This article is licensed under CC-BY 4.0 pubs.acs.org/estwater Article Feasible H2S Sensing in Water with a Printed Amperometric Microsensor Franc Paré, Rebeca Castro, David Gabriel, Xavier Guimera,̀ Gemma Gabriel, and Mireia Baeza* Cite This: ACS EST Water 2023, 3, 1116−1125 Read Online ACCESS Metrics & More Article Recommendations * sı Supporting Information ABSTRACT: Concern over pollution has led to an increase in wastewater treatment systems, which require constant monitorization. In particular, hydrogen sulfide (H2S) is a toxic gas, soluble in water, commonly found in industrial and urban effluents. For proper removal control, fast, durable, and easy-to-handle analytical systems, capable of on-line measurements, such as electrochemical sensors, are required. Moreover, for a proper monitoring of said treatment processes, analysis must be carried out through all steps, thus needing for an economic and highly reproducible method of sensor fabrication. Digital printing have risen in the last few years as technologies capable of mass producing miniaturized electronical devices, allowing for the fabrication of amperometric sensors. Here, a 2 mm2 graphite (Gr) electrode, modified with different dispersions of single-walled carbon nanotubes (SWCNTs), poly(vinyl alcohol), poly(diallyl dimethylammonium chloride), and polylactic acid (PLA), is presented as a H2S sensor. SWCNTs allow for lower oxidation potentials, higher sensitivity, and a reduced rate of sulfur poisoning, while polymer dispersion of PLA increases mechanical stability and as a result, electrochemical performance. This microsensor presents an optimal pH working range between 7.5 and 11.0, a limit of detection of 4.3 μM, and the capacity to operate on complex matrices for H2S contamination detection. KEYWORDS: amperometric sensor, hydrogen sulfide, inkjet-printed electrodes, single-walled carbon nanotubes, direct ink writing ■ INTRODUCTION H2S is a corrosive and extremely toxic gas, perceptible at very development of biotechnological H2S removal processes.10 However, these treatment processes require adequate tracking low concentrations as rotten egg smell.1−3 H2S is naturally systems that can be applied for complex sample measurement. emitted from several volcanic and geothermal areas of the A variety of methods for H2S determination have been world, but it can also be generated in oil and gas production as established, including chromatography, fluorescence, and well as refining industries.4−6 Above concentrations of 100 optical measurements.11−14 However, electrochemical sensors ppm, H2S generates saturation of olfactory nerves and poses provide a suitable platform due to its rapid detection, high health risks including respiratory paralysis, irritation, and sensitivity, and potential application for on-line measure- shock. Above 500 ppm, it has fatal consequences since it affects oxygen uptake in blood.2 Furthermore, its exposure is related ments.7,9 In addition, electrochemical sensors have the to Alzheimer’s disease, traumatic brain injury, and hyper- tension.7 Received: November 18, 2022 While H2S frequently appears as gas, it co-exist as H2S, HS−, Revised: March 14, 2023 and S2− in aqueous media where the percentage fraction of Accepted: March 21, 2023 species is related to medium pH (Figure S1).8,9 Due to its Published: April 4, 2023 poisonous nature, environmental hazardous effects, and toxicity for human health, many efforts are focused on © 2023 American Chemical Society https://doi.org/10.1021/acsestwater.2c00589 1116 ACS EST Water 2023, 3, 1116−1125 ACS ES&T Water pubs.acs.org/estwater Article advantage of using H2S oxidation at low potential for fabrication processes. Furthermore, the final dispersion measuring, allowing for a higher selectivity in complex samples. selected for ink preparation was used for the H2S microsensor In recent times, printed electronics have gradually replaced fabrication. Even though metal electrodes, such as Au, are a the conventional microtechnology of electrode fabrication.15 reliable base on which to test modifications, these are Among them, direct ink writing (DIW) has generated high unsuitable for the final design as a CE made of metal would interest due to advantages such as low cost, reduced still get sulfur poisoning. The H2S microsensor was electro- manufacturing time, high reproducibility, mass production, chemically characterized, and its pH range of application was and feasible miniature designs.16,17 These features make DIW a evaluated. Finally, its validation was performed using real suitable alternative for miniaturized electrochemical sensors’ samples of a sulfate-reducing bioreactor and H2S production fabrication. Moreover, in contrast to other printing technolo- was measured. gies, DIW has a small number of fabrication steps, including substrate and ink preparation, in addition to the printer setup. DIW can be performed under room conditions; thus, a wide ■ MATERIALS AND METHODS Reagents and Chemicals. All chemicals were commer- range of substrates are compatible for printing, including cially available and were used as received. flexible polymers.16,18 In addition, it does not require masks or For the construction of the Gr integrated electrodes, a direct contact between the ink ejector (nozzle, needle, etc.) screen-printing silver microparticle ink (DuPont 5029 from and the substrate for the printing process, leading to an easy Dupont, USA) for the conductive tracks, Gr ink (C2030519P4 deposition of different materials for multilayer platform from Gwent Electronic Materials, UK) for the WE and counter fabrication.17,19 Moreover, in biotechnological applications, electrode (CE), and a commercial Ag/AgCl ink (DuPont 5874 microsensors can be printed on different biocompatible from Dupont, USA) for the reference electrode (RE) were materials and their designs can be adapted to the shape of utilized, and a dielectric ink (LOCTITE EDAG PP 455 BC reactors without substantial modifications.20 Therefore, DIW is from Henkel Ibérica, ES) was included for passivation. All inks a promising alternative for H2S microsensor fabrication for were printed over a 125 μm-thick polyethylene terephthalate tracking applications. (PET) sheet (Q65HA from DuPont Teijin Films, USA). Regarding materials used for H2S sensor fabrication that can For the Gr-WE modification, an ink was prepared from a be implemented as inks for any of the DIW processes, one- dispersion of single-walled carbon nanotube carboxylic acid dimensional carbonaceous nanomaterials such as single-walled functionalized (SWCNT-COOH), sodium dodecyl sulfate carbon nanotubes (SWCNTs) have attracted attention due to (SDS), PVA, PDDA, and PLA, all purchased from Sigma- high chemical and thermal stability. This material has a large Aldrich and anhydrous glycerol (>98%, Honeywell Fluka, surface ratio and many active sites for H2S oxidation.9,21,22 In USA). For the RE membrane, an ink was prepared by addition, good conductivity, charge transference, and great dissolving 10 wt % polyvinyl butyral (PVB) from Sigma- mechanical strength are SWCNT features that are essential for Aldrich, saturated sodium chloride (NaCl, Scharlab, ES) in highly sensitive electrode fabrication.23 methanol (Sigma-Aldrich). Furthermore, polymers such as polyvinyl alcohol (PVA), Sodium sulfide (98%, Na2S·9H2O), phosphate buffer saline poly(diallyl dimethylammonium chloride) (PDDA), and (PBS) (0.01 M KH2PO4/Na2HPO4, 0.0027 M NaCl, 0.137 M polylactic acid (PLA) are feasible materials for ink preparation KCl, 7.2−7.6 pH), potassium ferricyanide (99%, K3(Fe- concerning the improvement of its mechanical stability and (CN)6)), potassium ferrocyanide trihydrate (99%, K4(Fe- dispersion. PVA is a polyhydroxy-type polymer, which has (CN)6)·3H2O), and potassium chloride (98%, KCl) were all hydrophilic, non-toxicity, and biodegradability properties.23,24 purchased from Sigma-Aldrich. Sodium hydroxide (99%, It can be prepared easily, has a high optical transparency, and NaOH) was acquired from Alfa Aesar (ES). For H2S stock has an excellent film-forming capacity.23,25,26 On the other solution standardization, a Pb2+ (0.1 M, standard solution) hand, PDDA is a water-soluble cationic polyelectrolyte, which from Thermo Scientific Orion, USA, was used. All solutions can offer anions as charge carriers.27 Because of the cyclic were prepared with deionized water for the Milli-Q system quaternary ammonium structure in PDDA, it has an excellent (Millipore, Billerica, MA, USA). chemical stability and plays an important role in nanoparticle Instrumentation. A DIW digital material depositor dispersion; thus, it can be used as a combination agent.24,25,28 (DMD100 from Kellenn Technologies, FR) was used for the Also, it is an environmentally friendly polymer, which is widely fabrication of bare Gr electrodes. used in industrial applications due to its easy operation. Finally, The printed devices were morphologically characterized by PLA is a biodegradable polymer, commonly used for 3D means of a digital microscope (AM4815ZTL from DinoLite, printing and biomedical applications, highly insoluble in water, NE) and a field emission-scanning electron microscope and with good mechanical properties. These properties make it (MerlinFE-SEM from Carl Zeiss, GE) with an energy- remarkably interesting for monitoring wastewater treatment dispersive X-ray spectroscopy (EDX) analysis system. Electro- processes for its resistance to erosion on aqueous samples and chemical performance was tested by using a potentiostat its low toxicity toward species present in real samples or μAutolab (PGSTAT204 from Metrohm Autolab BV, NE). complex matrices. Stock solutions of H2S were standardized using a commercial In this work, a H2S amperometric microsensor was S2− ion-selective electrode (Thermo Scientific Orion Star, developed using the cost-effective and reproducible DIW USA) coupled to a pH/ISE SB90M5 measurement system technology to achieve a sensitive, selective, and stable sensor (SympHony, USA). feasible for H2S analysis on environmental and bioreactor Real samples were compared to a SULF-10 commercial H2S samples. The optimization of the sensing ink using SWCNTs, gas sensor coupled to a X-5 UNIAMP multimeter both from PVA, PDDA, and PLA as well as each material’s ratios was Unisense, DK. studied onto a graphite (Gr) working electrode (WE) to SWCNTs/Polymer Composite Preparation. Several obtain a highly stable and conductive ink, suitable for DIW transductor inks were tested for this sensor fabrication. The 1117 https://doi.org/10.1021/acsestwater.2c00589 ACS EST Water 2023, 3, 1116−1125 ACS ES&T Water pubs.acs.org/estwater Article Figure 1. Step-by-step three-electrode integrated configuration (Gr-WE, Ag/AgCl-RE, and Gr-CE) fabrication process by a DMP printer: (a.i) Ag tracks and pads, (a.ii) Gr-WE and Gr-CE, (a.iii) Ag/AgCl-RE, (a.iv) electrode passivation using insulator dielectric, and (a.v) electrode modification by drop casting the SWCNTs/PLA transductor. Also, (b.i) pictures of the final printed platforms 1-Ag/AgCl-RE, 2-Gr-WE, and 3-Gr- CE, (b.ii) close-up of the Ag/AgCl-RE and Gr-WE, and (b.iii) modified electrodes 1-Ag/AgCl-PVBmembrane-RE and 2-Gr-SWCNTs/PLA-WE. first tryout was produced by dispersing SWCNT-COOH (1 wt All inks were then sintered at 150 °C for 1 h or until resistance %) and SDS (0.7 wt %) in deionized Milli-Q water (18.2 MΩ became lower than 100 Ω. cm). To improve the mechanical stability, polymers were Insulator Printing. A commercial ink was used to cover the added to the SWCNTs. First, PVA (0.25, 0.5, 0.75, and 1 wt tracks with an impermeable dielectric polymer. With it, the %) was added to the previous composition to test different silver tracks of the electrodes were protected from corrosion, SWCNTs/polymer ratios. Separately, PDDA (0.5 wt %) was which could lead to high resistance tracks and faulty added as PVA previously. Later, PVA (0.4 wt %) and PDDA measurements. One layer of the LOCTITE dielectric was (0.1 wt %) were mixed and added over the SWCNT printed with a PP of 30 kPA, TP of 0 kPa, and PS of 225 mm/ dispersion. Finally, a different dispersion with PLA (0.5 wt min. It was dried by curing under UV light for 30 s (Figure %) using the previously specified SWCNTs and SDS 1a.iv). composition was prepared. Dispersion was achieved by PVB Membrane Deposition. To grant the electrode better sonicating for 5 min using an ultrasound probe. Between electrochemical stability, the RE was protected with a uses, the ink was stored at low temperature (5 °C). Storing the membrane saturation of Cl−. Drops of 1 μL of PVB saturated PLA dispersions at low-temperature conditions considerably with NaCl in methanol were casted over the printed RE. This extends their lifetime. PLA is only soluble in THF, which has a guarantees a constant concentration of Cl− and, as such, very low boiling point, and is present in only 5 wt % mixed potential for the RE as well as protecting it from S2− and other possible chemical species that might attack the electrode. with water and glycerol. If THF evaporates, then PLA starts Transducer Deposition. The SWCNT-based inks were precipitating before the ink’s usage. drop-casted over the bare Gr electrode area. Drops of 1 μL Electrode Fabrication. To produce each electrode, several were used to achieve such small geometries while fully covering materials must be printed over a PET substrate. Since PET the bare electrode beneath. Moreover, transductor inks films were already pretreated for an enhanced hydrophilicity, (SWCNTs/PVA/PDDA and SWCNTs/PLA) were expected no extra treatment was required to achieve good ink adhesion. to enhance the bare electrodes’ electrochemical capabilities. All printing was carried out in a standard laboratory in ambient Finally, the ink was left to dry at room temperature, with no conditions. A platform with a three-electrode integrated further treatment applied. configuration was designed. Electrode Characterization. Modification of the Gr Three-Electrode Integrated Platform (Gr-WE, Ag/AgCl-RE, electrode was followed by optical microscopy and scanning and Gr-CE). As shown in Figure 1a.i, the first single Ag layer electron microscopy (SEM). Millimetric images allow for the was deposited using a printing pressure (PP) of 80 kPA and verification of optimal geometry, possible gaps in the printed traveling pressure (TP) of 10 kPA at a printing speed (PS) of surfaces, and long-term variations. SEM was done using a Zeiss 100 mm/min to produce the pads and tracks of the WE, RE, Merlin microscope. Samples were prepared by sticking a cut and CE. Those inks were then dried at 40 °C for 10 min. portion of the electrode onto the holder and placing an Afterward, a layer of Gr ink was deposited to form a 1 mm aluminum tape strip to help in the dissipation of the excess of diameter disc for the WE and the CE (Figure 1a.ii). This was electrons. done using a PP of 80 kPa, a TP of 20 kPA, and a PS of 80 Electrochemical performance was studied by an Autolab mm/min. They were then dried at 40 °C for 10 min. Next, the potentiostat using cyclic voltammetry (CV) and chronoamper- Ag/AgCl mixture was deposited to form the RE using a PP of ometry (CA) techniques. The external RE of Ag/AgCl (3 M 65 kPa, a TP of 10 kPa, and a PS of 80 mm/min (Figure 1a.iii). KCl inner filling) and a CE made of a platinum wire, both from 1118 https://doi.org/10.1021/acsestwater.2c00589 ACS EST Water 2023, 3, 1116−1125 ACS ES&T Water pubs.acs.org/estwater Article ItalSens, PalmSens (NE), were used to complete the system. CV in a 0.01 M [Fe(CN)6)]3−/[Fe(CN)6)]4− solution, which ■ RESULTS AND DISCUSSION Optimization of Composite Composition. Different leads to a highly reversible monoelectronic charge transfer dispersions of SWCNTs on polymers (SWCNTs/PVA, reaction, is highly used to study the electrodes’ properties. SWCNTs/PDDA, SWCNTs/PVA/PDDA, and SWCNTs/ These include the electrode’s electroactive surface, charge PLA) have been studied within this work to overcome the transfer reversibility, and exchange current. Signals were mechanical limitations that the SWCNTs present. SWCNTs recorded at a scanning speed of 0.01 V/s. are hard to disperse in aqueous solvents due to their low Sensor response was studied using CA, with no stirring while measuring. All calibrations were performed in batch con- polarizability and require a stabilizing agent. In this case, SDS ditions, with subsequent additions of 0.01 M H2S standards was used as a stabilizing agent to increase the dispersion over 25 mL of PBS adjusted with NaOH to pH 8−8.5. quality.30 However, the problem was rather with the stability To standardize H2S stock solutions, they were prepared at once the ink had dried and the SWCNTs had been deposited pH 14, as to form predominantly S2−, with a 0.1 M over a different material. Thus, the polymers were added as a concentration. Then, a potentiometric titration using a support to the SWCNTs. However, as to not decrease the standard Pb2+ solution and an ion-selective S2− electrode was electroanalytical response, the ratio of the polymer could not performed. exceed that of the SWCNTs. Therefore, the chosen SWCNTs/ Limit of detection (LD) was calculated using the relation polymer ratios tested were 1/0.25, 1/0.5, 1/0.75, and 1/1. To determine the most optimal one, several dispersions were LD = 3S b (1) drop-casted over a PET substrate and they were observed after where Sb is the standard deviation of the blank. 72 h submerged in water to test their improvement of H2S Response. The electrochemical reaction associated to printability (Figure S3) compared to the SWCNTs alone. After the measurements is 72 h, a pure SWCNT deposition is completely removed; thus, this was the minimum time to overcome the mechanical H 2S S0 + 2H+ + 2e (2) stability. 29 From the durability tests, it was decided that a SWCNTs/ which has a standard potential Eo = −0.476 V vs H2/H+ ≈ polymer ratio of 1/0.5 would work the best as higher polymer −0.080 V vs Ag/AgCl. Consequently, the reaction is ratios lead to very few SWCNTs being deposited, and lower spontaneous under standard conditions and stock solutions ones did not improve the stability enough. must be prepared daily. On one hand, PVA composites with SWCNTs were The concentration at the sensor’s surface and intensity discarded as the rapid apparition of large aggregates in the current can be related through Cottrell’s equation: prepared inks was observed. These aggregates made the nFAc j0 Dj composites easily peel off (Figure S3) and not last enough for a i= full electrochemical characterization. On the other hand, t (3) PDDA composites present their own oxidation and reduction with cj0 as the species activity at the electrode’s surface, if signals (Figure S4). The oxidation peak potential around 0.08 performed under diffusion conditions. F is Faraday’s constant, V vs Ag/AgCl has the direct inconvenient that interferes with A is the electrode’s electroactive area, n is the number of the analyte measurements as H2S is oxidized at −0.08 V vs Ag/ electrons exchanged, Dj is the diffusion coefficient of the AgCl and gradually degraded the sensor. However, the degree reacting species, and t is time. Measured at a given time, i vs c of dispersibility and stability of dispersions of the SWCNTs in results in a linear correlation. PDDA was adequate. Thus, it was decided that mixing at the The parameters for the measurements are the response time same time could be interesting to improve the ink’s properties. (tr), which is the time required to accurately carry out The SWCNTs/PVA:PDDA ratio was maintained in 1/0.5 measurements at 20 s. Also, given that random noise is always (total SWCNTs/polymer), and from all the tested dispersions, present in measurements, the best option is to average several 1/0.4/0.1 was the only ratio that both minimized the signals. Since the signal is time-dependent, the averaged values formation of aggregates and simultaneously prevented the must be taken in a short range of time to avoid reflecting this apparition of significant currents from the oxidation of PDDA. dependence. As such, a 0.1 s gap (10 values) is used for the As the initial problems with aggregates were not fully solved, chronoamperometric measurements. Finally, 0.1 V vs Ag/AgCl PVA and PPDA were substituted for PLA. The first test with was used as the working potential for the H2S measurements PLA, over a PET substrate using a 1/0.5 SWCNTs/PLA ratio, (Figure S2). yielded far better results than any previous composite in its Interference Study. To measure the possible interference durability when submerged in water for 72 h (Figure S5). of other species commonly found in water samples, different Electrochemical Performance Evaluation. We assess solutions were spiked with stocks of known concentration. All the electrochemical response improvement by modifying the solutions were measured with the same sensor consecutively bare Gr electrode with the previously tested ratios of 1/0.4/0.1 and by triplicate. SWCNTs/PVA/PDDA and 1/0.5 SWCNTs/PLA composites. Sample Preparation. Samples, both spiked and real, were The peak current, voltage separation, and the calculated prepared simultaneously to ensure minimum variability. Each electroactive area, calculated using Randles−Š evčiḱ (Table S1) was adjusted to the optimal pH for the commercial or our of the performed voltammograms (Figure 2), were used as the sensor, respectively. Spiked samples, consisting of Milli-Q and parameters of reference for comparison. tap water, were doped with H2S to known concentrations. Real As expected, the addition of SWCNTs over the bare samples come from a sulfate-reducing reactor and are expected electrode (Gr) increases the electroactive area (Table S1), to contain a lower concentration of hydrogen sulfide at the shown as an increase in oxidation and reduction current entrance and a higher one at the exit. intensities. This is clearly seen with the modification of Gr with 1119 https://doi.org/10.1021/acsestwater.2c00589 ACS EST Water 2023, 3, 1116−1125 ACS ES&T Water pubs.acs.org/estwater Article SWCNTs/PVA/PDDA composite, upon deposition over the electrode, results in the formation of an external layer of the polymer, covering the outmost CNTs. As a non-conductive material, it was not possible to obtain any quality image of the surface at a nanoscale. The best results were obtained by applying a very high voltage, with which the electrons managed to sufficiently penetrate the external membrane and get very little detailed images of the inner surface (Figure S6A). Nevertheless, little information can be obtained from these micrographs. The only alternative was metallizing the electro- des. Still, a layer at least 20 nm thick was needed to obtain quality images (Figure S6B). Confocal microscopy measures the roughness of a surface. In our case, this is correlated to the electroactive area. As this was an important criterion upon choosing the material, further Figure 2. Electrochemical response of the printed bare Gr electrode confirmation of the difference between SWCNTs/PVA/PDDA and modified with SWCNTs/PVA/PDDA or SWCNTs/PLA in a and SWCNTs/PLA helps in explaining this result. Profiles of solution of 0.01 M [Fe(CN)6)]3−/[Fe(CN)6)]4− at a scanning speed the surfaces are obtained for each electrode and, from there, of 0.01 V/s. Intensity currents are normalized with the geometrical the value of roughness calculated (Figure S7). These profiles area. show that the roughness of SWCNTs/PLA is the highest among all electrodes, with Gr being the smoothest. This is the SWCNTs/PLA. In that case, the electroactive to geometric same tendency that was seen previously with electroactive area is almost 2 times bigger with respect to bare Gr electrodes. areas using CVs (Table S1). Thus, the correlation is as On the other hand, for Gr- SWCNTs/PVA/PDDA, the ratio is predicted. approximately the same, though still larger than that of Gr. Infrared spectroscopy can be used to differentiate the Potential peak differences of 75 and 59 mV were obtained presence of the polymers added in each mixture. This is purely when Gr-SWCNTs/PVA/PDDA and Gr-SWCNTs/PLA a qualitative measurement; nonetheless, it helps in highlighting electrodes were used, respectively, indicating that electron the differences between the polymers PVA/PDDA and PLA. transfer reactions presented a greater degree of electrochemical As can be seen in Figure S8, SWCNTs present an overall reversibility than for the Gr electrode (239 mV). Thus, the higher transmittance compared to the polymer mixtures. The smallest peak separation is achieved when SWCNTs/PLA are PVA/PDDA composite presents a very broad band between drop-casted over Gr. This electrode presents the best 3000 and 3600 cm−1, which is characteristic of the O−H and electrochemical behavior and the most reversible (59 mV) N−H bonds present in PVA and PDDA, respectively. In electron-transference process. contrast, for the composite with PLA, this band is barely Electrode’s Structure and Morphology. SEM micro- visible, as is in the case of pure SWCNTs. Given that PLA does graphs of the bare electrode and the final chosen material, as not have the O−H or N−H bonds, a smaller signal is expected. shown in Figure 3, allow us to observe the change in The small band that appears belongs to the surfactant used in morphology between Gr (a) and Gr-SWCNTs/PLA (b). As the dispersion, SDS, which also presents an O−H bond after can be seen, Gr is much flatter, even though it still presents dissolving due to the substitution of the sodium cation for a many small moieties, than Gr-SWCNTs/PLA. The latter one proton. The lower transmittance for the composites suggests has a much rougher surface, with pores and crevices. The thicker layers than pure SWCNTs, despite the same volume polymer helps in the formation of irregular structures, and the being used. images reveal a random distribution of entangled SWCNTs SWCNTs/PVA/PDDA and SWCNTs/PLA Analytical perfectly integrated in the polymer matrix improving the Response. First, to determine which potential H2S should SWCNTs’ own rugosity. This explains the electrodes’ larger be measured, linear sweep voltammetry (LSV) was performed electroactive surface compared to the bare electrodes. The at 5 mV/s, from 0 to 0.5 V vs Ag/AgCl, in a 10 mM (pH 8.5) Figure 3. Morphological characterization by SEM of the (a) Gr electrode and (b) Gr-SWCNTs/PLA electrode’s surface. 1120 https://doi.org/10.1021/acsestwater.2c00589 ACS EST Water 2023, 3, 1116−1125 ACS ES&T Water pubs.acs.org/estwater Article Figure 4. (a) Gr-SWCNTs/PVA/PDDA sensor stability and its decay along the first 20 days of usage. (b) Sensitivity comparison evolution of two Gr electrodes, one modified with SWCNTs/PVA/PDDA and another with SWCNTs/PLA. Figure 5. (a) Characterization of the response of the Gr-SWCNTs/PLA integrated sensor to H2S species on a 3−12 pH range. (b) Chronoamperometry measurements at different H2S concentrations with a close-up into the smaller ones (inset). (c) Calibration curve (n = 3) of the Gr-SWCNTs/PLA-modified sensors for H2S oxidation at an 8 < pH < 9 range with a zoom-in into the smaller concentrations (inset). H2S solution (0.1 M KCl), and the represented voltammogram reaction is limited by mass transfer. However, at around 0.1 V is obtained using a SWCNTs-PLA sensor (Figure S2). This is vs Ag/AgCl, the current is at half height of the maximum. As to find an applied voltage sufficiently high to allow the selectivity is achieved by using the smallest oxidation potential electrochemical H2S oxidation to take place but low enough to offer a high selectivity toward other ions. The maximum possible, this was found as a compromise value between current is achieved around 0.17 V vs Ag/AgCl before the selectivity and sensitivity. There is enough signal to detect H2S 1121 https://doi.org/10.1021/acsestwater.2c00589 ACS EST Water 2023, 3, 1116−1125 ACS ES&T Water pubs.acs.org/estwater Article Figure 6. Sensor’s performance study. (a) Repeatability: repeated measurements (n = 5) of the same concentration, using the same electrode, consecutively. (b) Reproducibility: comparison of the sensitivities of the first calibration of nine different sensors and their average, built in different days but under the same conditions. (c) Short-term stability: decrease in sensitivity when several calibrations (n = 5) were performed in a single day (120 measurements). Experimental errors were calculated as standard deviation. at the desired concentrations, with a reduced sulfur-poisoning variations in a pH range between 7.5 and 11.0, which rate and an increased selectivity. corresponds to a dominant presence of HS− over H2S and Following the chronoamperometry method with the S2− (Figure S1). In consequence, the sensor will operate at parameters previously described, several calibrations of four optimal sensitivity in the pH-independent region. Nonetheless, different Gr-SWCNTs/PVA/PDDA sensors were performed. it would still be capable of measuring at pHs lower than 7.5, so As calibrations were repeated thrice (n = 3), after several days, far as the pH dependence is quantified and sensor response it was found that the SWCNTs/PVA/PDDA dispersion lacks compensated accordingly by normalizing. any long-term stability (Figure 4a). After days of measure- Electroanalytical Performance. With all parameters ments, under the previously described conditions, sensors lose considered, it was proceeded to evaluate the response of the up to 80% sensitivity (from ∼6 to ∼1 mA/M), with shorter electrode to variations of the H2S concentration. Calibrations linear ranges and higher limits of detection as calibrations were were carried out at a starting pH of 8−8.5, using a PBS buffer, done. As PVA and PDDA would progressively erode from the and NaOH to adjust pH, to ensure that there was no electrode, a large portion of the SWCNTs was also lost. This significant change in the response of the sensor. As previously caused the decay in response toward H2S. Consequently, PLA stated, stock solutions need to be prepared every few hours to was proposed to overcome the solubility problems observed prevent a severe decay of the starting concentration. However, with PVA/PDDA. To keep the ratio that was previously found by preparing stock solutions at a pH of 12, oxidation is much as optimal, a dispersion of 1/0.5 SWCNTs/PLA was chosen slower, allowing them to last unaltered through the entire for the final validation for H2S sensing. As can be seen in calibration procedure. Figure 4b, the sensitivity of the new SWCNTs/PLA dispersion Different concentrations, prepared on the spot by additions decreases much slower, from a sensitivity of 7.7−5.9 mA L of different volumes of stock solution, were measured in mol−1 (23%), than in the SWCNTs/PVA/PDDA one. Given separated chronoamperometries with no stirring (Figure 5b). the stability results and the better ratio of the electroactive The signal was taken between 19.9 and 20 s in a range from 0 area, PLA is chosen to carry on with the final sensor. to 2380 μM H2S. Each measurement was repeated three times Three-Electrode Integration. To complete the full (n = 3). The sensor shows a good linear correlation between integration of the Gr electrode as the WE, a functional RE concentrations of 16.3 and 2380 μM H2S and a LD of 4.3 μM and CE were integrated. The RE is made of a Ag/AgCl paste (Figure 5c). and protected from sulfur attacks with a PVB membrane31 that Additionally, repeatability, reproducibility, and short-term is NaCl-saturated to increase the stability of the electrode by stability were also studied to fully characterize the Gr- keeping a constant concentration of Cl−. To verify that the SWCNTs/PLA integrated sensor response (Table S2). integrated CE and RE were good enough to replace the Repeatability consists of the consecutive measurement of previously used commercial ones, CVs were performed under different samples of equal concentration. After five measure- the same conditions (Figure S9) using a Gr. Results showed no ments at 31 μM, with an RSD of 2.3%, and at 430 μM, with an variations in the intensity, even as the CE is swapped, but a RSD of 2.7%, it was proven that, for a given sensor at a certain decrease in peak separation when using the integrated RE concentration, there was no significant signal changes (Figure compared to an external commercial reference electrode. As 6a). Moreover, this proves that, while a working potential of such, the systems are not only interchangeable perfectly, but 0.1 V vs Ag/AgCl could lead to noise due to variations in the the integrated electrodes also lower the overpotential, applied voltage, this noise is small and does not affect the improving the electrochemical performance sensor. measure. Optimizing the Working pH. Of even greater importance Data shows that the optimized method of measurement was the response variance at different pH, given that H2S allows for a small dispersion of signal values when measuring a partially appears as a gas at acidic pH (Figure S1). Under this specific analyte’s concentration. Reproducibility is the condition, stripping can occur. As a result, all measurements calibration of several electrodes, equally fabricated and must be carried out in a limited amount of time. Thus, finding measured, under the same conditions, consecutively. With an an optimal range of operation is of paramount importance. average sensitivity of 1.65 mA/M and a standard deviation of With the optimized conditions, oxidation current is measured 0.29 mA/M (18%), the fabrication method of the electrodes at different pH values at a fixed analyte activity and equal time bears some difference between their responses (Figure 6b). (Figure 5a). As can be seen, the response presents the smallest This is most likely attributable to the manual nature of the last 1122 https://doi.org/10.1021/acsestwater.2c00589 ACS EST Water 2023, 3, 1116−1125 ACS ES&T Water pubs.acs.org/estwater Article fabrication step by drop casting the SWCNTs/PLA dispersion. of working under real conditions with complex samples Short-term stability consists of continuous calibrations matrices. (analyzing n = 9 H2S concentrations) of a sensor for 1 day. As can be seen in Figure 6c, sensitivity is only lightly affected by consecutive measurements, going from 2.14 to 1.52 mA/M ■ * ASSOCIATED CONTENT sı Supporting Information or decreasing its response by 29% over 120 measurements. The Supporting Information is available free of charge at This shows that sulfur poisoning is close to negligible, and the https://pubs.acs.org/doi/10.1021/acsestwater.2c00589. sensor suffers little strain over continuous usage. Interference Study. To assess the selectivity of the sensor, Hydrogen sulfide species in aqueous media (Figure S1); several species commonly present in water samples were study of the electroactive area and voltage separation measured (Figure S10). All the measured species give a signal (Table S1); determination of working potential to of the same magnitude compared to that of the LD of H2S. oxidize H2S using a SWCNTs-PLA sensor (Figure When measured at an optimal pH and equal concentration, S2); images of different proportions of SWCNTs/PVA none of the interferents gives a current close to that of the (Figure S3); CV of a Gr electrode modified with analyte. This corroborates that by using a low working SWCNTs/PDDA 1/0.5 (Figure S4); photographs of the potential, the sensor manages to achieve good selectivity for deposition of SWCNTs/PLA (Figure S5); micrographs the target medium. obtained by SEM of PVA/PDDA electrodes (Figure Real Sample Study. The final assessment of the S6); surface 3D profiles obtained using confocal applicability of the sensor was studied by measurements on microscopy to study the roughness (Figure S7); infrared spiked and real samples of increasing matrix complexity. To analysis for qualitative differentiation of the composition contrast the results, they are compared with a commercial H2S (Figure S8); electrochemical characterization and sensor. comparison of the integrated electrode (Figure S9); Sensors are calibrated right before measuring the samples. interference study of several species common in water Results show (Table 1) that the sensor is capable of samples (Figure S10); study of repeatability, reproduci- bility, and short-term stability (Table S2) (PDF) Table 1. Comparison Response between Our Printed Sensor and the Commercial H2S Sensora sample commercial sensor measured t-test (n = ■ AUTHOR INFORMATION Corresponding Author type [H2S] (μM) [H2S] (μM) 2, 95%) Mireia Baeza − Department of Chemistry, Faculty of Science, printed Milli-Q 180.6 ± 0.3 193.6 ± 0.1 0.62 Edifici C-Nord, Universitat Autónoma de Barcelona, 08193 sensor water Bellaterra, Spain; GENOCOV Research Group, Universitat tap water 197.1 ± 0.6 210.4 ± 0.7 1.89 Autónoma de Barcelona, 08193 Bellaterra, Spain; reactor 178.5 ± 0.7 206.9 ± 0.5 2.82 orcid.org/0000-0002-2240-6410; media Email: mariadelmar.baeza@uab.cat a The measurement error was expressed as the standard deviation (n = 2). Authors Franc Paré − Department of Chemistry, Faculty of Science, quantitatively measuring H2S in complex matrix samples Edifici C-Nord, Universitat Autónoma de Barcelona, 08193 since all the calculated t-Student values are lower than the Bellaterra, Spain; GENOCOV Research Group, Universitat corresponding limit for 2 degrees of freedom at a confidence Autónoma de Barcelona, 08193 Bellaterra, Spain interval of 95% (2.92). Rebeca Castro − Department of Mining Engineering and Natural Resources, Universitat Politècnica de Catalunya, ■ CONCLUSIONS In this work, a miniaturized, fully printed H2S sensor is 08240 Manresa, Spain David Gabriel − Departament of Chemical, Biological and presented. The fabrication and characterization steps are Environmental Engineering, Escola d’Enginyeria, Universitat thoroughly described. Results show that printing technologies Autónoma de Barcelona, 08193 Bellaterra, Spain; can be used to produce electroanalytical devices. In particular, GENOCOV Research Group, Universitat Autónoma de an easy and cheap to mass-produce sensor capable of operating Barcelona, 08193 Bellaterra, Spain for over 25 days with minimal sensitivity decrease is shown. Xavier Guimerà − Department of Mining Engineering and Nevertheless, future work can further improve carbon-based Natural Resources, Universitat Politècnica de Catalunya, material inks for fully printable electrodes. Still, the 08240 Manresa, Spain modifications applied using a SWCNTs/PLA ink over bare Gemma Gabriel − Instituto de Microelectrónica de Barcelona, Gr electrodes show a clear improvement in sensitivity (up to IMB-CNM (CSIC), Campus Universitat Autónoma de 2.2 mA/M in the best sensors) and sulfur-poisoning resistance, Barcelona, 08193 Bellaterra, Spain; CIBER de Bioingeniería, performing up to 135 measurements with less than 30% signal Biomateriales y Nanomedicina, Instituto de Salud Carlos III, loss. PLA helps to increase the mechanical resistance of the Gr- Madrid, Spain; orcid.org/0000-0003-2140-6299 SWCNT sensor, resisting erosion far longer than 72 h of the Complete contact information is available at: SWCNTs by themselves. This allows the sensor to https://pubs.acs.org/10.1021/acsestwater.2c00589 continuously operate for a few days or during measurement- intensive periods without needing replacement. Moreover, the Author Contributions optimal pH range for measurements, the integration of all the The manuscript was written through contributions of all electrodes into a single platform, and the high selectivity authors. All authors have given approval to the final version of achieved by the low oxidation potential of H2S make it capable the manuscript. CRediT: Franc Paré investigation (equal), 1123 https://doi.org/10.1021/acsestwater.2c00589 ACS EST Water 2023, 3, 1116−1125 ACS ES&T Water pubs.acs.org/estwater Article methodology (equal), validation (equal), writing-original draft (12) Kaushik, R.; Sakla, R.; Ghosh, A.; Selvan, G. T.; Selvakumar, P. (equal), writing-review & editing (equal); Rebeca Castro M.; Jose, D. A. Selective Detection of H2S by Copper Complex Carrasco investigation (equal), methodology (equal); David Embedded in Vesicles through Metal Indicator Displacement Gabriel funding acquisition (lead), project administration Approach. ACS Sens. 2018, 3, 1142−1148. (equal); Xavier Guimerà investigation (equal), supervision (13) Dulac, M.; Melet, A.; Galardon, E. Reversible Detection and Quantification of Hydrogen Sulfide by Fluorescence Using the (equal); Gemma Gabriel conceptualization (equal), super- Hemoglobin i from Lucina Pectinata. ACS Sens. 2018, 3, 2138−2144. vision (equal), writing-original draft (equal), writing-review & (14) Roda, B.; Zhang, N.; Gambari, L.; Grigolo, B.; Eller-Vainicher, editing (equal); Mireia Baeza conceptualization (lead), C.; Gennari, L.; Zappi, A.; Giordani, S.; Marassi, V.; Zattoni, A.; funding acquisition (lead), project administration (lead), Reschiglian, P.; Grassi, F. Optimization of a Monobromobimane supervision (lead), writing-original draft (lead), writing-review (MBB) Derivatization and RP-HPLC-FLD Detection Method for & editing (lead). Sulfur Species Measurement in Human Serum after Sulfur Inhalation Funding Treatment. Antioxidants 2022, 11, 939. Authors acknowledge the grant PID2021-126253OB-C21 (15) Zub, K.; Hoeppener, S.; Schubert, U. S. Inkjet Printing and 3D funded by MCIN/AEI/10.13039/501100011033 and by Printing Strategies for Biosensing, Analytical and Diagnostic Applications. Adv. Mater. 2022, 2105015. “ERDF A way of making Europe”, by the “European Union”, (16) Xu, Z.; Dong, Q.; Otieno, B.; Liu, Y.; Williams, I.; Cai, D.; Li, for the financial support provided to perform this research. Y.; Lei, Y.; Li, B. Real-Time in Situ Sensing of Multiple Water Quality Finally, we thank Fusion360 and Cura for their free software. Related Parameters Using Micro-Electrode Array (MEA) Fabricated Notes by Inkjet-Printing Technology (IPT). Sens. Actuators, B 2016, 237, The authors declare no competing financial interest. 1108−1119. (17) Jović, M.; Zhu, Y.; Lesch, A.; Bondarenko, A.; Cortés-Salazar, ■ ACKNOWLEDGMENTS The authors acknowledge Servei de Microscopia of UAB for F.; Gumy, F.; Girault, H. H. Inkjet-Printed Microtiter Plates for Portable Electrochemical Immunoassays. J. Electroanal. Chem. 2017, 786, 69−76. the SEM and confocal microscopy analysis. We thank Míriam (18) Moya, A.; Ortega-Ribera, M.; Guimerà, X.; Sowade, E.; Zea, González for her help with polymer dispersion experiments M.; Illa, X.; Ramon, E.; Villa, R.; Gracia-Sancho, J.; Gabriel, G. Online and to Joel Pérez for his help with the three-electrode platform Oxygen Monitoring Using Integrated Inkjet-Printed Sensors in a integration tests. Liver-on-a-Chip System. Lab Chip 2018, 18, 2023−2035. (19) Khan, Y.; Pavinatto, F. J.; Lin, M. C.; Liao, A.; Swisher, S. L.; ■ REFERENCES (1) Wang, Y.; Yan, H.; Wang, E. Solid Polymer Electrolyte-Based Mann, K.; Subramanian, V.; Maharbiz, M. M.; Arias, A. C. Inkjet- Printed Flexible Gold Electrode Arrays for Bioelectronic Interfaces. Adv. Funct. Mater. 2016, 26, 1004−1013. Hydrogen Sulfide Sensor. Sens. Actuators, B 2002, 87, 115−121. (20) Moya, A.; Gabriel, G.; Villa, R.; Javier del Campo, F. Inkjet- (2) Sarfraz, J.; Fogde, A.; Ihalainen, P.; Peltonen, J. The Performance Printed Electrochemical Sensors. Curr. Opin. Electrochem. 2017, 3, of Inkjet-Printed Copper Acetate Based Hydrogen Sulfide Gas Sensor 29−39. on a Flexible Plastic Substrate - Varying Ink Composition and Print (21) Zuo, P.; Wang, R.; Li, F.; Wu, F.; Xu, G.; Niu, W. A Trace Ppb- Density. Appl. Surf. Sci. 2018, 445, 89−96. Level Electrochemical H2S Sensor Based on Ultrathin Pt Nanotubes. (3) Yu, C.; Wang, Y.; Hua, K.; Xing, W.; Lu, T. Electrochemical H2S Talanta 2021, 233, No. 122539. Sensor with H2SO4 Pre-Treated Nafion Membrane as Solid Polymer (22) Zhou, T.; Wang, M.; He, X.; Qiao, J. Poly(Vinyl Alcohol)/ Electrolyte. Sens. Actuators, B 2002, 86, 259−265. Poly(Diallyldimethylammonium Chloride) Anion-Exchange Mem- (4) Carapezza, M. L.; Ranaldi, M.; Tarchini, L.; Ricci, T.; Barberi, F. brane Modified with Multiwalled Carbon Nanotubes for Alkaline Origin and Hazard of CO2 and H2S Emissions in the Lavinio-Tor Fuel Cells. J. Mater. 2019, 5, 286−295. Caldara Zone (Metropolitan City of Rome Capital, Italy). J. Volcanol. (23) Ben, J.; Song, Z.; Liu, X.; Lü, W.; Li, X. Fabrication and Geotherm. Res. 2020, 402, No. 106985. Electrochemical Performance of PVA/CNT/PANI Flexible Films as (5) Kenreck, G. Manage Hydrogen Sulfide Hazards with Chemical Electrodes for Supercapacitors. Nanoscale Res. Lett. 2020, 15, 1. Scavengers. Hydrocarb. Process. 2014, 73−76. (24) Zhang, J.; Zhou, T.; Qiao, J.; Liu, Y.; Zhang, J. Hydroxyl Anion (6) Ahmad, W.; Sethupathi, S.; Kanadasan, G.; Lau, L. C.; Conducting Membranes Poly(Vinyl Alcohol)/ Poly- Kanthasamy, R. A Review on the Removal of Hydrogen Sulfide from Biogas by Adsorption Using Sorbents Derived from Waste. Rev. (Diallyldimethylammonium Chloride) for Alkaline Fuel Cell Chem. Eng. 2021, 37, 407−431. Applications: Effect of Molecular Weight. Electrochim. Acta 2013, (7) Shang, H.; Xu, H.; Liu, Q.; Du, Y. PdCu Alloy Nanosheets- 111, 351−358. Constructed 3D Flowers: New Highly Sensitive Materials for H2S (25) Samsudin, A. M.; Hacker, V. Preparation and Characterization Detection. Sens. Actuators, B 2019, 289, 260−268. of PVA/PDDA/Nano-Zirconia Composite Anion Exchange Mem- (8) Syed, M.; Soreanu, G.; Falletta, P.; Béland, M. Removal of branes for Fuel Cells. Polymers 2019, 11, 1399. Hydrogen Sulfide from Gas Streams Using Biological Processes - A (26) Maity, D.; Rajavel, K.; Thangavelu, R.; Kumar, R. Polyvinyl Review. Ind. Eng. Chem. Res. 2019, 58, 22133−22164. Alcohol Wrapped Multiwall Carbon Nanotube (MWCNTs) Network (9) Li, C.; Zhang, D.; Wang, J.; Hu, P.; Jiang, Z. Magnetic MoS2 on on Fabrics for Wearable Room Temperature Ethanol Sensor. Sens. Multiwalled Carbon Nanotubes for Sulfide Sensing. Anal. Chim. Acta Actuators, B 2018, 261, 297−306. 2017, 975, 61−69. (27) Marcelo, G.; Tarazona, M. P.; Saiz, E. Solution Properties of (10) Blázquez, E.; Gabriel, D.; Baeza, J. A.; Guisasola, A. Treatment Poly(Diallyldimethylammonium Chloride) (PDDA). Polymer 2005, of High-Strength Sulfate Wastewater Using an Autotrophic 46, 2584−2594. Biocathode in View of Elemental Sulfur Recovery. Water Res. 2016, (28) Wang, Y.; Xiong, C.; Qu, H.; Chen, W.; Ma, A.; Zheng, L. 105, 395−405. Highly Sensitive Real-Time Detection of Tyrosine Based on Organic (11) Jornet-Martínez, N.; Hakobyan, L.; Argente-García, A. I.; Electrochemical Transistors with Poly-(Diallyldimethylammonium Molins-Legua, C.; Campíns-Falcó, P. Nylon-Supported Plasmonic Chloride), Gold Nanoparticles and Multi-Walled Carbon Nanotubes. Assay Based on the Aggregation of Silver Nanoparticles: In Situ J. Electroanal. Chem. 2017, 799, 321−326. Determination of Hydrogen Sulfide-like Compounds in Breath (29) Vany, P.; Vany, P. ELECTROCHEMICAL SERIES Petr Samples as a Proof of Concept. ACS Sens. 2019, 4, 2164−2172. Vanysek. J. Phys. Chem. Ref. Data 1989, 23−33. 1124 https://doi.org/10.1021/acsestwater.2c00589 ACS EST Water 2023, 3, 1116−1125 ACS ES&T Water pubs.acs.org/estwater Article (30) Jiang, L.; Gao, L.; Sun, J. Production of Aqueous Colloidal Dispersions of Carbon Nanotubes. J. Colloid Interface Sci. 2003, 260, 89−94. (31) Moya, A.; Pol, R.; Martínez-Cuadrado, A.; Villa, R.; Gabriel, G.; Baeza, M. Stable Full-Inkjet-Printed Solid-State Ag/AgCl Reference Electrode. Anal. Chem. 2019, 91, 15539−15546. 1125 https://doi.org/10.1021/acsestwater.2c00589 ACS EST Water 2023, 3, 1116−1125