Sustainable hydrocarbon production via simultaneous condensation-hydrodeoxygenation of propionic acid with furfural over red mud-supported noble metal catalysts

Article abstract of DOI:10.1016/j.catcom.2018.11.015

A new catalytic system for single-step alkane synthesis from furans and carboxylic acids was investigated using noble metals-supported red mud (RM) catalysts. Maximum alkane selectivities were 100%, 87%, and 56% over Pt/RM, Pd/RM, and Ru/RM respectively. Potassium doping prolonged the catalyst life by suppressing methanation reaction, decreased coking, and enhancing the stability of noble metals towards oxidation. Five functionalities were accomplished on the modified catalysts: partial reduction of propionic acid on noble metals; moderating coking reactions on K, suppressing methanation on K, ketonization of acid on RM, and condensation-hydrodeoxygenation reaction on RM. Magnetite played significant role in catalyzing ketonization reaction.

Full text of DOI:10.1016/j.catcom.2018.11.015

Accepted Manuscript
Sustainable hydrocarbon production via simultaneous
condensation-hydrodeoxygenation of propionic acid with furfural
over red mud-supported noble metal catalysts
Antonie Castille, Claudie Bessette, Francois Thomas, Mania
Etemad
PII:
S1566-7367(18)30393-5
DOI:
Reference:
https://doi.org/10.1016/j.catcom.2018.11.015
CATCOM 5559
To appear in:
Catalysis Communications
Received date:
Revised date:
Accepted date:
27 August 2018
10 November 2018
25 November 2018
Please cite this article as: Antonie Castille, Claudie Bessette, Francois Thomas, Mania
Etemad , Sustainable hydrocarbon production via simultaneous condensation-
hydrodeoxygenation of propionic acid with furfural over red mud-supported noble metal
catalysts. Catcom (2018), https://doi.org/10.1016/j.catcom.2018.11.015
This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.

ACCEPTED MANUSCRIPT
Sustainable hydrocarbon production via simultaneous condensation-hydrodeoxygenation
of propionic acid with furfural over red mud-supported noble metal catalysts
Antonie Castille, Claudie Bessette, Francois Thomas, Mania Etemad
French Environment and Energy Management Agency ADEME, France

ACCEPTED MANUSCRIPT
Abstract
A new catalytic system for single-step alkane synthesis from furans and carboxylic acids was
investigated using noble metals-supported red mud (RM) catalysts. Maximum alkane
selectivities were 100%, 87%, and 56% over Pt/RM, Pd/RM, and Ru/RM respectively.
Potassium doping prolonged the catalyst life by suppressing methanation reaction, decreased
coking, and enhancing the stability of noble metals towards oxidation. Five functionalities were
accomplished on the modified catalysts: partial reduction of propionic acid on noble metals;
moderating coking reactions on K, suppressing methanation on K, ketonization of acid on RM,
and condensation-hydrodeoxygenation reaction on RM. Magnetite played significant role in
catalyzing ketonization reaction.
Keywords: red mud, hydrocarbon synthesis, condensation-hydrodeoxygenation, biomass,
potassium-doped catalyst, noble metals

ACCEPTED MANUSCRIPT
1. Introduction
Red mud (RM) is a caustic waste (pH ~ 10-15) from aluminium industry. RM consists of
metal oxides including Fe2O3, Al2O3, SiO2, TiO2, ZrO2, and some trace metals such as Pd, V, Sr,
and Ru [1-4]. About 77 million tons of red mud are produced annually in France, which is one of
the most pressing disposal problems in the aluminium industry [5, 6]. In most countries where red
mud is produced, it is pumped into holding ponds [7]. Farming or construction can no longer take place
on the land, once it's used for RM, even after the mud is dry. In October 2010, approximately one
million cubic meters of this RM were accidentally released in Hungary that caused a severe
disaster for human life as well as aquatic life in that region. Since 1996, over 20 million tons of
this mud have been pumped into a 2400 square kilometer section of the seabed in the
Cassidaigne submarine canyon (France) [5, 6].
RM has been studied as catalyst for different processes such as hydrogen production,
pyrolysis, hydrogenation, liquefaction, exhaust gas clean-up, and hydrodechlorination reactions
[8-11]. Although RM is reported as an effective pyrolysis catalyst, it produces large amount of
small oxygenated molecules which are soluble in water. 10-30% of aqueous phase pyrolysis oil
is consisted of small oxygenated compounds that have no viable applications and are a
significant source of net carbon loss in pyrolysis-hydrodeoxygenation process. These low
molecular weight oxygenated compounds mostly include C1 to C5 aldehydes and ketones, as well
as C1 to C4 carboxylic acids and furans that are partially/completely soluble in the aqueous phase
pyrolysis oil. These compounds can potentially increase the cost of wastewater treatment
because of their toxic and corrosive nature.

ACCEPTED MANUSCRIPT
An attempt to design a sustainable catalytic process for continuous synthesis of alkanes
from these oxygenates has been made in this work based on a new red mud-supported nickel
catalyst (Ni/RM) that was firstly developed by Jahromi and Agblevor (2018) [12]. The Ni/RM
catalyst deactivated after three 30-minute batches (overall 90 minutes) due to coking and
formation of nickel iron oxide, however, unlike commercial Ni/SiO2-Al2O3, the Ni/RM catalyst
was regenerable [13]. These researchers discovered that red mud in reduced form (RRM) can
catalyze hydrocarbon production reactions from the reaction of furans and anisole with
aldehydes and ketones [14, 15] but the catalyst life was too short to be considered applicable in
continuous industrial processes. Although noble metals are more costly than traditional Ni, Co,
and Mo hydrodeoxygenation catalysts, these metals are less active for polymerization reactions,
thus, produce less coke [16-19]. In this short communication paper we report the effect of Pt, Pd,
and Ru and subsequent potassium doping on HDO activity of RM-supported catalysts with the
ultimate objective of prolonging the catalyst useful life in continuous process rather than batch
process as well as minimizing hydrogen consumption by inhibiting the undesired reactions. The
influence of reduction temperature on RM activity was also investigated. The ultimate goal of
this manuscript is to produce liquid hydrocarbons from biomass oxygenated compounds.
2. Experimental
2.1. Materials and catalyst preparation
Reagent grade furfural and propionic acid were obtained from Merck (Germany) to
represent derivatives of cellulose and hemicellulose respectively in aqueous phase pyrolysis oil.
Platinum (IV) nitrate, palladium nitrate dehydrate, and ruthenium (III) nitrosyl nitrate solution

ACCEPTED MANUSCRIPT
were purchased from Merck (Germany) as Pt, Pd, and Ru sources respectively for catalyst
preparation. Calculated amount of these salts were added to 50 ml deionized water to prepare the
catalyst precursor using wet impregnation method according to wet impregnation method
described by Jahromi and Agblevor [12, 13].
RM support was calcined at 580 °C for 4 hours, ground mechanically, and passed
through series of four sieves with mesh sizes of 90, 180, 270, and 360 μm. The portion with
particle size between 90 and 180 μm was used for catalyst preparation. Metal analysis of RM
support was performed using an Agilent 7700 ICP-MS (Agilent Technologies, USA). In the
catalyst screening stage, RM-supported catalysts with noble metal loadings of 0.01, 0.05, 0.1,
0.2, and 0.3 wt. % were prepared and denoted as yX/RM where y indicates the metal loading and
X shows the noble metal (Pt, Pd, or Ru). The catalysts were reduced prior to hydrodeoxygenation
(HDO) experiments in a tubular reactor using a gas mixture of 5%H2 in helium for 6 hours.
Scanning Electron Microscopy (SEM) of catalysts were recorded on a JEOL NEOSCOPE JCM-
6000 Plus using EDS detector to obtain surface morphology of catalyst samples. pH of catalyst
aqueous slurries and reaction medium was measured using a METTLER TOLEDO pH meter.
The effect of reduction temperature on the activity of the RM support was studied at 400,
500, and 600 °C (samples were labeled RRM-400, RRM-500 and RRM-600 respectively). These
samples were also analyzed via 57Fe Mössbauer Spectroscopy to determine the relative
abundance of iron redox states (hematite, magnetite, and zero valent iron). This method is further
outlined in Tishchenko et al. (2015) [20]. Temperature-programmed reduction (TPR) was
conducted using a u-tube reactor linked to a TCD detector. Potassium-doped catalysts were
prepared by wet impregnation of the catalyst precursor sample with an aqueous solution of

ACCEPTED MANUSCRIPT
potassium nitrate with the K:X atomic ratio of 4:1. The resulting powders were dried overnight at
373 K and finally heat treated at 773 K for 1 h in a N2 stream before reduction.
2.2. Hydrodeoxygenation experiments and product analysis
HDO reactions were performed in a continuous packed bed reactor system custom
designed by Parr Instrument Company (USA). The reactor consists of a stainless-steel tube with
inner diameter of 1.8 cm and length of 20 cm. Catalysts were loaded into the top of the reactor,
supported by a stainless-steel screen and quartz wool. The reaction temperature (350 °C) was
controlled by a Thermcraft Lab- Temp 1760-watt tube furnace powered by a Parr 4875 Power
Controller. Aqueous model compound mixtures (equimolar) were injected downward into the
reactor tube at a rate of 1.5 ml/min using a Lab Alliance HPLC pump to achieve mean residence
time of about 30 minutes. Through a pre-heated (~250 °C) side stream, hydrogen was co-fed at
flow rate of 10 ml/min with reaction mixture while the operating H2 pressure was maintained at
1400 psi using a secondary HPLC pump and a backpressure regulator was installed in reactor
outlet to control the pressure. Conversion, and product selectivity were determined according to
the method described by Jahromi and Agblevor [14, 21]. Liquid and gas products were analyzed
using a Shimadzu GC/MS-QP5000 (Shimadzu, USA) and a 436 SCION GC-TCD (SCION
Instruments, USA) respectively. Sampling was carried out using an online configuration through
two separate sampling lines. The gas release line was directly connected to the GC, while the
liquid line passed through a series of three ice-chilled condensers prior to the HPLC inlet port.
Sample analyses were performed in 20-min intervals. Hydrogen consumption was calculated
according to the following equation [14, 21]:

ACCEPTED MANUSCRIPT
(
)
퐻2 푐표푛푠푢푚푝푡푖표푛 푚표푙푒 = (푛푖 − 푥푓 . 푛푓 )
(1)
퐻2
퐻2
푡표푡
Where 푛푖 is the inlet number of mole of hydrogen, 푥푓 is the outlet mole fraction of hydrogen,
퐻2
퐻2
and 푛푓 is the total number of mole of gas at the end of experiment.
푡표푡
3. Results and discussion
3.1. Effect of reduction temperature
ICP analysis of RM is provided in Table S1 (supporting information). The major metal
components of RM were Al, Ca, Fe, Na, Si, and Ti. Oxides of these metals have been reported to
be promising supports for HDO catalysts. TPR profile of RM support showed two unresolved
reduction peaks between 300 and 500 °C with maximum hydrogen uptake at about 410 °C (Fig.
1) and another broad reduction peak between 500 and 700 °C with maximum hydrogen uptake at
about 600 °C. These peaks could be due to reduction of hematite and magnetite respectively [12,
13]. Impregnation of RM support with Pt, Pd, and Ru resulted a new reduction peak to appear
between 450 and 500 °C (Fig. 1), thus, this peak was attributed to a possible interaction between
noble metals and RM components, because TPR of Pt oxide, Pd oxide, and Ru oxide showed
sharp reduction peaks at lower temperatures (Fig. S1 in supporting information). Addition of
noble metals caused the first two RM reduction peaks (between 300 and 500 °C) to either
decrease or disappear depending on the noble metal. Therefore, this could suggest another
different type of interaction between these metals and the support material. Based on the TPR
profiles (Fig. 1), the effect of reduction temperature (400, 500, and 600 °C) on catalytic activity
was investigated only on the support material because noble metal reduction peaks appeared in
relatively the same temperature ranges. SEM images of RM and noble metal catalysts (Fig. S2)

ACCEPTED MANUSCRIPT
showed great dispersion of noble metals on support surface. The average metal particle size of
Pt, Pd, and Ru were 35, 43, and 71 nm respectively. These results appeared to suggest that Pt
formed the smallest particles among the three metals which is further confirmed by BET surface
area results as explained later in the manuscript.
For ketonization and condensation-hydrodeoxygenation (CONHDO) reactions, aqueous
solution of propionic acid and aqueous solution of furan/propanal (equimolar) were used as
reaction mixtures respectively, because RM catalyzed ketonization and CONHDO reactions [14,
15]. Reactant(s) conversion and product selectivity of these experiments are documented in
Table 1. The conversion of propionic acid for RRM-400 was 25% with 50% selectivity to its
corresponding ketone (3-pentanone). The only gas product was CO2 [14, 15]. Furan and propanal
conversion were both 35% on RRM-400 over a period of 30 minutes with 100% selectivity to n-
heptane. RRM-500 was excellent in catalyzing both ketonization and CONHDO reactions with
100% conversion (Table 1), however, RRM-600 did not catalyze ketonization although it was as
effective as RRM-500 in catalyzing CONHDO reaction (Table 1). This result could suggest that
ketonization reaction was catalyzed by an oxide that was possibly reduced to elemental metal at
600 °C. Regardless of the reduction temperature, all RRM samples catalyzed only ketonization
and/or CONHDO reactions and no other side reactions took place on RRM. In summary, RRM-
400 was weak in catalyzing ketonization and CONHDO reactions, RRM-500 was excellent in
catalyzing both ketonization and CONHDO reactions, and RRM-600 only catalyzed CONHDO
(Table 1).
Mössbauer spectra for RM indicated that all iron in the raw red mud existed as hematite
(Fig. 2). After reduction at 400 °C, hematite was partially reduced to magnetite. Increasing
reduction temperature to 500 °C resulted in complete reduction of hematite to magnetite and

ACCEPTED MANUSCRIPT
partial reduction of the magnetite to elemental iron. After reduction at 600 °C, a larger portion of
the magnetite was reduced to zero valent iron and only small amount of magnetite was
conserved. Therefore, Fe3O4 probably had a great contribution to catalyze ketonization reaction
[22]. In a different experiment we evaluated the ketonization of propionic acid over Fe3O4 (Table
S2 in supporting information). Although magnetite showed high catalytic activity (81%
conversion) under similar experimental conditions, RM was more effective for this reaction
(100% conversion). Thus, in addition to magnetite, other components of RM such as MgO and
ZrO2 could possibly participate in ketonization of propionic acid [23-25]. The increase in BET
specific surface area with increase in reduction temperature (Table 1) was attributed to enhanced
loss of oxygen at higher temperatures. The BET specific surface area of RRM-400, RRM-500,
and RRM-600 were 68, 73, and 90 m2/g respectively.
3.2. Catalyst screening
We screened a series of catalysts represented by different levels of noble metal loadings
(0.01, 0.05, 0.1, 0.2, and 0.3 wt. %). All catalyst samples were reduced at 500 °C because the
RM support had the highest level of activity (for ketonization and CONHDO reactions) at 500°C
and the reduction of all three metals occurred between 450 and 500 °C (Fig. 1). Increase in
catalyst BET specific surface area with increase in metal loading was expected (as presented in
Table 2) because noble metals appeared to form nano particles on RM support with average
particle size between 35-75 nm. No conversion of reactants were observed in absence of catalyst
so we concluded that the reactor body did not have any influence on the reactions (Table 2). As
presented in Table 2 propionic acid conversion was 100% in all experiments because of
ketonization reaction. Increasing the noble metal content from 0.01 to 0.1 wt. % improved the

ACCEPTED MANUSCRIPT
catalytic activity and hydrocarbon (n-heptane and 3-ethyloctane) selectivity in the case of all
three noble metals. However, metal loadings of 0.2 and 0.3 wt. % did not have a significant
effect on hydrocarbon selectivity. At these metal loading levels the selectivity to furfuryl alcohol
(FOL), furan, and 2-methylfuran (2-MF) slightly increased that caused increased hydrogen
consumption. Overall, 0.1%Pt/RM showed best results with total hydrocarbon selectivity of
100%. 0.1%Pd/RM and 0.1%Ru/RM catalysts gave 87% and 56% selectivity to hydrocarbons
(n-heptane and 3-ethyloctane selectivities combined). For these catalysts, 3-pentanone and 2-MF
were detected in product stream. This could suggest that CONHDO catalyzing component(s)
deactivated more rapidly over Pd/RM and Ru/RM than Pt/RM. This conclusion was further
confirmed by coke determination. The coke yields on 0.1Pt/RM, 0.1Pd/RM, and 0.1Ru/RM were
2.8%, 6.6%, and 8.5% of the initial feed material respectively, thus, coke formation could be the
main factor for deactivation of CONHDO-catalyzing components.
It was interesting to note that, in contrast to Ni/RM [14], no ring hydrogenation took
place on noble metal catalysts because no THF and no methyl-THF were detected in the product
stream. This phenomena was beneficial in both reducing the hydrogen consumption and
increasing the hydrocarbon yield. In the same reaction of furfural and propionic acid, Ni/RM
consumed 1.32 mole hydrogen but 0.1Pt/RM consumed 0.82 mole H2 which was about 30% less
than the value reported for Ni/RM [14]. In addition to hydrogenation of double bonds, Ni/RM
increased gasification and methanation that can increase the overall hydrogen consumption. The
hydrocarbon selectivity on Ni/RM was about 72 % (n-heptane 38.1%, 3-ethyloctane 34.7%) [14],
whereas this value for 0.1Pt/RM catalyst increased to 100% (58% n-heptane, 42% 3-ethyloctane)
(Table 2).

ACCEPTED MANUSCRIPT
In order to understand the reaction mechanism, reduction intermediates (furfuryl alcohol,
furan, and 2-methylfuran) were exposed to HDO reaction on noble metal catalysts (for more
details, see Table S3 in supporting information). Also when propionic acid was reacted over
noble metal catalysts in absence of furfural, the only liquid products were propanal and 3-
pentanone. The reaction pathways of furfural and propionic acid was therefore proposed based
on the analysis of products and literature review. These pathways are depicted in Fig. 3. Pathway
1 is partial hydrogenation of the carbonyl group of furfural on Pt, Pd, or Ru to produce furfuryl
alcohol (FOL). FOL was then reduced to furan or 2-methylfuran (2-MF) on the noble metal
(pathways 2 and 3 in Fig. 3 respectively). On the other hand, propionic acid was either
hydrogenated to propanal (pathway 4) on noble metals, or ketonized to 3-pentanone on RM
support (pathway 5). Furan and 2-MF reacted with propanal and 3-pentanone respectively on
RM, and produced n-heptane and 3-ethyloctane respectively (pathways 6 and 7 respectively).
The major gas products of these reactions were CO2 (due to ketonization of propionic acid) and
methane.
3.3. Effect of potassium doping
As 0.1Pt/RM gave the best selectivity to liquid hydrocarbons, potassium doping results
are reported only for this catalyst, even though it had similar effect on Pd/RM and Ru/RM as
well. Trends of reactants conversion, and hydrocarbon and methane selectivity versus time on
stream for Pt/RM and K-Pt/RM are showed in Fig. 4. The catalytic activity of Pt/RM for furfural
and acetic acid conversion died out in 100 min and 120 min respectively, while over the K-
PT/RM catalyst the catalyst life for these conversions significantly increased to 380 min and 400
min respectively. Hydrocarbon production time was 80 min and 360 min over Pt/RM and K-

ACCEPTED MANUSCRIPT
Pt/RM catalysts respectively (Fig. 4). It was interesting to note that, by addition of K, the average
selectivity to methane decreased by more than 50%. Furthermore, in contrast to bare Pt/RM, the
selectivity to 3-ethyloctane was significantly higher than n-heptane in all sampling points (Fig.
S3 in supporting information) which appeared to suggest that K additive suppressed
demethylation of 2-MF or dehydroxylation of furfuryl alcohol. Hence, the major methanation
reaction was probably due to pathway 2 (Fig. 3) that was suppressed to some extents in presence
of potassium. Inhibition of methanation reaction by potassium doping was in agreement with
literature [26].
After complete catalyst deactivation, TPR analysis of the used catalysts were carried out
(Fig. S4 in supporting information) which showed partial oxidation of Pt, however, Pt oxidation
in K-Pt/RM was about 30% less than that of bare Pt/RM. Therefore it appeared that, in addition
to suppressing methanation reaction, K additive could enhance the resistance of noble metals
towards oxidation, enhancing the catalytic stability. Carbon balance comparison over 0.1Pt/RM,
0.1Pd/RM, 0.1Ru/RM, and K-Pt/RM catalysts is reported in Table S4 in supporting information.
Carbon loss in the form of coke was clearly reduced by about 45% after K addition. As presented
in Table S4 in supporting information, carbon recovery in liquid form increased from 67.3% (on
Pt/RM) to 86.4 (on K-Pt/RM). These results could suggest that coke formation and oxidation
were possibly the main deactivation mechanisms of RM and noble metal respectively.
Overall, this research showed that by modifying RM-supported catalysts from transition
metals to noble metals, it is possible to narrow down the reaction pathways in order to inhibit the
production of undesired products such as THF and methyl-THF [14] (which increases the final
hydrocarbon yield) and decrease the amount of required hydrogen compared to previous studies

ACCEPTED MANUSCRIPT
using Ni/RM [14, 21]. Moreover, potassium doping could potentially help this novel process to
make a significant transition from laboratory test tubes to industrial scales.
4. Conclusion
Continuous mode hydrocarbon production from the reaction of furfural and propionic
acid over multifunctional bare and modified red mud-supported noble metal catalysts were
investigated. Ketonization of propionic acid was catalyzed on RM support, partial hydrogenation
of furfural (to furan and 2-methylfuran) occurred on noble metal, and subsequent simultaneous
condensation-hydrodeoxygenation (CONHDO) of furanic compounds with aldehydes and
ketones took place on the support material. An investigation of reduction temperature effect on
catalytic activity of RM support revealed that RRM-400 weakly catalyzed ketonization and
CONHDO reactions, and RRM-500 was excellent in catalyzing both reactions, whereas RRM-
600 lost its activity for ketonization due to reduction of magnetite to elemental iron. Among Pt,
Pd, and Ru catalysts Pt showed 100% selectivity to liquid hydrocarbons (n-heptane and 3-
ethyloctane). Potassium doping increased the catalyst useful life from 120 min to 400 min by
reducing coke formation, suppressing methanation, and enhancing the stability of noble metal
towards oxidation. This study provided a viable catalytic system for sustainable hydrocarbon
production from biomass-derived oxygenated compounds, however, preventing coke formation
must be studied more extensively.

ACCEPTED MANUSCRIPT
Acknowledgement
The authors graciously thank Mathews Miller for his invaluable contributions of time and
effort in analyzing materials and liquid compositions. We also thank Dr. Christa Candle for
supplying the red mud. Support for this research was provided in part by French Environment
and Energy Management Agency ADEME and a research initiative grant from EPF Graduate
School of Engineering.

ACCEPTED MANUSCRIPT
References
[1] W.G. Shim, J.W. Nah, H.Y. Jung, Y.K. Park, S.C. Jung, S.C. Kim, Recycling of red mud as a
catalyst for complete oxidation of benzene, J. Ind. Eng. Chem. 60 (2018) 259–267.
[2] L.C.A. De Oliveira, A.P. Heitmann, L.D. Almeida, S. Herman, A.A. Mansur, C.S. De Castro,
U.F. De Alfenas, Red mud as catalysts for oxidative desulfurization reactions, J. Environ. Sci. 57 (2015)
3–5.
[3]
production, J. Environ. Sci. (China). 25 (2013) 823–829.
[4] S. Hamid, S. Bae, W. Lee, Novel bimetallic catalyst supported by red mud for enhanced nitrate
reduction, Chem. Eng. J. 348 (2018) 877–887.
[5] S. Patel, B.K. Pal, Current status of an industrial wasteꢀ: red mud an overview, Ijltemas (Issn
2278-2540). IV (2015) 1–16.
Q. Liu, R. Xin, C. Li, C. Xu, J. Yang, Application of red mud as a basic catalyst for biodiesel
[6]
V. Dentoni, B. Grosso, G. Massacci, Environmental sustainability of the alumina industry in
Western Europe, Sustainability 6 (2014) 9477–9493.
[7]
of red mud in China, J. Clean. Prod. 84 (2014) 606–610.
[8] S. Sushil, V.S. Batra, Catalytic applications of red mud, an aluminium industry waste: A review,
Appl. Catal. B Environ. 81 (2008) 64–77.
[9] J. Yang, X. Xu, S. Liang, R. Guan, H. Li, Y. Chen, B. Liu, J. Song, W. Yu, K. Xiao, H. Hou, J.
W. Liu, X. Chen, W. Li, Y. Yu, K. Yan, Environmental assessment, management and utilization
Hu, H. Yao, B. Xiao, Enhanced hydrogen production in catalytic pyrolysis of sewage sludge by red mud:
Thermogravimetric kinetic analysis and pyrolysis characteristics, Int. J. Hydrogen Energy. 43 (2018)
7795–7807.

ACCEPTED MANUSCRIPT
[10]
J. Song, J. Yang, S. Liang, Y. Shi, W. Yu, C. Li, X. Xu, J. Xiao, R. Guan, N. Ye, X. Wu, H. Hou,
J. Hu, J. Hu, B. Xiao, Red mud enhanced hydrogen production from pyrolysis of deep-dewatered sludge
cakes conditioned with Fenton’s reagent and red mud, Int. J. Hydrogen Energy. 41 (2016) 16762–16771.
[11]
mud, J. Hazard. Mater. 81 (2001) 103–114.
[12] H. Jahromi, F.A. Agblevor, Hydrotreating of guaiacol: A comparative study of Red mud-
supported nickel and commercial Ni/SiO2-Al2O3catalysts, Appl. Catal. A Gen. 558 (2018) 109–121.
[13] H. Jahromi, F.A. Agblevor, Hydrodeoxygenation of pinyon-juniper catalytic pyrolysis oil using
red mud-supported nickel catalysts, Appl. Catal. B, Environ. 236 (2018) 1–12.
[14] F.A. Agblevor, H. Jahromi, Aqueous phase synthesis of hydrocarbons from furfural reactions
with low molecular weight biomass oxygenates, Energy and Fuels. 32 (2018) 8552-8562.
[15] F.A. Agblevor, H. Jahromi, Aqueous phase synthesis of hydrocarbons from reactions of guaiacol
and low molecular weight oxygenates, ChemCatChem. (2018). 10.1002/cctc.201800982.
S. Ordóez, H. Sastre, F. V. Díez, Catalytic hydrodechlorination of tetrachloroethylene over red
[16]
[17]
E. Furimsky, Catalytic hydrodeoxygenation, Appl. Catal. A Gen. 199 (2000) 147–190.
S. Ahmadi, E. Reyhanitash, Z. Yuan, S. Rohani, C. (Charles) Xu, Upgrading of fast pyrolysis oil
via catalytic hydrodeoxygenation: Effects of type of solvents, Renew. Energy. 114 (2017) 376–382.
[18] M. Mohammad, T. Kandaramath Hari, Z. Yaakob, Y. Chandra Sharma, K. Sopian, Overview on
the production of paraffin based-biofuels via catalytic hydrodeoxygenation, Renew. Sustain. Energy Rev.
22 (2013) 121–132.
[19]
H. Pourzolfaghar, F. Abnisa, W. Mohd, A. Wan, M.K. Aroua, Atmospheric hydrodeoxygenation
of bio-oil oxygenated model compounds: A review, J. Anal. Appl. Pyrolysis. 133 (2018) 117–127.

ACCEPTED MANUSCRIPT
[20]
V. Tishchenko, C. Meile, M.M. Scherer, T.S. Pasakarnis, A. Thompson, Fe2+catalyzed iron atom
exchange and re-crystallization in a tropical soil, Geochim. Cosmochim. Acta. 148 (2015) 191–202.
[21] H. Jahromi, F.A. Agblevor, Hydrodeoxygenation of aqueous-phase catalytic pyrolysis oil to
liquid hydrocarbons using multifunctional nickel catalyst, Ind. & Eng. Chem. Res. 57 (2018) 13257–
13268.
[22] J.A. Bennett, C.M.A. Parlett, M.A. Isaacs, L.J. Durndell, L. Olivi, A.F. Lee, K. Wilson, Acetic
Acid Ketonization over Fe3O4/SiO2 for Pyrolysis Bio-Oil Upgrading, ChemCatChem. 9 (2017) 1648–
1654.
[23]
Mechanisms, catalysts, and implications for biomass conversion, Catalysis 3 (2013) 2456-2473.
[24] R. Kumar, N. Enjamuri, S. Shah, A.S. Al-Fatesh, J.J. Bravo-Suárez, B. Chowdhury, Ketonization
of oxygenated hydrocarbons on metal oxide based catalysts, Catal. Today. 302 (2018) 16–49.
[25] R.A.L. Baylon, J. Sun, K.J. Martin, P. Venkitasubramanian, Y. Wang, Beyond ketonization:
T.N. Pham, T. Sooknoi, S.P. Crossley, D.E. Resasco, Ketonization of carboxylic acidsꢀ:
Selective conversion of carboxylic acids to olefins over balanced Lewis acid-base pairs, Chem. Commun.
52 (2016) 4975–4978.
[26]
X. Hu, D. Dong, X. Shao, L. Zhang, G. Lu, Steam reforming of acetic acid over cobalt catalysts:
Effects of Zr, Mg and K addition, Int. J. Hydrogen Energy. 42 (2017) 4793–4803.

ACCEPTED MANUSCRIPT
Fig. 1: TPR patterns of RM support, 0.1Pt/RM, 0.1Pd/RM, and 0.1Ru/RM catalysts
Fig. 2: 57Fe Mössbauer spectra of the unprocessed RM (top spectra), RM reduced at 400 °C
(RM-400) (second from top spectra), RM reduced at 500 °C (RM-500) (second from bottom
spectra), and RM reduced at 600 °C (RM-600) (bottom spectra).
Fig. 3: Reaction pathways of propionic acid-furfural mixture on Pt, Pd, and Ru supported on
RM.
Fig. 4: Effect of potassium doping on catalytic activity of Pt/RM catalyst. Triangle markers show
conversion and square markers show selectivity. In figure legend: FAL= Furfural, P.A=
Propionic Acid, CH4= Methane, H.C= Hydrocarbon, C= Conversion, S= Selectivity, Pt/RM=
0.1Pt/RM catalyst, and KPt/RM= potassium-doped 0.1Pt/RM catalyst.

ACCEPTED MANUSCRIPT
Table 1: Effect of reduction temperature on catalytic activity of RM support.
SBET
pH*
(catalyst slurry)
pH**
(reaction media)
Catalyst
Conversion (%)
Product selectivity (dry basis) (%)
(m2//g)
Ketonization reaction
Propionic acid
3-pentanone
RRM-400
RRM-500
RRM-600
68
73
90
8.6
7.5
6.4
5.7
4.5
3.1
25
100
0
100
100
0
CONHDO reaction
Furan
35
100
100
Propanal n-heptane
RRM-400
RRM-500
RRM-600
68
73
90
8.6
7.5
6.4
8.1
7.3
6.2
35
100
100
100
100
100
* pH of aqueous catalyst slurry.
** Initial pH of the reaction mixture.

ACCEPTED MANUSCRIPT
Table 2: Preliminary experiments for catalyst screening at 350°C and reaction time 30 min
(catalysts were reduced at 500°C).
pH**
(reactio
n
H2
Conversion (%)
Selectivity (%)
SB2ET
(m /g
)
pH*
(catalys
t slurry)
consumptio
n
(mole)
0
3-
2-
M
F
n-
heptan
e
-
3-
Catalyst
Propioni
c acid
Furfura
l
Propana
l
FO
L
Fura
n
pentanon
e
-
ethyloctan
e
-
media)
No
-
-
0
0
-
-
-
-
catalyst
RRM-500
0.01Pt/RM
0.05Pt/RM
0.1Pt/RM
0.2Pt/RM
0.3Pt/RM
73
79
92
118
135
143
76
7.5
7.2
6.9
7.1
7.0
6.8
7.3
5.4
5.2
5.5
5.1
5.0
4.9
5.2
0
100
100
100
100
100
100
100
0
38
73
100
100
100
21
0
100
31
42
0
0
0
0
28
7
0
0
0
0
0
0
0
0
10
0
0
0
0
0
0
8
0
0
9
0.21
0.45
0.82
0.87
0.90
0.18
21
11
0
0
0
11
23
58
58
56
7
17
42
42
44
6
0
33
0.01Pd/R
M
17
19
0.05Pd/R
M
85
7.0
5.2
0.34
100
43
14
30
18
4
6
17
11
0.1Pd/RM
0.2Pd/RM
0.3Pd/RM
0.01Ru/R
M
103
118
121
81
6.8
6.8
6.9
7.0
4.8
5.3
5.0
4.8
0.67
0.71
0.74
0.15
100
100
100
100
81
83
84
15
0
0
0
11
4
2
1
15
9
7
7
39
2
2
2
14
3
3
2
12
51
51
52
5
36
35
36
4
0.05Ru/R
M
94
6.7
4.9
0.27
100
27
8
29
24
7
8
14
10
0.1Ru/RM
0.2Ru/RM
0.3Ru/RM
98
109
117
6.6
7.1
6.9
5.4
4.9
5.2
0.53
0.55
0.58
100
100
100
53
56
56
0
0
0
22
23
26
13
13
11
5
4
4
4
4
3
37
36
36
19
20
20
* pH of aqueous catalyst slurry.
** Initial pH of the reaction mixture.

ACCEPTED MANUSCRIPT
Graphical abstract
Highlights
 New noble metal catalysts supported on red mud are synthesized.
The catalysts catalyze ketonization, alkylation, and hydrodeoxygenation simultaneously.

 Reduction temperature has a significant influence on red mud activity.
 Potassium doping prolongs the catalyst life by three times.