Vibrational Spectroscopic Characterization of Glycerol Reaction Pathways over Metal-Modified Molybdenum Carbide Surfaces

Article abstract of DOI:10.1002/cctc.201901629

As the production of biodiesel increases, it is economically preferable to upgrade the excess by-product, glycerol, into value-added products. Metal-modified molybdenum carbide (Mo₂C) is a class of promising catalysts for this application due to their promising hydrodeoxygenation (HDO) activity. In this work, Pt, Fe and Cu were used to modify the Mo₂C surface and tune the product selectivity of glycerol. Pt/Mo₂C showed reforming activity to produce syngas, Fe/Mo₂C cleaved all the C?O bonds of glycerol to produce propylene, while Cu/Mo₂C broke only one C?O bond of glycerol to form acetol. Consecutive temperature-programmed desorption (TPD) experiments demonstrated an enhanced stability of all surfaces when compared to Mo₂C. High-resolution electron energy loss spectroscopy (HREELS) characterization indicated that the Fe/Mo₂C surface weakened the C?O bonds of glycerol, while the Pt/Mo₂C surface showed no activity towards C?O bond cleavage. This study demonstrated that the Pt/Mo₂C, Fe/Mo₂C and Cu/Mo₂C surfaces were active, selective and stable for converting glycerol into syngas, propylene, and acetol, respectively. The combination of TPD and HREELS also served as an example of utilizing surface spectroscopy to identify the reaction pathways and intermediates on model catalyst surfaces. This should, in turn, provide insights to guide the design of metal-modified carbide catalysts for the upgrading of biomass-derived oxygenates.

Full text of DOI:10.1002/cctc.201901629

Accepted Article
Title: Vibrational spectroscopic characterization of glycerol reaction
pathways over metal-modified molybdenum carbide surfaces
Authors: Weiming Wan, Zhexi Lin, and Jingguang G. Chen
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201901629
Link to VoR: http://dx.doi.org/10.1002/cctc.201901629
A Journal of
www.chemcatchem.org

10.1002/cctc.201901629
ChemCatChem
Vibrational spectroscopic characterization of glycerol reaction
pathways over metal-modified molybdenum carbide surfaces
Dr. Weiming Wan, Zhexi Lin, Prof. Jingguang G. Chen*
Department of Chemical Engineering, Columbia University, New York, NY, 10027, USA
Corresponding author: jgchen@columbia.edu (J. G. Chen)
1
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
Abstract
As the production of biodiesel increases, it is economically preferable to upgrade the excess by-
product, glycerol, into value-added products. Metal-modified molybdenum carbide (Mo2C) is a
class of promising catalysts for this application due to their promising hydrodeoxygenation (HDO)
activity. In this work, Pt, Fe and Cu were used to modify the Mo2C surface and tune the product
selectivity of glycerol. Pt/Mo2C showed reforming activity to produce syngas, Fe/Mo2C cleaved
all the C-O bonds of glycerol to produce propylene, while Cu/Mo2C broke only one C-O bond of
glycerol to form acetol. Consecutive temperature-programmed desorption (TPD) experiments
demonstrated an enhanced stability of all surfaces when compared to Mo2C. High-resolution
electron energy loss spectroscopy (HREELS) characterization indicated that the Fe/Mo2C surface
weakened the C-O bonds of glycerol, while the Pt/Mo2C surface showed no activity towards C-O
bond cleavage. This study demonstrated that the Pt/Mo2C, Fe/Mo2C and Cu/Mo2C surfaces were
active, selective and stable for converting glycerol into syngas, propylene, and acetol, respectively.
The combination of TPD and HREELS also served as an example of utilizing surface spectroscopy
to identify the reaction pathways and intermediates on model catalyst surfaces. This should, in turn,
provide insights to guide the design of metal-modified carbide catalysts for the upgrading of
biomass-derived oxygenates.
2
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
1. Introduction
The upgrading of biomass-derived molecules has attracted significant research attention
recently.[1]–[7] Among the pathways to convert biomass, biodiesel production is a promising
approach for mitigating current environmental issues and reducing dependence on fossil fuels.[8]
In biodiesel production, glycerol is generated as a major by-product. For every 100kg of biodiesel,
10kg of glycerol is produced.[9] As biodiesel production continues to increase,[10] it is economically
preferable to convert the excess glycerol to value-added products. Various reactions have been
studied to explore the utilization of glycerol.[11] The reforming reaction breaks the C-H and C-C
bonds of glycerol to form syngas (CO + H2) for methanol production and Fisher-Tropsch
synthesis.[12]–[14] The complete hydrodeoxygenation (HDO) reaction can cleave all the C-O bonds
of glycerol to produce propylene, an industrially important building block.[15]–[17] The propylene
produced from glycerol can contribute to the on-purpose production to fill the current gap between
propylene market demand and supply.[18] On the other hand, the selective HDO reaction can break
one or two C-O bonds of glycerol to produce value-added C3 oxygenates such as acetol,[19][20]
propanediol,[21][22] and propanol.[23] To upgrade glycerol on an industrial scale, it is necessary to
design low-cost catalysts with high stability, activity, and selectivity. There have been many
research efforts on the upgrading of glycerol, but spectroscopic studies on the reaction pathways
and adsorbed intermediates on well-defined surfaces are still lacking. The understanding gained
from such fundamental studies is crucial to the rational design of efficient HDO catalysts.
Mo2C has ordered crystalline structure, extreme hardness, high melting temperature and electric
conductivity similar to metals.[24][25] The clean Mo2C surface has strong oxophilicity, which leads
to high activity for C-O bond cleavage. Therefore, Mo2C has been studied for CO2 reduction[26][27]
and furfural HDO reaction.[28][29] However, the oxophilicity of Mo2C is too strong, which makes
3
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
it difficult to remove surface oxygen or oxygen-containing surface intermediates, resulting in rapid
deactivation of the Mo2C surface.
Novel catalytic performance has been observed by tuning the oxophilicity of Mo2C with
admetals. For example, Ni-modified Mo2C surfaces show the reforming activity of ethanol[30] and
ethylene glycol,[31] Cu-modified Mo2C surfaces promote the dehydrogenation of ethanol,[30] and Co-
and Fe-modified Mo2C surfaces show HDO activity of furfural.[32][33] Though Mo2C has been
studied in the reactions of many oxygenates, very limited studies have focused on the reactions of
glycerol. Since glycerol contains two types of -OH groups (primary and secondary), it can also
serve as a model compound for the mechanistic study of selective C-O bond scission for larger
polyols.
In a previous study, the selective HDO reaction of glycerol was investigated on Cu-modified
Mo2C catalysts.[16] The combination of a strongly oxophilic site, Mo, and a weakly oxophilic sites,
Cu, can selectively break a different number of C-O bonds of glycerol, leading to the selective
production of propylene, allyl-alcohol, propanal, and acetol on Cu/Mo2C surfaces with different
Cu coverages. In this work, Pt/ Mo2C and Fe/Mo2C surfaces have been investigated to determine
the effect of metal modifiers in controlling the reaction pathways of glycerol. The results from Cu-
modified Mo2C surface were also referred to for comparison.[16] Consecutive TPD measurements
were used to determine the surface activity and stability, and vibrational spectroscopic studies
using high-resolution electron energy loss spectroscopy (HREELS) were performed to identify the
surface intermediates. This work reveals the distinct bond scission sequences of glycerol on
Pt/Mo2C, Fe/Mo2C and Cu/Mo2C and provides a guideline for designing catalysts for different
glycerol conversion pathways. The combination of TPD and HREELS also provides an example
of using surface spectroscopy to identify promising catalytic systems for biomass upgrading.
4
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
2. Methods
The TPD experiments were conducted in an ultra-high vacuum (UHV) chamber equipped with
a quadrupole mass spectrometer (UTI-100C), a cylindrical mirror analyzer for Auger analysis, a
metal source for physical vapor deposition and an electron gun for sputtering. The Mo(110) single
crystal (Princeton Scientific, 99.99%) was spot-welded to two tantalum posts for resistive heating.
A K-type thermocouple was spot-welded to the back of the crystal for temperature measurements.
The synthesis procedure for Mo2C was described previously.[34] Briefly, the Mo(110) single crystal
was cleaned by cycles of Ne+ sputtering and annealing, followed by oxygen treatment to remove
surface carbon. The carbide-modified Mo(110) was synthesized by cycles of exposing Mo(110) to
ethylene at 600K and subsequently annealing to 1200K until the Mo (186 eV):C (275 eV) Auger
Electron Spectroscopy (AES) atomic ratio reached approximately 2:1. Metal modifiers with
controlled coverage were deposited onto the as-synthesized Mo2C surface by the physical vapor
deposition (PVD) method. The metal coverage was 1 monolayer (ML) for all the metal modified
Mo2C surfaces and the coverage was quantified using the atomic ratio of admetal and substrate
from AES measurements. [35]
In a typical TPD experiment, the surface was exposed to 5 Langmuir (L, 1L = 10-6 Torr*s) of
H2 and 4L of glycerol at 200K and subsequently heated to 800K at 3K/s. The gas-phase products
desorbed from the surface were detected by mass spectrometry.
The HREELS experiments were performed in a separate UHV chamber equipped with a HREEL
spectrometer in addition to other capabilities described above in the TPD chamber. The HREEL
spectra were recorded with a primary beam energy of 6 eV with an incidence and reflection angle
of 60o with respect to surface normal. In a typical HREELS experiment, the surface was exposed
to the same amount of H2 and glycerol as in the TPD experiments. The surface was ramped to the
5
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
desired temperature, held for 1 minute and cooled down to 100K before the HREEL spectra were
recorded.
Glycerol (Fisher Scientific, 99%)) was transferred into a glass cylinder, purified with cycles of
freeze-pump-thaw and pre-heated to 400K before introducing glycerol into the chamber. Hydrogen,
oxygen, ethylene, neon, and carbon monoxide were of research purity and used without further
purification. The purity of each gas was verified by mass spectrometry.
3. Results and discussions
3.1.TPD studies of reaction pathways on metal modified Mo2C surfaces
TPD experiments were performed to investigate the reaction pathways of glycerol reaction on
various metal modified Mo2C surfaces. Propylene, CO, acetol and unreacted glycerol were
observed in the gas phase. Propylene and acetol are HDO products, CO is from both reforming
reaction and the recombinative desorption of surface carbon and oxygen. The desorption of H2 and
water were also observed.
6
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
Figure 1. TPD spectra of a) glycerol (m/z=61), b) propylene (m/z=39), c) CO (m/z=28) and d) acetol (m/z=43) with
an exposure of 4 L glycerol on hydrogen pre-dosed Mo2C, Cu/Mo2C, Fe/Mo2C, and Pt/Mo2C surfaces
Fig.1 summarizes the desorption of glycerol, propylene, CO, and acetol from Mo2C, Cu/Mo2C,
Fe/Mo2C, and Pt/Mo2C surfaces. The unreacted glycerol desorbed at 262 K (Fig.1(a)), and its
cracking patterns contributed to the peak at 262 K in all the other spectra. Propylene was produced
on both Fe/Mo2C and Mo2C surfaces (Fig.1(b)) with desorption temperatures of 354K and 471K,
respectively, indicating that both surfaces were active for the complete HDO reaction that broke
all C-O bonds of glycerol to form propylene. The lower propylene desorption temperature on
Mo2C suggested a higher C-O bond scission activity compared to that on Fe/Mo2C. The Pt/Mo2C
surface was active in reforming to produce H2 and CO. As shown in Fig. 1 (c), CO desorption peak
was observed on Pt/Mo2C with a peak temperature of 351K, indicating that the glycerol underwent
the reforming pathway. On the other hand, even though CO desorption peaks were also observed
on the other surfaces at a higher temperature, such desorption peaks were likely resulting from the
recombinative desorption of surface carbon and oxygen, which were produced from the total
decomposition of the unreacted surface adsorbates.
7
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
With the addition of Pt modifier on Mo2C, the reaction pathway of glycerol changed from HDO
to reforming. On Cu/Mo2C, glycerol only underwent one C-O bond scission to form acetol
(Fig.1(d)),[16] likely due to the lower oxophilicity of the Cu modifier. Comparing the reaction
pathways on these four surfaces (Scheme 1), Mo2C and Fe/Mo2C were able to cleave all the C-O
bonds in glycerol to produce propylene, Cu/Mo2C broke only one C-O bond of glycerol to form
acetol, while Pt/Mo2C favored the C-C and C-H bond scissions to produce CO and H2.
Scheme 1 – Reaction pathways of glycerol over the Mo2C, Pt/Mo2C, Fe/Mo2C, and Cu/Mo2C surfaces.
8
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
Figure 2. Sequential TPD spectra of a) propylene (m/z=39) and b) CO (m/z=28) with an exposure of 4 L glycerol on
hydrogen pre-dosed Mo2C, Fe/Mo2C and Pt/Mo2C surfaces
Consecutive TPD experiments were performed to further investigate the stability of the Pt/Mo2C,
Fe/Mo2C and Mo2C surfaces, the results of which were summarized in Figure 2. The unreacted
glycerol desorbed at 262 K and contributed to the peaks at 262 K in all other spectra. On Mo2C,
the first TPD spectrum showed a propylene desorption peak at 357 K (Fig.2(a)). The peak shifted
to 447 K in the second and third TPD experiments with decreased peak area. The lower peak area
and the higher desorption temperature indicated that the Mo2C surface was deactivated, most likely
by adsorbed carbon and oxygen. On Fe/Mo2C, propylene desorption was observed in the first TPD
experiment with a desorption temperature at 470 K. In the second and third TPD experiments, the
desorption temperature slightly shifted to 447 K and the peak areas of the three spectra were similar.
Comparing the first TPD spectra between Fe/Mo2C and Mo2C, the propylene spectrum from the
Fe/Mo2C surface had higher desorption temperature and smaller desorption peak than those of the
Mo2C surface, which suggested a weaker glycerol HDO activity on the Fe/Mo2C surface. However,
9
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
Fe/Mo2C showed better stability and higher activity than the Mo2C surface in the second and third
TPD spectra. The reason for the enhanced HDO stability of the Fe/Mo2C surface could be
attributed to the easier removal of surface oxygen. Under reactor condition, the surface oxygen
produced from the HDO reaction of glycerol could be removed by gas-phase H2.[36] While under
the UHV condition, due to the low partial pressure of gas-phase H2, surface oxygen was removed
by surface carbon.
On the other hand, no propylene was produced from Pt/Mo2C (Fig.2(b)). Instead, a strong CO
desorption peak at 350 K was observed in the first TPD experiment and this peak shifted to 332 K
in the subsequent experiments (Fig.2(c)), indicating a high reforming activity of the Pt/Mo2C
surface. Due to the low oxophilicity of Pt, surface oxygen could be easily removed. The similar
peak areas of CO desorption in three TPD cycles indicate good stability of the surface.
CO and propylene observed in the gas-phase were quantified using the following method. Since
the products of glycerol over Cu/Mo2C has been quantified previously, only two reaction pathways
regarding the Mo2C, Pt/Mo2C, and Fe/Mo2C surfaces were concerned here. The net reaction
pathways of glycerol on the three surfaces can be summarized as follows:
a C3H8O3 → a C3H6 + 3a O + a H2
b C3H8O3 → 3b CO + 4b H2
a = C3H6 yield
b = (CO yield)/3
The yield of propylene and CO were calculated by the following equations:
푠푎푡
퐶푂
퐶푂
푆
Propylene yield = 휃
푃
28 푎푚푢 = a → a = C3H6 yield
푔푙푦푐푒푟표푙
푝푟표푝푦푙푒푛푒
푠푎푡
퐶 퐻
3 6
푃
푆
퐶푂
39 푎푚푢
10
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
푠푎푡
퐶푂
CO yield = 휃
푃
퐶푂
푔푙푦푐푒푟표푙
푠푎푡
푃
퐶푂
Where the 푃푔푙푦푐푒푟표푙 and 푃푔푙푦푐푒푟표푙 are TPD areas of the 39 amu and 28 amu peaks from the
푝푟표푝푦푙푒푛푒
퐶푂
glycerol reaction. The CO desorption at 700 K was not considered as the glycerol reforming
퐶 퐻
39 푎푚푢
퐶푂
28 푎푚푢
3
6
reaction. 푆
and 푆
are mass spectrometer sensitivities for 28 amu and 39 amu,
푠푎푡
퐶푂
푠푎푡
퐶푂
respectively. 푃 was the peak area of 28 amu from saturation CO desorption and 휃 was the
saturation coverage of CO on Mo2C.
The quantification results (Table 1) show that the fresh Mo2C surface had a propylene yield of
0.072 in the first TPD experiment and the yield was reduced to 0.021 and 0.014 in the subsequent
experiments. On the other hand, the Fe/Mo2C surface, with better stability, had a propylene yield
of 0.04 in the three TPD experiments. Pt/Mo2C also showed high stability in glycerol reforming
reaction and the yield slightly increased in the third TPD experiment.
Table 1. Reactivity of glycerol on different surfaces (molecules per metal atom)
Mo2C
Fe/Mo2C
Pt/Mo2C
HDO reforming HDO reforming HDO reforming
1st
2nd
3rd
0.07
0.02
0.01
0
0
0
0.04
0.04
0.04
trace
0
0
0
0.05
0.05
0.06
0
0
11
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
3.2.Vibrational studies of surface intermediates of glycerol reactions
Figure 3. HREELS spectra of glycerol on hydrogen pre-dosed Fe/Mo2C at 300 K, 400 K,500 K, and 700 K
To identify the changes of surface intermediates as a function of temperature, a thermal
sequence of HREELS measurements were performed. Figure 3 shows the temperature dependence
of the spectra after exposing Fe/Mo2C to glycerol. When the temperature increased from 300 K to
400 K, the intensity of each peak remained relatively constant. The 흂(CO) mode at 1040 cm-1
became broader, suggesting that some C-O bonds were weakened on the surface. As the
temperature increased from 400 K to 500 K, the intensity of the 흂(CO) mode at 1040 cm-1
decreased significantly, indicating the C-O bond scission of glycerol. This was consistent with the
desorption of propylene starting at 400 K, as shown in the TPD experiments. As temperature
further increased to 700 K, the intensity of the peaks associated with the CH species at 1403 cm-1
and 2859 cm-1 decreased significantly, suggesting the desorption of most of the hydrocarbon
12
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
products. The peak at 537 cm-1 increased, which indicates a strong M-O bond from adsorbed
atomic oxygen on the surface.
Figure 4. HREELS spectra of glycerol on four surfaces at a) 300 K, b) 400 K
Fig.4 compares the surface intermediates on the four surfaces. The vibration mode assignments
are summarized in Table 2. Since unreacted glycerol desorbed at 262 K (Fig.1), the spectra at 300
K (Fig.4(a)) showed the geometries of chemisorbed glycerol or partially reacted intermediates.
The vibrational modes from the metal-oxygen bond appeared on Mo2C, Fe/Mo2C, and Pt/Mo2C,
suggesting the cleavage of O-H bond of molecular glycerol. On Cu/Mo2C, the absence of the
metal-oxygen bond suggested that the O-H bonds of glycerol were at least partially preserved. The
휈(CO) mode on Pt/Mo2C at 1067 cm-1, was higher than that on Fe/Mo2C and Mo2C at 1040 cm-1,
suggesting a stronger C-O bond of the adsorbed glycerol on Pt/Mo2C. The weakened C-O bond
on Fe/Mo2C and Mo2C facilitated C-O bond scission, leading to the formation of propylene. The
13
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
detection of the 휈(C=O) modes at 1859 cm-1 and 1946 cm-1 indicated the formation of CO. The
higher intensity of the 휈(C=O) mode on Pt/Mo2C was consistent with the TPD results that Pt/Mo2C
showed higher reforming activity (Fig.2(c)).
Table 2. Vibrational mode assignment for adsorbed glycerol molecule
Frequency (cm-1)
Mode
Mo2C
416
Fe/Mo2C
403
Pt/Mo2C
436
휹(푴푶)
흂(푴푶)
570
--
--
흂(CO)
1040
1329
1409
2859
3403
1040
1329
1409
2859
--
1067
1329
1409
2859
3403
흆w(CH2)
휹(CH2)
흂as(CH)
흂(OH)
Abbreviation: 훿 – deformation, 휈 – stretching, 휌w - wagging, as-asymmetric
As the surface was further annealed to 400 K (Fig.4(b)), the intensities of the 휈(CO) mode (1067
cm-1) and 휈(C=O) mode (1946 cm-1) on Pt/Mo2C were reduced, consistent with the high reforming
activity toward glycerol on Pt/Mo2C and with the CO desorption peak at 332 K(Fig.2(c)). On the
Fe/Mo2C surface, the intensity of the 휈(CO) mode (1040 cm-1) did not show significant change,
indicating that the C-O bonds had not been cleaved. In contrast, on the Mo2C surface, the intensity
of the 휈(CO) mode was reduced, confirming the C-O bond cleavage and being consistent with the
lower onset temperature of the propylene desorption from the fresh Mo2C surface than on Fe/Mo2C
(Fig.1(b)). On Cu/Mo2C, the decrease of the v(CO) mode intensity was consistent with the
formation of C-O bond scission products and the desorption of acetol, as reported previously.[16]
14
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
In summary, the HREELS results indicated that Mo2C interacted with the oxygen in glycerol more
strongly than Fe/Mo2C and Cu/Mo2C, and Pt/Mo2C showed reforming activity.
4. Conclusions
This work combined TPD and HREELS measurements to compare the reaction pathways and
adsorbed intermediates for the reaction of glycerol on Fe-, Pt- and Cu-modified Mo2C surfaces.
The Pt-modified Mo2C surface showed reforming activity, while the Mo2C, Fe-and Cu-modified
Mo2C surfaces showed HDO activity. On Mo2C and Fe/Mo2C, all the C-O bonds of glycerol were
cleaved to form propylene. The Fe modifier slightly reduced the surface oxophilicity of Mo2C,
which improved surface stability. The Cu modifier reduced the oxophilicity of Mo2C to a further
extent such that only one C-O bond of glycerol was broken to form acetol. This work revealed that
different metal modifiers could be used to tune the selectivity of the Mo2C catalyst to convert
glycerol into industrially important chemicals, such as syngas, propylene or acetol. In addition,
using glycerol as a model compound, the reaction pathways revealed in this work could provide
information on the bond-breaking sequences for larger polyols. The combination of TPD and
HREELS also demonstrated the feasibility of using spectroscopic methods to identify promising
catalytic systems for upgrading biomass-derived oxygenates.
Acknowledgements
This work was sponsored by the National Science Foundation under Grant No. CHE-1565964.
15
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
References
[1]
[2]
[3]
[4]
[5]
[6]
J. N. Chheda, G. W. Huber, J. a. Dumesic. Angew. Chemie - Int. Ed. 2007, 46 (38), 7164.
H. Wang, J. Male, Y. Wang. ACS Catal. 2013, 3 (5), 1047.
A. M. Robinson, J. E. Hensley, J. Will Medlin. ACS Catal. 2016, 6 (8), 5026.
Z. Lin, R. Chen, Z. Qu, J. G. Chen. Green Chem. 2018, 20 (12), 2679.
D. M. Alonso, S. G. Wettstein, J. A. Dumesic. Chem. Soc. Rev. 2012, 41 (24), 8075.
Y. Kwon, K. J. P. Schouten, J. C. Van Der Waal, E. De Jong, M. T. M. Koper. ACS Catal.
2016, 6, 6704.
[7]
[8]
L. Du, Y. Shao, J. Sun, G. Yin, C. Du, Y. Wang. Catal. Sci. Technol. 2018, 8 (13), 3216.
L. Lin, Z. Cunshan, S. Vittayapadung, S. Xiangqian, D. Mingdong. Appl. Energy 2011, 88
(4), 1020.
[9]
P. U. Okoye, A. Z. Abdullah, B. H. Hameed. Renew. Sustain. Energy Rev. 2017, 74, 387.
[10] F. Bauer, C. Hulteberg. Biofuels, Bioprod. Biorefining 2013, 7 (1), 43.
[11] S. Bagheri, N. M. Julkapli, W. A. Yehye. Renew. Sustain. Energy Rev. 2014, 41, 113.
[12] P. D. Vaidya, A. E. Rodrigues. Chemical Engineering and Technology. 2009, pp 1463–
1469.
[13] A. M. Karim, C. Howard, B. Roberts, L. Kovarik, L. Zhang, D. L. King, Y. Wang. ACS
Catal. 2012, 2 (11), 2387.
[14] I. Esteve-Adell, B. Crapart, A. Primo, P. Serp, H. Garcia. Green Chem. 2017, 19 (13),
16
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
3061.
[15] V. Zacharopoulou, E. S. Vasiliadou, A. A. Lemonidou. Green Chem. 2015, 17 (2), 903.
[16] W. Wan, S. C. Ammal, Z. Lin, K.-E. You, A. Heyden, J. G. Chen. Nat. Commun. 2018, 9
(1), 4612.
[17] E. Gomez, S. Kattel, B. Yan, S. Yao, P. Liu, J. G. Chen. Nat. Commun. 2018, 9 (1), 1398.
[18] A. Akah, M. Al-Ghrami. Appl. Petrochemical Res. 2015, 5 (4), 377.
[19] A. K. Kinage, P. P. Upare, P. Kasinathan, Y. K. Hwang, J.-S. Chang. 2010.
[20] C.-W. Chiu, M. A. Dasari, G. J. Suppes, W. R. Sutterlin. AIChE J. 2006, 52 (10), 3543.
[21] S. Zhu, X. Gao, Y. Zhu, J. Cui, H. Zheng, Y. Li. Appl. Catal. B Environ. 2014, 158–159,
391.
[22] M. Akiyama, S. Sato, R. Takahashi, K. Inui, M. Yokota. Appl. Catal. A Gen. 2009, 371
(1–2), 60.
[23] X. Lin, Y. Lv, Y. Xi, Y. Qu, D. L. Phillips, C. Liu. Energy and Fuels 2014, 28 (5), 3345.
[24] W. Wan, B. M. Tackett, J. G. Chen. Chem. Soc. Rev. 2017, 46, 1807.
[25] S. T. Oyama. Catal. Today 1992, 15 (2), 179.
[26] M. D. Porosoff, X. Yang, J. A. Boscoboinik, J. G. Chen. Angew. Chemie - Int. Ed. 2014,
53 (26), 6705.
[27] S. K. Kim, Y.-J. Zhang, H. Bergstrom, R. Michalsky, A. Peterson. ACS Catal. 2016, 6,
2003.
17
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
[28] K. Xiong, W. S. Lee, A. Bhan, J. G. Chen. ChemSusChem 2014, 7 (8), 2146.
[29] Y. Shi, Y. Yang, Y. W. Li, H. Jiao. ACS Catal. 2016, 6 (10), 6790.
[30] T. G. Kelly, J. G. Chen. 2014, 777.
[31] W. Yu, M. Salciccioli, K. Xiong, M. A. Barteau, D. G. Vlachos, J. G. Chen. ACS Catal.
2014, 4 (5), 1409.
[32] Z. Lin, W. Wan, S. Yao, J. G. Chen. Appl. Catal. B Environ. 2018, 233, 160.
[33] W. Wan, Z. Jiang, J. G. Chen. Top. Catal. 2018.
[34] B. Frühberger, J. G. Chen. J. Am. Chem. Soc. 1996, 118 (46), 11599.
[35] P. J. Cumpson, M. P. Seah. Surf. Interface Anal. 1997, 25, 430.
[36] H. Ren, W. Yu, M. Salciccioli, Y. Chen, Y. Huang, K. Xiong, D. G. Vlachos, J. G. Chen.
ChemSusChem 2013, 6 (5), 798.
18
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901629
ChemCatChem
Table of Content
The combination of TPD and HREELS identified the various reaction pathways and intermediates
of glycerol on metal modified Mo2C surfaces. Glycerol was converted to acetol on Cu/Mo2C, and
to syngas on Pt/Mo2C. Propylene was produced from glycerol on both Fe/Mo2C and Mo2C, and
Fe/Mo2C had better stability due to the modified surface oxophilicity.
19
This article is protected by copyright. All rights reserved.