Exploration of New Biomass‐Derived Solvents: Application to Carboxylation Reactions

Abstract A range of hitherto unexplored biomass‐derived chemicals have been evaluated as new sustainable solvents for a large variety of CO2‐based carboxylation reactions. Known biomass‐derived solvents (biosolvents) are also included in the study and the results are compared with commonly used solvents for the reactions. Biosolvents can be efficiently applied in a variety of carboxylation reactions, such as Cu‐catalyzed carboxylation of organoboranes and organoboronates, metal‐catalyzed hydrocarboxylation, borocarboxylation, and other related reactions. For many of these reactions, the use of biosolvents provides comparable or better yields than the commonly used solvents. The best biosolvents identified are the so far unexplored candidates isosorbide dimethyl ether, acetaldehyde diethyl acetal, rose oxide, and eucalyptol, alongside the known biosolvent 2‐methyltetrahydrofuran. This strategy was used for the synthesis of the commercial drugs Fenoprofen and Flurbiprofen.


Introduction
The vast majority of known organic transformations require use of as olvent. Solvents are essential not only for running a reaction, but also for the separation and purification of target products.A saresult, solvents usually constitute over 80 %o f all materials needed for the successful accomplishment of a typical synthetic transformation. [1] However,m ost commonly used organic solvents are derived from fossil resources, are not renewable, andh ave high toxicities. This can cause serious environmental and economic issues forl arge-scale chemical processes. One of the key directions of modern green chemistry is the minimization,e limination, or replacement of these solvents. [2] In this regard, so-called "solvent-free'' reactions have significant potential. [3] However,m ost of them are not really solvent-free and requirel arge amounts of organic solvents for the workup and purification.I nm ost cases, the research tasks cannotb ea chieved without solvents. Nevertheless, undesirable solvents can be replacedb ys ustainable/renewable alternatives. For instance, liquidso rc hemicals with low melting points, available from renewable resources, can fill the gap. [4] Particularly,m any chemicals derived from biomasss hare common properties with organic solvents derived from fossil resources. Importantly,m ost biomass-derived chemicals fulfill many of the criteria for green solvents as proposed by Gu and Jerome, such as availability,r enewability,l ow toxicity,b iodegradability,and reasonable prices. [4] The main biomass-derived solvents used in organic synthesis today are glycerola nd its acetals,s everal low-meltingm ixtures of carbohydrates, esters of lactic acid andg luconic acid, 2methyltetrahydrofuran (2MeTHF), cyrene (Cyr), limonene (Lim), p-cymene (Cym), and g-valerolactone (GVL;F igure 1). [2,4,5] The polar protic solvents like glycerol, carbohydratesa sw ell as esters of lactic acid and gluconic acid are mainly used in condensation reactionsa nd have found limited application in transformations involvingo rganometallics. [4] The polar aprotic (2MeTHF,G VL, Cyr)a nd nonpolara protic (Lim, Cym) solvents are far more popular and have been used for many classical transformations including transition metal (TM)-catalyzed cross-couplings [2, 4, 5c] and several CÀHa ctivations. [5c-e, 6] We should also highlight ethyl acetate (EtOAc), whichi sr eadily availablef rom biomass and is often overlooked in the context of biomass-derived solvents.
In addition, some biomass-derived chemicals have been proposed as green solvents, such as isosorbide dimethyl ether (Me 2 Isos) and diethyls uccinate (Et 2 Suc), both derived from cellulose but, to our knowledge, have not been examined for this purpose. [7g, 8] Further new andy et-unexplored candidates for biomass-derived solvents in organic synthesis may include acetaldehyde diethyla cetal (Acetal), which is readily available from ethanol, g-terpinene( g-Terp) and a-pinene( a-Pin) available from various coniferous plants, eucalyptol (Euc) from eucalyptuso il, and rose oxide (RoseOx) from rose oil ( Figure 1). The availability of these solvents can be judged based on their prices, which are comparable with those of common organic Ar ange of hitherto unexplored biomass-derived chemicals have been evaluated as new sustainable solvents for al arge variety of CO 2 -based carboxylation reactions. Known biomassderived solvents (biosolvents) are also included in the study and the resultsa re compared with commonlyu sed solvents for the reactions. Biosolventsc an be efficientlya pplied in av ariety of carboxylation reactions, such as Cu-catalyzed carboxylation of organoboranes and organoboronates, metal-catalyzed hydrocarboxylation, borocarboxylation, and other relatedr eac-tions. For many of theser eactions, the use of biosolvents provides comparable or better yields than the commonly used solvents. The best biosolvents identified are the so far unexplored candidates isosorbide dimethyl ether,a cetaldehyde diethyl acetal, rose oxide, and eucalyptol, alongside the known biosolvent 2-methyltetrahydrofuran. This strategy was used for the synthesis of the commercial drugs Fenoprofen and Flurbiprofen.
solvents and can decrease with further development of technologiesi nbiorefinery (see the Supporting Information, SchemeS2). It should be emphasized that most of the solvents considered in the work here are quite safe and are used in large quantities in the food industry as flavora nd fragrance ingredients. [9] Low toxicity and biodegradability is particularly inherentt oA cetal, Me 2 Isos, Et 2 Suc, Euc, RoseOx, and GVL. Unfortunately,i nformation on overall environmental impacts and full life-cyclea ssessments (LCAs) of the solvents introduced here remainlimited. [7,10] In the frame of our ongoing research program on CÀC bond-forming reactions involving CO 2 , [11] we became particularly interested to examine the use of biomass-derived solvents in av ariety of carboxylative transformations (Scheme 1B). Utilization of CO 2 , [12] and particularlyd evelopment of CÀCb ondforming reactions involving CO 2 , [13] is ah ighly promising field of research that potentially can solve many globali ssues,s uch as replacement of depleting natural resources. [14] Previously reported carboxylations were typically performed in DMF,d ioxane, toluene, and other related solvents, [12,13] which are highly undesirable from the perspective of industry and green chemistry (Scheme 1A). To our knowledge,b iomass-derived solvents have not been appliedf or any transformationi nvolving CO 2 .
The main goal of the presents tudy was to examine the suitability of aw ide range of biomass-derived chemicals as solvents for carboxylation reactions, including the known solvents 2MeTHF,G VL, Cyr,E tOAc, Lim, and Cym, and the unexplored solvents Acetal, Me 2 Isos, Et 2 Suc, g-Terp, a-Pin, Euc, and RoseOx. We startedo ut by screening the above listed biosolvents in two model reactions-carboxylations of in situ-generated organoboranes and organoboronates. The carboxylative transformationo fo rganoboronates to carboxylic acids in biosolvents was proven to be useful in the preparation of two commercial drugs.F inally,s ome of the bests olvents were evaluated in aw ide range of carboxylation reactionsu sing CO 2 .A mong others, these reactions included hydrocarboxylations,b orocarboxylation, and carbocarboxylation. The biomassderived solvents were also successfully applieda se xtraction media during product isolation. Overall,b iomass-derived solvents, and in particulars ome of the solvents tested for the first time in this study,h ave a high potentialt or eplacec ommono rganic solvents in the near future.

Results and Discussion
As as tartingp oint, we have examined the carboxylation of organoboranes, which can be easily generated in situ by hydroboration of the corresponding olefins with 9-BBN (9-borabicyclo[3.3.1]nonane dimer). [15] Analysiso ft he influence of various parameters on the outcomeo ft he reactionw ere conducted on 4methylstyrene 1a using 2MeTHF as as olvent, which already provedt ob easuitable media for reactions involving organometallics (Table 1). [2,4,5] Carefule x-
Having established that 2MeTHF is as uitable solventf or hydrocarboxylations, we continued to investigate the scope of this reaction( Scheme 2). Styrenes 1a-h and primary olefins 1i and 1j consistently provided moderate to excellent yields of the corresponding acids (2a-j, 7 3-98 %). For theses ystems, the hydroboration step proceeded as an anti-Markovnikova ddition, eventually leadingt ot erminal carboxylic acids with excellent regioselectivity. [15,17] Other regioisomers were not observed. Further studies showed that internal alkenes 1k-m are far less reactive than terminal olefins (2k-m,5 2-71 %). However,f or these substrates the reactions proceeded with excellent regioselectivities owing to the steric controlo ft he hydroboration step. [15,17] The reduced reactivity of internal alkenes allowed us to conductr egioselective hydrocarboxylation on nonconjugated dienes possessing one internal and one terminal double bond, 1n and 1o.I nt his case, we used 0.7 equivalents of 9-BBN,w hich allowed us to prepare only the hydrocarboxylation product of the terminal double bond (2n,5 8%; 2o, 73 %). These observations may explain why the hydrocarboxylation of 1a workedi nR oseOx and g-Terp, which have internal double bonds (Figure 2). TheC u-freeh ydrocarboxylation of stilbenes and b-substituted styrenes 1p-s also worked well in 2MeTHF and the hydrocarboxylation products 2p-s were observed in good yields (48-82 %) and excellent regioselectivity.
Next, we examined the substrate dependence of carboxylations of organoboronates in biomass-derived solvents (Scheme 3). The best solvents identified for organoboronates ( Figure 2) were screened on several substrates. These studies indicated that depending on the substrate, the efficiency of the used solventd iffers. For most aromatic systems, the best solventw as Me 2 Isos. However, for thiophene 4j,2 MeTHF performed slightly better (78% vs. 84 %). For benzylboronic acid pinacol esters, 2MeTHF proved to be the best solvent, outperforming Me 2 Isos by 20 %( 4p,Scheme 3).
The developed methodology for benzylboronic acid pinacol esters could be furthera ppliedf or the synthesis of the commercialn onsteroidal anti-inflammatory drugs Fenoprofen (4r, 60 %) and Flurbiprofen( 4s,5 3%). Notably,t he starting materi-als of thesed rugsw ere prepared in two steps from the corresponding commercially available aldehydes.T hese steps involved sequential Wittig olefination and Cu-catalyzed hydroboration reactions, which were conducted in biomass-derived solvents (2MeTHF and Cym, respectively;s ee the Supporting Information).
The observed excellent performance of various biomass-derived solvents for Cu-catalyzed carboxylations of in situ-generated organoboranes and organoboronates prompted us to test these solvents on ar ange of other CÀCb ond-forming reactions involving CO 2 (Scheme4 and SchemeS4).W eb egan with the examination of Cu-catalyzedh ydroboration/carboxylation of phenylacetylene 5a, [15c] which was here performed in biomass-derived ethers (Scheme 4A). Using the conditions developedb yS krydstrup and co-workers, but applying them in 2MeTHF insteado fd ioxane, we observed am ixture of benzylmalonic acid 6a with decarboxylative hydrocarboxylation product 2b in a1 :0.4 ratio. Further analysiso ft he reaction showedt hat the decarboxylation of 6a can be complete when the reaction is performed at 150 8Cf or 36 h. This improvement allowedu st oo btain 2b as am ajor product in 80 %y ield by using 2MeTHF.A cetal and Euc also gave good yields of 2b but were less effective than 2MeTHF;w hereas, the yield of the reaction in dioxane (84 %) was comparable with the yield observed in 2MeTHF.O verall,d ecarboxylative hydrocarboxylation was not described earlier.
This was followed by examination of the hydrocarboxylation of styrenes in biomass-derived solvents. Among others, these studies involved Fe-catalyzed hydrocarboxylation of 4-methylstyrene 1a by using EtMgBr as as toichiometricr eductant (Scheme 4B). Notably,s imilarh ydrocarboxylations were already reported in ether and THF. [18] Our studies showedt hat for Fecatalyzed hydrocarboxylation of 1a,i ti sp ossible to apply biomass-derived solvents. The best results wereo btainedb y using the Fe(acac) 3 /TMEDA( tetramethylethylenediamine) system as the catalysti n2 MeTHF.I nt his case, the yield of 7a was 27 %, applicationo fo ther biomass-derived ethers did not improvet he outcomeo ft he reaction; whereas the use of THF slightly enhanced the yield of hydrocarboxylation( 27 % 2MeTHF vs. 39 %T HF). [18c] It shouldb ee mphasized that EtMgBr is now available as a3 .4 m solution in 2MeTHF,a nd this typeo f experiment can be conducted by applyinge xclusively biomass-derived solvents. The hydrocarboxylation of styrenes was also examined by using different Ni-based catalysts, which unfortunately were not successful (Scheme S4 C,D).
Further,w eh ave explored the hydrocarboxylation of acetylenes in biomass-derived solvents (Scheme 4C-E). [19] We employedd iphenylacetylene 8a to test different catalytic systems based on Ni, Cu, and Fe. Promising resultsw ere observed when using the CuF 2 /IMesHCl/NaOtBu catalytic system with triethoxysilane as ar educing agent (Scheme4D). The best solvent proved to be Acetal (61 %), whereas the yields of the hydrocarboxylation product 9a were slightly lower in 2MeTHF (41 %) and Euc (43 %). The reactionp erformed in dioxane( solvent used in the originalw ork) gave 9a in 57 %y ield. [19a] Hydrocarboxylation of 8a was also possible with the Ni(cod) 2 /CsF catalytic system and using Et 2 Zn as the reductant. In this case, the yield of 9a was only 21 %i n2 MeTHF,w hereas in MeCN (solventu sed in the original work) the product was obtained in 49 %y ield (Scheme 4C). [19b] Among Fe-based catalytic systems, moderate yields of 9a were observed with FeCl 2 used in combination with 3.4 m EtMgBr in 2MeTHF (42 %i n2 MeTHF, Scheme 4E). Similarc onditions were tested in Et 2 O( solvent used in the originalw ork) where 9a waso btained in 14 % yield. [19e] Next, we examined other carboxylative transformations. Excellent results were observed for the Cu-catalyzed borocarboxylationo fs tyrenes (Scheme 4F). [20] Particularly,w ef ound that the catalytic system based on CuCl andI CyHCl (1,3-dicyclohexylimidazolium chloride), originally developed by Popp and coworkers, [20b] operates well in biomass-derived ethers, initiating efficient borocarboxylation of 1a.I nt his case, the best solvent was Euc (85 %), but good yields of borocarboxylation product 10 a were also observed in 2MeTHF (81 %), whereas Acetal (44 %) wasf ar less effective. For comparison, the borocarboxylation of 1a performed in THF( solvent used in the original work) gave 10 a in 78 %y ield. [20b] We also explored the carbocarboxylation of olefins, whichi s knownt oproceed under the influence of aw ide range of cata-lysts based on both early and late transition metals. [21] Screening of several catalysts derived from Zr and Ti as well as reducing agents showed that carbocarboxylation of 4-methylstyrene 1a can be performed in biomass-derived solvents (Scheme 4G). The best results were observed when using Cp 2 ZrCl 2 (zirconocene dichloride) as catalyst precursor combined with EtMgBr in Acetal (24 %). Application of other ethers as solvents did not improve the yield of 11 a.U sing THF under otherwisei dentical conditions gave 11 a in comparable 28 % yield. [21c] Similar to organoboronates, the carboxylation of organosilicon reagents can be performedi nb iomass-derived solvents (Scheme 4H). [22] The best results wereo bserved with triethoxyphenylsilane 12 a when using Cu-based catalysts. Particularly, we found that biomass-derived ethers are not the best solvents for this reaction( 2MeTHF 16 %, Acetal 0%). The best yields of benzoica cid 4a were observed when using the esters GVL (42 %) and Et 2 Suc (36 %) as solvents, CuBr as catalystp recursor,a nd CsF as ab ase. In this case, the yield of 4a could be notably improved when running the reactioni nD MA (62 %). It should be noted that the reaction does not work without the Cu catalyst.
Finally,w ee xamined TM-catalyzed direct CÀHc arboxylations. [23,24] To date, directC ÀHc arboxylations have been per-formed on azoles possessing an acidic CÀHb ond, arenes with appropriate directingg roups, [23] and terminala cetylenes. [24] Our studies on phenylacetylene 5a indicated that Cs 2 CO 3 alone can initiate direct CÀHc arboxylation in 2MeTHF,a lbeit with only 20 %y ield of the isolated product (Scheme 4I). The yield was improved to 31 %b ys witching to GVL. Furtheri mprovements were achieved by using the catalytic system developed for carboxylation of organoboranes and organoboronates. The best yields of 4o were observed in 2MeTHF and Acetal( 76 and 63 %, respectively), whereas GVL turned out to be far less effective with the Cu catalyst( 27 %). The best conditions were also tested with THF where 4o was obtainedi n3 8% yield. Unfortunately,a ll attempts to accomplish CÀHc arboxylation of azoles in biomass-derived solvents failed (Scheme S4 A,B).
For isolation andp urificationo ft he obtained carboxylic acids, we mainly used acid-basee xtraction techniques. Analysis of different renewable solvents for extraction showedt hat Et 2 O, which is readily availablef rom ethanol, but is not popular in industry owing to its volatility and flammability,c an be replacedb yr enewable 2MeTHF, Acetal, diethoxymethane, and dimethoxymethane without any noticeable drop in yields. Columnc hromatography,w ith mixtures of heptane/EtOAc or Et 2 O/pentane/HCO 2 Ha se luent, was only necessary for the pu-rificationo ft he products of decarboxylative hydrocarboxylation of phenylacetylene (Scheme 4A).

Conclusions
We have shown that av ariety of CO 2 -basedc arboxylations can be performed in biomass-derived solvents, including ar ange of previously unknown solvents. The studied media included polar aprotic biomass-derived ethers (2MeTHF,A cetal, Me 2 Isos, Cyr,E uc, RoseOx) and esters (GVL, Et 2 Suc, EtOAc), as wella s nonpolara protic unsaturated terpenes and their derivatives (g-Te rp, a-Pin, Lim,C ym). Initial studies on Cu-catalyzed carboxylation of in situ-generated organoboranes and -boronates revealed that most of the biosolvents are suitable for carboxylative transformations, with biomass-derived ethers showing the best efficiency.O ur methodology was successfully appliedt o organoboranes generated from styrenes and internal alkenes, as well as for carboxylation of aryl-, alkenyl-, alkynyl-, and benzylboronic acid pinacol esters. On the basis of the latter,w e have synthesized the commercial drugs Fenoprofen and Flurbiprofen.
Biomass-derived solvents were further applied for the hydrocarboxylation of acetylenes and styrenes, using catalysts based on Cu, Ni, or Fe. We observed moderate to good yields and excellent regioselectivities. Very good results were obtained for the Cu-catalyzed borocarboxylation of styrenes and CÀHc arboxylation of phenylacetylene.B iomass-derived ethers can also be used for the Cu-catalyzed carboxylation of triethoxyphenylsilane and the Zr-catalyzed carbocarboxylation of styrenes. Most of the reactions were examined in traditional organic solvents as ac omparison. These studies revealed thatt here is no advantage in using traditional solvents for the reactions described herein. In mostc ases, the yields obtained in traditional solvents were comparable with those in biosolvents, whereas in some cases,b iomass-deriveds olvents performed even better.B iomass-derived ethers showed the best performance, with 2MeTHF generallyb eing superior.H owever,i ti sn ot au niversal solvent. In several cases,e xcellent results were instead observed when using Me 2 Isos, Acetal, RoseOx, or Euc solvents. We believe that the biomass-derived solvents introduced herein will find broad applications in many processes currently based on traditional organic solvents.

Experimental Section
General experimental procedure for Cu-catalyzed hydroboration/carboxylation of olefins (Scheme 2) Inside of ag lovebox, a4 5mLp ressure tube was charged with the appropriate olefin (1.5 mmol), (9-BBN) 2 (1 equiv or 0.7 equiv in the case of dienes), and the corresponding dry solvent (4 mL). The flask was closed with as uitable cap, removed from the glovebox, and heated to 70 8Cf or 24 h. Afterwards, the pressure tube was transferred back to the glovebox. To the reaction mixture at 20 8C was added CsF (3 equiv) and ap reviously prepared solution of catalyst (mixture of CuI (5 mol %), IPrHCl (6 mol %), and NaOtBu (6 mol %) in appropriate dry solvent (2 mL) stirred at 20 8Cf or 30 min) was added. The pressure tube was closed with the cap and removed from the glovebox. Afterwards, CO 2 (120 mL) was added via as yringe, which was followed by stirring of the reaction mixture at 120 8Cf or 24 h. Next, the reaction mixture was diluted with Et 2 O( 30 mL) and transferred into a5 00 mL separating funnel. The resulting mixture was extracted with saturated NaHCO 3 solution (3 30 mL). The resulting basic solution was washed with Et 2 O (15 mL), acidified (50-55 mL 6 m HCl), and extracted with Et 2 O( 3 30 mL). The resulting solution of Et 2 Ow as distilled to dryness to give the corresponding acid.
In the cases of Me 2 Isos, GVL, and Et 2 Suc, the basic solution was washed with either CH 2 Cl 2 or Et 2 O( 3 15 mL), and the final Et 2 O solution was washed with distilled water (3 15 mL) before evaporation. Other renewable solvents such as 2MeTHF,A cetal, diethoxymethane, or dimethoxymethane can replace Et 2 Ow ithout any noticeable difference (the difference was in the range AE 3%). Similarly,t he saturated solution of NaHCO 3 can be replaced by a2m solution of KOH.
Generale xperimental procedure for Cu-catalyzedc arboxylation of organoboronates (Scheme3 ) Inside of ag lovebox, a4 5mLp ressure tube was charged with the appropriate organoboronate (0.8 mmol), CsF (3 equiv), and corresponding dry solvent (2 mL). This was followed by addition of a previously prepared solution of the catalyst (mixture of CuI (5 mol %), IPrHCl (6 mol %), and NaOtBu (6 mol %) in an appropriate dry solvent (2 mL) was stirred at 20 8Cf or 30 min). The pressure tube was closed with the cap and removed from the glovebox. Afterwards, CO 2 (120 mL) was added via as yringe, which was followed by stirring of the reaction mixture at 120 8Cf or 24 h. Next, the reaction mixture was diluted with Et 2 O( 30 mL) and transferred into a5 00 mL separating funnel. The resulting mixture was extracted with saturated NaHCO 3 solution (3 30 mL). The resulting basic solution was washed with Et 2 O( 15 mL), acidified (50-55 mL 6 m HCl), and extracted with Et 2 O( 3 30 mL). The resulting solution of Et 2 Ow as distilled to dryness to give the corresponding acid.
In the cases of Me 2 Isos, the basic solution was washed with either CH 2 Cl 2 or Et 2 O( 3 15 mL), and the final Et 2 Os olution was washed with distilled water (3 10 mL) before evaporation. Other renewable solvents such as 2MeTHF,A cetal, diethoxymethane, or dimethoxymethane can replace Et 2 Ow ithout any noticeable difference (the difference was in the range AE 3%). Similarly,t he saturated solution of NaHCO 3 can be replaced by a2m solution of KOH.