Journal Pre-proof

A 3D Origami Paper-Based Analytical Device Combined with PVC Membrane for

Colorimetric Assay of Heavy Metal Ions: Application to Determination of Cu(II) in
Water Samples

Hoda Sharifi, Javad Tashkhourian, Bahram Hemmateenejad

PII: S0003-2670(20)30654-1
DOI: https://doi.org/10.1016/j.aca.2020.06.006
Reference: ACA 237721

To appear in: Analytica Chimica Acta

Received Date: 10 March 2020

Revised Date: 30 May 2020
Accepted Date: 1 June 2020

Please cite this article as: H. Sharifi, J. Tashkhourian, B. Hemmateenejad, A 3D Origami Paper-
Based Analytical Device Combined with PVC Membrane for Colorimetric Assay of Heavy Metal Ions:
Application to Determination of Cu(II) in Water Samples, Analytica Chimica Acta, https://doi.org/10.1016/
j.aca.2020.06.006.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2020 Elsevier B.V. All rights reserved.

1 A 3D Origami Paper-Based Analytical Device Combined with PVC Membrane for

2 Colorimetric Assay of Heavy Metal Ions: Application to Determination of Cu(II) in

3 Water Samples

5 Hoda Sharifi, Javad Tashkhourian*, Bahram Hemmateenejad*

6 Chemistry Department, Shiraz University, Shiraz, 71454, Iran

10

11

12

13

14

15

16

17

18

19 Corresponding Author

20 Chemistry Department, Shiraz University, Shiraz, 71454, Iran

21 *E-mail: [email protected] (B. H.); [email protected] (J. T)

22

1
1 ABSTRACT

2 Microfluidic paper-based analytical devices (µPADs) as a potentially powerful analytical platform

3 have recently gained significant attention for on-site monitoring of heavy metal ions, which are one

4 of the most significant environmental concern because of non-degradability and high toxicity. The

5 commonly applied µPADs suffers from some defects, such as heterogeneous deposition of reagent,

6 resulting in poor detection limits and low sensitivity. So, in this work, a three-dimensional origami

7 µPAD combined with PVC Membrane was developed, which can manage problems of movement

8 of colored products or leaching out the dye and leading to color heterogeneity in the detection

9 zones. Furthermore, a waste layer was added to µPAD for loading of more amounts of the analyte,

10 which results in improvement of detection limit. AS a proof of concept, the µPAD was used for the

11 analysis of Cu2+ ion. For this purpose, pyrocatechol violet and chrome azurol S as colorimetric

12 reagents were doped into PVC membrane and injected in the detection zone. The proposed µPAD

13 was presented good linearity in the ranges of 5.0‒1400.0 and 5.0‒200.0 mg L‒1, and the limits of

14 detections of 1.7 and 1.9 mg L‒1 in presence of chrome azurol S and pyrocatechol violet,

15 respectively.

16
17
18 Keywords: Paper-based microfluidic devices; Origami; PVC membrane; Heavy metal ions;
19 Colorimetric detection
20

21

22

23

24

2
1

2 1. Introduction

3 Recently, microfluidic paper based analytical devices (µPADs) have gained extensive attention

4 because they are low cost, portable, easy to manufacture/use/store/transport and require the minimal

5 reagent and sample [1,2]. Paper has many advantages over traditional device materials that can

6 allow applying as a substrate material for microfluidic assays, including availability and low cost,

7 lightweight, high porosity, flexibility, compatibility, ease of transport, disposability, and

8 biodegradability and it is easy to transfer complex patterns on paper [3,4]. The first µPAD was

9 reported by Whitesides research group in 2007, for fluid handling and fluid analysis [5]. µPADs or

10 lab-on-a-paper devices are fabricated on a filter paper and consisting of hydrophilic/hydrophobic

11 microchannel networks that fluid wicks by capillary action without external power sources. The

12 above-mentioned advantages make these devices attractive for medical diagnosis, environmental

13 and food quality assays.

14 Heavy metal ions (i.e. Cu2+, Cd2+, Pb2+, Hg2+, Cr4+, Ni2+, etc.) as an increasingly important

15 environmental problem, are non-biodegradable, highly toxic, persistent and hazardous even at very

16 low concentrations [1]. These heavy metals can be spread among and accumulate in plants and

17 animals, entering into the human body via the food chain, binding easily to vital cellular

18 components, and leading to serious diseases, disorders (e.g., cancers, kidney malfunction,

19 Alzheimer’s disease, Parkinson’s disease, multiple sclerosis, osteoporosis, etc.), and organ failures

20 [1,6]. Therefore, monitoring of heavy metals levels in the environment, drinking water, food, and

21 biological fluids is of vital importance [1,6]. Numerous analytical methods have been developed for

22 identification and quantification of heavy metals, including: atomic absorption spectroscopy (AAS)

23 [7,8], electrochemistry [9-11], colorimetric analysis [12,13], inductively coupled plasma mass

24 spectrometry (ICP-MS) [14,15], inductively coupled plasma atomic emission spectrometry (ICP-
3
1 AES) [16], and visible and near-infrared reflectance spectroscopy (VNIR) [17]. While these

2 analytic methods are capable of providing high selectivity, sensitivity, and low detection limits for

3 the determination of trace amounts of metals, they are time-consuming and require complex and

4 expensive instruments, and trained specialists for operation. Furthermore, these methods are not

5 easily used as portable analytical tools for on-site detection of metal ions. Thus, researchers have

6 been striving to develop the simple, portable and low-cost sensors for rapid, sensitive and selective

7 detection of heavy metals that can be used for on-site assay and point-of-care tests [1,2].

8 Recent studies have reported the detection of very low concentrations of heavy metal ions in

9 real samples using microfluidic paper-based analytical devices. Mentele and coworkers employed a

10 flower-shaped µPAD for simultaneous colorimetric digestion of heavy metals (Cu2+, Fe3+, Ni2+) in

11 airborne particulate matter [18]. A two-layer µPAD was designed by Rattanarat et al. for

12 simultaneous colorimetric and electrochemical detections of metal ions [19]. A distance-based

13 µPAD using a water-soluble porphyrin derivative was reported for Cu2+ determination [20]. In

14 acidic conditions, the used porphyrin showed a high selectivity for reacting with Cu2+ that caused to

15 form a pink colored band along the channel of the µPAD. This device did not need any readout

16 equipment and worked based on measuring the length of the formed color.

17 The above-mentioned paper-based devices which commonly applied for colorimetric

18 detection of metal ions, suffer from some limitations. Movement of colored products from the

19 center of the spot to edge of the detection zone causes leaching out the dye and/or forming color

20 heterogeneity in the detection zones. In addition, the color reagents, which are spotted in the

21 detection zone can move into the fluid channel due to the direct connection of the detection zones

22 and fluid channels. This would lead to lose part of the signal. Moreover, some metal ions do

23 interaction with carboxyl groups present on the surface of the paper and hence do not reach totally to the

4
1 reaction (detection) zone. These phenomena result in low sensitivity and consequently poor detection

2 limits.

3 In this paper, we employed two simple strategies to lower down the effect of the mentioned-

4 above limitations. Firstly, polyvinylchloride (PVC) membrane was used to immobilize the

5 indicators on the paper surface. PVC has been used for a long for fabricating optical and

6 potentiometric sensors [21-26]. Many bulk optode systems based on a plasticized polymer

7 membrane for selective determination of cations were reported [27-33]. Shibata et al. recently

8 reported a paper-based colorimetric sensor for Na+ by attaching the PVC-based ion-selective optode on

9 the surface of paper [34]. Very recently, Soda and co-workers introduced distance-based µPAD

10 coupled to K+-selective film capillary sensor containing the potassium ionophore valinomycin

11 incorporated into a plasticized PVC membrane [35].

12 Here we showed that PVC can be used for fixing the indicators on the surface of paper such

13 that the movement of the reagent upon mixing with the analyte solution is lowered down

14 significantly. The results show that PVC membrane presented excellent uniform distribution for colored

15 products generated from the reaction between doped sensors and target analyte into test zones.

16 Moreover, the µPADs were fabricated based on a three-dimensional (3D) origami design, which enables

17 loading more amounts of the analytes. Combining this design with PVC-based immobilization of the

18 reagent resulted in a lower limit of detection. This new µPADs was used for assay of Cu(II).

19

20 2. Experimental

21 2.1. Reagents and Equipment.

22 Pyrocatechol violet (PV), murexide, eriochrome cyanin R (ECR), calcon, titan yellow (TY), xylenol

23 orange (XO), aluminon, congo red (CR), pyrogallol red (PR), calmagite, dithizone, 5-(4-

24 dimethylamino benzylidene) rhodanine (ABR), chrome azurol S (CAS), eriochrome black T (EBT),
5
1 1-(2-pyridylazo)-2-naphthol (PAN), 4-(2-pyridylazo)resorcinol (PAR), 1,5-diphenylcarbazone,

2 tetrahydrofuran (THF), high molecular weight poly(vinyl chloride) (PVC) and

3 tricaprylmethylammonium chloride (Aliquat 336) were purchased from Sigma-Aldrich.

4 Zinc (II) nitrate hexahydrate (Zn(NO3)2.6H2O), cadmium nitrate tetrahydrate

5 (Cd(NO3)2.4H2O), iron (III) nitrate nonahydrate (Fe(NO3)3.9H2O), manganese (II) nitrate

6 tetrahydrate (Mn(NO3)2.4H2O), lanthanum (III) nitrate hexahydrate (La(NO3)3.6H2O), cobalt (II)

7 nitrate hexahydrate (Co(NO3)2.6H2O), cerium (III) nitrate hexahydrate (Ce(NO3)3.6H2O), calcium

8 (II) nitrate tetrahydrate (Ca(NO3)2.4H2O), gallium (III) nitrate octahydrate (Ga(NO3)3.8H2O),

9 zirconium (IV) nitrate (Zr(NO3)4), dimethyl sulfoxide (DMSO), acetic acid, phosphoric acid and

10 ethanol were acquired from Merck chemical company.

11 Copper (II) nitrate trihydrate (Cu(NO3)2.3H2O), aluminum (III) nitrate nonahydrate

12 (Al(NO3)3.9H2O), nickel (II) nitrate (Ni(NO3)2), bismuth (III) nitrate (Bi(NO3)3), chromium(III)

13 nitrate nonahydrate (Cr(NO3)3.9H2O), dibutyl phthalate (DBP) were purchased from Fluka. Lead

14 (II) nitrate (Pb(NO3)2), mercury (II) nitrate dihydrate (Hg(NO3)2 . 2H2O), and silver nitrate

15 (AgNO3) were obtained from Riedel-de Haën. Boric acid and thorium (IV) nitrate hexahydrate

16 (Th(NO3)4.6H2O) were bought from BDH Laboratory.

17 All chemicals and materials were of analytical grade and used as received without further

18 purification.

19 Whatman # 2 filter paper supplied from Whatman LTD Company was used as a substrate

20 for making microfluidic devices. The pattern of the devices was created on the paper by AutoCAD

21 software and then printed on filter paper using HP LaserJet 1320 Printer. CanoScanLiDE 700F color

22 scanner was used to record Optical signal. Image analysis was done using ImageJ software.

23

24
6
1
2 2.2. Design and fabrication of the µPADs

3 For fabrication of hydrophobic barriers, the technique suggested by Dey et al., with minor

4 modifications, was employed [36]. Dey et al. used two-side printing, however, alignment of

5 channels in this method is challenging. By changing the backing time, we could make a

6 hydrophobic barrier by single-side printing [37]. Hydrophobic barrier pattern, defining the fluidic

7 path, was designed by AutoCAD and was printed on the paper using HP LaserJet 1320 Printer

8 contained HP Q5949A toner cartridge with a resolution of 600 dpi. The paper was then baked in the

9 oven at 200 °C for 50 min, which causes curing the Inkjet binder and penetrating it into the paper

10 pores to create the hydrophobic barriers [36].

11 Two types of 3D µPADs were fabricated as depicted in Figure 1. In the first type, µPAD

12 included two paper layers. The top layer contained a central zone (sample loading zone, diameter 6

13 mm) that directly connected with ten hydrophilic circular zones (pretreatment zones) via fluid

14 channels with 4.5 mm in length and 1.5 mm in width. The bottom layer consisted of ten hydrophilic

15 circular zones as detection zones with diameters of 3.5 mm. In another 3D µPAD design, the device

16 consisted of three paper layers. The first two layers were similar to the previous design. However, a

17 third layer was included as a waste layer, which led to a shared sample outlet, allowing larger

18 sample volume to arrive into the µPAD.

19

20 2.3. Preparation of reagent-doped plasticized PVC membrane

21 Plasticized PVC membranes were prepared by the procedure described by Shamsipur et al. [38]. The

22 colorimetric reagents were incorporated into a polymeric matrix prepared by dissolving 35 mg PVC

23 and 60 mg dibutyl phthalate as a plasticizer in 3.0 ml THF. To obtain the uniform distribution of

24 some of these indicators in the membrane, aliquat 336 was added to the above solution.
7
1

2 2.4. Assay procedures

3 For both designs of the 3D µPADs, a 0.35 µL portion of the colorimetric reagent-doped plasticized PVC

4 membrane was added to the detection zones, placed into the second layer, and allowed to dry completely.

5 Then, a buffer solution was pipetted onto the same zones. The µPAD was folded and placed in the

6 external glass apparatus as shown in Figure 2. After that, 30.0 µL (for µPAD with two layers) or 40.0

7 µL (for µPAD with three layers) of the sample solution was introduced to the sample reservoir. Because

8 of the capillary action of the paper, the solution flow toward the sensing areas, the analyte reacts with the

9 sensors followed by a visual color change. Finally, changes in the color of the sensing areas were

10 captured with the image scanner and the RGB color differences of each selected image were analyzed

11 using ImageJ software.

12 It should be noted that doping the indicator into PVC created a relatively hydrophobic

13 environment in the detection zone, which led to slowly penetrating the sample solution in the

14 sensing area. So, a controlled flow of sample solution is required [39,40]. Hence, the sample solution

15 was slowly injected (~ in 10 s) in the central zone for getting enough time to enter the analyte in the

16 detection zone. It was observed when the sample solution (40.0 µL) was quickly introduced to PAD, the

17 analyte was fast flowed down the channel and driven to the sensing area. Owing to a slower penetration

18 rate of the analyte in relatively hydrophobic sensing area, µPAD capacity was completely filled and

19 additional amounts (beyond capacity) were overflowed on the surface of hydrophobic areas of the

20 top layer.

21

22 2.5. Distance-based µPADs

23 To evaluate how metal ions movement on paper, distance-based µPADs were used [37]. These devices

24 consist of a long rectangular hydrophilic channel (30 mm in length and 3.6 mm in width)
8
1 connecting to a circular sample reservoir (diameter 3 mm). At first, 0.6 µL of the sample was added to

2 the sample zone, which then flowed along the channel. After µPAD was dried, the colorimetric reagent

3 was sprayed onto the paper surface. The appeared color showed the length of movement of the analyte.

5 2.6. Image processing

6 To obtain the colorimetric response, a CanoScan 700F color scanner was used to capture the images of

7 the sensing areas with 600 dpi resolution. The ImageJ color analysis software was used to calculate the

8 color intensity of sensing areas on the RGB color scale (first selecting the whole coloring areas and then

9 taking an average for color elements of R, G and B). The obtained color intensity for each sensing

10 area was subtracted from the intensity of the corresponding area related to the blank µPAD and was then

11 used as an analytical signal.

12

13 3. Results and Discussion

14 3.1. Assessment of stability and Color homogeneity

15 According to the previous studies, the most commonly used paper-based device was based on 2D flat

16 design containing a loading zone connecting to test zones via fluid channels [18,41,42]. In these µPADs,

17 the reagents flow backed from test zone to channels during reaction with the sample. So, during the

18 image analysis, since the selected area applied only on test zones, parts of the reagents flowing to channel

19 did not consider in the signal and led to create an error in the quantitative analysis. For resolving this

20 problem, here the direct connection was broken between the detection zones and fluid channels by

21 placing the multi detection zones in a layer separated from the layer including fluid channels. Moreover,

22 in the paper-based devices, color inhomogeneity, i.e., colored products are inconsistently distributed in

23 the detection zones, and dye leaching is an important factor that has a great influence on the efficiency of

24 the analytical method. Therefore, to lower down these unwilling effects, before spotting the colorimetric
9
1 reagents on the paper, they were immobilized within the dibutyl phthalate (DBP) plasticized PVC

2 membrane.

3 In our experimental work, the colorimetric reagents were incorporated into a polymeric

4 matrix. The used colorimetric reagents also in the absence of PVC were dissolved in water or

5 ethanol. For evaluating and comparing the performance of the prepared sensors in the absence or

6 presence of PVC, 15 colorimetric reagents, mostly used metallochromic reagents, were added to the

7 detection zones of two-layer devices and 30.0 µL of deionized water was introduced to the sample

8 reservoir. This amount of water can reach the detections zones and mix with the reagents. The

9 images of the devices arriving distilled water to the reagents are presented in Figure 3. As it is

10 obvious, when the reagents are not immobilized with PVC, they either moved to the edge of the

11 detection zone (resulting in inhomogeneous color spots) or leached out to the sampling paper layer.

12 Also, ring color has been observed for some indicators. Interestingly, the PVC immobilization

13 prevents leaching out of the reagents to the sampling PAD and at the same time it results in forming

14 homogeneous and accumulated color spots with lower radius. Also, as shown in Figure S2, when a

15 metal ion such as Cu2+ is added to chrome azurol S solution, PVC strongly prevents leaching back

16 of the reagent to sample solution and caused forming of very clear and focused spots.

17

18 3.2. Improving detection limit

19 It is not worth mentioning that the presence of more analyte in the detection zone would result in

20 more intense color changes and consequently achieving higher sensitivity and lower detection

21 limits. We used a simple experiment to show that the movement of metal ions into paper is varied

22 between cations. For this purpose, 18 metal ions in the form of nitrate salt were prepared and a little

23 amount of them were injected into the loading zone of the two-dimensional µPAD. The results

24 shown in Figure S1 of supplementary material confirm that the studied metal ions did not move in
10
1 the same distance through the µPAD channels. Because of the presence of carboxyl and hydroxyl

2 groups in the paper, the surface of the paper is negatively charged over a wide pH range [34]. Carboxylic

3 groups present in the paper substrate with pKa ≈ 4–5 are weakly dissociated and as anionic groups

4 interact with metal ions [34]. The degree of interactions depends on the charge density of metal ions. As

5 seen from Figure S1, metal ions with low charge density (e.g., Ni2+, Co2+, Cd2+, Mn2+, Zn2+) are weakly

6 interacted with the paper substrate and hence can move a longer distance on paper. In contrast, metal

7 ions of high charge density (e.g., Fe3+, Al3+, Ce3+, Ga3+, La3+, Th4+, Zr4+, Bi3+) are strongly interacted with

8 carboxylic groups and do no move a long distance.

9 So, application of µPADs for quantitative analysis of some cations has been limited to

10 determinate high concentrations owing to trapping them into the paper network before arriving in the

11 detection zone. To handle this limitation, larger volumes of the sample solution should be applied.

12 However, this usually doesn’t work out for limited capacity of the detection zone. Using larger volumes

13 results in overflow of the sample solutions on the surface of hydrophobic areas. Herein, to help more

14 metal ions reach to the detection zone, a third layer was suggested to add to the two-layer design

15 used in the previous section. This layer, which acted as waste, could wick the sample solutions and

16 prevent their overflowing. As it is shown in Figure 4, the 2-layer pad can tolerate around 20.0 µL

17 whereas the 3-layer pad can handle larger volumes such that the color intensity of the latter is

18 almost 2.5 times higher than that of the former.

19

20 3.3. Capillary flow rate through hydrophilic channels

21 Capillary flow through hydrophilic channels in the paper is an impotent issue in paper-based

22 microfluidics and has been studied extensively by Songok et al. [40, 43-45]. Analysis of the capillary

23 flow rate of reagents’ solution was achieved by recording the liquid front movement of two reagents

24 (pyrocatechol violet (PV) and chrome azurol S (CAS), which used for sensing of Cu (II) in the next
11
1 section) as a function of time in the rectangular channels (3.6 mm width). To do so, the µPAD was held

2 horizontally between two stands and 10.0 µL aliquots of the reagent solutions of two different

3 concentrations (2.5 ×10-4 mol L-1 and 1.7 ×10-4 mol L-1) were injected to the sampling zone. The

4 movement of the solution front was recorded by a digital camera in video recording mode. As

5 shown in Fig 5, the liquid movement front is fitted to the Washburn’s power law of the capillary flow

6 dynamics of l = Z t n [43-47]. All fitted lines have the correlation coefficient (R2) larger than 0.98, which

7 confirms the validity of the regression lines. The values of Z (7.2 and 5.6 for PV and 5.4 and 4.9 for CAS

8 for concentrations of 2.5 ×10-4 mol L-1 and 1.7 ×10-4 mol L-1, respectively) are more dependent on the

9 nature of indicators than the indicator’s concentration. Lesser dependency on concentration can be

10 attributed to very low concentrations used for the indicators such that the viscosities of the solutions do

11 not differ significantly for two concentrations. The power value (n) is almost constant around 0.3 for all

12 four experiments.

13 According to the explained power law of the capillary flow dynamics and as is observed from

∆௟ ௟ᇲ ି௟
14 Fig. 5, the flow velocities, which can be calculated as ‫∆ = ݒ‬௧ = ௧ ᇲ ି௧ are decreasing by time. The plots

15 shown in Fig. 5 can be divided in two linear regions at early and late times. Accordingly, the flow

16 velocities can be reported for these two regions. It is obvious that the slopes of the plots at early times are

17 much higher than those of the late times. Moreover, the early time velocities represent lesser dependency

18 on the type and concentration of the indicators since the early velocities might be more dependent on the

19 solution injection power. However, the late velocities represent obvious dependencies on both type and

20 concentration of the indictors.

21 The average flow velocities, where ݈′ and ‫ݐ‬′ are the last points in the plots and ݈ and ‫ ݐ‬are the first

22 points, are calculated as 0.5 and 0.45 mm/s for PV and 0.39 and 0.38 mm/s for CAS for concentrations of

23 2.5 ×10-4 mol L-1 and 1.7 ×10-4 mol L-1, respectively. These values are in the same order of Z values.

12
1 Concentrated solutions represent faster movement and so it is expected that at a fixed time the

2 concentrated solutions move longer. In our previous paper, we used such interesting dependency of

3 solution movement on concentration and used the length of movement as a measure of concentration

4 [37]

5 3.4. Assessment of analytical performance of the µPAD

6 To validate our approach (3-layer design combined with PVC immobilization), it was used for sensing of

7 Cu (II) ion using pyrocatechol violet and chrome azurol S reagents. In the preliminary

8 investigations, these reagents gave obvious color changes on the paper substrate in the presence of

9 copper ion. The used µ-PAD in this section is composed of 8 channels, every 4 channels for one

10 reagent. In this case, each sample solution of Cu(II) reaches to both of the reagents 4 times. This allows

11 more precise measurements by averaging of color intensities over 4 detection zones. By using two

12 reagents, one can derive two calibration curves for one analyte on a single µ-PAD, simultaneously.

13 In this manner, more accurate results will be achieved by averaging the estimated concentrations by

14 two calibration curves.

15 Before deriving calibration curves, some experimental parameters including reagent volume

16 and pH of the detection zone were optimized to get more sensitive results. According to the data

17 shown in Figure 4 and to be on a safe side, the volume of the sample solution was chosen 40.0 µL.

18 For preventing colorimetric reagent-doped plasticized PVC membrane prepared in THF from

19 penetrating in hydrophobic area of the PAD, 0.35 µL from each regent was selected to add in the

20 detection zone as the largest possible volume. The dependence of the sensor response on the pH

21 was studied in the absence and the presence of 100.0 mg L-1 of Cu2+ ion (Figure 6). According to

22 shown data, larger responses were found at pHs 5.0 and 9.0 for the pyrocatechol violet and chrome

23 azurol S reagents, respectively. In the case of chrome azurol S, the blue color of the formed complex

24 seems to be more intense at pH 6-8 and after that it quite decreasing. The color intensity of the free
13
1 reagent has also higher intensity at pH 6-8. However, image analysis showed that the color

2 difference (color intensity of complex – color intensity of free reagent) is maximum at pH 9.0. On

3 the other hand, the color intensity of the azurol S reagent was not stable at pH 9.0 and it was faded

4 by time whereas at pH 10.0, it represented more stability. Since the signal intensities at pH 10.0 and

5 9.0 were not differ significantly, because of higher color stability, pH 10.0 was chosen as optimum

6 for this reagent.

7 Figure 7 depicts the calibration curves for Cu(II) using two different indicators. For each

8 detection zone, the average of color intensity overall colored pixels was calculated and the red element of

9 RGB color space was used as response data for deriving calibration curve. For comparison, the

10 calibration curves derived in the absence of PVC but with the same experimental conditions are shown in

11 Figure 7, too. The analytical appraisals for the calibration curves of both approaches (in the presence and

12 absence of PVC) are given in Table 1. More sensitive and more precise results have been obtained in the

13 presence of PVC immobilized reagents. The lower concentration limits in the calibration curves for the

14 PVC-based sensor are 5.0 mg L-1 for both reagents whereas those obtained in the absence of PVC are

15 higher (10.0 and 30.0, mg L-1 for the chrome azurol S and pyrocatechol violet, respectively). Also, the

16 relative standard deviations of the PVC-based µ-pads are in the range of 8.02 and 4.84 whereas those of

17 non-PVC µ-pads are higher (in the range of 38.25 and 15.40). As shown in Figure S2, S3 and according

18 to the previous explanations, the higher performance of the PVC-based µ-pads can be attributed to the

19 formation of more concentrated and homogeneous color spots by PVC immobilization of the reagents on

20 the paper substrate.

21 The comparison of the results obtained in the present study and some previously reported paper-

22 based methods for Cu(II) sensing are summarized in Table 2. The bold advantage of the employed

23 µPADs in this study is wider dynamic range. Also, in some instances it reported lower DLs.

24
14
1 3.4. Analysis of the real sample

2 To demonstrate the practical applicability of the developed 3D µPAD for determining Cu2+ levels in

3 the real water samples, rain water and two tap water samples were collected and spiked with

4 different Cu2+ concentrations (30.0 and 70.0 mg L-1). Each solution was analyzed by the proposed

5 µPAD consisted of PVC-doped pyrocatechol violet. Four measurements were achieved for each

6 concentration. As shown in Table 3, the recoveries and the relative standard deviations (RSD) are in

7 the range of 100.27%‒105.86% and 0.18%‒2.22%, respectively.

9 4. Conclusions

10 The commonly used µPAD had certain limitations which caused conditions was become difficult to

11 achieve lower detection limits and wide dynamic range. To overcome these limitations, a 3D

12 origami paper-based analytical device in a new design was prepared for heavy metal ions

13 monitoring. It was found that PVC doping of indicators could fix the reagent on the paper, such that

14 the indicator did not leach even in the presence of larger sample volumes. By incorporation of a

15 waste layer, more volumes of the samples could be delivered. Integration of PVC-doping and waste

16 layer resulted in better reproducibility and lower detection limits. AS proof of concept, the designed

17 µPAD was successfully used as Cu(II) sensor employing pyrocatechol violet and chrome azurol S. In

18 this way, two calibration curves were derived on one PAD, simultaneously.

19

20 Conflicts of interest

21 The authors declare no conflict of interest.

22

23 Acknowledgments

15
1 We thankfully acknowledge financial support from Shiraz University Research Council.

3 Appendix A. Supplementary material

4 Supplementary material related to this article can be found in the web version.

6 REFERENCES

7 [1] X. Sun, B. Li, A. Qi, C. Tian, J. Han, Y. Shi, B. Lin, L. Chen, Improved assessment of accuracy

8 and performance using a rotational paper-based device for multiplexed detection of heavy metals,

9 Talanta 178 (2018) 426–431.

10 [2] J.P. Devadhasan, J. Kim, A chemically functionalized paper-based microfluidic platform for

11 multiplex heavy metal detection, Sens. Actuators, B 273 (2018) 18–24.

12 [3] M.I.G.S. Almeida, B.M. Jayawardane, S.D. Kolev, I.D. McKelvie, Developments of

13 microfluidic paper-based analytical devices (µPADs) for water analysis: Areview. Talanta 177

14 (2018) 176–190.

15 [4] G. Sriram, M.P. Bhat, P. Patil, U.T. Patil, H.Y. Jung, T. Altalhi, T. Kumeria, T.M.

16 Aminabhavi, R.K. Pai, Madhuprasad, M.D. Kurkuri, Paper-based microfluidic analytical devices

17 for colorimetric detection of toxic ions: A review, Trends Anal. Chem. 93 (2017) 212–227.

18 [5] A.W. Martinez, S.T. Phillips, M.J. Butte, G.M. Whitesides, Patterned paper as a platform

19 for inexpensive, low-volume, portable bioassays, Angew. Chem. Int. Ed. Engl. 46 (2007) 1318–

20 1320.

21 [6] M. Li, R. Cao, A. Nilghaz, L. Guan, X. Zhang, W. Shen, “Periodic-table-style” paper

22 device for monitoring heavy metals in water, Anal. Chem. 87 (2015) 2555–2559.

16
1 [7] M. Tuzen, Determination of heavy metals in soil, mushroom and plant samples by atomic

2 absorption spectrometry, Microchem. J. 74 (2003) 289–297.

3 [8] S. Hamida, L. Ouabdesslam, A.F. Ladjel, M. Escudero, J. Anzano, Determination of

4 cadmium, copper, lead, and zinc in Pilchard Sardines from the Bay of Boumerdés by atomic

5 absorption spectrometry, Anal. Lett. 51 (2018) 2499–2506.

6 [9] C. Topcu, G. Lacin, V. Yilmaz, F. Coldur, B. Caglar, O. Cubuk, I. Isildak,

7 Electrochemical determination of copper(II) in water samples using a novel ion-selective electrode

8 based on a graphite oxide–imprinted polymer composite, Anal. Lett. 51 (2018) 1890–1910.

9 [10] Y. Yao, H. Wu, J. Ping, Simultaneous determination of Cd(II) and Pb(II) ions in honey

10 and milk samples using a single-walled carbon nanohorns modified screen-

11 printed electrochemical sensor, Food Chem. 274 (2019) 8–15.

12 [11] Z. Zhai, N. Huang, H. Zhuang, L. Liu, B. Yang, C. Wang, Z. Gai, F. Guo, Z. Li, X. Jiang,

13 A diamond/graphite nanoplatelets electrode for anodic stripping voltammetric trace determination

14 of Zn(II), Cd(II), Pb(II) and Cu(II), Appl. Surf. Sci. 457 (2018) 1192–1201.

15 [12] R. Sedghi, B. Heidari, M. Behbahani, Synthesis, characterization and application of

16 poly(acrylamide-co-methylenbisacrylamide) nanocomposite as a colorimetric chemosensor for

17 visual detection of trace levels of Hg and Pb ions, J. Hazard. Mater. 285 (2015) 109–116.

18 [13] L. Chen, J. Li, L. Chen, Colorimetric detection of mercury species based on

19 functionalized gold nanoparticles, ACS Appl. Mater. Interface 6 (2014) 15897–15904.

20 [14] Ş. Tokalıoğlu, B. Çiçek, N. Đnanç, G. Zararsız, A. Öztürk, Multivariate statistical analysis

21 of data and ICP-MS detection of heavy metals in different brands of spices consumed in Kayseri,

22 Turkey, Food Anal. Methods 11 (2018) 2407–2418.

17
1 [15] M. Shen, L. Chen, W. Han, A. Ma, Methods for the determination of heavy metals in

2 indocalamus leaves after different preservation treatment using inductively-coupled plasma mass

3 spectrometry, Microchem. J. 139 (2018) 295–300.

4 [16] J.M. Matong, L. Nyaba, P.N. Nomngongo, Determination of As, Cr, Mo, Sb, Se and V in

5 agricultural soil samples by inductively coupled plasma optical emission spectrometry after simple

6 and rapid solvent extraction using choline chloride-oxalic acid deep eutectic solvent, Ecotoxicol.

7 Environ. Saf. 135 (2017) 152–157.

8 [17] H. Cheng, R. Shen, Y. Chen, Q. Wan, T. Shi, J. Wang, Y. Wan, Y. Hong, X. Li,

9 Estimating heavy metal concentrations in suburban soils with reflectance spectroscopy, Geoderma

10 336 (2019) 59–67.

11 [18] M.M. Mentele, J. Cunningham, K. Koehler, J. Volckens, C.S. Henry, Microfluidic paper-
12 based analytical device for particulate metals, Anal. Chem. 84 (2012) 4474-4480.
13 [19] P. Rattanarat, W. Dungchai, D. Cate, J. Volckens, O. Chailapakul, C.S. Henry, Multilayer
14 paper-based device for colorimetric and electrochemical quantification of metals, Anal. Chem. 86
15 (2014) 3555–3562.
16 [20] R. Pratiwi, M.P. Nguyen, S. Ibrahim, N. Yoshioka, C.S. Henry, D.H. Tjahjono, A selective
17 distance-based paper analytical device for copper (II) determination using a porphyrin derivative, Talanta
18 174 (2017) 493–499.
19 [21] M. Akhond, M.B. Najafi, J. Tashkhourian, Lanthanum-selective membrane electrode based on
20 2,2′ -dithiodipyridine, Anal. Chim. Acta. 531 (2005) 179–184.
21 [22] V. Zare-Shahabadi, M. Akhond, J. Tashkhourian, F. Abbasitabar, Characterization of a new
22 uranyl selective bulk optode; utilizing synergistic effect in optical sensor, Sens. Actuators, B 141 (2009)
23 34–39.

18
1 [23] M. Akhond, M.B. Najafi, J. Tashkhourian, A new cerium (III)-selective membrane electrode
2 based on 2-aminobenzothiazole, Sens. Actuators, B 99 (2004) 410–415.
3 [24] Z. Yan, S. Wang, H. Wang, S. Wu, Bismuth (III) PVC membrane ion selective electrodes based
4 on two compounds: Acylhydrazone and thiosemicarbazone with 1,3,4-thiadiazole, Mater. Sci. Eng., C 33
5 (2013) 2562-2568.
6 [25] K. Zhuo, Y. Wei, J. Ma, Y. Chen, G. Bai, Response of PVC membrane ion-selective
7 electrodes to alkylmethylimidazolium ionic liquid cations, Sens. Actuators, B 186 (2013) 461–465.
8 [26] Y. Lee, Y.K. Hahn, J.Y. Park, H. Seo, J. Jung, E. Cho, Y.S. Choi, S.S. Lee, An integrated
9 ion-selective optode sensor cartridge for directly detecting electrolyte ions in blood plasma without
10 pretreatment to adjust pH, Sens. Actuators, B 280 (2019) 256–262.
11 [27] F.B.M. Suah, Preparation and characterization of a novel Co (II) optode based on polymer
12 inclusion membrane, Anal. Chem. Res. 12 (2017) 40–46.
13 [28] F.B.M. Suah, M. Ahmad, L.Y. Heng, Highly sensitive fluorescence optode based on
14 polymer inclusion membranes for determination of Al (III) ions, J Fluoresc 24 (2014) 1235–1243.
15 [29] F.B.M. Suah, M. Ahmad, L.Y. Heng, A novel polymer inclusion membranes based optode
16 for sensitive determination of Al3+ ions, Spectrochim. Acta, Part A 144 (2015) 81-87.
17 [30] A. Safavi, M. Sadeghi, Design and evaluation of a thorium (IV) selective optode, Anal.
18 Chim. Acta 567 (2006) 184–188.
19 [31] A. Yari, M.B. Gholivand, F. Rahhedayat, Development and characterization of a new
20 nickel (II) ion selective optode based on 2-amino-1-cyclopentene-dithiocarboxylic acid.
21 Measurement 44 (2011) 1691–1696.
22 [32] N. Alizadeh, A. Moemeni, M. Shamsipur, Poly(vinylchloride)-membrane ion-selective
23 bulk optode based on 1,10-dibenzyl-1,10-diaza-18-crown-6 and 1-(2-pyridylazo)-2-naphthol for
24 Cu2+ and Pb2+ ions, Anal. Chim. Acta 464 (2002) 187–196.
19
1 [33] W.H. Chan, R.H. Yang, T. Mo, K.M. Wang, Lead-selective fluorescent optode membrane
2 based on 3,3´,5,5´-tetramethyl-N-(9-anthrylmethyl)benzidine, Anal. Chim. Acta 460 (2002) 123–
3 132.
4 [34] H. Shibata, T.G. Henares, K. Yamada, K. Suzuki, D. Citterio, Implementation of a
5 plasticized PVC-based cation-selective optode system into a paper based analytical device for
6 colorimetric sodium detection, Analyst 143 (2018) 678–686.
7 [35] Y. Soda, D. Citterio, E. Bakker, Equipment-free detection of K+ on microfluidic paper-based
8 analytical devices based on exhaustive replacement with ionic dye in ion-selective capillary sensors,
9 ACS Sens. 4 (2019) 670-677.
10 [36] R. Dey, S. Kar, S. Joshi, T.K. Maiti, S. Chakraborty, Ultra-low-cost ‘paper-and-pencil’
11 device for electrically controlled micromixing of analytes, Microfluid Nanofluid 19 (2015) 375-
12 383.
13 [37] M. Taghizadeh-Behbahani, B. Hemmateenejad, M. Shamsipur, A.R. Tavassoli, A paper-
14 based length of stain analytical device for naked eye (readout-free) detection of cystic fibrosis,
15 Anal. Chim. Acta 1080 (2019) 138–145.
16 [38] M. Shamsipur, M. Sadeghi, M.H. Beyzavi, H. Sharghi, Development of a novel
17 fluorimetric bulk optode membrane based on meso-tetrakis(2-hydroxynaphthyl) porphyrin
18 (MTHNP) for highly sensitive and selective monitoring of trace amounts of Hg2+ ions, Mater. Sci.
19 Eng., C 48 (2015) 424–433.
20 [39] S. Bhattacharya, S. Kumar, A.K. Agarwel, Paper Microfluidics: Theory and Applications,
21 First ed., Springer, Singapore, 2019.
22 [40] J. Songok, M. Toivakka, Modelling of capillary-driven flow for closed paper-based
23 microfluidic channels, J. Micromech. Microeng. 27 (2017) 065001.

20
1 [41] W. Dungchai, O. Chailapakul, C.S. Henry, Use of multiple colorimetric indicators for
2 paper-based microfluidic devices, Anal. Chim. Acta 674 (2010) 227–233.
3 [42] S. Karita, T. Kaneta, Rapid acid-base titrations using microfluidic paper-based analytical
4 devices, Anal. Chem. 86 (2014) 12108–12114.

5 [43] J. Songok, M. Tuominen, H. Teisala, J. Haapanen, J. Mäkelä, J. Kuusipalo, M. Toivakka,

6 Paper-based microfluidics: fabrication technique and dynamics of capillary-driven surface flow, ACS

7 Appl. Mater. Interfaces 6 (2014) 20060-20066.

8 [44] J. Songok, M. Toivakka, Controlling capillary-driven surface flow on a paper-based

9 microfluidic channel, Microfluid Nanofluid 20 (2016) 63.

10 [45] J. Songok, M. Toivakka, Enhancing capillary-driven flow for paper-based microfluidic

11 channels, ACS Appl. Mater. Interfaces 8 (2016) 30523-30530.

12 [46] R.R. Rye, J.A. Mann, F.G. Yost, the flow of liquids in surface grooves, Langmuir 12 (1996)

13 555-565.

14 [47] E.W. Washburn, The dynamics of capillary flow, Phys. Rev. 17 (1921) 273.

15 [48] L. Liu, M.R. Xie, F. Fang, Z.Y. Wu, Sensitive colorimetric detection of Cu2+ by

16 simultaneous reaction and electrokinetic stacking on a paper-based analytical device, Microchem J

17 139 (2018) 357-362.

18 [49] R. Pratiwi, M.P. Nguyen, S. Ibrahim, N. Yoshioka, C.S. Henry, D.H. Tjahjono, A selective
19 distance-based paper analytical device for copper (II) determination using a porphyrin derivative, Talanta
20 174 (2017) 493–499.

21 [50] S. Chaiyo, A. Apilux, W. Siangproh, O. Chailapakul, High sensitivity and specificity

22 simultaneous determination of lead, cadmium and copper using µPAD with dual electrochemical and

23 colorimetric detection, Sens. Actuators, B Chemical 233 (2016) 540-549.

21
1 [51] J. Rull-Barrull, M. d’Halluin, E.L. Grognec, F.X. Felpin, A paper-based biomimetic device for

2 the reduction of Cu(II) to Cu(I) - Application to the sensing of Cu(II), Chem Comm 52 (2016) 6569-

3 6572.

4 [52] S. Chaiyo, W. Siangproh, A. Apilux, O. Chailapakul, Highly selective and sensitive paper-

5 based colorimetric sensor using thiosulfate catalytic etching of silver nanoplates for trace determination

6 of copper ions, Anal. Chim. Acta 866 (2015) 75–83.

7 [53] A. Sadollahkhani, A. Hatamie, O. Nur, M. Willander, B. Zargar, I. Kazeminezhad, Colorimetric

8 disposable paper coated with ZnO@ZnS Core–Shell nanoparticles for detection of copper ions in

9 aqueous solutions, ACS Appl. Mater. Interfaces 20 (2014) 17694-17701.

10

11

22
1

2
3 Table 1. The analytical features for the calibration curves of both approaches (in the presence and
4 absence of PVC)

In the presence of PVC In the absence of PVC

CS PV CS PV

Linear range 5.0-1400.0 5.0-200.0 10.0-1000.0 30.0-400.0

(mg L-1)

Sensitivity 51.35 38.87 30.79 12.60

Regression 0.9816 0.9863 0.9195 0.9591

coefficient (R2)

Detection limit 1.7 1.9 2.3 4.1

(mg L-1) (n=4)

Repeatability 4.84 8.02 15.40 38.25

(For 70.0 mg L-1) (For 30.0 mg L-1) (For 70.0 mg L-1) (For 30.0 mg L-1)
(RSD %) (n=4)
5

23
1 Table 2. Comparison of the performance of the proposed µPAD with the previously reported paper-
2 based methods for determination of Cu2+ ion.

Paper-based LOD Linearity range

Indicator [Ref]
methods (mg L−1) (mg L−1)

Paper-based
Bathocuproine 1.60 5.00-80.00 [1]
device

Paper-based
Polyethyleneimine 1.90 6.35-63.54 [48]
analytical device

Distance-based Meso-tetrakis(1,2-
paper analytical dimethylpyrazolium-4-yl) Not reported 1.00-6.00 [49]
device porphyrin sulfonate (TDMPzP)

Paper-based
CTAB/thiosulfate/silver
colorimetric 5.00 10.00-350.00 [50]
nanoplates
sensor

Paper-based
Thioglycolic acid 2.10 6.35-63.54 [51]
biomimetic device

“Periodic-Table-
Style” paper Bathocuproine Not reported 0.80- 8.00 [6]
device

Paper-based
CTAB/thiosulfate/silver
colorimetric 0.30 0.50–200.00 [52]
nanoplates
sensor

ZnO@ZnS Core-shell
Paper sensor 0.96 0.96-96.00 [53]
Nanoparticles
Paper-based
Chrome azurol S 1.7 5.0-1400.0 In this
colorimetric
Pyrocatechol violet 1.9 5.0-200.0 work
sensor
3

24
1
2

3 Table 3. Recovery assays of the proposed method for the determination of Cu2+ in water samples (n =4).

sample Spiked Founded Recovery RSDb

(mg L−1) (mg L−1) (%) (%)

Rain water 0 NDa - -

30 30.67±0.68 102.23 2.22

70 70.43±0.79 100.61 1.12

Tap water 1 0 ND - -

30 31.13±0.29 103.77 0.93

70 70.19±0.13 100.27 0.18

Tap water 2 0 ND - -

30 31.76±0.23 105.86 0.72

70 70.37±0.75 100.53 1.06

4 a
Not detectable
5 b
Relative standard deviation

25
1 Figure captions
2 Figure 1. Design of the µPADs in this study. The black area is hydrophobic Inkjet and the white
3 region is hydrophilic paper. (a) distance-based µPAD for measuring the movement length of
4 analytes, (b): three-dimensional two-layer µPAD and (c): three-dimensional three-layer µPAD.
5

6 Figure 2. The whole process of designing and using µPADs. The colorimetric reagent-doped
7 plasticized PVC membrane and buffer are added to the detection zones placed into the second layer

8 (detection layer). The layers of µPAD are folded such that the detection layer is located in between
9 the waste layer and the sample loading layer. Then µPAD is placed in the external glass apparatus.
10 The sample solution is introduced to the sample reservoir. After sample solution reaching into detection
11 zones, colored product generates due to the reaction between the sample and colorimetric reagent.
12

13 Figure 3. Comparing between non-immobilized and PVC-immobilized reagent’s spots with respect
14 to the spot shape and homogeneity for different metallochromic reagents. The backside images of
15 µPAD before and after loading 30 µL distilled water into the sample reservoir are shown. The left-
16 and right-hand side panels show the images of µPAD for non-immobilized and PVC-immobilized
17 reagents, respectively. Each reagent was spotted on two neighboring circles. The investigated
18 metallochromic reagents are titan yellow (TY), calcon (Calc), eriochrome cyanin R (ECR),
19 murexide (Mur), pyrocatechol violet (PV), calmagite (Calm), pyrogallol red (PR), congo red (CR),
20 aluminon (Alum), xylenol orange (XO), 1-(2-pyridylazo)-2-naphthol (PAN), eriochrome black T
21 (EBT), chrome azurol S (CAS), 5-(4-dimethylamino benzylidene) rhodamine (ABR) and dithizone
22 (Dith).
23

26
1 Figure 4. Investigating loading capacity of µPADs: back-side view of the two- (upper panel) and

2 three-layer (lower panel) µPAD after loading different volumes of 0.005 M xylenol orange into the

3 sampling zone. The numbers on the µPADs denote the volume of the loaded solution in µL.

5 Figure 5. Dependency of solution front over time in the hydrophilic rectangular channels of 3.6

6 mm width with two different concentrations of PV and CAS indicators (2.5 ×10-4 mol L-1 and 1.7

7 ×10-4 mol L-1). The data points represent the experimental data and the dashed lines are the best fit to the

8 power law of l = Z t n.

10 Figure 6. Color change of the µPAD over pH and changing in the averaged R color intensity over

11 pH in the absence (a) and presence (b) of 100.0 mg L-1 Cu2+ ion for the PVC-doped indicators

12 pyrocatechol violet and chrome azurol S. The plotted responses are the average of four detection

13 zone on each µPAD and the error bars are the standard deviations.

14

15 Figure 7. The comparison of the calibration curves generated from µPAD combined with PVC

16 membrane and non-PVC membrane µPAD for Cu (II) determination. Response of chrome azurol S

17 (a) and pyrocatechol violet (b) in the presence (blue) and absence (orange) of PVC membrane

18 against different Cu (II) concentration. Each data point represents an average calculated from four

19 replicate measurements, and the error bars are the standard deviations.

20

21

22

23

27
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26 Figure 1
27
28

28
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18 Figure 2
19
20

29
1
2
3
4
5
6
7
8
9
10
11
12
13 Figure 3
14
15
16

30
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 Figure 4
25
26

31
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21 Figure 5
22

32
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19 Figure 6
20
21
22

33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 Figure 7
25
26
27
28
29
34
 Highlights
• Development of a new design of paper-based microfluidic device for heavy metal
ions.

• PVC doping of indicators prevents leaching out the dye and fix it on the paper.

• The µPADs combined with PVC immobilization shows higher sensitivity.

• Presence of waste layer into µPAD allows loading of more amounts of analyte.
 CRediT author statement

Hoda Sharifi: Investigation, methodology, writing original draft

Javad Tashkhourian: Conceptualization, supervision, writing-review-editing
Bahram Hemmateenejad: Conceptualization, supervision, writing-review-editing
Declaration of interests

☒ 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.

☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests: