New AB type monomers from lignocellulosic biomass

Introduction

An overwhelming amount of chemicals consumed by polymer industries are derived from petroleum resources. The fossil precursors of these materials are nonrenewable and thereby lack long-term sustainability. In general, the types of structures available from petroleum feedstocks are limited to non-oxidized electron rich aromatics, olefins, alkanes and derivatives thereof. The development of efficient methods for the synthesis of monomers for polymer applications using renewable feedstocks has received intense scrutiny [1], [2], [3], [4]. This increased activity addresses issues related to long-term sustainability and also opportunistic utilization of the diversity inherent in biomass-derived structures. There are three major classes of biomass that provide excellent feedstocks for the synthesis of useful monomers. They are lignin [5], oil seeds [6], and carbohydrates [7] (Fig. 1).

Fig. 1:

Paradigm of biomass utilization.

Lignin is a heterogeneous polymer and is generally derived from three main phenolic building blocks [8]. Depolymerization [9] and/or modification [10] of lignin provides access to monomers which contain an aromatic core structure with multiple sites of oxygenated functionality. Indeed, the major structural difference between petroleum derived (non-oxidized) aromatic monomers and those from lignin is the presence of phenolic hydroxyl groups and other oxidized functional group substitutions on the aromatic ring.

Triglycerides, prepared in nature from glycerol and fatty acids, are the major components of the oil pressed from oil seeds such as soy but also including nonedibles such as crambe. The composition (degree of unsaturation or olefin functionalization) of the fatty acid depends on the species of plant from which the oil is derived. Researchers have employed fatty acids, their esters, and compounds derived from the fatty acids as monomers or building blocks in polymer chemistry [11], [12], [13], [14], [15], [16]. Key features of monomers derived from oil seeds are: (1) flexible aliphatic chains spacing functionalization sites from polar attachment points (ester linkage) (2) diversity in structure modulated by altering the polyol core or secondary olefin functionalization.

Cellulosic biomass provides access to compounds with a furan skeleton [17]: a structural feature not accessible from fossil sources. Two compounds derived from cellulose, 5-(hydroxymethyl)furfural [18] (HMF, 1) and 2,5-furandicarboxylic acid [19], [20], [21], [22] (FDCA, 2), have been identified as the top value added feedstocks for monomer synthesis [23]. HMF has two functional groups at different oxidation states that can be selectively manipulated to provide access to other furan-based monomers (Fig. 2). Hydroxyaldehyde 1 is an aromatic dehydration product derivative of ketohexoses or their polymers such as the cellulose contained in agricultural wastes [24].

Fig. 2:

Diversified platform chemicals from lignocellulosic biomass: 5-(hydroxymethy)lfurfural (1), 2,5-furandicarboxylic acid (2), 5-(hydroxymethyl)furan-2-carboxylic acid (3), 2,5-bis(aminomethyl)furan (4), 2,5-dicyanofuran (5), 2,5-diformylfuran (6), 2,5-bis(hydroxymethyl)furan (7).

The HMF core structure offers opportunity for modification into symmetrically substituted products such as 2,5-furandicarboxylic acid (FDCA, 2, Fig. 2, AA type monomer) or for chemoselective modification to 5-(hydroxymethyl)furan-2-carboxylic acid [25] (HMFCA, 3, Fig. 2, AB type monomer). The chemical literature contains examples wherein attention was paid to AA (diester [26], [27], [28], diacid 2 [29], dicyano 5 [30], dialdehyde 6 [31]) and BB (diamine 4 [32], diol 7 [33], [34]) type monomers for copolymerization. Particularly of note, the toxicity of HMF and some of its derivatives has been established and is considered low [35].

In our work, we have sought unconventional building blocks [36] by utilizing three important renewable materials: (1) lignin [10], (2) oil seeds [12], [37], [38]], [39], and (3) carbohydrates [40], [41], [42], [43], [44]. Our group’s research is focused on the development of novel methods for the conversion of renewable resources to feedstock [42], [45] chemicals and for their use in polymer synthesis [12], [14], [37], [38], [39], [40], 43], [44]. We have recently developed a method for the direct valorization of cellulose derived renewables by aromatic-upgrading (Scheme 1) [42]. The protocol utilizes a Diels–Alder reaction between benzyne and a furanic monomer. The benzyne generation employs diazotization of anthranilic acid (8) to form benzenediazonium-2-carboxylate (9): the precursor of benzyne (10) under thermolytic conditions. Benzyne’s reactivity as a dienophile is driven by strain and less by its electronic nature. Its reactions are not easily characterized according to the electronic nature of the diene (normal vs. inverse electron demand).

Scheme 1:

Valorization by Diels–Alder reaction with lignocellulosic biomass-derived furan-dienes (12–17): preparation of differentially 1,4-disubstituted-7-oxabenzonorbornadienes (18–23).

We have now combined our valorization by cycloaddition strategy with AB type furanics derived from 3 to prepare novel AB type monomers (mixed diesters, hydroxyesters or hydroxyacids), which are amenable to homopolymerization [46], [47], [48]. The development of such monomers from readily available furanic-substrates should facilitate further developments in bio-based materials applications; the execution of polycondensation reactions using AB type monomers is simplified by the perfect reaction stoichiometry of pure samples. Herein we describe the valorization of a series of differentially substituted AB type furan-dienes (12–15) framed by two symmetrically substituted furan-dienes (16 and 17, Fig. 3).

Fig. 3:

Furan-dienes described in this study.

Results and discussion

The preparation of furan-dienes 12–15 from commercially available 5-(hydroxymethyl)furan-2-carboxylic acid (3), began by chemoselective Fisher esterification [46]. The aromatic carboxy moiety underwent acid-mediated methanolic dehydration and only minor products of etherification or chlorodehydration at the primary benzylic alcohol were isolated after 6 h of reaction (Scheme S1, ESI). The product, methyl 5-(hydroxymethyl)furan-2-carboxylate (S-1) was isolated by flash chromatography [49] in good yield (87%). A similar reaction wherein ethanol was the solvent led to isolation of ethyl 5-(hydroxymethyl)furan-2-carboxylate (S-2) in lower yield (60%, Scheme S2, ESI).

During the benzyne-Diels–Alder reaction, primary hydroxyl groups can undergo redox reaction with 9 which destroys both reactive-intermediate-precursor and substrate [50]. Since protection was required, a mixed ester was prepared from S-1 by 4-(dimethylamino)pyridine (DMAP) catalyzed reaction with acetic anhydride in dry acetone (HPLC grade, Scheme S3, ESI). The product, methyl 5-(acetoxymethyl)furan-2- carboxylate (12), was isolated by liquid–liquid extraction with ethyl acetate in excellent yield (97%). The same reaction conditions afforded ethyl 5-(acetoxymethyl)furan-2-carboxylate (13) from S-2 in similar excellent yield (98%, Scheme S4, ESI).

The hydroxyl group of S-1 underwent reaction with tert-butyldimethylsilyl chloride in an imidazole-mediated coupling at room temperature in dichloromethane. As such, S-1 was orthogonally protected with a robust and bulky silyl ether in good yield (88%, Scheme S5, ESI). Isolation of methyl 5-(((tert-butyldimethylsilyl)oxy)methyl)furan-2-carboxylate (14) required flash chromatography. The hydroxyl group of S-1 underwent pyridinium p-toluenesulfonate catalyzed reaction with 3,4-dihydro-2H-pyran to afford methyl 5-(((tetrahydro-2H-pyran-2-yl)oxy)methyl)furan-2-carboxylate (15) in good yield (89%, Scheme S6, ESI). This compound also required flash chromatography for isolation.

Dimethyl furan-2,5-dicarboxylate (16) was prepared as described in our previous report [42]. The final substrate employed in this study, 2,5-bis(acetoxymethyl)furan (17), was prepared with no flash chromatography from HMF (87% over two steps, Scheme S7, ESI). A complete experimental section including supporting spectra has been prepared in the Electronic Supporting Information (ESI).

With a series of differentially 2,5-disubstituted furan-dienes in hand, the established protocol [42] for cycloaddition with benzyne was slightly modified (ESI) and applied to methyl 5-(acetoxymethyl)furan-2-carboxylate (12) while varying the stoichiometric excess of anthranilic acid (8, Scheme 1). The results have been presented in Table 1 (entries 1 and 2). A greater excess of 9 (two molar equivalents) gave good yield (89%, Table 1, entry 2) of Diels–Alder adduct 18. The ethyl ester derivative 13 (78%, Table 1, entry 3) afforded more side products than the methyl ester 12 at the reaction stoichiometry examined.

Table 1:

Benzyne Diels–Alder reaction: preparation of 7-oxabenzonorbornadienes.

Entry	Substrate			Product			Yielda
	R1	R2	Compound	R1	R2	Compound
1	COOMe	CH2OAc	12	COOMe	CH2OAc	18	72b
2	COOMe	CH2OAc	12	COOMe	CH2OAc	18	89
3	COOEt	CH2OAc	13	COOEt	CH2OAc	19	78
4	COOMe	CH2OTBDMS	14	COOMe	CH2OTBDMS	20	85
5	COOMe	CH2OTHP	15	COOMe	CH2OTHP	21	85
6	COOMe	COOMe	16	COOMe	COOMe	22	86
7	CH2OAc	CH2OAc	17	CH2OAc	CH2OAc	23	63c
8	CH2OAc	CH2OAc	17	CH2OAc	CH2OAc	23	85b
9	CH2OAc	CH2OAc	17	CH2OAc	CH2OAc	23	92

1. ^aIsolated yield; ^b1.5 molar equivalents of 8; ^c1.0 molar equivalents of 8.

The substrate scope and substituent-effects were investigated by carrying out reactions with 14 (85%, Table 1, entry 4) and 15 (85%, Table 1, entry 5); greater steric bulk in the hydroxyl protecting group had a slight negative impact upon the yield of 7-oxabenzonorbornadienes. Substrates 16 (86%, Table 1, entry 6) and 17 (92%, Table 1, entry 9) were also subjected to the cycloaddition to establish the electronic demands of this strain-driven Diels–Alder reaction. The electronic affects of methylcarboxylate versus acetoxymethyl substituents was compared. Better yields were observed with relatively electron-rich diacetate 17 as would be expected for a normal electron demand Diels–Alder reaction. The relationship between reaction stoichiometry and adduct yield with the optimal substrate (17) was further explored (Table 1, entries 7–9). The results from these studies suggest that the reaction outcome of the benzyne-Diels–Alder reaction with AB type furan-dienes was more similar to electron deficient AA type furan-dienes (Table 1, compare 2 with entry 6) as opposed to electron rich BB type furan dienes (Table 1, compare entry 1 with entry 8).

To achieve access to a series of biosourced hydroxyester AB type monomers, it was important to define conditions which could selectively unmask the hydroxyl substituent from protected bicyclic intermediates. Inspired by facile deprotection of THP ethers with mild Dowex resin [51], methyl 4-(hydroxymethyl)-1,4-epoxynaphthalene-1(4H)-carboxylate (24) was procured following acidic methanolysis of 21 (Scheme 2). Facile isolation required filtration to remove the Amberlyst 15 [52] resin and concentration under reduced pressure to remove solvent and volatile 2-methoxytetrahydropyran. This deprotection strategy compared very favorably with an initial attempt to utilize pyridinium p-toluenesulfonate in methanol (Scheme S8, ESI).

Scheme 2:

Preparation of methyl 4-(hydroxymethyl)-1,4-epoxynaphthalene-1(4H)-carboxylate (24).

Symmetrically disubstituted 7-oxabenzonorbornadienes are known to undergo efficient catalytic hydrogenation at ambient temperature and pressure with minute palladium catalyst loadings [42]. A series of novel 7-oxabenzonorbornenes was prepared (Table 2, entries 1–4 and 6) in excellent yield (>90%). In a delightful turn of events, methyl 4-(hydroxymethyl)-3,4-dihydro-1,4-epoxynaphthalene-1(2H)-carboxylate (28) was isolated from the hydrogenation mixture of 21 also in excellent yield (98%, Table 2, entry 4).

Table 2:

Catalytic hydrogenation: preparation of 7-oxabenzonorbornenes.

Entry	Substrate			Product			Yielda
	R1	R2	Compound	R1	R2	Compound
1	COOMe	CH2OAc	18	COOMe	CH2OAc	25	94
2	COOEt	CH2OAc	19	COOEt	CH2OAc	26	97
3	COOMe	CH2OTBDMS	20	COOMe	CH2OTBDMS	27	96
4	COOMe	CH2OTHP	21	COOMe	CH2OH	28	98
5	COOMe	COOMe	22	COOMe	COOMe	29	95
6	CH2OAc	CH2OAc	23	CH2OAc	CH2OAc	30	94

1. ^aIsolated yield.

Amberlyst 15 mediated dehydroaromatization of 7-oxabenzonorbornenes [42] is facilitated by substituents with increasing aptitude to stabilize a developing positive charge at their benzylic positions. Previously, bicycle 29 with electron withdrawing substituents at its benzylic bridgehead required elevated temperature and toluene as solvent to afford complete conversion of substrate. In that case, the isolated yield of 34 could be maximized to 74% with concurrent formation of 1,4-naphthalenedicarboxylic acid [42]. Initially, it was presumed that AB type bicycles would smoothly aromatize upon treatment with Amberlyst 15 owing to the stabilizing influence of the oxymethylene substituents.

Surprisingly, this technique was not particularly amenable to obtaining oxymethyl substituted naphthalenes. Concentration seemed to have little effect on the reaction outcome (Table 3, compare entry 1 with entry 2). Steric bulk (Table 3, compare entry 1 with entry 3) and hydrolytic infirmity (Table 3, entry 1 compare with entry 4) both negatively affected the reaction’s outcome. As described in a previous report [42], and repeated here, electron withdrawing groups at both bridgeheads significantly limits acid mediated ring opening reactions (compare Table 3, entries 5 and 6). In some cases (Table 3, entries 1–3 and 5) a large amount of the mass balance was recovered starting material. Silyl ether 33 (Table 3, entry 4), afforded neither product nor recovered substrate. In the case of 35, which contained small-stabilizing substitutions at both bridgeheads (Table 3, entry 6), conversion of starting material was complete in 90 min.

Table 3:

Amberlyst 15 mediated dehydroaromatization: preparation of naphthalenes.

Entry	Substrate			Product			Yielda
	R1	R2	Compound	R1	R2	Compound
1	COOMe	CH2OAc	25	COOMe	CH2OAc	31	49
2	COOMe	CH2OAc	25	COOMe	CH2OAc	31	55b
3	COOEt	CH2OAc	26	COOEt	CH2OAc	32	17b
4	COOMe	CH2OTBDMS	27	COOMe	CH2OTBDMS	33	0b
5	COOMe	COOMe	29	COOMe	COOMe	34	22
6	CH2OAc	CH2OAc	30	CH2OAc	CH2OAc	35	75

1. ^aIsolated yield; ^breactions carried out at 0.1 M in DCE.

A naphthalene analog of 5-(hydroxymethyl)furan-2-carboxylic acid (3) was prepared (Scheme 3) by one pot deprotection and dehydroaromatization of 25. The isolation of 4-hydroxymethyl-1-napthoic acid (36) was simple; as the mixture cooled, light colored solid separated from the reaction mixture and was isolated by suction filtration. Supposing the low yield for the product was due to partial water solubility, the filtrate was extracted repeatedly with ethyl acetate. Disappointingly, the extract was a mixture of oligomerized side products, not pure hydroxyacid 36.

Scheme 3:

Preparation of 4-(hydroxymethyl)-1-naphthoic acid (36).

Since polysulfonic acid containing macroreticular resin failed to afford high yields due to degradation of the benzylic oxymethylene moiety (Table 3), and since aqueous sulfuric acid was also inefficient (Scheme 3), another aromatization strategy was evaluated. Specifically, the conversion of AB type bicycle 26 into AB type naphthalene 32 was selected as the most important transformation in this work. While isolated yield of naphthalene product in this case was only 17% (Table 3, entry 3), the yield based on recovered starting material was 99%. This was an indication that selective conversion to product was possible, but not properly facilitated by the polysulfonic acid resin.

Inspired by the mixed anhydride dehydroaromatizations used in the preparation of phthalic anhydride [53], 7-oxabenzonorbornene 26 was successfully converted to mixed ester naphthalene 32 in good yield (82%, Scheme 4). A similar reaction employed diacetate substrate 30, and afforded naphthalene 35 in 68% isolated yield. The lower yield of BB type naphthalene 35 as compared to AB type naphthalene 32 (Scheme 4) indicates a protective influence against formation of benzylic carbeniums by extrusion of acetate anion. This can be attributed to the mesomeric effect connoted by para carboxy substitution.

Scheme 4:

Improved preparation of ethyl 4-(acetoxymethyl)naphthalene-1-carboxylate (32).

Conclusions

In this work we have demonstrated the valorization of cellulosic biomass derived chemicals. This work has extended the range of functionalized monomers accessible from non-edible biomass and highlighted the utility of Diels–Alder cycloaddition for upgrading furanic platform chemicals. Importantly, chemoselective transformations yielding HMFCA (3) and ester derivatives enabled the synthesis of unsymmetrical monomers including hydroxyesters of interest for homopolymerization. 7-Oxabenzonorbornadienes and 7-oxabenzonorbornenes were obtained in good to excellent yields, and dehydroaromatization of the latter demonstrated the potential of this approach for accessing a variety of unsymmetrical (AB type) naphthalenic monomers from biomass. Continuing efforts to prepare novel monomers from biomass are underway in our laboratory.

Supporting Information

Details and supporting spectra have been collected in the Electronic Supporting Information and are available from the publisher’s website free of charge.