Steric Effect and Evolution of Surface Species in the

Subscriber access provided by UNIV OF PITTSBURGH

Article

Steric Effect and Evolution of Surface Species in the Hydrodeoxygenation of Bio-oil Model Compounds over Pt/HBEA Guo Shiou Foo, Allyson K. Rogers, Matthew M. Yung, and Carsten Sievers ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.5b02684 • Publication Date (Web): 11 Jan 2016 Downloaded from http://pubs.acs.org on January 18, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Steric Effect and Evolution of Surface Species in the Hydrodeoxygenation of Bio-oil Model Compounds over Pt/HBEA Guo Shiou Foo,1 Allyson K. Rogers,1 Matthew M. Yung,2 Carsten Sievers1*

1

School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive NW, Atlanta, Georgia 30332, United States

2

National Renewable Energy Laboratory, 15013 Denver West Parkway, Golden, Colorado 80401, United States * Email: [email protected]

1 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 52

Abstract The formation and evolution of surface species during the hydrodeoxygenation of various bio-oil model compounds (anisole, m-cresol and guaiacol) over Pt/HBEA and HBEA is investigated. Anisole and m-cresol form phenate and cresolate species on Lewis acid sites, while guaiacol can chemisorb more strongly forming bidentate surface species. The position of functional groups within these molecules has a strong influence on the degree of hydrodeoxygenation over Pt/HBEA due to steric hindrance of the C-O scission step. Consequently, the highest yield of deoxygenated products is formed over anisole, followed by m-cresol and guaiacol. No deoxygenation products are produced from HBEA. Based on operando transmission FTIR spectroscopy experiments at 400 °C and 1 atm of hydrogen pressure, a timeline for the formation of polynuclear aromatics and graphitic coke from aromatics with different substituents is established for Pt/HBEA. The early formation of relatively small amounts of graphitic coke and polynuclear aromatics results in pronounced catalyst deactivation. In addition, the formation of strongly adsorbed monomeric species appears to restrict transport processes within the zeolite pores and contribute to deactivation.

Keywords: Deactivation, Coke formation, Polynuclear aromatics, Operando FTIR Spectroscopy, Biomass


Page 3 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1. Introduction Biomass is a promising renewable energy source and can provide an alternative to depleting fossil fuels. Furthermore, its utilization can reduce the overall carbon footprint of fuel.1-2 An attractive approach for biomass utilization involves the fast pyrolysis of biomass to obtain bio-oils. Bio-oils are a complex mixture of several types of oxygenated compounds, such as carboxylic acids, aldehydes, ketones, phenolics, esters and furans.3 Unfortunately, bio-oils cannot directly be used as a fuel because of their high oxygen content.4 The latter leads to low stability, high viscosity, and a low heating value, which is unsuitable for current energy systems. One way of upgrading bio-oils is through catalytic hydrodeoxygenation (HDO), where the oxygen-based functional group is removed as water in the presence of hydrogen gas.5 Early work focused on supported CoMo and NiMo sulfide catalysts.6-7 These commercial catalysts are commonly used for hydrodesulphurization (HDS) and hydrodenitrogenation (HDN) in oil refineries. However, they display limited activity in HDO reactions.8 In addition, hydrogen sulfide must be added into the feedstock to maintain their catalytic activity. To overcome these problems, several noble and transition metal catalysts were studied extensively in HDO reactions.9-24 Many supported metals (Ru, Re, Rh, Pd, Pt, Ni, Fe) display activity in HDO reaction. However, most of these catalysts require a high hydrogen pressure, and they convert aromatics to products with saturated rings. Bimetallic catalysts, such as Pt-Sn,25 Ni-Cu,26 RhPt,27 Pd-Fe,28 and Fe-Mo, were also utilized.29 Some of these catalysts are active and selective in the apparent direct cleavage of C-O bond in phenolics, but a high hydrogen pressure is also required. Zhu et al. studied the hydrodeoxygenation of anisole and m-cresol over a bifunctional catalyst (Pt/HBEA) in vapor-phase under atmospheric pressure and produced deoxygenated aromatics such benzene, toluene and xylene.18, 23 Reducible metal oxides such as MoO3 and 3 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 52

ceria-zirconia were also investigated.30-32 Even in the absence of a metal, these catalysts display activity in the direct C-O bond cleavage of phenolic compounds. It is postulated that HDO is carried out via a reverse Mars-van Krevelen reaction with the aid of oxygen vacancies. Recently, it was reported that metallic Mo2C is a selective HDO catalyst under atmospheric hydrogen pressure, producing a high yield of deoxygenated aromatics.33 While most of the research has focused on catalytic reactivity of HDO catalysts, there is limited literature on specifics of catalyst deactivation, which is commonly observed. Coke formation is the most commonly mentioned deactivation path in HDO studies.34-35 While it is conceivable that surface reactions of aromatics with functional groups play a key role in coke formation, little is known about such surface reactions. Popov et al. studied the adsorption of various phenolic compounds on different catalyst supports.36-38 The studies indicated that the formation of phenate species on Lewis acid sites could be the source of catalyst poisoning. However, there is still a lack of insight for supported metal catalysts, which are extensively used in HDO reactions. More importantly, the evolution of these surface species has yet to be elucidated as there are no reports on in-situ or operando spectroscopic studies on HDO reactions. The understanding of the deactivation mechanism provided by such studies will be critical to the development of suitable catalysts and optimizing reaction and catalyst regeneration processes. In this work, we focus on Pt supported on HBEA, as it was reported that this catalyst displays high HDO activity under atmospheric hydrogen pressure.18, 22 Most importantly, it was shown that metallic Pt reduces coke formation during HDO of anisole and m-cresol.18, 23 By understanding this phenomenon, this study offers insight to develop catalyst with improved resistance against deactivation. The surface chemistry of oxygenates on HBEA is also studied for comparison. Anisole, m-cresol, and guaiacol are used as bio-oil model compounds as they 4 ACS Paragon Plus Environment

Page 5 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

represent a large fraction of phenolic compounds found in bio-oils. Furthermore, the effect of different types and positions of functional groups on the degree of deoxygenation will be highlighted, providing insight into the various reaction paths for the conversion of oxygenates. By using operando transmission FTIR spectroscopy we are able to monitor the performance of the catalyst and the evolution of surface species simultaneously. This allows us to establish a timeline for the optimal use and rapid regeneration of the catalysts. 2. Experimental 2.1 Materials. Zeolite beta (CP814E, SiO2/Al2O3 = 25) was purchased from Zeolyst. Silicon carbide (200-450 mesh particle size), anisole (>99.7%), guaiacol (>98%), and tetraamineplatinum(II) chloride hydrate (99.99%) were purchased from Sigma Aldrich. Hydrogen and helium (UHP Grade 5) were purchased from Airgas. Chloroform (99.8%) and mcresol (99%) were purchased from VWR. A Barnstead NANOpure ultrapure water system was used to further purify deionized water to 18.2 MΩ/cm. 2.2 Catalyst Synthesis. Zeolite Beta (HBEA) was calcined at 550 °C for 4 h. Pt/HBEA was prepared via wetness impregnation method of HBEA using a dilute aqueous solution of Pt(NH3)4Cl2. The slurry was stirred for 12 h, followed by drying at 110 °C for 12 h. Subsequently, the catalyst was calcined at 420 °C for 4 h. 2.3

Characterization.

Nitrogen

physisorption

and

hydrogen

chemisorption

measurements were taken using a Micromeritics ASAP 2020 physisorption/chemisorption analyzer. For N2 physisorption, the catalysts were degassed at 200 °C for 4 h prior to measurement. Surface areas and pore volumes were calculated based on the BET method39 and


ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 52

the BJH method,40 respectively. For H2 chemisorption, the sample was degassed at 100 °C for 1 h and 400 °C for 15 min. Next, the sample was reduced under hydrogen flow at 400 °C for 30 min and evacuated for another 30 min. The temperature was decreased to 35 °C and H2 chemisorption analysis was performed. A H2/Pt stoichiometry of 2 was used to calculate metal dispersion. To determine the amount of Pt metal deposited, Pt/HBEA was sent to Galbraith Laboratories for inductively coupled plasma-atomic emission spectroscopy (ICP-AES) analysis. X-ray diffraction (XRD) patterns were obtained using a Philips X’pert diffractometer equipped with an X’celerator module using Cu Kα radiation. Diffractograms were collected at incident angles from 2θ = 5 to 70° with a step size of 0.0167°.

27

Al MAS NMR spectroscopy was

performed with a Bruker DSX 400 spectrometer. The samples were inserted into a 4 mm zirconia rotor and spun at 12 kHz. At this spectrometer, 27Al has a resonance frequency of 104.2 MHz. A π/12 pulse was used for excitation, and a recycling delay of 250 ms was used. Pyridine adsorption followed by FTIR spectroscopy was performed using a Nicolet 8700 FTIR Spectrometer with an MCT/A detector. Each spectrum was recorded with 64 scans with a resolution of 4 cm-1. Each sample was pressed into a self-supported wafer and loaded into a vacuum FTIR transmission cell. The sample was activated at 400 °C for 1 h under high vacuum and cooled down to 100 °C. A background spectrum was taken. The chamber was dosed with 0.10 mbar of pyridine for 30 min. Subsequently, the cell was evacuated for 12 h to remove physisorbed pyridine and a spectrum was taken. To determine the strength of acid sites, the sample was heated to 200 °C, 300 °C and 400 °C for 1 h, and a spectrum was taken at 100 °C. After each experiment, the density of the wafer was determined by using a circular stamp of 6.35 mm to cut a disc of specific size from the wafer. The concentration of Lewis and Brønsted acid


Page 7 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

sites were determined by the integral of the peaks at 1445 cm-1 and 1540 cm-1, respectively. Extinction coefficients used were reported by Datka et al.41 2.4 Catalytic Performance. Reactivity experiments for the hydrodeoxygenation of anisole, m-cresol and guaiacol were performed in a trickle-bed reactor setup. A stainless steel tube (0.25 inch outer diameter with a wall thickness of 0.035 inch) was used as the reactor. The reactor was mounted in an insulated furnace. A thermocouple was placed at the middle of the reactor and it was connected to a Eurotherm 2416 Temperature Controller. The catalyst (HBEA or 1.3 wt% Pt/HBEA) was mixed with silicon carbide. The catalyst bed with a volume of 1 ml was used in all experiments. Quartz wool was placed at both ends of the catalyst bed to keep the catalyst in place. The organic model compound was fed into the reactor using an Agilent 1100 Series HPLC pump. Each experiment was performed at 400 °C and atmospheric pressure. HBEA and Pt/HBEA was reduced in the reactor for 1 h at 400 °C with 80 ml/min of hydrogen flow before the model compound was fed. All of the lines were heated to 300 °C to prevent condensation. A total hydrogen gas flow rate of 80 ml/min was used during reaction. The H2/reactant molar ratio was kept above 50. The weight to feed ratio (W/F) is defined as the ratio between the mass of the catalyst (gcat) and the molar flow rate of the reactant (mmolfeed h-1). The products were directly analyzed and quantified online using an Agilent 7890A Gas Chromatograph equipped with a HP-5 column and a flame ionized detector. The Weisz-Prater criterion was applied to ensure that all experiments were performed under conditions free of mass transfer limitation (see Supporting Information). The following definitions are used:

% =

× 100%

(1)


ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

!" % =

#

× 100%

Page 8 of 52

(2)

The catalyst deactivation profile of Pt/HBEA for the three model compounds were calculated by assuming a first order decay kinetic model.30, 42 $ = $ exp −) *

(3)

where X is the fractional conversion of the model compound at a given time t, X0 is the initial conversion of the model compound, and kd is the first order deactivation rate constant. 2.5 Adsorption of Model Bio-oil Compounds on HBEA. The adsorption of anisole, mcresol and guaiacol on HBEA was observed using a Nicolet 8700 FTIR spectrometer with an MCT/A detector. Each spectrum was recorded with 64 scans at a resolution of 4 cm-1. For the adsorption of anisole, HBEA was pressed into a self-supported wafer and loaded into a vacuum FTIR transmission cell. The sample was activated at 400 °C for 1 h under high vacuum. Subsequently, a background spectrum was taken at 400 °C, 300 °C, 200 °C and 100 °C. At 100 °C, anisole was introduced into the vacuum cell and maintained at a pressure of 0.100 mbar for 15 min. The cell was evacuated for 30 min and a spectrum was taken. The sample was heated at 200 °C, 300 °C and 400 °C for 1 h and a spectrum was taken at each temperature. For the adsorption of m-cresol and guaiacol, HBEA was initially impregnated with the model compounds. About 250 mg of HBEA was mixed with 10 ml of chloroform solution that had 6 mg of m-cresol or guaiacol dissolved in it. The slurry was dried in a fume hood to allow chloroform to evaporate. Subsequently, the impregnated samples are pressed into a selfsupported wafer and loaded into a vacuum transmission FTIR cell. The sample was evacuated under high vacuum ( 5nm).18 Figure S3 shows the 27

Al MAS NMR spectra of HBEA and Pt/HBEA. Both samples displayed peaks at 53 ppm and 0

ppm, which are assigned as tetrahedrally and octahedrally coordinated aluminum species, respectively.45 Framework aluminum species are tetrahedrally coordinated, while extra framework aluminum species are mostly octahedrally coordinated. Peak fitting showed that the fraction of extra framework aluminum species decreased from 37% to 31% after impregnation of Pt metal, suggesting that some of the octahedrally coordinated extra framework aluminum is 10 ACS Paragon Plus Environment

Page 11 of 52

reintroduced into the zeolite framework during calcination following the wet impregnation step.46 The concentration of Lewis (LAS) and Brønsted acid sites (BAS) on Pt/HBEA was lower compared to HBEA (Figure 1). This is probably due to anchoring of the Pt metal particles on some of the acid sites.47

700 HBEA BAS HBEA LAS Pt/HBEA BAS Pt/HBEA LAS

600 Concentration (µmol/g)

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

500 400 300 200 100 0 Fresh

Anisole

m-Cresol

Guaiacol

Figure 1. Concentration of Lewis (LAS) and Brønsted acid sites (BAS) of fresh and spent HBEA and Pt/HBEA catalysts. Reaction conditions: 400 °C, 80 ml/min H2, W/F = 0.0109 gcat (mmolfeed h-1)-1. 3.2 Reactivity of Bio-Oil Model Compounds. Figure 2A shows the conversion of anisole over HBEA and Pt/HBEA. The conversion over HBEA started at 60% and decreased with increasing time on stream. For Pt/HBEA, the conversion was a 100% in the first 2 h, followed by slow deactivation. The products over HBEA were phenol, cresol, methylanisole and xylenol (Figure S3A). In addition to these products, Pt/HBEA produced deoxygenated aromatics such as benzene, toluene and xylene (Figure S3B and S3C). The yield of deoxygenated aromatics was 28% at the first measurement after 0.5 h time on stream and declined to 16% within 8 h 11 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 52

(Figure 2D). In the case of m-cresol, Pt/HBEA initially showed a slightly higher conversion than HBEA (Figure 2B). With increasing time on stream, HBEA deactivated at a faster rate than Pt/HBEA. The products from m-cresol over HBEA were only p-cresol and o-cresol (Figure S4A). However, in the presence of Pt metal, toluene and trace amounts of methane and phenol were produced (Figure S4B). The yield of deoxygenated aromatics from m-cresol over Pt/HBEA was lower than during the conversion of anisole, with an initial yield of 13% that corresponds to a selectivity of 24% (Figure 2D). Figure 2C shows the conversion of guaiacol over HBEA and Pt/HBEA. The conversion of guaiacol over HBEA was less than 5%. The addition of Pt metal increased the initial conversion increased to 60%, and it slowly decreased to 35% after 8 h. The products from HBEA were only isomers of methylcatechol (Figure S5A). In the reaction over Pt/HBEA, a wide variety of products was observed. Deoxygenated aromatics such as benzene, toluene and xylene were produced along with methane (Figure S5B). Among the oxygenated products, catechol was produced in a higher yield compared to phenol, cresol, xylenol, methylcatechol and high molecular weight products (Figure S6C). The yield of deoxygenated aromatics dropped from 5% to 3% over the course of the reaction. Note that only 8% of the initial products were deoxygenated. Using a first order deactivation model, the deactivation rate constants (kd) for the conversion of anisole, m-cresol and guaiacol over Pt/HBEA were calculated to be 0.016 h-1, 0.044 h-1, and 0.057 h-1, respectively.


Page 13 of 52

60 40 20 0

1

2

3

4 5 TOS (h)

40 20

HBEA 1.3 wt% Pt/HBEA 0

6

7

0

8

0

1

40 Yield of Deoxygenated Aromatics (%)

HBEA 1.3 wt% Pt/HBEA

80

(C)

60 40 20

0

1

2

3

4 5 TOS (h)

(B)

60

100

0

HBEA 1.3 wt% Pt/HBEA

80

80

Conversion (%)

Conversion (%)

100

(A)

100

Conversion (%)

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

6

7

8

2

3

4 5 TOS (h)

6

Anisole m-Cresol Guaiacol

30

7

8

(D)

20

10

0

0

1

2

3

4 5 TOS (h)

6

7

8

Figure 2. Conversion of (A) anisole, (B) m-cresol, (C) guaiacol over HBEA and 1.3 wt% Pt/HBEA, (D) yield of deoxygenated aromatics from anisole, m-cresol and guaiacol over 1.3 wt% Pt/HBEA. Reactions conditions: 400 °C, 80 ml/min H2, W/F = 0.0109 gcat (mmolfeed h-1)-1. 3.3 Temperature Programmed Reaction of Anisole on HBEA followed by FTIR Spectroscopy. At room temperature, the FTIR spectrum of anisole on HBEA was similar to the spectrum of liquid anisole (Figure 3). The peaks at 1602 cm-1, 1587 cm-1 and 1500 cm-1 are attributed to ν(C=C)ring vibrations.36, 48 The peaks at 1470 cm-1 and 1454 cm-1 are due to the δasymCH3 vibrations of the methoxy group, while the peak at 1443 cm-1 is assigned as the δsymCH3 vibration. As the temperature increased, the peaks at 1602 cm-1 and 1587 cm-1 merged into a band at 1592 cm-1, while the peak at 1500 cm-1 red-shifted to 1495 cm-1, indicating that the aromatic ring is interacting with Lewis acid sites to a certain extent.36 The triplet peaks at 14701443 cm-1 merged into a band at 1458 cm-1, which has been assigned as the δasymCH3 vibration of 13 ACS Paragon Plus Environment

ACS Catalysis

surface methyl groups attached to an oxygen atom.36, 49 Its δsymCH3 vibration (expected around 1100 cm-1) could not be observed due to the structural vibration of HBEA. When the sample was heated to 400 °C, major changes were observed. Shoulders at 1525 cm-1 and 1508 cm-1 emerged, and they are due to the aromatic ring vibrations of para-disubstituted, and ortho- and metadisubstituted benzenes, respectively.48, 50 The peaks due to aromatic stretching vibrations at 1592 cm-1 and 1495 cm-1 shifted to 1596 cm-1 and 1491 cm-1, respectively. This is probably due to the emergence of the new aromatic species. The emerging weak peak at 1383 cm-1 is attributed to the deformation vibration of CH3 groups that are directly attached to aromatic rings. Similar results were obtained for anisole adsorbed on Pt/HBEA.

1468

1587

1454

1497 1599

1440 Anisole

1470

1602 1587

1454

1500

Intensity (a.u.)

1443

1495 25 °C 1592 1458 100 °C 200 °C 300 °C

1491 1596 1528

1600

1508

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 52

1500 Wavenumber (cm-1)

1383

400 °C

1400

Figure 3. FTIR spectra of anisole adsorbed on HBEA at various temperatures. 3.4 Temperature Programmed Reaction of m-Cresol on HBEA followed by FTIR Spectroscopy. Figure 4 shows the FTIR spectra of m-cresol adsorbed on HBEA. At room temperature, a broad peak at 1635 cm-1 was observed, and this is due to physisorbed water. Additional peaks at 1489 cm-1 and 1452 cm-1 were also observed. As the temperature increased, 14 ACS Paragon Plus Environment

Page 15 of 52

the broad peak at 1635 cm-1 vanished, revealing three peaks at 1635 cm-1, 1599 cm-1 and 1583 cm-1. Furthermore, a broad peak at 1527 cm-1 was observed. The peaks at 1599 cm-1, 1583 cm-1 and 1489 cm-1 are due to the ν(C=C)ring vibrations of m-cresol.48 The asymmetrical δCH3 vibrations of the methyl group of pure m-cresol were observed at 1462 cm-1 and 1437 cm-1. However, these two peaks became a broad peak at 1452 cm-1, indicating that the methyl group appears to be perturbed when m-cresol interacts with the surface. In addition, three more peaks were observed at 1389 cm-1, 1352 cm-1 and 1327 cm-1. The two former peaks are most likely related to the asymmetrical and symmetrical deformation vibration of a methyl group attached to an aromatic ring, respectively.48 These two peaks are not resolved for pure m-cresol due to the broad peak at 1333 cm-1, which is assigned as the phenolic δOH vibration.51 When m-cresol was adsorbed on HBEA, this broad peak vanished. The rise of a peak at 1327 cm-1 with increasing temperatures could be attributed to a ν(C-O) vibration of cresolate species.52 The two new peaks at 1635 cm-1 and 1527 cm-1 are most likely related to ν(C=C)ring vibrations of new aromatic species that have been formed.

1589 1491 1462

1614

1437

m-Cresol

1380

1333

1635

1599

25 °C

Intensity (a.u.)

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

1583

1527 1452

100 °C

1389 1352 1327

200 °C 300 °C 400 °C

1700

1600

1500 Wavenumber (cm-1)

1400

1300

Figure 4. FTIR spectra of m-cresol adsorbed on HBEA at various temperatures. 15 ACS Paragon Plus Environment

ACS Catalysis

3.5 Temperature Programmed Reaction of Guaiacol on HBEA followed by FTIR Spectroscopy. After adsorption of guaiacol on HBEA at room temperature, a peak at 1632 cm-1 was observed, and it is due to the presence of physisorbed water (Figure 5). The shoulder at 1610 cm-1 and the intense peak at 1497 cm-1 are assigned as the ν(C=C)ring vibrations of the aromatic ring. The peak at 1467 cm-1 is attributed to the δCH3 vibrations of the methoxy group. The δOH peak at 1360 cm-1 in the spectrum of pure guaiacol vanished when guaiacol was adsorbed on HBEA at room temperature. This indicates that the hydrogen atom on the phenol has dissociated.36 At the same time, two peaks at 1327 cm-1 and 1292 cm-1 appeared, which are due to ν(C-O) vibration.36 The observation of at least two new ν(C-O) vibrational modes suggests the formation of bidentate surface species. However, the formation of two distinct surface species cannot be ruled out based on the spectrum alone. As the temperature increased, the peak due to physisorbed water vanished, while the intensity for the rest of the peaks decreased slightly.

1498

Guaiacol

1614

1597

1470 1444

1360 1292

Intensity (a.u.)

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 52

1632 25 °C

1327

1610 1467

100 °C 200 °C 300 °C 400 °C

1700

1600

1500 1400 Wavenumber (cm-1)

1300

Figure 5. FTIR spectra of guaiacol adsorbed on HBEA at various temperatures.


Page 17 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

3.6 Operando FTIR Spectroscopic Study of HDO of Anisole over HBEA and Pt/HBEA. To observe and understand the evolution of surface species over time, an operando transmission FTIR gas flow cell was used under the same reaction conditions as the trickle bed reactor. It was shown that performance of Pt/HBEA for HDO of anisole in the operando FTIR cell was in good agreement with that in the trickle bed reactor (Figure S1). The operando FTIR spectra showed that multiple surface species were formed from anisole conversion over HBEA in the first 10 min on stream (Figure 6A). Peaks due to asymmetric aromatic compounds (1595 cm-1), characteristic aromatic vibration (1493 cm-1) and surface methyl groups (1455 cm-1) were observed (Table 2).36, 48-50, 53 The presence of various substituted monoaromatics (1530 cm-1 and 1507 cm-1) and aromatic methyl groups (1379 cm-1) indicated that transalkylation was occurring. After 4 min on stream, a small shoulder was observed at 1560 cm-1, which is assigned to the ν(C=C)ring vibration of strongly adsorbed polynuclear aromatic species.48,

50, 54

These polynuclear aromatic species are similar to

anthracenes and naphthalenes. Subsequently, the intensity of the peak at 1560 cm-1 increased with increasing time on stream.


ACS Catalysis

0.6

1507 1560

1493

1530 1455

0.4

1379

0.2

10min–2min 1h – 10min 2h – 1h 4h – 1h 8h – 1h

1597

1561

1595

(C)

0.8

0.6

1537

1544 1528

0.4

1374

Intensity (a.u.)

0.8

10 min 8 min 6 min 4 min 2 min

1554

(A)

Intensity (a.u.)

1.0

0.2

0.0 1580

0.0

1.0

8h 1558 7h 1507 6 h 1593 5h 4h 1530 1495 3h 2h 1455 1h

1350

1650

(B)

1377

0.6 0.2

1.0


1350

(D)

0.9 1558 cm-1 1530 cm-1 -1 1495 cm Phenol

0.8 0.7 0.6 0.5 0.4

1650

1550 1450 1350 Wavenumber (cm-1)

0

1

2

3 4 5 TOS (h)

6

7

8

100 90 80 70 60 50 40 30 20 10 0

Yield (%)

1.4


Normalized IR Intensity

1650

Intensity (a.u.)

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 52

Figure 6. FTIR spectra of anisole over HBEA in (A) 10 min, (B) 8 h, (C) Difference FTIR spectra at various time scales, (D) Normalized IR intensity of peaks and yield of phenol,. Reactions conditions: 400 °C, 80 ml/min H2, W/F = 0.0109 gcat (mmolfeed h-1)-1.


Page 19 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Table 2. FTIR peak assignments for hydrodeoxygenation of bio-oil model compounds -1

Frequency (cm )

Assignment

1600

ν(C=C)ring

1593 - 1583

ν(C=C)ring

1500 - 1483

ν(C=C)ring

1580 - 1570

Group Asymmetrically substituted aromatics

References 48, 50

Aromatic six-membered rings

36, 48, 50

ν(C=C)ring

Graphitic coke

55, 56, 57

1560 - 1540

ν(C=C)ring

Polynuclear aromatics

1537 – 1518

ν(C=C)ring

Para-disubstituted aromatics

48, 50

48, 50 1,2,4-trisubstituted aromatics 1507 – 1504

ν(C=C)ring

Ortho- and meta-disubstituted aromatics 48, 50 1,2,3-trisubstituted aromatics

1450

δCH3

Surface methyl groups

1390 - 1370

δasymCH3

Methyl groups attached to aromatic

1350

δsymCH3

rings

36, 49

36, 48, 50, 51

Over 8 h of time on stream, the intensity of all the peaks increased and approached an asymptotic limit, indicating that the surface coverage was approaching saturation (Figure 6B). The peaks at 1593 cm-1 and 1377 cm-1 became very intense, revealing that most of the aromatic monomers were asymmetrically substituted with methyl and phenolic groups. To further elucidate the evolution of surface species over time, difference FTIR spectra were obtained on various time scales for anisole conversion over HBEA (Figure 6C). On a short time scale (10 min – 2 min), the formation of monoaromatics (1597 cm-1 and 1537 cm-1) and methyl groups (1374 cm-1) was evident. A small amount of polynuclear aromatics (1561 cm-1) was deposited. Within 1 h, an intense peak at 1544 cm-1 appeared. Peaks between 1540-1550 cm-1 have been assigned to strongly adsorbed polynuclear aromatic coke species that are not affected by N2 19 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 52

flushing after reaction.54 The peak of substituted aromatics appeared at 1528 cm-1 instead of 1537 cm-1, indicating that heavier or more substituted species were deposited with increasing time on stream.48 At longer times on stream (between 2 and 8 h), the intensity of the peak at 1580 cm-1 increased, which is attributed to the ν(C=C)ring vibration of graphitic coke species.55-57 Figure 6D shows the normalized intensity of various IR peaks and the yield of phenol, the only product with reduced molecular mass. In the first 2 h of reaction, the yield of phenol decreased rapidly. However, the integrated IR area only increased by 9.7% from 1 h to 2 h time on stream, and the area of the peaks between 1580 cm-1 to 1540 cm-1 (heavy surface species) only makes up approximately 27% of the increased area. The rest of the increased area is due to relatively light surface species such as monoaromatics. During this period (1 h to 2 h), the yield of phenol decreased by 26%. It is likely that only a small amount of polynuclear aromatics or graphitic coke is sufficient to deactivate the catalyst during this period, as these surface species are large enough to block the pores of the catalyst compared to monoaromatics. The intensity of the polynuclear (1558 cm-1) and substituted aromatics (1530 cm-1) peaks followed an opposite trend to the yield of phenol as it increased rapidly during the first 2 h. Afterwards, they continued to grow, which occurred at the same rate as the peak corresponding to aromatic six-membered rings (1495 cm-1). This indicates that most of the surface of the catalyst is still covered with monoaromatics, while moderate amounts of polynuclear aromatics and graphitic coke form. The operando FTIR spectra during HDO of anisole over Pt/HBEA contained most of the same peaks as the spectra recorded during the conversion of anisole over HBEA (Figure 7A), and they are assigned to the same species (Table 2). The intensity of the peak due to asymmetrical aromatic species (1593 cm-1) was comparable to the general six-membered aromatics (1492 cm-1), as most of the deoxygenated products produced are symmetric (benzene 20 ACS Paragon Plus Environment

Page 21 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

and toluene). The difference FTIR spectra of surface species from anisole over Pt/HBEA on a short time scale (10 min – 2 min) contained characteristic peak for monoaromatics (1596 cm-1, 1533 cm-1 and 1493 cm-1) and methyl groups (1375 cm-1) (Figure 7B). Between 10 min and 1 h time on stream, polynuclear aromatics (1546 cm-1) were formed, but the peak in the different spectrum was not as intense as during the conversion of anisole over HBEA indicating that these species accumulate more slowly. The signal of substituted aromatics also appeared at 1527 cm-1 instead of 1533 cm-1, indicating the formation of heavier or more substituted deposits. At longer times on stream, coke species (1580 cm-1) and heavily substituted aromatics (1518 cm-1) were formed. The yield of deoxygenated aromatics (benzene, toluene and xylene) decreased rapidly in the first 2 h of reaction (Figure 7C), but the integrated IR area only increased by 7.7% from 1 h to 2 h time on stream. About 18% of the increased area consists of graphitic coke and polynuclear aromatic species, while the rest is due to relatively light surface species. During this period (1 h to 2 h), the yield of deoxygenated aromatics decreased by 27% (Figure 7C). The normalized IR peak intensity of the substituted aromatic monomers (1525 cm-1) increased more rapidly relative to the peak at 1492 cm-1 (Figure 7C).


Page 22 of 52

2.0 8 h 7h 1507 6h 1.5 5 h 1492 1593 4h 3h 1525 1.0 2 h 1452 1h

(A)

1375

0.5

1650


1533 1546 1527

1375

1596

Intensity (a.u.)

0.8

1493

0.6

1350


(B)

0.4 0.2

1650

1.0

1518


1350

(C)

0.9

1525 cm-1 1492 cm-1 Deoxygenated Aromatics

0.8 0.7 0.6 0.5

0

1

2

3 4 5 TOS (h)

6

7

8

100 90 80 70 60 50 40 30 20 10 0

Yield (%)

1580

0.0


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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Intensity (a.u.)

ACS Catalysis

Figure 7. (A) FTIR spectra of anisole over 1.3 wt% Pt/HBEA in 8 h, (B) Difference FTIR spectra at various time scales, (C) Normalized IR intensity of peaks and yield of deoxygenated aromatics. Reactions conditions: 400 °C, 80 ml/min H2, W/F = 0.0109 gcat (mmolfeed h-1)-1. 22 ACS Paragon Plus Environment

Page 23 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

3.7 Operando FTIR Spectroscopic Study of HDO of m-Cresol over HBEA and Pt/HBEA. In the first 10 min of reaction, the FTIR spectra of surface species formed from mcresol on HBEA showed the rise of several bands (Figure 8A). Intense peaks due to asymmetrically substituted aromatic species (1600 cm-1), six-membered aromatic rings (1583 cm-1 and 1486 cm-1) and methyl groups (1439 cm-1) were observed.48, 50 Similarly, the growing shoulder and peak at 1527 cm-1 and 1506 cm-1 indicated that the presence of various disubstituted monoaromatics, along with aromatic methyl groups (1390 cm-1 and 1354 cm-1). This suggests the isomerization of m-cresol to p- and o-cresol. After about 4 min of reaction, a shoulder at 1568 cm-1 appeared, and it is also assigned to the presence of polynuclear aromatics.54 As time progressed, the peak due to asymmetric aromatics (1600 cm-1) became increasingly intense. Figure 8B shows the FTIR spectra of m-cresol over HBEA over the course of 8 h. The intensity of all the peaks increased and approached an asymptotic limit, indicating that the surface of the catalyst was approaching saturation. The peak at 1600 cm-1 was the most intense, as the surface was largely populated by asymmetrically substituted aromatic species. The difference FTIR spectra collected during conversion of m-cresol over HBEA indicated the formation of asymmetrically substituted monoaromatics (1600 cm-1) in the first 1 h time on stream (Figure 8C). Substituted monoaromatics (1591 cm-1, 1506 cm-1 and 1485 cm-1), methyl groups (1439 cm-1, 1390 cm-1 and 1354 cm-1), and polynuclear aromatics (1570 cm-1) were also formed. At longer times on stream (1 h to 8 h), the polynuclear aromatic peak was observed at 1550 cm-1 instead of 1570 cm-1, probably indicating that the newly formed species were heavier or had more substituents.48 Furthermore, the formation of coke-like species (1580 cm-1) was more pronounced in this period. The integrated IR area only increased by 8% from 1 h to 2 h 23 ACS Paragon Plus Environment

ACS Catalysis

time on stream, and approximately 35% of the increased area consists of graphitic coke and polynuclear aromatics, while the rest is due the presence of monoaromatics and methyl groups attached to aromatic rings. The intensity of the peak corresponding to substituted monoaromatics (1506 cm-1) reached its maximum intensity after 5 h (Figure 8D), followed by the peak assigned to general ν(C=C)ring vibration (1487 cm-1).

1486

1568 1506

0.5

1527

1439


(C) 1600 1.5 1570 1506 1485

1439

1583 1.0

10 min 8 min 6 min 4 min 2 min 1 min 0.2 min

1600

1608 1591

(A)

Intensity (a.u.)

Intensity (a.u.)

1.5

1390 1354

1.0

1390 1354 0.5

0 1580 1550


1350

1527

8h 7h 6h 1600 1506 2.0 5 h 1583 4h 3h 1486 1568 2h 1.0 1 h 1439

(B)

1390 1354

1650

1.0 Normalized IR Intensity

1650

Intensity (a.u.)

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 52


1350

1350

(D)

0.9 0.8 0.7 1487 cm-1 1506 cm-1

0.6 0.5

1650


0

1

2

3

4 5 TOS (h)

6

7

8

Figure 8. FTIR spectra of m-cresol over HBEA in (A) 10 min, (B) 8 h, (C) Difference FTIR spectra at various time scales, (D) Normalized IR intensity of peaks over time. Reactions conditions: 400 °C, 80 ml/min H2, W/F = 0.0109 gcat (mmolfeed h-1)-1. During the conversion of m-cresol over Pt/HBEA, the FTIR spectra showed the same peaks as during the reaction over HBEA (Figure 9A, Table 2). However, the peak due to 24 ACS Paragon Plus Environment

Page 25 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

polynuclear aromatics (1568 cm-1) was only obvious in the difference spectra (Figure 9B). The intensity of the peak due to asymmetrically substituted aromatics at 1600 cm-1 was almost equivalent to 1487 cm-1 (general ν(C=C)ring vibration). The overall intensity of the peaks also increased approaching a limit. The difference FTIR spectra measured during the conversion of m-cresol over Pt/HBEA showed that substituted monomers (1603 cm-1, 1527 cm-1, 1506 cm-1) and methyl groups (1435 cm-1, 1390 cm-1 and 1354 cm-1) were formed between 10 min and 2 min on stream (Figure 9B). Compared to HBEA, the intensity of the peak due to polynuclear aromatics (1568 cm-1) was lower. Coke-like species (1585 cm-1) began to grow after 2 h, while monoaromatics (1493 cm-1) were accumulated throughout the experiment. Figure 9C shows the normalized intensity of the peaks at 1487 cm-1 (general ν(C=C)ring vibration) and 1504 cm-1 (substituted monomers), and the yield of deoxygenated aromatics (toluene) over time. Both peaks increased at the same rate. After 3 h of time on stream, the normalized intensity of the peak at 1487 cm-1 increased at a slower rate. The yield of toluene decreased by 22% in the first 3 h, followed by a slower decay. While the integrated IR peak area increased by 31% from 1 min to 3 h time on stream (Figure S7), graphitic coke and polynuclear aromatics only accounted for approximately 10% of the increased area, whereas the rest of the increase was due to monoaromatics.


ACS Catalysis

8h 7h 6h 5h 4h 3h 2h 1h

(A) 1487

1600

1.0

1585 1504

1529

Intensity (a.u.)

1.4

0.6

1439 1389

0.2 1650


Intensity (a.u.)

1568

1527

1603

0.6

1354 1350


(B) 1506 1435

1390 1354

1614 0.4

0.2

1493 1585 1650


1350

10

1.0 (C)

9 0.9

1487 cm-1 1504 cm-1 Deoxygenated Aromatics

0.8

8 7

Yield (%)


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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 52

6 0.7 5 0

1

2

3

4 5 TOS (h)

6

7

8

Figure 9. (A) FTIR spectra of m-cresol over 1.3 wt% Pt/HBEA in 8 h, (B) Difference FTIR spectra at various time scales, (C) Normalized IR intensity of peaks and yield of deoxygenated aromatics. Reactions conditions: 400 °C, 80 ml/min H2, W/F = 0.0109 gcat (mmolfeed h-1)-1.


Page 27 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

3.8 Operando FTIR Spectroscopic Study of HDO of Guaiacol over HBEA and Pt/HBEA. The FTIR spectra of surface species from guaiacol on HBEA showed four major peaks (Figure 10A). Specifically, aromatic species (1583 cm-1 and 1500 cm-1), surface methyl groups (1450 cm-1) and methyl groups bound to aromatic rings (1374 cm-1) were observed.36, 49 The peaks of products from guaiacol on HBEA are extremely broad compared to the surface species from anisole and m-cresol, suggesting that they contain contributions from multiple similar but distinct chemisorbed surface species, such as polynuclear aromatics (1560 cm-1). The broad peak in the region below 1350 cm-1 is attributed to a distribution of ν(C-O) vibrational modes suggesting the formation of various species that are chemisorbed through their oxygen atoms as observed in the adsorption and temperature programmed reaction of guaiacol on HBEA (vide supra). The contribution from other surface species is even more apparent in the difference FTIR spectra obtained during guaiacol conversion over HBEA (Figure 10B). At short times on stream, peaks corresponding to monoaromatics (1600 cm-1) and coke (1570 cm-1) increased at a similar rate. There was also a shoulder at 1540 cm-1, which has been assigned as polynuclear aromatics.54 With increasing time on stream, the difference peak at 1570 cm-1 became more intense, indicating that more coke-like species were formed on the surface. Figure 10C shows the normalized IR peak intensity of the ν(C=C)ring vibration (1583 cm-1) and δCH3 vibration (1374 cm-1) of the methyl group bound to aromatic rings. Both normalized IR peaks increased at the same rate, suggesting that the aromatic ring vibration is likely due to the presence of the transalkylated species. Between 1 h and 2 h time on stream, the integrated IR area only increased by 1.2%. Approximately 75% of the increased area is due to graphitic coke and polynuclear aromatic species, while the rest is due to asymmetrically substituted aromatics.


ACS Catalysis

8h 7h 6h 5h 4h 1.5 3h 2h 1h

Intensity (a.u.)

2.5

(A)

1500

1583

1450 1374

0.5

1600


0.9 1600 1570

Intensity (a.u.)

1300


(B)

0.6

0.3

0.0 1650

Normalized Peak Intensity

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 52

1.00

1550 1450 1350 Wavenumber (cm-1)

(C)

0.95 0.90 0.85 0.80 1583 cm-1 1374 cm-1

0.75 0.70

0

1

2

3

4 5 TOS (h)

6

7

8

Figure 10. (A) FTIR spectra of guaiacol over HBEA in 8 h, (B) Normalized IR intensity of peaks over time, (C) Difference FTIR spectra at various time scales. Reactions conditions: 400 °C, 80 ml/min H2, W/F = 0.0109 gcat (mmolfeed h-1)-1.


Page 29 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

The FTIR spectra of surface species from guaiacol on Pt/HBEA contained the four broad peaks 1589 cm-1, 1502 cm-1, 1452 cm-1 and 1380 cm-1 at short and long time on stream (Figure S8 and 11A). These peaks have the same assignment as the products from guaiacol on HBEA. Furthermore, aromatic and methyl vibrational modes from various phenolics (Figure S6C) contributed to the broad peaks. At short times on stream, the difference spectra of products from guaiacol on Pt/HBEA contained a shoulder at 1532 cm-1 (substituted monoaromatics), in addition to peaks assigned to asymmetrically substituted aromatics (1606 cm-1) and coke (1577 cm-1) (Figure 11B). On longer time scales, the intensity of the difference peak due to substituted monoaromatics was less pronounced, while the intensity of the peak due to polynuclear aromatics increased. Figure 11C shows the increase in peak intensity of the aromatic (1589 cm-1) and methyl group (1380 cm-1) over time. The yield of deoxygenated aromatics (benzene, toluene and xylene) decreased from 3.8% to 3.5% from 1 h to 2 h time on stream, while the integrated IR area only increased by 1.3% from 1 h to 2 h time on stream. Approximately 38% of the increased area is due to the presence of graphitic coke and polynuclear aromatic species.


8h 7h 6 h 1589 5h 4h 3h 2h 1h

1.5 1.0 0.5

Page 30 of 52

(A)

1502 1452 1380

0.0 1600 1500 1400 Wavenumber (cm-1)

(B)

1577

Intensity (a.u.)

0.3

1532

1300


1606 0.2

0.1

0.0

1.00


1350

8

(C)

0.95

0.85

1589 cm-1 6 1380 cm-1 Deoxygenated Aromatics

0.80

4

0.90

Yield (%)

1650

Normalized Peak Intensity

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Intensity (a.u.)

ACS Catalysis

0.75 0.70

0

1

2

3

4 5 6 TOS (h)

7

8

2

Figure 11. (A) FTIR spectra of guaiacol over 1.3 wt% Pt/HBEA in 8 h, (B) Normalized IR intensity of peaks and yield of deoxygenated aromatics, (C) Difference FTIR spectra at various time scales. Reactions conditions: 400 °C, 80 ml/min H2, W/F = 0.0109 gcat (mmolfeed h-1)-1.


Page 31 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

3.9 Analysis of Spent Catalysts. The coke content of the spent HBEA and Pt/HBEA catalysts were analyzed by temperature programmed oxidation (TPO). For the spent Pt/HBEA catalysts, reactions with guaiacol produced the highest coke content (18.1%), followed by mcresol (13.7%) and anisole (10%), while all of the spent HBEA catalysts accumulated more coke than spent Pt/HBEA. The reactions of guaiacol, m-cresol, and anisole over HBEA produced 22.8%, 15.2%, and 14.4% of coke, respectively. The temperature-programmed oxidation (TPO) curves of spent HBEA had larger CO2 evolution peaks that occurred at a higher temperature (560 °C) compared to spent Pt/HBEA (390 °C to 420 °C) (Figure S9). Carbon species that are oxidized between 500 °C and 600 °C is classified as hard coke with a very low hydrogen to carbon ratio.58 CO2 that is derived around 400 °C is related to the combustion of carbonaceous species that can be transformed at lower temperatures (soft coke). The peaks at 120 °C is related to the loss of water from the samples. Based on pyridine adsorption followed by IR spectroscopy, it was found that all of the spent catalysts had lower concentrations of accessible acid sites compared to the fresh catalysts (Figure 1). Furthermore, higher concentrations of Lewis acid sites remained accessible in the spent Pt/HBEA catalysts compared to the corresponding spent HBEA samples. 4. Discussion 4.1 Adsorption and Temperature Programmed Reaction of Bio-Oil Model Compounds. There are few reports on the adsorption of oxygenates that are typical bio-oil model compounds. Popov et al. studied the adsorption of phenolic compounds on silica, alumina and sulfided CoMo catalysts.36-38 The aromatic compounds interact with silica via hydrogenbonding with surface silanol groups. However, it was also reported that this interaction could be


ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 52

the result of a weakly-bound physisorbed state, even for aromatics without functional groups.59 Interactions between phenolic compounds and Lewis acid sites on alumina result in the formation of phenate species.36 It was observed that the phenate species formed on alumina have limited interaction with the supported CoMo sulfide phase.37-38 The adsorption of m-cresol on sodium-exchanged zeolite was also reported.51 Ion exchange between the acidic hydrogen atom of m-cresol and the sodium cation in the zeolite was observed. However, none of these catalysts had Brønsted acid sites. Since zeolites showed great potential in many studies on HDO, it is crucial to understand the interaction of these phenolic compounds in the presence of Brønsted acid sites. The chemisorption of anisole on HBEA occurred through an interaction between the oxygen atom of anisole and catalyst surface above 100 °C, while no chemisorption was observed at room temperature. This was indicated by the replacement of the triplet bands of the methoxy group by a single peak at 1458 cm-1 (Figure 3). It was suggested that the PhO-CH3 bond was cleaved, resulting in the formation of a surface methyl group.36 This would also result in the formation of phenate species containing an oxygen atom that is coordinated to a Lewis acid site (Scheme 1). The formation of such a species causes the frequencies of the aromatic ring vibrations to red-shift due to donor-acceptor interaction with the Lewis acid site.36 Unfortunately, the corresponding ν(C-O) vibration could not be observed due to the strong structural vibrations of HBEA under 1300 cm-1. The ν(C-O) vibrations of the phenate group on alumina were observed at 1295 cm-1 and 1248 cm-1, which are attributed to the formation of a monodentate and bridging phenate species, respectively.36 When the sample was heated to 400 °C, new aromatic species were formed as indicated by peaks at 1525 cm-1 and 1508 cm-1. This indicates that reactions occurred in the presence of Brønsted acid sites, as this was not observed for alumina.36 32 ACS Paragon Plus Environment

Page 33 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

The rise of the δCH3 peak at 1383 cm-1 (methyl group bound to aromatic rings) suggests that transalkylation occurred. This observation agrees with the reactivity data of anisole over HBEA that showed the formation of products such as cresols, xylenols and methylanisoles (Figure S4A). In another study, it was reported that an increase of Brønsted to Lewis acid site ratio resulted in an increase of alkylation and transalkylation activity over MCM-22 zeolite,60 which further supports that Brønsted acid sites can catalyze these reactions. However, in other catalytic systems such as sulfided CoMo and NiMo supported on ɣ-alumina, transalkylation was also observed to a certain extent, indicating that Lewis acid sites could also play a role.61-62

Scheme 1. Interaction of anisole, m-cresol and guaiacol on HBEA Changes of the aromatic ring vibrations of m-cresol were observed when it was adsorbed on HBEA (Figure 4). The absence of the broad peak (δOH vibration) at 1333 cm-1 and the appearance of a sharp peak (ν(C-O) vibration) at 1327 cm-1 suggest that a cresolate was formed,52 which involves dissociation of the hydrogen atom from the phenol group and 33 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 52

coordination of the oxygen atom to a Lewis acid site (Scheme 1). This interaction with the Lewis acid site caused the peaks of the aromatic ring vibrations to red-shift and change in intensity due to donor-acceptor effect. The ν(C-O) vibrational peak of the cresolate has a higher frequency compared to a phenolate, probably due to the presence of a methyl group that can donate electrons to the π orbital of the aromatic ring.23 The formation of two new peaks at 1635 cm-1 and 1527 cm-1 can be attributed to the isomerization of m-cresol to o- and p-cresol, indicating that the methyl group can be transferred (Scheme 1). p-Cresol has an intense ν(C=C)ring vibration at 1514 cm-1. In this case, it appears that its interaction with the surface caused a blue-shift in its aromatic vibrational frequency, indicating an increased donation of electrons to the aromatic ring.50, 63 The adsorption of guaiacol on HBEA led to significant changes in the IR spectrum at room temperature compared to its liquid form (Figure 5). The disappearance of the δOH peak at 1360 cm-1 indicates that the hydrogen atom on the phenol group was dissociated. In addition, the appearance of two ν(C-O) peaks at 1327 cm-1 and 1292 cm-1 suggests that guaiacol has formed a bidentate surface species, or two different types of chemisorbed surface species. The continued existence and decreasing intensity of the methoxy vibrations around 1467 cm-1 indicates that some of the functional group remains intact while others are converted. Thus, it is speculated that two surface species could be present. The first species still contains a methoxy group, and the oxygen atom of the dissociated phenol group is coordinated to a Lewis acid site, while the oxygen atom of the methoxy group is engaged in a Lewis acid/base interaction, forming a methoxy phenate species (Scheme 1). Similar interactions were observed for glycerol adsorbed on various metal oxides.64-66 The formation of the second species involves the decomposition of the methoxy group, and both of the oxygen atoms appear to be coordinated to one or more Lewis 34 ACS Paragon Plus Environment

Page 35 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

acid sites, forming a bidentate catecholate. It was reported that these bi- or multidentate surface species interact with metal oxides much more strongly compared to monodentate species, which can slow down the transformation of ɣ-alumina into boehmite in hot liquid water.67-68 In addition, vibrational modes due to methyl groups attached to an aromatic ring were not observed, suggesting that transalkylation did not occur to a significant extent. The results in our work are different from a study on guaiacol adsorption on alumina, where the methoxy group of guaiacol was quantitatively removed to form a surface methyl group at 100 °C.36 Since Brønsted acid sites were not present in alumina, additional aromatic species were not formed. 4.2

Hydrodeoxygenation

of

Different

Bio-oil

Model

Compounds.

The

hydrodeoxygenation of bio-oil model compounds has been studied over a wide variety of catalysts. These catalysts include sulfided NiMo and CoMo,6-8 noble and base metals,13-15, 18-23, 27 reducible metal oxides,30-32 as well as carbides.33 However, most studies reported the HDO of only one model compound, such as anisole or guaiacol. Since bio-oil consists of different types of phenolics, it is critical to compare the HDO of various bio-oil model compounds under the same reaction condition. In this way, the effect of the nature of different functional groups and their position on the aromatic ring can be correlated to the reactivity of specific compounds in the HDO reaction as well as their role in different deactivation paths. Several HDO reaction mechanisms have been proposed for different types of catalysts. For the HDO of anisole over Mo2C catalyst, it was found that the Ph-OMe bond was preferentially cleaved, producing benzene with high selectivity.33 Based on reactivity results, the authors proposed that two separate active sites are required for the activation of hydrogen and anisole. However, in the case of reducible metal oxides such as MoO3 and ceria-zirconia, a reverse Mars-van Krevelen mechanism was proposed.30-32 Specifically, the oxygenated aromatic 35 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 52

compounds interact with surface oxygen vacancies, and the Ph-O bond is preferentially cleaved. Subsequently, the oxygen vacancy is regenerated by reduction with hydrogen. A similar HDO mechanism was proposed for sulfided NiMo and CoMo catalysts with sulfur vacancies.4-5 In literature, many studies on HDO were performed over supported metals on metal oxides. On metals such as Pd, Ru, and Ni, the aromatic rings of reactants are preferentially hydrogenated. 12-15, 20, 27 This is followed by dehydration of the resulting alcohol group on an acid sites of the support (e.g. a zeolite), and the unsaturated carbon-carbon bond is hydrogenated again. Fe particles supported on poorly-acidic SiO2 converted guaiacol to HDO products that retained their aromaticity.69 It was reported that guaiacol adsorbs on the silica support via its oxygen atoms rather than its aromatic ring.36 Based on this result, Olcese et al. proposed that the Ph-O bond is preferentially attacked by spilt-over hydrogen species from the Fe metal particles. Supported noble metals were found to catalyze demethylation first, followed by hydrodeoxygenation.18, 25, 35, 70-71 In this study, phenol and catechol are formed from anisole and guaiacol, respectively, while methane is produced from both reactants (Figure S4 and S6). This result agrees with literature that the hydrogenolysis of the PhO-CH3 bond is the preferential step over Pt.18, 25, 35, 7273

The HBEA support can transfer the cleaved CH3 to an aromatic ring, resulting in intra- or

intermolecular rearrangement to form transalkylated or dealkylated products. In the presence of Pt metal, the dissociated methyl group can be hydrogenated to form methane. During the reaction of m-cresol over HBEA, some methyl groups are transferred to a different position on the aromatic ring, resulting in isomerization products such as o- and p-cresol (Figure S5B). Since mcresol does not contain a methoxy group, it is not surprising that only a very small amount of methane (< 0.5%) was produced from m-cresol over Pt/HBEA. 36 ACS Paragon Plus Environment

Page 37 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Following demethylation, four different pathways have been proposed for the hydrodeoxygenation of the phenolic oxygen atom. The first major pathway is direct hydrodeoxygenation to form aromatics and water. However, it has been shown that direct deoxygenation without activation of the aromatic ring is at least five orders of magnitude slower than the same reaction of an activated ring over Pt metal.74 This is because the bond strength of Ph-OH is 20 kcal/mol stronger compared to an aliphatic C-OH bond.4 The second proposed pathway is the hydrogenation of the aromatic ring, followed by dehydration to remove the oxygen atom and another hydrogenation step.75 However, cycloalkanes or partially hydrogenated products were not produced in this study. The third proposed pathway involves the partial hydrogenation of the aromatic ring.18, 23, 75 This would enable a less energy intensive cleavage of the aliphatic C-O bond as the delocalization of the oxygen lone pair orbital is inhibited. Subsequently, dehydration of the oxygen atom would restore the aromaticity of the compound. The final pathway involves the tautomerization of the phenol intermediate to 2,4cyclohexadienone, followed by hydrogenation of the carbonyl group and dehydration.76 Since deoxygenated aromatics are produced as the only relevant products in this study, the third or fourth HDO mechanism could be applicable for the Pt/HBEA catalyst used in this study. Comparing the yields of deoxygenated aromatic compounds provides insight into pronounced difference of the reactivity of different model compounds (Figure 2D). Anisole had a yield of deoxygenated aromatics that was 15 percentage points higher compared to m-cresol even though both compounds have only one oxygen-containing functional group. Since demethylation of the methoxy group in anisole is fast (vide supra), the deoxygenation of phenol is assumed to be the kinetically relevant step for anisole like for m-cresol. It was suggested that the presence of a methyl group on an aromatic ring in m-cresol could strengthen the Ph-OH 37 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 52

bond.23 However, the bond dissociation energies of a Ph-OH bond in phenol (106.1 kcal/mol) and m-cresol (106.0 kcal/mol) were calculated to be essentially identical.30 Thus, the significant difference in the degree of hydrodeoxygenation between anisole and m-cresol can be largely attributed to steric effects. Specifically, the additional methyl group in m-cresol could sterically hinder the approach of this reactant or a m-cresol derived surface species to the active site, or the formation of the transition state within a zeolite cage could become hindered. Lastly, the presence of increasingly bulky, strongly adsorbed surface species could restrict the transport of reactants to a significant part of the pore network of the zeolite. The yield of deoxygenated aromatics from guaiacol was even lower than that from anisole and m-cresol. Similarly to anisole, the first step toward hydrodeoxygenation of guaiacol is the kinetically fast decomposition of the methoxy group, followed by deoxygenation of the two phenol groups involving Pt metal sites.74 The decomposition of the methoxy group in guaiacol is slower compared to anisole as shown by the low yield of catechol (Figure S6). Compared to guaiacol, the PhO-Me bond dissociation energy in anisole was calculated to be around 58 kcal/mol, which is 8 kcal/mol higher than guaiacol.30 Nevertheless, the methoxy group in anisole decomposed easily. The low yield of entirely deoxygenated aromatics from guaiacol is not surprising due to the presence of two oxygen functional groups. However, even the yield of partially deoxygenated phenolics (phenol, cresols, and xylenols) from guaiacol was lower compared to the yield of deoxygenated aromatics from anisole. Also, the yield of phenol from guaiacol is lower compared to the yield of toluene from m-cresol. However, the bond dissociation energy of a Ph-OH bond is similar between guaiacol, phenol, and m-cresol.30 These observations support the fact that steric hindrance plays an important role in the degree of hydrodeoxygenation, especially when an adjacent functional group is present on the aromatic 38 ACS Paragon Plus Environment

Page 39 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

ring. It was calculated that the adsorption energy of guaiacol on a metal particle is lower compared to benzene, indicating that the presence of a methoxy and phenol functional group induces steric effect, weakening the binding between adsorbate and metal surface, while changes in the electronic π system could be affected to a certain extent.77 In addition to steric effects of the HDO reaction itself, the formation of strongly bound, bidentate surface species from molecules like guaiacol could block some of the active sites or pore mouths in the zeolite, leading to low activity in deoxygenation. 4.3 Evolution of Surface Species observed by Operando FTIR Spectroscopy. Even though hydrodeoxygenation of various bio-oil model compounds have been studied and coking has been considered the most common route of deactivation, there is a lack of knowledge of the timescale of the evolution of surface species and their impact on catalytic activity. To our knowledge, there have been no reports on operando spectroscopy for any HDO reactions. Understanding the evolution of surface species will allow bio-refineries to anticipate and monitor the formation of polynuclear aromatics or hard coke and the associated deactivation of the catalysts, which requires harsh conditions (typically combustion above 800 °C) for regeneration.78 In contrast, the initial hydrogen-rich surface species in the earlier stages of the process can be stripped off more easily, restoring the activity of the catalysts. Such a mild regeneration protocol could prevent irreversible transformations, such as dealumination of zeolites during combustion of coke at high temperature, and thus, result in a longer lifetime of the catalyst. During the conversion of anisole over HBEA and Pt/HBEA, substituted monoaromatics were initially observed (Figure 6 and 7). This is in agreement with the reactivity and temperature programmed reaction results, which show that anisole undergoes transalkylation or dealkylation 39 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 40 of 52

to form various aromatics. In the case of HBEA, polynuclear aromatics were only observed after 4 min time on stream, indicating that these species are secondary products that only formed when the surface began to saturate. With increasing time on stream, the yield of phenol and deoxygenated aromatics produced from anisole over HBEA and Pt/HBEA, respectively, decreased rapidly in the first 2 h, suggesting that the deactivation of the catalysts mainly took place during this period. The shift of the peaks at 1533 cm-1 and 1561 cm-1 to lower frequencies revealed that the mono- and polynuclear aromatics with heavier or more substituents were formed in the first hour (Figure 6C and 7B). In addition, polynuclear aromatics were observed on Pt/HBEA, but its relative peak intensity was low compared to HBEA. After 1 h on stream, carbonaceous species such as graphitic coke were observed, and the intensity of the corresponding peak increased at longer time scales from 2 h to 8 h. This suggest that monoaromatics are formed initially, which then condense to form simple polynuclear aromatics between 10 min and 1 h on stream. With increasing time on stream, these species became heavier and had more substituents. After 1 h, more of these polynuclear aromatics continued to condense and form heavy polynuclear aromatics and coke, which blocked most of the active sites and deactivated the catalyst. This explains the rapid decrease in the yield of deoxygenated aromatics in the first 2 h of reaction (Figure 7C). However, the integrated IR area only increased by 9.7% and 7.7% between 1 h and 2 h time on stream for anisole over HBEA and Pt/HBEA, respectively. Only 27% and 18% of this increase are associated with graphitic coke and polynuclear aromatics for HBEA and Pt/HBEA, respectively. The rest of the increased area is due to monoaromatics and methyl groups. Although different surface species likely have different extinction coefficient, this suggests that only a small amount of polynuclear aromatics and coke is needed to deactivate the catalysts. However, coke and heavy polynuclear aromatics


Page 41 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

continue accumulating on the surface. Interestingly, this occurs with only a minor effect on the yield of deoxygenated products. Scheme 2A illustrates the timeline on the formation of surface species from anisole over Pt/HBEA and their impact on the activity of the catalyst.

Scheme 2. Timeline on the formation of surface species (A) anisole over Pt/HBEA, (B) m-cresol over Pt/HBEA, (C) guaiacol over Pt/HBEA. During the conversion of m-cresol over HBEA, the formation of polynuclear aromatics was also observed after 4 min time on stream (Figure 8A), suggesting the condensation of monoaromatics. During the reaction over Pt/HBEA, the peak due to polynuclear aromatics was observed at short times on stream between 2 min and 10 min (Figure 9B). Between 10 min and 1 h, the difference FTIR spectrum showed that the intensity of the polynuclear aromatics peak increased on both HBEA and Pt/HBEA. Between 1 h and 2 h on stream, coke-like species began 41 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 42 of 52

to form on both HBEA and Pt/HBEA, while a lower amount of polynuclear aromatics was accumulated. This suggests that the polynuclear aromatics condensed to form coke which led to deactivation as indicated by the decrease of the yield of deoxygenated aromatics (Scheme 2B). The rapid increase in the intensity of the peak at 1568 cm-1 in the first 2 h followed by a slower increase provides further evidence for the condensation of polynuclear aromatics (Figure 8D). Furthermore, the increase in the integrated IR area was 31% between 1 min and 3 h time on stream for m-cresol over Pt/HBEA, but only about 10% of the increase is associated with graphitic coke and polynuclear aromatics, while the rest is due to monoaromatics. During the same period, the yield of deoxygenated product decreased by 22%. This suggests that a small amount of polynuclear aromatics and coke is needed to deactivate the catalysts as it is large enough to block the pores of the catalyst. Interestingly, no changes of the peak positions of mono and polynuclear aromatics were observed for both catalysts. This indicates that the size of the substituents did not increase over time as it was observed for anisole. The growth of such substituents typically occurs via acid catalyzed alkylation reactions.79 In the case of m-cresol, the required reagents for these alkylation steps are not available because the methyl group is only cracked from the aromatic ring to a very minor extent. The almost negligible yield of demethylated products (e.g. phenol) is in agreement with this observation. During the conversion of guaiacol, the FTIR spectra of surface species on HBEA and Pt/HBEA were very similar (Figure 10 and 11), showing peaks that are assigned to monoaromatics, surface methyl group, aromatic methyl group, and surface bound phenates (Table 2). The significant breadth of all peaks in the region of 1650 – 1300 cm-1 (Figures 10 and 11) and the observation of ν(C-O) vibrational modes that are strongly blue shifted from their positions in pure guaiacol indicate that strongly chemisorbed surface species are formed on both 42 ACS Paragon Plus Environment

Page 43 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

catalysts within the first 10 min time on stream (Figure S8). Based on the observations during the temperature programmed reaction of guaiacol (Figure 5) and previous studies,66-68 it is suggested that bidendate catecholate species are formed on the Lewis acid sites of the catalysts. Considering that the pore mouths of zeolite BEA are only slightly larger than the reactants, it is expected that such strongly adsorbed species will form “roadblocks” within the zeolite framework restricting the diffusion of reactants and products in and out of the pores. As a consequence, intermediates may have a higher residence time in proximity to active sites within the framework and undergo secondary reactions including the formation of heavier deposits. The difference FTIR spectra (Figures 10B and 11B) showed that coke-like species were already formed on both catalysts within the first 10 min. Based on the present data, it is not conclusive whether the accelerated formation of coke is due to the intrinsic reactivity of guaiacol or the increased residence time of intermediate in the pores. The early formation of coke and/or “roadblocks” severely reduced the yield of deoxygenated aromatics from guaiacol over Pt/HBEA compared to anisole and m-cresol (Scheme 2C). Alkyl substituted aromatics (1532 cm-1) were only observed on Pt/HBEA. This is tentatively attributed to the catalytic activity of the Pt metal, which accelerates the cleavage of the PhO-CH3 bond.18 With increasing time on stream, polynuclear aromatics continued to grow on both catalysts. The relative intensity of the peak around 1580-1570 cm-1 for HBEA was higher compared to Pt/HBEA. Even though transalkylation was not observed over HBEA, guaiacol is still able to condense to form polynuclear aromatics and coke, probably via protonation and hydrogen abstraction on the aromatic ring.80-81 The timeline for the evolution of surface species over Pt/HBEA indicates that the formation of polynuclear aromatics and coke is dependent on the bio-oil model compound 43 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 44 of 52

(Scheme 2). It is noted that the timeline may change under different reaction conditions, but it correlates the decrease in catalytic activity with the formation of polynuclear aromatics and coke. Interestingly, the results of this study show that relatively small amounts of polynuclear aromatics and coke can lead to a significant reduction of the activity of the catalysts, while most surface species remain relatively light. The formation of heavy surface species occurs latest in the HDO of anisole, followed by m-cresol. Moreover, the spent catalysts had the lowest amount of carbon deposits after the conversion of anisole, followed by m-cresol and guaiacol, indicating that a delayed formation of heavy deposits also slows down and reduces the overall amount of deposits after a specific time on stream. This result also agrees with the calculated first order deactivation rate constant (kd), where the conversion of anisole over Pt/HBEA displayed the lowest deactivation rate, followed by m-cresol and guaiacol. One possible reason for the higher coke content from m-cresol is the presence of a methyl group on the aromatic rings. Hydrogen abstraction is a key process in coke formation.80 It was calculated that the presence of methyl groups can lower the activation energy for hydrogen abstraction at various position on the aromatic ring, allowing the coke matrix to grow rapidly.81 Even though parts of the zeolite framework appear to be blocked during the conversion of guaiacol over HBEA and Pt/HBEA, the high coke carbon on these spent catalysts suggests that a certain fraction of acid sites remain accessible to catalyze aromatic condensation reactions involving protonation and hydrogen abstraction on the aromatic ring.80-81 The TGA results indicate that the presence of Pt metal on HBEA reduces the amount of coke formation, which agrees with the relatively lower intensity of the bands due to graphitic coke and polynuclear aromatics on Pt/HBEA observed in the operando FTIR experiments. In addition, the temperature of the CO2 evolution peak from the spent HBEA catalysts is 44 ACS Paragon Plus Environment

Page 45 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

consistently higher compared to spent Pt/HBEA. This could be due to the ability of Pt metal to catalyze the oxidation of coke at a lower temperature,82 or the coke formed on Pt/HBEA is less graphitic in nature (soft coke).58 The operando FTIR spectra indicated that the presence of Pt can retard the formation of polynuclear aromatics and coke. The lower content of coke could result from hydrogen spillover. Since hydrogen abstraction is involved in coke formation,81 atomic hydrogen species that are formed on Pt sites could migrate to the surface of the HBEA support and suppress coke formation by reversing the hydrogen abstraction reaction. Spillover of hydrogen from metal particles to reducible support and graphitic carbon has been demonstrated.83 However, spillover onto non-reducible and defective supports such as zeolites has not been directly shown. In this case, it is hypothesized that the slow deactivation rate of Pt/HBEA is related to Lewis acidity, where the hydrogen species can migrate to the Lewis acid sites of the HBEA support. Hydrogen spillover has been calculated to be energetically viable for aluminosilicates.84 The concentration of accessible Lewis acid sites on spent Pt/HBEA catalysts was higher compared to spent HBEA (Figure 1), which further supports the hypothesis that spilt over hydrogen on Lewis acid sites can reduce the rate of coking. Since water is an abundant component in crude bio-oil (15-35 wt%),85 the effect of water on the performance and stability of the catalysts must also be considered. Zhu et al. introduced the addition of water in the conversion of anisole over Pt/HBEA and found that the conversion of anisole was immediately enhanced.18 However, the product distribution remained similar and conversion decreased when the flow of water was stopped. The increase in conversion was attributed to the hydrolysis of the methoxy group by water, and it could also be expected in the case of guaiacol. Although the activity of the catalyst can be enhanced in the presence of water,


ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 46 of 52

the HBEA support could undergo dealumination, resulting in decreased crystallinity and acidity.45, 86 4.4 Potential Implications. The results obtained in this study are of great importance in guiding biomass pyrolysis and its subsequent hydrodeoxygenation to transportation fuels and chemicals. Since phenolics with multiple oxygen-based functional groups result in a lower degree of hydrodeoxygenation and rapid catalyst deactivation of the zeolite catalysts used in this study, it would be ideal to perform the HDO reaction in a two-stage process. The first stage could employ different types of catalysts with high activity for removing oxygen from molecules with multiple functional groups. It was reported that ceria-zirconia catalysts are effective in deoxygenating guaiacol to phenol with limited benzene yield.32 Subsequently, phenolics with a limited number of oxygen functional groups can be completely deoxygenated in the second stage by Pt/HBEA. In this way, hydrogenation of the aromatic rings can be minimized, and the lifetime of the catalysts can be extended. This study has also demonstrated that the formation of relatively small amounts of polynuclear aromatics and graphitic coke leads to rapid deactivation. It is suggested that this deactivation could be reversed if the HDO catalysts are subjected to frequent but mild regeneration treatments (below 700 °C). Along these lines, TPO showed that a temperature of 500 °C is sufficient to burn off carbonaceous deposits from spent Pt/HBEA catalysts, while higher temperatures are needed to combust the heavier coke from HBEA. For regeneration, the bio-oil feed can be stopped and the catalyst can be regenerated close to reaction conditions in the presence of oxygen flow. Alternatively, a high hydrogen pressure flow could also be used for


Page 47 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

regeneration.87 In addition, catalyst supports without Lewis acid sites should be considered, as these sites appear to be responsible for the formation of “roadblocks”. 5. CONCLUSIONS Anisole, m-cresol and guaiacol can adsorb on the Lewis acid sites of HBEA in different ways. Anisole and m-cresol form phenate and cresolate surface species, respectively. Guaiacol can adsorb more strongly by forming bidentate catecholate or methoxy phenate species. The type and position of functional groups affects the degree of hydrodeoxygenation, where steric hindrance plays an important factor. Consequently, the highest yield of deoxygenated products is achieved during HDO of anisole, followed by m-cresol and guaiacol. Operando FTIR spectroscopic experiments show that the strong deactivation of the zeolite catalysts in the initial hours of HDO of anisole and m-cresol over Pt/HBEA appears to be caused by the formation of relatively small amounts of graphitic coke and polynuclear aromatics, while most of the surface remains occupied by monoaromatics. The yield of deoxygenated products from guaiacol is significantly lower in every stage of the reaction. This behavior is attributed to the formation of strongly adsorbed species (“roadbloack”), which severely restrict mass transfer of the reactants and products in the zeolite pores. The increased residence time of reactive compounds in the pores could result in accelerated coke formation that was evidenced by the formation of graphitic coke within the first 10 min of the reaction and a larger amount of carbonaceous deposits on the catalyst after 8 h on stream. Based on these results, it appears that coke formation is only one of several factors contributing to the deactivation of zeolites in HDO. The reduced carbon content on spent Pt/HBEA compared to spent HBEA is attributed to the reversal of hydrogen abstraction steps in coke formation on acid sites by spilt-over hydrogen. 47 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 48 of 52

As a result the growth of some carbonaceous species is delayed to the extent that might still be removed from the catalyst. ASSOCIATED CONTENT Supporting Information. Mass transfer limitation test, validation of operando FTIR cell, XRD diffractograms, 27Al MAS NMR spectra, yield of products, TGA of spent catalysts. This material is available free of charge via the Internet at http://pubs.acs.org. AUTHOR INFORMATION Corresponding Author *Email: [email protected] Funding Sources U.S. Department of Energy grant DE-AC36-08-GO28308. Notes The authors declare no competing financial interests. ACKNOWLEDGMENT The Renewable Bioproducts Institute (RBI) is acknowledged for the use of its facilities. Funding from the U.S. Department of Energy is gratefully acknowledged. The authors thank Professor Fabio Ribeiro for providing the CAD drawings of the operando IR cell, as well as Jeffrey Andrews and Brad Parker for constructing the IR cell body. We thank Prof. Andreas Heyden for a helpful discussion. REFERENCES


Page 49 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

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. 29. 30. 31.

Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106, 4044-4098. Bu, Q.; Lei, H. W.; Zacher, A. H.; Wang, L.; Ren, S. J.; Liang, J.; Wei, Y.; Liu, Y. P.; Tang, J. M.; Zhang, Q.; Ruan, R. Bioresour. Technol. 2012, 124, 470-477. Czernik, S.; Bridgwater, A. V. Energy Fuels 2004, 18, 590-598. Furimsky, E. Appl. Catal., A 2000, 199, 147-190. Elliott, D. C. Energy Fuels 2007, 21, 1792-1815. Laurent, E.; Delmon, B. J. Catal. 1994, 146, 281-291. Laurent, E.; Delmon, B. Appl. Catal., A 1994, 109, 77-96. Lin, Y. C.; Li, C. L.; Wan, H. P.; Lee, H. T.; Liu, C. F. Energy Fuels 2011, 25, 890-896. Fisk, C. A.; Morgan, T.; Ji, Y. Y.; Crocker, M.; Crofcheck, C.; Lewis, S. A. Appl. Catal., A 2009, 358, 150-156. Wildschut, J.; Mahfud, F. H.; Venderbosch, R. H.; Heeres, H. J. Ind. Eng. Chem. Res. 2009, 48, 10324-10334. Wildschut, J.; Iqbal, M.; Mahfud, F. H.; Cabrera, I. M.; Venderbosch, R. H.; Heeres, H. J. Energy Environ. Sci. 2010, 3, 962-970. Zhao, C.; Kou, Y.; Lemonidou, A. A.; Li, X. B.; Lercher, J. A. Angew. Chem.-Int. Edit. 2009, 48, 3987-3990. Zhao, C.; Kou, Y.; Lemonidou, A. A.; Li, X. B.; Lercher, J. A. Chem. Commun. 2010, 46, 412-414. Zhao, C.; Lercher, J. A. Angew. Chem.-Int. Edit. 2012, 51, 5935-5940. Zhao, C.; Song, W. J.; Lercher, J. A. ACS Catal. 2012, 2, 2714-2723. Zhao, C.; Lercher, J. A. ChemCatChem 2012, 4, 64-68. Kubicka, D.; Kaluza, L. Appl. Catal., A 2010, 372, 199-208. Zhu, X. L.; Lobban, L. L.; Mallinson, R. G.; Resasco, D. E. J. Catal. 2011, 281, 21-29. Sitthisa, S.; Sooknoi, T.; Ma, Y. G.; Balbuena, P. B.; Resasco, D. E. J. Catal. 2011, 277, 113. Sitthisa, S.; Resasco, D. E. Catal. Lett. 2011, 141, 784-791. Ausavasukhi, A.; Huang, Y.; To, A. T.; Sooknoi, T.; Resasco, D. E. J. Catal. 2012, 290, 90100. Nie, L.; Resasco, D. E. J. Catal. 2014, 317, 22-29. Zhu, X. L.; Nie, L.; Lobban, L. L.; Mallinson, R. G.; Resasco, D. E. Energy Fuels 2014, 28, 4104-4111. Saidi, M.; Samimi, F.; Karimipourfard, D.; Nimmanwudipong, T.; Gates, B. C.; Rahimpour, M. R. Energy Environ. Sci. 2014, 7, 103-129. González-Borja, M. Á.; Resasco, D. E. Energy Fuels 2011, 25, 4155-4162. Khromova, S. A.; Smirnov, A. A.; Bulavchenko, O. A.; Saraev, A. A.; Kaichev, V. V.; Reshetnikov, S. I.; Yakovlev, V. A. Appl. Catal., A 2014, 470, 261-270. Gutierrez, A.; Kaila, R. K.; Honkela, M. L.; Slioor, R.; Krause, A. O. I. Catal. Today 2009, 147, 239-246. Sun, J. M.; Karim, A. M.; Zhang, H.; Kovarik, L.; Li, X. H. S.; Hensley, A. J.; McEwen, J. S.; Wang, Y. J. Catal. 2013, 306, 47-57. Rensel, D. J.; Rouvimov, S.; Gin, M. E.; Hicks, J. C. J. Catal. 2013, 305, 256-263. Prasomsri, T.; Shetty, M.; Murugappan, K.; Roman-Leshkov, Y. Energy Environ. Sci. 2014, 7, 2660-2669. Prasomsri, T.; Nimmanwudipong, T.; Roman-Leshkov, Y. Energy Environ. Sci. 2013, 6, 1732-1738. 49 ACS Paragon Plus Environment

ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 50 of 52

32. Schimming, S. M.; LaMont, O. D.; König, M.; Rogers, A. K.; D'Amico, A. D.; Yung, M. M.; Sievers, C. ChemSusChem 2015, 8, 2073-2083. 33. Lee, W.-S.; Wang, Z.; Wu, R. J.; Bhan, A. J. Catal. 2014, 319, 44-53. 34. Li, Y.; Zhang, C.; Liu, Y.; Hou, X.; Zhang, R.; Tang, X. Energy Fuels 2015, 29, 1722-1728. 35. Mu, W.; Ben, H.; Du, X.; Zhang, X.; Hu, F.; Liu, W.; Ragauskas, A. J.; Deng, Y. Bioresour. Technol. 2014, 173, 6-10. 36. Popov, A.; Kondratieva, E.; Goupil, J. M.; Mariey, L.; Bazin, P.; Gilson, J. P.; Travert, A.; Mauge, F. J. Phys. Chem. C 2010, 114, 15661-15670. 37. Popov, A.; Kondratieva, E.; Gilson, J.-P.; Mariey, L.; Travert, A.; Maugé, F. Catal. Today 2011, 172, 132-135. 38. Popov, A.; Kondratieva, E.; Mariey, L.; Goupil, J. M.; El Fallah, J.; Gilson, J. P.; Travert, A.; Mauge, F. J. Catal. 2013, 297, 176-186. 39. Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309-319. 40. Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951, 73, 373-380. 41. Datka, J.; Turek, A. M.; Jehng, J. M.; Wachs, I. E. J. Catal. 1992, 135, 186-199. 42. Possato, L. G.; Diniz, R. N.; Garetto, T.; Pulcinelli, S. H.; Santilli, C. V.; Martins, L. J. Catal. 2013, 300, 102-112. 43. Wang, J.; Kispersky, V. F.; Delgass, W. N.; Ribeiro, F. H. J. Catal. 2012, 289, 171-178. 44. Prokešová, P.; Mintova, S.; Čejka, J.; Bein, T. Microporous Mesoporous Mat. 2003, 64, 165-174. 45. Ravenelle, R. M.; Schussler, F.; D'Amico, A.; Danilina, N.; van Bokhoven, J. A.; Lercher, J. A.; Jones, C. W.; Sievers, C. J. Phys. Chem. C 2010, 114, 19582-19595. 46. Omegna, A.; Haouas, M.; Kogelbauer, A.; Prins, R. Microporous Mesoporous Mat. 2001, 46, 177-184. 47. Kwak, J. H.; Hu, J. Z.; Mei, D.; Yi, C. W.; Kim, D. H.; Peden, C. H. F.; Allard, L. F.; Szanyi, J. Science 2009, 325, 1670-1673. 48. Socrates, G. Infrared and Raman Characteristic Group Frequencies. 3rd ed.; John Wiley & Sons Ltd: England, 2001; p 157. 49. Jones, A. J.; Iglesia, E. Angew. Chem.-Int. Edit. 2014, 53, 12177-12181. 50. Colthup, N. B.; Daly, L. H.; Wiberley, S. E. Introduction to Infrared and Raman Spectroscopy. 3rd ed.; Academic Press: Waltham, MA, 1990; p 261. 51. Murphy, B.; Davis, M.; Xu, B. Top. Catal. 2015, 58, 393-404. 52. Krishnakumar, V.; Kumar, M.; Prabavathi, N.; Mathammal, R. Spectrochim. Acta, Part A 2012, 97, 144-154. 53. Mathew, T.; Vijayaraj, M.; Pal, S.; Tope, B. B.; Hegde, S. G.; Rao, B. S.; Gopinath, C. S. J. Catal. 2004, 227, 175-185. 54. Fernandez, C.; Stan, I.; Gilson, J.-P.; Thomas, K.; Vicente, A.; Bonilla, A.; Pérez-Ramírez, J. Chem. - Eur. J. 2010, 16, 6224-6233. 55. Park, J. W.; Seo, G. Appl. Catal., A 2009, 356, 180-188. 56. Rozwadowski, M.; Lezanska, M.; Wloch, J.; Erdmann, K.; Golembiewski, R.; Kornatowski, J. Chem. Mat. 2001, 13, 1609-1616. 57. Palumbo, L.; Bonino, F.; Beato, P.; Bjorgen, M.; Zecchina, A.; Bordiga, S. J. Phys. Chem. C 2008, 112, 9710-9716. 58. Torres-Mancera, P.; Rayo, P.; Ancheyta, J.; Marroquin, G.; Centeno, G.; Alonso, F. Fuel 2015, 149, 143-148.


Page 51 of 52

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

59. Reitmeier, S. J.; Mukti, R. R.; Jentys, A.; Lercher, J. A. J. Phys. Chem. C 2008, 112, 25382544. 60. Liu, K. F.; Xie, S. J.; Wei, H. J.; Li, X. J.; Liu, S. L.; Xu, L. Y. Appl. Catal., A 2013, 468, 288-295. 61. Bui, V. N.; Laurenti, D.; Afanasiev, P.; Geantet, C. Appl. Catal., B 2011, 101, 239-245. 62. Bui, V. N.; Laurenti, D.; Delichère, P.; Geantet, C. Appl. Catal., B 2011, 101, 246-255. 63. Melián, E. P.; Díaz, O. G.; Araña, J.; Rodríguez, J. M. D.; Rendón, E. T.; Melián, J. A. H. Catal. Today 2007, 129, 256-262. 64. Foo, G. S.; Wei, D.; Sholl, D. S.; Sievers, C. ACS Catal. 2014, 3180-3192. 65. Copeland, J. R.; Santillan, I. A.; Schimming, S. M.; Ewbank, J. L.; Sievers, C. J. Phys. Chem. C 2013, 117, 21413-21425. 66. Copeland, J. R.; Shi, X.-R.; Sholl, D. S.; Sievers, C. Langmuir 2012, 29, 581-593. 67. Ravenelle, R. M.; Copeland, J. R.; Van Pelt, A. H.; Crittenden, J. C.; Sievers, C. Top. Catal. 2012, 55, 162-174. 68. Jongerius, A. L.; Copeland, J. R.; Foo, G. S.; Hofmann, J. P.; Bruijnincx, P. C. A.; Sievers, C.; Weckhuysen, B. M. ACS Catal. 2013, 3, 464-473. 69. Olcese, R. N.; Bettahar, M.; Petitjean, D.; Malaman, B.; Giovanella, F.; Dufour, A. Appl. Catal., B 2012, 115, 63-73. 70. Nimmanwudipong, T.; Runnebaum, R. C.; Brodwater, K.; Heelan, J.; Block, D. E.; Gates, B. C. Energy Fuels 2014, 28, 1090-1096. 71. Nimmanwudipong, T.; Aydin, C.; Lu, J.; Runnebaum, R. C.; Brodwater, K. C.; Browning, N. D.; Block, D. E.; Gates, B. C. Catal. Lett. 2012, 142, 1190-1196. 72. Saidi, M.; Rahimpour, M. R.; Raeissi, S. Energy Fuels 2015, 29, 3335-3344. 73. Saidi, M.; Rostami, P.; Rahimpour, H. R.; Fallah, M. A. R.; Rahimpour, M. R.; Gates, B. C.; Raeissi, S. Energy Fuels 2015, 29, 4990-4997. 74. Lu, J.; Behtash, S.; Mamun, O.; Heyden, A. ACS Catal. 2015, 5, 2423-2435. 75. Massoth, F. E.; Politzer, P.; Concha, M. C.; Murray, J. S.; Jakowski, J.; Simons, J. J. Phys. Chem. B 2006, 110, 14283-14291. 76. de Souza, P. M.; Rabelo-Neto, R. C.; Borges, L. E. P.; Jacobs, G.; Davis, B. H.; Sooknoi, T.; Resasco, D. E.; Noronha, F. B. ACS Catal. 2015, 5, 1318-1329. 77. Dwiatmoko, A. A.; Lee, S.; Ham, H. C.; Choi, J.-W.; Suh, D. J.; Ha, J.-M. ACS Catal. 2015, 5, 433-437. 78. Bai, D.; Zhu, J. X.; Jin, Y.; Yu, Z. Chem. Eng. J. 1998, 71, 97-109. 79. Zhang, K.; Huang, C. H.; Zhang, H. B.; Xiang, S. H.; Liu, S. Y.; Xu, D.; Li, H. X. Appl. Catal., A 1998, 166, 89-95. 80. Bartholomew, C. H. Appl. Catal., A 2001, 212, 17-60. 81. Hemelsoet, K.; Van Speybroeck, V.; Waroquier, M. Chemphyschem 2008, 9, 2349-2358. 82. Neri, G.; Bonaccorsi, L.; Donato, A.; Milone, C.; Musolino, M. G.; Visco, A. M. Appl. Catal., B 1997, 11, 217-231. 83. Prins, R. Chem. Rev. 2012, 112, 2714-2738. 84. Im, J.; Shin, H.; Jang, H.; Kim, H.; Choi, M. Nat Commun 2014, 5:3370. 85. Westerhof, R. J. M.; Kuipers, N. J. M.; Kersten, S. R. A.; van Swaaij, W. P. M. Ind. Eng. Chem. Res. 2007, 46, 9238-9247. 86. Maier, S. M.; Jentys, A.; Lercher, J. A. J. Phys. Chem. C 2011, 115, 8005-8013. 87. Chen, C. Y. US Patent 6632765 B1, Oct 14, 2003.


ACS Catalysis

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 52 of 52

Table of Contents Figure


Steric Effect and Evolution of Surface Species in the

Recommend Documents