ADSORPTIVE REMOVAL OF H 2 S IN BIOGAS CONDITIONS FOR HIGH TEMPERATURE FUEL CELL SYSTEMS

Transcription

1 FiDh European Fuel Cell Technology & Applica0ons Conference - Piero Lunghi Conference December 11-13, 2013, Rome, Italy ADSORPTIVE REMOVAL OF H 2 S IN BIOGAS CONDITIONS FOR HIGH TEMPERATURE FUEL CELL SYSTEMS E. Sisani, A. Castelli, G. Cin0, G. Discepoli, D. Penchini, U. Desideri Fuel Cell Laboratory, Department of Industrial Engineering University of Perugia, Italy

2 CONTENTS Background Objec0ve Characteriza0on of the materials Experimental test bench Methodology Results Conclusions

3 BACKGROUND Biogas, from biomass or waste, can be used as fuel to feed fuel cell systems, because of the possibility to produce H 2 through CH 4 reforming. The advantages are: - low environmental impact (no posi0ve contribu0on to CO 2 emissions is done); - use of a renewable power source. Biogas is poor in energy content, so it is be]er to use it with a high efficiency system, for example a micro- CHP plant, cons0tuted by high temperature fuel cell stack. Biogas, depending on its origin (agricultural waste, industrial wastewater, sewage sludge, landfill) can have different composi0on, in par0cular in terms of pollutants: - sulphur compounds: Ø H 2 S ( ppm) Ø mercaptanes (0-100 ppm) Ø COS, CS 2 (traces); - halogenated hydrocarbons; - siloxanes. H 2 S is one of the most harmful sulphur compound for equipment, because of its reac0on with Ni to form Ni- S. The tolerance limit for high temperature fuel cell is < 1 ppm: thereore H 2 S has to be removed from the gas stream to allow its safe use. In literature there are many clean- up technologies to remove sulphur (wet processes, membrane processes, dry processes), but not all can reach ultra- low sulphur level, as required for this applica0on.

4 OBJECTIVE The most reliable clean- up method for HTFC micro- CHP applica0on is the adsorp0on, using materials like ac0vated carbons (impregnated or not), zeolites, metal oxides (metal as Al, Fe, Mn, Co, Cu and Zn, depending of their acid- base proper0es). For this type of applica0on, the efficiency and the compactness of the system represent a crucial factor. Therefore it is mandatory to op0mize the opera0ng parameters of the system, to obtain the smaller filter necessary for 1 year of opera0ve ac0vity. We selected different types of commercial adsorbent materials and we tested them at different condi0ons (in terms of space velocity, temperature, inlet H 2 S concentra0on, humidity, gas matrix composi0on and reactor geometry), simula0ng a biogas. We observed the influence of these parameters on H 2 S adsorp0on capacity and we selected the best set of parameters, characterized by the best performance. For the best and for the worse set of parameters, we calculated the resul0ng filter volume and filter height and we determined if they are suitable for this applica0on.

5 SELECTED ADSORBENT MATERIALS FOR H 2 S REMOVAL Seven commercial materials were been selected for the test campaign: Norit RB1 is a non- impregnated steam ac0vated carbon, with a par0cle diameter of 1 mm; it is characterized by high adsorp0on capacity and removal efficiency for contaminants present at moderate concentra0ons in gas flows. Norit RBAA1 is a steam ac0vated carbon impregnated with KOH, used to remove small amount of acidic vapors from waste air. Norit RGM1 is a steam ac0vated carbon impregnated with Cu and Cr salts, with a par0cle diameter of 1 mm, designed for the removal of low concentra0ons of sulphur compounds like hydrogen sulphide and mercaptans from gases AIRPEL ULTRA DS is a specially doped ac0vated carbon for desulphuriza0on of gases. Under the right condi0ons, loading rates of over 100 wt% sulfur can be achieved. Galipur S is an ac0vated alumina impregnated with KMnO 4, designed for gas purifica0on of both organic and inorganic sulphur compounds, combining physical and chemical adsorp0on, even at low values of residence 0me. Zeolite ATZ is a natural zeolite (cabasite, phillipsite), with par0cle size in the range 0,2-1,5 mm, used as adsorbent material for gas purifica0on. Sepiolite is a natural clay, used for gas adsorp0on and deodoriza0on. Commercial adsorbent Material Shape - size (mm) Bulk density (kg/m 3 ) RB1 Non- impregnated ac0vated carbon Pellets - 1 mm 480 RBAA1 Impregnated ac0vated carbon (KOH) Pellets - 1 mm 530 RGM1 Impregnated ac0vated carbon (Cu- Cr) Pellets - 1 mm 485 Airpel Ultra DS Impregnated ac0vated carbon Pellets - 4 mm 460 Galipur S Impregnated ac0vated alumina (8% KMnO 4 ) Pellets - 5 mm 850 Zeolite ATZ Natural zeolite Granular mm 875 Sepiolite Natural sepiolite Granular mm 630

6 BET ANALYSIS Volume adsorbed (cm 3 /g STP) AC RGM1 Airpel Ultra DS AC RB1 AC RBAA1 Alumina Galipur S Zeolite ATZ Sepiolite 0 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 RelaRve pressure (P/P 0 ) Sample BET surface area (m 2 /g) Micropore volume (cm 3 /g) AC RGM Airpel Ultra DS AC RB AC RBAA Zeolite ATZ Alumina Galipur S Sepiolite Nitrogen adsorp0on- desorp0on isotherms at 77 K (BET method) show that a high contribute to the BET surface area is due to the Nitrogen adsorbed in micropores. The presence of a relevant micropores volume represent a crucial factor for the kine0cs of H 2 S adsorp0on.

7 EXPERIMETAL LAY- OUT N2+H2S GFC-1 N2 Air GFC-2 EV CO2 CH4 CH LFC-1 N2 GFC-3 water RH GC FPD H2S ES CD REF N GFC 3 LFC 1 DESCRIPTION GAS MASS FLOW CONTROLLER LIQUID MASS FLOW CONTROLLER EV BV TC CD TR EVAPORATOR BALL VALVE THERMOCOUPLE CONDENSER TEMPERATURE INDICATOR & CONTROLLER RH 1 REACTOR HEATER CH SV ES CABLE HEATER SAMPLE VALVE ELECTROCHEMICAL SENSOR

8 METHODOLOGY The adsorbent materials, originally pellet- shaped (ac0vated carbons and ac0vated alumina) or in granular form (zeolite and sepiolite), were ground and sieved, obtaining homogeneous powder with par0cle size in the range μm. The materials were tested for H 2 S adsorp0on, obtaining breakthrough curves for different opera0ng condi0ons. The inves0gated parameters were: - Gas Hourly Space Velocity (GHSV), defined as Q tot / V filter [h - 1 ] - Rela0ve Humidity (R.H.) [%] - Gas Matrix (N 2, CH 4, CO 2, CH 4 +CO 2 ) - Temperature - Inlet H 2 S concentra0on - Filter geometry, h/d (h = filter height; D = filter diameter). CalculaRon of H 2 S adsorpron capacity The H 2 S adsorp0on capacity (C ads in mg/g) is calculated from the breakthrough curves using the following equa0on: C ads = Q tot MW [C in t 1 (t 1 t 0 ) 0,5] V m m 10 3 where: Q tot = total gas flow rate (nl/h); MW = molecular weight (H 2 S = 34 g/mol); C in = inlet H 2 S concentra0on (ppmv); t 1 = breakthrough 0me when the outlet H 2 S concentra0on is 1 ppmv (h); t 0 = breakthrough 0me at 0 ppmv (h); V m = molar volume (24,414 nl/mol); m = mass of adsorbent material (g).

9 BREAKTHROUGH CURVES OF ADSORBENT MATERIALS H 2 S outlet concentraron (ppmv) AC RGM1 AC RBAA1 AC Ultra DS Alumina Galipur S AC RB1 Zeolite ATZ Sepiolite Gas H 2 S + N 2 h/d 0.32 Temperature 30 C H 2 S inlet concentraron 200 ppm GHSV h - 1 R.H. 0 % Filter height 0.58 cm Filter volume 1.47 cm 3 Gas flow rate nl/h Time (min) The best performed material is the steam ac0vated carbon Norit RGM1 impregnated with Cu- Cr salts. Adsorbent material t 0 (h) t 1 (h) C ads (mg/g) AC RGM AC RBAA AC Ultra DS Ac0vated alumina Galipur S AC RB zeolite ATZ < 0.1 Sepiolite < 0.1 The other impregnated materials (Airpel Ultra DS, Norit RBAA1, treated with KOH, and Galipur S, impregnated with KMnO 4 ) exhibited significant values of H 2 S adsorp0on capacity. Zeolite and sepiolite presented very low ac0vity in terms of H 2 S adsorp0on: therefore, they are not suitable for this type of applica0on, even if they present good selec0vity for natural gas odorants, in par0cular for mercaptans.

10 EFFECT OF GHSV 100 AdsorpRon capacity (mg/g) Airpel Ultra DS Norit RGM1 Gas H 2 S + N 2 h/d 0.32 Temperature 30 C H 2 S inlet concentraron 200 ppm R.H. 0 % Filter height 0.58 cm Filter volume 1.47 cm GHSV (h - 1 ) GHSV (h - 1 ) C ads (mg/g) Airpel Ultra DS Norit RGM The effect of space velocity was inves0gated for AC Norit RGM1 and AC Airpel Ultra DS in the range h - 1. The two ac0vated carbons present the same behavior in terms of GHSV varia0on. H 2 S adsorp0on capacity is strongly influenced by GHSV in the range h - 1. For all the tested values, the adsorp0on capacity of Ultra DS is 18-75% lower than RGM1.

11 EFFECT OF GHSV Airpel Ultra DS AdsorpRon capacity (mg/g) Norit RGM1 RGM GHSV (h - 1 ) Ultra DS Using the op0miza0on method of least squares (op0miza0on technique of numerical analysis that minimizes the sum of squares of the distances between the experimental data and those of the func0on), the two ac0vated carbons presented a hyperbolic behavior. We are also able to furnish the explicit law of the func0ons.

12 EFFECT OF GHSV 12 Adsorbent Material AC RB1 C ads (mg/g) H2S DMS Gas H 2 S + N 2 /DMS + CH 4 h/d 0.32 Temperature 30 C H 2 S inlet concentraron 10 ppm R.H. 0 % Filter height 0.58 cm Filter volume 1.47 cm 3 2 H 2 S GHSV (h - 1 ) GHSV (h - 1 ) H 2 S C ads (mg/g) DMS , ,18 0, ,86 0, , , ,08 - A similar hyperbolic behavior was found for the non- impregnated steam ac0vated carbon AC RB1 in the case of H 2 S adsorp0on. A very different trend was found for the adsorp0on of an organic compound, dimethylsulphide (DMS), for which the C ads doesn t present a significant varia0on in the range of GHSV h - 1.

13 EFFECT OF HUMIDITY RGM1_dry H 2 S outlet concentraron (ppm) Ultra DS_dry RGM1_wet Ultra DS_wet Gas H 2 S + N 2 h/d 0.32 Temperature 45 C H 2 S inlet concentraron 100 ppm GHSV h - 1 R.H. 0 (dry)- 50 (wet) % H 2 O vap concentraron 3.5 %mol Filter height 0.58 cm Filter volume 1.47 cm 3 Gas flow rate nl/h Time (h) In dry condi0ons, RGM1 adsorp0on capacity is more than two 0mes Ultra DS one. Material t 0 t 1 C ads (mg/g) Ultra DS_dry RGM1_dry Ultra DS_wet RGM1_wet The presence of humidity determines an increase of C ads for both the ac0vated carbons, probably because the H 2 S was dissolved in the water matrix and therefore is captured by water. Going from dry to wet condi0ons, RGM1 slightly enhances its performance, while Ultra DS shows a significant increase of C ads, reaching a value similar to RGM1. It is possible that a further increase of R.H., un0l 100%, could lead to a performance of Ultra DS be]er than RGM1.

14 EFFECT OF GAS MATRIX N2 H 2 S outlet concentraron (ppmv) CO2 CH4 CH4(70%)+CO2(30%) Adsorbent material AC Norit RGM1 h/d 0.32 Temperature 30 C H 2 S inlet concentraron 100 ppm GHSV h - 1 R.H. 0 % Filter height 0.58 cm Filter volume 1.47 cm 3 Gas flow rate nl/h Time (h) The use of N 2 as carrier gas leads to the highest adsorp0on capacity. The adsorp0on capacity decreases using CO 2, CH 4 (simula0ng a natural gas) and a mixture composed of 70% CH 4 and 30% CO 2 (simula0ng a biogas) respec0vely. Matrice gas t 0 (h) t 1 (h) C ads (mg/g) CH CO N CH 4 (70%)+ CO 2 (30%) This behavior can be explained considering a compe00ve adsorp0on between H 2 S and CO 2, that is enhanced by the simultaneous presence of CH 4. For engineering scopes, using an inert gas as N 2 it is possible to overes0mate the adsorp0on capacity of the filter.

15 EFFECT OF TEMPERATURE Outlet H 2 S concentraron (ppm) T=30 C T=60 C T=90 C Adsorbent material AC RGM1 Gas H 2 S + N 2 h/d 0.32 H 2 S inlet concentraron 200 ppm GHSV h - 1 R.H. 0 % Filter height 0.58 cm Filter volume 1.47 cm 3 Gas flow rate nl/h Time (h) Temperatura ( C) t 0 (h) t 1 (h) C ads (mg/g) Increasing the temperature, the curves present an earlier breakthrough and the H 2 S adsorp0on capacity decreases. This behavior could be explained considering that, for this temperatures, the prevalent mechanism is the physical adsorp0on. Therefore, increasing the temperature, also the kine0c energy of the molecule inside the pores increases.

16 EFFECT OF INLET H 2 S CONCENTRATION 40 H 2 S adsorpron capacity (mg/g) Adsorbent material AC RGM1 Gas H 2 S + N 2 h/d 0.32 Temperature 30 C GHSV 10,000 h - 1 R.H. 0 % Filter height 0.58 cm Filter volume 1.47 cm 3 Gas flow rate nl/h H 2 S inlet concentraron (ppm) H 2 S inlet concentarron (ppm) t 0 (h) t 1 (h) C ads (mg/g) 50 20,57 21,28 33, ,89 11,14 34, ,21 4,36 27, ,733 1,773 27, ,867 0,888 27,79 H 2 S adsorp0on capacity increases from 50 to 100 ppm, probably due to the incresing of par0al pressure of H 2 S in the mixture. Then it is quite constant for a wide range of H 2 S inlet concentra0on ( ppm).

17 EFFECT OF FILTER GEOMETRY (h/d) 0,2 h/d < 1 h C out / C in 0,15 0,1 0,05 0 h/d > Time (h) D Adsorbent material AC RGM1 Gas H 2 S + N 2 Temperature 30 C Inlet H 2 S concentraron 10 ppm GHSV 10,000 h - 1 R.H. 0 % Filter height 0.58 cm Filter volume 1.47 cm 3 Gas flow rate nl/h h/d Breakthrough Rme (h) H 2 S sorpron capacity (mg/g) < ,84 > ,71 The configura0on with h/d > 1 lead to a bigger breakthrough 0me and to an increase of 38% of H 2 S sorp0on capacity. With this configura0on is in fact possible to obtain a be]er gas distribu0on inside the filter, allowing the gas mixture to reach also its peripheral parts.

18 FILTER VOLUME CALCULATION P = 1 kw 1 year of filter acrvity 3 1( kw ) 1( kj / s) 5 m s/ Q = = = 4, = ( kj / m ) 0, ( kj / m ) 0,6 s/ h 3 m/ l l 1h/ l = 0, = 173,9 = 2,9 3 h m/ h/ 60 min min 3 m Vgas = 0, h/ = 1391m h/ 3 C ads = 93 mg/g C ads = 25 mg/g Adsorbent Material AC RGM1 Gas H 2 S + N 2 GHSV 1000 h - 1 Temperature 30 C H 2 S inlet concentraron 200 ppm R.H. 0 % Adsorbent material AC RGM1 Gas CH 4 (70%) + CO 2 (30%) GHSV h - 1 Temperature 30 C H 2 S inlet concentraron 100 ppm R.H. 0 % D = 20 cm V filter = 9330 cm 3 V filter = cm 3 h = 29 cm h = 111 cm

19 CONCLUSIONS The impregnated commercial adsorbent materials achieves the best performance in terms of H 2 S adsorp0on capacity. H 2 S adsorp0on capacity is influenced by opera0ng parameters (GHSV, R.H., gas matrix composi0on, temperature, H 2 S inlet concentra0on, filter geometry). C ads enhances decreasing GHSV and temperature (in the case of predominant phisical adsorp0on), increasing R.H. and for a filter configura0on characterized by h/d>1. A li]le change in the set of opera0ng parameters can determine a significant varia0on of H 2 S adsorp0on capacity and, therfore, of the fiter volume and height, for 1 year of ac0vity. In the future work, other combina0ons of parameters will be tested and the effect of the simultaneous presence of different contaminants will be inves0gated.

20 AKNOWLEDGMENTS This work was carried out with the support of the European Community. We appreciate the support of the project MCFC- CONTEX, founded under the FP7 (Grant Agreement Number ).

21 Thank you for your a,en.on.