase ratio (L/G). So that, with de-
degradation by BF was begun sooner than H2S in the contaminated air. creasing liquid phase ratio, the removal efficiency declines too (Chen
After a stable condition was reached in the BF system, in EBRT = 60 s et al., 2001). Moreover, pressure drop which plays a key role in vertical
the removal rate of H2S and NH3 were 99% and 97% respectively. upward flows is affected significantly by superficial gas velocity (USG).
Further diminished EBRTs (15 s) caused deterioration in the removal Accordingly, when USG increases, the scrubbing liquid acts as a hamper
ability of BF. Decreasing EBRT increases the ΔP value (from 4 to 6 mm to the gas circulation from the bottom to the top of the contactor by
H2O) and a gradual drying of the biomass materials occurs in the BF forming a thick layer. Therefore, water drops are scattered into small
bed, following that H+ ions increase in the system which created the drops, which are driven by the gas flow, a thin layer can be observed
acidification event in packing materials, as a result, this phenomenon is along the pipe walls and whereupon, the CS performance drops
a hamper to biological activity. Hence, with the reduction of bacteria (Sanchez et al., 2007). Moreover, odor compounds are less water-so-
activities in the BF, the removal efficiency dropped as well (Lebrero luble and then, with decreasing the gas contact time to 2.58 s, the H2S
et al., 2013; Ma et al., 2006). Therefore, with increasing the airflow rate and NH3 molecules could not completely transfer to the liquid phase
from 27.6–36 L min−1 and EBRT from 20 to 15 s the BF removal effi- and caused a negative impact on the CS removal efficiency (Albright,
ciency for H2S decreases from 98 to 92% Fig. 4a. Additionally, in- 2009; Moosavi et al., 2005). The high performance of CS for all oxidants
creasing airflow rate in the system enhances mass loading, which is in EBRTs of 4.3 and 3.23 s is attributed to the use of media with high
mass of contaminant entering the BF per unit volume of filter materials specific surface area (plastic is one of the selective media in the CS
per unit time, and consequently disturbing adsorption processes be- because of less susceptibility to collapsing and clogging (Feilberg et al.,
tween bed bacteria and pollutant (Fulazzaky et al., 2014; Lebrero et al., 2010), hence, in this study PE was used with a high surface area), op-
2013). Thus, with an airflow rate of 36 L min−1 the amount of H2S that timum rate of scrubbing liquid flow and mainly long retention time of

39
E. Alinezhad, et al. Journal of Environmental Management 241 (2019) 32–43

Fig. 7. Effect of scrubbing liquid on CS performance in removal of H2S & NH3 in airflow rates of 200 and 250 L min−1. (a) NaOCl and KMnO4, (b) H2O2.

Table 4 Table 5
Initial investment cost in systems, (a): BF and (b): CS. Initial operating cost in systems, (a): BF and (b): CS.
Initial investment cost parameters Cost (€) Initial operating cost parameters Cost (€.year−1)

a a
Design and construction 133 Electricity 210
Packing material 76 Water consumption 0.0744
Supportive equipment costs (diffusing pump) 21 Labor 4380
Total initial investment cost 230 Total yearly operating cost 4590.0744
b b
Design and construction 120 Electricity 70.5
media 12 Water consumption 122
Humidifier & other equipment 47.5 Labor 4380
Total initial investment cost 179.5 Chemical usage 12768
Total yearly operating cost 17340.5

40
E. Alinezhad, et al. Journal of Environmental Management 241 (2019) 32–43

Table 6 the diving force. Second, the scrubbing solution in high USL is re-
Total annualized cost and annualized cost parameters for each system, generated, which under these conditions oxidant concentration is high
(a): BF with airflow rate of 36 L min−1 and EBRT 15 s; (b): CS with air and reaction velocity improved consequently.
flow rate of 250 L min−1 and EBRT 2.58 s. Furthermore, for avoiding dry spots, which could be a bypass for
Annualized parameters Cost (€.year−1) untreated air leaving the CS, the scrubbing solution should be uni-
formly distributed at the top of the packing material (Sanchez et al.,
a
2007). Then, in QL below 7 L min−1, these dry areas decrease the CS
AIC 40.244
AOC 29835.48
performance against gases treatment. In contrast, in Q = 7 Lmin−1 and,
APMRC 32.48 above the CS bed has overcome the gas to liquid mass transfer limita-
Total annualized cost 29908.2 tion, which is wetting media ratio and reaches to high removal effi-
b ciency. According to the results presented in Fig. 7, it is obvious that the
AIC (€/year) 51.566
scrubbing flow rate = of 7 L min−1 is suitable QL for achieving high
AOC (€/year) 112713.5
APMRC (€/year) 18.56 performance in the CS and is in good agreement with that reported in
Total annualized cost (€/year) 112783.62 the literature (Moussavi et al., 2008).

gas in the CS. These claims have been confirmed by previous studies 3.4. Economic analysis
(Charron et al., 2004; Chen et al., 2001). As can be seen from Fig. 6 (a &
b), H2O2 has a low removal efficiency among other oxidants in 3.4.1. Investment cost
EBRT = 2.58 s for H2S (45%) and NH3 (89%). The oxidative properties The cost analysis for both systems (CS and BF) at Yazd WWTP was
of H2O2 have a direct bond with the pH of the scrubbing liquid and its divided into three chief parts: initial investment cost, operating cost,
high performances have been observed in alkaline PHS. With increasing and replacement packing material cost. All these costs and their details
the pH in the scrubbing liquid, the concentration of HS− increases are presented in Table 4 (a & b). Investment cost mainly depends on the
corresponding with improved alkalinity in the oxidant liquid. Thus, this volume of reactors, which is determined by various parameters such as
the system's EBRT, airflow rate, concentration and type of the pollutant,
causes not only the enhancement of the H2S solubility but also increases
mass transfer between the gas and water phases (Moussavi et al., 2008). and required removal efficiency. In this study, investment cost was
classified into the three main parts; design & construction, media, and
As a result, the pH range of 7.8–8.1 in the scrubbing solution according
to Table 3 is not favorable to have high removal efficiency as a result of other supportive equipment (water bath in the BF and diffusing pump
in the CS and so on) cost. Design & construction cost was estimated 120
using H2O2. These results are consistent with previous findings
(Moussavi et al., 2008). Also, similar observations have already been and 133 € for the BF and CS, respectively. Forasmuch as volumes, and
design factors were approximately close to each other and similar
reported in various studies (González-Sánchez et al., 2008; Jiang et al.,
2009; Karageorgos et al., 2010). This study was accomplished without material was used to construct both systems, there was no significant
difference in this part of the costs. Considering that the main portion of
the addition of any pH adjusters into the scrubbing liquid because one
of the aims of this study was that all comparative data of two systems the BF bed is made using compost, which was cheaper than PE material
(CS packing material), the cost of this section was lower in the BF
are collected in real circumstances in the field. The results indicated
that EBRT associated with inlet airflow rate (QInlet air) and type of system. Other supportive equipment costs were the last factor for esti-
mated total investment cost in systems (47.5 € for BF and 21 € for CS).
oxidant are the main parameters in reactor design and its performance
from a technical and economic point of view. In this area, piping cost and other details cost were similar in two
methods, but the water bath in the BF versus a diffusing pump in the CS
made a difference in this section of fees. Plexiglas®, which was applied
3.3.2. Effect of scrubbing liquid rate on reactor performance
to the constructed water bath in the BF, was worthy rather than dif-
Adsorption is directly proportional to the content of the liquid ex-
fusing pump in the CS. All in all, the design and construction sector was
posed to the gas stream. Therefore, the scrubbing liquid rate is a critical
a highly expensive one above investment cost, and the total initial in-
factor to achieve an optimum operation in the CS designing.
vestment cost was 179.5 € and 230 € for the BF and CS, respectively. As
In this regard, the effect of scrubbing liquid flow rate in the range of
a result, this part of the cost can be very different study-by-study be-
QL = 4–10 L min−1 was surveyed and presented in Fig. 7.
cause it mostly it depends on process design parameters, commodity
According to Fig. 7a, increasing the liquid flow rate from 4 to 7
prices, time, and location of study.
L.min−1 in Qair = 200 and 250 L min−1 raises the H2S and NH3 removal
efficiency from 65% to 55% to above 95% for NaClO and KMnO4 oxi-
dants, respectively. A similar trend in the removal efficiency was also 3.4.2. Operating cost
observed with the increase of the H2O2 flow rate, as shown in Fig. 7b. Operating cost is expressed as the yearly cost in Euros (€.year−1). In
The removal efficiencies culminate from low ranges to above 40% and the CS system operating costs include the electricity, water consump-
above 70% in QInlet air = 250 L min−1 for H2S and NH3, respectively. tion, labor, and chemical usage costs, while the BF system also includes
Thus, the findings indicate that the changes in QL had a direct effect on all of these costs except chemicals. The BF worked more times than the
the CS performance in the simultaneous removal of H2S and NH3 gases. CS and, its electricity usage cost (210 €.year−1) was more than its
Considering that QL and superficial liquid velocity (USL) are directly opponent (70.5 €.year−1). Moreover, the water consumption costs were
related, the USL affects the removal process in two ways. First, with 0.0744 €.year−1 and 122 €.year−1 for the BF and CS, separately. In
increasing USL, mass transfer parameters enhance due to the supported fact, the CS is a water-based system, hence, for better performance, the

Table 7
Economic comparison of CS and BF systems in different literature.
System EBRT (s) Airflow rate (m3/h) Annualized cost (€/year) Index (€/m3.h) Reference

BF 60 20000 42000 1.8 Prado et al. (2009)

BF 60 50000 – 2 Estrada et al. (2012)
CS 70 17000 14000-23000 – (Gabriel and Deshusses, 2004)
CS 4 50000 – 3.6 Estrada et al. (2010)

41
E. Alinezhad, et al. Journal of Environmental Management 241 (2019) 32–43

whole surface of media needs to be covered with water. Also, chemical and BF, respectively. Although the BF system is preferred from the
usage is a highly expensive sector in the CS and affects total operating economic viewpoint, it suffers from some of the operational drawbacks
costs in this system. Because of the high fraction of NaOCl consumed mg/ such as performance fluctuation, low biodegradation rate, maintenance
(H2S + NH3) mg 6.5–7.1 and the liquid quantity of flow rate of biomass, and disinfection adequacy of the BF effluent which hinders
(7 L min−1), it was an expensive one of the operating costs. Labor cost the process from being widely employed and established. Accordingly
(4380 €.year−1) was the same in both systems because the work cost and regarding technical results, the CS seems to be more efficient ap-
unit was equal. In summary, chemical usage in the CS and work cost in plicable odor control technology in treating off-gas of Yazd WWTP.
the BF were identified as a costly part in reactors, and total initial op-
erating costs were 4590.0744 €.year−1 and 17340.5 €.year−1 for the Acknowledgements
BF and CS, respectively. In conclusion, operating costs are an expensive
segment in both systems costs Table 5 (a & b). This claim has been The authors gratefully acknowledge Yazd water and wastewater
confirmed by previous studies (Estrada et al., 2012; Gao et al., 2001). department office for the financial support of the project as well as
Tehran University of Medical Science for complementary financial
3.4.3. Packing material replacement cost support.
Packing material replacement cost considers a price to remove old
bed materials, purchasing and installing the new one into the systems. Appendix A. Supplementary data
The durability of the bed for the CS and BF was estimated 10 and 2
years, respectively. Based on the findings, the packing material re- Supplementary data to this article can be found online at https://
placement costs were 25 € and 89 € for the BF and CS systems, re- doi.org/10.1016/j.jenvman.2019.04.003.
spectively. Considering the above, the main reason for the difference
between costs in this section can be due to packing bed components. References

3.4.4. Determining annualized costs in systems Lebrero, R., Bouchy, L., Stuetz, R., Muñoz, R., 2011. Odor assessment and management in
The annualized worth (AW) method is customarily utilized for wastewater treatment plants: a review. Crit. Rev. Environ. Sci. Technol. 41, 915–950.
Albright, L., 2009. Albright's Chemical Engineering Handbook. CRC Press, New York.
comparing alternatives. In this context, we estimated all annualized Alfonsín, C., Lebrero, R., Estrada, J.M., Muñoz, R., Kraakman, N.B., Feijoo, G., Moreira,
costs for all price section of rectors and finding shown in Table 6 (a & M.T., 2015. Selection of odour removal technologies in wastewater treatment plants:
b). Total annualized cost for CS in airflow rate of 250 L min−1 and EBRT a guideline based on Life Cycle Assessment. J. Environ. Manag. 149, 77–84.
Arellano-Garcia, L., González-Sánchez, A., Van Langenhove, H., Kumar, A., Revah, S.,
2.58 s is 112783.62 €.year−1, and for the BF it is 29980.2 €.year−1 in 2012. Removal of odorant dimethyl disulfide under alkaline and neutral conditions in
airflow rate of 36 L min−1 and EBRT 15 s. According to the results, the biotrickling filters. Water Sci. Technol. 66, 1641–1646.
annual costs related to the CS are dramatically higher than BF. Re- Charron, I., Feliers, C., Couvert, A., Laplanche, A., Patria, L., Requieme, B., 2004. Use of
hydrogen peroxide in scrubbing towers for odor removal in wastewater treatment
garding the volume of treated waste air by CS during a year and some plants. Water Sci. Technol. 50, 267–274.
operating costs (chemical usage), despite the shorter working time of Chen, L., Huang, J., Yang, C.L., 2001. Absorption of H2S in NaOCl caustic aqueous so-
the CS, this phenomenon is justifiable. The (€.m−3. h−1) was applied as lution. Environ. Prog. Sustain. Energy 20, 175–181.
Chen, C.-Y., Tsai, T.-H., Chang, C.-H., Tseng, C.-F., Lin, S.-Y., Chung, Y.-C., 2018. Airlift
an index for comparing two alternative systems with the economic
bioreactor system for simultaneous removal of hydrogen sulfide and ammonia from
aspect for removal of odor from WWTP. This index was obtained 1.58 synthetic and actual waste gases. J. Environ. Sci. Health - Part A Toxic/Hazard. Subst.
€.m−3. h−1 for the BF with 18921 m3 year−1 volume of treated air in Environ. Eng. 1–8.
EBRT 15 s and 2.57 €.m−3. h−1 for the CS with 43800 m3 year−1 vo- Cheng, H.-H., Hsieh, C.-C., 2010. Integration of chemical scrubber with sodium hypo-
chlorite and surfactant for removal of hydrocarbons in cooking oil fume. J. Hazard
lume of treated air in EBRT 2.85 s. Table 7 indicates the result of var- Mater. 182, 39–44.
ious researches for economic analysis of the BF and the CS in different Chung, Y.-C., Huang, C., Tseng, C.-P., Pan, J.R., 2000. Biotreatment of H2S-and NH3-
conditions. containing waste gases by co-immobilized cells biofilter. Chemosphere 41, 329–336.
Estrada, J.M., Kraakman, N.B., Muñoz, R.l., Lebrero, R., 2010. A comparative analysis of
odour treatment technologies in wastewater treatment plants. Environ. Sci. Technol.
4. Conclusions 45, 1100–1106.
Estrada, J.M., Kraakman, N.B., Lebrero, R., Muñoz, R., 2012. A sensitivity analysis of
process design parameters, commodity prices and robustness on the economics of
The techno-economic comparison of the CS and BF systems em- odour abatement technologies. Biotechnol. Adv. 30, 1354–1363.
ployed for treatment of the real waste air streams containing NH3 and Feilberg, A., Adamsen, A.P., Lindholst, S., Lyngbye, M., Schäfer, A., 2010. Evaluation of
H2S was conducted over a 45-day time period at Yazd wastewater biological air filters for livestock ventilation air by membrane inlet mass spectro-
metry. J. Environ. Qual. 39, 1085–1096.
treatment plant. The experimental results showed that the removal ef-
Fulazzaky, M.A., Talaiekhozani, A., Ponraj, M., Abd Majid, M., Hadibarata, T., Goli, A.,
ficiencies of NH3 and H2S in both systems were affected by their re- 2014. Biofiltration process as an ideal approach to remove pollutants from polluted
spective loading rate. Additionally, the NaOCl solution, especially with air. Desalin. Water Treat. 52, 3600–3615.
Gabriel, D., Deshusses, M.A., 2004. Technical and economical analysis of the conversion
the optimum flow rate of 7 L min−1 was found to give the best results in
of a full-scale scrubber to a biotrickling filter for odor control. Water Sci. Technol 50
term of removal efficiency among the various scrubbing liquids (NaOCl, (4), 309–318.
H2O2, and KMnO4) examined in the CS. The technical analysis clearly Gabriel, D., Deshusses, M.A., Gamisans, X., 2013. Desulfurization of biogas in biotrickling
demonstrated the effectiveness of both systems in treating actual waste filters. Air Pollution Prevention and Control: Bioreactors and Bioenergy 513–523.
Galera, M.M., Cho, E., Tuuguu, E., Park, S.-J., Lee, C., Chung, W.-J., 2008. Effects of
gases containing H2S and NH3. The CS is able to simultaneously treat of pollutant concentration ratio on the simultaneous removal of NH3, H2S and toluene
H2S and NH3 with efficiency above 97% for each compound at gases using rock wool-compost biofilter. J. Hazard Mater. 152, 624–631.
Qair = 250 L min−1 in the presence of NaOCl as an oxidant. However, Gao, L., Keener, T.C., Zhuang, L., Siddiqui, K.F., 2001. A technical and economic com-
parison of biofiltration and wet chemical oxidation (scrubbing) for odor control at
the BF had the robustness to remove a yield above 92% and 99.5% for wastewater treatment plants. Environ. Eng. Policy 2, 203–212.
H2S and NH3, respectively, at Qair = 36 L min−1 under suitable oper- González-Sánchez, A., Revah, S., Deshusses, M.A., 2008. Alkaline biofiltration of H2S
ating circumstances. Regarding the volume of treated contaminated air, odors. Environ. Sci. Technol. 42, 7398–7404.
Hou, J., Li, M., Xia, T., Hao, Y., Ding, J., Liu, D., Xi, B., Liu, H., 2016. Simultaneous
it is obvious that the performance of the CS is better than BF. The removal of ammonia and hydrogen sulfide gases using biofilter media from the
economic index (€.m−3. h−1) was calculated by the annualized worth biodehydration stage and curing stage of composting. Environ. Sci. Pollut. Res. 23,
method and utilized the economic assessment of both control systems. 20628–20636.
Hvitved-Jacobsen, T., Vollertsen, J., Tanaka, N., 2000. An integrated aerobic/anaerobic
The economic index for the CS and BF were estimated to be 2.57 and approach for prediction of sulfide formation in sewers. Water Sci. Technol. 41, 107.
1.58 €.m−3. h−1, respectively, indicating the cost-effectiveness of the Jaber, M.B., Anet, B., Amrane, A., Couriol, C., Lendormi, T., Le Cloirec, P., Cogny, G.,
bio-filtration system compared to the CS ones. The results also revealed Fillieres, R., 2014. Impact of nutrients supply and pH changes on the elimination of
hydrogen sulfide, dimethyl disulfide and ethanethiol by biofiltration. Chem. Eng. J.
that chemical usage and operating costs were expensive parts of the CS

42
 Journal of Cleaner Production 242 (2020) 118523

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Pistachio waste management using combined composting-

vermicomposting technique: Physico-chemical changes and worm
growth analysis
Abbas Esmaeili a, Mahdi Reyahi Khoram b, Majid Gholami c, *, Hadi Eslami d
a
Occupational Environment Research Center, Department of Environmental Health Engineering, School of Health, Rafsanjan University of Medical Sciences,
Rafsanjan, Iran
b
Department of Environmental, Hamadan Branch, Islamic Azad University, Hamadan, Iran
c
MSc in Environment and Environmental Pollution, Hamadan Branch, Islamic Azad University, Hamadan, Iran
d
Pistachio Safety Research Center, Department of Environmental Health Engineering, School of Health, Rafsanjan University of Medical Science, Rafsanjan,
Iran

a r t i c l e i n f o a b s t r a c t

Article history: In this study a combined composting-vermicomposting process with Eisenia fetida was investigated for
Received 2 February 2019 pistachio waste (PW) treatment mixed with cow dung (CD) in different ratio (100% CD (T1), 25% PW (T2),
Received in revised form 50% PW (T3), 75% PW (T4), and 100% PW (T5)) as a substrate in experimental scale. The process was
17 September 2019
performed in two stages, including pre-composting, and vermicomposting which operate in 45d for each
Accepted 20 September 2019
Available online 20 September 2019
stage in the pilot scale. Physico-chemical parameters and worm growth ratio in different treatments
were measured and compared. The results showed that the combined process significantly increased the
Handling editor: CT Lee quality of final product and the produced fertilizer was homogeneous, granular, odorless, and nutrient
rich. Physico-chemical analysis showed a significant reduction in the C:N ratio (41.13% in T2 to 69.14% in
Keywords: T5), total organic carbon (TOC) (12.26% in T2 to 37.06% in T5), and total potassium (K) (14.63% in T1 to
Composting 42.64% in T3) and increase in total nitrogen (N) (55.1% in T4 to 140.47% in T1), available phosphorus (P)
Cow dung (4.41% in T2 to 51.22% in T5), and pH (2.74% in T5 to 6.8% T3). The number of worms during the process has
Eisenia fetida also increased in all treatments and the net weight and growth rate of worms in T2 and T5 were higher
Vermicomposting
than other treatments. Therefore, combined Composting-vermicomposting process can be used as an
Pistachio waste
effective method for managing pistachio waste and CD.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction paper, fruits, and vegetables, produce a vast amount of solid waste
after product processing stages (Kumar et al., 2012). Conventional
Solid waste production is inevitable in any society and its rate waste disposal methods, such as sanitation, storage and burning
has increased with population growth and economic and social can have adverse effects on health, environment, and biodiversity
development (Ebrahimi et al., 2018; Eslami et al., 2018; Guerrero due to the production of toxic and dangerous gases. They can lead
et al., 2013). Around the world, it has been estimated that 1.3 to climate change and ultimately ecological balance disruption (Lim
billion metric tons of waste is annually generated, which will in- and Wu, 2016; Pe rez-Godínez et al., 2017).
crease to 38e67% by the end of 2025. It includes municipal, in- Recovery and reuse of agricultural waste has many economic,
dustrial, agricultural, and sewage sludge waste (García-Sa nchez environmental, and health advantages (Thiruneelakandan and
et al., 2017). Some agricultural products, such as sugar, dairy, Subbulakshmi, 2014). Biotechnology methods have many bene-
fits, including high speed, ease of control, low costs, high envi-
ronmental and health acceptability, and potential to be used at any
scale (Yadav and Garg, 2011a; Yadav et al., 2010). Composting and
* Corresponding author. Department of Environment, Islamic Azad University-
vermicomposting are biological processes, which can be used for
Hamadan Branch, Hamadan, Iran.
E-mail addresses: [email protected] (A. Esmaeili), [email protected] organic waste management (Adhikary, 2012; Kumar et al., 2012).
(M.R. Khoram), [email protected] (M. Gholami), Haddieslami1986@ Composting is a biodegradation process for organic wastes in
yahoo.com (H. Eslami).

https://doi.org/10.1016/j.jclepro.2019.118523
0959-6526/© 2019 Elsevier Ltd. All rights reserved.
2 A. Esmaeili et al. / Journal of Cleaner Production 242 (2020) 118523

aerobic or anaerobic conditions carried out by mesophilic and 2. Materials and methods
thermophilic microorganisms, in which organic wastes are con-
verted to mineral compounds, such as CO2, H2O, NHþ 4 , as well as 2.1. Providing the source of substrate
humus-rich organic compounds. They are suitable for improving
the physical properties of soil (Lim et al., 2016; Qian et al., 2014). The CD needed for the study was obtained from one of the farms
Vermicomposting process, similar to composting process, is a bio- in Rafsanjan (or the local villages) and stored in plastic containers at
logical process in which the nutrients in organic waste will turn ambient temperature. The PW includes the soft and colored shell of
into a stable, available, and nutritious product for plants growth the pistachios (epicarp) as well as clusters that pistachios are
(Rupani et al., 2017b). In vermicomposting process, biological attached to them. This waste is separated from the pistachio after
oxidation and stabilization of organic waste depend on the inter- using the peeling machine (HL-1117, HONGLE, China) under 3 cm
action between microorganisms and earthworms in mesophilic and was collected from the pistachio processing industry in Raf-
process (Lazcano et al., 2008; Oluseyi et al., 2016). In fact, worm's sanjan, Iran. Then, the PW was dried in ambient temperature and
activity increases the level of organic waste for consumption of stored for the experiments. The physicochemical characteristics of
microorganisms. The microbial activity and organic compounds CD and PW are presented in Table 1. Eisenia fetida earthworms, with
decomposition are improved and it changes the physical, chemical, an average weight of 100±5 g for 500 worms, were obtained from
and biological properties of the waste and turns them into vermi- vermicomposting groups Company, Iran.
compost (Lim et al., 2015; Pigatin et al., 2016). Vermicompost is a
stabilized and homogeneous material containing high nutrients
2.2. Composting - vermicomposting process
and low toxic materials which has many advantages, such as cation
exchange capacity (CEC), water holding capacity, suitable particle
Two phases of combined process could be clarified as 1. Initial
size distribution, and a high concentration of nutrients (Adi and
digestion or pre-composting, and 2. Secondary digestion or ver-
Noor, 2009; Punde and Ganorkar, 2012). Grading uniformity of
micomposting and harvesting from the final product and biomass.
vermicompost and being odorless are the highlights of this fertil-
In the initial digestion phase, windrows of 0.6  0.6  1 m from the
izer (Adhikary, 2012; Chattopadhyay, 2012).
PW and CD were separately given moisture (water spray) and
In recent decades, vermicomposting process has been used to
suitable aeration (stirring). These substances were allowed to
stabilize various agricultural wastes such as Potato peels, Bagasse,
decompose (preliminary) for 45d to become more homogeneous
Sago waste, Sericulture waste, Onion residue/waste and Vegetable
and their toxic compounds (phenol and cyclic compounds and
waste (Sobana et al., 2016). The advantage of combining com-
substances) decreased, and became edible for worms (Kumar et al.,
posting with vermicomposting process is that in composting pro-
2012; Mainoo et al., 2009). In the secondary digestion phase or
cess, a wide range of organic and resistant to decomposition wastes,
vermicomposting, 5 treatments (T1-T5) with 3 replicates were
such as phenol are decomposed without affecting the process ef-
placed in plastic baskets for 45d. Each treatment contained 10 kg of
ficiency. Combining composting with vermicomposting process is
raw compost material including the CD alone and in mixture with
suitable for biodegradation of solid waste with high salinity or low
different ratios (25%, 50% and 75%) of the PW, and PW alone as
pH, such as pistachio skin wastes (Lim et al., 2016; Sa nchez-
presented in Table 2 (Yadav and Garg, 2011b).
Monedero et al., 2010).
To stabilize the heat, allow microbial decomposition, and reduce
The main agricultural product of Rafsanjan, Iran, is pistachio.
odor and toxic compounds of the treatments, they were put aside
The annual average production of dry pistachio and dry waste is
for 5d. Then, 500 worms (with an average weight of 100 ± 5 g) were
250 and 132e165 thousand tons. For every 3 kg of pistachio im-
added to each experimental chamber with total volume of 11 L
ported to pistachio processing industries, 1 kg of dried pistachio
(37  26  12 cm, length, width and height). The treatments were
and 2 kg of waste is produced; two-thirds of which is wastewater,
kept in darkness (away from direct light) and at ambient temper-
and nearly 660 g of pure dry substance is obtained from this pis-
ature (25-30  C). To maintain a humidity of 60%e70% in the
tachio waste (PW) (Hosseinifard et al., 2010; Mehrnejad, 2001).
experimental chamber, water was sprayed on them. To prevent loss
Every year during the harvest season, a tremendous amount of
of moisture and keep away pests, the culture dishes were covered
waste is produced during the peeling stage and disposed through
with wet hemp fibers.
inappropriate methods (they are either burned or discarded into
the environment) leading to environmental pollution. Failure to
properly manage pistachio waste can release toxic compounds, 2.3. Worm growth analysis
such as phenol into the soil and environment as well as annoying
odors (Jalili et al., 2019). Pistachio processing waste can be used as a A week before the end of the experiments, water sprinkling on
bulking agent in composting or vermicomposting process in com- the bed of worms (treatment) was stopped. The earthworms and
bination with animal wastes (Demirer, 2016; Malakootian et al., cocoons were separated easily from the product. In the third stage,
2014). the worms and cocoons were manually separated and counted.
Considering the large volume of pistachio waste production and Worms were washed by water and the total worms in each
the need for proper management and transformation of this waste
into a reusable and eco - friendly product as well as converting PW
Table 1
to an edible food for worms in vermicomposting process, this study Initial physico-chemical characteristics of CD and PW.
aimed to investigate the new combined compositing-
Parameter CD PW
vermicomposting process for managing of PW in combination
with animal manure in experimental scale. Effects of vermicompost pH 7.7 ± 0.01 8.1 ± 0.01
TOC (%) 24.8 ± 0.01 45.7 ± 0.01
on physicochemical changes of the produced fertilizer, as well as
K (%) 0.3 ± 0.01 0.5 ± 0.01
the rate of growth and multiplication of worms and their produced N (%) 0.4 ± 0.01 0.7 ± 0.01
biomass during the process were also evaluated. In this study, P (%) 0.5 ± 0.01 0.4 ± 0.01
earthworm species of Eisenia fetida were used to assess vermi- C:N ratio 59.1 ± 2.13 65.1 ± 1.87
compost under appropriate circumstances and with two substrates, TOC: Total organic carbon; K: Total potassium.
including PW and CD. N: Total nitrogen; P: Available phosphorus.
 A. Esmaeili et al. / Journal of Cleaner Production 242 (2020) 118523 3

Table 2 rez-Godínez et al., 2017). pH is one of the most important pa-

Pe
Composition of CD and PW in different treatments or design of treatments. rameters in determining the quality of the compost and vermi-
Treatment No. Description CD (kg) PW (kg) compost. In different studies, the best pH range for the final product
T1 CD (100%) 10.0 0.0
applied to most plants was between 6 and 8.5 (Campitelli and
T2 CD (75%) þ PW (25%) 7.5 2.5 Ceppi, 2008). The final products of combined composting-
T3 CD (50%) þ PW (50%) 5.0 5.0 vermicomposting process in this study were in an appropriate pH
T4 CD (25%) þ PW (75%) 2.5 7.5 range in all the treatments and were partially in alkaline range. pH
T5 PW (100%) 0.0 10.0
values in neutral and relatively alkaline range indicate the stabili-
zation of the final products (Pe rez-Godínez et al., 2017). Also, this
final product is suitable for plants that grow in alkaline conditions
treatment were weighed. The growth rate of the worms was such as maize, sugarcane, sorghum, and agave plants (Estrada-de
calculated based on the initial and final weight changes as shown in los Santos et al., 2012).
Eq. (1).

Initial weight before vermicompossting  Final weight of vermicompost

Growth rate ¼ (1)
Total vermicomposting time

The final product was passed through an iron sieve with holes of 3.2. Total organic carbon changes
1  6 mm and homogeneous samples of the vermicompost were
obtained for physico-chemical analysis. According to Fig. 2, TOC in the final product decreased compared
to initial treatments (substrate). The TOC reduction in different
2.4. Physico-chemical analysis treatments were (12.26%) T2 < (12.97%) T3 < (14.37%) T1 < (24.97%)
T4 < (37.06%) T5 and these reductions were statistically significant
Homogeneous samples (without earthworms and cocoon) were in all treatment (p  0.001) as presented in Table 3. The PW content
obtained from each treatment based on dry weight, before com- in T4 and T5 were higher than other treatments in the initial sub-
posting and after 14 week (end of the experiment), for analysis of strate. The more carbon available for microorganisms and worms
the parameters including pH, total nitrogen (N), total organic car- lead to an increase in the decomposition rate during the
bon (TOC), available phosphorus (P), and total potassium (K). composting-vermicomposting process (Singh and Kalamdhad,
Fifteen samples were collected and analyzed from 5 types of
treatments (each treatment with 3 replications) and the results
were reported as mean. The pH was measured by using aqueous
extract 1:10 (sample: distilled water v/w) and pH meter (SG2
model, Metter Toledo, Switzerland). To determine TOC, wet di-
chromate oxidation method was used (Cucina et al., 2018). N and P
were measured using Kjeldahl method and Olsen method with
spectrophotometer (6330 model, Jenway, England). K was extrac-
ted with amino acetate and measured using a photoelectric flame
photometer (Jalili et al., 2019).

2.5. Statistical analysis

To compare the mean value of Physico-chemical parameters in

initial substrate before the composting-vermicomposting process
and final products was used paired t-test statistical analysis in SPSS
version 18.0 software program. Fig. 1. pH changes in different treatment of pistachio waste (PW)þ cow dung (CD)
(T1:100% CD, T2: 75% CD and 25% PW, T3: 50% CD and 50% PW, T4: 25% CD and 75% PW,
T5: 100% PW).
3. Results and discussion

3.1. pH changes Table 3

Statistical analysis between Physico-chemical parameters measured in the initial
The initial pH of different treatments ranged from 7.65 ± 0.1 to substrate and the final product in different treatment (Significant level was at
P  0.05).
8.03 ± 0.08, and in the final vermicompost ranged from 7.98 ± 0.07
to 8.32 ± 0.01 as shown in Fig. 1 pH changes, in the combined Parameter P-value
composting-vermicomposting process, before and after the process T1 T2 T3 T4 T5
were incremental and varied from 2.74% in T5 to 6.8% in T3. Table 3
pH 0.218 0.070 0.078 0.094 0.161
indicated that these changes are not statistically significant in all K 0.088 0.090 0.009 0.038 0.094
the treatments (p > 0.05). This increase in pH can be related to the N 0.079 0.096 0.079 0.062 0.059
decomposition of organic acids and volatile fatty acids, as well as P 0.007 0.485 <0.001 0.353 0.027
the mineralization of organic nitrogen to ammonium during the TOC <0.001 <0.001 <0.001 <0.001 <0.001
C:N <0.001 <0.001 <0.001 <0.001 <0.001
composting-vermicomposting process (García-Sa nchez et al., 2017;
4 A. Esmaeili et al. / Journal of Cleaner Production 242 (2020) 118523

Fig. 2. Initial and final total organic carbon content in different treatment of pistachio Fig. 3. Initial and final total nitrogen content in different treatment of pistachio waste
waste (PW) þ cow dung (CD) (T1:100% CD, T2: 75% CD and 25% PW, T3: 50% CD and 50% (PW) þ cow dung (CD) (T1:100% CD, T2: 75% CD and 25% PW, T3: 50% CD and 50% PW,
PW, T4: 25% CD and 75% PW, T5: 100% PW). T4: 25% CD and 75% PW, T5: 100% PW).

2012). Decrease in TOC during composting-vermicomposting pro- vermicomposting process (Fatehi and Shayegan, 2010; Suthar and
cess in this study was coincided with similar studies (Fatehi and Singh, 2008). Given that in composting-vermicomposting process
Shayegan, 2010; Garg et al., 2006; Jadia and Fulekar, 2008). the mineralization rate of organic compounds was higher, the
Reduction in the amount of organic carbon may be due to the joint amount of N was higher than vermicomposting studies (Garg et al.,
activity of worms and microorganisms, in which organic materials 2012; Khatua et al., 2018).
are converted to carbon dioxide (Sharma and Garg, 2018; Yadav and
Garg, 2011a). Earthworms release mucus and enzymes that in- 3.4. C:N ratio changes
crease the activity of microorganisms; moreover, microorganisms
release the extracellular enzyme into the earthworm intestine Fig. 4 shows that the initial C:N ratio before composting-
rez-Godínez et al., 2017). Due to the pre-composting process,
(Pe vermicomposting process in different treatments ranges from
the amount of organic carbon reduction is higher than vermi- 37.26 to 65.1 and in the final product ranges from 18.02 to 22.8. This
composting process, which can be due to the breakdown and rate indicated an overall reduction of 41.13% in T2 to 69.14% in T5 at
decomposition of organic compounds by bacteria in composting the end of composting-vermicomposting process from the initial
process, as well as their easier consumption by worms in vermi- value. The reduction of C/N ratio was statistically significant in all
composting process (Lazcano et al., 2008; Soobhany et al., 2017). treatments (p  0.001) as presented in Table 3. Decrease in C:N
Consuming some of organic compounds by worms and converting represented the high level of mineralization and stabilization of the
them into biomass, the amount of primary raw materials decreased organic waste during composting-vermicomposting process
(Khatua et al., 2018). Research studies have shown that earth- (Alidadi et al., 2016; Jadia and Fulekar, 2008; Muthukumaravel
worms, in vermicompost process, can more efficiently use and et al., 2008). This high rate of C:N ratio reduction in the final
decompose organic compounds with longer and more complex product could be due to the carbon loss as CO2 through microbial
carbon chains than composting process. Therefore, it improves respiration and conversion to biomass by absorption process.
organic compounds decomposition and significantly reduces Simultaneous with increase in population of earthworms, the
organic carbon in composting-vermicomposting process (Ali et al., addition of N content in the form of mucus, growth hormones and
2015; Soobhany et al., 2017). enzymes, and fecal material by the earthworms was increased and
the C:N ratio in the final product was reduced (Lazcano et al., 2008;
Soobhany et al., 2017). The C/N ratio lower than 20 represents the
3.3. Total nitrogen changes

The initial and final N content after composting-

vermicomposting process is shown in Fig. 3. N content in the
final product increased from 55.1% in T4 to 140.47% in T1. Table 3
reveals that these changes are not statistically significant in all
treatments (p > 0.05). The result of this study shown that with
increasing the amount of CD in initial substrate, N content in final
products increased. The amount of N in the final product is affected
by the initial N content in the raw material and the rate of
decomposition (Pe rez-Godínez et al., 2017). The initial N content in
raw materials in T4 and T1 was respectively higher and lower than
other treatments. Therefore, N content in final product in T4 was
higher and in T1 was lower than other treatments. Increase of N
content in composting process can be due to mineralization process
of organic compounds containing protein by microorganism (Garg
et al., 2006). The activity of nitrogen-fixing bacteria can increase
the N content in the final product (Sharma et al., 2017). Organics Fig. 4. Initial and final C/N ratio in different treatment of pistachio waste (PW) þ cow
matter decomposition by E. fetida accelerated the mineralization dung (CD) (T1:100% CD, T2: 75% CD and 25% PW, T3: 50% CD and 50% PW, T4: 25% CD
process, and subsequently, increased N content in and 75% PW, T5: 100% PW).
 A. Esmaeili et al. / Journal of Cleaner Production 242 (2020) 118523 5

decomposition of plant materials by worms (Suthar, 2007). During

vermicomposting process, earthworms with the help of the activity
of microorganisms, especially alkaline bacteria in their digestive
tract, dissolve the insoluble phosphorus and increase accessibility
(Alidadi et al., 2016; Raphael and Velmourougane, 2011; Yadav and
Garg, 2011b). Mineralization of Phosphorous during vermi-
composting process by Phosphatase secretion using earthworms
can increase the P content in the final product (Raphael and
Velmourougane, 2011). The results regarding the increase of P
content in this study were consistent with that of the other studies
(Lim and Wu, 2016; Sharma and Garg, 2018). It is indicative of the
positive impact of earthworms on the nutrient content which is
essential for plant growth.

3.6. Total potassium changes

Fig. 5. Initial and final available phosphorus content in different treatment of pistachio
waste (PW) þ cow dung (CD) (T1:100% CD, T2: 75% CD and 25% PW, T3: 50% CD and 50%
PW, T4: 25% CD and 75% PW, T5: 100% PW). The K in the initial mixture and the final product of combined
composting-vermicomposting process is shown in Fig. 6. The
amount of K in the final product was lower than the initial level in
all the treatments (a reduction between 14.63% in T1 and 42.64% in
T3). Table 3 shows that the reduction of K in T3 and T4 are statisti-
cally significant (p  0.05). The reduction of K could be due to
washing of these elements by water during watering of the treat-
ments. (Sharma et al., 2011). Reduction of K concentration during
composting and vermicomposting process can be due to the
enzymatic activity and secretion of enzymes in the earthworm
digestive system, such as phosphatase (Alavi et al., 2017). Some
researchers have observed a similar process in K concentration for a
variety of organic waste during vermicomposting process (Rupani
et al., 2017a).

3.7. Worm growth and fecundity

Growth rate, number of worms (E. fetida), and cocoons in ver-

Fig. 6. Initial and final total potassium content in different treatment of pistachio
micompost process in different treatments are shown in Table 4.
waste (PW) þ cow dung (CD) (T1:100% CD, T2: 75% CD and 25% PW, T3: 50% CD and 50%
PW, T4: 25% CD and 75% PW, T5: 100% PW). The results reveal that, not only mortality did not occur during the
process, but also the number of worms during the process
increased. This increase in T2 and T3, containing 75% and 50% of the
maturity and stabilization of the organic materials (Pe rez-Godínez CD was higher than the other treatments, and in T5, which did not
et al., 2017). The C/N ratio in all treatment of this study were lower have CD, the number of worms was the lowest. This can be due to
than 22.8 and it shows that organic matter stabilization in the final the presence of nutrients for the growth of worms in the CD,
product is well done. making T2 a favorite feed for the earthworms (Pe rez-Godínez et al.,
2017). The net weight and growth rate of worms in T2 was higher
3.5. Available phosphorous changes than the other treatments. The number of cocoons produced by
worms (E. fetida) in T5, containing only PW, was higher and the
The amount of P in the initial mixture and final product after number of worms was lower than the other treatments. Due to
composting-vermicomposting process is shown in Fig. 5. P content some harmful compounds in the PW, such as phenol, Cocoons were
of the final product was higher than the initial substrate, and it not able to grow well and mature (Sonmezdag et al., 2017; Yadav
increased from 4.41% in T2 to 51.22% in T5. The increase of P content and Garg, 2011b).
in T1, T3 and T5 were statistically significant (p  0.001) (Table 3).
This increase in P content during the composting process can be 4. Conclusion
due to loss of organic material (Sharma et al., 2017) and in vermi-
composting process can be due to the breakdown and The results of the present study showed that all C:N ratio, TOC,

Table 4
Growth and fertility of Eisenia fetida on treatments during vermicomposting in different treatment.

Treatment Cocoon Mean initial weight (g) Mean final weight (g) Net weight gained (g) Growth rate Total population (number of worms)

T1 70 ± 15 102 ± 50 206 ± 11 104 2.3 565 ± 53

T2 73 ± 21 104 ± 30 258 ± 24 154 3.4 643 ± 75
T3 177 ± 25 97 ± 31 183 ± 13 86 1.9 641 ± 74
T4 52 ± 15 105 ± 62 183 ± 13 81 1.7 561 ± 40
T5 185 ± 22 95 ± 41 217 ± 15 113 2.7 516 ± 50

Note: Mean value of three replicates ± S.E.

Net weight gained/total process time.
6 A. Esmaeili et al. / Journal of Cleaner Production 242 (2020) 118523

and K decreased and N, P, and pH increased. The final product after Dynamics of organic matter decomposition during vermicomposting of banana
stem waste using Eisenia fetida. Waste Manag. 79, 287e295.
composting-vemicomposting process was nutrient rich. Moreover,
Kumar, V.V., Shanmugaprakash, M., Aravind, J., Namasivayam, S.K.R., 2012. Pilot-
the number of worms during the process increased in all the scale study of efficient vermicomposting of agro-industrial wastes. Environ.
treatments and the net weight and growth rate of worms in T2 Technol. 33, 975e981.
Lazcano, C., Go  mez-Brando  n, M., Domínguez, J., 2008. Comparison of the effec-
(containing 75% CD and 25% PW) and T5 (containing 100% PW) was
tiveness of composting and vermicomposting for the biological stabilization of
higher than the other treatments. Therefore, the combination of cattle manure. Chemosphere 72, 1013e1019.
composting-vermicomposting process improved the quality of the Lim, S.L., Lee, L.H., Wu, T.Y., 2016. Sustainability of using composting and vermi-
final production and can be used as an effective method for PW composting technologies for organic solid waste biotransformation: recent
overview, greenhouse gases emissions and economic analysis. J. Clean. Prod.
treatment with CD in the ratio of 1:3. 111, 262e278.
Lim, S.L., Wu, T.Y., 2016. Characterization of matured vermicompost derived from
valorization of palm oil mill byproduct. J. Agric. Food Chem. 64, 1761e1769.
Acknowledgements https://doi.org/10.1021/acs.jafc.6b00531.
Lim, S.L., Wu, T.Y., Lim, P.N., Shak, K.P.Y., 2015. The use of vermicompost in organic
Thanks are owed to Hamedan Azad University and Rafsanjan farming: overview, effects on soil and economics. J. Sci. Food Agric. 95,
1143e1156.
University of Medical Sciences for their financial supports in this Mainoo, N.O., Barrington, S., Whalen, J.K., Sampedro, L., 2009. Pilot-scale vermi-
study. composting of pineapple wastes with earthworms native to Accra, Ghana.
Bioresour. Technol. 100, 5872e5875.
Malakootian, M., Mobini, M., Nekoonam, G.A., 2014. Evaluation of the compost
References produced from mixed sludge of municipal wastewater treatment plant and
pistachio hull waste. J Mazand Univ Med Sci 24, 172e183. Available online at:
Adhikary, S., 2012. Vermicompost, the story of organic gold: a review. Agric. Sci. 3, http://jmums.mazums.ac.ir/article-1-4326-en.html.
905e917. Mehrnejad, M., 2001. The current status of pistachio pests in Iran. Cah. Options
Adi, A., Noor, Z., 2009. Waste recycling: utilization of coffee grounds and kitchen Mediterr. 56, 315e322. Available online at: http://om.ciheam.org/article.php?
waste in vermicomposting. Bioresour. Technol. 100, 1027e1030. IDPDF¼1600196.
Alavi, N., Daneshpajou, M., Shirmardi, M., Goudarzi, G., Neisi, A., Babaei, A.A., 2017. Muthukumaravel, K., Amsath, A., Sukumaran, M., 2008. Vermicomposting of
Investigating the efficiency of co-composting and vermicomposting of vinasse vegetable wastes using cow dung. J. Chem. 5, 810e813.
with the mixture of cow manure wastes, bagasse, and natural zeolite. Waste Oluseyi, E.E., Ewemoje, T., Adedeji, A., 2016. Comparative analysis of pit composting
Manag. 69, 117e126. and vermicomposting in a tropical environment. Agric. Biosyst. Eng. 731, 39757.
Ali, U., Sajid, N., Khalid, A., Riaz, L., Rabbani, M.M., Syed, J.H., Malik, R.N., 2015. rez-Godínez, E.A., Lagunes-Zarate, J., Corona-Herna
Pe ndez, J., Barajas-Aceves, M.,
A review on vermicomposting of organic wastes. Environ. Prog. Sustain. Energy 2017. Growth and reproductive potential of Eisenia foetida (Sav) on various zoo
34, 1050e1062. animal dungs after two methods of pre-composting followed by vermi-
Alidadi, H., Hosseinzadeh, A., Najafpoor, A.A., Esmaili, H., Zanganeh, J., Takabi, M.D., composting. Waste Manag. 64, 67e78.
Piranloo, F.G., 2016. Waste recycling by vermicomposting: maturity and quality Pigatin, L.B.F., Atoloye, I.A., Obikoya, O.A., Borsato, A.V., Rezende, M.O.O., 2016.
assessment via dehydrogenase enzyme activity, lignin, water soluble carbon, Chemical study of vermicomposted agroindustrial wastes. Int. J. Recycl. Org.
nitrogen, phosphorous and other indicators. J. Environ. Manag. 182, 134e140. Waste Agric. 5, 55e63.
Campitelli, P., Ceppi, S., 2008. Chemical, physical and biological compost and ver- Punde, B., Ganorkar, R., 2012. Vermicomposting-recycling waste into valuable
micompost characterization: a chemometric study. Chemometr. Intell. Lab. Syst. organic fertilizer. Int. J. Eng. Res. Afr. 2, 2342e2347.
90, 64e71. Qian, X., Shen, G., Wang, Z., Guo, C., Liu, Y., Lei, Z., Zhang, Z., 2014. Co-composting of
Chattopadhyay, G.N., 2012. Use of vermicomposting biotechnology for recycling livestock manure with rice straw: characterization and establishment of
organic wastes in agriculture. Int. J. Recycl. Org. Waste Agric. 1, 8. https://doi. maturity evaluation system. Waste Manag. 34, 530e535.
org/10.1186/2251-7715-1-8. Raphael, K., Velmourougane, K., 2011. Chemical and microbiological changes during
Cucina, M., Tacconi, C., Sordi, S., Pezzolla, D., Gigliotti, G., Zadra, C., 2018. Valoriza- vermicomposting of coffee pulp using exotic (Eudrilus eugeniae) and native
tion of a pharmaceutical organic sludge through different composting treat- earthworm (Perionyx ceylanesis) species. Biodegradation 22, 497e507.
ments. Waste Manag. 74, 203e212. Rupani, P.F., Embrandiri, A., Ibrahim, M.H., Shahadat, M., Hansen, S.B., Ismail, S.A.,
Demirer, G., 2016. Biogas production from pistachio (pistaciavera L.) de-hulling Kadir, M.O.A., 2017a. Recycling of palm oil industrial wastes using vermi-
waste. Int. J. Green Energy 13, 1320e1324. composting technology: its kinetics study and environmental application. En-
Ebrahimi, A., Hashemi, H., Eslami, H., Fallahzadeh, R.A., Khosravi, R., Askari, R., viron. Sci. Pollut. Res. 24, 12982e12990.
Ghahramani, E., 2018. Kinetics of biogas production and chemical oxygen de- Rupani, P.F., Embrandiri, A., Ibrahim, M.H., Shahadat, M., Hansen, S.B.,
mand removal from compost leachate in an anaerobic migrating blanket Mansor, N.N.A., 2017b. Bioremediation of palm industry wastes using vermi-
reactor. J. Environ. Manag. 206, 707e714. composting technology: its environmental application as green fertilizer. 3
Eslami, H., Ehrampoush, M.H., Esmaeili, A., Ebrahimi, A.A., Salmani, M.H., Biotech 7, 155.
Ghaneian, M.T., Falahzadeh, H., 2018. Efficient photocatalytic oxidation of nchez-Monedero, M.A., Serramia
Sa , N., Civantos, C.G.-O., Ferna ndez-Hern
andez, A.,
arsenite from contaminated water by Fe2O3-Mn2O3 nanocomposite under UVA Roig, A., 2010. Greenhouse gas emissions during composting of two-phase olive
radiation and process optimization with experimental design. Chemosphere mill wastes with different agroindustrial by-products. Chemosphere 81, 18e25.
207, 303e312. Sharma, D., Katnoria, J.K., Vig, A.P., 2011. Chemical changes of spinach waste during
Estrada-de los Santos, P., Martínez-Aguilar, L., Lopez-Lara, I.M., Caballero-Mellado, J., composting and vermicomposting. Afr. J. Biotechnol. 10, 3124e3127. https://doi.
2012. Cupriavidus alkaliphilus sp. nov., a new species associated with agricul- org/10.5897/AJB10.046.
tural plants that grow in alkaline soils. Syst. Appl. Microbiol. 35, 310e314. Sharma, D., Varma, V.S., Yadav, K.D., Kalamdhad, A.S., 2017. Evolution of chemical
Fatehi, M., Shayegan, J., 2010. Vermicomposting of organic solid waste with the E. and biological characterization during agitated pile composting of flower waste.
Fetida in different bedding materials. J. Environ. Stud. 36, 10e12. Int. J. Recycl. Org. Waste Agric. 6 (1), 89e98.
García-Sa nchez, M., Tausnerova , H., Hanc, A., Tlustos, P., 2017. Stabilization of Sharma, K., Garg, V., 2018. Comparative analysis of vermicompost quality produced
different starting materials through vermicomposting in a continuous-feeding from rice straw and paper waste employing earthworm Eisenia fetida (Sav.).
system: changes in chemical and biological parameters. Waste Manag. 62, Bioresour. Technol. 250, 708e715.
33e42. Singh, J., Kalamdhad, A.S., 2012. Concentration and speciation of heavy metals
Garg, P., Gupta, A., Satya, S., 2006. Vermicomposting of different types of waste during water hyacinth composting. Bioresour. Technol. 124, 169e179.
using Eisenia foetida: a comparative study. Bioresour. Technol. 97, 391e395. Sobana, K., Agnes Sharmila, M., Jegadeesan, M., 2016. A review on vermicomposting
Garg, V.K., Suthar, S., Yadav, A., 2012. Management of food industry waste of bio waste using different earthworm species. J Environ Sci, Comp Sci Eng
employing vermicomposting technology. Bioresour. Technol. 126, 437e443. Technol Sec. A. 5, 193e200. Available online at: http://www.jecet.org/paper_
Guerrero, L.A., Maas, G., Hogland, W., 2013. Solid waste management challenges for description.php?category¼Env%20Science&volume¼5&issue¼1.
cities in developing countries. Waste Manag. 33, 220e232. Sonmezdag, A.S., Kelebek, H., Selli, S., 2017. Characterization and comparative
Hosseinifard, S.J., Khademi, H., Kalbasi, M., 2010. Different forms of soil potassium as evaluation of volatile, phenolic and antioxidant properties of pistachio (Pistacia
affected by the age of pistachio (Pistacia vera L.) trees in Rafsanjan, Iran. Geo- vera L.) hull. J. Essent. Oil Res. 29, 262e270.
derma 155, 289e297. Soobhany, N., Gunasee, S., Rago, Y.P., Joyram, H., Raghoo, P., Mohee, R., Garg, V.K.,
Jadia, C.D., Fulekar, M., 2008. Vermicomposting of vegetable waste: a bio- 2017. Spectroscopic, thermogravimetric and structural characterization ana-
physicochemical process based on hydro-operating bioreactor. Afr. J. Biotechnol. lyses for comparing Municipal Solid Waste composts and vermicomposts sta-
7, 3723e3730. Available online at: http://www.academicjournals.org/AJB. bility and maturity. Bioresour. Technol. 236, 11e19.
Jalili, M., Mokhtari, M., Eslami, H., Abbasi, F., Ghanbari, R., Ebrahimi, A.A., 2019. Suthar, S., 2007. Vermicomposting potential of Perionyx sansibaricus (Perrier) in
Toxicity evaluation and management of co-composting pistachio wastes com- different waste materials. Bioresour. Technol. 98, 1231e1237.
bined with cattle manure and municipal sewage sludge. Ecotoxicol. Environ. Suthar, S., Singh, S., 2008. Vermicomposting of domestic waste by using two epigeic
Saf. 171, 798e804. earthworms (Perionyx excavatus and Perionyx sansibaricus). Int. J. Environ. Sci.
Khatua, C., Sengupta, S., Balla, V.K., Kundu, B., Chakraborti, A., Tripathi, S., 2018. Technol. 5, 99e106.