The use of naturally occurring starting materials in chemistry combines several benefits. It can be first seen as a way to design new compounds or propose alternative processes towards chemical products with a lower carbon footprint. It is also an opportunity for providing added-value to biomass or byproducts of biomass production. However using what Nature offers as available resources is not limited to the search for cheaper, safer, greener chemicals and processes, but is also a basis for bioinspiration towards new architectures with subtle behavior and original properties.
In this respect, being the most abundant part of biomass, carbohydrates represent a fantastic resource, either as polysaccharides or smaller molecules. For several decades already, a lot of work has been devoted to the design and the synthesis of a wide range of carbohydrate-based compounds, with significant successes in key economic sectors such as polymers, surfactants and cosmetics [1–14]. In most cases, the properties of the products are directly related to those of the carbohydrates themselves, in particular their polarity which induces physico-chemical properties bringing either solubility assistance to low water-soluble systems or providing amphiphilic character to hydrophobic ones .
Such targets can be reached either by direct functionalization of carbohydrates, or via the use of platform molecules or scaffolds easily obtained from very available resources. Whatever the final goals are bio-based chemicals for industrial applications or analogues of glycoconjugates dedicated to biological applications, the key issue remains the connection of a sugar with another functional moiety which necessitates overcoming the specific reactivity and selectivity issues of carbohydrate chemistry. Overall, carbohydrates are a mine for bioinspiration, either as resources for green chemistry or as models for biological purposes.
Among platform molecules, hydroxymethylfurfural (HMF) has reached a status of choice, being now in the famous “10 + 4” list of most important biobased chemicals as proposed by Bozell and Petersen in the “Top 10 + 4 chemicals” list [7, 14, 16, 17]. HMF is indeed a very interesting chemical system which can be obtained in one step from several carbohydrates and is the starting point of many processes towards other useful chemical intermediates. Although HMF is a naturally occurring molecule and also found in processed foods and beverages, its availability as a chemical resource relies on the multiple dehydration of different carbohydrates. Possible mechanisms of the multiple dehydration of sugars to HMF are divided in “cyclic” or “acyclic” pathways as illustrated in Fig. 1 for fructose and glucose. Fructose is definitely the sugar which exhibits the best ability to dehydrate to HMF. However, for evident economic concerns, there is still intense research on more efficient processes able to provide HMF from other more available, cheaper, carbohydrates such as glucose or cellulose. Furthermore, the moderate stability of HMF itself and of the starting carbohydrates used to prepare it result in many possible side reactions and byproducts, which limit often the yields and the overall feasibility of the process.
HMF can undergo several types of reactions, by reduction or oxidation reactions of the two substituents at position 2 and 5, or by transformations of its aromatic backbone (Fig. 2). Several of its derivatives, such as furandicarboxylic acid (FDCA) are considered as promising precursors of biobased polymers. Recent reviews [7–9, 17–19] cover the field of HMF chemistry and provide updated information on all catalysts, media and processes for its synthesis and uses.
A glucosylated analogue of HMF exists and is referred to as glucosyloxymethylfurfural (GMF). Like HMF, GMF has been found in some natural resources and different types of food products [20–24], but it is easier to prepare it chemically [25–27] by multiple dehydration reaction of isomaltulose, a disaccharide available on the industrial scale [28–30] (Fig. 3).
The purpose of this short account is to give a brief overview of GMF occurrence, preparation and uses towards diverse types of targets. All these targets combine a carbohydrate moiety attached to another functional backbone, therefore incorporating the properties that Nature has given to the carbohydrates into the final bioinspired products.
Occurrence and synthesis of GMF and analogous glycosylated derivatives of HMF
α-d-Glucosyloxymethylfurfural (GMF) is obtained by dehydration of isomaltulose (6-O-α-d-glucopyranosyl-d-fructofuranose, also referred to as Palatinose®. Isomaltulose, used as a food in Japan since 1985,  is a very available disaccharide synthesized from sucrose by bioconvesion on the industrial scale [28, 29]. It is used industrially for the synthesis of its hydrogenated derivative, isomaltitol (Palatinit®) [31–33]. GMF’s systematic name is “5-(α-glucopyranosyloxymethyl)-2-furancarboxaldehyde” or “(5′-formyl-2′-furyl)methyl-α-d-glucopyranoside” though it is also found under the name “HMF α-d-glucopyranoside”. For clarity and consistency with most former reports in literature, the abbreviation GMF will be used throughout this account, GMF being specifically the α-glucopyranoside member of the family, while others would be referred to as “β-galacto-GMF or α-glucofurano-GMF” for example. Apart from its dehydration to GMF, isomaltulose is a precursor of other compounds which can also lead to other biobased products. This has been illustrated by several groups, notably those of Lichtenthaler and Kunz who reported work on isomaltulose hydrogenation, reductive amination and oxidative degradation to glucosylated carboxylic acids [34–42] (Fig. 4).
In the framework of our research dedicated to the use of carbohydrates as organic raw materials [1, 2, 43, 44], our group has also investigated some applications of isomaltulose [45–47] and reported an oxidative degradation of isomaltulose leading to carboxymethyl glucoside (CMG), a precursor of a bicyclic lactone which was further used as a scaffold offering alternative routes towards pseudoglycoconjugates [48–57].
With respect to its occurrence in natural resources, GMF and several analogs have been isolated in several natural products, plants, and foods, or at least their presence has been identified (Fig. 5). Like HMF, GMF is also one of the components of commercial caramel and caramel candy . Those analogues can arise either from the dehydration of some oligosaccharides via a similar pathway as for the formation of GMF from isomaltulose, or by glycosylation reactions between HMF and other untransformed sugars during a thermal treatment. Thus, all possible anomers arising from glucose, galactose, lactose, showing either an α or a β glycosidic bond were identified in such products. In the case of lactose, an acetal formed by reaction of the 4′,6′-extremity and the aldehyde of HMF was also formed, either on lactose itself or on the lacto-GMF glycoside. GMF and its 6-O-glucosylated analog were found in the extracts of steamed Rehmanniae Radix (Di Huang), an important species in Traditional Chinese Medicine [20–22]. A pyrrolo analogue with a galactosyl backbone was isolated, together with a pyrididol analogue, from the kernel of Prinsepia uniflora which is another traditional medicine used in the treatment of eye diseases . The sorbitol and mannitol ethers of HMF were identified in the extracts of the fruit of Amelanchier Canadensis . During the course of these analytical studies, other GMF analogues were obtained, notably the corresponding furanosidic derivatives (all anomers from glucose and galactose).
The availability of pure GMF does not rely on its natural occurrence but on the dehydration of isomaltulose. The conditions for its generation require careful prevention of an undesired cleavage of the intersaccharide linkage of isomaltulose. Anhydrous conditions are preferred for isomaltulose dehydration to GMF. Indeed, if aqueous acidic conditions are used, dehydration of isomaltulose mostly leads to glucose and HMF, resulting either from the late hydrolysis of the GMF glycosidic bond, or from an early isomaltulose hydrolysis followed by fructose transformation to HMF. Subsequent degradation to levulinic and formic acid can then occur. The best results were obtained in dimethyl sulfoxide (DMSO) which proved to be a better solvent than those commonly used for the synthesis of HMF from fructose such as dimethylformamide (DMF), acetonitrile or quinolone. Lichtenthaler et al.  reported that GMF was obtained in 70 % yield when isomaltulose was heated at 120 °C in DMSO in the presence of a strongly acidic ion-exchange resin (Dowex 50 WX4, H+ form) and molecular sieves. Other products of the reaction were identified as dimeric isomaltulose anhydrides (10 %), HMF, glucose (5–10 %), and some untransformed isomaltulose itself (10 %). The reaction process was adapted from batch to continuous using a flow reactor , offering reasonable availability of GMF for future uses as a bio-based building block. Koenig and co-workers  recently reported that GMF could be obtained by treating isomatulose-choline chloride melts in different acidic conditions, the best results being obtained with ZnCl2 (52 %, 1 h) or Montmorillonite (46 % yield in 15 min), thus proposing a more environmentally friendly access to HMF compared to those using organic solvents.
Synthetic approaches towards GMF analogues have also been reported, leading to compounds similar to those found in vegetal species of processed foods as shown aboved. There are two ways for accessing such sugar-furfural hybrids, either the dehydration pathway from other types of disaccharides possessing a non-reducing fructose end, or by classical glycosylation reactions from an activated glycosyl donor with HMF.
With respect to disaccharide dehydration, the classical Lichtenthaler protocol efficiently provided α-galacto-GMF, β-gluco-GMF, and β-xylo-GMF from melibiose, gentiobiose, and primeverose, after a first aluminate-promoted isomerization step of the reducing end to a fructose one (Fig. 6) . Alternatively, the glycosylation pathway was reported by Urashima showing that , heating HMF with glucose or galactose in 1,4-dioxane led to all 4 possible isomers of galactosyloxymethylfurfural, either α or β, and in the pyranosic (as major products) or furanosic forms (Fig. 6). The reaction was also possible using HMF as the solvent. The Koenigs Knorr type glycosylation of HMF could also be achieved by Ag2O or BF3-promoted condensation with glucosyl donnor by Cottier et al.  who reported the synthesis of peracetylated β-gluco-GMF. The use of the silylated derivative of HMF was found to be preferable. The reaction was recently applied to other starting monosaccharides, namely glucose, galactose and xylose leading, after base catalyzed deactetylation to the corresponding fully deprotected HMF glycosides .
GMF chemistry and derivatives thereof
GMF can be transformed into several types of compounds because it is a multifunctional substrate, with a free glucosyl moiety, and the furfural one, on which transformations can occur either on the aldehydic function or on the furanic part, and several combinations of those different reactivities have been reported and are detailed in the following sections.
Carbon–heteroatom bond formations
Reactions on the glucose moiety
The acetylation of GMF by acetic anhydride in pyridine logically provides peracetylated GMF. However, when the reaction was performed in a way to get partial esterification, the triacetyl compound having the 3-OH can be isolated among other partially acetylated products, including the 3,6-diacetylated derivative. The triacetyl derivative, having only one available OH group, could be further oxidized to the keto derivative (Fig. 7) .
Oxidation and reduction of the aldehyde function of GMF
The aldehyde function of GMF can be selectively oxidized by chlorite into the carboxylic acid leading to the corresponding furancarboxylic acid (89 %) which can in turn be transformed to the amide or to various esters  (Fig. 8). The longer alkyl chain esters were found to exhibit interesting liquid crystal properties . The corresponding nitrile could also be obtained, either via treatment of the oxime by acetic anhydride, leading to the peracetylated nitrile, or by direct reaction with hydroxylammonium chloride in DMSO for 30 min at 110 °C leading to the unprotected nitrile. This latter compound proved to readily provide a tetrazolium heterocycle by [3 + 2] dipolar cycloaddition with sodium azide .
Reduction of the aldehyde function of GMF by NaBH4 provides the GM-furfuryl alcohol (85 %), and the reductive amination leads to the GM-furfurylamine (97 %), which could be acylated with long chain acid chlorides leading to amphiphilic amides which were compared to the long chain esters previously mentioned with respect to their thermotropic liquid crystal properties  (Fig. 9). The amination was also performed with methyl and tetradecylamine leading to the corresponding furfuryl secondary amines .
Oxidations of the heterocyclic moiety of GMF, derivatives and analogues
The oxidation of the furan moiety of GMF or Ac4-GMF provides γ-keto-carboxylic acids (or their hemiketalic counterpart) in very good yields (85–90 %) in various oxidation conditions (Fig. 10) [63, 64]. Some of these oxidized derivatives were further used for the synthesis of N-heterocycles by reaction with amines or diamines. Reaction of the unsaturated ketoacid with phenyl hydrazine 1,2-diaminobenzene or 1,2-diamino-4,5-dichloro-benzene led to the corresponding pyridazinone and benzodiazepinones in 70 %, 55 % and 48 % yields, respectively.
Starting from acetylated GMF, reaction of the acetylated glucosylated hydroxybutenolide (in its cyclic form) with o-phenylenediamine led to the glucosylated benzodiazepinone (65 %) and subsequently to the glucosylated benzodiazepinone by dehydrogenation using 2,3-dichloro-5,6-dicyano-benzoquinone (DDQ) in 50 % yield. From either GMF or its β-analogue, the intermediate hydroxy butenolides could be easily generated by photooxygenation . These compounds could be either reduced to a mixture of diastereoisomers of acetylated ranunculin (60 ∼ 80 %) or transformed into the α and β 5-glucosyloxymethyl methyl levulinate in 75 % and 70 % respectively via a hydrogen transfer reaction.
In Fig. 11 are depicted some oxidations at the furanyl moiety conducted on the reduced GMF products GM-furfuryl alcohol and GM-furfurylamines (their synthesis is mentioned in a previous section, see Fig. 9). Notably, the oxidative transposition of furans into cyclohexenones (Achmatowicz reaction ) applied to GM-furfuryl alcohol generates a dihydropyranone in 95 % yield by treatment with bromine in water, via the intermediate glucosyloxy-cis-hexenedione which cyclizes spontaneously. Subsequent reaction of the dihydropyranone with hydrazine provides the glucosylated 3,6-dihydroxymethyl-pyridazine (68 %). Oxidation of perbenzylated GMF analogue with 3-chloroperbenzoic acid led to the glucosylated cis-hexenedione (79 %). Reduction of the carbon–carbon double bond by Zinc in acetic acid followed by treatment with ammonium acetate, aniline, benzylamine, or dodecylamine provided the corresponding glucosylated pyrroles in 70 %, 75 %, 84 % and 83 % yields respectively . From GM-furfurylamine, using bromine in water, the analogous “aza-Achmatowicz” reaction  efficiently provided the 6-(glucosyloxymethyl)-3-pyridinol (75 %) (Fig. 11), whereas from the corresponding secondary amine (methyl or tetradecyle), the oxidation led to the N-alkyl pyridinium betaines (71 and 70 %). Selective carbamoylation of the betaine anion of the methylpyridinium betaine led in 86 % yield to a pyridostigmine which was found to exhibit cholinesterase inhibitory activity .
Carbon–carbon bond formations at the aldehydic carbon of GMF
In the earlier investigations of GMF chemistry, only a few carbon–carbon bond forming reactions had been studied, most of the work being carbon–heteroatom bond formations as shown in previous sections. This was limited to examples reported by Lichtenthaler showing that base catalyzed aldol-type reaction of GMF with nitromethane or acetophenone could provide in good yields the corresponding 2-nitrovinyl and 2-benzoylvinyl compounds  (Fig. 12). Because olefinic or acrylic systems are potential monomers, several other derivatives having such a carbon–carbon double bond were targeted. Thus, the dicyanovinylderivative could also be obtained by aluminium oxide catalyzed Knoevenagel reaction with malodinitrile in 80 % yield. The acrylic analogue could also be obtained in 52 % yield by reaction with malonic acid in quinoline at 120–170 °C. Further heating of this acrylic derivative in the presence of copper sulfate promoted the decarboxylation to the vinylfuran analogue which proved to be unstable at this stage and could not be isolated as a pure product. Such a vinylfuran-GMF hybrid could be however obtained in the peracetylated GMF series, by carbonyl olefination of Ac4GMF using the CH2Br2/Zn/TiCl4 (Lombardo reagent ).
The other approach towards new GMF derivatives made by carbon–carbon bond formation using the Baylis–Hillman reaction was recently reported by Queneau and coworkers  (Fig. 13). The Morita–Baylis–Hillman reaction is a tertiary amine (or phosphine)-promoted reaction of olefins activated by an electron-withdrawing group with an electrophilic double bond such as an aldehyde . This reaction is a very interesting one for several reasons: i) it is an addition reaction, therefore exhibiting favorable atom-economy, ii) it involves activated olefins such as acrylic derivatives which are also potentially biobased, iii) the reaction is known to be possibly performed in water or aqueous media, iv) the products retain the activated unsaturation which provide useful reactivity for further uses. All ingredients were thus there for having an interesting new approach towards original new biobased acrylic systems. The reaction was found to proceed in pure water, though in a moderate isolated yield (33 %, 47 % based on starting material recovery). Several classical features of the Baylis–Hillman reaction were confirmed in this case, that is the slow reaction rates which cannot be improved by increasing the reaction temperature, the fact that it is an equilibrated process, and that, in aqueous medium, an equimolar amount of the promoter DABCO was necessary to get reasonable yields . Surprisingly, phosphines, even water soluble ones, did not lead to satisfactory results. Though the aldehyde is chiral, no diasteroselectivity was observed for the reaction, all products being 1:1 mixtures of the two possible diatereoisomers on the newly formed CHOH linkage (formerly the aldehydic carbon atom). The best yields (56 %) were obtained when a 1:1 (v/v) ratio of dimethylisosorbide (DMI) and water was used. DMI has been reported to find applications in pharmaceutical and cosmetic industries, and as a substitute for DMSO and DMF in some reactions [7, 70]. Here, DMI was found to efficiently substitute dioxane or THF which are often used for the MBH reaction. MBH adducts were obtained with a variety of acrylates showing that the strategy could also be used for the design of very diverse derivatives including new amphiphilic derivatives (Fig. 13). The reaction was also applied to other GMF analogues having the β-gluco, β-xylo and β-galacto configuration leading to the corresponding Baylis–Hillman adducts in similar yields as observed for GMF.
The tetraacetate Ac4-GMF was also used as starting material in MBH reactions, and adducts were obtained in fair yields (Fig. 14). The reaction proceeded with methyl acrylate and acrylonitrile in 52 and 79 % yield respectively using DABCO as the promoter, whereas with the more reactive methyl vinyl ketone (MVK), the Baylis–Hillman adduct was obtained when using the less reactive DMAP. The reaction of GMF with MVK could be perfomed in DMI leading to the corresponding adduct in 62 % yield.
GMF, an easily available scaffold which combines the polarity of an unprotected sugar and the rich chemical potential of the furfural system, can lead by short sequences to very diverse and chemically elaborated products. Prominent examples of reported GMF-derived compounds include surfactants, liquid crystalline derivatives, monomers and various elaborated heterocyclic compounds. GMF is likely to stand as a promising substrate for the design of new valuable chemicals, and will benefit from the wide range of recent studies dealing with the synthesis and the uses of HMF for diversifying its applications in green and bioinspired chemistry.
Authors thank the Ministère de l’Enseignement Supérieur et de la Recherche (MESR) and CNRS for the financial support, as well as the China Scholarship Council (UT-INSA-CSC call) for a fellowship to JNT.
A. P. Rauter, P. Vogel, Y. Queneau. Carbohydrates in Sustainable Development I, Topics in Current Chemistry, vol. 294, Springer, Heidelberg (2010).Google Scholar
A. P. Rauter, P. Vogel, Y. Queneau. Carbohydrates in Sustainable Development II, Topics in Current Chemistry, vol. 295, Springer, Heidelberg (2011).Google Scholar
F. W. Lichtenthaler. Carbohydrate as Organic Raw Materials, VCH, Weinheim (1991).Google Scholar
G. Descotes. Carbohydrate as Organic Raw Materials, vol 2, VCH, Weinheim (1993).Google Scholar
H. van Bekkum, H. Röper, A. G. J. Voragen. Carbohydrate as Organic Raw Materials, vol. 3, VCH, Weinheim (1996).Google Scholar
W. Praznik, H. Huber. Carbohydrate as Organic Raw Materials, vol. 4, WUV-Univ, Vienna (1998).Google Scholar
P. Gallezot. Chem. Soc. Rev.41, 1538 (2012).Google Scholar
A. Corma, S. Iborra, A. Velty. Chem. Rev.107, 2411 (2007).Google Scholar
W. von Rybinski, K. Hill. Angew. Chem. Int. Ed. Engl.37, 1328 (1998).Google Scholar
F. Rodrigues, Y. Canac, A Lubineau. Chem. Commun. 2049 (2000).Google Scholar
T. Werpy, G. Petersen. NREL/TP-510-35523; National Renewable Energy Laboratory: Golden, CO (2004).Google Scholar
R. Xu, Y. Queneau. Carbohydr. Chem.40, 31 (2014).Google Scholar
R. J. van Putten, J. C. van der Waal, E. de Jong, C. B. Rasrendra, H. J. Heeres, J. G. de Vries. Chem. Rev.113, 1499 (2013).Google Scholar
B. Saha, M. M. Abu-Omar. Green Chem.16, 24 (2014).Google Scholar
T. H. Won, L. Liao, S. S. Kang, J. Shin. Food Chem.142, 107 (2014).Google Scholar
F. W. Lichtenthaler, D. Martin, T. Weber, H. Schiweck. D. E. Patent 3936522, filed 2 Nov 1989, issued 8 May 1991.Google Scholar
C. Russ, C. Luff, A. H. Begli, B. Koenig. Synth. Commun.42, 3112 (2012).Google Scholar
R. Weidenhagen, S. Lorenz. Angew. Chem.69, 641 (1957).Google Scholar
H. Schiweck, M. Munir, K. M. Rapp, B. Schneider, P. Vogel, in Carbohydrates as Organic Raw Materials, F. W. Lichtenthaler (Ed.), pp. 57–94,VCH, Weinheim (1991).Google Scholar
B. A. R. Lina, D. Jonker, G. Kozianowski. Food. Chem. Toxicol.40, 1375 (2002).Google Scholar
D. Schwengers, H. Benecke. D. E. Patent 3446380, filed 19 Dec 1984, issued 22 May 1986.Google Scholar
D. Schwengers, in: Carbohydrate as Organic Raw Materials, F. W. Lichtenthaler (Ed.), pp. 183–195, VCH, Weinheim (1991).Google Scholar
M. Munir, B. Schneider, H. Schiweck. Carbohydr. Res.164, 477 (1987).Google Scholar
F. W. Lichtenthaler, S. Peters. C. R. Chimie7, 65 (2004).Google Scholar
F. W Lichtenthaler. Carbohydr. Res.313, 69 (1998).Google Scholar
L. Guderjahn, M. Kunz, M. Schuettenhelm. Tenside Surfactant Det.31, 146 (1994).Google Scholar
M Kunz, in Carbohydrate as Organic Raw Materials, F. W. Lichtenthaler (Ed.), pp. 127–153, VCH, Weinheim (1991).Google Scholar
F. W. Lichtenthaler. Acc. Chem. Res.35, 728 (2002).Google Scholar
F. W. Lichtenthaler, R. G. Klimesch. U.S. Patent 4618675, filed 16 Dec 1983, issued 21 October 1986.Google Scholar
Y. Queneau, J. Fitremann, S. Trombotto. C. R. Chimie, 7, 177 (2004).Google Scholar
Y. Queneau, S. Jarosz, B. Lewandowski, J. Fitremann. Adv. Carbohydr. Chem. Biochem.61, 217 (2008).Google Scholar
Y. Queneau, S. Chambert, C. Besset, R. Cheaib. Carbohydr. Res.343, 1999 (2008).Google Scholar
N. Villandier, I. Adam, F. Jérôme, J. Barrault, R. Pierre, A. Bouchu, J. Fitremann, Y. Queneau. J. Mol. Catal A: Chem.259, 67, (2006).Google Scholar
N. M. Xavier, A. P. Rauter, Y. Queneau. in Carbohydrates in Sustainable Development II, Topics in Current Chemistry, vol. 295, A. P. Rauter, P. Vogel, Y. Queneau (Eds.), pp. 19–62, Springer, Heidelberg (2011).Google Scholar
S. Trombotto, E. Violet-Courtens, L. Cottier, Y. Queneau. Top. Catal.27, 31 (2004).Google Scholar
R. Pierre, S. Chambert, F. Alirachedi, M. Danel, S. Trombotto, A. Doutheau. Y. Queneau. C. R. Chimie11, 61 (2008).Google Scholar
C. Barsu, R. Cheaib, S. Chambert, Y. Queneau, O. Maury, D. Cottet, H. Wege, J. Douady, Y. Bretonniere, C. Andraud. Org. Biomol. Chem.8, 142 (2010).Google Scholar
J. Chen, Y. Miao, S. Chambert, J. Bernard, E. Fleury, Y. Queneau. Sci. China Chem.53, 1880 (2010).Google Scholar
O. Abdelkader, S. Moebs-Sanchez, Y. Queneau, J. Bernard, E. Fleury. J. Polym. Sci., Part A: Polym. Chem.49, 1309 (2011).Google Scholar
L. Cottier, G. Descotes, Y. Soro. J. Carbohydr. Chem.24, 55 (2005).Google Scholar
J. N. Tan, M. Ahmar, Y. Queneau. RSC Adv.3, 17649 (2013).Google Scholar
T. Hanemann, E. Schumacher, W. Haase, F. W. Lichtenthaler. Liq. Cryst.22, 47 (1997).Google Scholar
A. Brust, F. W. Lichtenthaler. Green Chem.15, 1368 (2013).Google Scholar
L. Lombardo, Org. Synth.65, 81 (1987).Google Scholar
D. Bassavaiah, G. Veeraraghavaiah. Chem. Soc. Rev.41, 68 (2012).Google Scholar
M. Durand, V. Molinier, T. Féron, J. M. Aubry. Prog. Org. Coat.69, 344 (2010).Google Scholar