Modulation of Self‐Separating Molecular Catalysts for Highly Efficient Biomass Transformations

Abstract The energetically viable fabrication of stable and highly efficient solid acid catalysts is one of the key steps in large‐scale transformation processes of biomass resources. Herein, the covalent modification of the classical Dawson polyoxometalate (POMs) with sulfonic acids (‐SO3H) is reported by grafting sulfonic acid groups on the POM's surface followed by oxidation of (3‐mercaptopropyl)trimethoxysilane. The acidity of TBA6‐P2W17‐SO3H (TBA=tetrabutyl ammonium) has been demonstrated by using 31P NMR spectroscopy, clearly indicating the presence of strong Brønsted acid sites. The presence of TBA counterions renders the solid acid catalyst as a promising candidate for phase transfer catalytic processes. The TBA6‐P2W17‐SO3H shows remarkable activity and selectivity, excellent stability, and great substrate compatibility for the esterification of free fatty acids (FFA) with methanol and conversion into biodiesel at 70 °C with >98 % conversion of oleic acid in 20 min. The excellent catalytic performance can be attributed to the formation of a catalytically active emulsion, which results in a uniform catalytic behavior during the reaction, leading to efficient interaction between the substrate and the active sites of the catalyst. Most importantly, the catalyst can be easily recovered and reused without any loss of its catalytic activity owing to its excellent phase transfer properties. This work offers an efficient and cost‐effective strategy for large‐scale biomass conversion applications.


Introduction
Phase transfer catalysts (PTCs) are widelyu sed in the industrial production of aw ide range of chemicals. This is ah ighly desirable approach as it combines the advantages of both homogeneous and heterogeneous catalytic processes. The advantages of the former include higha ctivity, mild reaction conditions, fast reaction rates,a nd good accessibility to the catalytic active sites by the substrate; [1] whereas the latter demonstrates excellent recovery and recycling features. [2] Polyoxometalates (POMs) are ac lass of discrete anionic metalo xideso fV ,M o, W, etc. [3] andh ave been widely used in acid-catalyzed reactions such as esterification, alkylation, fructose conversion, and hydroxylation of olefins, owing to their highlya cidic properties and high thermals tability. [4] Additionally,t he combination of acidic properties, high proton mobility, and stability,r ender them excellent candidatesf or the conversions of biomass. [4] Nevertheless, the low surface area (< 10 m 2 g À1 )a sas olid catalyst, the high solubility in polar reactionm edia, the ease of agglomeration, and the difficulty of separation significantly limit their applicationi nc atalytic reactions.I ng eneral, the common strategy employedi nt hese cases is the "immobilization"o r" solidification"o fc atalytically activeh eteropoly acids (HPAs) [5] on appropriate supports. For example, in the case of HPA-immobilized heterogeneous acid catalysts in acid-catalyzed reactions, differentt ypes of supports have been reported,s uch as silica, [6] zirconia, [7] and alumina. [8] Recently,Juan et al. prepared aseries of materials based on immobilizing 12-tungstophosphorich eteropolyacid on az irconia supporta nd applied these as the heterogeneous acid catalysts for the esterification of palmitic acid with methanola sabiodieselm odel. [7] Although the immobilization of acid catalysts leads to larger BET (Brunauer-Emmett-Teller) surface areas, improved catalytic activity,a nd easy separationp rocesses, quite often the immobilization generatesaseries of otheri ssues such as reduced acidd ensity leadingt od ecreased acidity of [a] L. Lian, + X. Chen, + Y. Liu the POMs. [5] An alternative approach could help us overcome these disadvantages, which is the preparation of POM-based PTCs by careful modulation of the POM-based catalyst's solubility.T he most common strategy to modify the solubility of the catalyst is the careful consideration of the POM's counterions such as alkali anda lkali earth metals and their replacement with organic cations sucha si onic liquids, quaternary ammonium salts, oligomers, and so on. [9] It was recently reportedt hat the ionic liquids( IL)-POM systems "IL-POMs" exhibit high-density acidic sites and superior catalytic performancei nl iquid-phase organic reactions. [10] For example, Wang et al. synthesized as eries of solid non-conventional IL compounds composed of propanes ulfonate functionalized organic cations and heteropolyoxoanionsa nd used them as "reaction-induced self-separation catalysts" for various esterification reactions, [11] even though some mechanical and chemicals tability issues and occasionally an egative influence on the acidity of the catalyst may occur. [12] Moreover,s olidification of POMs can be realized by cationic surfactant encapsulation. [13] For example, Mizuno and co-workersr eported as eries of highly efficient POM-based Lewis acid catalysts containing rare-earth metals (TBA 6 RE-POM, TBA = tetrabutyl ammonium, RE = Y 3 + ,N d 3 + ,E u 3 + ,G d 3 + ,T b 3 + ,o rD y 3 + )m odified with quaternary ammonium salt. In this case, the incorporated rareearth metal cation performsa saLewis acidic site and exhibits significant catalytic properties in the cyanosilylation of ketones and aldehydes. [14] However,t he modification effect of the POMs in PTC systems using quaternary ammonium saltsh ave seldom been investigatedi nB rønsted acid-catalyzed reactions. This is due to the fact that the interaction between the organic ammonium cations and the inorganic polyoxoanion is greater than the one between H + and POMs. [9e] Protons can be easily exchanged with cations,l eadingt ot he decrease of the POM's acidic properties. [9g] In this work, we report an ovel approachw hichl ed to the formationo famolecular solid acid catalyst, TBA 6 -P 2 W 17 -SO 3 H, by covalentm odification of the Dawsonp olyoxometalate cluster with sulfonic acids (-SO 3 H). The structural properties and acidity of the TBA 6 -P 2 W 17 -SO 3 Ha re determinedb y 31 PNMR spectroscopy,E SI-MS,X -ray photoelectron spectroscopy( XPS), and high-resolution (HR)TEM,e tc. Use of the solid catalyst TBA 6 -P 2 W 17 -SO 3 Hi narange of catalytic biomass transformations revealed superiorc atalytic activityt ot he corresponding classical POM archetypes (such as H 3 PW 12 O 40 and K 10 -P 2 W 17 ) and in somec ases even highert han inorganic strong acids such as H 2 SO 4 under the same reactionc onditions. Mosti mportantly,t he emulsification effect of the TBA-modified amphiphilic catalysti nduces increased catalytic efficiency in the esterification of oleic acid and methanolo wing to effective interactions between substrates andt he catalyst. At the endo f the reaction, the catalysts elf-separates by precipitation;i tc an then be easily recovered and reused in multiple catalytic cycles.
The FTIR spectrum of TBA 6 -P 2 W 17 -SH ( Figure S3 in the Supporting Information) showed the characteristics tretching vibrationband of the SÀHbond located at 2571 cm À1 ,which disappearedu pon oxidationo ft he startingm aterial. Comparison of the FTIR spectra of the oxidized product and the parent molecule (TBA 6 -P 2 W 17 -SH), revealed as et of new bands located at 1043 and 1170 cm À1 associated with the stretching vibrations of the CÀSa nd S=Ob onds, indicative of the successful oxidation of the -SH functional group to -SO 3 H. Furthermore, the band centered at 1220 cm À1 was attributed to the stretching vibration of the -SO 3 Hg roup. [15] As can be seen from Figure 1b,t he 1 HNMR spectra of TBA 6 -P 2 W 17 -SH andT BA 6 -P 2 W 17 -SO 3 Hs howed the characteristics ignals at 1.02, 1.42, 1.65, and 3.15 ppm, corresponding to four kinds of hydrogen atoms associatedw ith the TBA + cation. [13] The peaks at 0.71, 1.85, and 2.64 ppm for TBA 6 -P 2 W 17 -SH can be assigned to the -Si-CH 2 -, -CH 2 CH 2 CH 2 -, and -CH 2 -SH, which are shifted to 1.13, 2.32, and 3.44 ppm for TBA 6 -P 2 W 17 -SO 3 H, respectively.T he 31 PNMR spectra of K 10 -P 2 W 17 ,T BA 6 -P 2 W 17 -SH, and TBA 6 -P 2 W 17 -SO 3 Hs how the characteristic two-lines ignals. For K 10 -P 2 W 17 ,t wo 31 PNMR resonances can be observed at À7.36 and À14.39 ppm [16] owing to two non-equivalent phosphorous atoms. In contrast, these resonances are shiftedt o À10.79 and À13.73 ppm [17] for the TBA 6 -P 2 W 17 -SH and À10.21 and À13.29 ppm fort he TBA 6 -P 2 W 17 -SO 3 Hc luster ( Figure 1c). The downfield resonance can be attributed to the phosphorus close to the organosilyl sites, whereas the upfield resonance was due to the phosphorus atom located close to the W 3 cap. [17] XPS study of the TBA 6 -P 2 W 17 -SH clusterr evealed ab and located at 163.5 eV,a ttributed to the binding energy of the S2p [18a] ( Figure S5 in the Supporting Information). After oxidation to TBA 6 -P 2 W 17 -SO 3 H, the bindinge nergy of the S2pshifted to higher energy and two closely spacedb ands are located at 168.9 and 169.9 eV (Figure 1d), which can be assigned to two different chemical environments of the covalently grafted -SO 3 Hg roups.T he observed increaseo ft he bindinge nergy in the XPS spectrum indicatesadecrease in electron density on the sulfur atom. [18b] The binding energy observed in the case of the TBA 6 -P 2 W 17 -SO 3 Hc lustera ppearst ob eh ighero wing to the more electronegative oxygen atoms on the POM shell adjacent to the -SO 3 Hg roup comparedw ith conventional cata-lyst materials such as SiO 2 -SO 3 H. [15b] The ESI-MS helped us confirm the composition of the synthesized clustera sw ell as its relevant stabilityi nt he relevant solventm edium. [19] The ESI-MS spectrumr evealed ac omplex isotope pattern ( Figure S6, Ta ble S1 in the Supporting Information) and all of the signals can be clearly assigned.T he isotopic distribution envelopes of the intact [TBA 6 -P 2 W 17 -SO 3 ] 2À and [TBA 4 -P 2 W 17 -SO 3 H] 2À cluster were located at m/z = 2967.2 and 2726.0, respectively (Figure 1e).
SEM images of TBA 6 -P 2 W 17 -SO 3 Hs howed irregular particles, which were uniformly distributed ( Figure S8 ai nt he Supporting Information) with ad iameter ranging from 30 to 50 nm. HRTEMi mages of TBA 6 -P 2 W 17 -SO 3 H( Figure S8 bi nt he Supporting Information) exhibitedh omogeneously distributed dark dots of approximately1nm in diameter, [20] which can be ascribed to the POM clusters. HAADF-STEM of the as-prepared TBA 6   Acid-base titrations weree mployed to analyze the acidic groups quantitatively( Ta ble S2 in the Supporting Information). As determined by using the Hammett indicators, the TBA 6 -P 2 W 17 cluster gave an H 0 value > À0.2 whereas the corresponding value in the case of the TBA 6 -P 2 W 17 -SO 3 Hc luster was found to be < À11.4 (Table S2 in the Supporting Information), which was comparable to that of the concentrated H 2 SO 4 (H 0 = À11.9). [21] As such, the acidity of TBA 6 -P 2 W 17 -SO 3 Hw as higher than that of the non-modified cluster, TBA 6 -P 2 W 17 .Furthermore, the acid properties of TBA 6 -P 2 W 17 -SO 3 Hw ere characterizedb y 31 PMAS (magic angle spinning) NMR probe techniquesi nvolving adsorbed trimethylphosphine (TMP) and trimethylphosphine oxide (TMPO), which is as ensitive and reliable approach to determine the type of acidity (Brønsted or Lewis acid) and the acid strength of solid acid catalysts. [22] As shown in Figure 2, the 31 Pr esonance at À2.5 ppm of adsorbed TMP confirmed the Brønsted acidity of the TBA 6 -P 2 W 17 -SO 3 Hc luster. Moreover, the strength of the Brønsted acidity was explored by TMPO adsorption, where two 31 Pr esonance peaks centered at 85 and 80 ppm clearly indicatet he presence of Brønsted acid sites with different acid strengths( Figure 2). As the threshold d 31 Pv alue of TMPO for superacidity wasd emonstrated to be approximately 86 ppm (with an acid strength similar to 100 %H 2 SO 4 ), [23] it can be concluded that the TBA 6 -P 2 W 17 -SO 3 H modified cluster exerted superacidity,which may facilitateasuperior catalytic performance.
Based on the above observations, we explored the catalytic efficiency of the modifiedc atalysti nt he esterification reaction of oleic acid with methanola si ti savery importantp retreatment step in the production of biodiesel from high free fatty acid feedstocks (Figure 3a). During the course of the catalytic reaction, the generationo ft he emulsion owing to the presence of the amphiphilic molecule provedt ob eb eneficialf or the catalytic performance owing to improved interaction of the substrate with the catalytic sites of the POM derivative.W e investigated the phase transition during the reaction in the presence of the reactant organic matrix and our modified catalyst TBA 6 -P 2 W 17 -SO 3 H. At the beginning of the reaction, oleic acid and methanol were mixed, to whicht he TBA 6 -P 2 W 17 -SO 3 H was added as al ight-yellow solid (Figure 3b)g enerating ah eterogeneous mixture. Interestingly,asafunctionoftime, the solution becameg radually turbid (Figure 3c), andastable emulsion was formed. The emulsion was developeda saresult of the formation of hydrophobic POM-based micelles containing the product of the catalytic reactiona sd epicted schematically in Figure S10 (in the Supporting Information). As the catalytic reactionp rogressed,t he micelles became unstable, leadingt o separation of the reaction mixture into two liquid phases and subsequentp recipitation of the catalysta sawhite powder (Figure 3d). The phase separation and regeneration of the heterogeneous system induced the separation of the solid catalyst as well as the phase containing the final product of the catalytic reaction. Overall, the TBA 6 -P 2 W 17 -SO 3 Hc lusterp rovedt ob ea very efficient catalyst, giving an excellent yield and selectivity of 98.7 and 99.0 %, respectively,a t7 08Ci n2 0min, which appeared to be largely enhanced compared with other examples reported so far. [24][25][26][27][28][29] To determine the optimum reactionc onditions, we studied the effect of the reaction temperature and time on the esterification of oleic acid with methanol( Figure 3, Figures S11, S12 in the Supporting Information). Generally,t he yield of methyl oleate increased as af unctiono ft ime. In 3min, the methyl oleate yield increased slowly to 16.5 %a t5 0 8C, and it increasedq uicklyt o7 9.5 %a t7 0 8C. In 20 min, the yield of methyl oleate could reach 31.8 %a t3 0 8C, 57.0 %a t5 0 8C, 81.0 %a t608C, and 98.7 %a t7 08C,.
The yield of methyl oleate and ln(C t /C 0 )w ere plotted against the reaction timea ss hown in Figure 4g,i nwhich C 0 and C t are the initial oleic acid concentration and concentration at time t,r espectively.T he linear fit of the datar evealed that the catalytic reaction exhibited ap seudofirst-order kinetic constant for the esterification reaction (R 2 = 0.9942). The rate constant k of the conversion of oleic acid was determined to be 0.0166min À1 based on Equations (1) and (2).
The above results obtained from our system along with data of previously reported catalysts are summarized in Ta [29,30] esterification hardly occurred in the presence of the K 10 -P 2 W 17 ,T BA 6 -P 2 W 17 ,a nd TBA 6 -P 2 W 17 -SH catalysts under the employed conditions. The relevant yield of the methyl oleate in this case was found to be only 0.6, 1.0, and 0.7 %, respectively (entries 4-6, Ta ble S4 in the Supporting Information). In marked contrast, the presence of the modified TBA 6 -P 2 W 17 -SO 3 Hc atalysti nduced as elf-separating liquid-solid heterogeneous reactions ystem and demonstrated as uperior yield of 98.7 %( entry 7, Table S4 in the Supporting Information). The observede fficiency of the modified catalytic system clearly outperforms the one observed in the case of the non-modified adduct (TBA 6 -P 2 W 17 )aswell as the top performingexamples reported previously.
Ta ble 1s ummarizes the conditions and the catalytic performance of different catalysts used for the catalytic esterification reaction. It is evident that the modified TBA 6 -P 2 W 17 -SO 3 H catalystr evealed ah igh conversion rate with at urnover frequency( TOF) of 52.8 h À1 and 546.0 h À1 at 298 and 343 K, re-spectively( entries 8a nd 9, Ta ble 1). The grafting of sulfonic acid (SO 3 H) functional groups on the POM shell modified the acidityo ft he catalyst, which clearly benefited the catalytic efficiency.
To investigate further the general applicabilityo ft he TBA 6 -P 2 W 17 -SO 3 Hc atalysti ne sterification reactions, as eries of various combinationso ff attya cid and alcohols ubstrates were evaluated. Ta ble 2a nd Table S5 (in the Supporting Information) summarize the findings of this effort. More specifically, for small molecular weight alcohols such as methanol, ethanol, propanol, butanol, and pentanol, the yield of the esterification reactionu sually reached av alue of more than 97 %w ithin 90 min (entries 1-5, Ta ble 2). The time required to reachayield of 97 %i ncreased according to the increaseo ft he alcohol's molecular weight. On the other hand, with the use of small molecular acids, such as propionic, butyric, valeric, and caprylic acid, the esterification reactions proceeded rapidly, reaching more than 97 %i n3 0min (entries 6-10, Table 2). Interestingly, equally excellent catalytic activity and selectivityw ere obtained in the esterification of long-chain acids and methanol ( Figure S14 in the Supporting Information) as demonstrated in the synthesis of benzyll aurate, benzyl hexanoate, methyl 5-hexanoate,a nd methyl methacrylate( entries 11-14, Table 2). These resultsd emonstrated the general applicability of the modified TBA 6 -P 2 W 17 -SO 3 Ha cid catalyst in the esterification of av ariety of acids andalcohols for the production of biodiesel.
In an effort to investigate the recyclability of the TBA 6 -P 2 W 17 -SO 3 H, the catalyst was separated by filtrationa fter the first run, washedw ith methanol, and dried under vacuum before use in the next catalytic cycle. The yield of methyl oleate decreased slightly from 98.67 to 94.35 %a fter five successive runs, whereas negligible loss of reactivity could be detected. In addition, the 31 PNMR, XPS, and elemental (C, N, O, P, Si, S, and W) mapping data obtainedf or the recycled catalystw ere found to be the same as that of the fresh one, whichi si ndicative of the structuralstability during the course of the catalytic cycles (Figure S15 in the Supporting Information). 5-Hydroxymethylfurfural( 5-HMF) is ap otentially promising platform molecule that can be converted into several valuable chemicals, including2 ,5-dimethylfuran, 2,5-diformylfuran,1 ,6hexanediol, formic acid, and levulinic acid. [31] Considering the efficiency observed in the esterification reactions, we investigated the potential use of TBA 6 -P 2 W 17 -SO 3 Hi nt he catalytic transformation of different carbohydrates into 5-HMF.
In this case, as eries of different organic solvents were evaluated for their potentiale ffect on the fructose dehydration at 100 8C( Figure S16 in the Supporting Information). 1,4-Dioxane provedt ob et he most effective solvent medium, reaching a yield of 99.0 %f or the production of 5-HMFa t1 00 8Ci n2h, whereas the obtained yields when using DMSO, DMF,m ethanol, ethanol, and water as solvents were the 94.9, 88.5, 1.5, 39.1, and 2.8 %, respectively.F urthermore, the effect of the reaction temperature ( Figure S17 in the Supporting Information) and catalystd osage (FigureS18 in the Supporting Information) on the catalytic activity of fructosed ehydration were investigated and they optimum values found to be 100 8Ca nd   Chem.E ur.J.2020, 26,1 1900 -11908 www.chemeurj.org 2020 The Authors. Published by Wiley-VCH GmbH 150 mg, respectively.I ts hould be noted that TBA 6 -P 2 W 17 -SO 3 H showedi mproved catalytic conversion than the one observed in the case of strongi norganic acids such as H 2 SO 4 [32] and HCl. [33] To improve further our understanding of the fructosed ehydration reaction, we monitored the catalytic reactionu sing 13 CNMR spectroscopy.A ss hown in Figure 4b,a tt he beginning of the catalytic reaction, the signals located in the range 50-120 ppm can be assigned to the cyclic forms of fructose (the 68.5 ppm peak corresponds to the 1,4-dioxane solvent). [34] Ad ecrease of the signal's intensity corresponding to the fructose molecules was observed as af unction of the time, whereas new peaks gradually appeared at 180.4, 161.5,1 52.0, 126.7, 111.0, and 56.1 ppm, which can be assigned to the production of 5-HMF. [35] Finally, 13 CNMR spectroscopy revealed the complete transformation of the fructose within ap eriod of 2h, during whicht he only detectable productsi nt he reaction mixture were 5-HMF and 1,4-dioxane solvent. During the catalytic transformation of fructose, the color of the reaction mixture turned gradually from colorless to orange-yellow.C atalytic recycling experimentss howed the decrease of 5-HMF yield from 94.9 to 90.2 %a fter four consecutiver uns, indicating minor leachingo ft he catalyst (FigureS19 in the Supporting Information).
The broad utility of the catalyst was furtherd emonstrated by investigating the efficiency during the catalytic transformation of different substrates (Figure 4c)o ver TBA 6 -P 2 W 17 -SO 3 Hi n 1,4-dioxane. Using aw ide range of carbohydrates as substrates such as glucose, sucrose,a nd inulin, we were also able to obtain decent yields of 57.9, 60.3, and 47.5 %, during their catalytic transformation to 5-HMF.H owever,o nly 1.6 %o fH MF product was obtained when cellulose wasu sed as the substrate. This observation is indicative of the catalyst's high efficiency and selectivity in the case of monosaccharides or disaccharidesb ut poor performance in the case of polysaccharide substrates. It is worth noting that the differencei ny ields observed for the dehydration of glucose (57.9 %) and fructose (99.0 %) could be due to the lack of co-existence of Brønsted (B) and Lewis (L) acidic sites in the catalytic system,w hich seem to be required for the efficient transformation of glucose or cellulose to HMF. [36] Conclusion The covalent tethering of sulfonic acids on the shell of the Dawson cluster was achieved by surfacegraftinga nd oxidation of (3-mercaptopropyl)trimethoxysilane. The employed approach led to the modulation of the Brønsted acidity of this self-separating phase transfer molecular catalyst, which exhibits superior performance in biomass transformationso wing to its superacidic properties. Thea cidity of the catalyst was determined by Hammett indicators, potentiometric titration,a nd 31 PMAS NMR spectroscopy,c onfirming its approximate superacidity.T he modified molecular catalyst, TBA 6 -P 2 W 17 -SO 3 H, showede xcellent catalytic activity and selectivity in aw ide range of acid-catalyzed reactions, such as the esterification of oleic acid with ay ield of 99.0 %. Interestingly,t he emulsification effect of the modified amphiphilic catalystn ot only induced an increased catalytic efficiency during the catalytic transformation of the substrates owing to the homogeneity of the system but also led to as elf-separating catalytic system at the end of the catalytic cycle owing to the destabilization of the emulsion and self-precipitation of the catalyst. The embedded emulsification-precipitation cycle induces excellent self-recycling properties to the catalytic system, leading to facile and low-cost recovery of the catalyst at high yields. The design approach described herein paves the way for further development of cost-effective highly efficient solid acid catalysts engineered for targeted catalytic transformations of biomass-derived raw materials to high value-added chemicals.