Abstract-- Rhenium modified Pt/TiO2 catalyst has been tested during erythritol (ERY) hydrogenolysis reaction in a slurry reactor. N2 physisorption, temperature programmed reduction and CO chemisorption characterizations were performed. After reduction treatment, Pt species were fully reduced with a dispersion of 26% whereas Re species remained partially oxidized. ERY may convert through four pathways: isomerization, dehydration and C-O or C-C hydrogenolysis. Pt-Re/TiO2 was active and selective to butanetriols and butanediols formation. The influence of temperature (423–498 K) and reactant concentrations (15-35 barH2 and 0.2-0.6 MERY) on catalytic activity and products distribution were evaluated; the activation energy and the reaction orders for each route were estimated. Dehydration pathway displayed the highest activation energy whereas isomerization showed the lowest value. The order to ERY and H2 were 0.23 and 0.97 respectively, for C-O hydrogenolysis route and is consistent with two different mechanisms proposed for the removal of primary or secondary OH group.
Keywords-- bimetallic catalysts – C-O hydrogenolysis – Kinetic study – C4 polyols.
I. INTRODUCTION
The study of lignocellulosic biomass and derivatives as raw materials for the production of valuable compounds has gained great interest in the last decades. Among this biomass derived molecules can be found a C4 polyol: erythritol (ERY, C4H10O4), that is a stable sugar alcohol widely present in fruits, fermented foods, and animals. Erythritol has a O/C=1 ratio, and therefore the excess of oxygen must be removed to produce fuels and valuable chemicals. One of the most interesting methods to remove oxygen atoms is deoxygenation (Nakagawa et al., 2020), involving the dissociation of C−O bonds, and replacing it by C−H bonds. For this reaction, the use of some reducing agents like molecular hydrogen (H2) is required. Deoxygenation of alcohols using molecular hydrogen is called hydrodeoxygenation or C−O hydrogenolysis and typically, supported noble metals (Pt, Ru, Rh, and Ir) are capable of catalyze these reactions. A large number of works have been published using Rh or Ir catalysts together with a promoter oxide (Re or Mo) for C-O hydrogenolysis of polyols such as glycerol (GLY) or ERY. The effect of the support, the addition of acid co-catalysts, the metal content and the reaction conditions strongly affect the product distribution since several competitive routes usually get involved when an aqueous polyol solution and H2 reacts in presences of metal catalysts (Amada et al., 2012; Said et al., 2017; Tomishige et al., 2014; Virgilio et al., 2021). In a previous work (Virgilio et al., 2021), TiO2 was selected as the most convenient support for Ir-based catalysts, to favor C-O scissions of ERY to render butanediols (BDO).
Pt catalysts were less studied for polyols hydrogenolysis. Actually, Pt-Re/SiO2 (4%wt. Pt, Re/Pt=0.25) resulted less active than Ir-Re/SiO2 during GLY hydrogenolysis at 373 K although the products distribution were similar forming mainly 1,3-propanediol (13PDO) (Amada et al., 2011). Some discrepancy was found when comparing the catalytic activity of Pt-W catalysts during GLY hydrogenolysis. Liu et al. (2019) informed that W-modified Pt (5-38 %wt. of W and 1-9 %wt. of Pt) catalysts supported on different solids displayed very low conversion (less than 5%) at 403-493 K and 10-80 barH2 resulting unattractive when comparing with Ir catalysts. On the other hand, it was also reported a 13PDO selectivity of 46% for GLY conversion of 70% on Pt-WOx/ZrO2 (3 %wt. Pt and 10 %wt. W) at 403 K and 40 barH2 (Gong et al., 2010). Finally, Daniel et al. have reported that Pt-Re (5.5%wt of each metal) supported on carbon have been selective in the conversion of GLY to PDO at 443 K and 40 barH2, while monometallic Pt/C only favored the breaking of C-C bonds (Daniel et al., 2010).
Regarding ERY hydrogenolysis, it has been reported that Pt-WOx/SiO2 can produce 1,4-butanediol (14BDO) from ERY at 413 K and 80 barH2 (40% selectivity and 83% conversion) at 24 h (Liu et al., 2020) but the time evolution of products were not informed. Moreover, rhenium promoted Pt has not been tested in this reaction according to our best knowledge.
Herein, we present the catalytic performance of Pt-Re/TiO2 in ERY conversion under different reaction temperatures and reactant concentrations. Furthermore, we estimate kinetic parameter as activation energy and reaction orders of routes involved in the ERY transformation.
II. EXPERIMENTAL SECTION
A. Catalyst preparation and characterization
Pt-Re/TiO2 catalyst was prepared by sequential incipient wetness impregnation with Ir and Re aqueous solution. Prior impregnation, the support (TiO2, Degussa P-25) was treated in air flow at 723 K for 4 h. Then, the support was impregnated with H2PtCl6 and HReO4 in order to obtain 2%wt Pt and Re/Pt=1 molar ratio. After each impregnation the solid was dried at 373 K for 16 h. Finally, the catalyst was treated in air at 723 K for 4 h. The bimetallic Pt-Re/TiO2 catalyst was characterized by N2 physisorption, temperature programmed reduction (TPR) and CO chemisorption. The specific surface area (SBET) was measured by N2 physisorption at 77 K in a Micromeritics Model ASAP 2020 apparatus. The TPR was carried out in a fixed-bed reactor using H2 diluted in Ar, in the range from 298 K to 773 K and a mass spectrometry unit (Hiden Analytical QGA). Pt0 dispersion and average particle size were determined by CO chemisorption (CO/Pt = 1) at 298 K in vacuum (10-5 torr) (Virgilio et al., 2021).
B. Catalytic tests
A slurry batch reactor (Parr 4565) was used for catalytic experiments. The catalyst was reduced ex-situ in H2 flow before reaction and then was loaded into the reactor with appropriate amount of water. The reactor was purged with N2 at 6 bar for 10 min and heated up to 373 K. An aqueous erythritol solution was injected into the reactor to ensure an ERY concentration of 0.2 M, 0.4 M and 0.6 M into the reactor, then the temperature was raised to desired value (448 K, 473 K and 498 K) under stirring. Finally, the desirable H2 pressure was admitted (15 bar, 25 bar and 35 bar). Liquid products and unconverted ERY were analyzed periodically by HPLC. A Shimadzu 20A HPLC with refractive index detector (RID-20A) was used with a Bio-Rad Aminex HPX-87H ion exchange column and a flow rate of 0.6 mL min-1 (H2SO4 5 mM).
Initial reaction rate (
) of each
reaction path was determined from the slope at t=0 of liquid products
concentration (
) vs time
curve for each experiment:
(1)
where 
is the
catalyst concentration. It has been ensured that the
reaction was carried out under chemical control conditions. Mears (CM),
Ramachandran (CR) and Weisz-Prater (CWP) Criterions were
used to verify the absence of diffusional limitations (CM=1.6·10-5,
CR=0.097 and CWP=0.085).
The conversion of erythritol (
) was calculated
as:
(2)
where 
is the initial
ERY concentration and 
is the ERY
concentration at a t time of the reaction.
The carbon balance (CB) is defined by Eq. 3, where 
and 
are number of
C atom in product i or ERY:
(3)
III. RESULTS AND DISCUSSION
A. Catalysts characterization
BET surface of the support (50 m2 g-1)
and the bimetallic catalyst (46 m2 g-1) did not show a
significant change, indicating that the impregnations and thermal treatments
did not affect the TiO2 textural properties. TPR profile showed
three H2 consumption peaks: a low temperature peak (˂ 373 K), a
broad band between 373 K and 523 K, and a peak centered at 613 K. A reduction
temperature of
Figure 1: Reaction network for erythritol hydrogenolysis.
613 K was chosen in order to ensure that the noble metal is in metallic state whereas the oxide promoter is partially reduced, similarly to Ir-Re/TiO2 (Virgilio et al., 2021).
Pt dispersion estimated from CO chemisorption of the catalyst reduced at 613 K was 26% (particle size: 3.5 nm) suggesting a good dispersion of Pt on the support with a small coverage with Re species.
B. Catalytic results
In previous work of ERY hydrogenolysis using bimetallic M-Re/TiO2
(M: Ir or Rh) catalysts (Gu et al., 2021; Virgilio et al., 2020), it has been reported four main reactions routes (Fig. 1). ERY may be
isomerized to threitol (TRE) by a reversible reaction. Hydrogenolytic cleavages
can occur between C-C or C-O bonds. C-C bond scissions lead to compounds with 2
and 3 carbon atoms such as GLY, ethylene glycol (EG) and PDO, whereas C-O hydrogenolysis
forms butanetriols (BTO), BDO and butanols (BuOH). Furthermore, acid sites may
favor dehydration to cyclic compounds such as 1,4-anhydroerythritol or
tetrahydrofuran (THF). Finally, some products can be transferred to the gas
phase under the reaction conditions (butane, methanol, propanol, methane and
THF). Figure 2A shows ERY conversion and liquid products concentrations (by
route) vs reaction time at 473 K and 25 barH2 for an ERY initial
concentration of 0.6 M. Pt-Re/TiO2 was active for the transformation
of the reactant under those conditions, achieving XERY = 19% after 1
h and 67% at 8 h of reaction. Furthermore, the CB de-
Figure 2: Temporal evolution of (A) ERY conversion and products concentrations [(!) XERY (%), (") CB (%), (7) CCO (M), (x) CDH (M), (L) CCC (M) and (,) CISO (M)] and (B) C-O hydrogenolysis products [() CBTO (M), (7) CBDO (M) and (b) CBuOH (M)]. [473 K, 0.6 M ERY, 25barH2 and 12.5 gc L-1].
creased from 95% to 70% during reaction as some products transfered to gas phase. Concentrations of liquid products are also shown in Fig. 2A. C-O scission products were the most favored followed by cyclic products (dehydration route), with concentration of 116 mM and 51 mM at 8 h respectively. In contrast, at low reaction time, TRE (isomerization route) was the product with highest concentration (17 mM at 0.5 h), while in the 4-8 h range its concentration remained almost constant (≈ 34 mM) due to the reversible reaction ERY ↔ TRE (Virgilio et al., 2021). C-C bond cleavage products were the less favored, achieving 18 mM at 8 h. In Fig. 2B the concentration of C-O hydrogenolysis products (BTO, BDO and BuOH) vs. time are shown. BTO and BDO concentrations were similar for the first 2 h of reaction (≈ 20 mM) but while BDO monotonically increased over the time reaching 78 mM at 8h (13% yield), BTO concentration increase up to 6 h and then remain almost constant (≈ 30 mM). The breaking of C-O bonds in ERY leads to the formation of 123BTO or 124BTO depending on whether the C-O scission proceeds on terminal or secondary OH, respectively. Although the concentration of both isomers were similar during the first 2 h, and therefore also their initial rates, between 2 and 8 h the concentration of 124BTO was higher than 123BTO in good agreement with results using Re modified Ir or Rh catalysts (Virgilio et al., 2020, 2021). 12BDO and 23BDO were the predominant isomers formed and their concentrations were similar throughout the reaction. Considering that 12BDO and 13BDO isomers can be formed from both BTO, while 23BDO only comes from 123BTO and 14BDO only from 124BTO, a rapid conversion: 123BTO→ 23BDO explain the lower concentration of 123BTO after 2 h. BuOH concentration was low during the 8 h of reaction.
Activity of Pt-Re/TiO2 catalyst on ERY
hydrogenolysis at different reaction temperatures, ERY initial concentration
and H2 pressure are displayed in Fig. 3A-C. Figure 3A shows the ERY
conversion in the range 448 K – 498 K. The XERY
increased with reaction temperature as expected. Total
conversion was achieved at 8 h at the
