~ Pergamon War. se: Tech. Vol. 32, No.8. pp. 85-90.1995.

Copyright C 19961AWQ
Printed in Great Bnlalll. All rights reserved,
0273-1223195 59 'SO + 000
pn:S0273-l223(96)OOO n-x

AGGREGATION OF
SULPHATE-REDUCING BACTERIA AND
HOMO-ACETOGENIC BACTERIA IN A
LAB-SCALE GAS-LIFT REACTOR

R. T. van Houten*, A. C. van Aelst** and G. Lettinga"

* Departmentof Enviromental Technology. Wageningen Agricultural University,
Bomenweg 2. 6703 HD Wageningen. The Netherlands
** Departmentof Plant Cytologyand Morphology,
Wageningen Agricultural University, Arboretumlaan 4, 6703 BD Wageningen.
The Netherlands
ABSTRACT

This paper discusses aspects of the formation, growth and detachment of sulphate-reducing aggre•
gates, growing on Hz/COz and Hz/CO gas mixtures in a 4.5 L lab-scale gas-lift reactor. The biomass
in the aggregates consisted predominantly of Desulfovibrio sp. and Acetobacterium sp.
Our experiments with H/COz showed that aggregates with pumice particles, i.e, biofilms, as well as
carrier-free aggregates were formed. At pH 7.0, all the aggregates had a very hairy structure. This
structure, however, was pH dependent. At lower pH-levels the aggregates became smooth. At pH•
values higher than 7.0 the hairy structure became even more irregular. The predominant microor•
ganisms were rather heterogeneously distributed throughout the aggregates at all pH-values.
Growth on gas-mixtures of Hz/CO led to the formation of smooth aggregates. Presence of CO also
led to formation of a layered biomass structure in which the Acetobacterium sp. were more at the
outside of the aggregates and the Desulfovlbrio sp. were located at the inner part of the aggregates.
The observed changes in the surface structure of the sulphate-reducing aggregates are related to the
growth rate of the microorganisms present. A high growth rate yields a rough and hairy aggregate
surface. A low growth rate yields a smooth aggregate surface, since the low growth rate of the
microorganisms allows detachment of the large protrusions from the aggregate surface.
The surface structure is not dependent on the type of bacteria present at the surface, i.e. sulphate•
reducing bacteria or homo-acetogenic bacteria.

KEYWORDS

Sulphate-reducing bacteria; aggregation; immobilization; biofiIm; gas-lift reactor; detachment.

INTRODUCTION

Many industries, particularly chemical, metallurgical and rmnmg industries, produce wastewater
containing sulphate or sulphuric acid. Most, if not all, of this wastewater is discharged to local
sewage systems or treated with lime resulting in high amounts of solid gypsum waste. Because of the
negative environmental impact, however, both options have to be reconsidered.
85
86 R. T. van HOUTEN tt al:
There are essentially only two alternatives to the current situation: either reduction of the discharged
amount of sulphate at the source or recovery of sulphur from the wastewater, which can be reused in
several chemical processes. Since the first alternative is not always a short term matter - e.g. in case
of mine leachate of already dumped mine waste - the second alternative is quite viable.
A two step biological treatment is such a viable alternative. In this biological process, sulphate-reduc•
ing bacteria first convert the sulphate into, hydrogen sulphide. After that, colorless sulphur bacteria
convert the produced hydrogen sulphide into elemental sulphur. Our research is focused at the first
biological step.

Because the previously mentioned wastewater hardly contains organic compounds, biological treat•
ment is only possible when an electron donor and carbon source are added. We have chosen to use
synthesis gas, i.e., a gas mixture of H/CO/C02 , as the electron donor and carbon source.
Now, there are three boundary conditions for reactor design. First, the reactor should contain a large
amount of active biomass. Thus, immobilization of the biomass is necessary. Second, the system
consists of three phases, i.e., gas-liquid-solid. Besides, the reactor should be well mixed, since acidic
wastewater needs to be treated. Therefore, we have chosen to use a gas-lift reactor.

Obviously, the phenomenon of biomass retention is very important to the operation of such gas-lift
reactors. Although in recent years large improvements have been made (Heijnen, 1984; Beeftink,
1987; Tijhuis, 1994), several aspects of biomass retention are still poorly understood. So, further
research on this topic is necessary. In our case it is of particular importance, because the ability of
sulphate-reducing bacteria to form biofilms or aggregates in turbulent three phase systems has not
been described before. Thus, in our laboratory research has been started at the aggregation of
sulphate-reducing bacteria in gas-lift reactors fed with synthesis gas (van Houten et al., 1994; van
Houten et aJ., 1995a; van Houten et aJ., 1995b).

The paper presented here discusses some aspects of the formation, growth and detachment of sul•
phate-reducing aggregates, growing on H2/C02 and H/CO gas mixtures. Particularly, the influence of
growth rate on aggregation and aggregate appearance is reviewed.

RESULTS

The biofilm formation capacity of sulphate-reducing bacteria was tested with use of pumice as the
carrier material. Main reason for using this material was that carrier particles with a rough surface
provide better conditions for attachment of biomass (Mulder and Heijnen, 1992). The used experi•
mental set-up, start-up procedures and analytical techniques have been described before (van Houten
et al., 1994; van Houten et aJ., 1995a; van Houten et al., 1995b).

The results of our experiments performed at pH 7.0 showed that sulphate-reducing bacteria formed
stable biofilms on pumice particles. In addition, carrier-free aggegates were formed. The average
Sauter mean diameter of the aggregates was 1.57 mm. Moreover, phase-contrast microscopy and
SEM showed highly branched surfaces with sometimes large protrusions. A SEM photograph is given
in Fig. lao Examination of the bacterial composition of the aggregates by phase-contrast microscopy,
SEM as well as enrichments showed that Desulfovlbrio species and Acetobacterium species were the
most abundant microorganisms (van Houten et al, I995a). The ratio of Desulfovibrio sp. to Acetob•
acterium sp. was about 20: 1. These predominant microorganisms were rather heterogeneously
distributed throughout the aggregates.
 Aggregation of sulphate-reducing bacteria 87
Further experiments, at different pH-values (van Houten et al., 1995a), showed that the pH affected
aggregate configuration and diameter. A pH increase led to increased surface irregularity , due to
reduced hydrogen sulphide toxicity, without affecting the particle diameter. A pH decrease caused a
decreased surface irregularity and changed the aggregate Sauter mean diameter from 1.50 mm at pH
7.0 to 2.26 at pH 5.5. However, the pH did not have a significant effect on the biomass composition.

Moreover, experiments were conducted in which the effect of CO on aggregation, growth and
population dynamics was studied (van Houten et ai, 1995b). The results showed that CO affected the
aggregate shape and diameter. SEM photographs showed that rough aggregates, pre-grown on
H/C02, changed into smooth aggregates upon addition of CO. An example is given in Figure lb.
Addition of CO also changed the aggregate Sauter mean diameter from 1.67 mm at 5% CO to 2.06
mm at 20% CO. Additionally, although the hydrodynamics in the reactor were unchanged, the
detachment rate was half of the rate observed before CO addition.
After addition of CO, a layered aggregate structure developed. Acetobacterium sp. were mainly
located at the outside of the aggregates , whereas Desulfovibrio sp. were located inside the aggegates .

(a)

(b)

F'lgUre I. SEM photographs of sulphate -reducing aggregates. (a) Aggregate grown on H, /CO,. (b) Aggregate grown on
H,/CO. Clearly the surface if the aggregate changed upon addit ion of CO . Bar length 500 I'm .

JllST 32:8.'
88 R. T. van HOlITEN et al;
DISCUSSION

The observations made during all our experiments show a clear trend. During optimal growth
conditions, i.e . pH ~ 7.0 and no carbon monoxide present, the biofilm or aggregate surface is rough
with sometimes large protrusions. However, when the growth decreases by some inhibitory com•
pound, i.e. excess of H2S, excess of H+ (low pH) or carbon monoxide, the biofilm or aggregate
surface becomes smooth. Additionally, the detachment rate of the aggregates is lower in case of a
reduced growth rate due to inhibition. For example, the detachment rate reduced by 50% upon the
change from CO 2 to CO although the hydrodynamics remained constant . The observed phenomena
are most apparent in case of inhibition with carbon monoxide.

In recent years, several publications have discussed the same phenomena . Beeftink (Beeftink and
Staugaard,1986) showed that aggregates producing butyric acid from glucose had a loose and hairy
appearance. The propionate and acetate producing aggregates had a much smoother aggregate surface.
They related the observed hairy structure to the vigorous production rate - and therefore high
substrate consumption rate - of gaseous metabolites during butyrate product ion. Such an open struc•
ture would facilitate substrate transport into the interior of the aggregates.
Additionally, the articles of Tijhuis (Tijhuis et al. 1995a; 1995b) showed a relationship between
surface substrate loading rate and aggregate morphology. They found that, with a high surface
substrate loading rate and low applied detachment forces, hairy and rough aggregates developed. It
appeared also that these hairy aggregates were weaker than smooth aggregates. Moreover, they postu•
lated that due to the high surface to volume ratio, which probably reduces diffusion limitation of the
substrates, the growth rate of the bacteria in the protrusions is high. Higher surface substrate loading
rates could therefore result in higher growth rates, weaker aggregates and higher detachment rates.

As there was a clear difference in the detachment rate between the growing and non-growing aggre•
gates under the same hydrodynamic conditions, Tijhuis (Tijhuis ff al. 1995b) proposed two hypoth•
eses to explain the observed phenomena.
The first hypothesis (A) was based on the assumption that a gradient in strength in the aggregate
exists due to a gradient in growth rate over the depth of the biofilm. The fast growing outer shell of
the biofilm is assumed to produce less exopolymer than the inner part of the biofilm. This implies
that the outer biofilm is weaker than the inner part, since exopolymers are supposed to be responsible
for the strength of the aggregates. Therefore, the outer part of the aggregate is much more susceptible
to shear forces than the slowly growing inner part of the aggregate. As a result, the fast growing
outer shell of the aggregate is sheared off during the substrate depleted period. After that, the
stronger , inner part of the aggregate is left, resulting in a lower detachment rate.
The second hypothesis (B) was based on their observation that most of the aggregates observed, have
a knobby or rough surface with sometimes large protrusions. These protrusions are exposed to higher
shear forces and are easily removed from the aggregate. Due to the absence of growth during
substrate depletion the number of protrusions then decreases and, therefore, the rate of detachment
decreases. During such a process the aggregate surface will change from a rough to a smooth
appearance. This is in contrast to hypothesis A, where the roughness of the aggregate surface does
not necessarily change.

Our experimental results reveal, at least partly, the discussion about the two different hypotheses. In
our experiments the growth rate of the microorganisms was clearly intlucnced by either low pH or
addition of CO. Particularly, the addition of CO causes a decrease in the growth rate of the bacteria
(M6rsdorf et al., 1992; Davidova et al., 1994). Simultaneously, the protrusions disappeared from the
aggregate surface, as shown in Fig. la.b, Clearly, there is a relationship between the growth rate of
the bacteria and the aggregate morphology, since the hydrodynamics in the reactor remained constant.
Therefore, according to our experimental results, hypothesis B is most obvious.
 Aggregationof sulphate-reducingbacteria 89
It might be argued that, since we did not measure the amount of exopolymer in the aggregate, hypo•
thesis A still can be valid. However, according to hypothesis A the diameter of the aggregates should
decrease after substrate depletion or growth inhibition. This is not observed. In contrast, upon
addition of CO or growth at a pH below 7.0, the Sauter mean aggregate diameter increased from 1.6
mm to 2.0 mm (van Houten et 01., 1995ab). Thus, hypothesis A is not valid.

Although hypothesis B and our experimental results seem very complementary, there are still some
questions left. Interesting, for example, is the question why the non-inhibited microorganisms tend to
grow as aggregates with a rough or hairy surface. One hypothesis is given in one of our previous
articles (van Houten et al., 1995a) in which we postulated that the microorganisms show a kind of
chemotaxis, by which they tend to increase the specific surface area in order to postpone possible
substrate limitation. This hypothesis is in agreement with the observation made by other authors
(Christensen, 1989; van Ede, 1994). that this type of aggregate only prevails during substrate
limitation. More detailed studies, however, are required for the verification of this hypothesis.

Another related phenomenon observed is the change in the spatial distribution of the microorganisms
upon addition of CO. As we have discussed before (van Houten et al. 1995b). the Desulfovibrlo sp.
cannot survive in the outer region of the aggregates at higher CO levels. Therefore, they will be
replaced by Acetobacterium sp. This results in a relative increase of the amount of Acetobacterlum
sp. As these bacteria can convert CO either into acetate or into CO2, CO will be removed when it
diffuses into the inner region of the aggregate. In this way growth of Desulfovibrlo sp. in the inner
part of the aggregate becomes possible. For the aggregates grown on HlCO this might also cause the
observed change in surface structure. However, during the experiments in which the pH was changed
to values below 7.0 (van Houten et 01., 1995a), we observed also the development of a smooth
aggregate surface without any measurable change in the heterogeneous population distribution.
Therefore, it is not plausible that the change in surface structure during growth on HlCO is deter•
mined by the change in population distribution.

CONCLUSIONS

The surface structure of the sulphate-reducing aggregates depends on the growth rate of the microor•
ganisms present. A high growth rate yields a rough and hairy aggregate surface. A low growth rate
yields a smooth aggregate surface. since the low growth rate of the microorganisms allows detach•
ment of the large protrusions from the aggregate surface.
The surface structure is not dependent on the type of bacteria present at the surface, i.e. sulphate•
reducing bacteria or homo-acetogenic bacteria.

ACKNOWLEDGEMENTS

We thank Dr L.W. Hulshoff Pol for valuable discussion regarding the manuscript. This work was
financially supported by the European Community STEP-programme. grant no. CT 90-0073 and the
IOP-Milieutechnologie programme, grant no. lOP 90-209 and Paques Environmental Technology
B.V.

REFERENCES

Beeftink, H.H. and Staugaard, P. (1986) Structure and dynamics of anaerobic bacterial aggregates in a gas-lift
reactor. Appl. Environ. Microbial. 52:1139-1146.
90 R. T. van HOUlEN el aL
Beeftlnk, H.H. (1987). Anaerobic bacterial aggregates: variety and variation. Doctoral thesis, Department of
Chemical Technology, Amsterdam University, Amsterdam, The Netherlands.
Christensen, F.R., Holm Kristensen, G. and la Cour Jansen, J. (1989) Biofilm structure . an important and
neglected parameter in wastewater treatment. Water. Sci. Technol. 21(8/9):805-814.
Davidova, M.N., Tarasova, N.B., Mukhitova, F.K. and Karpilova, I.U. (1994). Carbon monoxide in meta•
bolism of anaerobic bacteria. Can. J. Microbiol. 40:417-425.
van Ede, C.J. (1994) Bioconversions catalyzed by growing immobilized bacteria. Doctoral thesis, Department
of Chemical Egineering, Groningen State University, The Netherlands.
Heijnen, J.J. (1984). Biological industrial wastewater treatment: minimizing biomass production and maximi•
zing biomass concentration. Doctoral thesis, Department of Biochemical Engineering, Delft Technical
University, Delft, The Netherlands.
van Houten, R.T., Hulshoff Pol, L.W. and Lettinga, G. (1994). Biological sulphate reduction using gas-lift
reactors fed with hydrogen and carbon dioxide as energy and carbon source. Biotechnol. Bioeng.
44:586-594.
van Houten, R.T., Oude-Elferink, S.J.W.H., van Hamel, S.E., Hulshoff Pol, L.W. and Lettinga, G. (1995a).
Sulphate reduction by aggregates of sulphate-reducing bacteria and homo-acetogenic bacteria in a lab•
scale gas-lift reactor. (Submitted).
van Houten, R.T., van.der Spoel, H., van Aelst, A.C., Hulshoff Pol, L.W. and Lettinga, G. (l995b) Biolo•
gical sulphate reduction using synthesis gas as energy and carbon source. (Submitted).
MOrsdorf, G., Frunzke, K., Gadkari, D. and Meyer, O. (1992). Microbial growth on carbon monoxide.
Biodegradation 3:61-82.
Mulder, A. and Heijnen, J.J. (1982). The effect of carrier characterisitics on the biofilm development in airlift
suspension reactors. In: Proceedings of the 2"" Netherlands Biotechnology Congress, H. Breteler, Ph.
van Lelyveld, K.Ch.A.M. Luyben (Eds.), Netherlands Biotechnology Society, Amsterdam, pp. 102•
110.
Tijhuis, L. (1994). The biojilm airlift suspension reactor: biofilm formation. detachment and heterogeneity.
Doctoral thesis, Department of Biochemical Engineering, Delft Technical University, Delft, The
Netherlands.
Tijhuis, L., Hijman, B., van Loosdrecht, M.C.M. and Heijnen, J.J. (1995a). Influence of detachment,
substrate loading and reactor scale on the formation of biofilms in airlift reactors. Appl, Microbiol.
Biotechnol. (Accepted).
Tijhuis, L., van Loosdrecht, M.C.M. and Heijnen, J.J. (1995b). Dynamics of biofilm detachment in biofilm
airlift suspension reactors. Btotechnol. Bloeng. 45:481-487.