- Email: [email protected]
A novel test method for cracking catalysts
Journal of Analytical and Applied Pyrolysis 44 (1998) 193 – 204
A novel test method for cracking catalysts M.I. Nokkosma¨ki a,*, E.T. Kuoppala b, E.A. Leppa¨ma¨ki b, A.O.I. Krause a a
Helsinki Uni6ersity of Technology, Department of Chemical Technology, P.O. Box 6100, FIN-02015 HUT, Finland b VTT Energy, Energy Production Technologies, P.O. Box 1601, FIN-02044 VTT, Finland Received 24 August 1997; accepted 25 September 1997
Abstract A novel microscale test method was developed for cracking catalysts. The pyrolysis unit was a commercial pyrolyser connected to a gas chromatograph. The injection port of the gas chromatograph was used as a fixed bed catalyst reactor. The test method was applied to cracking pyrolysis vapours of Scots pine sawdust with zeolites. Cracking of vacuum gas oil on a commercial FCC catalyst was used as reference. Detection of reaction products was carried out with a mass selective detector to identify the compounds or with an atomic emission detector to quantify the various elements. The results, in agreement with the literature indicate that the microscale pyrolysis and vapour-phase catalyst reactor was suitable for screening catalysts. Pyrolysis vapours were converted mainly into gases and aromatic hydrocarbons with zeolites. The zeolite catalysts were effective in the removal of oxygen but the liquid yields were low. © 1998 Elsevier Science B.V. Keywords: Catalyst screening; Cracking; Catalysts; Zeolites; Biomass; Pyrolysis; Pyrolysis vapours
1. Introduction One alternative to using biomasses is to pyrolyse them to liquid products. The products, mechanism and kinetics of pyrolytic reactions have been discussed previously [1]. These pyrolysis oils can be used, e.g. as fuels for energy production. However, the major limitation is their instability, i.e. quality changes during * Corresponding author. Tel.: +358 9 4566582; fax: + 358 9 460493; e-mail: Milja.Nokkosmaki@ vtt.fi 0165-2370/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 2 3 7 0 ( 9 7 ) 0 0 0 8 0 - 6
194
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
storage. The pyrolysis oils can be upgraded via catalytic conversion with or without hydrogen [2]. Alternatives for catalytic upgrading are physical methods like addition of water or organic solvents and hot vapour filtration. The only way to evaluate the applicability of various catalysts is to measure their activity. In the preliminary evaluation of catalytic performance it is important to use a method that is fast and reliable. Further, freshly produced oils should be utilised in catalyst testing for upgrading of pyrolysis oils because of their instability. For these reasons we have developed a test method for the estimation of the catalytic activity of various materials for the treatment of pyrolysis vapours without hydrogen. The product quality is estimated on the basis of the main compounds of pyrolysis vapours. Catalyst testing is typically carried out by conventional test procedures like batch reactors, continuous stirred-tank reactors (CSTR) or turbular reactors. The testing of catalytic activity is time-consuming. Special ASTM methods, so-called microactivity tests (MAT) D 3907-92 and D 5154-91, have been developed for the evaluation of the performance of cracking catalysts. However, these methods are unsuitable for the pyrolysis oils because a separate feed injection is applied and the space time is too long. An adaption of a commercial analytical scale pyrolyser has been used earlier as a catalyst reactor without pyrolysis as reported by Perez et al. [3]. However, the method cannot be used for studying the catalytic cracking of pyrolysis vapours. In this study, the injection port of a gas chromatograph (GC) was used as a catalyst reactor which allowed the catalytic conversion of pyrolysis vapours. The test method was applied to cracking pyrolysis vapours with zeolites. The aim of catalyst screening was to identify the pyrolysis products using a mass selective detector (MSD) and to determine the mass yields for hydrogen and carbon with an atomic emission detector (AED) [4,5]. In addition, the effect of the catalyst temperature on product distribution and oxygen removal was studied using AED. Zeolite cracking of pyrolysis oils has been widely studied [6–12]. The results of zeolite cracking from this novel method was compared with those obtained using conventional test procedures. A commercial FCC catalyst and vacuum gas oil was used as reference in this study.
2. Experimental
2.1. Materials In pyrolysis studies air-dried, bark-free sawdust of Scots pine (Pinus syl6estris) was used. The sawdust was ground and sieved on a round-holed screen and the 105–125 mm fraction was collected for testing. The elemental composition of the pine sawdust (moisture and ash-free basis) was: 51.1 wt.% carbon, 5.4 wt.% hydrogen, 0.1 wt.% nitrogen and 43.4 wt.% oxygen. The oxygen content was calculated by the difference [13]. The vacuum gas oil used had a boiling point range of 300–550°C.
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
195
The catalysts tested were a commercial FCC-catalyst, H-ZSM-5 (PQ corporation Valfor CBV-1502, Si/Al 150, SA 420 m2g − 1) and H-mordenite (BDH, Poole, UK, BDH-29881, clay binder, particle size 2 mm). H-mordenite was ground and sieved (0.1–0.25 mm). The catalysts were activated in air at 600°C over night and stored in a desiccator.
2.2. Methods A microscale pyrolysis and vapour-phase catalyst reactor was developed from commercially available instruments. The pyrolysis unit was a CDS Instruments Pyroprobe 1000 pyrolyser connected to an HP 5890 Series II gas chromatograph. The injection port of the GC was used as a fixed-bed catalytic reactor. The catalyst was placed into the liner tube (7.8 cm× 4 mm i.d.) of the GC’s split/splitless injector. The placement of the catalyst bed was adjusted with glasswool. Fig. 1 shows the main parts of the pyrolyser and injection port. The injection port maximum temperature is normally 400°C. Some catalysts require higher temperatures. For this reason, the thermostat was modified to achieve a catalyst temperature of 450 or 500°C. A fixed resistor was added in parallel with the thermostats’ platinum resistance-sensor. The correct temperature versus set point was calculated to achieve the temperature desired by the catalyst. Detection was carried out either with an HP 5921A atomic emission detector or alternatively with an HP 5970 mass
Fig. 1. The pyrolyser and the placement of the catalyst.
196
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
Fig. 2. The formation of the cracked products from rapid pyrolysis with GC/AED.
selective detector. The system was controlled by an HP 9000 series Pascal ChemStation computer. The chromatographic separation was carried out using an HP Ultra 1 fused silica capillary column (50 m × 0.32 mm i.d., film thickness 0.52 mm). The oven temperature program was 2 min at 30°C, 10°C min − 1 to 300°C, and 25 min at 300°C. Helium was used as carrier gas and the column pressure was 125 kPa. The vacuum gas oil sample was vaporised and the pine sawdust sample of 1–2 mg was pyrolysed in a quartz tube (2.5 cm×1.0 mm i.d.) at 600°C (the heating rate was 1000°C s − 1 and the heating time was 20 s). The interface temperature of the pyrolyser was 300°C. Vapours from rapid pyrolysis were led through the liner tube with or without the catalyst for instant analysis by the GC. For quantitative carbon and hydrogen determinations the AED’s mass responses were calibrated by injecting benzene solution. The wavelengths used for hydrogen, carbon and oxygen were 486.1, 495.7 and 777.3 nm, respectively. In pyrolysis the pine sawdust was converted into GC-eluated compounds, pyrolysis residue, coke and loss as shown in Fig. 2. The term ‘loss’ was used for high-molecular-mass compounds not eluated through the GC column but released from the raw material. In this study the GC-eluated compounds were classified into light volatile compounds (light fraction) and heavy volatile compounds (heavy fraction). The retention time (4 min) of n-pentane (b.p. 36°C) was the upper limit of light fraction. The sample was weighed with a microbalance Mettler AT20 before the pyrolysis and the pyrolysis residue was weighed afterwards. The pyrolysis residue did not contain GC-eluated compounds and hence was assumed to be carbon. Total mass balances for carbon and hydrogen were calculated from the calibration results and the weighed masses. The mass spectra identification of the main compounds was based on automatic library search (NBS REV – F/HP 9000 series 310) and on literature data [14– 16].
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
197
3. Results and discussion
3.1. Verification of method The pyrolysis products of pine sawdust from Pyroprobe were compared with GC-analysis results of pyrolysis oil from a bigger scale (1 kg h − 1) pyrolyser. Both products of pyrolysis vapours had practically similar composition [17,18]. It was concluded that the microscale pyrolyser can be used to produce relevant pyrolysis vapours for catalytic upgrading. For verification of the catalytic converter, vacuum gas oil and a commercial FCC catalyst from an oil refinery were chosen for reference. The fluid catalytic cracking (FCC) process produces gases and a gasoline fraction. Vacuum gas oil and its catalytically cracked vapours were analysed by GC/MSD. Fig. 3 shows the main compounds of vacuum gas oil and cracking products with the FCC catalyst. The catalytic cracking was carried out at 500°C instead of the typical temperature of 525°C because of the service life of materials in the injection port. The GC-eluated compounds of the vacuum gas oil contained a series of n-paraffins (saturated straight chain hydrocarbons) C11H24 –C37H76 and a residue fraction (b.p. range 400 – 500°C). The FCC catalyst cracked the high-molecularweight compounds (residue fraction) of the vacuum gas oil and produced unsaturated gases (mainly propene and butene), aromatic hydrocarbons (b.p. range 80–400°C) and a series of n-paraffins (C13H28 –C24H50). The aromatic hydrocarbons were from one to four ring aromatic hydrocarbons consisting mainly of benzene, naphthalene and indene as well as their derivatives. The product contained 36 area.% gases and gasoline fraction (boiling points below 200°C) and 64 area.% heavier fraction. Under these conditions the results were in line with those of the FCC-process in oil refining. On the basis of the experiments the microscale pyrolysis and vapour phase catalyst reactor appears to give relevant results.
3.2. Identification and comparison of reaction products with or without zeolites The method developed was used for studying the conversion of biomass pyrolysis products with different zeolites. The rapid pyrolysis products of pine sawdust without a catalyst and with H-ZSM-5, FCC and H-mordenite were identified by Py-GC/MSD. The main compounds in pyrolysis vapours without a catalyst and with H-mordenite are shown in Fig. 4. Without a catalyst the pyrolysis products of pine sawdust consisted mainly of carbon monoxide, carbon dioxide, water and other degradation products of polysaccharides and lignin. Formation of the main degradation compounds from pine wood was reported earlier by Ale´n et al. [17]. H-mordenite as well as other zeolites produced mainly carbon monoxide, carbon dioxide, water and aromatic hydrocarbons, e.g. benzene, naphthalene and also their derivatives. However, with H-ZSM-5 and FCC more oxygenated compounds (degradation products of polysaccharides and lignin mainly acetic acid, 2-furaldehyde and guaiacol) were obtained than with H-mordenite. From the catalyst tested,
198
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
Fig. 3. Total ion chromatograms of vacuum gas oil (a) dissolved in dichloromethane and (b) cracking with FCC catalyst at 500°C.
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
199
Fig. 4. The main compounds of pine sawdust pyrolysed at 600°C (a) without a catalyst and (b) with H-mordenite catalyst at 500°C: 1. CO/CO2/H2O, 2. acetaldehyde, 3. 2-propanone, 4. methylfurans, 5. acetic acid, 6. 1-hydroxy-2-propanone, 7. 1-hydroxy-2-butanone, 8. 2-furaldehyde, 9. a-angelica lactone, 10. 1-acetyloxy-2-propanone, 11. 5-methyl-2-furaldehyde, 12. 2-hydroxy-3-methyl-2-cyclopenten-1-one, 13. guaiacol, 14. 4-methylguaiacol, 15. 5-(hydroxymethyl)-2-furaldehyde, 16. 4-vinylguaiacol, 17. isoeugenol, 18. levoglucosan, 19. coniferaldehyde, 20. benzene, 21. toluene, 22. ethylbenzene, 23. xylenes, 24. 1-ethyl-4-methylbenzene, 25. naphthalene, 26. methylnaphthalenes, 27. dimethylnaphthalenes, 28. trimethylnapphthalenes, 29. phenanthrene or anthracene, 30. methylphenanthrenes or methylanthracenes.
200
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
Fig. 5. Carbon distribution in pyrolysis vapours in ( ) light fraction and () heavy fraction using FCC catalyst at 500°C with different bed lengths.
the H-ZSM-5 products contained most degradation compounds of polysaccharides and lignin monomers.
3.3. Effect of catalyst bed length In the cracking of pyrolysis vapours by the method presented, the process conditions i.e. the placement and the bed length of the catalyst have a considerable effect on product distribution. The placement of the catalyst in the liner tube was critical due to the steep temperature profile in the injection port. The placement of the catalyst bed was adjusted with glasswool (Fig. 1). The suitable bed length was determined with FCC at the temperature of 500°C (Fig. 5). The carbon content of the heavy fraction increased and the light fraction decreased while the length of the catalyst bed was cut down. Simultaneously, the amount of oxygen-containing compounds like acetic acid and 2-furaldehyde increased in reaction products. On the basis of these results the bed length of 4 mm was chosen for the experiments. In this way the residence time of pyrolysis vapours in the catalyst bed remained constant (about 30 ms) in all tests.
3.4. Effect of catalyst temperature The effect of temperature on product distribution was studied for H-ZSM-5, FCC and H-mordenite catalysts (Fig. 6). The influence of the catalyst temperature on the quantity and quality of the heavy fraction varied with different zeolites. For the FCC catalyst the temperature (400–500°C) had only a limited effect on carbon distribution but the chemical composition of products was changed. The amount of oxygen-containing compounds increased in the heavy fraction at lower tempera-
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
201
tures. The carbon distribution in the heavy fraction for H-ZSM-5 increased at decreasing temperature while the heavy fraction contained more degradation products of polysaccharides and lignin monomers. With H-mordenite the product distribution of the heavy fraction was almost the same at different temperatures but carbon distribution in the heavy fraction diminished considerably at decreasing temperature. The cracking products did not desorb from the surface of the catalyst at 300°C. To achieve a considerable amount of heavy fraction and deoxygenation different catalyst temperatures were chosen. The temperatures used for FCC, H-ZSM-5 and H-mordenite were 500, 450 and 500°C, respectively. Quantitative analysis of carbon and hydrogen were determined at these temperatures.
3.5. Mass balances for carbon and hydrogen Mass balances for carbon and hydrogen were determined with zeolite catalysts. The carbon and hydrogen yields (wt.% of carbon or hydrogen in the raw material) are shown in Fig. 7. The first column presents the yields in the GC-eluated compounds (light and heavy fraction) and the second one the yields in the pyrolysis residue and the coke and loss. The carbon yields in GC-eluated compounds (light and heavy fraction combined) of pine sawdust were without any catalyst 61 wt.% and with H-ZSM-5 46 wt.%, FCC 39 wt.% and H-mordenite 34 wt.%. When the yields of zeolites cracking was compared to those without a catalyst, the carbon and hydrogen yields in the heavy fraction decreased while the yields in the light fraction increased. The carbon yields in the heavy fraction were: no catalyst 48 wt.%, H-ZSM-5 21 wt.%, FCC 12 wt.%, H-mordenite 11 wt.%. Carbon yields in the light fraction were about 25 wt.% for all zeolites. The carbon amounts of pyrolysis residue were about the same (25–29 wt.%) for all experiments. The yields of pyrolysis residue calculated for dry ash-free sawdust were 13–14 wt.% which correspond to the char yields reported by Diebold et al.
Fig. 6. Carbon distribution in pyrolysis vapours in heavy fraction using ("), FCC; ( ), H-ZSM-5 and () H-mordenite catalysts at different temperatures.
202
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
Fig. 7. The yields for (a) carbon and (b) hydrogen (wt.% of carbon or hydrogen in the raw material).
[19,20]. During pyrolysis the activated white-coloured catalyst changed to gray in the upper layer. The coke content of the catalyst was too low to be quantified. For these reasons, the coke and loss were combined and calculated by the difference. The carbon amounts in coke and loss were 11 wt.% without a catalyst and 27 wt.% with H-ZSM-5, 34 wt.% with FCC and 42 wt.% with H-mordenite. The hydrogen content in coke and loss were similar (21–25 wt.%) with or without a catalyst. On the basis of these results it was concluded that the oxygen-containing compounds in pyrolysis vapours were converted mainly into carbon oxides and water by zeolites, and the carbon yields in the heavy fraction decreased as previously reported by Horne et al. [21]. These results support the presumptions [22] that the coke consists mainly of carbon.
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
203
Table 1 Oxygen distribution in pyrolysis vapours in light and heavy fractions with different zeolites Catalyst
Temperature (°C) Oxygen in light fraction (wt.% of GC measured oxygen)
Oxygen in heavy fraction (wt.% of GC measured oxygen)
FCC H-ZSM-5 H-mordenite No catalyst
500 450 500 —
3 5 0 24
97 95 100 76
3.6. Oxygen remo6al Oxygen in reaction products was determined by AED in order to find out the efficiency of deoxygenation with different zeolites. The oxygen distributions of pyrolysis vapours are presented in Table 1. The zeolites were effective in the removal of oxygen, but the liquid yields decreased as also reported by Williams et al. [23]. The efficiency of deoxygenation with the fresh catalysts decreased in the order: H-mordenite\ FCC \ H-ZSM-5.
4. Conclusions The results indicate that the microscale pyrolysis and vapour-phase catalyst reactor was suitable for screening catalysts. The zeolites were effective in the removal of oxygen and the liquid yields were low as also reported from bigger scale processes. The advantages of the method were high speed and the small amount of materials needed in the test. The Py-GC combination with the catalyst in the injection port can be applied to study the behaviour of catalysts with various pyrolysis raw materials like biomasses, liquid products, plastics, rubbers, wastes, etc. In addition, different detection techniques, i.e. atomic emission, mass selective and flame ionisation detectors, can be used to determine amounts for various elements, identify and quantify compounds.
Acknowledgements This work was supported by the EU-Joule programme under Contract no. JOR3-CT95-0025 (Bio Fuel Oil for Power Plant and Boilers). The authors thank the project coordinator Prof. K. Sipila¨ and the Upgrading work group members R. Maggi, C. Lahousse, D. Meier, A.Oasmaa and Y. Solantausta for fruitful discussions.
204
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
References [1] F. Shafizadeh Fundamentals of thermochemical biomass conversion, in: P.P. Overend, Y.A. Milne, L.K. Mudge (Eds.), Elsevier, New York, 1985, pp. 183 – 217. [2] R.E. Maggi, D.C. Elliott, in: A.V. Bridgwater, D.G.B. Boocock (Eds.), Developments in Thermochemical Biomass Conversion, Blackie Academic and Professional, 1997, pp. 575 – 588. [3] G. Perez, M. Raimondo, A. De Stefanis, A.A.G. Tomlinson, J. Anal. Appl. Pyrolysis 35 (1995) 157 – 166. [4] B.D. Quimby, J.J. Sullivan, Anal. Chem. 62 (1990) 1027 – 1034. [5] J.J. Sullivan, B.D. Quimby, Anal. Chem. 62 (1990) 1034 – 1043. [6] A.V. Bridgwater, Catal. Today 29 (1996) 285 – 295. [7] R.J. Evans, T. Milne, in: E.J. Soltes, T.A. Milne (Eds.), Pyrolysis Oils from Biomass, Producing, Analysing and Upgrading, ACS Symposium Series No. 376, American Chemical Society, Washington DC. 1988, chap. 26, pp. 311–327. [8] P.A. Horne, P.T. Williams, J. Anal. Appl. Pyrolysis 34 (1995) 65 – 85. [9] P.T. Williams, P.A. Horne, Biomass Bioenergy 7 (1994) 223 – 236. [10] P. Chantal, S. Kaliaguine, J.L. Grandmaison, A. Mahay, Appl. Catal. 10 (1984) 317 – 332. [11] J.D. Adjaye, N.N. Bakhshi, Fuel Process. Technol. 45 (1995) 161 – 183. [12] R.K. Sharma, N.N. Bakhshi, Can. J. Chem. Eng. 71 (1993) 383 – 391. [13] R. Ale´n, P. Oesch, E. Kuoppala, J. Anal. Appl. Pyrolysis 35 (1995) 259 – 265. [14] O. Faix, D. Meier, I. Fortman, Holz. Roh- Werkst. 48 (1990) 281 – 285. [15] R.J. Evans, T.A. Milne, Energy Fuels 1 (1987) 123 – 137. [16] O. Faix, I. Fortman, J. Bremer, D. Meier, Holz. Roh- Werkst. 49 (1991) 213 – 219. [17] R. Ale´n, E. Kuoppala, P. Oesch, J. Anal. Appl. Pyrolysis 36 (1996) 137 – 148. [18] K. Sipila¨, E. Kuoppala, L. Fagerna¨s, A. Oasmaa, Biomass Bioenergy (accepted). [19] J. Diebold, J. Scahill, in: E.J. Soltes, T.A. Milne (Eds.), Pyrolysis Oils from Biomass, Producing, Analysing and Upgrading, ACS Symposium Series No. 376, American Chemical Society, Washington DC, 1988, chap. 23, pp. 264–276. [20] J. Diebold, J. Scahill, R. Bain, H. Chum, S. Black, T. Milne, R. Evans, B. Rajai, in: E. Hogan, J. Robert, G. Grassi, A.V. Bridgwater (Eds.), Biomass Thermal Processing, Proceedings of the first Canada/European Community R and D Contractors Meeting, cpl Press Ottawa, Canada, 1992, pp. 101 – 108. [21] P.A. Horne, P.T. Williams, Renew. Energy 5 (1994) 810 – 812. [22] J.T. Richardson, Principles of Catalyst Development, Plenum Press, New York, 1992, pp. 210 – 215. [23] P.T. Williams, P.A. Horne, J. Anal. Appl. Pyrolysis 31 (1995) 39 – 61.
.
A novel test method for cracking catalysts M.I. Nokkosma¨ki a,*, E.T. Kuoppala b, E.A. Leppa¨ma¨ki b, A.O.I. Krause a a
Helsinki Uni6ersity of Technology, Department of Chemical Technology, P.O. Box 6100, FIN-02015 HUT, Finland b VTT Energy, Energy Production Technologies, P.O. Box 1601, FIN-02044 VTT, Finland Received 24 August 1997; accepted 25 September 1997
Abstract A novel microscale test method was developed for cracking catalysts. The pyrolysis unit was a commercial pyrolyser connected to a gas chromatograph. The injection port of the gas chromatograph was used as a fixed bed catalyst reactor. The test method was applied to cracking pyrolysis vapours of Scots pine sawdust with zeolites. Cracking of vacuum gas oil on a commercial FCC catalyst was used as reference. Detection of reaction products was carried out with a mass selective detector to identify the compounds or with an atomic emission detector to quantify the various elements. The results, in agreement with the literature indicate that the microscale pyrolysis and vapour-phase catalyst reactor was suitable for screening catalysts. Pyrolysis vapours were converted mainly into gases and aromatic hydrocarbons with zeolites. The zeolite catalysts were effective in the removal of oxygen but the liquid yields were low. © 1998 Elsevier Science B.V. Keywords: Catalyst screening; Cracking; Catalysts; Zeolites; Biomass; Pyrolysis; Pyrolysis vapours
1. Introduction One alternative to using biomasses is to pyrolyse them to liquid products. The products, mechanism and kinetics of pyrolytic reactions have been discussed previously [1]. These pyrolysis oils can be used, e.g. as fuels for energy production. However, the major limitation is their instability, i.e. quality changes during * Corresponding author. Tel.: +358 9 4566582; fax: + 358 9 460493; e-mail: Milja.Nokkosmaki@ vtt.fi 0165-2370/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 2 3 7 0 ( 9 7 ) 0 0 0 8 0 - 6
194
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
storage. The pyrolysis oils can be upgraded via catalytic conversion with or without hydrogen [2]. Alternatives for catalytic upgrading are physical methods like addition of water or organic solvents and hot vapour filtration. The only way to evaluate the applicability of various catalysts is to measure their activity. In the preliminary evaluation of catalytic performance it is important to use a method that is fast and reliable. Further, freshly produced oils should be utilised in catalyst testing for upgrading of pyrolysis oils because of their instability. For these reasons we have developed a test method for the estimation of the catalytic activity of various materials for the treatment of pyrolysis vapours without hydrogen. The product quality is estimated on the basis of the main compounds of pyrolysis vapours. Catalyst testing is typically carried out by conventional test procedures like batch reactors, continuous stirred-tank reactors (CSTR) or turbular reactors. The testing of catalytic activity is time-consuming. Special ASTM methods, so-called microactivity tests (MAT) D 3907-92 and D 5154-91, have been developed for the evaluation of the performance of cracking catalysts. However, these methods are unsuitable for the pyrolysis oils because a separate feed injection is applied and the space time is too long. An adaption of a commercial analytical scale pyrolyser has been used earlier as a catalyst reactor without pyrolysis as reported by Perez et al. [3]. However, the method cannot be used for studying the catalytic cracking of pyrolysis vapours. In this study, the injection port of a gas chromatograph (GC) was used as a catalyst reactor which allowed the catalytic conversion of pyrolysis vapours. The test method was applied to cracking pyrolysis vapours with zeolites. The aim of catalyst screening was to identify the pyrolysis products using a mass selective detector (MSD) and to determine the mass yields for hydrogen and carbon with an atomic emission detector (AED) [4,5]. In addition, the effect of the catalyst temperature on product distribution and oxygen removal was studied using AED. Zeolite cracking of pyrolysis oils has been widely studied [6–12]. The results of zeolite cracking from this novel method was compared with those obtained using conventional test procedures. A commercial FCC catalyst and vacuum gas oil was used as reference in this study.
2. Experimental
2.1. Materials In pyrolysis studies air-dried, bark-free sawdust of Scots pine (Pinus syl6estris) was used. The sawdust was ground and sieved on a round-holed screen and the 105–125 mm fraction was collected for testing. The elemental composition of the pine sawdust (moisture and ash-free basis) was: 51.1 wt.% carbon, 5.4 wt.% hydrogen, 0.1 wt.% nitrogen and 43.4 wt.% oxygen. The oxygen content was calculated by the difference [13]. The vacuum gas oil used had a boiling point range of 300–550°C.
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
195
The catalysts tested were a commercial FCC-catalyst, H-ZSM-5 (PQ corporation Valfor CBV-1502, Si/Al 150, SA 420 m2g − 1) and H-mordenite (BDH, Poole, UK, BDH-29881, clay binder, particle size 2 mm). H-mordenite was ground and sieved (0.1–0.25 mm). The catalysts were activated in air at 600°C over night and stored in a desiccator.
2.2. Methods A microscale pyrolysis and vapour-phase catalyst reactor was developed from commercially available instruments. The pyrolysis unit was a CDS Instruments Pyroprobe 1000 pyrolyser connected to an HP 5890 Series II gas chromatograph. The injection port of the GC was used as a fixed-bed catalytic reactor. The catalyst was placed into the liner tube (7.8 cm× 4 mm i.d.) of the GC’s split/splitless injector. The placement of the catalyst bed was adjusted with glasswool. Fig. 1 shows the main parts of the pyrolyser and injection port. The injection port maximum temperature is normally 400°C. Some catalysts require higher temperatures. For this reason, the thermostat was modified to achieve a catalyst temperature of 450 or 500°C. A fixed resistor was added in parallel with the thermostats’ platinum resistance-sensor. The correct temperature versus set point was calculated to achieve the temperature desired by the catalyst. Detection was carried out either with an HP 5921A atomic emission detector or alternatively with an HP 5970 mass
Fig. 1. The pyrolyser and the placement of the catalyst.
196
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
Fig. 2. The formation of the cracked products from rapid pyrolysis with GC/AED.
selective detector. The system was controlled by an HP 9000 series Pascal ChemStation computer. The chromatographic separation was carried out using an HP Ultra 1 fused silica capillary column (50 m × 0.32 mm i.d., film thickness 0.52 mm). The oven temperature program was 2 min at 30°C, 10°C min − 1 to 300°C, and 25 min at 300°C. Helium was used as carrier gas and the column pressure was 125 kPa. The vacuum gas oil sample was vaporised and the pine sawdust sample of 1–2 mg was pyrolysed in a quartz tube (2.5 cm×1.0 mm i.d.) at 600°C (the heating rate was 1000°C s − 1 and the heating time was 20 s). The interface temperature of the pyrolyser was 300°C. Vapours from rapid pyrolysis were led through the liner tube with or without the catalyst for instant analysis by the GC. For quantitative carbon and hydrogen determinations the AED’s mass responses were calibrated by injecting benzene solution. The wavelengths used for hydrogen, carbon and oxygen were 486.1, 495.7 and 777.3 nm, respectively. In pyrolysis the pine sawdust was converted into GC-eluated compounds, pyrolysis residue, coke and loss as shown in Fig. 2. The term ‘loss’ was used for high-molecular-mass compounds not eluated through the GC column but released from the raw material. In this study the GC-eluated compounds were classified into light volatile compounds (light fraction) and heavy volatile compounds (heavy fraction). The retention time (4 min) of n-pentane (b.p. 36°C) was the upper limit of light fraction. The sample was weighed with a microbalance Mettler AT20 before the pyrolysis and the pyrolysis residue was weighed afterwards. The pyrolysis residue did not contain GC-eluated compounds and hence was assumed to be carbon. Total mass balances for carbon and hydrogen were calculated from the calibration results and the weighed masses. The mass spectra identification of the main compounds was based on automatic library search (NBS REV – F/HP 9000 series 310) and on literature data [14– 16].
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
197
3. Results and discussion
3.1. Verification of method The pyrolysis products of pine sawdust from Pyroprobe were compared with GC-analysis results of pyrolysis oil from a bigger scale (1 kg h − 1) pyrolyser. Both products of pyrolysis vapours had practically similar composition [17,18]. It was concluded that the microscale pyrolyser can be used to produce relevant pyrolysis vapours for catalytic upgrading. For verification of the catalytic converter, vacuum gas oil and a commercial FCC catalyst from an oil refinery were chosen for reference. The fluid catalytic cracking (FCC) process produces gases and a gasoline fraction. Vacuum gas oil and its catalytically cracked vapours were analysed by GC/MSD. Fig. 3 shows the main compounds of vacuum gas oil and cracking products with the FCC catalyst. The catalytic cracking was carried out at 500°C instead of the typical temperature of 525°C because of the service life of materials in the injection port. The GC-eluated compounds of the vacuum gas oil contained a series of n-paraffins (saturated straight chain hydrocarbons) C11H24 –C37H76 and a residue fraction (b.p. range 400 – 500°C). The FCC catalyst cracked the high-molecularweight compounds (residue fraction) of the vacuum gas oil and produced unsaturated gases (mainly propene and butene), aromatic hydrocarbons (b.p. range 80–400°C) and a series of n-paraffins (C13H28 –C24H50). The aromatic hydrocarbons were from one to four ring aromatic hydrocarbons consisting mainly of benzene, naphthalene and indene as well as their derivatives. The product contained 36 area.% gases and gasoline fraction (boiling points below 200°C) and 64 area.% heavier fraction. Under these conditions the results were in line with those of the FCC-process in oil refining. On the basis of the experiments the microscale pyrolysis and vapour phase catalyst reactor appears to give relevant results.
3.2. Identification and comparison of reaction products with or without zeolites The method developed was used for studying the conversion of biomass pyrolysis products with different zeolites. The rapid pyrolysis products of pine sawdust without a catalyst and with H-ZSM-5, FCC and H-mordenite were identified by Py-GC/MSD. The main compounds in pyrolysis vapours without a catalyst and with H-mordenite are shown in Fig. 4. Without a catalyst the pyrolysis products of pine sawdust consisted mainly of carbon monoxide, carbon dioxide, water and other degradation products of polysaccharides and lignin. Formation of the main degradation compounds from pine wood was reported earlier by Ale´n et al. [17]. H-mordenite as well as other zeolites produced mainly carbon monoxide, carbon dioxide, water and aromatic hydrocarbons, e.g. benzene, naphthalene and also their derivatives. However, with H-ZSM-5 and FCC more oxygenated compounds (degradation products of polysaccharides and lignin mainly acetic acid, 2-furaldehyde and guaiacol) were obtained than with H-mordenite. From the catalyst tested,
198
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
Fig. 3. Total ion chromatograms of vacuum gas oil (a) dissolved in dichloromethane and (b) cracking with FCC catalyst at 500°C.
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
199
Fig. 4. The main compounds of pine sawdust pyrolysed at 600°C (a) without a catalyst and (b) with H-mordenite catalyst at 500°C: 1. CO/CO2/H2O, 2. acetaldehyde, 3. 2-propanone, 4. methylfurans, 5. acetic acid, 6. 1-hydroxy-2-propanone, 7. 1-hydroxy-2-butanone, 8. 2-furaldehyde, 9. a-angelica lactone, 10. 1-acetyloxy-2-propanone, 11. 5-methyl-2-furaldehyde, 12. 2-hydroxy-3-methyl-2-cyclopenten-1-one, 13. guaiacol, 14. 4-methylguaiacol, 15. 5-(hydroxymethyl)-2-furaldehyde, 16. 4-vinylguaiacol, 17. isoeugenol, 18. levoglucosan, 19. coniferaldehyde, 20. benzene, 21. toluene, 22. ethylbenzene, 23. xylenes, 24. 1-ethyl-4-methylbenzene, 25. naphthalene, 26. methylnaphthalenes, 27. dimethylnaphthalenes, 28. trimethylnapphthalenes, 29. phenanthrene or anthracene, 30. methylphenanthrenes or methylanthracenes.
200
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
Fig. 5. Carbon distribution in pyrolysis vapours in ( ) light fraction and () heavy fraction using FCC catalyst at 500°C with different bed lengths.
the H-ZSM-5 products contained most degradation compounds of polysaccharides and lignin monomers.
3.3. Effect of catalyst bed length In the cracking of pyrolysis vapours by the method presented, the process conditions i.e. the placement and the bed length of the catalyst have a considerable effect on product distribution. The placement of the catalyst in the liner tube was critical due to the steep temperature profile in the injection port. The placement of the catalyst bed was adjusted with glasswool (Fig. 1). The suitable bed length was determined with FCC at the temperature of 500°C (Fig. 5). The carbon content of the heavy fraction increased and the light fraction decreased while the length of the catalyst bed was cut down. Simultaneously, the amount of oxygen-containing compounds like acetic acid and 2-furaldehyde increased in reaction products. On the basis of these results the bed length of 4 mm was chosen for the experiments. In this way the residence time of pyrolysis vapours in the catalyst bed remained constant (about 30 ms) in all tests.
3.4. Effect of catalyst temperature The effect of temperature on product distribution was studied for H-ZSM-5, FCC and H-mordenite catalysts (Fig. 6). The influence of the catalyst temperature on the quantity and quality of the heavy fraction varied with different zeolites. For the FCC catalyst the temperature (400–500°C) had only a limited effect on carbon distribution but the chemical composition of products was changed. The amount of oxygen-containing compounds increased in the heavy fraction at lower tempera-
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
201
tures. The carbon distribution in the heavy fraction for H-ZSM-5 increased at decreasing temperature while the heavy fraction contained more degradation products of polysaccharides and lignin monomers. With H-mordenite the product distribution of the heavy fraction was almost the same at different temperatures but carbon distribution in the heavy fraction diminished considerably at decreasing temperature. The cracking products did not desorb from the surface of the catalyst at 300°C. To achieve a considerable amount of heavy fraction and deoxygenation different catalyst temperatures were chosen. The temperatures used for FCC, H-ZSM-5 and H-mordenite were 500, 450 and 500°C, respectively. Quantitative analysis of carbon and hydrogen were determined at these temperatures.
3.5. Mass balances for carbon and hydrogen Mass balances for carbon and hydrogen were determined with zeolite catalysts. The carbon and hydrogen yields (wt.% of carbon or hydrogen in the raw material) are shown in Fig. 7. The first column presents the yields in the GC-eluated compounds (light and heavy fraction) and the second one the yields in the pyrolysis residue and the coke and loss. The carbon yields in GC-eluated compounds (light and heavy fraction combined) of pine sawdust were without any catalyst 61 wt.% and with H-ZSM-5 46 wt.%, FCC 39 wt.% and H-mordenite 34 wt.%. When the yields of zeolites cracking was compared to those without a catalyst, the carbon and hydrogen yields in the heavy fraction decreased while the yields in the light fraction increased. The carbon yields in the heavy fraction were: no catalyst 48 wt.%, H-ZSM-5 21 wt.%, FCC 12 wt.%, H-mordenite 11 wt.%. Carbon yields in the light fraction were about 25 wt.% for all zeolites. The carbon amounts of pyrolysis residue were about the same (25–29 wt.%) for all experiments. The yields of pyrolysis residue calculated for dry ash-free sawdust were 13–14 wt.% which correspond to the char yields reported by Diebold et al.
Fig. 6. Carbon distribution in pyrolysis vapours in heavy fraction using ("), FCC; ( ), H-ZSM-5 and () H-mordenite catalysts at different temperatures.
202
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
Fig. 7. The yields for (a) carbon and (b) hydrogen (wt.% of carbon or hydrogen in the raw material).
[19,20]. During pyrolysis the activated white-coloured catalyst changed to gray in the upper layer. The coke content of the catalyst was too low to be quantified. For these reasons, the coke and loss were combined and calculated by the difference. The carbon amounts in coke and loss were 11 wt.% without a catalyst and 27 wt.% with H-ZSM-5, 34 wt.% with FCC and 42 wt.% with H-mordenite. The hydrogen content in coke and loss were similar (21–25 wt.%) with or without a catalyst. On the basis of these results it was concluded that the oxygen-containing compounds in pyrolysis vapours were converted mainly into carbon oxides and water by zeolites, and the carbon yields in the heavy fraction decreased as previously reported by Horne et al. [21]. These results support the presumptions [22] that the coke consists mainly of carbon.
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
203
Table 1 Oxygen distribution in pyrolysis vapours in light and heavy fractions with different zeolites Catalyst
Temperature (°C) Oxygen in light fraction (wt.% of GC measured oxygen)
Oxygen in heavy fraction (wt.% of GC measured oxygen)
FCC H-ZSM-5 H-mordenite No catalyst
500 450 500 —
3 5 0 24
97 95 100 76
3.6. Oxygen remo6al Oxygen in reaction products was determined by AED in order to find out the efficiency of deoxygenation with different zeolites. The oxygen distributions of pyrolysis vapours are presented in Table 1. The zeolites were effective in the removal of oxygen, but the liquid yields decreased as also reported by Williams et al. [23]. The efficiency of deoxygenation with the fresh catalysts decreased in the order: H-mordenite\ FCC \ H-ZSM-5.
4. Conclusions The results indicate that the microscale pyrolysis and vapour-phase catalyst reactor was suitable for screening catalysts. The zeolites were effective in the removal of oxygen and the liquid yields were low as also reported from bigger scale processes. The advantages of the method were high speed and the small amount of materials needed in the test. The Py-GC combination with the catalyst in the injection port can be applied to study the behaviour of catalysts with various pyrolysis raw materials like biomasses, liquid products, plastics, rubbers, wastes, etc. In addition, different detection techniques, i.e. atomic emission, mass selective and flame ionisation detectors, can be used to determine amounts for various elements, identify and quantify compounds.
Acknowledgements This work was supported by the EU-Joule programme under Contract no. JOR3-CT95-0025 (Bio Fuel Oil for Power Plant and Boilers). The authors thank the project coordinator Prof. K. Sipila¨ and the Upgrading work group members R. Maggi, C. Lahousse, D. Meier, A.Oasmaa and Y. Solantausta for fruitful discussions.
204
M.I. Nokkosma¨ki et al. / J. Anal. Appl. Pyrolysis 44 (1998) 193–204
References [1] F. Shafizadeh Fundamentals of thermochemical biomass conversion, in: P.P. Overend, Y.A. Milne, L.K. Mudge (Eds.), Elsevier, New York, 1985, pp. 183 – 217. [2] R.E. Maggi, D.C. Elliott, in: A.V. Bridgwater, D.G.B. Boocock (Eds.), Developments in Thermochemical Biomass Conversion, Blackie Academic and Professional, 1997, pp. 575 – 588. [3] G. Perez, M. Raimondo, A. De Stefanis, A.A.G. Tomlinson, J. Anal. Appl. Pyrolysis 35 (1995) 157 – 166. [4] B.D. Quimby, J.J. Sullivan, Anal. Chem. 62 (1990) 1027 – 1034. [5] J.J. Sullivan, B.D. Quimby, Anal. Chem. 62 (1990) 1034 – 1043. [6] A.V. Bridgwater, Catal. Today 29 (1996) 285 – 295. [7] R.J. Evans, T. Milne, in: E.J. Soltes, T.A. Milne (Eds.), Pyrolysis Oils from Biomass, Producing, Analysing and Upgrading, ACS Symposium Series No. 376, American Chemical Society, Washington DC. 1988, chap. 26, pp. 311–327. [8] P.A. Horne, P.T. Williams, J. Anal. Appl. Pyrolysis 34 (1995) 65 – 85. [9] P.T. Williams, P.A. Horne, Biomass Bioenergy 7 (1994) 223 – 236. [10] P. Chantal, S. Kaliaguine, J.L. Grandmaison, A. Mahay, Appl. Catal. 10 (1984) 317 – 332. [11] J.D. Adjaye, N.N. Bakhshi, Fuel Process. Technol. 45 (1995) 161 – 183. [12] R.K. Sharma, N.N. Bakhshi, Can. J. Chem. Eng. 71 (1993) 383 – 391. [13] R. Ale´n, P. Oesch, E. Kuoppala, J. Anal. Appl. Pyrolysis 35 (1995) 259 – 265. [14] O. Faix, D. Meier, I. Fortman, Holz. Roh- Werkst. 48 (1990) 281 – 285. [15] R.J. Evans, T.A. Milne, Energy Fuels 1 (1987) 123 – 137. [16] O. Faix, I. Fortman, J. Bremer, D. Meier, Holz. Roh- Werkst. 49 (1991) 213 – 219. [17] R. Ale´n, E. Kuoppala, P. Oesch, J. Anal. Appl. Pyrolysis 36 (1996) 137 – 148. [18] K. Sipila¨, E. Kuoppala, L. Fagerna¨s, A. Oasmaa, Biomass Bioenergy (accepted). [19] J. Diebold, J. Scahill, in: E.J. Soltes, T.A. Milne (Eds.), Pyrolysis Oils from Biomass, Producing, Analysing and Upgrading, ACS Symposium Series No. 376, American Chemical Society, Washington DC, 1988, chap. 23, pp. 264–276. [20] J. Diebold, J. Scahill, R. Bain, H. Chum, S. Black, T. Milne, R. Evans, B. Rajai, in: E. Hogan, J. Robert, G. Grassi, A.V. Bridgwater (Eds.), Biomass Thermal Processing, Proceedings of the first Canada/European Community R and D Contractors Meeting, cpl Press Ottawa, Canada, 1992, pp. 101 – 108. [21] P.A. Horne, P.T. Williams, Renew. Energy 5 (1994) 810 – 812. [22] J.T. Richardson, Principles of Catalyst Development, Plenum Press, New York, 1992, pp. 210 – 215. [23] P.T. Williams, P.A. Horne, J. Anal. Appl. Pyrolysis 31 (1995) 39 – 61.
.