Reagent Free Electrochemical-Based Detection of Silver Ions at Interdigitated Micro Electrodes Using in Situ pH Control

Silver ions, the most toxic form of silver, can be present in drinking water due to their release from silver nanoparticles which are widely used in consumer products. Due to their adverse health effects, a quick portable approach for detection in drinking water is needed. Herein we report on the development of an electrochemical sensor for silver ions detection in tap water using linear sweep voltammetry with in situ pH control; enabled by closely space interdigitated electrode arrays. The in situ pH control approach, allows the pH of a test solution to be tailored to pH 3 thereby eliminating the current need for acid addition. A calibration curve between 0.2 - 10 µM was established for silver detection in sodium acetate when 1.25 V and 1.65 V was applied at the protonator electrode during deposition and stripping, respectively, as a proof of concept study. For the final application in tap water, 1.65 V was applied at the protonator electrode during deposition and stripping. The chlorine ions, present in tap water as a consequence of the disinfection process, facilitated the silver detection and no additional electrolyte had to be added. Combination of complexation of silver ions with chlorine coupled with in situ pH control resulted in linear calibration range between 0.25 and 2 µM in tap water without the need for acidification. Abstract Silver ions, the most toxic form of silver, can be present in drinking water due to their release from silver nanoparticles which are widely used in consumer products. Due to their adverse health effects, a quick portable approach for detection in drinking water is needed. Herein we report on the development of an electrochemical sensor for silver ions detection in tap water using linear sweep voltammetry with in situ pH control; enabled by closely space interdigitated electrode arrays. The in situ pH control approach, allows the pH of a test solution to be tailored to pH 3 thereby eliminating the current need for acid addition. A calibration curve between 0.2 - 10 µM was established for silver detection in sodium acetate when 1.25 V and 1.65 V was applied at the protonator electrode during deposition and stripping, respectively, as a proof of concept study. For the final application in tap water, 1.65 V was applied at the protonator electrode during deposition and stripping. The chlorine ions, present in tap water as a consequence of the disinfection process, facilitated the silver detection and no additional electrolyte had to be added. Combination of complexation of silver ions with chlorine coupled with in situ pH control resulted in linear calibration range between 0.25 and 2 µM in tap water without the need of acidification.

Interdigitated Micro Electrodes Using in Situ pH Control. ChemRxiv. Preprint. https://doi.org/10.26434/chemrxiv.12631727.v1 Silver ions, the most toxic form of silver, can be present in drinking water due to their release from silver nanoparticles which are widely used in consumer products. Due to their adverse health effects, a quick portable approach for detection in drinking water is needed. Herein we report on the development of an electrochemical sensor for silver ions detection in tap water using linear sweep voltammetry with in situ pH control; enabled by closely space interdigitated electrode arrays. The in situ pH control approach, allows the pH of a test solution to be tailored to pH 3 thereby eliminating the current need for acid addition. A calibration curve between 0.2 -10 µM was established for silver detection in sodium acetate when 1.25 V and 1.65 V was applied at the protonator electrode during deposition and stripping, respectively, as a proof of concept study.
For the final application in tap water, 1.65 V was applied at the protonator electrode during deposition and stripping. The chlorine ions, present in tap water as a consequence of the disinfection process, facilitated the silver detection and no additional electrolyte had to be added. Combination of complexation of silver ions with chlorine coupled with in situ pH control resulted in linear calibration range between 0.25 and 2 µM in tap water without the need for acidification.

Introduction
Silver nanoparticles have become ubiquitous in a wide variety of products ranging from electronic & medical devices, textiles, cosmetics through to home disinfectants, due to their antimicrobial effect and physical properties [1]. Their increased use in consumer products has, however, resulted in their unwanted release into the environment, particularly into water sources [2]. Dumont, Johnson [3] have reported the "predicted levels" of silver nanoparticles present in surface water across the European countries. It was observed that the highest levels occur in the areas with a high population density; especially in East & Central Europe, Germany, England and around Madrid in Spain.
The antimicrobial activity of silver nanoparticles involves the slow release of silver ions into solution, which is the most toxic form of silver [4]. Silver toxicity to aquatic life has been well documented, while bioaccumulation in humans may lead to a disease called argyria [5]. Despite the lack of robust data on silver toxicity in humans, the World Health Organisation (WHO) have, suggested 0.1 mg/L (~ 0.93 µM) as upper limits for silver in drinking water. It has been suggested that higher concentrations would constitute health risks with long-term consumption [6]. Accordingly, in recent Drinking Water Standards and Health Advisories Tables, the United States Environmental Protection Agency (EPA) have proposed the same permissible (0.1 mg/L) concentration of silver in drinking water [7]. There is therefore a need, on health grounds, for rapid methods for monitoring of silver concentrations in drinking water.
Several instrumental and non-instrumental methods for detection of silver in aqueous solutions have been described to date. These include: atomic absorbance spectroscopy, colorimetric, fluorescent, and electrochemical methods [8][9][10][11]. Recently, much attention has focused on electrochemistry as a detection technique, due to its low cost, suitability to device miniaturization & portability as well as simplicity of use [12][13][14]. Anodic sweep voltammetry is a popular electrochemical technique that have been employed for silver detection with various authors reporting very low limits of detection. Typically, carbon paste or glassy carbon electrodes modified with a variety of different ligands, such as, for example: N,N0-bis(2hydroxybenzylidene)-2,20(aminophenylthio) ethane [15]; CNT and (E)-4-(2hydroxyethylimino) pentan-2-one (EHPO) [16]; phenylthiourea-functionalized high ordered nanoporous silica gel [17] or 8-Mercaptoquinoline [18], are employed. The principal role of these ligands is to pre-concentrate the silver ions at the surface of an electrode and therefore increase its overall sensitivity [19]. However, the modification of the electrodes is laborious, of variable quality and reproducibility, and may be expensive.
As a first step, experimental conditions including the buffer, pH or deposition potential must be optimised. Optimising the pH of both the metal deposition and stripping steps is crucial for the development of an electrochemical based detection method [20,21]. The optimal concentration of H + ions can delay side reactions, e.g., a metal's complexation with other species and thus increase the availability of the metal for electrodeposition and consequently the final measured signal [22]. However, in many deployment scenarios, for real-time silver detection, pH adjustment of a solution prior to a measurement remains unfeasible. The optimal pH for silver detection is strongly affected by: electrode material composition, type of ligand, modification process, and the supporting electrolyte. The optimal pH for silver detection, reported by other authors, varied between pH 1.1 and 9.5 [15,23,24] depending on the parameters used. Tap water pH, typically varies between 6.5 and 8.5, thus prior to detection in real samples, reagents for example nitric acid or acetic acid are usually added [17,25]. Ideally, an electrochemical-based sensor for tap water should allow for detection and automatically correct within this pH range without manual adjustment.
Mineral acids are usually used to adjust the pH. However, more recently, an electrochemical based in situ pH adjustment method, using a boron-doped ring disc electrode system, was demonstrated for detection of mercury in water [22]. This approach was based on electrochemical driven decomposition of water achieved by applying a sufficiently high oxidising potential to the ring electrode. During this process, hydrogen ions were anodically produced at the ring electrode which caused acidification of the solution in the region of the 'sensing' disk electrode. In this work, we extend this approach by developing solid-state sensors on silicon chip substrates that incorporate interdigitated microelectrodes for detection of silver ions in tap water. Each sensor comprised two interdigitated electrodes arrays (IDA), one of gold and one of platinum. In our approach, the platinum IDA was used to electrogenerating H + ions i.e. a protonator electrode (for local pH control), while the other gold IDA was used to detect the silver (I) ions. Each chip contained six different sensors, each comprising two IDA microband electrodes. Using this approach, electrochemical based in situ pH control eliminated the requirement for addition of reagents, e.g., nitric acid, prior to measurement. By applying a constant potential to a pronator electrode, the in situ pH surrounding an electrode could be easily tailored within a range of pH 2-10. In addition, complexation of silver with chlorine ions, present in tap water following the disinfection process, created significantly sharper peaks that aide detection [23]. Finally, due to their low RC constants, the use of ultra microelectrode systems eliminated the need for additional electrolytes to be added to solution.

Apparatus
All the electrochemical measurements were undertaken using a CHI 920 potentiostat with a bipotentiostat function. A three electrode cell configuration was used for the silver detection in acidified solutions while a four electrode bipotentiostat cell configuration was used for silver detection employing in-situ electrochemical pH control. Gold interdigitated microband electrodes were employed as a working electrode and while platinum modified electrodes were used as protonator electrode, gold on-chip wire as a counter electrode and an external Ag/AgCl as a reference electrode.

Silicon chips fabrication
Interdigitated electrodes array (IDA) electrodes at silicon chips were designed and fabricated for silver detection. Fabrication of the chips were similar to those described by Wahl, Barry [26], [27]. Briefly, gold microband electrodes were fabricated on four-inch silicon wafer substrates bearing a ~300 nm layer of thermally grown silicon dioxide; see (C). As platinum is known to catalyse water electrolysis effectively, platinum was electrodeposited as a protonator on one IDA to promote proton flux production at lower over potentials. Platinum deposition was undertaken by immersing a chip with electrical connection made to one interdigitated comb in DNS plating solution and applying -0.5V (Vs Ag/AgCl) for 6 seconds.

2.4.Electrode characterisation
Following fabrication, optical microscopy was employed to identify any obvious defects or faults with faulty chips being discarded. Chips were cleaned by immersion and sonication for ten minutes, first in ethanol, then in de-ionized water, and dried in a flow of nitrogen.
Electrochemical characterisation was undertaken using a CHI 920 potentiostat. Cyclic voltammograms (CV) were performed from 0 V to 0.6 V at 100 mV/s in 1 mM ferrocene carboxylic acid (FCA). Generator-collector scans where the protonator IDA were held at 0 V and the working electrode swept as above, were also undertaken. All electrochemical characterisation measurements were recorded versus an Ag/AgCl external reference electrode.

2.5.Silver detection method
(iii) Finally, square wave stripping voltammetry was then used to strip the silver from the underlying gold electrode (see equation 2) with the corresponding oxidation peak current proportional to concentration of silver pre-deposited on a sensor.
In brief, the method comprised the following steps: a) acidification of the test solution either using (i) addition of mineral acid or (ii) in situ electrochemical based pH control. (b) Applying a bias of -0.2 V (vs Ag/AgCl) for 2-5 minutes (depending on the medium used, i.e. buffer or tap water) to reduce the silver ions at the working electrode. c) Square wave stripping voltammetry was recorded between -0.2 and 0.8 V vs Ag/AgCl, using following conditions: frequency 15 Hz, increment potential 0.004 V, amplitude 0.025 V and the silver strip peak was recorded. d) After the measurement, the electrode was potientio-dynamically cleaned by replacing the sample with 0.01 M sodium acetate solution and applying the potential of 0.5 V for 150 s, which is a slightly more oxidative potential than observed for silver stripping in sodium acetate. e) Square wave voltammetry was recorded between -0.2 and 0.8 V in blank, sodium acetate solution to confirm that all the silver was oxidised from the electrode and prevent carry over from previous experiments.

2.6.Silver detection method using pH control
For the silver detection using in-situ electrochemical pH control in tap water, the same conditions for deposition and stripping as described above were applied. In addition, a constant oxidising potential was applied to the protonator electrode, to produce protons (H + ions) according to equation 3, thus tailoring the pH in the vicinity of the sensor IDA.

Results and discussions
3.1.Sensor Characterisation.  Sensors were then characterised in Generator-Collector mode undertaken with the protonator IDA held at 0 V and the working IDA swept from 0.0 to 0.6 V, at 50 mV/s. A typical voltammogram is presented in Fig. 3 (B). In this approach, the generator IDA first oxidised FCA to FCA + species which then diffused across the gap to the collector electrode where it was subsequently reduced back to FCA establishing redox cycling; resulting in higher measured currents. As a result of redox cycling, the Generator-Collector voltammograms exhibited a quasi steady state, time independent behaviour typically associated with ultramicroelectrodes [27]. The generator IDA voltammogram shows a maximum current of 31 nA, which is significantly higher comparing to a single CV mode (~10 nA). The voltammogram also exhibited a small hysteresis in current between the forward and reverse scans. This hysteresis arose form an additional capacitance current component arising from the high scan rate applied. The collector IDA voltammogram exhibited a lower maximum current of -26 nA when compared to the protonator IDA. The difference between the two measured currents was due to some oxidised FCA + diffusing away from the sensor into the bulk solution.
Consequently, the collection efficiency of the sensor, which is a ratio of the collector to the generator currents, was determined to be ~84%. This thus suggested that 84% of protons, produced at a protonator IDA, would diffuse to the working IDA, and tailor the pH as desired.
No hysteresis in current between the forward and reverse scans was observed at the collector IDA, as it was held at a constant potential throughout the experiment. As a result, a much lower capacitive current component can be observed, which dissipated quickly after initial biasing.

Deposition optimisation
The influence of solution pH on silver deposition was evaluated by varying buffer pH between pH 2 and pH 4.5. 1 µM AgNO3 dissolved in 10 mM sodium acetate at different pH was deposited at the working electrode at -0.2 V for 3 minutes. Following deposition, the silver was then stripped using linear sweep voltammetry and the peak current was recorded. In Error! Reference source not found.. (A), representative stripping current peak heights measured for silver stripping at different pHs are presented. The measurements were undertaken in duplicate for each pH and the average value with standard deviation were plotted. Decreasing the pH from 4.5 led to a change in the measured stripping current with the maximum stripping peak current found to be between pH 2.5 and 3.0. The corresponding deposition currents are presented in Fig. S1. (A) with a change in deposition current magnitude observed with different pHs. The highest deposition currents (measured at 100 seconds) were observed for pH 3, followed by pH 2 and 3.5, respectively. pH 3 was selected as the pH of choice for further experiments as the most optimal pH for silver detection.

V was applied for 3 minutes. Dotted line is a guide to the eye only. (B) Influence of deposition potential on measured stripping current (peak height) for 1 µM of AgNO3 in 10 mM sodium acetate, pH 3, when deposition done for 3 minutes. Dotted line is a guide to the eye only.
Following optimisation of pH, the influence of reduction potential on the peak height was evaluated and optimised. A series of silver electro-reduction voltages were assessed by first electro-depositing silver at a selected voltage followed by a square wave stripping voltammetry where the peak current was measured. Electro-reduction was undertaken in the voltage range of -0.1 V and -0.5 V. It can be seen in Error! Reference source not found.4. (B) a maximum peak current was observed when using a reduction potential of -0.2 V. Although higher currents were measured at more cathodic potentials, this increase arose from superimposition of an oxygen reduction signal onto the electrodeposition current; see Fig. S1. (B). To this end, -0.2 V was selected as the optimal for silver detection and used in the further studies.
The duration used for silver ion deposition depended on the expected concentration of silver in solution. In theory, the longer the deposition time, the lower limit of detection for silver detection can be achieved. However, this must be offset and balanced by an electrode potentially becoming saturated at higher silver concentrations during longer deposition times.
A series of deposition times were explored.

In situ electrochemical pH control: potential selection
As discussed previously, the pH of the solution greatly influenced the silver deposition process.
Typically, nitric acid would be manually added to a sample to adjust the solution pH to create the optimal conditions for metal detection. However, this step requires trained personnel, proper protective equipment as well as special storage/carriage requirements for concentrated acids. Therefore, this approach is not very suitable for real time detection outside of the laboratory.
Electro-generated in situ pH control was explored, to eliminate the requirement of sample acidification prior to analysis. By applying a constant oxidizing potential to the protonator electrode, sufficient hydrogen ions (protons) can be produced to tailor the local pH surrounding an electrode. First, the right potential to be applied at the protonator electrode was selected. It is well known that when undertaking cyclic voltammetry at a gold electrode, the positions of the gold oxide and reduction peaks vary, depending on the solution's pH [28]. In this manner, the voltage at which the gold oxide reduction peak maximum occurs, may be used as an indicative measure of the pH of the solution at a sensor. To confirm this, cyclic voltammetry in 10 mM sodium acetate in the voltage range of 0.2 to 1.2 V was performed at pH 3 and pH 7.5 (the selected pH for silver detection and pH of sodium acetate without acidification, respectively). The CVs are presented in Fig. 6. (A). At pH 7.5 the gold oxide reduction peak maximum was observed at 0.55 V. On addition of nitric acid to acidify the sodium acetate buffer to pH 3, the gold oxide reduction peak moved to 0.76 V, as expected. The decrease in the reduction peak area arose from a limited amount of gold oxide formed due to the narrow potential window used; as the gold oxidation process would also have shifted to higher anodic voltages at this lower pH value. On increasing the protonator potential, the gold oxide reduction peak was observed to move to higher anodic voltages consistent with a decrease in pH. In these experiments, an applied potential of 1.65 V yielded a maximum gold reduction peak at 0.76 V; which was the same location as the acidified pH 3 sodium acetate as showed by the dashed line in Fig. 6. (B).
Although the acetate buffer has a solution pH of 7.5 in the bulk away from the electrode, the local in situ pH in the vicinity of the electrode was electrochemically tailored to pH 3.0 using this approach. Thus an applied voltage of 1.65 V was selected as the protonator voltage of choice.

In situ pH control: Acetate buffer
The initial experiments to assess and optimise in situ pH control were undertaken in 10 mM sodium acetate, as described previously. However, as shown in Fig. 7. (A), when 1.65 V was applied to a protonator IDA and a cathodic potential of -0.2 V was applied to the working IDA (during silver electrodeposition), the stripping peak for silver was significantly lower when compared to the stripping peak obtained using chemically acidified silver solutions. In addition, the measured current during deposition was significantly higher when using pH control (~ -40 nA at 100 s) compared to the deposition current in chemically acidified sodium acetate (~ -0.6 nA at 100 s); see Fig. S4. To explain this observation, we believe that this discrepancy may have arisen from parallel competitive electrochemical processes occurring at the sensors electrode in sodium acetate: namely silver electrodeposition and direct oxygen reduction. To explore this further, CVs were undertaken in 0.01 M sodium acetate using pristine gold working electrodes both with and without the protonator IDA biased at +1.65 V. The corresponding voltammograms are presented in Fig. S3. A significant increase in anodic current is observed at ~ 1.2 V without the pH control and ~ 1.5 V when in situ pH control (i.e., biasing the protonator) was applied. This current increase corresponds to the formation of a gold oxide layer at the gold IDA, which is in accordance with Burke and Nugent [28]. This suggests that setting the protonator IDA to 1.65 V during silver detection could lead to the production of a significant amount of molecular oxygen at the protonator. This argument is strongly supported by the significant current increase corresponding to gold oxide reduction at potential ~ 0.6 V when pH control was applied. The cyclic voltammogram in sodium acetate, with applied pH control, also exhibits an increase of cathodic current starting around -0.2 V corresponding to the potential at which the oxygen begins to be reduced in the acidic conditions [29]. Consequently, applying a bias of 1.65 V to a protonator IDA, results in the generation of molecular oxygen which diffuse to, and is reduced at, a sensor IDA biased at -0.2 V. This can explain the observed increase in the magnitude of deposition current (Fig. S4.). Concerning analyte mass transfer transport, diffusion of silver ions to a working electrode is diffusion limited, as shown in Fig. 3 (A). However, diffusion of oxygen, generated along the entire length of a protonator IDA, to a sensing IDA will be radial in nature and will thus be more efficient.
These competitive processes may thus result in the lower concentrations of silver deposited at the working IDA even though the observed deposition was higher Consequently, a trade off was required between applied protonator voltage versus the associated molecular oxygen formation. To this end, 1 µM AgNO3 dissolved in 10 mM sodium acetate at pH 7.5 was deposited at a working IDA while different potentials, varying from 1.25 V to 1.65 V, were applied to protonator IDA during deposition, followed by SWV, see Fig. 7.
(A). It can be observed that the lowest applied voltage of 1.25 V during deposition yielded the maximum stripping peak current and thus was selected for further experiments. The potential applied to a protonator IDA during stripping of silver electrodeposited at the working IDA was also optimised. It can be seen in Fig. 7. (B) that resetting the protonator voltage back to 1.65 V during SWV resulted in the highest stripping peak. This further supports the thesis that molecular oxygen formation during silver deposition was an interferent during deposition, as the applied anodic potentials during stripping were too positive to enable oxygen reduction.

In situ pH control: Tap Water
Having showed a proof of concept of silver detection using pH control in sodium acetate buffer as described above, the developed method was applied and optimised for potable tap water.
Typically, pH of drinking water varies between 6.5 to 8.5 [30]. and may need to be adjusted and optimised prior to silver detection, i.e., pH 3. CVs was undertaken in tap water with 1.65 V applied to the protonator IDA; see Fig. S5. It was observed that oxygen reduction in tap water requires a more cathodic potential (~ -0.5 V) compared to sodium acetate (~ -0.2 V) which suggested that the competitive process, described above, should not significantly interfere with silver deposition. To this end, in situ pH control was undertaken by, applying the desired 1.65 V to the protonator during both silver deposition and stripping steps.
For the optimisation process, tap water was spiked with 1 µM AgNO3 required as silver was not present in tap-water in our lab. No additional supporting electrolyte was added to test solutions in order to verify that the sensor could be used for a real time tap water detection. Fig. 9.Error! Reference source not found. shows the stripping peak SWVs in tap water both with and without pH control. The silver stripping peaks shifted to more cathodic values, and the peaks became higher and sharper when compared to measurement undertaken in sodium acetate buffer. The silver stripping peaks measured in sodium acetate with and without pH control are included in Fig. 9. for comparison. We believe, that the cathodic shift and change in the voltammetric profile are attributed to chlorine, typically present in tap water as a sterilising agent. Therefore, the stripping peak in tap water corresponded to AgCl -, instead of Ag + recorded in sodium acetate, following the mechanism previously described by Saterlay, Marken [23]: It is well known that chlorine ions are very reactive, especially with regard to silver. The process of a peak sharpening upon complexation of silver with chlorine, as well as its beneficial analytical applications, has been previously described by Saterlay, Marken [23]. In their work, chlorine in the form of 12.5 mM KCl was added as an electrolyte reagent to test solutions to improve the signal results for ASV. Other authors have also reported chlorine additions, e.g, using 0.1 M HCl as an electrolyte for the optimisation of silver detection [16,31]. In this work, chlorine was naturally present in tap water thus, no additional reagent had to be added to enhance silver detection in tap water as chlorine remains the most commonly used water disinfectant worldwide [32]. Despite the form of chlorine used (gaseous chlorine, calcium hypochlorite or sodium hypochlorite), a combination of hypochlorous acid (HOCl) and hypochlorite ion (OCl -) will be present in tap water as "free available chlorine" in line with statutory regulations. In acidic conditions, the HOCl species dominates, which is a more reactive form of chlorine [33,34].  (between 0.25 and 2 µM) in tap water, using in situ electro-generated pH control, is presented in Fig. 10. The data points represent the mean value of three replicate measurements, with the error bars representing 1 standard deviation. The calibration curve exhibited excellent with R 2 of 0.992. The EPA suggested limit of 0.93 M fits well within this linear dynamic range. Fig.   10.: inset shows the associated stripping voltammograms for the different concentrations. Although several authors have reported electrochemical approaches for silver detection in real water samples, most of these reports had to (i) first acidify the solution using an acid or (ii) add additional electrolyte prior to performing the measurements [15][16][17]35]. To the best of our knowledge, none of these papers to date have reported silver detection using in situ pH control.
Moreover, our results have suggested that silver detection efficiency was improved with the chlorine ions present in tap water. Based on Saterlay, Marken [23] the intensity of AgCl peak will depend on the concentration of chlorine in water which they have showed by undertaking measurements with different concentration of KCl. This suggests that the concentration of chlorine in the tap water could be a limiting factor for the presented technique. Based on WHO guidelines, chlorine is present in most disinfected tap water at a concentration between 0.2 mg/L and 1 mg/L [6]. This means there would be sufficient amount of chlorine present to allow detection of the target silver concentrations (0.1 mg/L). Using this approach and the sensors developed herein, silver detection may be undertaken in previously chlorinated tap water without addition of electrolyte, acid or base with a prior calibration is done depending on the chlorine concentration.

Conclusions
We present an easy and quick technique that employs IDA electrodes for silver detection in tap water using a electrogenerated in situ pH control method. Silver detection was undertaken using square wave voltammetry at a working IDA with simultaneous production of hydrogen ions at a protonator IDA which allowed the pH to be tailored in the vicinity of the sensor. In addition, complexation of the silver ions with chlorine enabled more sensitive detection and faster time-to-result without the need of electrolytes addition. The sensors have the potential to be deployed for real time detection in water utility systems as well as in estuarine or marine waters with the need for pre conditioning of a sample.
download file view on ChemRxiv 2020.07.02 Reagent free detection of silver ions using squ... (1.70 MiB) Supporting information for Reagent free electrochemical-based detection of silver ions at interdigitated micro electrodes using in situ pH control.