Impact of Macroporosity on Catalytic Upgrading of Fast Pyrolysis Bio‐Oil by Esterification over Silica Sulfonic Acids

Abstract Fast pyrolysis bio‐oils possess unfavorable physicochemical properties and poor stability, in large part, owing to the presence of carboxylic acids, which hinders their use as biofuels. Catalytic esterification offers an atom‐ and energy‐efficient route to upgrade pyrolysis bio‐oils. Propyl sulfonic acid (PrSO3H) silicas are active for carboxylic acid esterification but suffer mass‐transport limitations for bulky substrates. The incorporation of macropores (200 nm) enhances the activity of mesoporous SBA‐15 architectures (post‐functionalized by hydrothermal saline‐promoted grafting) for the esterification of linear carboxylic acids, with the magnitude of the turnover frequency (TOF) enhancement increasing with carboxylic acid chain length from 5 % (C3) to 110 % (C12). Macroporous–mesoporous PrSO3H/SBA‐15 also provides a two‐fold TOF enhancement over its mesoporous analogue for the esterification of a real, thermal fast‐pyrolysis bio‐oil derived from woodchips. The total acid number was reduced by 57 %, as determined by GC×GC–time‐of‐flight mass spectrometry (GC×GC–ToFMS), which indicated ester and ether formation accompanying the loss of acid, phenolic, aldehyde, and ketone components.


Introduction
Biofuelsh ave an important role to play in mitigating anthropogenic climate change arising from the combustion of fossil fuels. [1] In the context of energy,d espite significant growth in fossil fuel reserves, great uncertainties remain in the economic and environmental impact of exploitation, and crucially,a pproximately 65-80 %o fs uch carbon resources cannotb e burned without breaching the United Nationsf ramework conventiono nc limate change (UNFCC)t arget to keep the global temperature rise this centuryw ell below 2 8C. Biofuelsw ill prove critical in helping many countries meet their renewable energy commitments, which for the UK are 15 %b y2 020, alongside greenhouse gas (GHG) emission reductionso f3 4% by 2020 and8 0% by 2050 (compared with 1990 levels). They also represent drop-in fuels abletou tilize existing pipeline and filling station distribution networks. [2] Thermochemical processing of waste biomass such as lignocellulosic materials sourced from agriculture or municipal waste offers ap romising route to biofuels through pyrolysis. [3] Pyrolysis is aw idespreada pproach for bio-oil [4] synthesis, in which biomass is thermally decomposedi na no xygen-free or oxygen-limited environment. [5] The resultingc rude bio-oil is a complex mixture of acids, alcohols, furans, aldehydes,e sters, ketones,s ugars, and multifunctional compoundss uch as hydroxyacetic acid, hydroxyl-acetaldehyde and hydroxyacetone (derived from cellulose and hemicellulose), together with 3-hydroxy-3-methoxy benzaldehyde,p henols, guaiacols, and syringols derived from the lignin component. [1b, 6] Pyrolysis bio-oils thus require "upgrading" through deoxygenation and neutralization to enhance their energy density,s tability, and physical properties. [6a, 7] Ar ange of catalytic upgrading methods are known, [8] at least at the laboratory scale, including esterification, [9] ketonization, [10] hydrodeoxygenation, [11] and condensation. [12] Carboxylic acids comprise5 -10 wt %o fp yrolysisb io-oils, [9,13] and are largely responsible for their poor chemical stability. Hence, esterification (particularly employing bio-derived alcohols such as methanol, ethanol, or phenols [9,14] )o ffers an energy-efficient and atom-economical route to upgrading. [8b, 15] Homogeneous mineral acid catalysts are historically employed for esterification,h owever their process disadvantages and poor (environmental) E-factors are well-documented;h ence, strong drivers remain for the development of heterogeneous Fast pyrolysis bio-oils possess unfavorable physicochemical properties and poor stability,i nl arge part, owing to the presence of carboxylic acids, whichh inders their use as biofuels. Catalytic esterification offers an atom-and energy-efficient route to upgrade pyrolysis bio-oils. Propyl sulfonica cid (PrSO 3 H) silicas are active for carboxylica cid esterification but suffer mass-transport limitations for bulky substrates. The incorporation of macropores (200 nm) enhances the activity of mesoporousS BA-15a rchitectures( post-functionalized by hydrothermal saline-promoted grafting) for the esterification of linear carboxylic acids, with the magnitude of the turnover frequency( TOF) enhancementi ncreasing with carboxylic acid chain length from 5% (C 3 )t o1 10 %( C 12 ). Macroporous-meso-porousP rSO 3 H/SBA-15 also provides at wo-fold TOF enhancement over its mesoporous analogue for the esterification of a real, thermal fast-pyrolysisb io-oil derived from woodchips.T he total acid number was reduced by 57 %, as determined by GC GC-time-of-flight mass spectrometry (GC GC-ToFMS), which indicated ester and ether formation accompanying the loss of acid, phenolic, aldehyde, and ketone components. KGaA. This is an openaccessarticleunder the termsoft he Creative Commons AttributionL icense, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. solid acid counterparts. [11] Although base catalysts are widely used for the transesterification of vegetable oils (triacylglycerides) to yield biodiesel, they are unsuitable for catalytic esterification owing to neutralization/saponification. [1d] Diverse solid acids have been explored for esterification, including zeolites, [16] heteropolyacids, [17] sulfated metal oxides, [18] carbon-based acid catalysts, [19] and functionalized mesoporous silicas. [20] Research on the latter indicates that mesoporous SBA-15, [21] KIT-6, [22] and PMO [23] sulfonica cids, and am acroporous-mesoporous   [20g] analogue,a re among the most promising owing to their tunable pore architectures trong Brønsted acidity and hydrophobicity. [2a, 14a, 20g, 23, 24] 3-Propylsulfonic acid (PrSO 3 H)/SBA-15 has been reported as an efficient catalystf or acetic acid esterification with methanol [2a, 25] and other alcohols in simulated bio-oils, [26] and the most widely used sulfonic acid in solid acid catalyzed esterification. [27] Such catalysts exhibit improved water tolerance during esterification when the sulfonated silica surfacei sc o-functionalized with alkyl chains. [2a, 5, 25b] We recently reported ap ostmodification hydrothermal saline-promoted grafting( HSPG) route to introduce higher sulfonic acid loadingsi nto mesoporous silicas than those achievable by conventional grafting methods, [24a] andc onfer stability towards leachingd uring the esterification of model acids. [24b, 28] Hydrophobicity and catalytic reactivity,c an also be enhanced throughi ncorporating organic groups into the silica framework. [24b] Mesopore interconnectivity also plays ar ole in controlling esterification activity,w ith interconnectivity between the hexagonal cylindrical mesopores of PrSO 3 H/KIT-6 offering superior mass transporta nd active site accessibility to non-interconnectedP rSO 3 H/SBA-15. [20g] Mesopore expansion (from % 5t o1 4nm), [14a] and macroporei ncorporation [23] offer alternative approaches to enhancethe esterification activity of PrSO 3 H/SBA-15 for long chain fatty acid esterification.
With respect to bio-oil upgrading through catalytic esterification, most studies have employed only modelc ompounds owing to the complex nature of real pyrolysis bio-oils [7a] and the associateda nalytical challenge. We previously reported the application of PrSO 3 H/SBA-15 for acetic acid esterification of model bio-oils. [26,28] Here, we report the synthesis and application of HSPG-derivedm esoporous PrSO 3 H/SBA-15, and am acroporouscounterpart, for the esterification of simple carboxylic acids (C 3 ,C 6 ,a nd C 12 ), and the upgrading of thermal fast pyrolysis bio-oilderived from woodchips.

Catalyst characterization
The successful synthesis of an ordered mesoporous skeleton for SBA-15a nd am acroporous-mesoporous (MM) skeleton for MM-SBA-15 (with am eanm acropore diameter of % 200 nm, close to that of the polystyrene colloidal hard template, Figure S1 in the Supporting Information) supports was confirmed by TEM. An ordered, 2D hexagonal mesopore channel network was observed for the former,a nd aw ell-definedi nterconnecting macropore-mesopore network for the latter ( Figure S2). Formation of the desired p6mm pore architecturef or both SBA-15 and MM-SBA-15 was confirmed by low angle X-ray diffraction( Figure S3), which revealed reflections characteristic of hexagonally orderedm esostructures. Both supports retained hexagonal close packed pore architectures following functionalization by propylsulfonic acidi naH 2 O/NaCl mixture (the HSPG method). However, as hift in the diffraction peaks to highera ngle was observed post-functionalizationo wing to mesopore contraction. [23] Mesopore generation (and retention after sulfonation) was further evidenced by N 2 porosimetry, which showed type IV isothermsw ith H1 hysteresis loops for all materials ( Figure S4). The textural properties of PrSO 3 H/SBA-15 and PrSO 3 H/MM-SBA-15 are summarized in Ta ble 1. The BET surfacea reas decreased after sulfonic acid grafting over both silicas owing to micropore blockage, which was apparent as a dramatic drop in the micropore area and pore volume. These changes were accompanied by ad ecrease in pore diameter and an increaseinwall thickness, suggesting the uniformgrafting of sulfonic acid groups throughout both pore networks without distortion of their unit cells. Previous studies have shownt he macropores in such hierarchical frameworks are open andinterconnectedb ybottleneck pore openings. [23,29] Diffuse reflectancei nfrared fourier transform spectra (DRIFTS)o ft he parent silicas showed bands at 700-1400 cm À1 and 3000-3800 cm À1 ,w hich were indicative of framework Si-O-Si and surfaces ilanols, respectively ( Figure S5). [15] Additional bands appeared atapproximately 2960-2830 cm À1 after sulfonation of both materials, which were attributed to CH 2 vibrations of the propyl backbone, and an ew CH 2 ÀSi band centered at 1360 cm À1 .C HNS elemental analysiso ft he sulfonated silicas revealed that both contained approximately 6wt% sulfur (Table 1), which represented af ive-fold increase over conventional toluene grafting, [14a, 23] in good agreement with our preliminary results using the HSPG method. [24a] S2pX Ps pectra of both sulfonic-acid-functionalized materials in Figure S6 reveal two distinct Sc hemical environments;alow binding energy centered at 164.5 eV associated with unoxidized thiol, and a higher energy doublet arising from sulfonica cid groups centered at 168.9 eV. [30] Quantitative XPS analysis ( Table S1) showedt hat approximately 85 %o fSwas incorporated as sulfonic acid groups. Thermogravimetric analysis (Figure S7 b) highlighted two major weightl osses;o ne below 100 8C, which was attributedt op hysisorbed water;a nd the second between 250-650 8Co wing to propylsulfonic acidd ecomposition. [31] The bulk Sc ontent estimated from this second loss feature was approximately 5wt% in accordance with elemental analysis.A cid properties of both sulfonated silica were subsequently probed throughp yridinea nd propylamine adsorption. DRIFT spectra of pyridine-titrated materials ( Figure S8) evidenced only Brønsted acid sites. [26] Te mperature-programmed analysiso fr eactively formed propene from chemisorbed propylaminec onfirmed that PrSO 3 H/SBA-15 and PrSO 3 H/MM-SBA-15 possesseds imilar acid strengthsa nd loadings ( Figure S9 and Figure S10). Therefore, the incorporation of macropores into the SBA-15a rchitecture had minimal impact on silica functionalization;t he propylsulfonica cid functions graftedo ver silica surfaces in PrSO 3 H/ SBA-15 and PrSO 3 H/MM-SBA-15 catalysts were chemically identical. Therefore, any differences in TOFs betweent he two catalysts must arise solely from diffusion phenomena. However,despite their similara cid site loadings,t he surface coverage of acid sites was higher over the macroporous material (which possessed al ower surfacea rea). Note that the higher Sl oadings accessible through the HSPG method offer acid loadings of approximately 1.5 mmol g À1 ,a pproximately twice those obtained through sulfonic acid grafting in toluene (0.6-0.8 mmol g À1 ). [2a] Molecular dynamics simulations and adsorption calorimetry revealed that cooperative effects between silanol and sulfonic acid functions can weaken their acidity in PrSO 3 H/MCM-41 owing to hydrogen bondinga nd associate sulfonate reorientation. [32] However,s uch effects only operated for low sulfonic acid loadings, and were absent on crowded surfaces such as those employed in this work;h ence, cooperative effects were not expected to influence the catalytic performance.

Esterification of modelcarboxylic acids
The catalytic performance of mesoporousa nd macroporousmesoporoussulfonic acid silicas was evaluated in the esterification of propanoic (C 3 ), hexanoic (C 6 ), and lauric acids (C 12 )w ith methanol to explore the influence of the macroporeso nt he reactivityu nder previously optimized conditions. [2a] Because both catalysts possesseds imilara cid loadings and strength, any differences in activity must arise from their pore architecture. Both sulfonic acid catalysts were active for methylic esterification of the C 3 ,C 6 ,a nd C 12 acids ( Figure S11), which were 100 %s elective to their corresponding methyl esters. The rate of esterification decreased with increasing alkyl chain length owing to polar and stericeffects. [33] The associated turnover frequencies (TOFs) for carboxylic acid esterification were similaro ver both catalysts for the C 3 and C 6 acids (Figure 1), whereas the TOF for lauric acid over the hierarchical PrSO 3 H/MM-SBA-15 was twice that observed for the purely mesoporous PrSO 3 H/SBA-15 ( Figure S12). This rate enhancement for the bulky lauric acid esterification could be explained in terms of improved sulfonica cid accessibility through (i)faster in-pore diffusion of the reactant/ester product;( ii)shorter mesopore channel lengths owing to truncation by macropores;a nd (iii)ani ncreased number of mesopore openings, which may boost the sulfonic acid density at mesopore entrances. [23] Esterification of thermal pyrolysis bio-oil The performance of both sulfonic acid silicas was also assessed for the upgrading of ab io-oil produced by thermalf ast pyrolysis of oak woodchips at ab ench-scale, continuous fluidized bed reactor at 500 8C. Some physicochemical properties of the parentb iomass feedstock are presented in Ta ble S2,a nd of the crude bio-oil in Table S3. Although the bio-oil possessed asimilar calorific value to the woodchips, the volumetric energy density of the former was significantly highert han that of the originalb iomass,w hose density waso nly 600-900kgm À3 .T he bio-oil contained 23 wt %w ater,t ypical of fast pyrolysis biooil, [6b, 34] although the total acid number (TAN) of 61.6 mg KOH g À1 measured by the ModifiedD 664A acid number titration method [35] was relatively low. [34] Figure 2c ompares TOFsf or total acid removal (as determined by KOH titration)t hrough catalytic esterificationw ith methanol, and the corresponding reaction profiles for total acid conversion (Figure 2inset). The PrSO 3 H/MM-SBA-15 catalyst was almostt hree times more active in terms of TOF,a nd converted twice the amount of acid than the PrSO 3 H/SBA-15 after 6h.B ecause the pyrolysis oil contains numerous bulky compounds as described in Ta ble 2, Ta ble 3, and Ta ble S4, we attributed the superiorp erformance of the hierarchical catalyst to improved active site accessibility akin to that for laurica cid esterification. The carboxylic acid constituents of fast pyrolysis bio-oils may drive low level (< 5%)a utocatalytic esterification. [36] This was consistent with ac ontrol experiment in the absence of anys ulfonic acid catalyst, which revealed < 8% The chemical composition of the crude andu pgraded biooil following catalytic treatment by PrSO 3 H/MM-SBA-15 were analyzed in detail by GC GC-time-of-flight mass spectrometry (GC GC-ToFMS), and the resulting 2D chromatograms are shown in Figure 3. For both the crude and upgraded bio-oils, the chromatographic space was divided into six discreet mo-lecular groups:a cids and esters;a ldehydes and ketones( including furanoicsa nd cyclic carbonyls);h ydrocarbons (saturated and unsaturated non-aromatic);a romatic hydrocarbons; phenolic compounds;a nd sugars.C ompounds that could not be identified by the library and/ord id not meet the required identification criteria (as detailed in the Supporting Information) were classified as "unidentified". Am ore detailed classification of each molecularg roup and their relative chromatographic area is presented in Ta ble 2. Almost completel oss of organic acids (from 19.7 to 0.9 %) and as ignificant decrease in phenolics, ketones,a ldehydes, and sugars was observed followingc atalytic upgrading, accompanied by as ignificant increase in ester and alcohol components, consistent with esterification. Additional details on the removal/formation of specific phenolics, ethers, and carbonyls is presentedi n Ta ble S4.A cetic acid was the major organic acid in both the crude and upgraded bio-oils.E sters with relative areas > 0.1 in the crude andu pgraded bio-oils are presentedi nT able 3.
Methyl acetate accountedf or 10.8 %o ft he total chromatographic area of the esterified bio-oil, as compared to only 1.4 %o ft he crude bio-oil, alongside ar ange of methyl and dimethyl esters from C 3 -C 11 compounds. Identifiable ethers were mainly C 3 -C 6 methoxy-compounds, with 1,1,2,2-tetramethoxyethanep redominant. Considering phenolics, upgrading princi-   pally removed methoxy-phenols,w hereas cresol and catechol derivatives were recalcitrant.T he increase in alcohols appeared to arise from glycolaldehyde dimethyl acetal (GDA) formation from levoglucosan. [37] Previous studies have revealed that levoglucosan can be transformed in alcohol media by acid catalysts to methyll evulinate, through intermediate glycolaldehyde( GA) formation [38] (which may itselff orm glycolaldehyde dimethyl acetal). GA and GDA were detected in the upgraded bio-oil, supporting this proposed reaction pathway.F uture work will address the recyclability of PrSO 3 H/MM-SBA-15 for the esterification of real bio-oils, wherein we expect strong adsorption of organicst hat will requiret he development of low-temperature regeneration protocols that avoid decomposition of the grafted sulfonate.
In summary,G C GC-ToFMS analysisc onfirmed that PrSO 3 H/ MM-SBA-15 was an effective catalyst for the esterification of a real thermal pyrolysis bio-oil, significantly reducingt he bio-oil acidity throughe sterification of organic acids under mild reaction conditions.

Conclusions
Mesoporousa nd hierarchical macroporous-mesoporous (MM) propyls ulfonic acid (PrSO 3 H) silicas were synthesized by hydrothermals aline-promoted grafting of the pre-formed architectures. The textural properties of the parents ilicas were unperturbed by sulfonation, which resulted in similars ulfonica cid loadings and strengthsf or both pore networks. The turnover frequencies for catalytic esterification of model C 3 -C 12 carboxylic acids with methanold ecreased with alkyl chain length over both materials, however the introduction of 200 nm macropores into the SBA-15 framework doubled the activity per acid site for the bulkiest lauric acid,w hich was attributed to enhanced mass transport and actives ite access, and ah igher À PrSO 3 Hs urface density.M acropore incorporation also enhanced the esterification activity for the upgrading of ar eal bio-oil derived from thermalf ast pyrolysis of oak woodchips; the TOF for total organic acid removali ncreased three-fold relative to the mesoporouss ulfonic acid silica, which was also attributed to superior in-pore mass transport and actives ite accessibility.T he total acid number was reduced by 57 %o ver a 6h reaction at 85 8Cu sing the hierarchical PrSO 3 H/MM-SBA-15 catalyst. GC GC-time-of-flight mass spectrometry (GC GC-To FMS)c onfirmed that catalytic upgrading removed almost all organic acids, and significantly lowered the concentration of reactive,p henolic, aldehyde, and ketone components, accompanied by the formation of carboxylic acids methyl esters and ethers.

Experimental Section
Full details of the catalyst synthesis, bulk and surface characterization (TEM, XRD, N 2 porosimetry,D RIFTS, XPS, TGA, pyridine adsorption/DRIFTS, propylamine adsorption/TGA-MS), and catalytic esterification and bio-oil analysis protocols are provided in the Supporting Information.