The removal of dissolved sulfides in water and wastewater by nanoscale zero-valent iron (nZVI) was examined in the study. Both laboratory batch studies and a pilot test in a 50,000-pig farm were conducted. Laboratory studies indicated that the sulfide removal with nZVI was a function of pH where an increase in pH decreased removal rates. The pH effect on the sulfide removal with nZVI is attributed to the formation of FeS through the precipitation of Fe(II) and sulfide. The saturated adsorption capacities determined by the Langmuir model were 821.2, 486.3, and 359.7 mg/g at pH values 4, 7, and 12, respectively, for nZVI, largely higher than conventional adsorbents such as activated carbon and impregnated activated carbon. The surface characterization of sulfide-laden nZVI using XPS and TGA indicated the formation of iron sulfide, disulfide, and polysulfide that may account for the high adsorption capacity of nZVI towards sulfide. The pilot study showed the effectiveness of nZVI for sulfide removal; however, the adsorption capacity is almost 50 times less than that determined in the laboratory studies during the testing period of 30 d. The complexity of digested wastewater constituents may limit the effectiveness of nZVI. Microbial analysis suggested that the impact of nZVI on the change of microbial species distribution was relatively noticeable after the addition of nZVI.

1. Introduction

Anaerobic digestion of piggery wastewater is capable of producing methane-rich biogas and minimizing the environmental impact caused by the wastewater. The methane-rich biogas is a renewable energy resource that can be used to generate electricity. On the other hand, the Kyoto Protocol has explicitly defined methane as one of the six key greenhouse gases where the global warming potential of methane is 25 times higher than that of carbon dioxide [1]. Thus, the use of methane for electricity generation can not only provide energy but also reduce the methane emission. According to the most recent statistics from the Council of Agriculture in Taiwan, there are 128 pig farms with 5,000 or more swines, an economically sound scale for methane to energy projects [2]. With reference to the methodology based on the Kyoto Protocol mechanisms, this figure represents a potential of over 46,000 megawatt hour (MWh) annually and 800,000 tons of carbon dioxide equivalent () emission reductions for trading carbon credits.

The biogas consists mainly of methane (50–80%), carbon dioxide (20–50%), and a mix of trace gases including nitrogen (1–4%), hydrogen sulfide (50–5,000 ppm), and others. Hydrogen sulfide is highly corrosive and quickly fractures the cast iron and steel used for many electric generators. Farmers have shown little interest in applying biogas for electricity production because hydrogen sulfide has detrimental effects on the equipment. Therefore, hydrogen sulfide must be removed before biogas can be further utilized.

Several methods for hydrogen sulfide removal have been developed including biological fixation process, liquid-phase absorption with either water or NaOH solution, and adsorption on solids such as activated carbon, iron hydroxide or oxide [36]. Among these technologies, adsorption with high adsorption capacity is recognized to be an energy efficient technology for hydrogen sulfide removal. Because of nanotechnology providing high surface areas of nanomaterials, developing a nanoscale adsorbent with high adsorption capacity becomes promising for hydrogen sulfide removal [7, 8].

The nanoscale zero-valent iron (nZVI) represents an important nanotechnology of the environmental remediation that has been developed since 1996 [9]. The nZVI with a particle size of 1–100 nm has been implemented in the field remediation of contaminated groundwater by direct injection technology [1012]. Extensive studies have indicated the ability of nZVI to remove a wide variety of environmental contaminants including PCBs, chlorinated aliphatic and aromatic hydrocarbons, chlorinated pesticides, heavy metals, and inorganic ions (e.g., nitrate and perchlorate) [8, 9, 1315]. It has been reported that nZVI has an adsorption capacity towards arsenic ranging from 9 to 174 times greater than conventional ZVI because of its relatively high specific surface areas [16]. The high adsorption capacity of nZVI has been documented for the control of malodorous sulfide-containing compounds generated in biosolids [7, 8].

Though nZVI has been demonstrated to effectively remove sulfide species in the bench scale, the feasibility of nZVI for field applications of sulfide removal has not yet been investigated. In this work, we present here the use of nZVI for effective removal of hydrogen sulfide from the laboratory scale to the field pilot-scale in anaerobically digested piggery wastewater. The pilot test was carried out in a 50,000-pig farm in Taiwan. The objectives of this study include (1) investigating the mechanisms of dissolved sulfide removal by nZVI, (2) exploring the technical feasibility of nZVI for sulfide removal at a pilot scale, and (3) examining the potential impact of nZVI on the microbial community in the anaerobical wastewater.

2. Experimental Methods

2.1. Chemicals and Materials

For the laboratory study, all chemicals were reagent grade or above and used without further purification. Phosphoric acid (85%) and sodium nitrate (99%) were obtained from Scharlau and Showa, respectively. Sodium borohydride (NaBH4, 98%), sodium sulfide, and ferric chloride (, 98%) were obtained from Aldrich. However, commercial grade chemicals, purchased from Kanto Chemical Co. Inc., Japan, were applied to the pilot test in the field.

2.2. Synthesis of nZVI

In the laboratory studies, synthesis of nZVI was carried out by adding 1 : 1 volume ratio of NaBH4 (0.25 M) into (0.045 M) solution. The borohydride to ferric iron ratio was 7.4 times higher than that of the stoichiometric requirement (1). A detailed synthetic procedure has been described elsewhere [17].

Consider the following:

The suspension was mixed vigorously under room temperature (°C). The iron particles were magnetically collected and then washed with a large volume of Milli-Q water for at least three times and were used without further treatments unless indicated otherwise.

In the field test, nZVI was produced on site by a batch reactor where 250 g of nZVI was produced in each batch. Approximately 1210 g of was dissolved in 10 L water and slowly added into the 10 L NaBH4 aqueous solution (0.45 M). After the completion of the reaction, nZVI was magnetically collected and washed thoroughly with tap water.

It should be pointed out that the characterization of borohydride-synthesized nZVI, which possesses a core-shell morphology with zero-valent iron as the core and iron oxide/hydroxide in the shell has been widely published in many literatures. The detail characterization of nZVI synthesized by the borohydride reduction can be found elsewhere [1820].

2.3. Batch Experiments

For the study of sulfide removal, experiments were conducted in 150 mL serum bottles containing 1–5 g/L nZVI in 100 mL aqueous solution at °C. Initial concentrations of sulfide were about 1,000 mg/L. Reaction vessels were placed on a rotary shaker (150 rpm). The solution pH was adjusted at the beginning of the reaction using 1 M HCl or NaOH and monitored periodically throughout the experiment. The adsorption isotherm study was carried out by varying the initial sulfide concentration from 200 to 1,000 mg/L in 100 mL aqueous solution at the nZVI dose of 1.0 g/L at °C.

2.4. Pilot Studies

Pilot studies were performed in a pig farm in Pingtung, Taiwan. The pig farm has 50,000 pigs that produce 600–700 m3/d wastewater. The anaerobic digestion tanks have a total volume of 8,640 m3 resulting in a hydraulic retention time of approximately 15 d. In pilot tests, the anaerobically digested piggery wastewater was fed into a 1,000 L reaction tank containing 1 kg nZVI at the flow rate of 66 L/d to achieve the same hydraulic retention time. A second tank without nZVI was run in parallel under the same conditions (Figure 1). Physicochemical parameters of wastewater quality including temperature, pH, dissolved oxygen (DO), and oxidation-reduction potential (ORP) were measured in the field using a YSI 650 MDS-6600 probe (V2-4 Sonde, YSI Inc.).

2.5. Identification of Anaerobic Bacteria

All the materials of selection and identification of bacteria were purchased from Creative Microbiologicals, Ltd. Taiwan. The fresh specimens were collected from the pig farm in Pingtung, Taiwan. Fresh specimens were inoculated to anaerobic blood agar plate (anBAP) and the anBAP plates were put into MGC AnaeroPack Jar and cultured at 37°C 24–48 h. According to the morphology of colony growth on anBAP plate, the different colonies were selected and inoculated once again to anBAP plate until one single morphology of colony was cultured. Each fresh single colony was selected in creating slide smear and stained by Gram’s stain [21] according to the manufacturer. The fresh single colony was selected and applied to the tube of RapID ANA II system to make an equivalent McFarland turbidity suspension. The entire contents were then transferred into the right upper corner of the panel and incubated at 37°C for 4 h. The anaerobic bacteria were identified by RapID ANA II ID Code.

2.6. Gene Sequencing and Identification

Samples are stored at −20°C for further identification. To clone 16S rRNA, the polymerase chain reaction (PCR) was performed in MyCyclerTM Thermal Cycler (Bio-Rad) according to manual and primers were designed based on NCBI database and the literatures [22, 23]. PCR products were separated on an agarose gel, eluted, and sequenced (Protech Technology Enterprise, Taipei, Taiwan).

2.7. DNA Extraction

Samples (0.1 g) were extracted with Lysis buffer (100 mM Tris-HCl, 100 mM EDTA, 0.75 M sucrose), resuspended, and lysed with 1 μg/μL lysozyme and 0.1 μg/μL achromopeptidase for 30 min at 37°C. Proteinase K (0.2 μg/μL) and 20% sodium dodecyl sulphate (SDS) were added for reaction at 37°C for 2 h; then, 10% cetyl trimethylammonium bromide (CTAB) was added at 60°C for 30 min. Supernatant was collected and reacted with Proteinase K for 30 min at 60°C. Equal amount of phenol/chloroform/isoamyl alcohol (25 : 24 : 1) was added and centrifuged at 14,000 rpm for 5 min. Supernatant was collected, mixed with 700 μL of phenol/chloroform/isoamyl alcohol, and centrifuged for 10 min. Supernatant was mixed with 500 μL of chloroform/isoamyl alcohol and centrifuged for 10 min. Equal amount of isopropanol was added. DNA was precipitated at −20°C. Pellet was washed with iced alcohol (75%), dried, and dissolved in H2O.

2.8. Analytic Methods

Dissolved sulfide concentrations in the water were analyzed by a spectrophotometer (DR 2700, Hach Co.) using the methylene blue spectrophotometric method. Dissolved iron (Fe(II)) was detected by UV detector at 562 nm by adding ascorbic acid (1% w/v) and ferrozine (3-(2-pyridyl)-5,6-bis (4-phenylsulfonic acid)-1,2,4-triazine) to a 0.45 μm-membrane-filtered sample. Each analysis was performed in duplicate for sulfide and ferrous concentrations.

2.9. Solid Phase Characterization

Surface analysis of nZVI was performed with a high resolution X-Ray photoelectron spectrometer (Scienta ESCA 300 spectrometer). An Al rotating anode serves as the X-ray source generating an Al Kα X-ray beam at 1486.6 eV. The X-ray beam is monochromatized using seven crystals mounted on three Rowland circles. The energy was analyzed using a high-resolution 300 mm mean radius hemispherical electrostatic analyzer and detected by a multichannel plate-CCD camera. Samples that had been previously dried and stored in N2 glove bag were mounted on a stainless steel stub coated with conductive carbon tape before being transferred to the analysis chamber. Spectra were obtained using a takeoff angle of 90° with respect to the surface plane of the samples. Binding energies of the photoelectrons are correlated to the aliphatic adventitious hydrocarbon C(1s) peak at 284.6 eV.

Morphological and elemental analyses of nZVI were performed by a scanning electron microscope (SEM) (Hitachi S-4300, Hitachi Science Systems, Ltd.) equipped with energy-dispersive X-ray (EDX) at 10 kV. Thermogravimetric analysis (TGA) was conducted using a STA 6000 simultaneous thermal analyzer (PerkinElmer Co.) with a temperature ramp of 20°C/min. The system was purged with air at a flow rate of 20 mL/min.

3. Results and Discussion

3.1. Laboratory Studies

Effectiveness of nZVI for sulfide removal under various pH conditions is illustrated in Figure 2. The initial sulfide concentration was about 1,000 mg/L in an aqueous solution. Approximately 95% of initial sulfide concentration was removed at pH 4. However, the sulfide removal efficiency decreased as pH increased. The oxidation of nZVI leading to generation of Fe(II) is favorable under acidic conditions: Sulfide readily reacted with Fe(II) to form iron sulfide (FeS) [4, 24]: Accordingly, the pH effect on the sulfide removal with nZVI may be attributed to the formation of FeS through the precipitation of Fe(II) and sulfide. To verify the possibility, experiments were conducted to measure the Fe(II) concentrations in the reaction system containing nZVI and sulfide at various pH values. As shown in Figure 3, the Fe(II) concentration increased as pH decreased. This is in a good agreement with the theory of FeS formation.

Extensive studies have indicated that nZVI is an effective reagent for nitrate reduction in the aqueous solution under acidic conditions [14, 15]:

Nitrate is reduced to ammonium by nZVI that concurrently releases Fe(II). In other words, nitrate can serve as a promoter of the Fe(II) generation. On the other hand, phosphate is known to inhibit reaction rates of the nitrate reduction by ZVI [25] and to precipitate with Fe(II) to iron phosphate Fe3(PO4)2 [26]. Thus, phosphate can act as an Fe(II) scavenger. Figures 4 and 5 show the results for the effects of these anions on the sulfide removal and the Fe(II) formation, respectively, in the presence of nZVI at pH 4. It is clearly observed that an enhanced sulfide removal took place and the Fe(II) concentration increased significantly in the presence of nitrate. However, the sulfide removal efficiency and the Fe(II) concentration declined in the presence of phosphate.

The adsorption capacity of nZVI for sulfide removal at various pH values was determined by adsorption isotherm experiments at 25°C. A Langmuir model was tested: where is the amount of adsorbed sulfide at equilibrium (mg/g), is the equilibrium concentration of dissolved sulfide (mg/L), and and are Langmuir constants indicating the saturated capacity of nZVI and an energy term, respectively. Figure 6 illustrates the experimental data of the adsorption isotherm for sulfide removal corrected with the Langmuir equation. The coefficients of determination () were 0.871, 0.885, and 0.689, at pH values 4, 7, and 12, respectively. The saturated adsorption capacities determined by the Langmuir model were 821.2, 486.3, and 359.7 mg/g at pH values 4, 7, and 12, respectively. Compared to the commonly used adsorbents such as virgin activated carbon or impregnated activated carbon, nZVI exhibited a high capacity of sulfide adsorption. For example, the adsorption capacity was estimated to be 62 mg-H2S/g of 2%-KI impregnated activated carbon in a small pig farm [6] while Yan et al. showed that the capacity of H2S adsorption varied from 50 to 210 mg-H2S/g using alkaline activated carbon under different conditions [3].

The adsorption kinetics of sulfide on ZVI was investigated using pseudo-first order and pseudo-second order adsorption equations: where is the dissolved sulfide concentration (mg/L), is the initial sulfide concentration (mg/L), is the rate constant of the pseudo-first order adsorption (1/h), is the rate constant of the pseudo-second order adsorption (), and is time (h). Linear regression of the experimental data obtained at pH 4 for these two kinetic models is shown in Figure 6. Values of were all in the range of 0.95–0.98 for the pseudo-second order model while values of those were 0.78, 0.81, and 0.96 at the nZVI loadings of 1, 2.5, and 5 g/L, respectively, for the pseudo-first order model. Accordingly, the regression results suggested that the sulfide adsorption on nZVI exhibited a second-order kinetic model. This implies that a chemisorption is involved and two distinct sorption sites may be available for the adsorption of sulfide on nZVI [27] (see Figure 7).

3.2. Field Pilot Test

The use of nZVI for in-situ full and pilot scale remediation of contaminated groundwater has been implemented around the world including the US, Canada, Czech Republic, Germany, and Taiwan [28]. The removal efficiency of contaminants is generally dependent on the field conditions and the injection amount of nZVI. In this study, a pilot test was conducted to investigate the performance of nZVI for sulfide removal under field conditions in a pig farm. Experiments were carried out in the 1,000 L reaction tank containing 1 kg nZVI at the flow rate of 66 L/d and a second tank without nZVI under the same conditions as control.

Table 1 shows the monitoring results of wastewater quality in the experimental and control tank. The impacts of nZVI on the wastewater quality were insignificant except for the Fe(II) concentration measured at day 1. This is quite different from the result observed in the groundwater remediation with nZVI that influenced the groundwater quality such as pH, ORP, and Fe(II) concentrations largely. For example, the addition of nZVI resulting in an increase in pH and a decrease in ORP in the aqueous system was commonly found both at the laboratory and field studies [8, 11]. It is likely that the anaerobically digested wastewater may contain complex matrices buffering the impact of nZVI on the water quality.

Figure 8 shows the sulfide removal with nZVI in the pilot test during a period of 30 d. The sulfide concentration in the digested wastewater dropped from 8.4 mg/L to around 4.5 mg/L and remained relatively stable in the control tank, while it declined to 0.25 mg/L initially and then gradually increased to 1 mg/L at the end of the test in the experimental tank. This indicates that nZVI is an effective reagent for sulfide removal. However, the total amount of sulfide removed during the pilot test was calculated to be about 10,000 mg, corresponding to the adsorption capacity of 10 mg/g. The adsorption capacity is almost 50 times less than that determined in the laboratory studies at pH 7. According to the results of Figure 8, the breakthrough of sulfide did not actually occur in the pilot study; thus, the adsorption capacity may be underestimated. Nevertheless, the complexity of digested wastewater constituents may limit the effectiveness of nZVI.

The effect of nZVI on the microbial species distribution in the piggery wastewater was examined using gram staining and culture and further confirmed by gene sequencing identification. The study was performed for five months to collect the background information of microbial species before and after the addition of nZVI. The results of the anaerobic organisms identified in the pilot reaction system are listed in Table 2. There were totally 12 species observed before nZVI was added. Among them, only 4 species were identified after the addition of nZVI. In addition, C. limosum was only identified after the nZVI addition. The impact of nZVI on the change of the microbial species distribution was relatively noticeable.

Impact of nZVI on the microorganisms in environments has received increased attention since nZVI has been implemented in the full-scale field remediation [29, 30]. Studies have reported that no deleterious microbial effects of nZVI were observed when it was applied to aquifer materials or soil matrix [3133]. Additionally, nZVI coated with biodegradable polyaspartate stimulated microbial growth [31], which is consistent with a field study where an enhanced biodegradation by surfactant-modified nZVI was found [12]. Fajardo et al. reported a lack of a broad bactericidal effect of nZVI on the bulk soil microbial community; however, significant changes in the structure and composition of the soil bacteria population were detected in the presence of nZVI [33]. On the contrary, other studies showed divergent results. For example, Lee et al. found that nZVI had a strong bactericidal effect on E. coli under deaerated conditions [34] while complete inactivation of B. subtilis var. niger and P. fluorescens by nZVI (10 g/L) was observed with vigorous shaking under aerobic conditions [35]. It should be pointed out that our study examined only the identification of microbial species. There is a need to attain quantitative information of microorganisms in order to better understand the impact of nZVI on the environments.

3.3. Removal Mechanisms

Surface characterization of nZVI using EDX, TGA, and XPS was conducted to gain insight into the mechanism of sulfide removal with nZVI. The EDX analysis indicated a precipitate of sulfur species on the surface of nZVI. The TGA analysis of the sulfide-laden nZVI was performed over the temperature range 105–600°C in air (Figure 9). The first event of weight loss occurring at 195°C is caused essentially by water vaporization from the samples. The weight gain occurring at 300°C can be attributed to the transformation of FeS to FeSO4 [36], while another weight gain appearing at 550°C is likely due to the oxidation of iron to hematite [37]. The results of TGA analysis suggested that the formation of FeS took place on the nZVI surface when nZVI reacted with sulfide.

Figure 10 presents the S 2p survey of nZVI in reactions with and without sulfide after 17 d. The sulfur spectra were fitted by using doublets 2p1/2 and 2p3/2 with the same full width at half maximum (FWHM) separated by a spin-orbit splitting of 1.2 eV. The S 2p1/2 peak intensity was constrained to be half that of S 2p3/2 peak. It was decomposed into five subsets of doublets with the 2p3/2 peak binding energies at 161.3, 162, 163.4, 166.9, and 168.4 eV. These five peaks can be assigned to monosulfide (S2−), disulfide (), polysulfide (), sulfite (), and sulfate (), respectively [38]. The areas of the fitting curves suggested that approximately 25.8% of sulfur on the surface existed as monosulfide, 22.4% as disulfide, 35.9% as polysulfide, 8% as sulfite, and 7.9% as sulfate. Li et al. have reported the use of nZVI for sulfide removal [9]. The XPS analysis of nZVI after 24 h of reaction with hydrogen sulfide indicated that approximately 36% of sulfur on the surface existed as monosulfide and 64% as disulfide. In comparison to this study, it was found that monosulfide and disulfide were significantly decreased, whereas polysulfide became a dominant species when a longer reaction period was conducted. This suggests that a transformation of disulfide to polysulfide, which is a favorable process for the enhanced adsorption of sulfide, may be involved in the reaction of nZVI with sulfide.

Hydrosulfide that was anoxically converted into thiosulfate and polysulfides by iron/cerium oxide hydroxide has been found at pH 8.0–11.0 under the ambient pressure and temperature conditions [39]. The nZVI has a core-shell structure containing a metallic core with a highly reducing characteristic and a thin amorphous iron (oxy)hydroxide layer promoting oxidation reactions [8, 40]. The formation of polysulfides was attributed to an iron-polysulfide surface complex at the iron (oxy)hydroxide surface [39]. In addition, the second-order kinetic behavior for sulfide adsorption with nZVI was observed in this study. Taken together, the reaction mechanism of sulfide removal with nZVI can be considered as a homogeneous-heterogeneous process. The homogeneous process involves the iron dissolution to Fe(II) that reacts with dissolved sulfide to form the FeS precipitate on the surface of nZVI. The surface associated sulfide can further react with H2S and form FeS2: where the symbol denotes the reactive sites at the nZVI surface. This is consistent with the studies reported by Yan et al. [8]. On the other hand, this work provides new evidence for the formation of polysulfide, an aged product of nZVI reacting with sulfide. The formation of disulfide and polysulfide suggests that the iron-polysulfide surface complex may take place at the iron (oxy)hydroxide surface of nZVI, which is another process for sulfide removal by nZVI.

3.4. Cost Effectiveness of nZVI

The cost for synthesis of 1 g nZVI using the sodium borohydride reduction is about 1 USD based on the estimation of cost of industrial grade chemicals. Sodium borohydride is an expensive reductant, which accounts for nearly 99% of the total cost of borohydride-produced nZVI. Therefore, it is difficult to apply borohydride-produced nZVI for sulfide removal in the real-world situation. Nevertheless, we have found that the elemental aluminum rod is a low-cost reductant that may be used to synthesize nZVI. The total cost of 1 g aluminum-produced nZVI is estimated to be 0.01 USD, which is more competitive to be applied to the wastewater treatment in real conditions.

4. Conclusions

Nanoscale zero-valent iron (nZVI) developed since 1996 represents an important nanotechnology for environmental protection and remediation. We have applied nZVI for dissolved sulfide removal and performed a field pilot test in a 50,000-pig farm to treat dissolved sulfide in anaerobically digested wastewater. The field test suggests that nZVI is an effective reagent for sulfide removal; however, the adsorption capacity is calculated to be about 10 mg/g, almost 50 times less than that determined in the laboratory studies. The complexity of digested wastewater constituents may limit the effectiveness of nZVI. The impact of nZVI on microbial species distribution was relatively noticeable. Further studies are needed to better characterize the impact of nZVI on the environments. Laboratory studies revealed that the reaction mechanism of sulfide removal with nZVI can be considered as a homogeneous-heterogeneous process. The dissolved Fe(II) reacts with sulfide to form the FeS precipitate on the surface of nZVI and an iron-polysulfide surface complex may be involved to form disulfide and polysulfide.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.


The authors would like to thank the National Science Council (NSC) and Environmental Protection Administration (EPA), Taiwan, R.O.C., for the financial support under Grant nos. NSC 98-2622-E-390-003-CC3 and 99-2622-E-390-002-CC3 and EPA-101-U1U1-02-101, respectively. Laboratory assistance provided by Mr. Chia-Wei Yang, Ms. Nien-Chun Hsieh, and Mr. Kuo-Ming Chiang are greatly appreciated.