HYDROTALCITE-IMPREGNATED COPPER AND CHROMIUM-DOPED COPPER AS NOVEL AND EFFICIENT CATALYSTS FOR VAPOR-PHASE HYDROGENATION OF FURFURAL: EFFECT OF CLAY PRETREATMENT

A series of novel catalysts based on immobilization of copper and chromium-doped copper on calcined and uncalcined hydrotalcite support have been developed and used for promoting vapor-phase hydrogenation of furfural. The results indicated that a pre-calcination of the support has a negative role in changing the structural features and catalytic activity of the final catalysts. Moreover, the presence of the Cr promoter can improve the catalyst performance. The best performance (furfuryl alcohol selectivity of 94% and furfural conversion of 89%) was obtained over the Cr-doped copper catalyst supported on as-synthesized hydrotalcite material.

Layered double hydroxides (LDHs), also referred to as hydrotalcite-like compounds (Xu et al., 2011), are a class of clays with the general formula of , in which M 2+ and M 3+ are divalent and trivalent metal cations, and A n-is a counter anion (Liu et al., 2016).Although LDHs have been widely used in catalysis (Thirumalaiswamy et al., 2015;Basini et al., 2002;Wei et al., 2010;Baskaran et al., 2015;Thangaraj et al., 2016;Wang et al., 2017;Karthikeyan and Karuppuchamy, 2017;Hernández et al., 2017;D'Oca et al., 2017;Eskandari et al., 2017;Santos et al., 2017), there are only a few reports on their utility as the catalyst support for the gas-phase hydroconversion of FF to FFA (Manikandan et al., 2016b;Sangeetha et al., 2009); in none of them copper and chromiumdoped copper species were supported (impregnated) on Mg-Al hydrotalcite.To the best of our knowledge, the effect of a pre-calcination of the hydrotalcite support on the catalytic activity of copper catalysts for the above-mentioned process was not reported in the literature.
Taking into account the outstanding properties of hydrotalcite support and the importance of conversion of FF, in this paper we wish to report the utility of hydrotalcite for supporting copper and chromiumdoped copper species for developing novel catalysts for promoting catalytic hydrogenation of FF to FFA.Moreover, for the first time we study the effect of the pretreatment (pre-calcination) of hydrotalcite on the catalytic activity of the final catalyst to highlight the importance of the hydrotalcite pretreatment on its performance as a support.

Preparation of support
A conventional co-precipitation method was used for the synthesis of hydrotalcite (abbreviated as HT) (Li et al., 2006).Briefly, to a stirred mixture of 1 mol of Mg(NO 3 ) 2 .6H 2 O and 0.5 mol of Al(NO 3 ) 3 .9H 2 O (with an Mg/Al molar ratio of 2) in 700 mL of deionized water a solution of 3.5 mol of NaOH and 0.943 mol of Na 2 CO 3 was added dropwise to adjust the final pH to 13.Subsequently, the mixture was heated at 65 ºC for 18 h.The resulting precipitate obtained by vacuum filtration was washed thoroughly with deionized water to reach a neutral pH and then dried overnight at 130 ºC.Half of the solid was further used without calcination and the remaining part was calcined at 550 ºC for 5 h to obtain the final oxide catalyst.

Supporting the catalytic species
Wet impregnation was used for supporting the copper and copper chromite species on the prepared supports.In a typical procedure, the HT support was suspended in deionized water and stirred vigorously.Subsequently, a solution of Cu(NO 3 ) 2 .3H 2 O 11% (wt/ wt) in deionized water was added dropwise to the aforementioned suspension.The obtained mixture was then impregnated for 24 h at ambient temperature.Upon completion of the process, the catalyst was dried in an oven at 150 ºC for 12 h and afterward calcined at 550 ºC for 5 h.The obtained catalysts were denoted as CHT1 and CHT3 for the calcined and uncalcined supports, respectively.The hydrotalcite-supported chromium-copper catalysts were prepared via the same procedure except that Cr(NO 3 ) 3 .9H 2 O 5.5% (wt/ wt) was added together with the copper precursor during the impregnation.These two catalysts were named CHT2 and CHT4 as the counterparts of CHT1 and CHT3, respectively.

Characterization of the catalysts
The analytical methods employed for the characterization of the synthesized catalysts included SEM, EDX, BET, XRD, and TGA.A Tescan instrument, using Au-coated samples with an acceleration voltage of 20 kV was used for taking SEM/EDX images.The measurement of BET surface area was carried out after nitrogen physisorption using a Quantachrome Chem-BET 3000 sorption analyzer at 77 K. Degassing of the sample was performed at 120 ºC for 3 h prior to the BET analysis.Room temperature powder X-ray diffraction patterns were collected using a Siemens D5000.Cu Kα radiation was used from a sealed tube.Data were collected in the 2θ range of 30-65º with a step size of 0.02º and an exposure time of 2 s per step.Perkin Elmer Pyris 1 instrument with 10 ºC/min ramping in nitrogen atmosphere was applied for the thermogravimetric analysis (TGA) and obtaining the derivative thermogravimetric (DTG) curves of the spent catalysts.

Catalytic activity
The atmospheric vapor-phase conversion of FF over the hydrotalcite-supported catalysts was conducted in a continuous flow tubular reactor inside an electrically heated furnace.In each run, the fixed bed of catalyst pellets (0.3-1.0 mm) was placed between two plugs of quartz wool.The temperature of the reactor was maintained by two thermocouples and two programmable temperature controllers.Prior to the reaction, the catalyst was reduced in a dilute flow of hydrogen (33 vol.%) in nitrogen with a total flow rate of 6.4 Lg -1 h -1 at 250 ºC for 3 h.Then, the catalyst bed was cooled down to the reaction temperature in pure hydrogen.Afterwards, FF was injected into the reactor using a motorized syringe pump.The reaction was carried out at the temperature of 180 ºC and WHSV of 1.7 h -1 with a hydrogen-to-feed (H 2 /HC) volumetric ratio of 10.The hydrogenation process was followed by collecting the liquid products in an ice trap and analyzing on a gas chromatograph (GC) equipped with a capillary column and a flame ionization detector (FID).The carbon balance was higher than 95% in all experiments unless otherwise stated.The outlet molar flow rates were accordingly normalized assuming a closed carbon balance for the interpretations.Finally, the performance of the catalyst was estimated using the following definitions: where the performance data were reproducible to within ±2%.

RESULTS AND DISCUSSION
Fig. 1 compares the performance of the four synthesized catalysts in terms of FF conversion, FFA selectivity, and FFA yield for the vapor-phase hydrogenation process.As observed in this figure, the most selective catalyst was CHT3 with FFA selectivities higher than 98% in the whole range of the hydrogenation process.Although this catalyst showed sustained selectivity during the operation, its FFA yield and FF conversion declined moderately from 92.4% and 96.4% respectively to 50.8% and 51.1%.This trend showed how this catalyst suffered from a gradual deactivation with time.Comparatively, CHT4 proved to have a superior performance relative to other catalysts (see Fig. 1).In addition to its high and sustained selectivity towards FFA (more than 94%), there appeared only mild variations in its FF conversion and FFA yield levels during the 4 h experimental run in such a manner that the values of these parameters remained respectively above 89% and 87%.In contrast, the CHT1 and CHT2 catalysts exhibited poor performances in the vaporphase hydrogenation of FF, particularly in terms of FF conversion and FFA yield, which were lower than 3% on both of the catalysts during the operation.In fact, the non-calcined hydrotalcite-supported catalysts were much more efficient than those on the calcined supports in the FF hydrogenation process.This is an interesting observation, proving that a pre-calcination of the support is not beneficial for the favored reactions.Nevertheless, CHT2 with an about 80% selectivity towards FFA after 4 h of operation was more favorable than CHT1 with an FFA selectivity of ~50%.Moreover, by incorporating Cr as a promoter, the catalysts behaved more durably under the conditions employed.Therefore, as illustrated in Fig. 1, the best overall performance belonged to CHT4, the non-calcined hydrotalcite-supported chromiumpromoted copper catalyst.
A schematic diagram of the plausible reactions leading to the observed products is shown in Fig. 2.Moreover, Fig. S1 presents the trends of important byproducts which were produced on the synthesized catalysts during 4 h of operation.As can be seen, the only byproduct of the hydrogenation process over CHT1 was 5-methylfurfural (MFF), which was produced exactly with a reverse trend against FFA (see Fig. 1).More precisely, at the initial times of catalytic run, the selectivity of the CHT1 catalyst towards MFF was 100%, which then decreased sharply and reached the level of 50% and stayed almost constant at this plateau until the end of run.The byproducts of the CHT2 catalyst included MFF and 2-acetylfuran (AF).MFF increased gradually with time on CHT2.Although the variety of the byproducts obtained over CHT3 and CHT4 was more pronounced than on CHT1 and CHT2, the selectivities of the byproducts over the two former catalysts were significantly lower (up to ~3% of each) than those on the others.This led to higher FFA selectivities on CHT3 and CHT4 (see Fig. 1).The CHT3 catalyst mainly produced tetrahydrofurfuryl alcohol (THFA), the amount of which decreased rapidly during the operation.Other byproducts on this catalyst were 5-methylfurfuryl alcohol (MFFA) and γ-valerolactone (GVL).However, CHT3 also produced MFF with a gradual increasing trend.As can be seen in Fig. S1, the main byproduct of the hydrogenation process over CHT4 was THFA as observed in the case of CHT3.Other important byproducts on this catalyst were MF, difurfuryl ether (FFE) and 1-pentanol (1POL).Interestingly, the most active and durable catalyst (CHT4) did not produce MFF, in contrast with the other counterparts.This finding alludes to some detrimental role of MFF in the sustained production of FFA.
The significant effect of pre-calcination of the support, as well as the role of Cr promoter on the catalytic performance motivated us to study and compare the four catalysts from a structural point of view.For this purpose, all prepared catalysts were characterized by using the SEM, EDX, XRD and elemental mapping techniques and the BET method.Interestingly, the comparison of the BET surface areas of the catalysts (Table 1) revealed that a precalcination of the support remarkably reduces the BET surface area.As shown, the CHT1 and CHT2 samples exhibited very low surface areas (less than 2 m 2 g -1 ) while the uncalcined counterparts had high surface areas (more than 120 m 2 g -1 ).Moreover, the CHT4 sample, which contained ~5 wt% of Cr as measured by EDX analysis, showed a considerably higher surface area than CHT3.Noteworthy, the surface .This observation established that the incorporation of Cu with or without promoter and the subsequent calcination can dramatically decrease the surface area of the HT.Having these data, we attributed the superior catalytic activity of CHT3 and CHT4 to their higher surface area.This influence might be surprising considering the so-called memory effect (Shylesh et al., 2016;Kanezaki, 1998;Nishimura et al., 2013;Cavani et al., 1991) known for the layered double hydroxides.A reasonable explanation for this complexity would be a partial collapse of the hydrotalcite structure that reduced the internal surface area.Similar studies (Kowalik et al., 2013) reported structural transformations of HT material into crystalline nonlayered oxide phases such that the material lost its capability to recover its layered structure.In fact, hydrotalcites calcined in the temperature range of 350-550 ºC are known to be susceptible to partial/ total destruction by exposure to water (Shylesh et al., 2016;Nishimura et al., 2013;Wang et al., 2012;Zhu et al., 2014) which in turn leads to a loss of surface area (Shylesh et al., 2016;Cavani et al., 1991;Nishimura et al., 2013) followed possibly by the appearance of some broad, ill-defined diffraction peaks attributed to periclase (vide infra) (Shylesh et al., 2016).
To provide more insight into the structural features of the catalysts, SEM/EDX and elemental mapping analyses were also performed (Fig. S2).As depicted in this figure, the morphologies of the catalysts prepared from calcined and uncalcined supports are clearly distinguished.Furthermore, the presence of Cr as a promoter can induce changes in the morphology.However, the SEM images of CHT1 and CHT2 exhibited some degree of similarity and both catalysts showed aggregated morphologies.Analogous morphological changes have been reported previously (Wang et al., 2016;Zhu et al., 2014).Nevertheless, the introduction of Cr caused more roughness.In the case of CHT4 the catalytic species are presented in the form of small particulates while in CHT3 large particles are formed and aggregated.The EDX analyses of all samples demonstrate the presence of Mg, Al, O, and Cu atoms.Moreover, the Cr atom was detectable in the cases of CHT2 and CHT4.The elemental mapping analyses proved that the Cu species were more abundant in the cases of CHT3 and CHT4.It is worth mentioning that the stacking of Mg and Al was more pronounced in CHT3 and CHT4 and their dispersions were uniform.Considering these results, it can be concluded that the uncalcined support had higher stacking and uniformity, which could result in a better immobilization of Cu species and consequently a superior catalytic activity.
The XRD patterns (Fig. 3) were also used to disclose the effect of pre-calcination of the support on the structural features of the catalysts.As depicted in this figure, the XRD patterns of all catalysts exhibited the characteristic peaks of  and .Moreover, the intensities of the characteristic peaks for the uncalcined samples were more pronounced.Additionally, no characteristic peaks for Cr species were detectable on CHT2 and CHT4.According to the  previous reports (Mallik et al., 2006), this observation can be ascribed to the high dispersion of promoter on the catalyst.A well-crystallized periclase (MgO) structure was formed along with a tenorite phase in the pre-calcined samples, indicating the partial collapse of the hydrotalcite structure (Shylesh et al., 2016;Comelli et al., 2013).Moreover, the XRD patterns of CHT1 and CHT2 gave no reflection for the presence of additional alumina phases; the diffraction peaks of these catalysts presented slightly higher 2θ values compared to those of pure periclase, suggesting that the Al 3+ cations were incorporated into the MgO lattice (Shylesh et al., 2016;Kanezaki, 1998).Using the Scherrer equation, the CuO crystallite sizes were calculated for all samples (Table 1).As is obvious, this value changed in the order of CHT3 < CHT1 < CHT2 < CHT4 indicating that the incorporation of Cr can result in an increase in the CuO crystallite size.This effect is more obvious in the case of uncalcined support.
Fig. 4 illustrates the TGA and DTG curves of the spent catalysts within the temperature range of 25-800 ºC.For all synthesized catalysts, endothermic valleys related to different kinds of carbonaceous deposits were observed in the DTG curves.The weight loss in the range of approximately 200-450 ºC was attributed to the soft coke.However, any other weight losses at higher temperatures could be assigned to the hard coke (Zhang et al., 2014;Ghashghaee et al., 2017).A significant weight loss was associated with CHT1 with two considerable valleys between 300-650 ºC (see Fig. 4), the first valley regarded as soft coke and the second one attributed to hard coke.These observations were consistent with the extremely poor performance of CHT1 in hydroconversion of FF.More precisely, the catalyst suffered from deactivation at the beginning of the hydrogenation process (see Fig. 1).The lowest amount of soft coke belonged to CHT4, which displayed the highest catalytic performance in terms of activity, FFA selectivity and durability among others.Among the four fabricated catalysts, the least weight losses were observed for CHT2 and CHT4, which were chromium-promoted catalysts.From this observation, one may conclude that the presence of chromium inhibited the formation of carbonaceous residues (see Fig. 4(a)).Although CHT2 showed the second lowest amount of weight loss during coke burnout, it experienced the deposition of hard coke as associated with CHT1, but with relatively harder deposits (compare the temperature peaks in Fig. 4(b)).This observation supports the less favorable activity on CHT2 than on CHT4.

CONCLUSION
For the first time, copper and chromium-doped copper catalysts supported on hydrotalcite were applied to the gas-phase hydrogenation of furfural.In summary, the exertion of a pre-calcination of hydrotalcite support can affect the structural features of the resulting copper catalyst and consequently the catalytic performance.The presence of Cr promoter can also influence these parameters.In an excellent agreement with the surface area measurements, the Cu/HT catalyst with Cr promoter with no precalcination rendered the most active and stable catalyst towards furfuryl alcohol, with a selectivity of >94% and a furfural conversion of 89%.Interestingly, no 5-methylfurfurl was produced over this durable catalyst, in contrast with the other counterparts.

Figure 1 .
Figure1.The main performance data for the hydrotalcite-supported copper catalysts in the catalytic conversion of FF to FFA at 180 °C, 1 atm, WHSV of 1.7 h -1 , and H 2 /HC of 10.

Figure 2 .
Figure 2. A schematic diagram of the plausible reaction network for the activity of the four synthesized catalysts (the plain compounds are those that were not detected in the effluent stream).

Figure 3 .
Figure 3. XRD patterns of the four synthesized catalysts.

Table 1 .
Textural properties and EDX chemical compositions of HT and the impregnated catalysts.