Spontaneous electrochemical removal of aqueous sulfide

water research 42 (2008) 4965–4975

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Spontaneous electrochemical removal of aqueous sulfide Paritam K. Dutta, Korneel Rabaey, Zhiguo Yuan, Ju¨rg Keller* The University of Queensland, Advanced Water Management Centre (AWMC), St. Lucia, QLD 4072, Australia

article info

abstract

Article history:

Most of the existing sulfide removal processes from wastewaters and waste gases require

Received 22 February 2008

substantial amounts of energy inputs. Here we present an electrochemical method by

Received in revised form

means of a fuel cell that removes sulfide while producing energy. A lab scale fuel cell was

5 September 2008

operated at ambient temperature and neutral pH, which was capable of removing aqueous

Accepted 6 September 2008

sulfide continuously for 2 months at a rate of 0.62  0.1 kg S m3 d1 of net anodic

Published online 27 September 2008

compartment (NAC) (0.28  0.05 kg S m3 d1 of total anodic compartment, TAC). During continuous operation, on average, the power generated was 12  2 W m3 NAC (5  1 W m3

Keywords:

TAC), with a maximum capacity of the cell of 166 W m3 NAC (74 W m3 TAC). Potassium

Sulfide

ferricyanide was used as cathodic electron acceptor. Elemental sulfur was identified as the

Wastewater

predominant final oxidation product that was deposited on the anode. In this abiotic fuel

Fuel cell

cell, the sulfide oxidation rate was not diminished by the presence of an organic electron

Electricity

donor (acetate) during batch experiments while the acetate concentration remained

Removal

unchanged. This is particularly important for selective sulfide removal from wastewater

Elemental sulfur

where organics are essential for downstream nutrient removal. Elemental sulfur deposited on the anode appeared to limit the operation of the fuel cell after 3 months of operation, necessitating periodic removal of the accumulated sulfur from the electrode. ª 2008 Elsevier Ltd. All rights reserved.

1.

Introduction

Sulfide is found in the environment as dissolved sulfide in wastewaters and as gaseous hydrogen sulfide in waste gases. Sewage and industrial wastewaters from e.g. petrochemical plants and tanneries (Vaiopoulou et al., 2005; Janssen et al., 1999) are important sources of aqueous sulfide. Gaseous sources such as biogas, natural gas as well as off-gases from wastewater treatment systems (e.g. inlet works of sewage treatment plants) are important sources of gaseous sulfide. Sulfide is toxic, corrosive and odorous. Hence, its removal from wastewaters and waste gases is required from both an environmental and an economical standpoint. A variety of physicochemical methods such as chemical oxidation and catalytic conversion have been used to oxidize sulfide to either elemental sulfur or sulfate, thus achieving sulfide removal

from wastewaters and waste gases. Since these methods require substantial energy and chemical inputs, the development of more cost-effective techniques is needed. Electrochemical processes offer several advantages over the aforementioned methods, including good energetic efficiency, environmental compatibility, amenability to automation, versatility and cost effectiveness (Rajeshwar et al., 1994). Sulfide (H2S, HS, S2), an electrochemically active component, can react at an anodic electrode and directly donate electrons to the electrode. Depending on the experimental conditions elemental sulfur, polysulfides, sulfate, dithionate, and thiosulfate may be produced during oxidation of sulfide. Elemental sulfur has been found to be the main electrochemical product (Ateya et al., 2003; Szynkarczuk et al., 1994; Farooque and Fahidy, 1977). Most of the studies to date were performed in alkaline media by electrolysis at high anode

* Corresponding author. Tel.: þ61 7 3365 4727; fax: þ61 7 3365 4726. E-mail address: [email protected] (J. Keller). 0043-1354/$ – see front matter ª 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2008.09.007

4966

water research 42 (2008) 4965–4975

potentials and through catalytic conversion in a single chamber reactor. The electrochemical oxidation of pyrite (FeS2), the most common sulfide mineral, was also reported (Kelsall et al., 1999; Zhu et al., 1994). Zaman and Chakma (1995) and Kameyama et al. (1981) extracted two valuable products, hydrogen and sulfur, from gaseous hydrogen sulfide using electrochemical oxidation. None of the aforementioned studies considered spontaneous electrochemical oxidation of aqueous sulfide. Both electrolysis and catalytic conversion need net energy to run the process, while a spontaneous electrochemical process can generate electricity. Researchers working on seafloor or marine sediment fuel cells have reported that a significant portion of the anodic current of these fuel cells resulted from abiotic oxidation of dissolved and mineral sulfides (Reimers et al., 2006; Ryckelynck et al., 2005). Aguilar et al. (2004) and Chuang et al. (2000) achieved the oxidation of gaseous H2S in fuel cells at high temperature and pressure with anode catalysts. Recently, Rabaey et al. (2006) reported that sulfide present in an anaerobic digester effluent can be oxidized in a microbial fuel cell. Their study, which was performed in a biological system, demonstrated sulfide removal in the presence of micro-organisms and residual organics. Hence, biological factors were not uncoupled from abiotic spontaneous electrochemical sulfide oxidation. In addition, the causes for the observed low charge recovery (between 22 and 54%) were not addressed, while the sulfide oxidation products were also not fully identified and quantified. According to our knowledge no detailed studies have been reported on the oxidation of aqueous sulfide using abiotic fuel cells at ambient temperature, pressure and neutral pH. The main aim of this study is to examine spontaneous electrochemical aqueous sulfide oxidation using a non-catalysed high surface area graphite electrode. The impact of several key parameters, such as the anode potential, the presence of alternate electron donors and the pH on sulfide removal is investigated. In addition, the long-term performance of the fuel cell is evaluated to further demonstrate the application potential of the process.

2.

Materials and methods

2.1.

Fuel cell design and operation

The abiotic fuel cell was constructed according to Freguia et al. (2007a). Fig. 1 gives a schematic diagram of the electrochemical cell. It consisted of a rectangular anode chamber and a rectangular cathode chamber, which were separated by a cation exchange membrane (Ultrex, CM17000, Membranes International Inc., USA). The volumes of the total anode compartment (TAC) and the total cathode compartment (TCC) were both 450 mL. Both chambers were filled with graphite granules (El Carb 100, Graphite Sales Inc., USA) which reduced the net anode compartment (NAC) and the net cathode compartment (NCC) volumes to 200 mL each. The graphite granules had diameters between 1.5 mm and 6 mm. The reactor bed had a porosity of approximately 0.44. Prior to use, graphite granules were pre-treated with acid and base to remove impurities. At first, granules were

kept in 1 M HCl for 24 h and then washed 4–5 times with reverse osmosis water. After that granules were kept in 1 M NaOH solution for 24 h and finally washed 8–10 times again. A graphite rod (5 mm diameter) was used in both the anode and cathode compartments to connect the electrodes to the external circuit. A variable resistor (max 100 U) was used as external load. Sodium sulfide (Na2S$9H2O) was supplied to the anode compartment, either as a pulse or continuously depending on the mode of experiments. Before using, Na2S$9H2O crystals were washed with MilliQ (18 MU) water to remove oxidized sulfur species at the surface of the crystals. A phosphate buffered medium (per litre 4 g Na2HPO4; 5 g KH2PO4; 1 g NaCl) was used as electrolyte and buffer. A 100 mM K3Fe(CN)6 solution was used as electron acceptor at the cathode. An external buffer flask of 1 L was used in the anodic recirculation line during batch experiments. Similarly, the cathodic external buffer flask of 1 L was included in the cathodic recirculation line for all experiments to minimize periodic replenishment of catholyte. The recirculation flow rates at both anode and cathode were kept at 6 L h1. PVC tubing of 3 mm internal diameter was used for the recirculation lines. A pH probe (Ionode Pty Ltd., Aus) was placed in the anodic recirculation line to measure pH of the anode compartment. In all experiments, an Ag/AgCl (RE-5B, BioAnalytical, USA) electrode was used as the reference electrode. Its potential was estimated as þ197 mV versus standard hydrogen electrode (SHE), this value may fluctuate slightly. This fluctuation depends on the chloride concentration and the temperature as defined by the Nernst equation. The extent of the fluctuation is predicted to be quite limited (approximately 5 mV) due to the constant temperature and chloride concentration within the experimental timeframe.

2.2.

Batch sulfide oxidation experiments in the fuel cell

Batch tests, divided into four different sets, were conducted to verify the feasibility of spontaneous sulfide oxidation and to investigate the impact of various factors on the performance of the fuel cell. Graphite granules were changed after 14–16 batches of experiments except in one case where granules were used for 30 batches. The latter allowed a higher amount of elemental sulfur per unit mass of graphite granules and hence facilitated quantification (see further Section 2.5). The first set of batch tests were conducted at different external resistances of 5, 10, 20, 50 U to observe spontaneous sulfide oxidation at different external loads and to measure the rates of sulfide oxidation, the Coulombic efficiency, as well as the products of sulfide oxidation. A batch experiment was started by sparging the anodic reactor for about half an hour with nitrogen to achieve anaerobic conditions. A sulfide solution was then spiked into the anode compartment, resulting in an initial sulfide concentration of 100  5 mg SL1. The pH in the anode compartment was generally maintained at around 6.75 due to the presence of the phosphate buffer. The experimental time for each test lasted for 5–24 h, depending on the external resistance applied. The anode and cathode potentials as well as the current were recorded

4967

water research 42 (2008) 4965–4975

e

e

-

4Fe(CN)6

0

S Effluent Buffer tank

Buffer tank pH meter

3Fe(CN)6

H2S/HS Reference Electrode

Influent

ANODE

CATHODE

Voltage/Potential Data Acquistion Unit Potentiostat

Fig. 1 – Schematic diagram of the electrochemical sulfide oxidation fuel cell. Sulfide containing medium is recirculated through a buffer tank and the anode compartment, where oxidation occurs. Likewise, ferricyanide is reduced to ferrocyanide at the cathode. The catholyte is recirculated through a buffer tank and the cathode compartment. In continuous mode, medium containing sulfide is pumped into the recirculation loop.

continuously using an Agilent 34970A data acquisition unit (to be further described below). Liquid phase samples were collected from the reactor for the measurement of various sulfur compounds (methods to be detailed below). Each batch test was performed in triplicate. The second set of batch tests were performed to verify if acetate oxidation occurs spontaneously in the cell and to measure the sulfide oxidation rate in the presence of organic compounds. Acetate was spiked simultaneously with sulfide, resulting in initial acetate concentrations of 100–1000 mg COD L1. The initial sulfide concentration used in these tests was 100  5 mg S L1. The external resistance used was 50 U. In a third set of batch tests, the buffer was omitted to observe the pH change during sulfide oxidation and the impact of pH variation on sulfide oxidation. Granules used in this set of batch tests were subjected to three previous batch experiments during which approximately 300 mg of sulfur was precipitated on the granules. These pH change observation tests were run with 1 g L1 NaCl as the only electrolyte. The sulfide spiked into the anode compartment was 100  5 mg S L1 and the external resistance applied was 50 U. A fourth set of batch tests aimed to investigate the influence of the anode potential on electrochemical sulfide oxidation. This was achieved by controlling the potential at fixed levels between 300 mV and 100 mV (vs SHE) at intervals of 50 mV every 24 h. At the start of each potential cycle (except 300 mV cycle), 120  5 mg S L1 of sulfide was spiked into the anode. In addition, sulfide oxidation at 300 mV controlled potential, was also examined with higher concentration (300  5 mg S L1) to observe the effect of the concentration.

2.3. Continuous sulfide oxidation experiments in the fuel cell The continuous experiments were carried out in the same cell as used for the batch tests. The cell was started with freshly treated granules and operated for a period of 3 months. The continuous experiments were conducted to examine the long term performance of the fuel cell. Sulfide was fed at a loading rate of 0.98  0.09 kg S m3 NAC d1 using a peristaltic pump (Watson Marlow, UK) with PVC tubing of 3 mm internal diameter as the feed line. This variability in loading was due to the washing step required after weighing sodium sulfide crystals. A nitrogen balloon was connected to the top of the feed tank to avoid the leakage of oxygen into the feed due to pumping. The anode potential and cell voltage were monitored continuously as in the batch tests. The sulfide concentrations in the feed and anode compartment effluent were monitored through manual sampling and off-line chemical analysis. The cell was operated in open circuit for 8 h every 10 days to investigate possible sulfide oxidation by pathways other than electrochemical.

2.4.

Potentiostatic measurements

Potentiostatic measurements and controls were performed using a VMP3 potentiostat (Princeton Applied Research, USA). An Agilent 34970A data acquisition unit was used to record voltage data every 60 s. Polarization curves during sulfide oxidation were obtained by imposing a linear cell voltage decrease from open circuit to 0 mV at a rate of 0.1 mV s1 and then a linear increase at the same rate. The ohmic resistance, which includes both the resistance to the flow of electrons

4968

water research 42 (2008) 4965–4975

through electrodes and interconnections, and the resistance caused by the movement of ions through electrolytes and the membrane, was determined by the current interrupt method (Lee et al., 1998). The Coulombic efficiency (3cb), defined as the ratio of the charge (as electrons) transferred to the anode from the reactant and the charge that would be yielded if all the removed reactants transferred electrons to the anode (Rabaey et al., 2005), was calculated by assuming the end product of sulfide oxidation was elemental sulfur (to be further discussed in Sections 3 and 4). All these calculations were performed according to Logan et al. (2006).

JSM 6460 LA, Tokyo, Japan). The Secondary Electron (SE) detector provided the shape of the specimen whereas a Back Scattered Electron (BSE) detector produced images with compositional contrast. The apparatus was equipped with Energy Dispersion Spectrometry (EDS) which was used for detection of different elements. SEM images were taken to identify sulfur species and to examine the sulfur deposition pattern on granule surfaces. The surface of the graphite granules was also analysed by a Kratos Axis X-ray Photoelectron Spectroscope (XPS) to identify the oxidation state of sulfur species.

2.5.

2.7.

Chemical analyses

Sulfide, sulfate, thiosulfate and sulfite concentrations were measured by Ion Chromatography (IC), using the Dionex 2010i system, according to Keller-Lehman et al. (2006). Samples collected from the reactor were immediately preserved in previously prepared Sulfide Antioxidant Buffer (SAOB) solution prior to ion chromatography analysis. SAOB solution was also used to dilute the samples in the cases when it was necessary. SAOB solution was prepared using helium purged MilliQ (18 MU) water, 3.2 g L1 NaOH and 2.8 g L1 a-ascorbic acid. After preparation, the solution was kept refrigerated, shielded from light and not used beyond 24 h. The acetate concentration was determined using a LC-10ADVP HPLC system (Shimadzu, Japan). Polysulfide was analysed with a UV/VIS spectrophotometer at 295 nm wavelength (Kleinjan et al., 2005a; Danielsson et al., 1996) (see details Supplementary data, S1). Total dissolved sulfur was calculated by adding the measured sulfate, sulfide, thiosulfate, sulfite and polysulfide concentrations. To quantify elemental sulfur attached to the surface of the graphite granules, granule samples were collected from the top, bottom and middle portions of the chamber after several batches. These granules were grinded manually and microwave digested in PTFE digestion vessels in aqua-regia (3:1 HCl/HNO3) at 150  C and 3000 kPa using an Anton Paar Multiwave microwave system. The contents were heated using a power regime based on US EPA Method 3051. Digest solutions and water samples were diluted to a sample/solution ratio of 1:w10 with 4% aqua-regia prior to analysis and total sulfur concentrations were determined by Inductive Coupled Plasma-Optical Emission Spectroscopy (ICPOES, Perkin Elmer Optima 3300DV). This enables the quantification of the total sulfur in the granules and attached to the surface of the granules. The sulfur content in the granules was measured by subjecting freshly treated granules to the above procedure. The possible attachment of sulfur species other than elemental sulfur to granules was studied in an adsorption experiment. 2 2 Adsorption of sulfate (SO2 4 ), thiosulfate (S2O3 ) and sulfite (SO3 ) on graphite granules in phosphate buffered solution were performed in an air-tight flask. The total surface area of the graphite granules was measured by mercury porosimetry with the method of Favas and Jackson (2003). This method measures the surface area of pores with a minimum size of 6 nm.

2.6.

Electron microscopy and spectroscopy

The microscopic images of graphite granule surfaces were collected using a Scanning Electron Microscope (SEM) (JOEL

Electrode kinetics

The kinetics of an electrode can be described using the Butler– Volmer equation (Barbir, 2005):      aRd;a FðEa  Er;a Þ aOx;a FðEa  Er;a Þ  exp (1) ia ¼ i0;a exp RT RT with ia being the current density (A m2) and i0,a the exchange current density of the anode (A m2) which measures the readiness of an electrode to proceed the electrochemical reaction and its high value represents the high activity of the electrode surface. F is the Faraday constant (96,485 C mol1), R the ideal gas constant (8.31 J mol1 K1), T the absolute temperature (K), Ea the anode potential and Er,a the equilibrium potential of the anode. Ea  Er,a (also denoted as hanode) gives the overpotential of the anode, aa is the transfer coefficient of the anode and Rd and Ox refer to reduction and oxidation. At higher current densities, the equation can be simplified to the Tafel equation (Larminie and Dicks, 2003):   ia aa Fhanode ¼ (2) ln i0;a RT The above equation was used in this study to calculate overpotential at different current densities. The symmetry factor, b, an intrinsic characteristic of the given charge-transfer reaction at the given interface, is also sometimes used instead of the transfer coefficient, a (Bockris and Reddy, 1970). However, the symmetry factor may be used strictly only for a single-step reaction involving a single electron (Barbir, 2005). The relation between a and b is further discussed in Supplementary data, S2.

3.

Results

3.1. Spontaneous sulfide oxidation and electricity generation in batch experiments Sulfide reacted spontaneously at granular graphite electrodes in the anode compartment. Fig. 2 depicts the evolution of the cell voltage, anode and cathode potentials during spontaneous sulfide oxidation in a batch test at pH 6.75 with an external resistor of 50 U. Under these conditions, a maximum power output of 35 W m3 NAC (voltage 0.59 V, current 11.8 mA, anode potential 0.250 V, SHE) was obtained. The cell voltage and consequently the current decreased over time as aqueous sulfide was removed from the solution. As oxidation progressed, the pH of the solution decreased due to the

4969

Cell

Potential/Voltage (V)

0.6 Cathode 0.4 0.2 0.0

0

10

20

30

40

-0.2

Concentration (sulfide-S, mg L-1)

0.8

120

Current (A m-3 NAC)

water research 42 (2008) 4965–4975

200

Anode -0.4

Time (hr) Fig. 2 – Evolution of anode and cathode potentials and cell voltage (V) vs. time (h) during electrochemical sulfide oxidation at pH 6.7 ± 0.05. Dashed arrows indicate sulfide spiking. The initial sulfide-S concentrations after the two doses were 101.2 and 99.3 mg S LL1, respectively. The sulfide concentrations at the end of the two batches were below 0.1 mg S LL1. The external resistance applied was 50 U.

release of protons generated as the sulfide is removed from the solution. In batch mode, the pH dropped from 6.75 to 6.65. The calculated coulombic efficiency for all external resistances was found to be 90  4% (assuming sulfide oxidation to elemental sulfur – see Sections 4.1 and 4.2) Both the sulfide oxidation rate and current production increased with decreased external resistance. This is demonstrated in Fig. 3, which shows a much higher sulfide oxidation rate and current at 10 U in comparison to 50 U. The concentrations of various dissolved sulfur species were measured during sulfide oxidation. The concentrations of 2 both sulfite (SO2 3 ) and sulfate (SO4 ) were negligible (<2 mg S L1) at all times during the batch tests, 5  2% (n ¼ 12) of the sulfide was converted to thiosulfate (S2O3) (all measured as S). It is worthwhile to mention that almost the same amount of thiosulfate was measured during open circuit experiments, suggesting that the formation of thiosulfate was not due to the electrochemical oxidation of sulfide. The formation could have been caused by a minor leakage of oxygen into the anode chamber. At the end of several batch experiments, the sulfur deposited on the graphite granules was quantified. The results are summarised in Table 1. The mass balance indicates that over 95% of the sulfide removed in the reactor was converted to products deposited on the surface of the graphite granules. The X-ray Photoelectron Spectroscopic results (see Supplementary data, Fig. S1) confirmed the presence of elemental sulfur on the surface of the granules after batch tests.

3.2.

Parameters influencing sulfide oxidation

Spontaneous sulfide oxidation occurs in a broad pH range (Fig. 4). The dissolution of sodium sulfide in the absence of a buffer resulted in an initial pH value of 11.4. Polysulfide was

100 80 60 40 20 0

150

100

50

0 0

10

20

30

40

50

Time (hr) Fig. 3 – Profiles of aqueous sulfide concentration and corresponding current at different resistances in batch mode (,) 10 U, (C) 50 U. (pH 6.75), arrows indicate sulfide spiking. The cell voltage in the start up phase was 0.59 V (anode potential – 0.250 V, SHE) for 50 U and 0.4 V (anode potential – 0.082 V, SHE) for 10 U.

formed in the solution during the oxidation at that pH. While the polysulfide concentration increased at the beginning of the oxidation, it decreased after 5 h, likely due to the drop in pH (from 11.4 to 7.3) during the oxidation process. The influence of organic electron donors on sulfide oxidation was examined by adding acetate in a concentration range of 100–1000 mg COD L1. In all cases, the sulfide oxidation rates were found to be identical to those measured in the absence of acetate, indicating the presence of acetate does not affect sulfide oxidation. The COD concentration remained unaltered during each test, confirming that acetate was not oxidized in the cell. The sulfide oxidation rate decreased with the decrease of the anode potential as shown in Fig. 5. Oxidation of sulfide could not be obtained at a potential of 300 mV and the current reversed at that point. At that potential, sulfur products previously deposited on the electrode surface were apparently reduced to sulfide. This was evidenced by the increased sulfide concentration measured in the liquid phase. The rate of release was dependent on the dissolved sulfide concentration. When the sulfide concentration was increased to 300  5 mg/L, a positive current could be obtained at 300 mV (SHE) (results not shown), which can be expected based on the thermodynamic driving forces as expressed in the Nernst equation with sulfide as electron donor. At 250 mV, sulfide was oxidized at a very slow rate in the experiment as shown in Fig. 5.

4970

water research 42 (2008) 4965–4975

Table 1 – Mass balances of sulfur components after oxidation (wt. of granules 430 g, total liquid volume 1.2 L) Total no. of batches

Cumulative feed sulfide conc., A (mg S)

Cumulative dissolved sulfur species 2 2 (S2O2 3 , SO3 , SO4 ) measured by IC, B (mg S)

Measured S attached to granules by ICP-OES, C (mg S)

Unaccounted for sulfur (%) ½ðABÞðCDÞ100 ðABÞ

438  41 1131  54 3587  564

4.3 2.9 5.5

Sulfur content of freshly treated granules, 9.8  1.8 mg S (D) 4 478 31 10 1238 83 30 3624 235

HS þ

x1 þ S8 5S2 x þH 8

3.3.

(3)

Continuous operation of the fuel cell

The electrochemical oxidation cell was run continuously for 3 months at a loading rate of 0.98  0.09 kg S m3 NAC d1 with an external resistance of 10 U. In the first 8 days, a sulfide removal rate of 0.80  0.09 kg S m3 NAC d1 was attained, giving rise to a removal efficiency of 80  2% (Fig. 6). The sulfide removal rate then decreased to 0.62  0.1 kg S m3 NAC d1 (62  4% removal) and remained at this level over the following 50 days, after which the sulfide removal rate decreased sharply. When the cell was operated at open circuit for 8 h every 10 days, no sulfide oxidation was observed suggesting that sulfide oxidation was caused by electrochemical reaction. The calculated coulombic efficiency during the continuous experiment was found to be 88  5%. In the first 2 months of continuous operation, on average power production of 12  2 W m3 NAC (5  1 W m3 TAC) was estimated (calculated from the data presented in Fig. 6), with a maximum of

52  3 W m3 NAC (23  2 W m3 TAC) (Fig. 7) at the loading rate of 0.98  0.09 kg S m3 NAC d1. However, this did not represent the maximum power production capacity of the cell. Fig. 7 compares the shapes and heights of the polarization curves and the power generated between day-1 and day-60. This figure also shows that a power density of 166 W m3 NAC (74 W m3 TAC) was obtained when 500  10 mg S L1 of sulfide was spiked to the cell at the beginning of the cell operation. The shape and height of the polarization curves remained almost unchanged during the first 30 days of operation (not shown separately as it overlaps the day-1 curve). However, the maximum value of power generation reduced to about 22  3 W m3 NAC (10  2 W m3 TAC) after 60 days of operation. This correlated well with the results that the ohmic resistance of the cell increased from 1.0  0.4 U to 9  1 U, as measured by the current interrupt method, after 60 days of operation. The total surface area of the graphite granules (including pores 6 nm size and above) was measured to be

250

Current ( A m-3 NAC)

Considerable concentrations of polysulfide were found directly after the sulfide spiking to the reactor. This occurred at all potentials except at 100 mV (SHE). Polysulfide (S2 x , x > 1) was likely formed from the chemical dissolution of elemental sulfur according to the following equation:

200 150 100 50 0

80

11

60

10

40

9

20

8

0

7

-100mV

-150mV

-200mV

-250mV

-300mV

250

Concentration (mg-S L-1)

12

pH

Concentration (mg-S L-1)

-50 100

200 150 100 50 0

0

5

10

15

20

25

Time (hr) Fig. 4 – Evolution of polysulfide concentration in a nonbuffered solution and pH profile during spontaneous sulfide oxidation, (C) polysulfide-S, (-) sulfide-S, and (:) pH [ext. resistance 50 U].

0

20

40

60

80

100

120

Time (hr) Fig. 5 – Concentration profiles of different aqueous sulfur species and current at different controlled anode potentials. (6) Total-S, (C) sulfide-S, (;) polysulfide-S, (B) thiosulfate-S, [sulfide spiked at the beginning of each cycle except L300 mV cycle, pH 6.75, n [ 4].

4971

water research 42 (2008) 4965–4975

100

A

0.8

0.6 a

60

Voltage (V)

% removal

80

40

b

0.4

20

c

0.2 0

0

15

30

45

60

75

90

Time (days)

0.0 0

Fig. 6 – Percentage removal of sulfide when the reactor was operated continuously at a loading rate of 0.98 ± 0.09 kg S mL3 NAC dL1, pH 6.75, ext. resistance 10 U.

200

B 180

4.

Discussion

4.1. Spontaneous aqueous sulfide oxidation and electricity generation Aqueous sulfide removal by spontaneous electrochemical oxidation with concomitant electricity generation was demonstrated at ambient temperature, pressure and neutral pH. The experimental results demonstrate that the rate of the reaction can be controlled by changing the external resistance as it determines the anode potential for a given current. Sulfur has many oxidation states; a number of aqueous species may have evolved in parallel or in consecutive redox reactions as their calculated theoretical potentials were closely matched. The quantitative analysis of the sulfide oxidation products and current strongly supports the finding that the reaction

600

800

a

Power (W m-3NAC)

160

6.2  106 m2 m3. This value was used for the calculation of the current density. The calculated overpotential of the anode and cathode at different current densities can be seen in Fig. 8. Based on linear sweep voltammetry, performed on the anode/ cathode (Rabaey et al., 2007), i0,a was 2.3  103 mA m2 and aa ¼ 0.69 during spontaneous sulfide oxidation at pH 6.75 at a loading rate of 1 kg S m3 NAC d1. After 60 days of continuous operation, i0,a was 1.7  103 mA m2 and aa ¼ 0.3. The values of the exchange current density of the cathode, i0,c and ac were 12.9  103 mA m2 and 0.85, respectively. For comparison purpose, in the calculations we have assumed that the surface area of the anode was not altered due to the precipitation of elemental sulfur. In contrast to the anode, the i0,c and ac values did not notably change after 60 days of continuous operation. Back Scattered Electron images (Fig. 9) taken of a fresh granule and a granule after 60 days of continuous reactor operation show that the latter (Fig. 9-B) retained very little open graphite surface available for direct contact between the electrode and the solution. The white color in the picture shows contrast due to the presence of sulfur on carbon, as confirmed by Electron Dispersion Spectrometry (results not shown).

400

Current (A m-3 NAC)

140 120 100 80 60

b

40 20 c

0 0

200

400

600

800

Current (Am-3NAC) Fig. 7 – (A) Polarization curve. (B) Power curve on different days during spontaneous sulfide oxidation when the cell was operated continuously at a loading rate of 0.98 ± 0.09 kg S mL3 NAC dL1, pH 6.75. (a) day-1 (with extra spiked sulfide, 500 ± 10 mg S LL1), (b) day-1 and (c) day-60.

H2S/HS / S(s) þ 2e þ 2Hþ/Hþ is primarily responsible for sulfide oxidation in this system (see further in Supplementary data, Table S1). During continuous operation, the cell showed good removal efficiency without regeneration of the electrode over an extended time period (up to 60 days). The observed removal rate could be further increased by lowering the external resistance. However, the electrode might have started to be passivated to some degree due to deposited sulfur even after 8 days of operation since the removal efficiency decreased from 80  2% to 62  4%. The shape and height of the polarization curve almost remained unaltered in the first 30 days of operation, corresponding to an unaltered removal efficiency (62  4%). This was likely due to the fact that the fuel cell was operated at the left side of the polarization curve (45  10 A m3 NAC), hence far below the attainable, maximal current. At that point, the cell performance was presumably not affected by the formation of a sulfur deposit on the granules. The operational point of the cell can easily be shifted by changing the value of the external resistance. However, after 60 days of operation, the height of the polarization curve changed considerably, which implies that even at the left side

4972

water research 42 (2008) 4965–4975

Overpotetnial,η (V)

0.5 0.4 0.3 0.2 0.1 0 0

0.1

0.2

0.3

0.4

0.5

Current density, i (mA/m2) Fig. 8 – Calculated anodic and cathodic overpotentials at different current densities. (,) Anode [day-1], (-) anode [day-60], and (B) Cathode.

of the curve the sulfur film started to negatively affect the removal rate and power generation. By that time, about 7.8 g of sulfur (equivalent to 17.3 kg m3 TAC) was stored in the reactor, as calculated from the sulfur balances. The ohmic resistance of the cell drastically increased (from 1.0  0.4 U to 9  1 U) likely due to the sulfur deposition, which caused a loss of direct contact between the granules (limiting electron transfer within the electrode bed) and between granules and the bulk liquid (limiting sulfide oxidation at the electrode). The sulfide removal significantly decreased to approximately 10% after 90 days. With an MFC system, Rabaey et al. (2006) achieved removal of sulfide available either as a single source of electron donor or contained in UASB effluents. Micro-organisms were found and a possible role of these organisms was postulated. In comparison, this study clearly shows that sulfide removal can be achieved without the presence of bacteria. The results presented in this paper could also indicate that the processes described by Rabaey et al. (2006) may have been electrochemical or at least partially electrochemical. However, it should be noted that the finding that sulfide can be electrochemically oxidized does not imply that it cannot happen bioelectrochemically. The fact that the presence of a biofilm does not notably change the sulfide removal rate does not

necessarily imply that bio-electrochemical sulfide oxidation does not happen. In an MFC system, electrochemical and bioelectrochemical processes might proceed at the same time but the combined sulfide oxidation rate might not be altered in case that the reaction rate is limited by the electrode or by the fuel cell operation. The focus of this study was solely on understanding the anodic processes in detail while the cathodic process was only utilized as an electron sink. Potassium ferricyanide, used as electron acceptor in the cathode, is highly useful for such studies as it provides a constant and easily replenished electron sink, which does not interfere with the anodic processes. But, it is clearly not suitable for any practical applications due to environmental concerns and the limited re-oxidation of ferricyanide by oxygen (Oh et al., 2004). Open air cathodes or biologically catalysed cathodes (Freguia et al., 2008; Clauwaert et al., 2007; Freguia et al., 2007b) appear better suited, provided any significant oxygen flux from the cathode to the anode can be prevented. Further research is needed to develop more practically suitable cathodes such as those that use oxygen or nitrate as the ultimate electron acceptor. The true surface area of the granular electrode was used to obtain the values of exchange current density, i0 both for the and the cathode anode (2.3  103 mA m2) (12.9  103 mA m2). After 60 days of operation, the value of i0,a decreased from 2.3  103 mA m2 to 1.7  103 mA m2 and aa from 0.69 to 0.3 respectively. The calculated anode overpotential clearly had increased (Fig. 8), likely due to deposited sulfur on the graphite granules, which eventually changed the kinetics of the interface reaction between the electrode and solution. This deposit can also be seen in Fig. 9. The value of i0 for the ferricyanide cathode was similar to the value for a platinized oxygen cathode (Barbir, 2005). Many important sulfide electrolysis studies used platinum (Szynkarczuk et al., 1994; Farooque and Fahidy, 1977) as catalyst, which entails a substantial cost. In contrast, the non-catalyzed graphite granules used in this study are quite cost effective (US$1.75 per kg) and can also work with oxygen as final electron acceptor (Freguia et al., 2007b). Recently, high cost (V1.9– 7.2 kg1 S removed) was mentioned as the main problem of the existing sulfide control technologies in sewage systems (Zhang et al., 2008). Even environmentally benign biotechnological or microbiological sulfide removal processes involve aeration costs

Fig. 9 – Back Scattered Electron (BSE) images of Scanning electron microscope (SEM) before and after sulfur deposition. (A) – fresh graphite granule and (B) – granules with sulfur deposition (white color).

water research 42 (2008) 4965–4975

and stringent control (Syed et al., 2006). The process described here can remove sulfide from the solution, potentially at a low operational cost. Detailed cost analysis of the entire process is required to compare the cost effectiveness of this process with other technologies. Besides this, the removal of particulate materials might be required before feeding certain types of wastewaters into the anode as those materials might block the flow channels or the electrode surface. Further research is needed to establish whether and to what extent these particulates would influence the cell performance.

4.2.

Final product of electrochemical sulfide oxidation

Elemental sulfur was the primary final product of electrochemical oxidation. Table 1 shows that the majority (>95%) of the oxidation products was deposited onto the surface of electrodes. The presence of elemental sulfur on the electrode surface was confirmed by ICP-OES, XPS and SEM analysis. More quantitative evidence can be found through the electron balance. The coulombic efficiency of the batch experiment is determined to be 90  4% if we assume elemental sulfur is the sole final product of oxidation. With the same assumption, the coulombic efficiency during the continuous experiment was 88  5%. Both strongly support the hypothesis that elemental sulfur is likely the primary product of electrochemical oxidation of sulfide. The approximate 10% loss of electrons was likely due to autooxidation of sulfide to other dissolved sulfur species such as thiosulfate. In each test, approximately 5% (by mass of sulfur) of the sulfide was oxidized to thiosulfate. The fact that a similar amount of thiosulfate was produced also during open circuit operation suggests the presence of other electron acceptors (e.g. oxygen) in the solution and electrons associated with this oxidation process are not transferred to the electrode. The good sulfur balance also indicates that there was negligible transfer of H2S gas from the anode to the cathode through the membrane. Chuang et al. (2000) also ruled out possible cross over of H2S gas through Nafion membranes as used in their study. An important key difference between Rabaey et al. (2006) and the current work is that we investigated the fate of sulfur and the effects of a number of operational parameters. The good mass and charge balances derived in this study suggest that the causes for the far lower charge recovery and the poor sulfur mass balances in the MFC experiments reported by Rabaey et al. (2006) might be either biological factors (biomass growth) and/or the result of incomplete balance analysis. High concentrations of polysulfide were detected as an intermediate when batch experiments were performed without pH control, during which the pH varied between 11.4– 7.3. Polysulfide could have been formed either by chemical dissolution of elemental sulfur by sulfide (Eq. (3)), or as an intermediate by the following two-step reaction proposed for sulfide oxidation in alkaline media (Szynkarczuk et al., 1994):  nHS þ nOH ¼ S2 n þ nH2 O þ 2ðn  1Þe

(4)

o  S2 n ¼ nS þ 2e

(5)

The concentration of polysulfide depends on the anode potential, pH, sulfide concentration and the amount of elemental sulfur present on the electrode surface. The polysulfide formed

4973

eventually spontaneously oxidized. Polysulfide anions can exist in water only at relatively high pH levels (>6–7) (Steudel, 2000; Kleinjan et al., 2005b). This probably explains the observation that polysulfide was negligible in batch tests where the pH was maintained at around 6.75. The anode potential may also play an important role in the formation of polysulfide.

4.3. Potential dependent sulfide oxidation and (re)active sulfur Controlling the anodic potential demonstrated that sulfide does not oxidize at potentials of 300 mV (SHE) at dissolved concentrations of up to 150 mg S L1. At these low potentials which are below the open circuit potential at that particular concentration, elemental sulfur present on the granular graphite was reduced to sulfide or polysulfide. These observations indicate that elemental sulfur produced by spontaneous electrochemical oxidation remains (re)active, i.e. it could be further oxidized or reduced (Zhu et al., 1994). Electrochemically converted elemental sulfur might not be S8, which was found for showing lack of electrochemical reactivity (Szynkarczuk et al., 1994). At higher anodic potentials, the sulfide oxidation rate increased and up to 100 mV (SHE), elemental sulfur remained the only significant final oxidation product. A further increase of the potential up to þ150 mV (SHE) did not create significantly different oxidation products (results not shown) compared to those obtained at 100 mV (SHE).

4.4.

Potential applications and future research

Sulfide removal from sewage systems and anaerobic digester effluents can be the first line applications of this process. Sulfide was often detected at low concentrations in sewage. Sharma et al. (2008) reported a maximum dissolved sulfide concentration of 15 mg S L1 in their study on two rising mains sewers in Australia. Higher sulfide concentration was also reported in literature, for instance 47 mg S L1 in a sanitary network in Kuwait (Tomar and Abdullah, 1994) and 400 mg S L1 in the effluent of a specific anaerobic digester (Velasco et al., 2008). Taking into account the often low concentrations of sulfide in wastewater, it appears unlikely that the generation of electricity from this removal becomes a significant goal. However, the results presented in this paper show that sulfide can be removed effectively and without the expense of energy using the electrochemical oxidation approach. The spontaneity of the sulfide oxidation reaction was unaffected by the presence of organic carbon compounds (acetate), which were not oxidized due to the absence of a catalyst, e.g. bacterial biofilm. Likewise, we do not expect interference by other organic electron donors due to the large overpotential required to achieve their abiotic oxidation at non-catalyzed anodes. These features offer an opening to make the process suitable for selective sulfide removal in sewage since the commonly used oxidants oxygen and nitrate also oxidize organic carbon compounds while oxidizing sulfide. The organic carbon compounds often need to be preserved for the downstream nutrient removal wastewater treatment plants. However, the formation of a biofilm on the electrode needs to be avoided in this case; otherwise the reactor will behave as a microbial fuel cell where sulfide and

4974

water research 42 (2008) 4965–4975

organic carbon compounds can be removed simultaneously (Rabaey et al., 2006). Research is underway to develop suitable strategies to avoid biofilm formation or initiate periodic biofilms removal when a fuel cell is continuously fed with wastewater. However, the presence of biofilms does not inhibit the ability of the fuel cell to oxidize sulfide as demonstrated by Rabaey et al. (2006). A major limitation of the method is the decrease in electrochemical activity over time due to the deposition of elemental sulfur on granules. An efficient method to re-activate the electrode and recover sulfur from the electrode surface needs to be developed. Different approaches, such as potential switching or acid/base treatment, are currently under investigation to regenerate the anode material.

5.

Conclusions

In this study, aqueous sulfide oxidation and simultaneous power generation was demonstrated in a fuel cell system with the following key characteristics: Sulfide oxidizes spontaneously in an abiotic fuel cell reactor while generating electricity. The presence of organic compounds did not affect this process. - Elemental sulfur is the major final product of the electrochemical oxidation of sulfide. The elemental sulfur deposited on the anode surface remains (re)active in the fuel cell and it reduces to sulfide at lower redox potential (w300 mV vs SHE at 150 mg S L1). 3 - A maximum power of 166 W m NAC was obtained and the power generation depends on the amount of sulfide in the liquid stream. - The precipitated sulfur forms a barrier towards the further oxidation of sulfides in longterm experiments. Sustainable solutions to remove the precipitate need to be developed in future research. -

Acknowledgements We sincerely thank Dr. Tony Jong for ICP-OES analysis, Dr. Barry Wood for XPS analysis, Ms Ying Yu for SEM analysis. Paritam K. Dutta and Korneel Rabaey thank the University of Queensland for scholarship and fellowship support. This work was funded by the Australian Research Council (Grant DP0666927).

Supplementary data This section contains the detailed method of polysulfide analysis, information on electrode kinetics parameter, XPS results of graphite granules, table for important sulfur half cell reactions. Supplementary data associated with this article can be found in the online version, at doi:10.1016/j.watres.2008.09. 007.

references

Aguilar, L., Zha, S., Cheng, Z., Winnick, J., Liu, M., 2004. A solid oxide fuel cell operating on hydrogen sulfide (H2S) and sulfurcontaining fuels. J. Power Sources 135 (1–2), 17–24. Ateya, B.G., AlKharafi, F.M., Al-Azab, A.S., 2003. Electrodeposition of sulfur from sulfide contaminated brines. Electrochem. Solid-State Lett. 6 (9), C137–C140. Barbir, F., 2005. PEM Fuel Cells: Theory and Practice. Elsevier Academic Press. Bockris, J.O.M., Reddy, A.K.N., 1970. Modern Electrochemistry. Plennum Press, New York. Chuang, K.T., Donini, J.C., Sanger, A.R., Slavov, S.V., 2000. A protonconducting solid state H2S-O2 fuel cell: 2. Production of liquid sulfur at 120–145  C. Int. J. Hydrogen Energy 25 (9), 887–894. Clauwaert, P., vanderHa, D., Boon, N., Verbeken, K., Verhaege, M., Rabaey, K., Verstraete, W., 2007. Open air biocathode enables effective electricity generation with microbial fuel cells. Environ. Sci. Technol. 41 (21), 7564–7569. Danielsson, L.G., Chai, X.S., Behm, M., Renberg, L., 1996. UV characterization of sulphide-polysulphide solutions and its application for process monitoring in the electrochemical production of polysulphides. J. Pulp. Pap. Sci. 22 (6), J187–J191. Farooque, M., Fahidy, T.Z., 1977. Low potential oxidation of hydrogen sulfide on a rotating tripolar wiper-blade electrode via continuous anode reactivation. Electrochem. Sci. Technol. 124 (8), 1191–1195. Favas, G., Jackson, W.R., 2003. Hydrothermal dewatering of lower rank coals. 1. Effects of process conditions on the properties of dried product. Fuel 82 (1), 53–57. Freguia, S., Rabaey, K., Yuan, Z., Keller, J., 2007a. Electron and carbon balances in microbial fuel cells reveal temporary bacterial storage behavior during electricity generation. Environ. Sci. Technol. 41 (8), 2915–2921. Freguia, S., Rabaey, K., Yuan, Z., Keller, J., 2007b. Non-catalyzed cathodic oxygen reduction at graphite granules in microbial fuel cells. Electrochim. Acta 53 (2), 598–603. Freguia, S., Rabaey, K., Yuan, Z., Keller, J., 2008. Sequential anodecathode configuration improves cathodic oxygen reduction and effluent quality of microbial fuel cells. Water Res. 42 (6–7), 1387–1396. Janssen, A.J.H., Lettinga, G., de Keizer, A., 1999. Removal of hydrogen sulphide from wastewater and waste gases by biological conversion to elemental sulphur: colloidal and interfacial aspects of biologically produced sulphur particles. Colloids Surf. A: Physicochem. Eng. Aspects 151 (1–2), 389–397. Kameyama, T., Dokiya, M., Fujishige, M., Yokokawa, H., Fukuda, K., 1981. Possibility for effective production of hydrogen sulfide by means of a porous vycor glass membrane. Ind. Eng. Chem. Fundam. 20 (1), 97–99. Keller-Lehman, B., Corrie, S.K., Ravn, R., Yuan, Z., Keller, J., 2006. Preservation and simultaneous analysis of relevant soluble sulfur species in sewage samples. In: Proceedings of Second International IWA Conference on Sewer Operation and Maintenance. pp. 339–346. Kelsall, G.H., Yin, Q., Vaughan, D.J., England, K.E.R., Brandon, N.P., 1999. Electrochemical oxidation of pyrite (FeS2) in aqueous electrolytes. J. Electroanal. Chem. 471 (2), 116–125. Kleinjan, W.E., de Keizer, A., Janssen, A.J.H., 2005a. Equilibrium of the reaction between dissolved sodium sulfide and biologically produced sulfur. Colloids Surf. B: Biointerf. 43 (3–4), 228–237. Kleinjan, W.E., Keizer, A.d, Janssen, A.J.H., 2005b. Kinetics of the chemical oxidation of polysulfide anions in aqueous solution. Water Res. 39 (17), 4093–4100. Larminie, J., Dicks, A., 2003. Fuel Cell Systems Explained. John Wiley & Sons Ltd, Chichester.

water research 42 (2008) 4965–4975

Lee, C.G., Nakano, H., Nishina, T., Uchida, I., Kuroe, S., 1998. Characterization of a 100 cm [sup 2] class molten carbonate fuel cell with current interruption. J. Electrochem. Soc. 145 (8), 2747–2751. Logan, B.E., Hamelers, B., Rozendal, R., Schroder, U., Keller, J., Freguia, S., Aelterman, P., Verstraete, W., Rabaey, K., 2006. Microbial fuel cells: methodology and technology. Environ. Sci. Technol. 40 (17), 5181–5192. Oh, S., Min, B., Logan, B.E., 2004. Cathode performance as a factor in electricity generation in microbial fuel cells. Environ. Sci. Technol. 38 (18), 4900–4904. Rabaey, K., Clauwaert, P., Aelterman, P., Verstraete, W., 2005. Tubular microbial fuel cells for efficient electricity generation. Environ. Sci. Technol. 39 (20), 8077–8082. Rabaey, K., Rodriguez, J., Blackall, L.L., Keller, J., Gross, P., Batstone, D., Verstraete, W., Nealson, K.H., 2007. Microbial ecology meets electrochemistry: electricity-driven and driving communities. ISME J. 1 (1), 9–18. Rabaey, K., VandeSompel, K., Maignien, L., Boon, N., Aelterman, P., Clauwaert, P., DeSchamphelaire, L., Pham, H.T., Vermeulen, J., Verhaege, M., Lens, P., Verstraete, W., 2006. Microbial fuel cells for sulfide removal. Environ. Sci. Technol. 40 (17), 5218–5224. Rajeshwar, K., Ibanez, J.G., Swain, G.M., 1994. Electrochemistry and the environment. J. Appl. Electrochem. 24 (11), 1077–1091. Reimers, C.E., Girguis, P., Stecher, H.A., Tender, L.M., Ryckelynck, N., Whaling, P., 2006. Microbial fuel cell energy from an ocean cold seep. Geobiology 4 (2), 123–136. Ryckelynck, N., Stecher, H.A., Reimers, C.E., 2005. Understanding the anodic mechanism of a seafloor fuel cell: interactions between geochemistry and microbial activity. Biogeochemistry 76 (1), 113–139. Sharma, K.R., Yuan, Z., de Haas, D., Hamilton, G., Corrie, S., Keller, J., 2008. Dynamics and dynamic modelling of H2S

4975

production in sewer systems. Water Res. 42 (10–11), 2527–2538. Steudel, R., 2000. The chemical sulfur cycle. In: Lens, P.N.L., Pol, L.H. (Eds.), Environmental Technologies to Treat Sulfur Pollution. IWA Publishing, London. Syed, M., Soreanu, G., Falletta, P., Beland, M., 2006. Removal of hydrogen sulfide from gas streams using biological processes – a review. Can. Biosyst. Eng. 48, 2.1–2.14. Szynkarczuk, J., Komorowski, P.G., Donini, J.C., 1994. Redox reactions of hydrosulphide ions on the platinum electrode – I. The presence of intermediate polysulphide ions and sulphur layers. Electrochim. Acta 39 (15), 2285–2289. Tomar, M., Abdullah, T.H.A., 1994. Evaluation of chemicals to control the generation of malodorous hydrogen sulfide in waste water. Water Res. 28 (12), 2545–2552. Vaiopoulou, E., Melidis, P., Aivasidis, A., 2005. Sulfide removal in wastewater from petrochemical industries by autotrophic denitrification. Water Res. 39 (17), 4101–4109. Velasco, A., Ramı´rez, M., Volke-Sepu´lveda, T., Gonza´lezSa´nchez, A., Revah, S., 2008. Evaluation of feed COD/sulfate ratio as a control criterion for the biological hydrogen sulfide production and lead precipitation. J. Hazard. Mater. 151 (2–3), 407–413. Zaman, J., Chakma, A., 1995. Production of hydrogen and sulfur from hydrogen sulfide. Fuel Process. Technol. 41 (9), 159–198. Zhang, L., De Schryver, P., De Gusseme, B., De Muynck, W., Boon, N., Verstraete, W., 2008. Chemical and biological technologies for hydrogen sulfide emission control in sewer systems: a review. Water Res. 42 (1–2), 1–12. Zhu, X., Li, J., Wadsworth, M.E., 1994. Characterization of surface layers formed during pyrite oxidation. Colloids Surf. A: Physicochem. Eng. Aspects 93, 201–210.