Selective Hydrogenation of Stearic Acid Using Mechanochemically Prepared Titania-Supported Pt and Pt–Re Bimetallic Catalysts

: A range of Pt and Pt − Re catalysts on titania were mechanochemically prepared and compared to conventional catalysts made by wet impregnation for the selective hydrogenation of stearic acid. Mechanochemically prepared catalysts showed superior selectivity toward stearyl alcohol compared to the conventional catalysts. It was possible to achieve 100% selectivity to stearyl alcohol using mechanochemically prepared Pt/TiO 2 .


■ INTRODUCTION
−12 The mechanochemical preparation of catalysts is currently far less developed than the more traditional metal impregnation methods such as incipient wetness, wet impregnation, and coprecipitation techniques.Mechanochemical catalyst preparation brings with it potential advantages over traditional synthesis by conventional solvent-based techniques such as solvent-less catalyst synthesis and higher metal dispersion on the support.−15 For example, in the hydrocarbon-selective catalytic reduction of NO x with noctane as the reducing agent using Ag/Al 2 O 3 catalysts, the Ag/ Al 2 O 3 catalyst prepared by milling was significantly more active than the conventionally prepared catalyst.The catalyst prepared by milling had a light off temperature (temperature at which 50% NO x conversion is achieved) of 240 °C, while the conventionally prepared catalyst had a light off temperature of 390 °C.
Current environmental concerns over the use of solvents now make it particularly timely to investigate mechanochemical synthesis more extensively than has previously been done.In practical terms, mechanochemical preparation can be as simple as grinding two solids together using a pestle and mortar, but it also extends to high-energy milling equipment such as ball mills of various kinds for increased reproducibility and potential scale-up.A distinct advantage of the mechanochemical preparation of catalysts is that it can be achieved under solvent-free conditions, decreasing the environmental impact of heterogeneous catalyst synthesis.Catalysts can also be directly prepared from metal oxides rather than metal nitrates, which avoid the generation of NO x during the calcination step in conventional supported metal catalyst preparation.The energy consumption of milling does need to be considered; however, it is generally believed to be a lowenergy ball milling process.In this study, the conventional Pt/ TiO 2 and Pt−Re/TiO 2 catalysts were prepared by traditional wet impregnation in methanol using metal nitrates as precursors.The mechanochemical synthesis of catalysts was done using metal oxides as precursors.
−18 It can also be used to synthesize industrially important fatty alcohols from their corresponding carboxylic acid or esters.Fatty alcohols are typically long-chain alcohols and are widely used in cosmetics, lubricants, resins, and perfumes.−21 The fatty alcohol market is projected to grow from 5.4 billion USD in 2020 to 7 billion USD by 2025 at a compound annual growth rate of 5.2% during the forecast period. 22he hydrogenation of carboxylic acids to alcohols is difficult as the CO group has weak polarizability and hence low activity compared with CO bonds in ketones and aldehydes, for example.The process, therefore, typically requires harsh reaction conditions.For example, copper chromate catalysts operating in the range of 523−573 K and 13−21 MPa have been reported. 23Other catalysts that do not contain chromium have also been developed; however, harsh reaction conditions are still required, for example, Ru−Sn/Al 2 O 3 (523 K, 5.6−8 MPa), 24 Re 2 O 7 (433−583 K, 5−27 MPa), 25 and ReO x /SiO 2 catalysts (413 K, 8 MPa). 26Tomishige et al. 26 investigated the hydrogenation of carboxylic acids using bimetallic Pd or Pt− ReO x /SiO 2 catalysts.The presence of Pd was found to increase the conversion of stearic acid and also the selectivity to stearyl alcohol to 95%.However, the presence of Pt increased the selectivity to 93 from 85% with a monometallic ReO x /SiO 2 catalyst.The Pd/SiO 2 catalyst was not active in the hydrogenation of stearic acid.The activity of ReO x /SiO 2 + Pd/SiO 2 was found to be marginally higher than that of ReO x / SiO 2 .Hence, it was deduced that a direct interaction of Pd and Re was required to promote both the rate and selectivity.It has been suggested that the hydrogen species activated on the noble metal is supplied to the ReO x desorbing site which affects the reaction rate.The enhanced selectivity has been suggested to be related to the low hydrogenolysis activity of Pd.
The promotional effects of Re have also been observed by Dumesic et al. 27 The conversion of glycerol to synthesis gas was investigated over carbon-supported platinum and platinum−rhenium catalysts.At low pressures of CO, the addition of an equimolar amount of Re to a Pt/C catalyst resulted in a 5-fold increase in the production rate of CO.The authors have suggested that the effect of Re is to weaken the interaction of CO with the catalyst surface, hence decreasing the coverage of CO which allows the catalyst to operate at high rates in the presence of CO.Temperature-programmed reduction (TPR) results indicated that a Pt−Re alloy was present in the catalyst.
Previously, we have shown that the hydrogenation of a range of carboxylic acids using conventionally prepared Pt/TiO 2 and Pt−Re/TiO 2 catalysts under much milder reaction conditions (333−403 K, 0.5−2 MPa) is possible. 28Over a Pt/TiO 2 catalyst, the hydrogenation of the carboxylic acid group proceeds via the interaction of the oxygen atom of the carbonyl group with oxygen vacancies created by hydrogen spillover by the Pt metal.This leads to a high selectivity of an alcohol product.However, in a Pt−Re bimetallic catalyst, the oxygen atom of the carbonyl group is thought to interact with the Re cations, which leads to decarboxylation and formation of an alkane product.In general, a metal Pt site is needed for the H 2 dissociation, but we have shown previously that Re can be oxidized or reduced to be active at least for the amide hydrogenation. 29n the present study, the catalysts were prepared mechanochemically using a planetary ball mill with PtO 2 or Pt(COD)Me 2 (COD = cis,cis-1,5,-cyclooctadiene) as platinum precursors and Re 2 O 7 or Re 2 (CO) 10 as rhenium precursors.In a typical synthesis, the required weight of Pt and Re precursors and TiO 2 (Hombikat UV-100) was placed into a 500 cm 3 sintered aluminum oxide grinding jar with seven 10 mm diameter sintered alumina grinding balls.Milling was performed in a Retsch PM100 planetary ball mill at a rotation speed of 150 rpm for 1 h.The resultant powders were then calcined at 500 °C for 4 h.As a benchmark, conventional catalysts were also prepared using the wet impregnation technique.The required weight of Pt(NO 3 ) 4 , Re(OH) 4 , and TiO 2 (Hombikat UV-100) was stirred in methanol for 24 h.Methanol was then removed by rotary evaporation, and the catalyst was dried at 120 °C for 4 h and calcined at 500 °C for 4 h.The catalysts were thoroughly characterized using X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), high transmission electron microscopy (TEM), and N 2 sorption techniques.
The evaluation of catalytic activity was performed using the selective hydrogenation of stearic acid as a model reaction.Liquid phase-selective hydrogenation reactions were carried out in a 100 cm 3 Autoclave Engineers' high-pressure reactor, which had a pressure range of 0−20 MPa and a maximum temperature of 573 K.In a typical experiment, catalysts were prereduced in situ before the reaction.0.5 g of the catalyst was suspended in 40 mL of dodecane, and the reactor was purged three times with N 2 .The mixture was agitated at 1500 rpm and heated to 393 K.After purging with H 2 three times, the reactor was pressurized to 2 MPa, and the catalyst was reduced for 1 h.After prereducing the catalyst, the reactor was charged with 1.42 g (0.05 mol) of stearic acid.As mentioned above, the reactor was purged with N 2 followed by H 2 and heated to 403 K, while the impeller was set to 1500 rpm.At the desired temperature, the reactor was pressurized to the reaction pressure, which corresponded to t = 0.The reaction was monitored by sampling at regular time intervals, with analysis of the samples performed using a gas chromatograph equipped with a Zebron ZB Wax capillary column and a flame ionization detector.
The XRD measurements in this work were carried out using a PANanalytical X'Pert Pro X-ray diffractometer.The X-ray source was copper with a wavelength of 1.5405 Å.All measurements were carried out ex situ using a spinning stage.The diffractograms were recorded from 4 to 75°with a step size of 0.017°.The catalysts were analyzed following calcination and prior to use.
The surface area, total pore volume, and average pore diameter were measured by N 2 adsorption−desorption isotherms at 77 K using Micromeritics ASAP 2010.The pore size was calculated on the adsorption branch of the isotherms using the Barrett−Joyner−Halenda method, and the surface area was calculated using the Brunauer−Emmett−Teller (BET) method.The catalysts were analyzed following calcination and prior to use.
XPS was carried out using a Kratos Axis Ultra X-ray photoelectron spectrometer with a monochromated Al Kα (hν = 1486.6eV) X-ray source.The catalysts were analyzed as received and were not subject to pretreatments prior to analysis.Survey spectra were recorded at a pass energy of 160 eV with a step size of 1 eV and a dwell time of 50 ms.Corelevel scans were acquired at a pass energy of 20 eV with a step size of 0.05 eV and a dwell time of 100 ms.The catalysts were analyzed following calcination and prior to use (Table 1).
The turnover frequency (TOF) was then calculated from the dispersion and the initial rate of reaction as shown in eq 1.

TOF
initial rate molar mass of platinum platinum dispersion = × (1) TEM analysis was performed using a JOEL 2100 electron microscope at an operating voltage of 200 kV.Samples were prepared by suspending a small amount of the material in methanol and sonicating for 15 min.The samples were then transferred to a Cu grid with a holey carbon membrane.The samples were left to dry overnight.The average nanoparticle size was calculated using ImageJ software by measuring the size of 100 nanoparticles and taking an average.The catalysts were analyzed following calcination and prior to use.The average diameter of Pt atoms was determined using ImageJ software.We were unable to differentiate between Pt and Re atoms.The dispersion of Pt was estimated from the particle size by assuming that the Pt nanoparticles are spherical.
The XRD analysis showed peaks corresponding to the anatase phase of TiO 2 (Figure 1).There is no notable difference between the XRD patterns of the ball-milled   catalysts and those prepared by wet impregnation.The ballmilled catalysts were found to have lower surface areas; in particular, the ball-milled catalyst prepared with Pt(COD)Me 2 and Re 2 CO 10 had a N 2 BET surface area of 71 m 2 g −1 (Table 2).Catalysts prepared at higher milling speeds exhibited significant reductions in the surface area, for example, a bimetallic catalyst prepared by milling at 150 rpm had a surface area of 124 m 2 g −1 , while the same catalyst prepared at 400 rpm had a surface area of 40 m 2 g −1 .
Figure 2 shows the Pt 4f and Re 4f core levels of monometallic Pt and bimetallic Pt−Re catalysts.
Catalysts prepared through ball milling comprise primarily of ReO 3 , with minor contribution of Re 2 O 7 and ReO 2 .The Re 4f of the ball-milled catalysts displays complex core levels with multiple oxidation states present.Re 7+ is observed at a B.E. of 46.3 eV and Re 4+ at a B.E. 42.6 eV, consistent with the presence of Re 2 O 7 and ReO 2 , respectively.The catalyst prepared from the oxide precursor (Figure 2g) has a greater contribution of ReO 2 compared to that from the organometallic precursor (Figure 2h).
Both ball-milled catalysts contain a third peak centered at a B.E. of 43.9 eV, which resides in between the binding energies reported for Re 6+ and Re 4+ oxides.A few cases of binding energies in this range are reported in Re alloys and have been tentatively assigned to Re 5+ . 33,34However, the multiple oxidation states which give rise to overlapping Re 4f 7/2 and 4f 5/2 peaks of the various oxides preclude a definitive assignment in this range.Differences between the 4f spectra of the ball-milled catalysts and the conventionally prepared catalyst may also be attributed to support interactions, 35 and this is evidenced from the analysis of the Ti 2p core level (see Supporting Information, Figure S16).The B.E. of the ballmilled catalysts is similar at 458.8 and 458.6 eV for 4%Pt-4% Re/TiO 2 _BM_org and %Pt-4%Re/TiO 2 _BM_oxide, respectively, while the conventionally prepared catalyst displays an upshifted peak position at 459.1 eV.
Previously, it has been reported that ball milling as a catalyst synthesis route has a positive effect on catalytic activity as a result of the presence of oxygen vacancies.For example, MnO x -based materials showed excellent catalytic activity for the removal of persistent organic pollutants (POPs). 36This was attributed to abundant surface vacancies that enhance the generation of adsorbed oxygen and massive bulk vacancies that promote the mobility of bulk lattice oxygen.The ball-milled samples were also found to have a large surface area and uniform pore size distribution which would facilitate the physisorption of hexachlorobenzene.Ball milling has also been used to synthesize bimetallic Cu−Mn/HT materials for the aerobic oxidative synthesis of 2-acylbenzothiazole and quinoxalines. 37It was found that the ball milling route caused the material to possess surface oxygen vacancies and much more surface-adsorbed oxygen.This is thought to contribute to the enhanced catalytic performance of the material.
For the conventional catalysts, TEM analysis showed welldispersed uniform nanoparticles (Figure 3).However, in the case of the bimetallic ball-milled catalyst prepared with metal oxide precursors and milled for 150 rpm for 1 h, a significant number of free unsupported metal nanoparticles on the grid were observed.In contrast, the ball-milled catalyst prepared with organometallic precursors showed a good uniform dispersion of nanoparticles similar to that in the conventional catalyst and was found to have a smaller particle size.

■ RESULTS AND DISCUSSION
In this study, we have investigated the hydrogenation of stearic acid over Pt/TiO 2 and Pt−Re/TiO 2 catalysts prepared mechanochemically and compared them to the analogous conventionally prepared materials (Table 3).Stearic acid was chosen as a model substrate.Conversion versus selectivity profiles of the hydrogenation using 4%Pt-4%Re/TiO 2 _conv, 4 % P t -4 % R e / T i O 2 _ B M _ o x i d e , a n d 4 % P t -4 % R e / TiO 2 _BM_org at 403 K and a hydrogen pressure of 2 MPa are shown in Figure 4 and all the reaction profiles with respect to the reaction can be found in the Supporting Information.After 5 h, the conversion of stearic acid with 4%Pt/TiO 2 _conv was 56%, and the selectivity toward stearyl alcohol was 90% with the balance made up of octadecane.
With a bimetallic conventionally prepared catalyst, 4%Pt-4% Re/TiO 2 _conv, it was possible to enhance the rate of reaction (87% conversion after 5 h); however, there was reduced selectivity toward stearyl alcohol.At 56% stearic acid conversion, the selectivity toward stearyl alcohol was 80%; this continued to decrease as the reaction progressed.As discussed in a previous work, 26 the increased activity in the presence of Re was attributed to the increased oxophilicity associated with Re oxide species at the Pt−Re energy barriers and thus increased the reaction rates, although at the expense of selectivity to the stearyl alcohol.The interaction of the substrate with rhenium on the catalyst surface favors the formation of alkanes.The reduction in selectivity was attributed to the close proximity of Pt and Re in the bimetallic catalyst.
In contrast, the bimetallic catalyst, 4%Pt-4%Re/ TiO 2 _BM_org, prepared with organometallic precursors, Pt(COD)Me 2 and Re 2 (CO) 10 , by ball milling was found to be highly selective to stearyl alcohol.After 5 h, the conversion obtained was 86%, and the selectivity to the alcohol product was 89% with the balance made up of octadecane.The catalyst showed equal activity to 4%Pt-4%Re/TiO 2 _conv (87 vs 86%) and superior selectivity (89 vs 69%) toward the desired alcohol product.This catalyst contained small, well-dispersed (1.2 nm)  Abbreviations: conv, conventionally prepared catalysts by wet impregnation; oxide, catalysts prepared by mechanochemical methods with metal oxide precursors; org, catalysts prepared by mechanochemical methods with organometallic precursors; and COD, cis,cis-1,5-cyclooctadiene.nanoparticles on the support, as evidenced by TEM (Figure 3).While the mean diameter of 4%Pt-4%Re/TiO 2 _conv and 4% Pt-4%Re/TiO 2 _BM_org was similar, the ball-milled catalyst displayed a more uniform dispersion of nanoparticles across the TiO 2 support compared to the conventionally prepared catalyst and a significantly higher proportion of particles below 1 nm, for example.These smaller well-dispersed nanoparticles are thought to be the reason for the enhanced catalytic activity and selectivity due to the increased metal−support interaction.The high selectivity to stearyl alcohol indicates that the Re interaction with Pt is small and is dominated by the interaction with the support leading to high alcohol selectivity.It is thought that the Re cations may increase the rate of reaction by increasing the oxophilicity of the surface by interacting with the lone electron pair on the carbonyl oxygen.This interaction is thought to result in the decarboxylation of the carboxylic acid, resulting in the formation of alkanes.In order for decarboxylation to occur, Re and Pt must be in close proximity/interact. 28%Pt-4%Re/TiO 2 _BM_org was also found to display enhanced activity compared to the ball-milled catalyst prepared with metal oxide precursors, 4%Pt-4%Re/TiO 2 _BM_oxide (Supporting Information, Figures S7 and S8).After 5 h with 4%Pt-4%Re/TiO 2 _BM_oxide, the conversion was 62% compared to the 86% conversion achieved with 4%Pt-4%Re/ TiO 2 _BM_org.At similar conversions (62%), the selectivity toward stearyl alcohol was similar, with 90% obtained for 4% Pt-4%Re/TiO 2 _BM_org and 83% for 4%Pt-4%Re/ TiO 2 _BM_oxide.The reduction in conversion is thought to be due to the poorer dispersion of the nanoparticles on the support and also the high quantity of nanoparticles found unsupported on the TEM grid.4%Pt-4%Re/TiO 2 _BM_oxide was found to have a slightly larger particle size of 3 nm.However, despite this, 4%Pt-4%Re/TiO 2 _BM_oxide was found to be considerably more selective than 4%Pt-4%Re/ TiO 2 _conv, which had well-dispersed and smaller nanoparticles than 4%Pt-4%Re/TiO 2 _BM_oxide.
TOFs were calculated using the initial rate of reaction and the available Pt surface atoms on the catalyst obtained from TEM analysis for a selection of catalysts (Figure 3).The 4%Pt-4%Re/TiO 2 _BM_oxide catalyst was found to have a TOF of 0.045 s −1 , which is approximately 2.5 times greater than that of 4%Pt4%Re/TiO 2 _conv and 4.5 times greater than that of 4% Pt-4%Re/TiO 2 _BM_org.Previous studies have shown 27 that upon the addition of Re to the catalyst, there is a dramatic rate enhancement; however, the selectivity suffers.Remarkably, this does not appear to be the case for the ball-milled bimetallic catalysts prepared with metal oxides and organometallic precursors.The ball milling process enables an enhancement in rate (compared to that of the monometallic catalysts), while the selectivity toward the desired alcohol remains high.Monometallic ball-milled catalyst synthesis was also investigated (Supporting Information, Figures S5 and S6).At a conversion of 36%, it was found that a 4%Pt/TiO 2 _oxide catalyst, prepared with PtO 2 , was 100% selective to stearyl alcohol compared to the 91% selectivity obtained with 4%Pt/ TiO 2 _conv.However, the conventional catalyst was found to be more active with 56% conversion compared to the 36% conversion for the ball-milled catalyst after 5 h.When the Re loading was reduced to 2%, 4%Pt-2%Re/TiO 2 _oxide, from 4%, 4%Pt-4%Re/TiO 2 _oxide, there was a considerable drop in the activity of the catalyst with the conversion dropping from 62% for the 4% Re catalyst to 39% for the 2% Re catalyst after 5 h.
The selectivity remained high at 87% for the 2% catalyst at a conversion of 36% (Supporting Information, Figures S13 and  S14).
Variation of the milling speed was also investigated for the ball-milled catalyst prepared from metal oxides.The milling speed was increased from 150 to 400 rpm which was accompanied by a dramatic decrease in activity.The bimetallic catalyst milled at 400 rpm, 4%Pt-4%Re/TiO 2 _oxide 400 rpm, had a reduction in conversion to 18 from 62% after 5 h compared to the catalyst milled at 150 rpm, 4%Pt-4%Re/ TiO 2 _oxide.The selectivity was also significantly affected; at similar conversions of 18%, the selectivity toward stearyl alcohol obtained with the catalyst milled at 400 rpm was just 35%, while for the catalyst milled at 150 rpm, the selectivity was 98% (Supporting Information, Figures S11 and S12).
It is thought that a higher milling speed is required to generate the Pt−Re bimetallic interaction.At 36% conversion, the 4%Pt/TiO 2 _BM_oxide catalyst is 100% selective to stearyl alcohol.The 4%Pt-4%Re/TiO 2 _BM_oxide catalyst at 36% conversion (Supporting Information, Figures S7 and S8) has low selectivity toward alkanes (6%), indicating that a slight Pt−Re interaction is present.It is not until a higher milling speed of 400 rpm is used, 4%Pt-4%Re/TiO 2 _BM_oxide 400 rpm, that there is a significant change in selectivity toward the alkane product.For example, at a low conversion of 10%, the catalyst is 55% selective to octadecane, and at a maximum conversion of 18%, the selectivity toward octadecane is 65%.This indicates that a significant Pt−Re interaction is present in the catalyst milled at higher ball milling speeds.

■ CONCLUSIONS
A range of titania-supported Pt and Pt−Re catalysts were prepared mechanochemically and compared to conventional wet-impregnated catalysts in the selective hydrogenation of stearic acid.Mechanochemical synthesis was able to form catalysts which had a Pt−Re interaction as indicated by the high selectivity to alkanes particularly at high agitation speeds.Highly active and alcohol-selective catalysts were also found to be facilitated by this method.In particular, the mechanochemically prepared catalysts prepared from organometallic precursors were found to be superior compared to those prepared conventionally.After 5 h, the selectivity with 4%Pt-4%Re/TiO 2 _BM_org toward the desired alcohol product was 89% compared to just 69% for 4%Pt-4%Re/TiO 2 _conv, with similar conversions and much higher activity than the conventionally prepared monometallic Pt catalyst as a result of the high dispersion of the Pt particles.Monometallic mechanochemically prepared catalysts were also able to produce stearyl alcohol in 100% selectivity, which was slightly higher than that found for the conventionally prepared catalysts at the same conversion.We have shown that catalysts that have a similar activity and display superior selectivity to the desired product can be prepared by a simple solvent-free mechanochemical method.
With regard to sustainable synthesis, mechanochemistry has some inherent advantages over normal wet-impregnation methods such as the avoidance of solvent and therefore avoiding solvent waste.In other areas of synthesis, life cycle analysis has concluded that mechanochemistry presents substantial savings in terms of environmental impact, energy usage, and cost compared to solvent-based methods. 38Given the growing evidence for the effectiveness of mechanochemical methods for the synthesis of heterogeneous catalysts, aspects such as scale-up and life cycle analysis would be of interest in this area too.

Table 1 .
XPS Peak Area
N 2 Sorption Analysis of Catalysts Used in This Study a catalyst Pt precursor Re precursor BET surface area (m 2 g −1 ) pore volume (cm 3 g −1 )a

Table 3 .
Hydrogenation of Stearic Acid under the Reaction Conditions in Figure 4 over Various Catalysts after 5 h a