Conversion of canola oil and canola oil methyl ester (CME) to green aromatics over a HZSM-5 catalyst: a comparative study

Abstract

The catalytic conversions of canola oil and canola oil methyl ester (CME) for the production of green aromatics over a HZSM-5 catalyst were investigated. General Factorial Design (GFD) of experiments was applied in order to evaluate the aromatic production statistically. The influence of reaction conditions such as reaction temperature and Weight Hourly Space Velocity (WHSV) on the yield of the aromatic products was studied in the experiments. The reaction temperature was set at 375, 400, 450 or 500 °C and the space velocity was selected to be either 2 or 4 h⁻¹. The products comprised liquid hydrocarbon product (LHP), exhaust gases and water for both canola oil and CME. Moreover, thermal cracking of the CME for the production of aromatics was conducted at 450 and 500 °C to compare the results with the corresponding catalytic route. The LHPs were analyzed using Gas Chromatography (GC) to determine the BTX yields. Temperature, space velocity and feed type were found to be significant parameters for the production of aromatics. A comparison of CME and canola oil identified that the catalytic cracking of CME leads to higher aromatic production. Catalytic conversion of CME as well as canola oil yielded toluene as the major aromatic compound followed by para–meta xylene and benzene. Thermal cracking of the CME yielded only minor amounts of aromatic products and could not compete with the catalytic route for aromatic production.

---

1. Introduction

Benzene, toluene, and xylenes (BTX) are among the very important aromatic products of the chemical and petrochemical industries. These aromatics are produced during different processes. However, the highest BTX production is achieved via catalytic reforming of naphtha which is a petroleum derivative.¹ Due to the depletion of conventional fossil energy resources such as petroleum, considerable attention has been devoted to alternative renewable fuels that are environmentally friendly.^2–4 In this respect, vegetable oils have attracted much attention. Among the various methods for the conversion of vegetable oils to biofuels, catalytic cracking has been investigated by many researchers.^5–8 Canola oil (a vegetable oil) has been converted to various hydrocarbons including aromatics over different catalysts.^9,10

On the other hand, Fatty Acid Methyl Esters (FAMEs) obtained via the transesterification reaction of vegetable oils (triacylglycerol) with methanol have attracted remarkable attention due to the different advantages they possess. The major benefits of FAMEs are renewability, environmentally benign nature, lubricity function, high cetane number, high flash point and biodegradability.¹¹ One of the major disadvantages of FAMEs is the existence of oxygen in their molecular structure, which makes them thermally and chemically unstable.¹² In order to overcome this drawback, the deoxygenation of oxygen-containing molecules has been examined using zeolite catalysts at mild operating pressures.^12–14

HZSM-5 catalysts have been applied for the conversion of alcohols to hydrocarbons with a high amount of aromatics.^15,16 Also, some research work has been allocated to the conversion of oxygenate compounds like aldehydes, ketones and acids to hydrocarbons comprising olefins and aromatics over HZSM-5.¹⁷

The conversion of methyl octanoate, a short chain methyl ester, to hydrocarbons has been studied in the presence of HZSM-5.¹⁸ The reaction of methyl octanoate over HZSM-5 catalysts led to the production of various hydrocarbons containing high amounts of aromatics.

In summary, although the catalytic conversion of other types of oxygenates in the presence of HZSM-5 has been extensively reported in the literature, there have been few studies on the conversion of fatty acid methyl esters (FAMEs). Hence, in this research work, the conversion of canola oil and Canola Methyl Ester (CME) for the production of aromatics was studied and the results were discussed. Due to the previous reported studies in the literature regarding the possibility of aromatic production from the thermal cracking of canola oil,¹⁹ the thermal cracking of CME for aromatic production was conducted at two temperatures and the results were compared to the catalytic route.

2. Experimental

2.1. Materials

The ZSM-5 catalyst (CBV-5524G, SiO₂/Al₂O₃ = 50, NH₄ form) was supplied by the Zeolyst Company. Prior to use, the as-received zeolite was calcined in air at 550 °C for 5 hours to convert it to its hydrogen form. Methanol (99.8% purity) and potassium hydroxide (84% purity) were obtained from the Merck Company, Germany. Canola oil was obtained from a local market and utilized in the experiments without any purification. The fatty acid composition of the canola oil as determined using gas chromatography is presented in Table 1. CME was produced by conducting a transesterification reaction between the canola oil and methanol in the presence of a potassium hydroxide catalyst. The temperature of the reaction, molar ratio of methanol to oil and catalyst content were found by preliminary experiments to be 60 °C, 6 and 0.6 weight percent of oil, respectively.

Table 1 Canola oil fatty acid composition

Fatty acid	Structure	Content (wt%)
Palmitic	16 : 0	4
Stearic	18 : 0	4
Oleic	18 : 1	62
Linoleic	18 : 2	20
Linolenic	18 : 3	10

---

2.2. Catalyst characterization

The surface morphology of the HZSM-5 catalyst depicting the size of the crystals was characterized by Scanning Electron Microscopy (SEM) (KYKY-EM3200). The particle size distribution of the catalyst was measured using a Laser Fritsch Particle Sizer Analysette 22. A powder XRD (X-ray diffraction) pattern of the catalyst was recorded on a STOE Stadi MP diffractometer using Cu as an anode material at 40 kV and 30 mA (λ = 1.54 Å, scanning rate of 0.015° min⁻¹ in the 2θ range of 5–50°). The acidity of the catalyst was measured by the method of NH₃-TPD (Temperature-Programmed Desorption of Ammonia).

2.3. Catalyst activity

The catalytic conversions of canola oil or CME to hydrocarbons were conducted in a continuous flow fixed-bed tubular reactor made of quartz with an inside diameter of 2 cm. The catalyst was placed over a thin layer of glass wool supported in a stainless steel grid centrally positioned within the reactor. The process flow diagram of the experimental setup is illustrated in Fig. 1. In order to study the influence of temperature and WHSV on the aromatic content of the liquid hydrocarbon product (LHP) and also the aromatic yield, the reactor was operated in the temperature range of 375–500 °C and with a WHSV of 2 or 4. All experiments were performed under atmospheric pressure. Before the initiation of the reaction, the temperature of the reactor was raised to 375, 400, 450 or 500 °C. Then the reactor was purged using nitrogen gas to remove the trapped oxygen inside the reactor. Thereafter, the canola oil or CME was introduced to the reactor at a predetermined space velocity and, after vaporization in the primary section of the reactor, reacted over the catalyst bed. The outlet product vapor passed through water-cooled and ice-cooled condensers consecutively to condense the vapor as much as possible. The products mainly comprised water, liquid hydrocarbon products (LHPs) and exhaust gases. The LHPs obtained by the catalytic cracking of canola oil and CME were called LHPO and LHPM, respectively. Thermal cracking of the CME was also performed using the aforementioned experimental setup. The thermal cracking experiments were conducted at 450 and 500 °C and 2 h⁻¹ WHSV. Instead of catalyst particles, the center of the tubular reactor was filled with inert pellets to increase the surface area for the thermal cracking reactions.

Fig. 1 Process flow diagram of the green aromatic production.

2.4. Design of experiments

General Factorial Design (GFD) was applied to study the influence of temperature and space velocity on the aromatic yield. When using GFD, it’s possible to have parameters with different numbers of levels. The GFD includes all possible combinations of the factor levels. Additionally, one of the major applications of GFD is the possibility of analyzing mixtures of categorical and numeric factors simultaneously. In this research, Design Expert software (version 7.1.3) was employed to design the experiments and analyze the results statistically. Temperature and WHSV were selected as independent numeric parameters. The reactant type, namely CME or canola oil, was chosen as a categorical factor. The total aromatic content measured from the LHP analysis was the response factor. Table 2 shows the levels of the variables investigated in this study.

Table 2 Levels of the selected variables

Factor	Unit	Low level	Intermediate levels		High levels
Reaction temperature	°C	375	400	450	500
WHSV	h^−1	2 and 4
Feed type	Categorical factor	Canola oil and CME

---

2.5. Analysis of the products

In order to identify the amount of aromatic hydrocarbons in the LHP, a Varian CP-3800 gas chromatograph equipped with a Flame Ionization Detector (FID) was used. The GC column was a CP-Sil 13 CB with a 25 m length, 0.53 mm I.D. and 2 μm film thickness. The temperature of the column oven was programmed to maintain at 80 °C for 12 min, then increase to 250 °C at a rate of 30 °C min⁻¹ and hold at 250 °C for 13 minutes. Eqn (1) and (2) describe the LHP yield and aromatic yield consecutively:

LHP yield (%) = 100 × (mass of LHP (g))/(mass of the canola oil or CME feed (g)) (1)

Aromatic yield (wt%) = (aromatic content (wt%)) × (LHP yield (wt%)) (2)

Moreover, as the BTX boiling points are in the range of 70–160 °C, the LHP was distilled into three boiling ranges of less than 70, 70–160 and 160+ °C to roughly estimate the cut of the product which comprised BTX.²⁰

3. Results and discussion

3.1. Catalyst characterization

The SEM image of the ZSM-5 catalyst is depicted in Fig. 2. The electron micrographs showed that the catalyst particles were in the micrometer scale.

Fig. 2 SEM image of the HZSM-5 catalyst used in the experiments.

The particle size distribution of the catalyst is shown in Fig. 3. Q₃(x) is the percentage of the complete sample volume filled with particles smaller than x μm while dQ₃(x) corresponds to the volume percentage of the sample particles with diameters between x and y μm. As displayed, almost 80% of the particle sizes are in the range of 0.1–0.5 μm. The arithmetic mean diameter of the particles, as measured by the particle size analysis method, is 1.17 μm.

Fig. 3 Particle size distribution of the HZSM-5 catalyst.

The XRD pattern of the catalyst as shown in Fig. 4 comprises peaks in the 2θ ranges of 7–10 and 20–25° in agreement with the previous studies in the literature.^21,22

Fig. 4 XRD pattern of the ZSM-5.

Fig. 5 shows the acid site distribution of the ZSM-5 catalyst as determined by the TPD of ammonia. Three desorption peaks are indicated at 230, 428 and 659 °C, which can be attributed to the NH₃ desorbed from the low, medium and high strength acid sites, respectively. The total acid sites of the ZSM-5 as measured by TPD was 1.14 mmol NH₃ per g.

Fig. 5 NH₃-TPD profile of the ZSM-5 catalyst.

3.2. Design of experiments (DOE) and analysis of variance (ANOVA)

The experimental design layout as well as the responses are presented in Table 3. In order to fit the experimental points, a modified quadratic model was suggested. Estimation models of aromatic yields for canola oil and CME in terms of actual factors are indicated respectively by eqn (3) and (4).

Aromatic yield = −185.90594 + 1.16001 × temperature − 31.24879 × WHSV + 0.057041 × temperature × WHSV − 1.44554 × 10⁻³ × temperature² (3)

Aromatic yield = −190.25344 + 1.16001 × temperature − 27.35129 × WHSV + 0.057041 × temperature × WHSV − 1.44554 × 10⁻³ × temperature² (4)

Table 3 Experimental design structure and the related responses

Run	A: feed type	B: temperature (°C)	C: WHSV (h^−1)	Total aromatic yield (wt%)
1	CME	375	4	16.5
2	CME	450	4	32.1
3	CME	450	2	34.9
4	CME	400	4	29.4
5	Canola oil	500	4	24.5
6	Canola oil	450	2	31.0
7	Canola oil	450	4	22.7
8	CME	500	2	30.7
9	CME	500	4	29.4
10	Canola oil	375	2	25.3
11	CME	400	2	32.8
12	Canola oil	400	2	30.8
13	Canola oil	375	4	6.6
14	Canola oil	400	4	8.5
15	CME	375	2	31.0
16	Canola oil	500	2	28.4

---

The Design Expert software was used to analyze the suggested model statistically, evaluate the ANOVA and check the suitability of the model as illustrated in Table 4.

Table 4 Analysis of variance for the model

Source	P-value, prob. > F
Model	<0.0001, significant
A	0.0008, significant
B	0.0020, significant
C	0.0003, significant
AC	0.0272, significant
BC	0.0049, significant
B^2	0.0113, significant
R^2	0.9265
Adjusted R^2	0.8776
Predicted R^2	0.7697
Adequate precision	14.956

---

According to Table 4, the model p-value was less than 0.0001, indicating high significance in predicting the response values. AB and C² were insignificant model terms that were removed from the model. The other model terms as shown in Table 4 are significant.

The R-squared, which is also known as the coefficient of determination, is the measure of how close the response values are to the fitted model equations. The R-squared value was calculated to be 0.9265 for the response. This value implies that 92.65% of the experimental data were consistent with the model predicted data.

The “Predicted R-Squared” of 0.7697 was close to the “Adjusted R-Squared” of 0.8776. The difference between these values should be lower than 0.2 to conclude that there is no problem with the model or the data.

An “Adequate Precision” of more than 4 is acceptable. Therefore, the value of 14.956 shows an adequate signal.

Fig. 6 shows the predicted values of total aromatic yields (by the model eqn (3) and (4)) versus the actual experimental data. As illustrated, the predicted responses are in good agreement with the observed experimental values.

Fig. 6 Predicted total aromatic yields vs. actual values.

3.3. Catalytic activity for aromatic hydrocarbon production

3.3.1. The effect of the reaction conditions on the LHP yield. The yields of the LHPs (LHPO and LHPM) under different operating conditions are illustrated in Fig. 7. Elevating the temperature causes more cracking reactions, leading to the production of more light gases. Hence, as expected, with rises in the temperature, both the LHPO and LHPM yields decreased.

Fig. 7 The LHPO and LHPM yields under different operating conditions.

The higher the space velocity, the lower the residence time of the hydrocarbons in the reaction medium. So, due to less residence time, lower light gases formed and, consequently, the yields reasonably increased.

In comparison with the CME, because of the fact that canola oil (triacylglycerol molecules) has heavier molecules, its cracking needs more severe operating conditions. Therefore, the yield of LHPO was higher than the LHPM yield under the same temperature and space velocity conditions.

3.3.2. The effect of the reaction conditions on the distillation cuts of the LHP. Fig. 8 depicts the comparison between the LHPO and LHPM distillation cuts for different operating conditions. As illustrated, the weight fraction of the cut with a boiling point range of 70–160 °C increased with increasing temperature and/or reduction in the space velocity whereas the heavier “160+ cut” and “below 70 cut” weight fractions were decreased. This phenomenon was ascribed to more cracking reactions occurring at higher temperatures and/or lower space velocities leading to lighter hydrocarbons with lower boiling points. In contrast with the LHPO, the weight fraction of the 70–160 °C boiling cut of the LHPM was higher and the portion of its heavier components (160+ °C boiling point cut) was lower. It can be explained such that the lighter the molecules of the inlet feed to the reactor, the higher the light fraction of the LHP.

Fig. 8 Comparison of the LHPO and LHPM distillation cuts (I–L).

3.3.3. The effect of the reaction conditions on the aromatic production. Fig. 9 depicts the total aromatic contents of the LHPO and LHPM at different reaction conditions as determined by GC analysis. According to the figure, the total aromatic content was increased with increasing temperature and/or lowering of the space velocity. As mentioned in the literature,⁹ the required reactions for oxygenate (methyl ester or vegetable oil) conversion include thermal-catalytic cracking and deoxygenation. Complementary reactions that convert the produced hydrocarbons to aromatics comprise oligomerization, acid catalyzed cyclization and H-transfer. Cracking and deoxygenation occur on Bronsted and Lewis acid sites. The HZSM-5 catalyst contains acid sites to conduct the above reactions and also secondary cracking, olefin oligomerization, cyclization and H-transfer, which are essential steps to produce aromatic hydrocarbons.^9,23 A different path that has been presented for methyl ester aromatization includes intermediate compounds produced during the conversion of oxygenates. These intermediates represent strong interactions with the catalysts.¹⁸ The aforementioned intermediates, due to their low mobility, remain on the catalyst surface long enough to be converted to aromatics especially when the space velocity is 2 h⁻¹. Additionally, the lower the space velocity, the higher the space time is for olefin oligomerization and cyclization reactions, leading to aromatic production. The total aromatic content of the LHPM was higher than the LHPO for the same operating conditions. The difference between the aromatic content of the LHPO and the LHPM was more significant when the space velocity was 4 h⁻¹. It can be explained that cracking is one of the major and primary steps for the aromatization of oxygenates. The higher aromatic content of the LHPM may be attributed to more cracking reactions of the CME because it consists of lighter molecules. The total aromatic content increased with increasing temperature and reached its maximum at approximately 450 °C. This trend was true for all operating conditions except when the space velocity of canola oil was 4 h⁻¹. In this case, the maximum total aromatic content of the LHPO was obtained at 500 °C. This may be due to increased cracking reactions of the higher molecular weight canola oil at elevated temperatures that enhanced the aromatization of the oil.

Fig. 9 Total aromatic contents of the LHPO and LHPM under different reaction conditions.

Besides, total aromatic yields were calculated using eqn (2). Fig. 10 demonstrates the total aromatic yields under different operating conditions. According to this figure, the total aromatic yields were increased with increasing of the temperature to its maximum value and then decreased. Further cracking of cracked methyl esters to light gases at higher temperatures may cause the diminishing of the total aromatic yield. The exception of this trend is the case of the LHPO at 4 h⁻¹ space velocity. The reason of this phenomenon may be increased cracking of the oil and enhancement of aromatization at elevated temperatures.

Fig. 10 Total aromatic yields under different reaction conditions.

The effect of interaction between reaction temperature and space velocity for both canola oil and CME is depicted in 3D plots of Fig. 11. As displayed, the total aromatic yield was increased with increasing temperature and decreasing space velocity. Increasing the temperature further than 450 °C caused the total aromatic yield to decrease. Therefore, operation at such high temperatures is not reasonable.

Fig. 11 3D plots of the predicted aromatic yield versus temperature and WHSV for canola oil (I) and CME (J).

Yields of the aromatic hydrocarbon products are provided in Table 5. BTX, which are critical aromatics, had relatively high yields as shown in the table. Also, both feeds had a similar distribution of aromatic products. The affinity for the production of C7–C9 aromatics was related to the shape selectivity of the HZSM-5 catalyst.²³ Toluene was the main aromatic hydrocarbon produced followed by para–meta xylene, benzene and ortho-xylene.

Table 5 Yields of aromatic hydrocarbons at different operating conditions

Feed	WHSV	Temperature	Benzene	Toluene	Para–meta-xylene	Ortho-xylene	C9 aromatics
Canola oil	2	375	2.27	7.89	6.21	1.66	7.27
		400	3.26	10.95	7.75	2.23	7.71
		450	3.76	11.51	8.03	2.46	5.31
		500	4.33	11.93	6.98	2.23	2.99
	4	375	0.57	1.72	1.57	0.37	2.39
		400	0.67	2.19	2.04	0.49	3.14
		450	1.97	6.6	5.98	1.6	6.62
		500	2.65	8.67	6.84	2.06	4.31
CME	2	375	2.34	9.43	8.61	2.62	8.02
		400	2.97	10.26	8.39	3.02	8.18
		450	3.48	11.77	9.82	3.24	6.53
		500	4.31	12.56	7.92	2.57	3.31
	4	375	1.12	4.1	4.25	1.04	6.03
		400	3.72	10.54	7.27	1.92	5.86
		450	4.73	12.05	8.07	2.29	5.02
		500	4.35	11.64	7.54	2.41	3.36

---

3.3.4. Thermal cracking of CME for aromatic production. The CME thermal cracking results are presented in Table 6. As shown, increasing the temperature caused more aromatic production. However, the aromatic yield through thermal cracking was very low in comparison with the catalytic cracking route. In conclusion, thermal cracking of CME alone is not adequate for aromatic production and the presence of a catalyst to enhance the yield is essential.

Table 6 Total aromatic content and yield of CME thermal cracking

Temperature (°C)	LHP (wt%)	Total aromatic content (wt%)	Total aromatic yield (wt%)
450	93.6	0.7	0.66
500	74.9	2.3	1.7

---

4. Conclusions

The conversions of canola oil and CME to aromatic rich hydrocarbons were conducted in the presence of a HZSM-5 catalyst. The influences of the reaction temperature and space velocity on the production of aromatics from canola oil and CME were investigated using General Factorial Design (GFD). The reaction temperature, space velocity and feed type were found to be significant parameters for aromatic production. In comparison with canola oil, the catalytic cracking of CME yielded a greater amount of aromatics. The difference between the aromatic yields of canola oil and CME was more pronounced at higher space velocities. The maximum aromatic yield was obtained at the temperature of 450 °C and space velocity of 2 h⁻¹. Increasing the temperature further led to a loss of the aromatic yield for both feeds. The distribution of aromatic products was similar for both canola oil and CME. The main aromatic product of the process was toluene followed by para–meta xylene and benzene. Thermal cracking of the CME yielded minor amounts of aromatics that could not compete with the catalytic route for aromatic production.

Acknowledgements

This contribution was based on research supported by a grant from the Tarbiat Modares University.

References

1. M. R. Rahimpour, M. Jafari and D. Iranshahi, Appl. Energy, 2013, 109, 79–93.
2. P. S. Rezaei, H. Shafaghat and W. M. A. W. Daud, Appl. Catal., A, 2014, 469, 490–511.
3. L. Wang, H. Lei, Q. Bu, S. Ren, Y. Wei, L. Zhu, X. Zhang, Y. Liu, G. Yadavalli, J. Lee, S. Chen and J. Tang, Fuel, 2014, 129, 78–85.
4. B. Zhang, Z. Zhong, K. Ding and Z. Song, Fuel, 2015, 139, 622–628.
5. P. Tamunaidu and S. Bhatia, Bioresour. Technol., 2007, 98, 3593–3601.
6. C. M. R. Prado and F. N. R. Antoniosi, J. Anal. Appl. Pyrolysis, 2008, 86, 338–347.
7. H. Li, P. Yu and B. Shen, Fuel Process. Technol., 2009, 90, 1087–1092.
8. H. Li, B. Shen, J. C. Kabalu and M. Nchare, Renewable Energy, 2009, 34, 1033–1039.
9. S. P. R. Katikaneni, J. D. Adjaye and N. N. Bakhshi, Can. J. Chem. Eng., 1995, 73, 484–497.
10. S. M. Sadrameli, A. E. S. Green and W. Seames, J. Anal. Appl. Pyrolysis, 2009, 86, 1–7.
11. N. Pasupulety, K. Gunda, Y. Liu, G. L. Rempel and F. T. T. Ng, Appl. Catal., A, 2013, 452, 189–202.
12. T. Sooknoi, T. Danuthai, L. L. Lobban, R. G. Mallinson and D. E. Resasco, J. Catal., 2008, 258, 199–209.
13. R. Cerny, M. Kubu and D. Kubicka, Catal. Today, 2013, 204, 46–53.
14. T. Danuthai, T. Sooknoi, S. Jongpatiwut, T. Rirksomboon, S. Osuwan and D. E. Resasco, Appl. Catal., A, 2011, 409–410, 74–81.
15. J. D. Adjaye and N. N. Bakhshi, Fuel Process. Technol., 1995, 45, 161–183.
16. A. A. Rownaghi, F. Rezaei and J. Hedlund, Microporous Mesoporous Mater., 2012, 151, 26–33.
17. A. G. Gayubo, A. T. Aguayo, A. Atutxa, R. Aguado, M. Olazar and J. Bilbao, Ind. Eng. Chem. Res., 2004, 43, 2619–2626.
18. T. Danuthai, S. Jongpatiwut, T. Rirksomboon, S. Osuwan and D. E. Resasco, Appl. Catal., A, 2009, 361, 99–105.
19. O. I. Raphael, S. P. R. Katikaneni and N. N. Bakhshi, Energy Fuels, 1996, 10, 1150–1162.
20. R. C. Weast, M. J. Astle and W. H. Beyer, CRC Handbook of Chemistry and Physics, CRC Press, Florida, USA, 69th edn, 1988–1989.
21. Y. Cheng, L. J. Wang, J. S. Li, Y. C. Yang and X. Y. Sun, Mater. Lett., 2005, 59, 3427–3430.
22. M. Khatamian and M. Irani, J. Iran. Chem. Soc., 2009, 6, 187–194.
23. N. Y. Chen, W. E. Garwood and F. G. Dwyer, Shape Selective Catalysis in Industrial Applications, Marcel Dekker, New York, USA, 1989.

---