High performance regenerative adsorption of hydrogen sulfide from biogas on thermally-treated sewage-sludge

Depósito de Investigación de la Universidad de Sevilla https://idus.us.es/ This is an Accepted Manuscript of an article published by Elsevier in Fuel Processing Technology, Vol. 145, on May 2016, available at: https://doi.org/10.1016/j.fuproc.2016.01.036 © 2016 Elsevier. En idUS Licencia Creative Commons CC BY-NC-ND 1 High performance regenerative adsorption of hydrogen sulfide from biogas on thermally-treated sewage-sludge P. G. Aguilera, F. J. Gutiérrez Ortiz* Departamento de Ingeniería Química y Ambiental, Universidad de Sevilla Camino de los Descubrimientos, s/n. 41092 Sevilla, Spain Phone: + 34 95 448 72 68 * Corresponding author, “Francisco Javier Gutiérrez Ortiz” [email protected] Abstract Biogas desulfurization can be performed by adsorption, although new materials are needed since commercial adsorbents are expensive. In this regard, three types of sewage-sludge were studied as precursors to obtain low-cost adsorbents in a previous paper, attaining the best precursor from a sewage-sludge that was thermally treated up to 700 ºC. However, it must be regenerated to make the process feasible. To find an economical and environmentally friendly regeneration process, an experimental design was performed aimed at minimizing the use of resources such as water consumption, time and the temperature required while achieving a high rate of regeneration. The selected in-situ regeneration consists of entering firstly steam at relatively low temperature (< 250 ºC), against most of published studies, followed by a second step with air. Besides, it can be performed in only 20 min, giving a large feasibility to the overall continuous adsorption process, with very low energy cost and duration for the regeneration. As a relevant result, the thermally treated sewage-sludge was regenerated up to 14 times, and although the adsorption capacity decreased 2.7% on average in each adsorption/regeneration cycle, the cost relative to the adsorbent may be reduced to 20% of the cost of using fresh adsorbent. Keywords Regeneration, adsorption, low-cost adsorbents, hydrogen sulfide, biogas, desulfurization 2 1. Introduction Biogas produced by anaerobic digestion of organic matter from municipal solid waste deposited in a landfill should not be emitted directly to the atmosphere because of its high methane content, which is a greenhouse gas. Instead, biogas can be utilized in internal combustion engines as a renewable source of energy to generate electricity. Nevertheless, biogas also contains H2S, which has to be removed before using it in engines to prevent the equipment corrosion and the formation of sulfur oxides when combusted [1, 2]. Adsorption by activated carbon is one of the most suitable methods for the removal of H2S from biogas. However, its high operational costs have promoted the search of low-cost alternative adsorbents using natural materials (wood, peat, coal, lignite, etc.) as well as industrial/agricultural/domestic wastes or by-products, such as slag, sludge, fly ash, bagasse fly ash or red mud [3-8]. In this regard, adsorbents from sewage-sludge produced by several thermal treatments have been previously tested, characterized and modeled [9, 10]. As a previous result, the adsorbent coded as LG700PA and obtained by pyrolyzing at 500 ºC and then by calcinating at 700 ºC showed an adsorption capacity twice that of a commercial unimpregnated activated carbon [9]. Furthermore, the scaling-up of the adsorption process concluded that a regeneration of the adsorbent is required to make it feasible [10]. Regeneration methods of activated carbon with solvents, as well as by biological or thermal treatments, are the most commonly used [11, 12]. There are other less common regeneration methods such as wet oxidation, electrochemical, supercritical fluid regeneration, microwave irradiation regeneration and regeneration by ultrasounds [13-15]. Chemical regeneration with sodium hydroxide, acids, or an appropriate solvent can be used to remove the solute from the porous structure of the activated carbons. The regeneration is carried out by passing a stream of the solvent through a bed that contains the adsorbent. Usually, these solvents are expensive and harmful, so they need to be recovered for its reuse or management. Biological regeneration 3 uses aerobic or anaerobic micro-organisms to remove biodegradable adsorbates. However, the main problem of this technology is that the temperature, pH and concentration of the solution need to be kept under tight control in order to preserve the colony of micro-organisms at an adequate level. On the other hand, thermal regeneration consists of heating the adsorbent at temperatures between 600-1000 ºC in different types of furnace such as a rotary kiln, a multiple hearth and a fluidized bed; a purge gas removes the adsorbate as it is being desorbed. As drawbacks, thermal regeneration requires high energy consumption, has to be carried out offsite and causes a loss of material by attrition of about 5% [11, 12]. The on-site steam regeneration process is an intermediate treatment between the heat regeneration and solvent extraction. Steam regeneration at high temperature (650-1000 ºC) had been proven to be very effective to regenerate different types of adsorbents [14, 16-17]. Adsorbents used for removing volatile organic compounds can be regenerated at low temperature (< 200 ºC) with high regeneration efficiencies (about 80%) [18], but drying is a necessary and costly step after the steam regeneration. The previous studies found in the literature dealing with regeneration of adsorbents used to remove hydrogen sulfide present in air, not in biogas, utilize cold/hot water washing or thermal treatment where low-to-moderate regeneration efficiencies were reached (about 40%) [19, 20]. GBH enterprises has performed a process where use steam and air at low temperature (< 300 ºC) to regenerate an activated carbons used at desulfurization of natural gas but the treatment takes at least 9 hours [21]. This study is aimed at reaching a high performance regeneration of sewage-sludge adsorbents previously obtained and used to remove hydrogen sulfide from biogas, optimizing the energy consumption that is one of the main drawbacks of regenerative adsorption. To achieve this objective, a systematic methodology was followed by using different stages involving steam and air at different temperature and duration, by limiting them to reduce the cost of regeneration and make feasible a continuous biogas desulfurization by regenerative 4 adsorption. To our knowledge, no other similar study has been addressed; this fact along with the final high performance regenerative process achieved for biogas desulfurization, which consumes a relatively low energy, confer a remarkable novelty to this research. 2. Experimental section 2.1. Materials The LG700PA adsorbent was used to investigate the steam regeneration of thermally treated sewage-sludge adsorbents previously used to remove H2S from biogas [9]. Briefly, this adsorbent was produced in a tubular furnace by heating 100 g of LG sludge at a rate of 5 ºC/min up to 700 ºC and holding this temperature for 30 min. The characterization of this sludge can be found elsewhere [9]. The thermal treatment was carried out under inert atmosphere up to 500 ºC, using a nitrogen flow-rate of 1 L/min and, beyond this temperature, under oxidant atmosphere at an air flow-rate of 0.125 L/min up to 700 ºC. In addition, two commercial activated carbons, labeled as CAT and CAA, were tested to compare the results of the selected adsorbent regeneration. CAT is a fresh activated carbon without impregnation, while CAA is activated carbon impregnated with a NaOH solution. Before the tests, the adsorbents were milled and sieved to obtain a particle size range from 1.41 to 2.83 mm. Experiments were carried out by using a certified mixture of CH4 (60 vol.%), H2S (2000 ppmv) and CO2 (balance) as simulated biogas. The simulated biogas would match a biogas to be dried before entering the adsorber, although some desulfurization experiments were carried out with a humid gas in a previous paper [9], obtaining a small impact on the adsorption process studied. The detailed procedure of adsorbent production is fully described elsewhere [9]. 5 2.2. Methods 2.2.1. Experimental adsorption/regeneration unit The hydrogen sulfide adsorption test was performed before and after each regeneration process to register the breakthrough curves and evaluate the regeneration efficiency of adsorbents. To perform this test, a determined amount of adsorbent (34 g) was packed in a horizontal reactor (22 mm ID, 220 mm fixed length) made of stainless steel AISI 304. Simulated biogas was passed through the adsorption fixed bed at a flow rate of 1.1 L/min and the H2S outlet concentration was monitored by a continuous gas analyzer, based on ultraviolet radiation. Table 1 shows the main parameters used in this study. Several adsorption tests allowed verifying that the results were similar to those conducted in the vertical glass column used in the previous study [9]. In order to carry out the in situ regeneration process, a flexible silicone wire heater was wrapped around the reactor providing 600 mm of heated length and two K-thermocouples were inserted in the reactor. The assembly was thermally isolated by glass ribbon and ceramic fiber with a total thickness of 400 mm, thus minimizing heat losses. Fig. 1 depicts a scheme of the experimental unit. Experiments were carried out at controlled room temperature (20 ± 2 ºC). An SO2 analyzer (based on non-dispersive infrared radiation) was used to detect and quantify the potential presence of SO2 during the regeneration. Regeneration treatments were performed using firstly a water flow-rate of 1.5 mL/min (at 20 ºC) heated and vaporized at different temperatures (120 ºC, 180 ºC and 250 ºC) for 5, 10, 15 and 20 min. The mentioned water volumetric flow-rate corresponds to a steam mass flow-rate of 1.5 g/min. Afterwards, as a second step, a dry air flow rate of 0.1 L/min was fed for 5, 10 or 15 min. Experimental levels of factors were selected by prior screening tests. The samples were coded using three numbers as follows: first, the regeneration treatment temperature (ºC), then the duration of steam feed (min), and finally, the duration of air feed (min). The 250_15_5 regeneration treatment, which 6 consists of feeding water and air for 15 and 5 minutes, respectively, at 250 ºC, was also tested over commercial activated carbons, above mentioned. The regenerated adsorbent samples in cycle n were denoted as 250_15_5 Rn. Adsorption/removal capacities, denoted as x/M (mg H2S/g material), were calculated from the corresponding breakthrough curves and by applying Eq. (1) [4]:           s t s M dttctc Vw MWQ M x 0 0)( [1] where Q is the total inlet flow rate (m3/s), w is the weight of sludge-based material inside the column (g), MW is the molecular weight of H2S (34 g/mol), VM is the molar volume at normal conditions (22.4 NL/mol), c0 is the inlet gas H2S concentration (ppmv), c(t) is the gas outlet H2S concentration (ppmv), and ts is the bed saturation/exhaustion time (s). The test was stopped at the breakthrough concentration, which was fixed in 500 ppmv; thus, ts is the time corresponding to this concentration. It must be taken into account that, rigorously, the real adsorption capacity of an adsorbent should be calculated at saturation, and not at the selected breakthrough point. Therefore, the calculated capacities should be taken just as reference values, useful to compare among different materials and to assess the degradation after several regeneration cycles. Nevertheless, in order to avoid confusion, the adsorption capacity computed at the breakthrough point is called as breakthrough adsorption capacity or, simply, breakthrough capacity. Eq. (2) defines the regeneration efficiency of the adsorbent samples: 100 0  q qn  [2] where q0 and qn are the adsorption capacity of fresh and n-times regenerated adsorbents, respectively. 7 2.2.2. Characterization of sewage-sludge adsorbents The texture characterization of fresh and regenerated adsorbent was carried out by measuring the nitrogen adsorption isotherm at 77 K in a Micromeritics ASAP 2420 System. The specific surface area, micropore volume as well as mesoand macropore volume of the adsorbents were calculated using the Brunauer-Emmett-Teller (BET) method, t-Plot method based on the thickness equation by the Harkins and Jura, and the Barrer Joyner Halenda (BJH) method, respectively. Scanning Electronic Microscopy (SEM) observations were carried out by a Philips XL30 microscope with secondary and backscattered electron imaging, equipped with an integrated Energy Dispersive X-ray Spectrometer (EDS) to analyze the elements and their contents. An X-Ray Diffraction (XRD) instrument Bruker D8 Advance A25 coupled with a Cu anticathode (40 KV, 30 mA) was used to obtain structural information on an atomic scale of both crystalline and non-crystalline (amorphous) materials. Powder samples were scanned from 3º to 70º (2θ), using a step size of 0.015º (2θ), and a scan step time of 0.1 s. Inductively coupled plasma-optical emission spectrometry (ICP-OES) (Horiba Jobin Yvon Ultima 2) was used in the determination of Si, Al, Fe, Ca, Mg, Na, K, P, S, Cu and Zn in the condensate collected in the Erlenmeyer flask (Fig. 1) and in the adsorbent before and after conducting the regeneration process. The solid samples were treated in a digester microwave (Anton Paar Multiwave 3000). 3. Results and discussion 3.1. Breakthrough curves and adsorption capacities Fig. 2a depicts the H2S breakthrough curves of the fresh and regenerated samples at 250 ºC for 20 min. These regeneration tests were conducted by feeding water from 5 to 20 minutes and air the remaining time to conclude the test, as specified in the test codes. The breakthrough 8 curves are shifted to the right when the duration of steam regeneration is increased and the duration of dry air feed is decreased in the regeneration test. Therefore, the adsorption capacities and regeneration efficiencies also increase correspondingly, except for the 250_20_0 sample. This sample was treated without air and, probably, the condensate was retained in the pores, obstructing them. On the contrary, when injecting air after steam, the dry air reduces the water content on the adsorbent surface, thus favoring the next hydrogen sulfide adsorption. Therefore, the duration of steam and air feed must be optimized. The sample regenerated by feeding steam and air, for 15 and 5 min, respectively (250_15_5 sample), provides the best results, so the duration of air feed was fixed at 5 min for the next tests. Table 2 shows the main adsorption/regeneration tests data: water consumption per amount of regenerated activated carbon, regeneration duration, breakthrough capacity and regeneration efficiency for the samples subjected to one adsorption/regeneration cycle. Thus, in this study, the longest on-site tested treatment takes 20 minutes with a water consumption of 0.87 L/kg of adsorbent, whereas in literature [19] five off-site washing cycles at 100 ºC were done with 91 L/kg of carbon, so the consumption of water would make its industrial application unfeasible. Fig. 2b illustrates the effect of the regeneration temperature on the H2S breakthrough curves using the same duration for the air and steam fed to the packed-bed (15 and 5 minutes, respectively). The treatment at 120 ºC could not regenerate the adsorbent sample, and only a slight regeneration capacity was obtained at 180 ºC. However, the breakthrough curves of samples treated at higher regeneration temperatures are very close to that of the fresh adsorbent. The condensates of regeneration tests at 120 ºC and 180 ºC had a whitish color and pH 6, while the condensates were yellowish and pH was between 3 and 4 when regenerating at a higher temperature (250 ºC). This latter indicates that the sulfur retained on the adsorbent surface was removed and entrained by the steam as H2S (pKa=7.02 at 298 K) to the condensate collected in the Erlenmeyer flask. 15 [11] J.A. Menendez-Diaz, I. Martin-Gullón. Types of carbon adsorbents and their production In: Bandosz TJ, editor. Activated carbon surfaces in environmental remediation. Oxford: Elsevier (2006). [12] Z. Chowdhury, R.S. Summers, G.P. Westerhoff. Activated Carbon: Solutions for Improving Water Quality, American Water Works Assoc. (2012). [13] J. Carratalá-Abril, M.A. Lillo-Ródenas, A. Linares-Solano, D. Cazorla-Amorós. Regeneration of activated carbons saturated with benzene or toluene using an oxygencontaining atmosphere. Chem. Eng. Sci. 65 (2010) 2190–2198. [14] S. Román, B. Ledesma, A. Álvarez-Murillo, J.F. González. Comparative study on the thermal reactivation of spent adsorbents. Fuel Process. Technol. 116 (2013) 358–365. [15] K.Y. Foo, B.H. Hameed. Microwave-assisted regeneration of activated carbon. Bioresource Technol. 119 (2012) 234–240. [16] D. Xin-hui, C. Srinivasakannan, L. Jin-sheng. Process optimization of thermal regeneration of spent coal based activated carbon using steam and application to methylene blue dye adsorption. J. Taiwan. Inst. Chem. E. 45 (2014) 1618–1627. [17] J. Wang, J. Guo, R. Parnas, B. Liang. Calcium-based regenerable sorbents for high temperature H2S removal. Fuel 154 (2015) 17–23. [18] S.G. Ramalingam, P. Pré, S. Giraudet, L. Le Coq, P. Le Cloirec, O. Baudouin, S. Déchelotte. Recovery comparisons—hot nitrogen Vs steam regeneration of toxic dichloromethane from activated carbon beds in oil sands process. J. Hazard. Mater. 205– 206 (2012) 222– 228. [19] A. Bagreev, H. Rahman, T.J. Bandosz. Wood-Based Activated Carbons as Adsorbents of Hydrogen Sulfide: A Study of Adsorption and Water Regeneration Processes. Ind. Eng. Chem. Res. 39 (2000) 3849-3855. 16 [20] A. Bagreev, H. Rahman, T.J. Bandosz. Study of H2S adsorption and water regeneration of spent coconut-based activated carbon. Environ. Sci. Technol. 34 (2000) 4587-4592. [21] Activated carbon air-steam regeneration procedure. Vulcan series VGP-102/103/104. Catalyst, process technology consultancy. GBH Enterprises, Ltd. (2013). [22] Energy price statistics - European Commission, 2015 http://ec.europa.eu/eurostat/statistics-explained/index.php/Energy_price_statistics (Accessed September 15, 2015) Tables and Figures Captions Table 1. Parameters used in the study Table 2. Some main adsorption/regeneration tests results Table 3. Regenerated activated carbon results as compared to sewage-sludge adsorbent Table 4. BET surface area and pore structure of the fresh adsorbent and the 14-timesregenerated adsorbent Table 5. EDS results of the adsorbent sample used once and the 14-times-regenerated adsorbent (note: LG700PAG refers to adsorbent sample used once) Table 6. ICP-OES analysis results of the precursor, the adsorbent before and after regeneration process, and the condensate Fig. 1. Scheme of the experimental unit for regeneration Fig. 2. H2S breakthrough curves for fresh and regenerated samples at 250 ºC for 20 min (a), at different temperatures (b), at 250 ºC by feeding steam during different periods of time (c) Fig. 3. H2S breakthrough curves of the regenerated activated carbon (a) and of the 14 times regenerated adsorbent (b) Fig. 4. Estimated costs and regeneration efficiencies of samples regenerated at different temperatures (a), and at 250 ºC by feeding steam and air during different periods of time (b) Fig. 5. Economical estimation of the scale up process using different number of regeneration cycles Fig. 6. SEM micrographs and EDS analysis of the adsorbent (LG700PA) sample (a) and 250_15_5 R14 sample: area outside the deposits (b), around the deposits (c), on the white deposits (d) Fig. 7. X-Ray Diffraction (XRD) patterns of the adsorbent sample used once (a) and after 14 cycles of regeneration (b) (note: LG700PAG refers to adsorbent sample used once) Parameter Value Unit Simulated biogas composition CH4 60 vol.% H2S 2000 ppmv CO2 (balance) vol.% Feed biogas flow-rate 1.1 L/min Temperature 20±2 ºC Bed length 220 mm Bed diameter 22 mm Particle size 1.41-2.83 mm Bulk density 481.0 kg/m3 Bed void fraction 0.808 - Gas density 1.101 kg/m3 Table 1 Sample Water consumption (L/kg adsorbent) Regeneration duration (min) Breakthrough adsorption capacity (mg/g adsorbent) limit: 500 ppmv Regeneration efficiency (%) 250_5_15 0.22 20 1.69 23.5 250_10_10 0.43 20 4.67 64.7 250_15_5 0.65 20 6.82 94.6 250_ 20_0 0.87 20 4.01 55.6 120_15_5 0.65 20 0.04 0.6 180_15_5 0.65 20 3.11 43.1 250_10_5 0.43 15 6.14 85.2 Table 2 CAT CAT 250_15_5 R1 CAA CAA 250_15_5 R1 LG700PA 250_15_5 R1 Breakthrough capacity (mg/g) 3.18 2.37 10.84 5.19 7.21 6.82 Time to reach 500 ppmv (min) 35 21 97 56 76 73.5 Regeneration efficiency (%) 74.6 47.9 94.6 Table 3 LG700PA 250_15_5 R14 BET Surface Area (m²/g) 89.93 39.38 t-Plot Micropore Area (m²/g) 60.90 12.49 t-Plot External Surface Area (m²/g) 29.03 26.88 Micropore Volume (cm³/g) 24.89 E-3 6.44 E-3 Mesoand macropore volume (cm³/g) 36.71 E-3 57.12 E-3 Vol total (cm³/g) 61.60 E-3 63.56 E-3 Vmicro/Vtotal 40.4 10.1 Vmeso/Vtotal 59.5 89.8 Table 4 LG700PAG 250_15_5 R14 % weight (a) (b) outside the deposits (c) around the deposits (d) on the white deposits C 8.2 n.d. n.d. n.d. O 29.5 34.9 47.2 45.8 Mg 12.6 3.3 n.d. n.d. Al 3.7 9.0 n.d. n.d. Si 7.0 23.6 n.d. n.d. P 5.1 n.d. n.d. n.d. Cu n.d. n.d. 2.7 0.9 S 3.2 10.6 12.2 26.2 Cl 0.4 n.d. n.d. n.d. K 0.9 n.d. n.d. n.d. Ca 18.5 12.9 3.1 26.0 Fe 10.9 5.7 34.8 1.0 Note: “n.d.” means non-detected Table 5 LG mg/g precursor LG700PA mg/g adsorbent 250_15_5 R14 mg/g adsorbent LR14 mg/g regenerated adsorbent Al 19.2 40.6 42.4 1.7E-4 Fe 31.1 75.6 69.2 7.9E-4 Ca 43.4 97.3 90.8 1.5E-3 Mg 9.2 20.4 19.3 3.6E-4 Na 1.6 3.7 4.1 3.2E-4 K 3.7 5.5 8.4 4.9E-4 P 16.5 40.0 34.0 2.0E-4 S 11.3 8.9 33.4 3.4 Cu 0.2 0.4 0.3 3.2E-5 Zn 0.4 1.2 1.0 1.2E-4 Table 6 Note: Steam and air are sequentially fed by using push-in fittings. Fig. 1 Fig. 2 0 50 100 150 200 250 300 350 400 450 500 0 20406080100 H2S outlet concentration (ppmv) Time (min) 250_5_15 250_10_10 250_15_5 250_20_0 LG 700 PA 0 100 200 300 400 500 600 700 800 900 1000 0 20406080100120 H2S outlet concentration (ppmv) Time (min) 120_15_5 180_15_5 250_15_5 LG 700 PA 0 100 200 300 400 500 600 700 800 900 1000 0 20406080100120 H2S outlet concentration (ppmv) Time (min) 250_10_5 250_15_5 LG 700 PA (a) (b) (c) Fig. 3 0 50 100 150 200 250 300 350 400 450 500 0 20 40 60 80 100 120 140 H2S outlet concentration (ppmv) Time (min) LG 700 PA 250_15_5 R1 CAT CAT 250_15_5 R1 CAA CAA 250_15_5 R1 0 50 100 150 200 250 300 350 400 450 500 0 20406080100 H2S outlet concentration (ppmv) Time (min) LG 700 PA 250_15_5 R1 250_15_5 R2 250_15_5 R3 250_15_5 R4 250_15_5 R5 250_15_5 R6 250_15_5 R7 250_15_5 R8 250_15_5 R9 250_15_5 R10 250_15_5 R11 250_15_5 R12 250_15_5 R13 250_15_5 R14 (a) (b) Fig. 4 120°C 180°C 250°C 0 10 20 30 40 50 60 70 80 90 100 0.055 0.056 0.057 0.058 0.059 0.060 0.061 0.062 Regeneration efficiency (%) Regeneration cost (€/kg adsorbent) 250_10_5 250_15_5 250_5_15 250_10_10 250_20_0 0 20 40 60 80 100 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 Regeneration efficiency (%) Regeneration cost (€/kg adsorbent) (a) (b) Fig. 5 0 500 1000 1500 2000 2500 3000 3500 4000 0.00 20.00 40.00 60.00 80.00 100.00 Cost (€) Time(days) Fresh adsorbent 1 reg. 2 reg. 4 reg. 6 reg. 8 reg. 10 reg. 12 reg. 14 reg. 0.00 0.05 0.10 0.15 0.20 0.25 0 2 4 6 8 10 12 14 Cost (€/m3biogas) Number of regeneration cycles (a) (b)