Biogas produced by waste biomass becomes one of the vital substitute energy sources in recent years as it is obtained from non-fossil fuels  -  . The biogas consists of CH4 and CO2 which are the main constituents, but also contains insufficient amount of contaminants such as H2S, NH3, siloxanes and halogenated volatile organics     . The hydrogen sulfide in biogas stream stands as extremely toxic gas which causes corrosion, erosion, fouling for metal devices such as, cooking stoves, biogas plants and steam turbines  . The presence of H2S in combustion process results in formation of sulfur dioxide which is harmful for environment   , and injurious for human and animal health as it causes irritation of mucous membranes, headaches, dizziness, nausea and sudden death. For these motives, substantial attention has been paid to removal of H2S from biogas stream prior to use.
Researchers focused on developing methods of H2S removal from biogas stream. These methods are mainly directed in fabrication of various materials to adsorb and capture H2S from gas streams       -  . Currently, biogas sequestration through the adsorption process is a striking method due to its simplicity and low-cost materials such as natural clay soils and red mud soils      rich with iron oxide; the latter being oxidizing agent to convert H2S to elemental sulfur which is a vital factor for biogas sequestration. Meanwhile, few studies have proposed the utilization of waste ashes (coal ash, biomass ash)      , kaolin and modified crude clay   , Penausende muscovite  , laterite  , montmorillonite   , which are economically feasible for CO2 and H2S removal.
Considering many benefits of biogas use as non-fossil fuel, it is imperative to develop cost effective adsorbents, easy way of removing toxic substances for application at medium and small-scale plants. Hence this current study aiming to investigate the removal of H2S and NH3 from biogas using Oldoinyo Lengai volcanic ash (OLA) at ambient temperature conditions. The Oldoinyo Lengai is the only active natrocarbonatite volcano and it is a unique source of alkaline ash in the world; the OLA is abundant and readily available     . The experiment was conducted on-site at The Banana investment company Ltd, Arusha, Tanzania where characteristics of H2S and NH3 removal from biogas stream were determined at natural variation of biogas composition.
2. Materials and Methods
2.1. Materials Collection and Preparation
Ashes used in this study were collected from Oldoinyo Lengai volcano which is located Tanzania, the Rift Valley at 2˚45'S, 35˚55'E and 2000 m above the Serengeti plains   . Several samples of ashes were randomly collected from various places on top, medium and bottom of the mountain. Previously it was found that the samples contain compounds of K, Na, Ca, Zn and Fe     which are potentially effective for H2S and NH3 removal from biogas stream. The OLA samples were air dried, ground with a mortar and pestle, sieved to obtain the acquired particle size, below 2 mm, and stored in vacuum desiccators until used. The samples were calcinated at different temperatures Tc, 550˚C, 650˚C, 750˚C, and 850˚C for one hour in the oven (Thermo Scientific 1200 Box furnace) and then allowed to cool; the heating rate was 10˚C/min and cooling rate 20˚C/min. After the preliminary test on H2S removal was performed, the best samples were retained for further investigations and detailed experiments. The selected samples were two from the top (OLA-TP and OLA-TN), two from medium (OLA-MN and OLA-MX) and one from bottom (OLA-BN), where TP and TN represent white ashes without and with plants, respectively; MN is white with plants and MX is a mixture of black and white ashes with plants. The notation of the samples also includes the calcination temperature, e.g. OLA-TP-850 implies the sample was collected at the top site of the mountain and calcinated at 850˚C. The raw materials taken at room temperature (30˚C) are denoted as OLA-TP-RT.
2.2. Materials Characterization
The oxides composition of the samples was determined by X-ray fluorescence spectrometer (XRF), model MiniPal4 (Pw4030)-Rh manufactured by PANalytical, using software provided with the instrument. A sample of 25 g of calcinated adsorbent was added to 100 ml of distilled water. The mixture was stirred and shaken for one hour by using mechanical shaker (model AS200, RETSH Company). The pH of samples was measured with a pH meter, model H199121. The moisture content of raw OLA ashes determined under standard procedures as has been reported elsewhere   was found to be 22.2% ± 0.5%. Mineral phase analysis and elemental oxide composition were analyzed by X-ray diffraction (XRD) technique using D2 phaser-Bruker model and X-ray diffraction meter with a Cu-Kα radiation source in a 2θ range between 10˚ and 80˚ at a scanning rate of 2˚ min−1 and analyzed using EVA software provided with the instrument.
The surface micromorphology of samples was investigated using a scanning electron microscope (SEM), modal JEOL JSM-6335F with resolution of 500 nm, at 200 kV at the Department of Materials Science and Engineering, University of Connecticut, USA.
2.3. Measurements of H2S and NH3 Removal and Evaluation of Sorption Capacity of Absorbents
The sorption experiment of H2S and NH3 was carried out at the Banana Investments company Ltd in Arusha, Tanzania where the biogas was produced from winery effluent banana industrial waste. The biogas was liberated from upflow anaerobic sludge blanket between 100 and 120 m3 daily, the composition of the gas was determined with the biogas 5000 g as analyzer to be 82% - 89% CH4, 12% - 15% CO2, <1% O2, 5 - 48 ppm of NH3, and 24 - 60 ppm of H2S  .
A reactor made of plastic tube, 6.5 cm length, 1.7 cm diameter, filled with adsorbent supported by cotton wool on both end sides of the reactor tube. The biogas allowed to flow through the reactor at ambient temperature varying the flowrate from 0.0004 to 0.008 m3/min, and mass of adsorbents from 0.5 to 2.0 g. The mass of adsorbent of 1 g and the biogas flowrate of 0.002 m3/min were used as standard parameters for this study. Schematic diagram and photos of the experimental setup are shown in Figure 1. The flowrate was controlled and monitored by using a flow meter, model JBD2.5-SA. The concentration of H2S and NH3 was recorded at the inlet and outlet of the reactor with different time intervals from 5 to 150 min. Breakthrough time (BT) was noted when the outlet concentration Cout of H2S and NH3 reached 50% of the inlet concentration Cin. The sorption tests were repeated at least twice for each sample.
The performance of material adsorbent was specified as percentage removal R:
Sulfur sorption capacity (SC)of the OLA samples, in grams of sulfur per 100 grams of sorbent, was determined as described in    ;
Figure 1. Photos and schematic diagram of the sorption experimental setup.
where GHSV is the gas hourly space velocity which is the volume of the gas flowing hourly through the reactor with 1 g of absorbent (L・h−1・g−1); Vmol is the molar volume of the gas (L・mol−1) under standard conditions, M is the atomic mass of sulfur; Cin and Cout are the H2S concentrations before and after sorption (ppm); t is the breakthrough time in hours.
3. Results and Discussion
3.1. XRF Analysis and pH of Raw OLA Samples
Results on the XRF analysis of the raw OLA samples (Table 1) show that Fe2O3, CaO and SiO2 are most abundant components of all samples; alkali metal oxides are in essential amount, K2O ~3% - 10% and Na2O ~2% - 5%. Some of transition and rare-earth elements were also found. As is seen the content of these oxides changes with height the samples were collected. The composition of the OLA samples is rather promising for H2S removal due to, as it was observed earlier    , Fe, K, Ca, Zn, Na, Ti, and Sr oxides may act as catalysts for oxidation of H2S.
The pH values of the OLA ashes are in the range of about 11 - 12 and indicate materials’ alkalinity which is evidently caused, according to  , by presence of iron, calcium, potassium and sodium compounds. The basicity of the OLA ashes is expected to favour the hydrogen sulfide removal as it was shown for example in  . The mixed metal fly ashes    , different metal oxides        , coal ashes   were successfully used for H2S
Table 1. The composition (wt%) and pH values of Oldoinyo Lengai volcanic ash samples.
removal from biogas stream at ambient temperature.
The mechanism of the H2S adsorption/oxidation processes was considered in     ; it was shown that inorganic materials containing active metal oxides participate in this process. The H2S was oxidized to sulfur due to presence of moisture, which enabled dissociation of H2S into HS− and then to elemental sulfur. These reactions were enhanced by basic pH of the adsorbent   .
3.2. XRD Analysis of Calcinated OLA Samples
The XRD spectra were measured for several samples; an example for the OLA-TP-850 is shown in Figure 2. The data indicate the crystalline solid form of the materials where compounds of iron, sodium, calcium and potassium are identified as main components by peaks of magnetite (2θ = 35.5˚, 43˚, 57˚, 63˚), analcime (2θ = 16˚, 18˚, 26˚), epidote (2θ = 31˚, 56˚) and nepheline (2θ = 21˚, 23˚, 27˚, 29.8˚, 37˚), respectively. These results are in accordance with the data reported previously  . For the sample after absorption, the additional peaks were recorded at 2θ = 22˚, 31˚, 32.5˚ and 49˚ which correspond to sodium hydrogen sulfate Na3H(SO4)2. The formation of this compound indicates interaction of hydrogen sulfide with the adsorbent and hence chemisorption occurrence. Our results differ from those published  in which red mud was used for H2S removal, and after sorption FeS2, FeS, CaSO4・2H2O, sulfur, sodium sulfide and bisulfide were formed.
Apparently, the distinction of our results with literature  may be attributed to the different mineralogical and chemical composition of the adsorbents.
3.3. SEM Surface Micromorphology Analysis
The surface morphology of samples was investigated using SEM; an example shown in Figure 3 for OLA-TP-850. The surface features indicate considerable
Figure 2. XRD patterns for OLA-TP-850 sample before adsorption.
Figure 3. Surface morphology of OLA-TP-850 sample before adsorption.
porosity which evidently may provide efficient physisorption.
It was observed earlier that the porous structure and surface chemistry of the adsorbent had significant effect on adsorption and oxidation of hydrogen sulfide  . Besides the morphology of ashes is controlled by calcination temperature, and porosity increases with Tc increase  . We may expect that our samples which were calcinated at higher temperatures would demonstrate better adsorption performance.
3.4. Adsorption Performance of OLA Samples
3.4.1. Effect of Calcination Temperature and Height the Samples Were Collected
The values of sorption capacity calculated using Equation (2) and breakthrough time are given in Table 2 for different OLAs. The effect of calcination temperature and height of sites the samples were collected on the sorption capacity is depicted also in the diagram (Figure 4). As is seen, the SC is increased with calcination temperature raise; e.g., the ashes taken without calcinations, OLA-TP-RT, and calcinated at 850˚C, OLA-TP-850, result in the SC of ~0.2 and ~1.0 g S/100g of adsorbent, respectively. It is also observed that the SC increases with height of site; the samples collected at the top and calcinated at 850˚C, OLA-TP-850, demonstrate the best performance. The breakthrough time relates to the life time of the adsorbent; the longer is the BT, the higher is the working capability of the material. One can see the BT increases greatly with height (from 5 min for OLA-BN-650 to 120 min for OLA-TN-650) and calcination temperature raise (from 55 min for the OLA-TP-RT up to 177 min for OLA-TP-850). Thus, the OLA-TP-850 samples possess the longest working time.
The adsorption efficiency removal R was measured both for H2S and NH3 and analyzed for different samples with respect to time interval between the measurements of initial and final concentrations of the adsorbates while biogas flowed through the reactor. The plots for Rvs time for the samples collected at different heights and calcinated at 850˚C are shown in Figure 5. Results indicate the removal efficiency decreases with time for all samples. For the OLA-TP-850, R reduces by ~10% both for H2S and NH3 after one hour of the adsorbent’s use.
Table 2. Adsorption performance of different OLA samples for H2S removal: sorption capacity and breakthrough time; mass of adsorbent 1 g and the biogas flowrate 0.002 m3/min.
Figure 4. Sorption capacity of Oldoinyo Lengai ash materials for hydrogen sulfide removal; mass of adsorbent 1 g and the biogas flowrate 0.002 m3/min.
Figure 5. Effect of the site location on the removal efficiency of H2S (a) and NH3 (b); for all samples Tc = 850˚C, flowrate 0.002 m3/min, mass of the adsorbent 1 g.
Similar trend is observed in Figure 6 where the plots of Rvs time are given for OLA-TP samples calcinated at different temperatures. The removal efficiency is increased with calcination temperature raise. Thus after 150 min of the absorbent’s use, the R of OLA-TP-RT descents to 37%, while the OLA-TP-850 holds 60% for H2S removal; for ammonia, the values of R are 40% and 80%, respectively. Based on the results of    , we suggest that the increase in adsorption efficiency with temperature Tc may be caused by increase of number of pores created in adsorbents. On the other hand, the removal efficiency decreases in time because the surface of adsorbent is being occupied with the adsorbate molecules and the porosity is reduced as the pores are clogging with the gas molecules  .
If compare the plots of Rvs time for H2S and NH3, they look differently: monotonic decay for the former and step-like for latter. The step-like behavior for ammonia may be attributed to rather low concentration (Cin~0.005 mg NH3 per 1 g of the biogas, that is one order less than of H2S) measured at the sensitivity limit of the gas analyzers. Moreover, according to Equation (2), the lower input concentration of ammonia also brings bigger value of R.
Therefore, among all OLA samples considered the ashes collected at the top site and calcinated at 850˚C, the OLA-TP-850, demonstrated the best removal efficiency retaining it high during the exploitation time of the adsorbent.
3.4.2. Effect of Biogas Flowrate
The effect of biogas flowrate on the samples adsorption performance was investigated; the flowrates were 0.0004, 0.001, 0.002, and 0.008 m3/min. The values of the removal efficiency of H2S and NH3 are plotted vs time for the top-site samples OLA-TP-850 in Figure 7. One can see, with the flowrate raise from 0.0004
Figure 6. Effect of calcination temperature on removal efficiency of H2S (a) and NH3 (b); OLA-TP samples, flowrate 0.002 m3/min, mass of the adsorbent 1 g.
Figure 7. Effect of biogas flowrate on removal efficiency of H2S (a) and NH3 (b); OLA-TP-850, mass of the adsorbent 1 g.
to 0.008 m3/min the R values decrease from 71% to 20% for H2S and from 100% to 58% for NH3, respectively, measured after 150 min adsorbent’s use.
It seems at low flowrate (0.0004 m3/min) the materials get enough contact time for interaction between gas molecules and adsorbent; whereas at high flowrate (0.008 m3/min), H2S and NH3 gas molecules can pass through without being adsorbed and reacted fully with adsorbent. Therefore, the contact time between gas and adsorbent is significant in determining the adsorption capacity. This is in accordance with findings reported in  .
3.4.3. Effect of Mass of Adsorbent Material
The effect of variation of the absorbent mass, from 0.5 to 2.0 g, on the removal efficiency of H2S and NH3; was considered for the samples OLA-TP-850 (Figure 8). The results indicate that R values evidently increase with mass. Four-fold enlargement in mass leads to essential increase in removal efficiency, from 23 to 80% for H2S and from 50% to 100% for NH3 measured after 150 min adsorbent’s use. When the mass of adsorbent was 2 g, the material retained 100% H2S removal after 40 min and efficiency decreased to 80% after 150 min; while for NH3 the adsorbent held 100% removal efficiency for all time measurements.
3.4.4. Comparison of OLA-TP-850 with Other Materials
The adsorption properties of our best sample OLA-TP-850 are compared to other materials reported in literature (Table 3).
Figure 8. Effect of adsorbent mass on removal efficiency of H2S (a) and NH3 (b); OLA-TP-850, flowrate 0.002 m3/min.
Table 3. Adsorption properties of OLA-TP-850 compared to other materials.
The sorption capacity of OLA ashes appeared to be close to the commercial montmorillonite clay and higher than other natural materials used for H2S removal, e.g. municipal waste bottom ash, coal ashes etc. Meanwhile, when different samples are compared it is worth to take into account the effect of experimental conditions like the biogas flowrate and pH; as it was mentioned above, the lower flowrate and basicity of the sorbent favour the hydrogen sulfide removal. Thus, the Oldoinyo Lengai volcanic ashes can be used as adsorbents for H2S and NH3 removal in biogas plant industries, however further improvement of the adsorption properties may be suggested, e.g., increasing of calcination temperature and activation with chemical reagents.
A potency of the Oldoinyo Lengai volcanic ashes on removal of hydrogen sulfide and ammonia from biogas stream was investigated. Sorption capacity and removal efficiency were measured for different OLA samples at ambient temperature. The materials were examined in the on-site experiment at real industrial conditions. The best materials demonstrated long-lasting sorption ability. The XRD and XRF spectra of the OLA samples were examined. Mineralogical composition analysis exhibited an abundance of compounds of iron, alkali and alkali-earthmetals; rare-earth metals compounds were also found. The high porosity of samples favoured physisorption of H2S and NH3. Moreover, chemisorption of H2S was confirmed via formation of sodium sulfate-sodium bisulfate. In our on-site experiment the biogas flowrate and mass of adsorbent were variable parameters. The inlet concentration of H2S in biogas was controlled but difficult to keep constant. For the next step, it may be recommended to evaluate the effect of concentration of H2S and NH3 on the sorption capacity of the samples in laboratory conditions.
The authors thank the German Academic Exchange Service DAAD, for the award of the scholarship which catered for paying tuition fees and accommodation to undertake the MSc studies at The Nelson Mandela African Institution of Science and Technology. Prof. K. Njau through VLIR-IUC/NM-AIST partnership programme is highly acknowledged for the financial support and Department of Materials Science and Engineering, University of Connecticut, USA, for scanning electron microscope (SEM) analysis allowed to render this study.
Conflict of Interest
The authors have not declared any conflict of interest.