Article pubs.acs.org/ac
In Situ Optimization of pH for Parts-Per-Billion Electrochemical Detection of Dissolved Hydrogen Sulfide Using Boron Doped Diamond Flow Electrodes Eleni Bitziou,† Maxim B. Joseph,† Tania L. Read,† Nicola Palmer,§ Tim Mollart,§ Mark E. Newton,‡ and Julie V. Macpherson*,† †
Departments of Chemistry and ‡Physics, University of Warwick, Coventry CV4 7AL, U.K. § Element Six Ltd., Element Six Global Innovations Centre, Harwell Campus, Didcot, OX11 0QR, U.K. S Supporting Information *
ABSTRACT: A novel electrochemical approach to the direct detection of hydrogen sulfide (H2S), in aqueous solutions, covering a wide pH range (acid to alkali), is described. In brief, a dual band electrode device is employed, in a hydrodynamic flow cell, where the upstream electrode is used to controllably generate hydroxide ions (OH−), which flood the downstream detector electrode and provide the correct pH environment for complete conversion of H2S to the electrochemically detectable, sulfide (HS−) ion. All-diamond, coplanar conducting diamond band electrodes, insulated in diamond, were used due to their exceptional stability and robustness when applying extreme potentials, essential attributes for both local OH− generation via the reduction of water, and for in situ cleaning of the electrode, post oxidation of sulfide. Using a galvanostatic approach, it was demonstrated the pH locally could be modified by over five pH units, depending on the initial pH of the mobile phase and the applied current. Electrochemical detection limits of 13.6 ppb sulfide were achieved using flow injection amperometry. This approach which offers local control of the pH of the detector electrode in a solution, which is far from ideal for optimized detection of the analyte of interest, enhances the capabilities of online electrochemical detection systems. ydrogen sulfide (H2S) is a colorless, flammable, and toxic gas with a distinct rotten egg odor.1 Immediate exposure to levels >300 ppm is considered dangerous to life or health, as is continuous exposure to levels >20 ppm.2 Industrial processes involving oil and natural gas, pulp and paper production, sewage treatment, and sulfur production plants are prone to potentially high levels of H2S, either occurring naturally or as a byproduct of the industrial method.1 In aqueous solution at 20 °C, the following pH-dependent protolytic equilibria exists:3
H
H 2S(aq) + H 2O(l) ⇌ HS−(aq) + H3O+(aq)
monitoring. The most notable electrochemical H2S(aq) sensors are based on potentiometric,10 galvanostatic,11 and amperometric12,13 techniques. Potentiometric approaches, such as the silver/silver sulfide ion selective electrode14 which monitor sulfide concentration typically require a suitably high pH for measurement (eq 1). They can also suffer from slow response times and drift.9a An attractive amperometric H2S measurement method, and the basis of a commercial sensor, employs a Clark type cell.15 In brief, dissolved H2S(g) passes through a H2S(g) permeable membrane into alkaline solution, ensuring complete conversion of H2S to HS−. Oxidation of HS− with an oxidizing reagent (ferricyanide) results in the formation of sulfur and ferrocyanide. The measurement signal, which correlates to the dissolved H2S concentration, is generated by reoxidation of ferrocyanide at a Pt working electrode. This methodology works best at pH < 9 due to the negligible levels of dissolved H2S(g) in solution under alkaline conditions.16 To simplify the experimental procedure, in particular negating the use of both additional chemical reagents and an electrode membrane (which is prone to clogging, a frequent problem of Clark type cell electrodes15) and extending the pH
pK a1 = 6.88 (1)
HS−(aq) + H 2O(l) ⇌ S2 −(aq) + H3O+(aq)
pK a2 = 14.15 (2)
Note HS− is electrochemically detectable and the pKa values will vary slightly with salt content and temperature. Additionally, the solubility of H2S decreases with increasing temperature and salinity.4,5 Various analytical techniques have been used for the detection of dissolved sulfide such as photometric methods,6 gas or ion chromatography,7 and fluorescence analysis.8 Electrochemical methodologies9 provide an attractive platform for industrial and environmental applications due to the simplicity of the approach and the potential for online © 2014 American Chemical Society
Received: August 6, 2014 Accepted: September 27, 2014 Published: September 28, 2014 10834
dx.doi.org/10.1021/ac502941h | Anal. Chem. 2014, 86, 10834−10840
Analytical Chemistry
Article
detection directly at the downstream electrode,12a,22 (iii) fast response times, and (iv) the use of successful in situ electrochemical cleaning protocols, paving the way for prolonged deployment in situ. (3) Controlled hydrodynamic flow in order to increase sensitivity and maximize detection signals at the detector electrode.23 (4) A sensor not limited by pH.
range over which the amperometric sensor will operate, we describe an alternative approach, which operates in flow, as shown in Figure 1. Herein we propose a dual band electrode, where sulfide is directly detected at the downstream electrode.
■
EXPERIMENTAL SECTION Solutions. All solutions were prepared using Milli-Q water (resistivity 18.2 MΩ cm at 25 °C). Preliminary electrode characterization experiments utilized ferrocenylmethyl trimethylammonium (FcTMA+), synthesized in house and hexaamineruthenium (Ru(NH3)63+) in 0.1 M potassium nitrate (KNO3) as supporting electrolyte. For sulfide detection in quiescent solution, buffered sodium sulfide (Na2S) solutions were prepared (borax for pH 10, phosphate for pH 7, and phthalate for pH 4) in 0.1 M KNO3. A glass sealed 1 mm diameter BDD macrodisc electrode, fabricated as described previously,20b from freestanding electroanalysis grade BDD (Element Six, Harwell),24 was used for the initial sulfide oxidation measurements. For flow injection analysis electrochemical (FIA-EC) experiments, the mobile phase contained 0.1 M KNO3 at pH 3.8, pH 6.1 or pH 10, where the pH was adjusted by adding 0.1 M HNO3 or 0.1 M NaOH. Stock solutions of 2 mM Na2S were prepared (deaerated) in 0.1 M KNO3 at pH 3.8, pH 6.1 or pH 10 and diluted as appropriate into solutions containing 0.1 M KNO3 at the required pH. The diluted solutions were injected in the FIA-EC system (n = 4) in order to determine limits of detection [LOD, 3 × standard deviation of the blank (sb) + blank] for sulfide oxidation.25 All-Diamond Band Electrode Fabrication. Fabrication of individually addressable all-diamond band electrodes, starting
Figure 1. Schematic illustrating the all-diamond coplanar dual band electrode applied to sulfide detection in nonalkaline aqueous solutions. The required [OH−] to convert H2S to sulfide (HS−) is generated at the upstream BDD electrode, while the downstream BDD electrode is poised at a potential to electrochemically detect HS−.
The key features of the sensor are (1) the use of an upstream electrode to generate OH− in order to optimize the local pH environment of the detector such that H2S is in the electrochemically detectable HS− form.17 (2) The use of robust conducting boron doped diamond (BDD) electrodes insulated in diamond (coplanar arrangement).18 These electrodes provide inherent ultrawide solvent windows in aqueous media, low capacitive currents, high resistance to fouling/ corrosion processes,19,20 and the ability to maintain structural integrity even at high applied potentials.20c,21 They enable (i) stable pH generation over long periods of time,17 (ii) sulfide
Figure 2. (a) Schematic showing the step-by-step fabrication (1−5) of an all-diamond dual band electrode. (b) FE-SEM image of the coplanar 90 μm BDD detector electrode, insulated with diamond. (c) MSL fabricated flow cell positioned on top of the dual BDD band electrode. (d) Relative proportion of hydrogen sulfide as a function of pH, described in detail in section ESI2 in the Supporting Information and defined assuming pKa1 = 6.88 and pKa2 = 14.15. 10835
dx.doi.org/10.1021/ac502941h | Anal. Chem. 2014, 86, 10834−10840
Analytical Chemistry
Article
deposited on the detector electrode.29,17,29a,30 To confine IrOx to the electrode surface and prevent significant overgrowth, electrodeposition took place in the flow cell, under stationary conditions, by repeatedly pulsing (10 s pulses) the potential between 0 V and +0.7 V for a total time of 120 s. The film was left to stabilize in water for 2 days and then calibrated using pH buffers in the range 2−10. The open circuit potential (OCP) was measured against a Ag|AgCl reference electrode, under flow conditions (1 cm3 min−1) pertinent to those described herein. The calibration plot, recorded over several days (n = 5), produced a super-Nernstian slope of 80.3 ± 2 mV pH−1 and an intercept of 56 ± 1 mV (R2 = 0.999) and was stable even after extensive use. [OH−] was galvanostatically generated at the upstream electrode using a Keithley current source (versus a Pt counter electrode located in the outlet tubing) in conjunction with a potentiostat equipped with a peripheral differential amplifier module that allowed simultaneous recording of OCP at the IrOx-BDD band electrode. For the simultaneous galvanostatic OH− generation and sulfide amperometric detection, a current source (vs Pt counter electrode) and potentiostat (vs second Pt counter and Ag|AgCl reference) were again employed; there was no evidence of instrumental interferences. The sulfide detection results with and without galvanostatic OH − generation were compared using a student’s t-test.25
from a wafer of insulating diamond, has been detailed previously and is outlined in Figure 2a.18 Briefly, insulating (polished) CVD grown diamond was laser micromachined (E355H-3-ATHI-O, Oxford Lasers) to create recessed structures, reflecting the resulting geometry of the dual band electrode. The structures were acid cleaned in boiling concentrated H2SO4 saturated with KNO320d and white light interferometry (WLI, WYCO NT-2000) performed to characterize the depth and roughness of the trenches. Electroanalysis grade polycrystalline BDD24,20c was grown into and over the lasered trenches using a microwave plasma CVD system (Element Six). The substrate was lapped to expose the coplanar BDD band structures (surface roughness
