Origin of catalyst deactivation in atmospheric hydrogenolysis of m-cresol over Fe/HBeta

Abstract

Zeolites are the most common catalysts used for atmospheric deoxygenation of biomass pyrolysis derived feedstocks. The catalytic performance of the zeolite and the yield of deoxygenation greatly depend on the nature of the feedstock. Lignin is the most difficult part of biomass to be deoxygenated and lignin derived phenolic compounds cause rapid deactivation of zeolites. The main purpose of this research was to study the origin of zeolite deactivation in atmospheric deoxygenation of phenolic compounds. Phenol and m-cresol were selected as model compounds for lignin. In order to investigate their effect on zeolite deactivation, catalytic conversion of a mixture of methanol with m-cresol or phenol and a mixture of m-cresol with phenol were carried out over HBeta and Fe/HBeta, respectively. Co-feeding phenol or m-cresol with methanol caused high deactivation of HBeta and significant reduction in the aromatics yield. Meanwhile, these phenols had low reactivity over HBeta. Catalytic performance was enhanced by iron impregnation on zeolite, and Fe/HBeta could considerably convert m-cresol into aromatic hydrocarbons through hydrogenolysis. However, this catalyst was not efficient for deoxygenation of phenol. Strong adsorption of phenol molecules on zeolite acid sites resulting in high formation of coke was the main source of zeolite deactivation which was attenuated by an increase in reaction temperature.

---

1. Introduction

Considering the depletion of fossil fuels as well as the environmental threats caused by large-scale consumption of this source of energy, there is an increasing concern about the future of the hydrocarbon industry.^1–4 Lignocellulosic biomass seems to be a potential alternative to replace fossil fuels since it is renewable and abundantly available.^5,6 Pyrolysis is an effective technique for high exploitation of biomass and production of liquid fuel, called bio-oil.⁷ However, pyrolysis bio-oil is highly oxygenated due to the high oxygen content of biomass.^8–10 In the last two decades, several different catalytic reactions operated at atmospheric pressure or high hydrogen pressure have been examined in order to achieve an efficient process for deoxygenation of biomass derived feedstocks. Atmospheric pressure processes seem more attractive due to lower operating costs. However, it is more difficult to achieve a high yield of deoxygenation at atmospheric pressure mainly due to a low hydrogen to carbon effective ratio of biomass derived feedstocks and catalyst deactivation caused by high coke deposition. Zeolites are the most widely used catalysts in the atmospheric transformation of biomass derived feedstocks for production of valuable hydrocarbons like aromatics and olefins.⁵ The studies reported in the literature indicate that the nature of the feedstock greatly affects the catalytic performance of zeolites and yield of deoxygenation. Therefore, the composition of the biomass should be taken into account in order to design an efficient catalytic system. Ana G. Gayubo et al.^11,12 showed that alcohols, ketones and acids could be significantly converted to hydrocarbons through catalytic transformation over HZSM-5, while phenols and aldehydes had low reactivity to hydrocarbons and caused a large amount of coke deposited on this zeolite. As a result, it was suggested to separate phenols and aldehydes prior to catalytic upgrading of pyrolysis bio-oil. It has been indicated in several researches that lignin is the most difficult fraction of biomass to be converted to hydrocarbons.^13–15 This is due to the phenolic structure of lignin and low reactivity of lignin derived phenolic compounds over zeolites. In a study by Charles A. Mullen et al.,¹⁶ it was revealed that the lignin with higher content of p-hydroxyphenyl units and lower contents of guaiacyl and syringyl units caused higher deactivation of HZSM-5. The reason for this was declared to be higher concentration of simple phenolics compared to that of guaiacols and syringols obtained by pyrolysis of the lignin with higher proportion of p-hydroxyphenyl units; simple phenols have more potential than guaiacols and syringols to be tightly bound to the active sites of zeolites causing higher catalyst deactivation. In transformation of methylcyclohexane over HZSM-5, it was revealed that addition of phenol to methylcyclohexane significantly increased catalyst deactivation due to strong adsorption of phenol on zeolite acid sites.¹⁷ The loss of catalytic activity was also observed by addition of phenol to methylcyclohexane and n-heptane in the transformation of these compounds over HY zeolite.^18,19

As mentioned above, simple phenols derived from p-hydroxyphenyl units of lignin cause higher zeolite deactivation compared to phenolic compounds obtained from pyrolysis of guaiacyl and syringyl units. The main purpose of this work was to compare the deactivating effects of m-cresol and phenol which are derivatives of p-hydroxyphenyl units of lignin. Phenol and m-cresol were co-fed with methanol in order to show how catalytic performance of HBeta zeolite could be affected by simple phenols derived from p-hydroxyphenyl units of lignin. The reason for selection of methanol was that it has high potential to be converted into aromatic hydrocarbons over zeolite at atmospheric pressure.^20–22 Furthermore, the possibility of atmospheric conversion of phenolic compounds into aromatic hydrocarbons was studied over iron impregnated Beta as a modified zeolite.

2. Experimental

2.1. Chemicals

Phenol (C₆H₆O, ≥99%) and m-cresol (C₇H₈O, ≥98%) were purchased from Sigma-Aldrich. Methanol (CH₄O, ≥99.9%) was procured from Merck. Ethyl acetate was purchased from R&M Chemicals. The chemicals were used as received without further purification. Purified hydrogen and nitrogen were supplied from Linde Malaysia Sdn. Bhd.

2.2. Catalyst preparation

The catalysts used in this study were HBeta zeolite and 1 wt% Fe/HBeta. HBeta was obtained by calcination of the ammonium form of Beta zeolite (Zeolyst, CP814C, SiO₂/Al₂O₃ molar ratio: 38) at 550 °C (with heating rate of 3 °C min⁻¹) for 12 h. The iron impregnated HBeta was prepared by incipient wetness impregnation of HBeta with an aqueous solution of Fe(NO₃)₃·9H₂O (Aldrich) for 3 h and then was dried at 100 °C for 12 h, followed by calcination at 550 °C for 12 h.

2.3. Catalyst characterization

The crystalline phase of the zeolites was verified by X-ray diffraction (XRD) on a Rigaku Miniflex diffractometer using Cu Kα radiation (λ = 1.54443 Å) at 45 kV and 40 mA. The XRD patterns were collected in the 2θ range of 5–80° with a step size of 0.026° and scan rate of 0.05° s⁻¹.

The chemical composition of the catalysts was determined by X-ray fluorescence (XRF) instrument (PANalytical Axios^mAX).

The surface area and pore size distribution of the catalysts were measured by N₂ isothermal (−196 °C) adsorption–desorption using Micromeritics ASAP 2020 surface area and porosity analyzer. The samples were degassed at 180 °C under vacuum for 4 h prior to the analysis.

The acidity of catalysts was analyzed by temperature programmed desorption of ammonia (NH₃-TPD) using Micromeritics ChemiSorb 2720 instrument. 200 mg of each sample was set in TPD cell. In a stream of He gas (20 mL min⁻¹), the sample was heated from ambient temperature to 700 °C at a heating rate of 20 °C min⁻¹ and was held at 700 °C for 1 h. Afterward, the sample temperature was brought down to 210 °C and ammonia was introduced into the cell in a stream of 10% NH₃/90% He (20 mL min⁻¹) for 30 min. After being flushed with He gas for 30 min for elimination of physisorbed NH₃, the sample was cooled down to 70 °C. When the thermal conductivity detector (TCD) signal was stable, ammonia desorption measurement was performed by heating the sample to 600 °C with a rate of 10 °C min⁻¹ under He flow (20 mL min⁻¹).

The amount of coke deposited on catalysts was measured by thermogravimetric analysis using a PerkinElmer STA 6000 Simultaneous Thermal Analyzer. In the flow of synthetic air at 100 mL min⁻¹, samples were heated from 30 to 750 °C with the rate of 5 °C min⁻¹ and kept at final temperature for 30 min. The weight loss in temperature range of 300–750 °C was considered as the amount of coke deposited on catalyst, and the weight loss below 300 °C was assigned to desorption of water and volatile components.

2.4. Catalytic activity measurement

The catalytic reactions were conducted in a continuous, down-flow, fixed-bed tubular reactor (ID: 6 cm; height: 60 cm) made of stainless steel 316L which was heated by a two-zone furnace. All runs were carried out at atmospheric pressure. In each run, 5 g calcined catalyst was loaded in a stainless steel cylindrical cup (ID: 3 cm; height: 10 cm) with screen of 400 mesh at the bottom side. The cup was placed inside the reactor. The activity of catalysts was studied at reaction temperatures of 350 or 450 °C. The temperature of catalyst bed was measured by a K-type thermocouple inserted into the catalyst bed. Feed was pumped to the reactor using a syringe pump (Fisher, KDS100). The compounds fed to the reactor were methanol, m-cresol and mixtures of methanol/m-cresol (90 : 10 wt%), methanol/phenol (90 : 10 wt%) and m-cresol/phenol (90 : 10 wt%). In the case of using pure HBeta zeolite, catalyst bed was heated to reaction temperature and nitrogen was purged to the reactor at flow rate of 2 L min⁻¹ for 30 min. Afterward, 10 g feed was introduced to the reactor with weight hourly space velocity (WHSV) of 2 h⁻¹ and N₂ flow rate was kept at 2 L min⁻¹. When Fe/HBeta was used as catalyst, catalyst bed was heated to 300 °C in a stream of N₂ (2 L min⁻¹). At this temperature, gas was changed to H₂ (0.4 L min⁻¹) for in situ reduction of catalyst for 2 h. Subsequently, temperature was raised to 350 or 450 °C for reaction. Then, H₂ flow rate was increased to 2 L min⁻¹ and 10 g feed was injected with WHSV of 2 h⁻¹. The liquid products were collected by two condensers maintained at −10 °C. All lines were heated to avoid any condensation. After each run, the catalyst bed was exposed to N₂ flow (2 L min⁻¹) at the reaction temperature for 30 min in order to remove the components which might remain on the catalyst. Qualitative and quantitative analysis of liquid products was performed by GC/MS (Shimadzu QP 2010, DB-5 30 m × 0.25 mm × 0.25 μm), equipped with flame ionization and mass spectrometry detection. The GC oven temperature program was as follows: temperature was held at 50 °C for 5 min, ramped to 300 °C at 10 °C min⁻¹, and kept at 300 °C for 10 min. The injector temperature was 290 °C and a split ratio of 50 : 1 was employed. Helium was used as carrier gas with flow rate of 1.26 mL min⁻¹. Peak identification was done using the NIST (National Institute of Standards and Technology) mass spectrum library. The organic phase of product was separated and diluted with ethyl acetate before injection to GC/MS.

3. Results and discussion

3.1. Physicochemical characteristics of catalysts

The XRD data show that there is no distinct difference in crystallinity between HBeta and Fe/HBeta indicating that crystalline structure of HBeta was relatively unchanged by addition of Fe (Fig. 1). No iron species were detected in Fe/HBeta signifying that Fe is well dispersed on HBeta surface. HBeta is a zeolite with three-dimensional framework formed of 12-membered ring pores with dimensions of 0.66 × 0.67 and 0.56 × 0.56 nm which are suitable for diffusion of relatively large phenolic compounds.²³ Textural properties of fresh HBeta and Fe/HBeta as well as HBeta used in different reactant systems evaluated from nitrogen isothermal adsorption–desorption are presented in Table 1. The fresh catalysts are mostly microporous since the surface area and volume of micropores are larger than those of mesopores. BET surface area of 1 wt% Fe/HBeta was 6% lower than that of pure HBeta, and this reduction in surface area was only occurred in micropores. As shown in Fig. 2, both zeolites displayed type IV isotherm with H4-type hysteresis loop which indicates the predominance of microporous structure in these catalysts. The BJH desorption pore size distribution showed that most of the mesoporosity of both zeolites was in the range below 4 nm. The acidity of fresh HBeta and Fe/HBeta determined by NH₃-TPD analysis is depicted in Fig. 3. HBeta exhibited two ammonia desorption peaks at 232 and 328 °C, while desorption peak temperatures for Fe/HBeta were 248 and 322 °C. The peak area of iron incorporated HBeta was lower than that of HBeta demonstrating the reduction in the density of free acid sites as a result of Fe loading.

Fig. 1 X-ray diffraction patterns of HBeta and Fe/HBeta.

Table 1 Chemical and textural properties of HBeta and Fe/HBeta

Sample	SiO2/Al2O3^a	SBET^b (m^2 g^−1)	Smeso^c (m^2 g^−1)	SBET/Smeso	Vtotal^d (cm^3 g^−1)	Vmicro^e (cm^3 g^−1)	Vmeso^f (cm^3 g^−1)	d^g (nm)
a Determined by XRF analysis.b Calculated in the range of relative pressure (P/P0) = 0.05–0.25.c Evaluated by t-plot method.d Total pore volume evaluated at P/P0 = 0.99.e Evaluated by t-plot method.f Vmeso = Vtotal − Vmicro.g BJH adsorption average pore width.h Used HBeta (WHSV, 2 h^−1; time on stream, 60 min; carrier gas, N2).i MeOH: methanol.
HBeta	38.2	502	118	4.25	0.294	0.188	0.106	6.672
Fe/HBeta	40.5	471	123	3.83	0.287	0.169	0.118	6.815
HBeta^h (MeOH^i-450 °C)		497	120	4.14	0.291	0.177	0.114	6.586
HBeta^h (MeOH-350 °C)		494	122	4.05	0.282	0.173	0.109	6.611
HBeta^h (MeOH/cresol-450 °C)		465	131	3.55	0.258	0.136	0.122	6.457
HBeta^h (MeOH/phenol-450 °C)		436	128	3.41	0.243	0.111	0.132	6.348
HBeta^h (MeOH/cresol-350 °C)		397	134	2.96	0.223	0.096	0.127	6.112
HBeta^h (MeOH/phenol-350 °C)		355	137	2.59	0.211	0.073	0.138	5.819

---

Fig. 2 Nitrogen adsorption–desorption isotherms of HBeta and Fe/HBeta.

Fig. 3 NH₃-TPD profiles of HBeta and Fe/HBeta.

3.2. Catalytic activity

The yields and selectivities of the products obtained from catalytic reactions at 350 °C are presented in Table 2. HBeta zeolite was efficient for transformation of methanol to aromatic hydrocarbons, and aromatics yield of 59.6 wt% was achieved using this catalyst. The main aromatic hydrocarbons detected in liquid product were xylene, trimethylbenzene, ethyl-methylbenzene and tetramethylbenzene. When m-cresol was co-fed with methanol at low amount of 10 wt%, aromatics yield was 24.3 wt% which was 59% lower than that obtained from pure methanol. Cresol, phenol, xylenol, ethylphenol, trimethylphenol and some other oxygenate compounds were also detected in the liquid product from conversion of mixture of methanol/m-cresol over HBeta. Compered to m-cresol, addition of 10 wt% phenol to methanol caused much more negative effect on catalytic performance of HBeta, and the aromatics yield achieved from mixture of methanol/phenol was only 6.2 wt%. Furthermore, as shown in Table 3, the amount of coke deposited on HBeta in the conversion of methanol at 350 °C was 0.22 wt% which was increased to 2.61 and 3.48 wt% in the conversion of mixtures of methanol/m-cresol and methanol/phenol, respectively. The data in Table 3 are given by TGA results presented in Fig. S1 and S2 (ESI†). Meanwhile in transformation of pure m-cresol over HBeta under hydrogen atmosphere, low yield of aromatics (2.6 wt%) was obtained. Iron incorporated HBeta showed to be effective for deoxygenation of m-cresol through hydrogenolysis. In the transformation of m-cresol over Fe/HBeta, aromatics yield of 17.5 wt% was achieved. At temperature of 350 °C, WHSV of 2 h⁻¹ and under hydrogen atmosphere, 58.5 wt% of m-cresol was converted mostly to phenol, trimethylbenzene, xylene and ethyl-methylbenzene. However, when 10 wt% phenol was co-fed with m-cresol, cresol conversion and aromatics yield were reduced to 30.1 and 3.2 wt%, respectively. This clearly shows the significance of negative effect of phenol on catalytic performance of Fe/HBeta. In another experiment under similar reaction conditions, Fe/HBeta was used for conversion of pure phenol, but the yield of hydrocarbons detected in liquid product was very low (below 0.3 wt%).

Table 2 Product yields and selectivities (wt%) obtained from catalytic conversion of different reactants over HBeta and Fe/HBeta. Reaction conditions: WHSV, 2 h⁻¹; reaction temperature, 350 °C; pressure, 1 atm

Feed	MeOH^a	MeOH–cresol	MeOH–phenol	Cresol	Cresol	Cresol–phenol
Catalyst	HBeta	HBeta	HBeta	HBeta	Fe/HBeta	Fe/HBeta
Gas	N2	N2	N2	H2	H2	H2
a MeOH: methanol.
%Yield of organic phase
	59.62	32.58	13.96	87.29	83.34	81.50
%Selectivity in organic phase
Toluene				0.23	1.53	0.22
Xylene	41.68	32.47	19.91	0.64	5.00	0.85
Ethyl-methylbenzene	15.93	12.12	7.59	0.34	4.24	1.09
Trimethylbenzene	28.65	20.35	10.96	0.57	5.51	0.89
Tetramethylbenzene	6.81	3.68	1.72	0.18	1.19	0.16
Naphthalenes				0.30	1.31	0.20
Other hydrocarbons	6.93	5.89	3.94	0.70	2.21	0.47
Phenol		8.48	42.34	25.83	11.43	17.97
Cresol		13.57	8.88	60.72	49.82	73.48
Other oxygenates		3.44	4.66	10.49	17.76	4.67

---

Table 3 Coke deposition on HBeta and Fe/HBeta for different reactants at reaction temperatures of 350 and 450 °C. Reaction conditions: WHSV, 2 h⁻¹; time on stream, 60 min; pressure, 1 atm

Feed	MeOH^a	MeOH–cresol	MeOH–phenol	Cresol	Cresol–phenol
Catalyst	HBeta	HBeta	HBeta	Fe/HBeta	Fe/HBeta
Gas	N2	N2	N2	H2	H2
a MeOH: methanol.
%gcoke/gcatalyst
350 °C	0.22	2.61	3.48	9.33	11.86
450 °C	0.28	1.52	1.94	6.11	7.02

---

Methanol could be considered as representative for that part of biomass which has high potential to be deoxygenated at atmospheric pressure; methanol is easily transformed into aromatics with small amount of coke deposited on zeolite. It is well described in literature that lignin derived phenolic compounds have very low reactivity over zeolite acid sites. Using zeolite, phenolic compounds are converted to aromatic hydrocarbons through cracking to olefins and subsequent aromatization of intermediate olefins.¹⁶ In fact, aromatic hydrocarbons are not produced by direct cleavage of C–O bond of phenolics over zeolite acid sites. Charles A. Mullen et al.¹⁶ revealed that phenols derived from p-hydroxyphenyl units of lignin are more difficult than other lignin derived phenols like guaiacols and syringols to be reacted over zeolite, and cause higher deactivation of catalyst. The reason for higher reactivity of guaiacols and syringols was mentioned to be the steric hindrance caused by the methoxy groups on the benzene ring of these compounds which prevents from tight bond with zeolite acid sites leading to less coke formation and catalyst deactivation. As it was observed in this work, when 10 wt% m-cresol or phenol which are derivatives of p-hydroxyphenyl units of lignin was co-fed with methanol, aromatics yield of HBeta zeolite had a significant decrease. TPD profiles of fresh HBeta and HBeta used in different reactant systems shown in Fig. 4 demonstrate that methanol transformation caused a slight reduction in the number of free acid sites. However TPD results for HBeta exposed to mixtures of methanol/m-cresol and methanol/phenol reveal that the number of zeolite acid sites occupied by catalytic coke significantly increased by co-feeding m-cresol or phenol. Furthermore, the data from nitrogen isothermal adsorption–desorption presented in Table 1 illustrate that mixtures of methanol/m-cresol and methanol/phenol led to much more reduction in surface area and volume of micropores of HBeta compared to pure methanol. These results are in agreement with the TGA data given in Table 3 showing that the amount of coke deposited on HBeta had a noticeable increase by addition of m-cresol or phenol to methanol. Therefore, the presence of these phenolic compounds result in high zeolite deactivation, and in turn less aromatization of methanol is occurred. It is also clear from the data in Fig. 4 and Table 1 that phenol caused a higher reduction in number of free acid sites and surface area of HBeta compered to m-cresol. This clearly indicates that phenol has more negative effect than m-cresol on catalytic performance of zeolite. The reason for this is the interaction between hydroxyl of phenol molecules and zeolite framework oxygen atoms producing phenolate ions which are strongly adsorbed on the oxygen atoms linked to framework aluminium.¹⁷ Meanwhile, the phenol molecules tightly bound to zeolite acid sites act as coke precursor and result in rapid formation of coke which is deposited on catalyst surface causing fast deactivation. But, m-cresol might have less potential to form a tight bond with acid sites due to the steric hindrance caused by the methyl group present on the phenolic ring of m-cresol. In fact, the steric bulk around the hydroxyl of m-cresol provided by the methyl group prevents from the interaction between the hydroxyl and zeolite framework oxygen atoms. The significant effect of this steric hindrance is due to the confined space inside the micropores of HBeta where the reaction occurs. Furthermore, one other reason for lower aromatization of methanol in the presence of m-cresol or phenol could be some competing reactions which might be occurred by addition of these phenols. Since methanol could be used as alkylating agent for alkylation of phenolic compounds over zeolite, a fraction of methanol might be involved in alkylation reactions and not undergo deoxygenation and aromatization.^24,25 Besides, transalkylation of aromatics produced from methanol with the co-fed phenols or the compounds produced from transformation of phenols could vary product distribution.^26,27 However, considering the significant reduction in number of free acid sites and surface area of HBeta exposed to m-cresol or phenol, it seems that catalyst deactivation is the main cause for less aromatization of methanol in the presence of these phenols.

Fig. 4 NH₃-TPD profiles of fresh HBeta and HBeta used in different reactant systems (WHSV, 2 h⁻¹; time on stream, 60 min; carrier gas, N₂).

It was shown in this study that bifunctional catalyst of HBeta impregnated with iron which promotes hydrogenolysis reaction was effective for cleavage of C–O bond of m-cresol under hydrogen atmosphere; m-cresol was transformed into aromatics with a noticeable yield of 17.5 wt% over Fe/HBeta. However, this catalyst was not efficient for deoxygenation of simple phenol molecule which, as mentioned above, is tightly bound to zeolite acid sites. It can be inferred that in transformation of m-cresol, the main source of catalyst deactivation is the adsorption of phenol molecules which are produced through demethylation of m-cresol. Therefore, reduction of reaction selectivity toward phenol production could increase lifetime of zeolite and its catalytic activity. Consequently, Fe/HBeta is expected to exhibit higher catalytic lifetime compared to HBeta when exposed to phenolic compounds; over Fe/HBeta and under hydrogen atmosphere, phenolic compounds could undergo hydrogenolysis and lower amount of phenol is produced and subsequently less adsorption of phenol on zeolite acid sites is occurred compared to the case of using pure HBeta zeolite as catalyst. This study clearly demonstrates the undesired effect of phenol molecule on catalytic performance of zeolite in deoxygenation of biomass derived feedstocks; phenolic compounds not only have low reactivity over zeolite but also high potential of phenol molecule to be tightly bound to zeolite acid sites causes rapid catalyst deactivation. Therefore, it seems to be essential to use modified zeolites for catalytic conversion of feedstocks derived from biomass with high content of lignin.

Table 4 demonstrates the catalytic activity of HBeta and Fe/HBeta at reaction temperature of 450 °C. In conversion of methanol at this temperature, the aromatics yield of HBeta was 51.3 wt% while it was 59.6 wt% at 350 °C. Comparison of the data presented in Tables 2 and 4 reveals that negative effect of co-feeding phenol or m-cresol is reduced by increase of reaction temperature, and aromatics yield is less decreased by addition of these phenols at higher temperature of 450 °C. By co-feeding m-cresol or phenol with methanol at 350 °C, the aromatics yield of HBeta was decreased from 59.6 to 24.3 and 6.2 wt%, respectively. However at 450 °C, the presence of m-cresol or phenol in feed mixture caused less influence on HBeta performance and aromatics yield decreased from 51.3 to 36.7 and 12.3 wt%, respectively. As can be seen from TPD profiles in Fig. 4, in the transformation of mixtures of methanol/m-cresol and methanol/phenol over HBeta, the reduction in the number of zeolite free acid sites at reaction temperature of 450 °C was lower than that at 350 °C. Meanwhile, the data in Table 1 indicate that surface area and pore volume of HBeta were less affected in the presence of m-cresol or phenol at 450 °C compared to 350 °C. Also as shown in Table 3, the coke content of spent HBeta was decreased at elevated temperature. Similarly, the increase of temperature attenuated the negative effect of phenol on catalytic activity of Fe/HBeta; in transformation of m-cresol over Fe/HBeta, phenol addition led to reduction of m-cresol conversion and aromatics yield from 58.5 and 17.5 to 30.1 and 3.2 wt% at 350 °C, and from 74.7 and 30.3 to 42.0 and 10.6 wt% at 450 °C, respectively. As can be observed in Table 3, the amount of coke deposited on catalyst is less increased by addition of phenol or m-cresol at reaction temperature of 450 °C compared to 350 °C. For instance, addition of 10 wt% phenol to m-cresol at 350 °C caused an increase of 2.53 wt% in the coke content of Fe/HBeta (from 9.33 to 11.86 wt%). However, presence of phenol led to less increase of 0.91 wt% of coke content (from 6.11 to 7.02 wt%) at 450 °C. Therefore, it can be inferred that increase of temperature led to lower adsorption of phenolic compounds on zeolite acid sites causing less catalyst deactivation. One reason for this is that increase of temperature leads to increase in diffusion rate of phenol molecules in the pores of catalyst and reduces the possibility of phenol adsorption on zeolite acid sites. Inês Graça et al.¹⁷ revealed that increase of temperature did not result in less adsorption of phenol on HZSM-5 zeolite in transformation of mixture of methylcyclohexane/phenol. They concluded that the 10-membered ring channels of HZSM-5 are too narrow which cause slow diffusion of phenol molecules even at higher temperature of 450 °C. However, HBeta zeolite used in this work contains 12-membered ring channels (0.66 × 0.67 and 0.56 × 0.56 nm) which are larger than HZSM-5 channels (0.51 × 0.55 and 0.53 × 0.56 nm).²³ Therefore, increase of reaction temperature might cause faster diffusion of phenol molecules in at least the larger channel (0.66 × 0.67) of HBeta resulting in less adsorption of phenol on zeolite acid sites and in turn lower formation of coke and catalyst deactivation. The other reason for the positive effect of temperature increase is the exothermic nature of phenol adsorption. Therefore at higher temperature, less phenol molecules can be tightly bound to acid sites. In fact, less number of zeolite acid sites, only those with high acidic strength, can adsorb and retain phenol molecules at elevated temperature. As can be seen from the TPD profiles shown in Fig. 3, desorption peak temperatures for HBeta and Fe/HBeta were below 350 °C illustrating that the majority of acid sites of HBeta or Fe/HBeta are not of high acidic strength. Therefore by increase of reaction temperature, the number of acid sites of these zeolites which could adsorb and retain phenol molecules was noticeably decreased (Fig. 4), resulting in less catalyst deactivation at 450 °C compared to 350 °C. This can be another reason why phenol adsorption on HZSM-5 was not decreased by increase of temperature in the study held by Inês Graça et al.¹⁷ HZSM-5 mostly shows two TPD peaks with one at temperature above 400 °C.^28–30 Therefore, HZSM-5 contains considerable density of strong acid sites which probably could still adsorb phenol molecules at higher temperature of 450 °C. The dependency of phenol adsorption on strength of acid sites is also shown in Fig. 4; it can be seen from TPD profiles that strong acid sites had higher reduction compared to weak acid sites, and the acid sites of very low strength were only affected at 350 °C. As a result, it can be concluded that higher reaction temperature as well as the use of zeolite with larger pore size and lower density of strong acid sites could be efficient for atmospheric deoxygenation of phenolic compounds.

Table 4 Product yields and selectivities (wt%) obtained from catalytic conversion of different reactants over HBeta and Fe/HBeta. Reaction conditions: WHSV, 2 h⁻¹; reaction temperature, 450 °C; pressure, 1 atm

Feed	MeOH^a	MeOH–cresol	MeOH–phenol	Cresol	Cresol–phenol
Catalyst	HBeta	HBeta	HBeta	Fe/HBeta	Fe/HBeta
Gas	N2	N2	N2	H2	H2
a MeOH: methanol.
%Yield of organic phase
	51.27	45.48	20.69	85.27	83.19
%Selectivity in organic phase
Toluene				3.39	1.11
Xylene	40.10	34.17	25.62	9.31	3.25
Ethyl-methylbenzene	22.12	19.81	15.51	7.49	3.32
Trimethylbenzene	24.52	16.14	10.97	9.17	2.86
Tetramethylbenzene	6.14	2.92	1.30	1.51	0.49
Naphthalenes				1.55	0.53
Other hydrocarbons	7.12	7.61	5.99	3.08	1.21
Phenol		8.14	30.98	15.80	20.87
Cresol		9.08	6.34	29.69	57.65
Other oxygenates		2.13	3.29	19.01	8.71

---

4. Conclusions

The composition of biomass and its amount of lignin content greatly affect the catalytic performance of zeolites in deoxygenation of biomass pyrolysis derived feedstocks. In this work, at reaction temperature of 350 °C, aromatics yield obtained from conversion of methanol over HBeta was 59.6 wt% which was decreased to 24.3 and 6.2 wt% by co-feeding 10 wt% m-cresol and phenol, respectively. Meanwhile, the coke content of HBeta used for conversion of pure methanol was 0.22 wt% which was increased to 2.61 and 3.48 wt% in the presence of m-cresol and phenol, respectively. It was revealed that atmospheric deoxygenation of m-cresol occurred over Fe/HBeta through hydrogenolysis under hydrogen atmosphere; aromatics yield of 17.5 wt% was achieved from conversion of m-cresol over Fe/HBeta. However, this yield was dramatically decreased to 3.2 wt% by addition of 10 wt% phenol since zeolite acid sites are rapidly occupied by phenol molecules due to high potential of phenol to be tightly bound with zeolite acid sites. But, the steric hindrance caused by the methyl group on the phenolic ring of m-cresol prevents from its strong adsorption on zeolite acid sites. Therefore, it could be concluded that the origin of zeolite deactivation in atmospheric conversion of pure m-cresol is the strong adsorption of simple phenol molecules which are produced from demethylation of m-cresol. Hydrogenolysis promoted by Fe active sites not only is effective for deoxygenation of m-cresol but also is a competing reaction with its demethylation on zeolite acid sites causing lower production of phenol and in turn less occupation of acid sites and catalyst deactivation. The negative effect of phenol on performance of zeolite showed to be attenuated at elevated temperature; addition of phenol to m-cresol reduced the aromatics yield of Fe/HBeta from 17.5 to 3.2 wt% at 350 °C, and from 30.3 to 10.6 wt% at 450 °C. Meanwhile, the adsorption of phenol molecules on zeolite acid sites can be reduced by using a zeolite type with larger pore size and less density of strong acid sites. It could be inferred from this study that pure zeolite such as HBeta is not suitable for deoxygenation of feedstocks derived from biomass with high content of lignin due to low reactivity of lignin derived phenolic compounds over zeolite acid sites as well as strong adsorption of phenolics especially simple phenol molecules on the acid sites. Zeolite impregnated with the metals which could promote hydrogenolysis reaction is efficient for atmospheric deoxygenation of such feedstocks.

Acknowledgements

The authors would like to acknowledge the Faculty of Engineering at University of Malaya for financial support through the HIR Grant (D000011-16001).

References

1. J. C. Serrano-Ruiz and J. A. Dumesic, Energy Environ. Sci., 2011, 4, 83–99.
2. K. L. Hew, A. M. Tamidi, S. Yusup, K. T. Lee and M. M. Ahmad, Bioresour. Technol., 2010, 101, 8855–8858.
3. C. Perego and A. Bosetti, Microporous Mesoporous Mater., 2011, 144, 28–39.
4. G. Fogassy, N. Thegarid, G. Toussaint, A. C. van Veen, Y. Schuurman and C. Mirodatos, Appl. Catal., B, 2010, 96, 476–485.
5. P. S. Rezaei, H. Shafaghat and W. M. A. W. Daud, Appl. Catal., A, 2014, 469, 490–511.
6. Q. Zhang, J. Chang, T. Wang and Y. Xu, Energy Fuels, 2006, 20, 2717–2720.
7. Y. Zhang, T. R. Brown, G. Hu and R. C. Brown, Chem. Eng. J., 2013, 225, 895–904.
8. F. Melligan, M. H. B. Hayes, W. Kwapinski and J. J. Leahy, Energy Fuels, 2012, 26, 6080–6090.
9. M. C. Samolada, A. Papafotica and I. A. Vasalos, Energy Fuels, 2000, 14, 1161–1167.
10. I. Graça, F. R. Ribeiro, H. S. Cerqueira, Y. L. Lam and M. B. B. de Almeida, Appl. Catal., B, 2009, 90, 556–563.
11. A. G. Gayubo, A. T. Aguayo, A. Atutxa, R. Aguado and J. Bilbao, Ind. Eng. Chem. Res., 2004, 43, 2610–2618.
12. A. G. Gayubo, A. T. Aguayo, A. Atutxa, R. Aguado, M. Olazar and J. Bilbao, Ind. Eng. Chem. Res., 2004, 43, 2619–2626.
13. X. Li, L. Su, Y. Wang, Y. Yu, C. Wang, X. Li and Z. Wang, Front. Environ. Sci. Eng., 2012, 6, 295–303.
14. H. Ben and A. J. Ragauskas, Energy Fuels, 2011, 25, 4662–4668.
15. W. Huang, F. Gong, M. Fan, Q. Zhai, C. Hong and Q. Li, Bioresour. Technol., 2012, 121, 248–255.
16. C. A. Mullen and A. A. Boateng, Fuel Process. Technol., 2010, 91, 1446–1458.
17. I. Graça, J.-D. Comparot, S. b. Laforge, P. Magnoux, J. M. Lopes, M. F. Ribeiro and F. Ramôa Ribeiro, Energy Fuels, 2009, 23, 4224–4230.
18. I. Graça, J. D. Comparot, S. Laforge, P. Magnoux, J. M. Lopes, M. F. Ribeiro and F. R. Ribeiro, Appl. Catal., A, 2009, 353, 123–129.
19. I. Graça, A. Fernandes, J. M. Lopes, M. F. Ribeiro, S. Laforge, P. Magnoux and F. Ramôa Ribeiro, Appl. Catal., A, 2010, 385, 178–189.
20. G. Q. Zhang, T. Bai, T. F. Chen, W. T. Fan and X. Zhang, Ind. Eng. Chem. Res., 2014, 53, 14932–14940.
21. H. A. Zaidi and K. K. Pant, Catal. Today, 2004, 96, 155–160.
22. M. Conte, J. A. Lopez-Sanchez, Q. He, D. J. Morgan, Y. Ryabenkova, J. K. Bartley, A. F. Carley, S. H. Taylor, C. J. Kiely, K. Khalid and G. J. Hutchings, Catal. Sci. Technol., 2012, 2, 105–112.
23. J. Jae, G. A. Tompsett, A. J. Foster, K. D. Hammond, S. M. Auerbach, R. F. Lobo and G. W. Huber, J. Catal., 2011, 279, 257–268.
24. M. E. Sad, C. L. Padró and C. R. Apesteguía, Catal. Today, 2008, 133–135, 720–728.
25. W. Wang, P. De Cola, R. Glaeser, I. Ivanova, J. Weitkamp and M. Hunger, Catal. Lett., 2004, 94, 119–123.
26. X. Zhu, L. L. Lobban, R. G. Mallinson and D. E. Resasco, J. Catal., 2011, 281, 21–29.
27. X. Zhu, R. G. Mallinson and D. E. Resasco, Appl. Catal., A, 2010, 379, 172–181.
28. Y. Ni, A. Sun, X. Wu, G. Hai, J. Hu, T. Li and G. Li, Microporous Mesoporous Mater., 2011, 143, 435–442.
29. J. Kim, M. Choi and R. Ryoo, J. Catal., 2010, 269, 219–228.
30. J. Liu, G. Jiang, Y. Liu, J. Di, Y. Wang, Z. Zhao, Q. Sun, C. Xu, J. Gao, A. Duan, J. Liu, Y. Wei, Y. Zhao and L. Jiang, Sci. Rep., 2014, 4, 7276.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra07420g

---