Res Chem Intermed (2011) 37:1293–1303 DOI 10.1007/s11164-011-0397-5
Upgrading of light cycle oil by partial hydrogenation and selective ring opening over an iridium bifunctional catalyst Dipali P. Upare • R. Nageswara Rao Songhun Yoon • Chul Wee Lee
•
Received: 31 May 2011 / Accepted: 21 July 2011 / Published online: 13 September 2011 Ó Springer Science+Business Media B.V. 2011
Abstract An iridium catalyst supported on USY zeolite has been evaluated for hydrotreating of light cycle oil (LCO) to reduce its aromaticity and improve its cetane index. The products were analyzed by GC and 13C NMR spectrometry to determine quantitatively the aromatic carbon content and the increase in cetane index. Addition of an appropriate amount of potassium was found to be an effective way of optimizing the acid properties of the catalyst; reducing the number of strongly acid sites reduced its cracking activity. The results confirm that 0.9% (w/w) Ir/USY zeolite catalyst doped with 0.75% (w/w) K was highly suitable for partial hydrogenation and ring opening of LCO to improve cetane quality, thus increasing the extent of its blending ratio with the diesel pool. Keywords LCO Selective ring opening (SRO) Bifunctional catalyst Cetane number (CN) 13C NMR spectra
Introduction Heavy and extra-heavy oils, including bitumen, are unconventional oil resources characterized by high viscosity and density. Because these are deficient in hydrogen, with high aromatic carbon, sulfur, and metal content, they require upgrading to D. P. Upare S. Yoon C. W. Lee (&) Green Chemistry Division, Petroleum Displacement Technology Center, Korea Research Institute of Chemical Technology, Daejeon 305-600, Korea e-mail:
[email protected] D. P. Upare School of Science, University of Science and Technology (UST), Daejeon 305-333, Korea R. Nageswara Rao Analytical Chemistry Division, IICT, Tarnaka, Hyderabad, Andhra Pradesh 500007, India
123
1294
D. P. Upare et al.
convert them into feedstock suitable for normal refineries [1]. According to the International Energy Agency (IEA), there are 8–9 trillion barrels of heavy crudes and bitumen throughout the world [2]. The energy demand of the world is continuously increasing while conventional oil resources are declining. Because of the shortage of oil in known petroleum reserves, less exploited energy resources, for example bitumen, and heavy and extra-heavy oils, have become more attractive [3, 4]. The demand for clean and high-quality transportation fuels is increasing, and environmental regulation of their specifications is becoming more stringent. These developments pose not only economic but also technological challenges for the global refining industry. For example, annual diesel consumption is currently growing at a rate of 5% worldwide. To meet the increased demand for diesel fuel, the most viable option is to blend light cycle oil (LCO) with the diesel pool. LCO accounts for 10–20% of the fluid catalytic cracking (FCC) products and was formerly used as a low-value product for heating purposes [5]. In the last few years, there has been a steep decline in fuel oil consumption worldwide, because of which disposal of LCO has become an issue for refineries. Thus value addition to LCO not only helps to meet the growing demand for diesel fuel but also the disposal of LCO. However, LCO has a low cetane index, high density, and high sulfur, unsaturated, and aromatic content, properties which adversely affect the quality of the resulting diesel fuel thus limiting its blending ratio [6]. There is, therefore, great interest in the development of suitable catalysts for upgrading LCO to increase its blending ratio with the diesel pool. Generally, hydrotreating is one of the main catalytic processes used for upgrading of LCO in petroleum refineries [7, 8]. It involves: (1) aromatic saturation (ASAT); (2) mild hydrocracking; and (3) deep hydrogenation followed by selective ring opening (SRO) of naphthenic structures. Although these processes lead to significant improvement in product quality, they suffer from several disadvantages. In ASAT, the aromatics are converted into naphthenes by aromatic ring saturation. However, this is not sufficient to improve fuel quality because naphthenes have low cetane numbers (CNs) compared with paraffins containing an equal number of carbon atoms [9]. Further, the naphthenes may reconvert into aromatics during combustion which in turn reduces fuel quality significantly. Deep hydrogenation followed by opening of naphthenic rings requires very high consumption of hydrogen that affects the economics of the process. Recently, as an alternative to overcome these problems, partial saturation of aromatic rings followed by SRO of naphthenic structures was suggested for production of alkanes of the same carbon number [10]. These are considered to be ideal reaction pathways for converting LCO to higher-quality diesel fuels [11, 12]. Thus, there is increased interest in SRO catalysis. Most investigations of ring opening reported in the literature were on C5–C10 naphthenes as model compounds [13, 14], with only a few studies of the effects of ring-opening catalysts on improvement of the fuel quality of real feedstocks.
123
Upgrading of light cycle oil
1295
In this paper we report an investigation of the catalytic performance of an Ir/USY zeolite catalyst for ring opening and cetane improvement of LCO. This was accomplished by modifying the acidic sites of the support by doping with small amounts of potassium. The effect is related to K-induced changes in the properties of the support which indirectly affect the metal, probably as a result of metal– support interactions [15]. Its effect on conversion and selectivity for ring-opening products of methylcyclopentane (MCP) as a model compound was studied initially. The best among the modifications was selected for hydrogenolysis of LCO in a batch-type reactor. The products were analyzed by GC and 13C NMR spectrometry and the results indicate that 0.9% (w/w) Ir/USY zeolite catalyst doped with 0.75% K was highly suitable for upgrading LCO to improve its cetane quality, thus increasing its blending ratio with the diesel pool.
Experimental Catalyst preparation The catalyst was prepared by impregnation with 0.9% (w/w) Ir and loading with 0.75% (w/w) K, using aqueous solutions of IrCl33H2O and KNO3 (Samchun Pure Chemicals, Korea). The support material USY zeolite with an Si-to-Al ratio of 40 (Zeolyst International, USA) was used as received. After impregnation, the catalysts were dried overnight in an oven at 100 °C and then calcined at 400 °C for 4 h in air. Catalyst characterization BET surface area, pore size, and pore volume measurement of the catalysts were carried out using liquid nitrogen and use of a Micromeritics Instruments (USA) ASAP2420 adsorption analyzer. All the samples were degassed at 250 °C for 2 h before the measurements to remove the adsorbed moisture from the catalysts’ surface and pores. The BET area and pore size distribution (PSD) were calculated from the BET and BJH equations, respectively, by the instrument software. The acidity of the catalysts was measured by temperature-programmed desorption (TPD) of ammonia using an Autochem II 2920 chemisorption analyzer (Micromeritics). The catalyst (0.1 g) was placed in an adsorption vessel and heated to 450 °C at a rate of 10 °C/min in an He flow for 1 h. Later it was cooled to 100 °C in an He flow. At this temperature 0.1% NH3 in N2 was passed through the sample for 1 h. Then, NH3 gas was changed to He and the sample was cooled to room temperature. TPD was run from room temperature to 600 °C at a heating rate of 10 °C/min with the He flow at 30 mL/min. The gases leaving the reactor were monitored by use of a thermal conductivity detector (TCD). For TPR measurement the same instrument was used. Catalyst sample (0.2 g) was heated slowly at 10 °C/min, then dried at 500 °C for 2 h. The TPR runs were performed with a heating rate of 10 °C/min to 800 °C for 30 min with 10% H2/Ar (flow rate = 30 cm3/min). Hydrogen consumption was monitored by use of a TCD.
123
1296
D. P. Upare et al.
Catalyst activity measurement Hydrogenolysis of LCO was carried out in a batch-type high-pressure 3/600 stainless-steel reactor placed inside an electric furnace equipped with a temperature controller. In a typical experiment, the reactor was charged with 0.5 g catalyst previously crushed and sieved to fine particles. Before starting the reaction, the catalyst was reduced in situ with H2 at 210 °C for 1 h. The optimum temperature for reduction was determined on the basis of the TPR analysis of the selected catalysts. LCO (15 g) was introduced into the reactor by use of a syringe. The reaction was carried at 300 °C under an H2 pressure of 28–30 bar for 1 h. The stirring rate was 300 rpm. The reactant-to-catalyst ratio was kept at a constant value of 30 (w/w), products were analyzed by GC and 13C NMR spectrometry. GC and
13
C NMR spectrometry
A Bruker Biospin (Billerica, USA) Avance500 500-MHz NMR spectrometer with wide-bore superconducting magnet and unmodified commercial 5 mm 13C probe was used. All data processing was performed with standard software. Quantitative spectra were recorded with a gated decoupled pulse sequence to ensure suppression of the nuclear Overhauser effect (NOE). Spin–lattice relaxation was leveled and a recycle time of 10 s between scans was found to enable quantitative measurement of samples. All chemical shifts were estimated in deuterated chloroform (CDCl3) with reference to tetramethylsilane (TMS) as internal standard. The instrument operating conditions were: spectral width 0–180 ppm, frequency 125.79 MHz, acquisition time 5 h. A GC (DS Chrom 6200, Donam Instruments, Korea) equipped with an FID was used. Separation was carried out on a fused silica capillary column (Alltech, USA; 30 m 9 0.2 mm 9 0.5 lm) using helium as a carrier gas. Conditions: initial temperature 100 °C, initial temperature hold time 5 min, heating rate 10 °C/min, final temperature 450 °C, final temperature hold time 10 min, and detector temperature 250 °C.
Results and discussion According to the mechanism of bifunctional catalysis, dehydrogenation, cracking, and isomerization can occur on acidic sites, whereas metal sites catalyze hydrogenation and hydrogenolysis. With the Ir/USY catalyst, acidic function is important. However, the strong acid sites in the USY would lead to severe hydrocracking of the reactants. For this reason, potassium was used as a basic component for tuning the acidic properties of the catalysts in this work. The modified catalysts were investigated previously [16] to find its effect on conversion and selectivity for ring opening products of MCP and MCH (methylcyclohexane) as model compounds. In this study, the best among the modifications was selected for hydrogenolysis of LCO in a batch-type reactor.
123
Upgrading of light cycle oil
1297
Catalyst characterization The catalysts were thoroughly characterized to gain insight into the relationship between catalyst surface properties and the selectivity for ring-opening products. Typical characteristics, viz. surface area, pore size, and total acidity of the catalysts were determined; the results are given in Table 1. The structures of the catalysts were confirmed by XRD pattern analysis (not shown here). The loading of K did not produce a new crystalline phase related to the potassium [17]. No distinct decrease in the relative crystallinity of the USY was observed. This means that K was highly dispersed on the USY. From the XRD patterns it can be concluded that introduction of K on USY did not cause a blocking of the pores, as shown in Table 1. The dependence of BET specific surface areas of the Ir/USY on K loadings is shown in Table 1. The specific surface areas of USY decrease somewhat with 0.9% (w/w) Ir loading. A decrease in surface area from 780 to 759 m2/g was observed. With a loading of 0.75% (w/w) K, surface area was increased to 811 m2/g. Only a slight decrease of pore volumes was observed, from 0.65 to 0.52 cm3/g for Ir/USY and from 0.52 to 0.51 cm3/g for K–Ir/USY. By taking into account the results of the BET surface area and pore volume, it can be concluded that the introduction of K on USY did not cause a blocking of the pores. Addition of potassium to the catalyst resulted in an increase in BET surface area and a decrease in pore volume. These results can be interpreted as follows: at 0.75% (w/w) potassium loading, potassium may be well dispersed on the surface of Ir-USY, and the adherence of potassium atoms to the metal precursor surface may increase the total surface area of the catalyst [18]. The activity of the catalysts increased with higher BET surface area. The results indicated that the USY zeolite was highly acidic in nature. The effect of K loading on the acid sites was evaluated by NH3-TPD; the results are shown in Fig. 1 and Table 1. It is well known that catalyst acidity decreases on addition of an alkali [17, 19]. We also found the same result, the acidity of the 0.9 Ir/USY catalyst decreases with addition of 0.75% (w/w) K. On loading with K, amounts of both weak and strong acid sites decreased obviously. The TPD profiles of adsorbed ammonia over different catalysts are shown in Fig. 1. Different desorption peaks were observed on the TPD profiles in different temperature ranges. Peaks shown at low and high temperatures can be attributed to ammonia desorbed from weak and strong acid sites, respectively [20]. The desorption peak at 150–200 °C represents Table 1 Textural and acidic properties of USY samples with 0.9% (w/w) Ir and 0.75% (w/w) K loadings Catalyst
BET surface area (m2/g)a
Pore volume (cm3/g)b
Total acidity (lmol/g)c
USY
780
0.65
487
0.9 Ir/USY
759
0.52
388
0.75 K/0.9 Ir/USY
811
0.51
183
a
Calculated by the BET method
b
Calculated from nitrogen adsorption branch at P/Po = 0.9
c
Amount of adsorbed NH3 measured by NH3-TPD; appeared in Fig. 1
123
1298
D. P. Upare et al.
TCD Signal (a.u.)
USY 0.9 wt.% Ir/USY 0.75wt%K/ 0.9wt%Ir/USY
100
200
300
400
500
600
Temp °C Fig. 1 NH3-TPD graph of USY, 0.9% (w/w) Ir/USY, and 0.75% (w/w) K/0.9% (w/w) Ir/USY catalysts
the weak acid sites, which may because of an octahedral Al ion formed by dealumination [21]. Impregnation with Ir results in a loss of acidic sites resulting from the interactions between iridium crystallites and acid sites. An interesting point to draw attention to is that impregnation with potassium leads to a significant loss in the number of acid sites, which may result from direct anchoring of K to proton sites. This demonstrates that K covered the acid sites of the USY catalyst [22]. To discover the effect of potassium addition on the reducibility of Ir/USY catalysts, we characterized these samples by hydrogen temperature-programmed reduction; the results are illustrated in Fig. 2. The Ir/USY catalysts furnished their H2 consumption peak at 190–210 °C; on addition of K the peak shifted to 180–200 °C; little shift in reduction temperature was found. The characteristic consumption peak at 180–210 °C with the K–Ir/USY catalysts is attributed to reduction of iridium oxide. Addition of potassium induced a shift of the main reduction peak of iridium oxide to lower temperature. However, H2 uptake by K–Ir/ USY was not much different than that of Ir/USY. H2 uptake implied the existence of reducible Ir species on the catalyst surface, which was active for hydrogenation reaction. Loadings of potassium markedly promote the reducibility of iridium oxide, probably because of a K–O–Ir interaction, as has been previously reported [23, 24]. From H2 TPR analysis, it can be concluded that metal particles are well dispersed on the support which could help upgrading of LCO. Hydrogenolysis of LCO The LCO used in this study boils in the range 108–428 °C. On the basis of the boiling point distribution 5% of the feed distills in the range IBP–180 °C, 85% between 180 and 340 °C, and the remainder 10% between 380 and 430 °C. The
123
Upgrading of light cycle oil
1299
0.9 wt.% Ir/USY
TCD Signals (a.u.)
0.75wt.%K/0.9wt.%Ir/USY
100
200
300
400
500
600
700
Temp °C Fig. 2 H2-TPR patterns of 0.9% (w/w) Ir/USY and 0.75% (w/w) K/0.9% (w/w) Ir/USY catalysts
material was produced in an HDS-FCC complex at the refinery using highly naphthenic vacuum gas oil. Its density was 0.9366 g/cm3. It contains high percentage of aromatics (72.3%) and sulfur (2,513 ppm). It is composed of 39% monoaromatics, 29% diaromatics, and 4.3% by weight polyaromatics. It has an extremely low cetane number (CN), i.e., 18. The low CN of LCO could be because of the presence of di and polyaromatics, which combust poorly in the diesel engine. This is one of the fundamental causes of the low blending quality of LCO. Typical characteristics of the feed are given in Table 2. LCO can be used as a diesel component if most of the aromatics are converted into paraffins. On the basis of the preliminary experimental results obtained on MCP as a model compound, 0.75 K/0.9 Ir/USY was selected for treatment of LCO [16]. Thus, hydrogenolysis of LCO over 0.9% Ir/USY catalyst doped with 0.75% K was tried to achieve partial aromatic ring saturation followed by SRO of naphthenes to produce paraffins of the same carbon numbers. GC and
13
C NMR spectrometry
Feed and product analysis was carried out by 13C NMR spectrometry. Quantitative spectra were obtained in Fourier-transform (FT) mode operating at a carbon frequency of 125.79 MHz. The 13C NMR spectra of LCO feed and LCO treated with 0.75% (w/w) K/0.9% (w/w) Ir/USY are shown in Fig. 3. The aromatic carbon (110–160 ppm) and the saturated carbon (1.5–40 ppm) regions were integrated quantitatively by suppressing NOE. From the integral data, structural data, for example aromatic carbon (Car), saturated carbon (Csat), and aromaticity (fa) of the feed and the product were determined. UV–visible tests provided an explanation of
123
1300
D. P. Upare et al.
Table 2 Characteristics of LCO before and after treatment Characteristics
LCOa
Treated LCO
Density (g/cm3)
0.9366
0.8895
GC R.T.b range (min)
10–20
1.5–18
Cetane number
18
27
Mono (MAH)
39.00
45.39
Di (DAH)
29.00
16.60
Poly (PAH)
04.30
0.39
Total (TAH)
72.30
62.38
Aliphatics (%)
27.70
37.62
Aromatics (%)c
a
Boiling range 108–428 °C, total sulfur 2,513 ppm, total nitrogen 574 ppm
b
GC R.T., gas chromatography retention time
c
MAH, monocyclic aromatic hydrocarbons; DAH, dicyclic aromatics hydrocarbons; PAH, polycyclic aromatic hydrocarbons; TAH, total aromatic hydrocarbons
the chemical composition of the LCO [25]. According to the literature [26], the onset wavelengths of the absorption peaks of benzene, naphthalene, anthracene, tetracene, and pentacene are near 270, 320, 380, 470, and 580 nm, respectively. The UV method compares sample absorbance at selected wavelengths with reference spectra of solutions of aromatics composed of representative compounds in the diesel boiling range. Because the absorbance is proportional to the aromatic rings, weight percent aromatic carbon. The aromaticity (fa) was found to decrease from 0.60 to 0.50 on treating the feed with 0.75 K/0.9 Ir/USY catalyst. Simultaneously, an increase of 10.23% of paraffinic carbons was also observed. The two major features of these spectra, i.e. the decrease of aromatic carbon and the increase of saturated carbon, indicate an increase of CN of the treated LCO. Many empirical relationships between the aromaticity and CN of diesel fuels have been reported elsewhere [27, 28]. On the basis of such correlations, the increase in CN of LCO was determined from the change in aromaticity (fa). It was found to increase by 9 units on hydrogenolysis with 0.75% (w/w) K/0.9% (w/w) Ir/USY catalyst. Figure 4 shows typical GC chromatograms obtained before and after treatment of LCO by 0.75% (w/w) K/0.9% (w/w) Ir/USY. It could be clearly seen that the products were shifted toward the low boiling range. For LCO the products were distributed over the retention time 10–20 min whereas for the treated LCO the product range was between 1.5 and 18 min. The change in aromatic product distribution was significant. The monoaromatics were increased to 45.39% and there was a significant reduction in diaromatic and polyaromatic compounds. These were determined to be 16.60 and 0.39%, respectively. The increase in monoaromatics could be explained in terms of ASAT followed by ring opening of di and polyaromatics. Further, the increase of products in the range of 1.5–10 min could be
123
Upgrading of light cycle oil
Fig. 3
1301
13
C NMR spectra of a LCO and b treated LCO with 0.75% (w/w) K/0.9% (w/w) Ir/USY
because of the formation of paraffinic and naphthenic compounds. The increase in their content was 9.92%, as determined by GC. These results were found to be in good agreement with those obtained by 13C NMR spectrometry. The density of the product was also determined and found to be reduced to 0.8895 g/cm3. The results are given in Table 2. These results clearly indicate that 0.75 K/0.9 Ir/USY catalyst was effective in reducing the aromaticity and increasing the paraffinic nature of LCO.
Conclusions Upgrading of LCO by partial hydrogenation and SRO on 0.9% (w/w) Ir/USY (Si-toAl ratio = 40) were studied in a batch-type reactor at T = 300 °C, P = 28–30 bar, and Time = 1 h. The catalysts were prepared by the impregnation method and their surface characteristics were evaluated by BET, XRD, NH3-TPD, and H2-TPR.
123
1302
D. P. Upare et al.
a
0
5
10
15
20
25
20
25
Time(min)
b
0
5
10
15
Time(min) Fig. 4 GC chromatograms obtained from a LCO and b LCO treated with 0.75% (w/w) K/0.9% (w/w) Ir/USY
Addition of an appropriate amount of potassium was found to be an effective way of optimizing the acid properties of the catalyst; reducing the number of strongly acid sites reduced its cracking activity. The results show that the 0.75% (w/w) K/0.9% (w/w) Ir/USY zeolite catalyst was effective in reducing the aromaticity and increasing the paraffinic nature of LCO. The catalyst was found to be effective in increasing its CN by 9 units thus increasing its blending ratio with the diesel pool. Acknowledgments This study was supported by the Creative Research Program of KRICT in 2011. Dr Rao acknowledges the financial support of the Brain Pool Program sponsored by Korean Federation of Science and Technology (KOFST), Seoul, Korea.
123
Upgrading of light cycle oil
1303
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.
T.C. Kaufmann, A. Kaldor, G.F. Stuntz, M.C. Kerby, L.L. Ansell, Catal. Today 62, 77 (2000) A. Nishijima, T. Kameoka, T. Sato, N. Matsubayshiand, Y. Nishimura, Catal. Today 45, 261 (1998) J. Kjarstad, F. Johnsson, Energy Policy 37, 441 (2009) J. Chang, Y. Fu, Y. Shibata, M. Yoshimoto, K. Fugimoto, N. Tsubaki, Fuel 84, 1661 (2005) V.P. Thakkar, S. Fabdo, V.A. Gembicki, J.F. Mcgehee, Proc. NPRA, annual meeting (San-Francisco, 2005). UOP LLC, Des Plaines, Illinois, USA, AM-05-53, pp. 13–15 A. Corma, V.G. Alfaro, A.V. Orchillies, J. Catal. 200, 34 (2001) S. Zhang, D. Liu, W. Deng, G. Que, Energy Fuels 21, 3057 (2009) R. Yang, X. Li, J. Wu, X. Zang, Z. Zang, Korean J. Chem. Eng. 27, 55 (2010) P.T. Do, W.E. Alvarez, D.E. Resasco, J. Catal. 238, 477 (2006) Mc.G.B. Vicker, M. Daage, M.S. Touvelle, C.W. Hudson, D.P. Klein, W.C. Baird, B.R. Cook Jr, G. Chen, S. Hantzer, D. Vaughan, E.S. Ellis, O.C. Feeley, J. Catal. 210, 137 (2002) S. Rossini, Catal. Today 77, 467 (2003) R.C. Santana, P.T. Do, M. Santikunaporn, W.E. Alvarez, J.D. Taylor, E.L. Sughrue, D.E. Resasco, Fuel 85, 643 (2006) M. Santikunaporn, W.E. Alvarez, D.E. Resasco, Appl. Catal. A 325, 175 (2007) S. Dokjampa, T. Riraksomboon, D.T. Phongand, E. Resasco, J. Mol. Catal. A 274, 231 (2007) J. Juan, M.C. Roman-Martinez, M.J. Illa´n-Go´mez, Appl. Catal. A 264, 169 (2004) R.N. Rao, N. You, D.P. Upare, S.H Yoon, C.W. Lee, Catal. Lett. (2011). doi: 10.1007/s10562-011-0573-7 Y.H. Park, G.L. Price, Ind. Eng. Chem. Res. 31, 2 (1992) S.T. Hussain, G. Hasan, Open Catal. J. 4, 36 (2011) K.B. Samson, R. Grabowski, M.A. Banares, E. LozanoDiz, Pol. J. Chem. 83, 1977 (2009) F. Lonyi, J. Valyon, Thermochim. Acta 373, 57 (2001) R. Nakao, Y. Kubota, N. Katada, N. Nishiyama, K. Kunimori, K. Tomishige, Appl. Catal. A 273, 63 (2004) H. Long, X. Wang, W. Sun, G. Xiong, W. Kunyuan, Fuel 87, 3660 (2008) S.L.G. Cortes, S. Ddorkjampa, P.T. Du, Z. Li, J.R. Lopez, F.G. Requejo, Chem. Eng. J. 139, 147 (2008) L. Feng, X. Li, D.B. Dadyburjor, E.L. Kugler, J. Catal. 190, 1 (2000) S. Yoon, S.D. Bhatt, W. Lee, H. Lee, S.Y. Jeong, J.O. Baeg, C.W. Lee, Korean J. Chem. Eng. 26, 71 (2009) O. Abbas, C. Rebufa, N. Dupuy, A. Permanyer, J. Kister, D.A. Azevedo, Fuel 85, 2653 (2006) T.H. Defries, D. Indritz, R.V. Kastrup, Ind. Eng. Chem. Res 26, 189 (1987) D.J. Cookson, P. Lliopoulos, B.E. Smith, Fuel 74, 70 (1995)
123