A LOW-COST PAPER-BASED MICROFLUIDIC IMPEDIMETRIC DEVICEC FOR THE DETECTION OF WATER HARDNESS

A microfluidic paper-based impedimetric device was developed as a water hardness sensor. This device is capable of performing the analysis with a sample volume of few microliters with no prior treatments. A phenol-formaldehyde graphene electrode modified with ethylenediaminetetraacetate was used as the working electrode. Ag pseudo reference and carbon electrodes were used to fabricate the device. Current simultaneous metal ion detection sensors are based on complex and expensive electrode setups. The proposed inexpensive, quick and portable device is capable of detecting Ca2+ and Mg2+ simultaneously. Electrode double layer-based charge transfer resistance and the maximum negative imaginary impedance produced a linear correlation with each metal ion concentration. The calculated limits of detection for Ca2+ and Mg2+ were 0.31 and 0.24 ppm, respectively. A set of samples containing Ca2+ and Mg2+ with a hardness of 2 ppm (as calcium carbonate) were used to test the device. The proposed tool is suitable as a semi-quantitative device for the determination of hardness in water.

has not been previously reported. However, few paper-based impedimetric devices were reported previously for the detection of biological parameters. [6], [7] Work reported in this paper describes μPad that uses impedimetric based detection for the simultaneous detection of Ca 2+ and Mg 2+ ions. Phenolformaldehyde graphene electrode modified with ethylenediaminetetraacetate (EDTA) is used as the working electrode to fabricate the three-electrode system. This EDTA-modified phenol-formaldehydegraphite electrode platform trap metal ions may increase the charge density of the electrode double layer when exposed to Ca 2+ and Mg 2+ ions. Compared to the double layer without chelated Ca 2+ /Mg 2+ ions, a layer chelated with metal ions has a lower double-layer capacitance. This type of capacitance change was observed with divalent copper and nickel ions for an electrode with a modified layer containing nitrilotriacetic acid groups. [8] Change in the double layer charge density due to metal ion chelation can be detected as an impedance change using electrochemical impedance spectroscopy (EIS). [9] Nyquist plots can be used to calculate this parameter. [9] The possibility of using the highest imaginary impedance (highest -Z") to estimate the concentration of the targeted analyte is reported in this paper. The selectivity of the detector can be improved by introducing selective chelating agents, varying pH, or adding masking agents. The ion density of the working electrode double layer may change in the presence of metal ions that can chelate with EDTA.

Preparation of working electrode material
Oxalic acid (0.6 g) was dissolved in a minimum amount of deionized water. The oxalic aqueous solution and a solution of phenol (phenol 6.4 g and formaldehyde 16 mL) were mixed and heated at 90 °C for an hour. The resultant mixture was cooled and the bottom phenol-formaldehyde polymer layer was collected. Phenol-formaldehyde, graphene and Na2EDTA (4:2:1) were mixed to develop the modified working electrode material and the bare working electrode material was developed by mixing phenol-formaldehyde and graphite (2:1) only. Graphite powder (200 μm) was obtained from Kahatagaha Graphite Lanka, Sri Lanka.

Development of the paper-based device
Commercially available varnish was used to draw the hydrophobic barrier on a Whatman No. 01 filter paper to construct the paper-based device platform. Silver conductive ink (CircuitWorks® Silver Conductive Pens) and 6HB pencil were used to fabricate the reference and counter electrodes, respectively. Phenol-formaldehyde-graphite-Na2EDTA (4:2:1) mixture was used to fabricate the working electrode. The developed device is shown in Figure 1. The working electrode developed with Phenol-formaldehyde-graphite (unmodified) was used as the control device.

Reagents
All reagents were analytical grade and all solutions were prepared using deionized water unless otherwise stated. A buffer solution (pH = 7.0) was used to prepare ion solutions of K + from KCl (Techno PharmChem, India), Ca 2+ from CaCl2 (Vickers, UK), Mg 2+ from MgCl2.6H2O (Merck, India) and Fe 3+ from FeCl3 (Merck, India).

Electrochemical characterization of the microfluidic paper-based electrochemical device
Electrochemical characterization was performed using ZIVE SP5 workstation (WonATech, Korea). Cyclic voltammograms for the solutions (2, 4, 6, 8, 10 mM) of K4Fe(CN)6 in 0.1 M KCl were collected using the developed μPad at a scan rate of 25 mV/s. The peak current variation with the concentration of K4Fe(CN)6 was plotted.

Measurement procedure of electrochemical impedances
Electrochemical impedance spectroscopy (EIS) studies were performed using ZIVE SP5 workstation (WonATech, Korea). Each run was conducted using a fresh sensor and a volume of 50 μl was used to collect EIS response of each solution in the range of 1 Hz to 1 MHz, with the amplitude potential of 60 mV. All electrochemical impedance spectrums were collected under the 0 V biased potential against the Ag pseudo reference electrode.

Preparation of laboratory samples
The proposed μPad was tested using the laboratory prepared solutions shown in Table 1. All solutions listed in Table 1 were prepared with a hardness of 2 ppm (as calcium carbonate) and buffered at pH = 7. A volume of 50 μl from each solution was tested as explained above.

Results and Discussion
The Whatman No. 01 was used as the base and commercially available hydrophobic paint was used to fabricate the paper-based device. The selection of these materials reduces the cost of a device. The electrochemically inactive phenol-formaldehyde polymer was used as the electrode matrix for both working and counter electrodes. Impedimetric paper supported immunosensors [10] and graphene paper impedimetric sensors [11] are previously reported. However, EIS detection using a modified electrode fabricated on paper-based devices is never been reported previously. An image of a developed microfluidic paper-based device (μPad) is shown in Figure 1. The device only requires 50 μL of sample for the detection of metal ion concertation. Compared to the other water hardness detection methods, the use of smaller sample volumes in these proposed methods reduces the use of chemical resources. This can be considered as a major advantage in this method. Compared to the other electrochemical [12] metal ion sensing methods, the proposed paper-based method is simple, small in size and inexpensive (cost per device = 0.1 USD).
Peak current variation with the concentration of K4Fe(CN)6 graph is shown in Figure 2. The proposed μPad is capable of producing a linear relationship between the concentration of K4Fe(CN)6 and the peak current. The developed μPad is capable of functioning as a typical electrochemical setup under standard electrochemical conditions.
. All the electrochemical impedance spectrums were collected keeping the bias potential of the working electrode kept at 0 V to minimize the ion migration due to the applied electric field. The typical electrochemical impedance spectrums collected using the paper-based tool with EDTA modified electrode for Fe 3+ ion solutions with the concentrations of 0.2, 0.4, 0.6, 0.8, 1.0, and 1.2 ppm are shown in Figure 3. The charge transfer resistance (Rct) and the highest negative imaginary impedance (highest -Z") of 0.2, 0.4, 0.6, 0.8, 1.0, and 1.2 ppm Fe 3+ solutions were recorded. These values were compared with the values obtained for the same set of solutions using an unmodified electrode. Both highest -Z" and Rct recorded using μPad with EDTA modified electrode showed greater responses and produced a liner variation against Fe 3+ concertation. Since the highest -Z" showed a linear relationship with Fe3+ concentration, the highest -Z" responses of 0.2, 0.4, 0.6, 0.8 1.0 and 1.2 ppm Fe 3+ were used to build the correlation between Fe 3+ concentration and the highest -Z" of the Nyquist plot constructed for each Fe 3+ concentration. The calibration graph for these Fe(III) concentrations is shown in Figure 4. Electrodes modified with EDTA can chelate with the metal ions in the solution. Chelation of metal ions increases the charge density of the electrode double layer of the working electrode. The accumulation of Fe 3+ ions on the modified electrode surface increases the impedance (-Z") and the Rct of the double layer compared to unmodified electrodes. Electrode interfacial charge-transfer rate between the solution and the electrode is controlled by the Rct. The proposed μPad with the modified electrode is capable of producing a change in the impedance in the presence of metal ions that can chelate with EDTA. Both Ca 2+ and Mg 2+ chelate with EDTA and these two metal ions are responsible for water hardness. Hardness in water can be estimated by determining the total Ca 2+ and Mg 2+ ion concentrations. Both these group 2 metal ions can chelate with the EDTA modified working electrode of the proposed μPad. The suitability of the proposed μPad to determine water hardess was evaluated. Rct and highest -Z" produced by the μPad with Ca 2+ and Mg 2+ solutions of varying concentrations buffered at pH = 7 are shown in Figures 5 and 6 Both Rct and highest -Z" produced linear correlations with Ca 2+ and Mg 2+ concentrations. The increase observed in the highest -Z" and the Rct is due to the accumulation of metal ions on the working electrode surface. Rct and -Z" responses of Mg 2+ solutions with the EDTA modified electrode were greater than those of Ca 2+ solutions, due to the greater charge density of Mg 2+ ions when chelated with EDTA.
The ability of the proposed device for the determination of water hardness when both Ca 2+ and Mg 2+ are available was tested using laboratory samples containing Ca 2+ and Mg 2+ . These solutions were prepared with a hardness of 2 ppm (as calcium carbonate hardness). The hardness of each solution was determined using the proposed μPad device and the recorded highest -Z" and Rct and values are shown in Figure 7.   Even though all five samples had a hardness of 2 ppm (as calcium hardness), the recorded Rct and highest -Z" values showed variations with the amounts of Ca and Mg ions in the solution. Hence the tool is not suitable for the determination of total hardness when both Ca and Mg are present. However, when both Ca and Mg are present, the device can be used as a semi-quantitative tool to determine whether the total hardness is larger than a pre-decided threshold value (priory previously decided -Z" value can be used as the threshold value). Advantages of such determination using the proposed device are; determination can be performed within a short period and no prior calibration is required. Further, only a minute volume (50 μl) is required for a single analysis. Based on these results, it can be observed that both Rct and highest -Z" produced similar correlations with the tested metal ion concentrations. This justifies the use of the highest -Z" in Nyquist plots to correlate directly to the total concentration of metal ions that can chelate with the EDTA-modified electrode surface. The hardness of samples contacting either Ca or Mg, with no significant amounts of other metal ions that can chelate with EDTA, can be estimated with this proposed μPad.

Conclusion
The proposed impedimetric μPad can be used as a sensor for the detection of Ca 2+ and Mg 2+ with the detection limits of 0.31 and 0.24 ppm, respectively. The device can be used to detect a single metal ion, however, the proposed impedimetric μPad is not ideal to detect the total Ca 2+ and Mg 2+ metal ion concentration of a solution. The device can be used successfully for the qualitative detection of hardness. .