Catalytic Oxidation of Lignins Into the Aromatic Aldehydes: General Process Trends and Development Prospects

International Journal of Molecular Sciences

Review Catalytic Oxidation of Lignins into the Aromatic Aldehydes: General Process Trends and Development Prospects

Valery E. Tarabanko * and Nikolay Tarabanko

Institute of Chemistry and Chemical Technology SB RAS, Federal Research Center “Krasnoyarsk Science Center SB RAS”, Akademgorodok 50/24, Krasnoyarsk 660036, Russia; [email protected] * Correspondence: [email protected]; Tel.: +7-391-205-1936

Received: 15 October 2017; Accepted: 12 November 2017; Published: 15 November 2017

Abstract: This review discusses principal patterns that govern the processes of lignins’ catalytic oxidation into vanillin (3-methoxy-4-hydroxybenzaldehyde) and syringaldehyde (3,5-dimethoxy-4-hydroxybenzaldehyde). It examines the influence of lignin and oxidant nature, temperature, mass transfer, and of other factors on the yield of the aldehydes and the process selectivity. The review reveals that properly organized processes of catalytic oxidation of various lignins are only insignificantly (10–15%) inferior to oxidation by nitrobenzene in terms of yield and selectivity in vanillin and syringaldehyde. Very high consumption of oxygen (and consequentially, of alkali) in the process—over 10 mol per mol of obtained vanillin—is highlighted as an unresolved and unexplored problem: scientific literature reveals almost no studies devoted to the possibilities of decreasing the consumption of oxygen and alkali. Different hypotheses about the mechanism of lignin oxidation into the aromatic aldehydes are discussed, and the mechanism comprising the steps of single-electron oxidation of phenolate anions, and ending with retroaldol reaction of a substituted coniferyl aldehyde was pointed out as the most convincing one. The possibility and development prospects of single-stage oxidative processing of wood into the aromatic aldehydes and cellulose are analyzed.

Keywords: lignin; lignosulfonates; wood; vanillin; syringaldehyde; retroaldol reaction; mechanism; technology

1. Introduction Traditionally utilized industrial technologies of wood chemical processing provide a limited assortment of primary products—cellulose, other carbohydrates—and substances derived from them. The subject of processing lignin into valuable chemicals has been attracting research for over 100 years, and presently remains actively discussed [1–7]. Among promising approaches to disposal of lignins, their processing into the aromatic aldehydes—vanillin (4-hydroxy-3-methoxybenzaldehyde) and syringaldehyde (4-hydroxy-3,5- dimethoxybenzaldehyde)—can be highlighted [1–10]. These substances are important in pharmaceutical, food, and fragrance industries [11,12]. Vanillin is used for the production of papaverine, ftivazide, and L-DOPA [13]. Active research is devoted to the vanilloid class of drugs, e.g., capsaicin—a component of cayenne pepper and the basis of medical topical formulations [13–15]. Syringaldehyde can be used for the synthesis of trimethoxybenzaldehyde, trimethoprim and other pharmaceuticals [16]. Syringaldehyde can also be transformed into substituted anthraquinones—the catalysts for improving the processes of alkaline deligniﬁcation [17]. Para-hydroxybenzaldehyde ﬁnds use in polymer synthesis and other branches of the chemical industry. Examples of polymer synthesis from lignin-derived monomers are polyferulic acid, polydihydroferulic acid, and copolymers of ferulic

Int. J. Mol. Sci. 2017, 18, 2421; doi:10.3390/ijms18112421 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2017, 18, 2421 2 of 29 or coumaryl acid with ethylene glycol—more environmentally benign alternatives to poly(ethylene terephthalate)Int. J. Mol. Sci. [18 2017]., 18, 2421 2 of 29 Vanillin production by oxidation of lignosulfonates (the byproduct of sulfite wood pulping) dominatedpolydihydroferulic the world market acid, and from copolymers the 1950’s of to ferulic the 1970’s, or coumaryl and is stillacid inwith use ethylene in Scandinavia. glycol—more Presently, vanillinenvironmentally is primarily producedbenign alternatives from guaiacol to poly(ethylene and glyoxylic terephthalate) acid, attaining [18]. around 15,000 tonnes per Vanillin production by oxidation of lignosulfonates (the byproduct of sulfite wood pulping) year at the price 6–15 USD/kg. Developing more efficient methods of vanillin production from lignins dominated the world market from the 1950’s to the 1970’s, and is still in use in Scandinavia. can lead to the price decrease and emergence of novel large-scale technologies of its consumption, Presently, vanillin is primarily produced from guaiacol and glyoxylic acid, attaining around 15,000 like productiontonnes per year of new at the polymers price 6–15 [19 USD/kg.–21]. Developing more efficient methods of vanillin production Afrom considerable lignins can lead amount to the ofprice experimental decrease and data emergence in the of fieldnovel oflarge-scale lignins technologies processing of has its been accumulatedconsumption, and like there production is a number of new of polymers modern [19–21]. reviews, this also includes the subject of catalytic oxidationA [ 2considerable–7]. However, amount the lately of experimental published reviewsdata in the barely field look of lignins into the processing principal has patterns been that governaccumulated the processes and inthere question: is a number how theof modern aldehyde review yields, andthis selectivityalso includes are the influenced subject of bycatalytic the nature of ligninoxidation and oxidant, [2–7]. However, temperature, the lately mass published transfer, reviews etc. The barely prospects look forinto developing the principal the patterns industry that of the aromaticgovern aldehydes the processes production in question: by oxidation how the of aldehyde lignins areyield not and discussed selectivity either. are influenced by the nature of lignin and oxidant, temperature, mass transfer, etc. The prospects for developing the The purpose of this review is the analysis of the modern literature data on catalytic oxidation industry of the aromatic aldehydes production by oxidation of lignins are not discussed either. of lignins by oxygen into vanillin and syringaldehyde, from basic patterns and mechanisms to the The purpose of this review is the analysis of the modern literature data on catalytic oxidation of prospectslignins of by industrial oxygen into implementation. vanillin and syringaldehyde, from basic patterns and mechanisms to the prospects of industrial implementation. 2. Lignin Structure Lignin2. Lignin is Structure one of the principal components of plant biomass. Its content in wood attains 30 wt. %. ConsiderableLignin amounts is one of of the technical principal ligninscomponents (Kraft of lignin,plant biomass. lignosulfonates, Its content hydrolysisin wood attains lignin, 30 wt. etc.; %. more aboutConsiderable their differences amounts from of nativetechnical lignin lignins later (Kraft in thislignin, Section) lignosulfonates, are generated hydrolysis after lignin, chemical etc.; processing more of woodabout by their the pulping,differences paper, from and native hydrolysis lignin later industries. in this Section) are generated after chemical Theprocessing plant of cell wood wall by the is pulping, an intricate paper, biochemicaland hydrolysiscomplex industries.formed primarily by cellulose, The plant cell wall is an intricate biochemical complex formed primarily by cellulose, hemicelluloses, and lignin [1,22]. The structure of lignin (Scheme1) is comprised of phenylpropane hemicelluloses, and lignin [1,22]. The structure of lignin (Scheme 1) is comprised of phenylpropane units (PPU, (I)). units (PPU, (I)).

C C R1 5 6 C C 1 C HO4 C C C 3 2 C R2 (I) CH O CH O 3 R 3 R 5 5 6 O O 4 1 CH O HO C C C 3 C C C 4 OH 3 2 CH3O (II) (III)

SchemeScheme 1. Lignin 1. Lignin structural structural unit, unit, and and thethe principalprincipal types types of of linkage linkage between between the theunits. units.

SoftwoodSoftwood lignin lignin has has simple simple composition composition mostly mostly consisting consisting of guaiacylpropane of guaiacylpropane structural structuralunits (R1 = H, R2 = OMe). Hardwood lignin contains, in addition to those, syringylpropane units (R1 = R2 = units (R1 = H, R2 = OMe). Hardwood lignin contains, in addition to those, syringylpropane units OMe). Besides that, grass lignin also includes p-hydroxyphenylpropane units (R1 = R2 = H). Softwood (R1 = R2 = OMe). Besides that, grass lignin also includes p-hydroxyphenylpropane units (R1 = R2 = H). and hardwood contain 28–30% and 18–24% of lignin, respectively [1,22]. Grassy plants contain Softwood and hardwood contain 28–30% and 18–24% of lignin, respectively [1,22]. Grassy plants relatively little lignin (5–15%), and its oxidation leads to the most complex aldehyde containmixture—vanillin, relatively little syringaldehyde, lignin (5–15%), and andp-hydroxybenzaldehyde. its oxidation leads For to these the mostreasons, complex oxidation aldehyde of mixture—vanillin,grass lignins is not syringaldehyde, discussed in detail and in pthis-hydroxybenzaldehyde. review. For these reasons, oxidation of grass ligninsThe isprincipal not discussed functional in groups detail in lignin this review. are methoxyl substituents at the benzene ring, alcoholic Theand principalphenolic hydroxyls, functional carbonyl groups (both in lignin aldehyde are methoxyl and ketone) substituents and carboxylic at the groups. benzene The ring,content alcoholic of and phenolic hydroxyls, carbonyl (both aldehyde and ketone) and carboxylic groups. The content of the

Int. J. Mol. Sci. 2017, 18, 2421 3 of 29 methoxyl groups is 15–16 wt. % in softwood lignins and 17–21% in hardwood lignins. The elemental composition of lignin is 61–64 wt. % carbon, 5–6% hydrogen, balanced with oxygen [22]. Among the numerous types of linkage between PPUs, it is worth pointing out the β–O–4 bonds (II) as the type that may be dominant in native lignins. Additionally, the 5–O–4 (III), 5–5 and 5–β bonds are created while pulping wood, processing and isolating the lignins of guaiacyl structure. These three bond types are the result of condensation reactions involving the benzene ring 5th position [1,22]. As the first approximation, the described bond types may suffice to explain the general patterns of lignin transformation in oxidation processes. Hardwood lignins have abundant syringyl PPUs that already have their aromatic fifth position substituted, and are unable to undergo condensation involving this position like guaiacyl PPUs in softwood lignins do. However, syringyl PPUs still can undergo condensation with formation of syringaresinol structures [22]. The structural units with free (non-etheric) phenolic functional groups have high reactivity, particularly in terms of oxidation in alkaline media. The carbohydrate components of wood, especially cellulose, are considerably more stable in these conditions. Oxidation of lignins in alkaline media leads first of all to destruction of the propane chains, while the aromatic rings retain their aromatic structure. Oxidation of lignins in acidic aqueous solutions leads primarily to the products of the benzene ring destruction [1].

3. Oxidation of Lignins into the Aromatic Aldehydes by Nitrobenzene Oxidation of lignins by nitrobenzene—or nitrobenzene oxidation (NBO)—is a long-known [23–25] and highly selective process; its results qualitatively—and even quantitatively—were used to devise the structure of lignins [1,8,22–24]. The use of nitrobenzene for oxidizing isoeugenol into vanillin was patented in 1927 [25]. NBO of lignins was actively developed in the works of Hibbert, Leopold, and other authors [23–37]; new inquiries into the process are still being made [27,28] (Table1). The process takes place in alkaline media at 160–180 ◦C lasting 2–4 h [23]. NBO causes oxidative lignin destruction with the α–β bond cleavage, leading to formation of the aromatic aldehydes as the principal products, together with their corresponding carboxylic acids, and a small portion of the aceto-derivatives of the main products—acetovanillone (4-hydroxy-3-methoxyacetophenone) and acetosyringone (4-hydroxy-3,5-dimethoxyacetophenone). The yields of these products characterize the quantities (and their ratios) of non-substituted guaiacyl and syringyl units in the macromolecules of the initial lignin.

Table 1. Nitrobenzene oxidation of various lignins into the aromatic aldehydes.

Yield, wt. % of the Lignin Lignin Type Reference No. Vanillin Syringaldehyde Silver fir wood (Abies alba) 24.3 - [29] Norway spruce wood (Picea abies) 27.5 0.06 [23] White spruce wood (Picea glauca Voss) 20.1 - [22,30] Quaking aspen wood (Populus tremuloides Michx) 12. 30. [30] Quaking aspen wood (Populus tremuloides Michx) 11 * 53 * [31] Birch wood (Betula pendula Roth.) 12 35 [29] Norway maple wood (Acer platanoides) 13 37 [32] Milled wood lignin of Loblolly pine (Pinus taeda L.) 26.6 - [33] Organosolv ethanol lignin ** 6.7 17.0 [32] Kraft softwood lignin 13.1 0.6 [32] Kraft hardwood lignin 5.3 7.9 [32] Sulfite softwood lignin 16.5 trace [32] Sulfite hardwood lignin 6.1 10.1 [32] Brown rotted spruce wood 13.9 - [34] Pine soda lignin 5.2 - [10] Japanese cedar soda-anthraquinone lignin (Cryptomeria japonica) 8–11.2 - [35] Indulin AT commercial pine kraft lignin 10.9 - [35] Pine prehydrolysis liquor precipitate 8.00 *** - [36] Klason hydrolysis lignin 1.5 - [24] * With the addition of o-phenanthroline; ** wood was treated by 0.012 M HCl in 50% aqueous ethanol; *** yield is relative to the mass of the whole substrate. Int. J. Mol. Sci. 2017, 18, 2421 4 of 29 Int.Int. J.J. Mol.Mol. Sci.Sci. 20172017,, 1818,, 24212421 44 ofof 2929

The aldehyde yield depends on the lignin source, the conditions of its isolation or preprocessing, TheThe aldehydealdehyde yieldyield dependsdepends onon thethe ligninlignin source,source, thethe conditionsconditions ofof itsits isolationisolation oror andpreprocessing,preprocessing, the conditions andand of the ligninthe conditionsconditions oxidation. ofof ligninlignin The oxidation. highestoxidation. aromatic TheThe highesthighest aldehyde aromaticaromatic yields aldehydealdehyde are attained yieldsyields areare in theattainedattained nitrobenzene inin thethe nitrobenzenenitrobenzene oxidation ofoxidationoxidation native ofof (in nativenative the original(in(in thethe originaloriginal timber) timber)timber) hardwood hardwoodhardwood lignins, lignins,lignins, up toupup toto 40–5040–5040–50 wt. %;wt.wt. even%;%; eveneven higher higherhigher figures figuresfigures (63–64 (63–64(63–64 wt. wt. %)wt. were%)%) werewere obtained obtainedobtained by usingbyby usinusin thegg catalyticthethe catalyticcatalytic systems systemssystems nitrobenzene-phenanthrolinenitrobenzene-phenanthrolinenitrobenzene-phenanthroline and nitrobenzene-anthraquinoneandand nitrobenzene-anthraquinitrobenzene-anthraqui [31nonenone,38]. Vanillin[31,38].[31,38]. yields VaniVani fromllinllin softwoodyieldsyields fromfrom ligninssoftwoodsoftwood do not lignins exceedlignins do 25–30do notnot wt. exceedexceed %. 25–3025–30 wt.wt. %.%. InsignificantlyInsignificantlyInsignificantly altered alteredaltered lignin, lignin,lignin, e.g., e.g.,e.g., the ligninthethe ligninlignin of milled ofof milledmilled pine pinepine wood, wood,wood, yields yieldsyields practically practicallypractically the samethethe samesame vanillinvanillinvanillin amount amountamount as the asas native thethe nativenative one (23.4oneone (23.4(23.4 wt. % wt.wt. and %% and 24.8and 24.8 wt.24.8 % wt.wt. respectively) %% respectively)respectively) [33,35 [33,35].[33,35].]. A small AA smallsmall vanillin vanillinvanillin yieldyieldyield decrease decreasedecrease (from (from(from 27 to 23–242727 toto 23–24 wt.23–24 %) wt.wt. [34 ]%)%) or [34] even[34] oror no eveneven decrease nono decreasedecrease at all [33 atat] are allall caused [33][33] areare by causedcaused switching byby switchingswitching from nativefromfrom softwood nativenative ligninsoftwoodsoftwood to its ligninlignin brown toto rotted itsits brownbrown form rottedrotted or enzymatic formform oror lignin. enzymaticenzymatic A more lignin.lignin. considerable AA moremore decreaseconsiderableconsiderable in thedecreasedecrease yield of inin the thethe aldehydes yieldyield ofof the isthe observed aldehydesaldehydes between isis observedobserved aspen betweenbetween wood and aspenaspen its enzymatic woodwood andand lignin itsits enzymaticenzymatic (from 44 toligninlignin 34 wt.(from(from %) [ 443944 ]. toto 3434 wt.wt. %)%) [39].[39]. OxidationOxidationOxidation of isolated ofof isolatedisolated technical technicaltechnical lignins ligninslignins yields yieldsyields 5–16 5–165–16 wt. % wt.wt. of %% the ofof aromatic thethe aromaticaromatic aldehydes. aldehydes.aldehydes. Among AmongAmong the thethe technicaltechnicaltechnical lignins ligninslignins outlined outlinedoutlined in Table inin 1 TableTable, the aromatic 1,1, thethe aromaticaromatic aldehydes aldehydesaldehydes yield decreases yieldyield decreasesdecreases along the along listalong of thethethe woodlistlist ofof thethe processingwoodwood processingprocessing techniques: techniques:techniques: sulfite > Kraft susulfitelfite > soda-anthraquinone >> KraftKraft >> soda-anthraquinonesoda-anthraquinone > soda > hydrolysis. >> sodasoda >> hydrolysis.hydrolysis. OneOne reasonOne reasonreason behind behindbehind these thesethese yield yield patternsyield patternspatterns is that isis the thatthat syringyl thethe syringylsyringyl structural structuralstructural units are unitunit moress areare stable moremore than stablestable the thanthan guaiacylthethe guaiacylguaiacyl ones because onesones because thebecause former thethe cannot formerformer undergo cannotcannot undeunde condensationrgorgo condensationcondensation (1) involving (1)(1) invoinvo thelvinglving unsubstituted thethe unsubstitutedunsubstituted fifth carbon atom in the benzene ring [40,41]: here R = CH O or H. The competition between Reactions (1) fifthfifth carboncarbon atomatom inin thethe benzenebenzene ringring [40,41]:[40,41]: 3 herehere RR == CHCH33OO oror H.H. TheThe competitioncompetition betweenbetween andReactions (2)Reactions explains (1)(1) the andand differences (2)(2) explainsexplains in the thethe aldehyde differencesdifferences yields inin between thethe aldehydealdehyde hardwood yieldsyields and betweenbetween softwood hardwoodhardwood oxidation, andand as wellsoftwoodsoftwood as (to some oxidation,oxidation, extent) as theas wewe decreasellll asas (to(to of somesome the aldehyde extent)extent) thethe yield decreasedecrease among of technicalof thethe aldehydealdehyde lignins [yieldyield40–42 amongamong]. The lignin technicaltechnical condensationligninslignins [40–42].[40–42]. extent isTheThe increasingly ligninlignin condensationcondensation affected along extentextent the listisis increasinglyincreasingly of technical ligninsaffectedaffected mentioned alongalong thethe above listlist ofof by technicaltechnical the harshligninslignins conditions mentionedmentioned of lignin aboveabove treatment byby thethe in harsh strongharsh conditionsconditions alkaline or acidofof ligninlignin media treatmenttreatment together withinin strongstrong high temperature;alkalinealkaline oror acidacid therefore,mediamedia the togethertogether vanillin withwith yield highhigh decreases temperaturtemperatur alonge;e; thistherefore,therefore, list. Higher thethe vanillinvanillin temperature yieldyield decreasesdecreases of Kraft cooking alongalong thisthis compared list.list. HigherHigher ◦ to thetemperaturetemperature sulfite process ofof KraftKraft (180 cookingcooking and 130 comparedcomparedC, correspondingly toto thethe sulfitesulfite processprocess [22]) may (180(180 cause andand 130130 deeper °C,°C, correspondinglycorrespondingly condensation of [22])[22]) Kraftmaymay lignin. causecause deeperdeeper condensationcondensation ofof KraftKraft lignin.lignin.

CHCH33OO CHCH33OO

'' "" '' HOHO RR ++ RR CH(OH)ArCH(OH)Ar HOHO RR (1)(1)(1)

" RR"CHArCHAr

CH O CH33O CHCH33OO

'' OO22 HOHO RR HO CHO (2) NaOHNaOH HO CHO (2)(2)

R R RR As aAsAs ﬁnal aa remarkfinalfinal remarkremark for this forfor Section, thisthis Section,Section, it should itit beshouldshould emphasized bebe emphasizedemphasized that the quantitative thatthat thethe quantitativequantitative results of ligninresultsresults ofof oxidationligninlignin byoxidationoxidation nitrobenzene byby nitrobenzenenitrobenzene can be used cancan as thebebe referenceusedused asas thethe value referencereference of the highest valuevalue ofof practically thethe highesthighest attainable practicallypractically vanillinattainableattainable and syringaldehyde vanillinvanillin andand syringaldehydesyringaldehyde yield in the catalytic yieldyield inin oxidation thethe catalyticcatalytic of the oxidationoxidation corresponding ofof thethe correspondingcorresponding ligneous feedstock ligneousligneous by oxygen.feedstockfeedstock byby oxygen.oxygen.

4. Catalytic4.4. CatalyticCatalytic Oxidation OxidationOxidation of Lignins ofof LigninsLignins by Oxygen byby OxygenOxygen In theInIn ﬁrst thethe half firstfirst of halfhalf the 20thofof thethe century, 20th20th itcentury,century, was discovered itit waswas discovereddiscovered that the processes thatthat thethe of lignosulfonateprocessesprocesses ofof lignosulfonatelignosulfonate oxidation intooxidation vanillinoxidation by intointo oxygen vanillinvanillin are catalyzed byby oxygenoxygen by areare oxides catalyzedcatalyzed of copper, byby oxidesoxides manganese, ofof copper,copper, cobalt, manganese,manganese, and silver [ 24cobalt,cobalt,], and andand these silversilver catalysts[24],[24], exhibitandand thesethese similar catalystscatalysts effectiveness. exhibitexhibit simisimi Forlarlar example, effectiveness.effectiveness. catalyzing ForFor example, theexample, oxidation cacatalyzingtalyzing of native thethe spruce oxidationoxidation lignin ofof by nativenative the oxidessprucespruce of ligninlignin either byby copper thethe oxidesoxides or manganese ofof eithereither yields coppercopper practically oror manganesemanganese the same yieldsyields vanillin practicallypractically amounts thethe (18–19 samesame wt. vanillinvanillin %, Tableamountsamounts2). In alkaline (18–19(18–19 media, wt.wt. %,%, the TableTable hydroxides 2).2). InIn ofalkalinealkaline the aforementioned memedia,dia, thethe hydroxideshydroxides metals—Cu(II), ofof thethe Ag(I), aforementionedaforementioned Mn(IV), Co(III)—havemetals—Cu(II),metals—Cu(II), relatively Ag(I),Ag(I), low Mn(IV),Mn(IV), redox Co(III)—haveCo(III)—have potential (below relativelyrelatively 0.35 V), lowlow among redoxredox which potentialpotential copper (below(below has 0.350.35 the lowest.V),V), amongamong whichwhich coppercopper hashas thethe lowest.lowest. AsAs willwill laterlater bebe discusseddiscussed inin thethe SectionSection aboutabout thethe processprocess mechanism,mechanism, catalystscatalysts forfor obtainingobtaining thethe aldehydesaldehydes indeedindeed shouldshould havehave lowlow redoxredox potential.potential.

Int. J. Mol. Sci. 2017, 18, 2421 5 of 29

As will later be discussed in the Section about the process mechanism, catalysts for obtaining the aldehydes indeed should have low redox potential. Using catalysts increases the aromatic aldehydes yield in oxidation of lignins by oxygen by a factor of 1.5–2 [1,22,24,42–49]. Comparing the yields of the main products in the nitrobenzene oxidation and the catalytic oxygen-consuming processes reveals that typically the latter are not much worse than the NBO. Across different lignin types (Table2), the average value of the ratios between the aldehyde yield in nitrobenzene and catalytic processes conducted under similar conditions (160–180 ◦C, 80–100 g/L NaOH) is 0.87 ± 0.07 (arithmetic mean error). Therefore, properly organized catalytic processes of various lignins oxidation are at most 10–15% inferior to nitrobenzene oxidation in terms of vanillin and syringaldehyde yields.

Table 2. Comparison of the aromatic aldehydes yield in oxidation of various lignins by molecular oxygen and nitrobenzene in alkaline media.

Lignin Type Oxidant, Catalyst, Other Remarks Yield, wt. % of Lignin Reference No. Spruce wood (28% of lignin) Air, no catalyst 11.4 [46] Spruce wood Air, Cu(OH)2 10% of wood weight 18.9 [46] Spruce wood Air, MnO2 10% of wood weight 18.2 [46] Spruce wood Nitrobenzene 20–27 Table1 Pine wood Oxygen, no catalyst 12.9 [42] Pine wood Oxygen, Cu(OH)2 23.1 [42] Brown rotted Pine wood Oxygen, Cu(OH)2 19.8 [42] Aspen wood Oxygen, CuO, ﬂow reactor 36 * [50] Aspen wood Nitrobenzene 43.6 * Table1 Birch wood Oxygen, Cu(OH)2 43 * [51] Birch wood Nitrobenzene 47 * Table1 Softwood lignosulfonates Air, no catalyst 5–7 [1,45] Softwood lignosulfonates Oxygen, Cu(OH)2 12 [43] Softwood lignosulfonates Nitrobenzene 16 Table1 Softwood lignosulfonates Oxygen, Co(OH)3, Mn3O4 10–15.5 [1,24] Eucalyptus lignosulfonates Oxygen, Cu(OH)2 13.4 * [52] Kraft lignin Oxygen, no catalyst 4.5–10.8 [49,53,54] Kraft lignin Nitrobenzene 13.1 Table1 Poplar precipitated hydrolysis lignin Oxygen, Cu(OH)2, Fe(OH)3 15 * 55 * Combined yield of vanillin and syringaldehyde.

Naturally, the lignin origin affects the aldehydes yield, and the trends observed for the catalytic oxidation are similar to those for the NBO. The relatively high aldehydes yield from the hydrolysis lignin of poplar in the presence of copper- and iron-containing hydroxides should be pointed out (up to 15 wt. %) [55]. This is quite a high ﬁgure for hydrolysis lignins, and it can be explained by an insigniﬁcant extent of intermolecular condensation in the lignin dissolved under mild percolation timber hydrolysis conditions (0.07 wt. % sulfuric acid, 210–220 ◦C). The soluble hydrolysis lignin of bagasse also yields up to 15 wt. % of vanillin and syringaldehyde, together with 3–4% of their aceto-derivatives [55]. There is also an earlier report about obtaining 8 wt. % of vanillin from the “caramel” (the precipitate from prehydrolysis liquor) of pine wood processing (Table1)[ 36]. Therefore, despite the fact that practically no aldehydes can be obtained from the lignins of harsh wood hydrolysis (Klason lignin), there exist certain hydrolysis conditions that leave lignin suitable for oxidation with high yields of the aromatic aldehydes. In oxidation of the enzymatic hydrolysis lignin from steam-exploded corn cobs, at 120 ◦C and 5 bar oxygen partial pressure (20 bar total pressure) in an alkaline aqueous media, catalyzed by the perovskite-structured oxides LaFe1−xCuxO3 (x = 0, 0.1, 0.2), the highest aldehyde yields (2.4% p-hydroxybenzaldehyde, 4.6% vanillin, 12% syringaldehyde, based on the lignin mass) were obtained with the catalyst of the highest copper content (x = 0.2) [56]. These yields are 1.5–2 times greater than in the non-catalytic reaction; adding the chlorides of Fe(III) or La(III) do not lead to the aldehydes yield increase in these conditions. The data on the effectiveness of the NBO or the catalytic activity of CuCl2 in the cited process are lacking [56]. Int. J. Mol. Sci. 2017, 18, 2421 6 of 29

The maximum aldehydes yield in oxidation of the lignin from Dedine fast hydrolysis of bagasse ◦ with the Pd/γ-Al2O3 catalyst attains 15–18 wt. % at 140 C[57]. The yield difference between the catalytic and non-catalytic variants of the described process is unique, 10–20-fold. Unfortunately, this study offers no comparison of this highly efficient palladium catalyst to copper oxide; data on oxidation by nitrobenzene are also lacking [57]. The latter circumstance makes it impossible to rate the palladium catalyst versus the maximum attainable yield—the figure that tends to greatly vary depending on the conditions of the ligneous material hydrolysis, as previously established (Tables1 and2). Research into the oxidation of lignins in weakly acidic organic solvent media [58–61] is worth a special mention. The clear advantage of the processes in acidic media is the lack of alkali consumption, but in practice this may be compensated by the loss of organic solvents to oxidation. These works focus on the oxidation of Kraft and organosolv lignins in methanol and acetic acid. In such media, oxidation of undissociated phenols takes place, as opposed to more reactive phenolate ions in alkaline aqueous media. For this reason, the processes in organic solvents attain the highest efficiency at higher temperatures (170–210 ◦C and higher) than necessary for the alkaline oxidation. In organic solvents, the yield of vanillin together with methyl vanillate attains 4–7 wt. % based on the lignin, whereas the nitrobenzene process yields 10.6%. Among the studied catalysts (phosphomolybdic acid, chlorides of copper, iron, and cobalt, sulfate of copper), CuSO4 and CoCl2 are the most effective, but the product yield is at most 30% higher in comparison to the non-catalytic process [59]. With the Co–Mn–Zr–Br catalyst, a similar process proceeds starting with a lower temperature of 140 ◦C; at 190 ◦C, up to 10.6 wt. % of vanillin, syringaldehyde, and their corresponding acids was obtained from a mixed organosolv lignin of various hardwoods. This is twice as much compared to the non-catalyzed process [62]. The use of catalysts in the processes of lignin oxidation by hydrogen peroxide [63], as well as lignin oxidation by electrochemical [64] and photochemical [65] methods do not result in high vanillin yield. For the same reason, the use of ionic liquids for the aromatic aldehydes synthesis does not look promising [66]. Next, some trends observed in the catalytic processes will be discussed to gain some insight into why the high results outlined in Table2 can be attained.

5. The Effects of Diffusion in the Processes of Lignin Oxidation Even in case of solid wood, the catalytic oxidation of lignin by oxygen proceeds considerably faster than with nitrobenzene (0.5–1 and 2–4 h respectively), and is 3–4 times faster when oxidizing lignosulfonates [24,33,43]. Naturally, a question arises: what is the role of oxygen diffusion in the processes of oxidizing lignins, and how significant is this role? In the works devoted to the kinetics of powdered aspen and pine wood oxidation in a shaking batch reactor with high mass exchange intensity (shaking rate 4 s−1)[44,67,68], it was found that chemical-controlled process kinetics can be implemented at temperatures no greater than 160 ◦C[67]. At 160 ◦C the process enters the combined control mode which is evidenced by the decreasing apparent activation energy as the temperature increases. This means that the effects of limitation by diffusion can be expected to play a considerable role in other processes of lignin oxidation in the temperature range around 160 ◦C. Indeed, significant influence of diffusion on the selectivity of powdered birch wood catalytic oxidation was discovered [51]. Decreasing the wood content in the reaction mass from 100 to 50 g/L led to almost two-fold greater yield of the aldehydes—43 wt. % combined based on the lignin, 30% syringaldehyde (Figure1)[51]. In many instances of oxidizing lignosulfonates or other ligneous materials, an increase in the rate of oxygen consumption is observed when the content of the lignin-containing material decreases in the reactor, and this can be explained by diffusion limitations of the process, and by decreased reaction mass viscosity [42,43,51,69,70]. Something less straightforward but quite doubtless is the influence of mass transfer rates on the selectivity of oxidation. In the purely diffusion-controlled situation, Int. J. Mol. Sci. 2017, 18, 2421 7 of 29 there are two regions in the reaction solution—the near-surface region that borders the gas phase, and the liquid bulk that represents the majority of the solution volume. Here it should be reminded that lignins in alkaline media at 150–170 ◦C are generally quickly oxidized even without catalysts. Therefore, in the described reaction situation, the near-surface region is affected by quick oxidation to great extent—potentially to carbon dioxide and water—and this leads to the destruction of the desired aldehydes instead of their accumulation (Scheme2). In this mode, the availability of oxygen in the solution bulk or on the catalyst surface is very limited. On the other hand, under the chemical control of the process, the concentrations of reactants and the aldehydes are similar across the entire liquid phase,Int. and J. Mol. the Sci. aldehydes 2017, 18, 2421 concentration can reach its practical maximum [69]. 7 of 29

40 1

30 2

Total yield, wt.% yield, Total 10

0 5 1015202530 Reaction time, min.

FigureFigure 1. Total 1. Total yield yield of theof the aromatic aromatic aldehydes aldehydes versusversus process duration duration in inthe the catalytic catalytic oxidation oxidation of of birchInt. J. Mol.birch wood Sci. wood by2017 oxygen ,by 18 ,oxygen 2421at 170at 170◦C[ °C51 [51].]. (1) (1) 50 50 g/Lg/L loading loading of of wood wood powder powder,, (2) (2)100 100g/L g/Lloading. loading. 8 of 29

In many instances of oxidizing lignosulfonates or other ligneous materials, an increase in the rate of oxygen consumption is observed when the content of the lignin-containing material decreases in the reactor, and this can be explained by diffusion limitations of the process, and by decreased reaction mass viscosity [42,43,51,69,70]. Something less straightforward but quite doubtless is the influence of mass transfer rates on the selectivity of oxidation. In the purely diffusion-controlled situation, there are two regions in the reaction solution—the near-surface region that borders the gas phase, and the liquid bulk that represents the majority of the solution volume. Here it should be reminded that lignins in alkaline media at 150–170 °C are generally quickly oxidized even without catalysts. Therefore, in the described reaction situation, the near-surface region is affected by quick oxidation to great extent—potentially to carbon dioxide and water—and this leads to the destruction of the desired aldehydes instead of their accumulation (Scheme 2). In this mode, the availability of oxygen in the solution bulk or on the catalyst surface is very limited. On the other hand, under the chemical control of the process, the concentrations of reactants and the aldehydes are similar across the entire liquid phase, and the aldehydes concentration can reach its practical maximum [69]. So, the selectivity of the heterogeneous processes of lignins oxidation in terms of the intended (but formally intermediate) products becomes higher as the kinetic mode gets closer to pure chemical control, and decreasing the wood-to-liquid ratio in the reactor intensifies mass transfer. This trend (the reduction of oxidation rate—potentially down to a complete halt—due to increasing lignin concentration) is observed in catalytic oxidation of soda cooking liquor from straw [70], lignosulfonates [43], and brown rotted pine wood [42]. The complete stoppage of the process is probably due to the lignin condensation type of side reactions that lead to the formation of an impenetrable polymeric film on the liquid-gas boundary that denies the access of oxygen for further

reactions. SchemeScheme 2. The 2. inﬂuenceThe influence of mass of mass transfer transfer intensity intensity on theon selectivitythe selectivity ofoxidation of oxidation of ligninsof lignins by oxygen.by For presentationoxygen. For presentation simplicity, thissimplicity, schematic this onlyschematic includes only the includes factors the that factors have that the have main the effect main on the selectivity.effect Thinon the solid selectivity. arrows Thin indicate solid chemical arrows indica transformations,te chemical transformation thin dashed arrowss, thin indicatedashed arrows slow mass transfer,indicate thick slow dashed mass arrows transfer, indicate thick dashed fastmass arrows transfer. indicate fast mass transfer.

6. The Influence of Temperature on Yield of the Aromatic Aldehydes in Oxidation by Oxygen The typical temperature range in which lignin oxidation by oxygen takes place is the same as with oxidation by nitrobenzene, 160–170 °C. Literature reveals no comprehensive insight into the influence of temperature on the effectiveness of oxidizing lignins into the aromatic aldehydes, and this Section summarizes the available fragmentary information. First of all, the patent by Schoeffel [71] should be pointed out. It reports the possibility of high vanillin yields when oxidizing lignins at 200–220 °C: non-catalytic oxidation under these conditions produces twice as much aldehydes compared to the process at 160 °C. This result is supported by other literature data [68,72] (Table 3). Increasing the temperature of aspen wood oxidation from 160 to 190–200 °C leads to almost two times higher yield of vanillin and syringaldehyde (up to 31 wt. % based on lignin). This is close to the best results of catalytic aspen wood oxidation. An even greater increase of the aromatic aldehydes yield in aspen wood oxidation (10-fold, from 1.5 to 15%) is observed when raising the temperature from 110 to 160 °C [43,50].

Table 3. The influence of temperature on yield of vanillin and syringaldehyde from the oxidation of aspen wood by oxygen in alkaline media (see references [68,72]).

Yields, wt. % Based on Lignin Temperature, °C Vanillin Syringaldehyde Total Yields 170 5.5 11.5 17.0 180 7.4 15.6 23.0 190 7.8 23.4 31.2 200 7.8 19.7 27.5

Int. J. Mol. Sci. 2017, 18, 2421 8 of 29

So, the selectivity of the heterogeneous processes of lignins oxidation in terms of the intended (but formally intermediate) products becomes higher as the kinetic mode gets closer to pure chemical control, and decreasing the wood-to-liquid ratio in the reactor intensiﬁes mass transfer. This trend (the reduction of oxidation rate—potentially down to a complete halt—due to increasing lignin concentration) is observed in catalytic oxidation of soda cooking liquor from straw [70], lignosulfonates [43], and brown rotted pine wood [42]. The complete stoppage of the process is probably due to the lignin condensation type of side reactions that lead to the formation of an impenetrable polymeric ﬁlm on the liquid-gas boundary that denies the access of oxygen for further reactions.

6. The Inﬂuence of Temperature on Yield of the Aromatic Aldehydes in Oxidation by Oxygen The typical temperature range in which lignin oxidation by oxygen takes place is the same as with oxidation by nitrobenzene, 160–170 ◦C. Literature reveals no comprehensive insight into the inﬂuence of temperature on the effectiveness of oxidizing lignins into the aromatic aldehydes, and this Section summarizes the available fragmentary information. First of all, the patent by Schoeffel [71] should be pointed out. It reports the possibility of high vanillin yields when oxidizing lignins at 200–220 ◦C: non-catalytic oxidation under these conditions produces twice as much aldehydes compared to the process at 160 ◦C. This result is supported by other literature data [68,72] (Table3). Increasing the temperature of aspen wood oxidation from 160 to 190–200 ◦C leads to almost two times higher yield of vanillin and syringaldehyde (up to 31 wt. % based on lignin). This is close to the best results of catalytic aspen wood oxidation. An even greater increase of the aromatic aldehydes yield in aspen wood oxidation (10-fold, from 1.5 to 15%) is observed when raising the temperature from 110 to 160 ◦C[43,50].

Table 3. The inﬂuence of temperature on yield of vanillin and syringaldehyde from the oxidation of aspen wood by oxygen in alkaline media (see references [68,72]).

Yields, wt. % Based on Lignin Temperature, ◦C Vanillin Syringaldehyde Total Yields 170 5.5 11.5 17.0 180 7.4 15.6 23.0 190 7.8 23.4 31.2 200 7.8 19.7 27.5

Increasing the temperature of the catalytic oxidation of aspen enzymatic lignin from 130 to 170 ◦C also results in two times greater vanillin and syringaldehyde yield [39]. Similarly, our study of the temperature’s influence on the selectivity of pine enzymatic lignin oxidation showed that the vanillin yield increases monotonously in the temperature range 90–160 ◦C[73]. Increased yield of the aldehydes caused by higher temperatures can have many underlying explanations, and we will point out two of them. On the one hand, there is quite an obvious thermodynamical trend: at high temperatures, the equilibrium of depolymerization reactions favors their products; to some extent, this can explain the discussed temperature effect on aldehyde formation. On the other hand, there is a kinetic explanation that follows from the reaction mechanism discussed in Section8: the side reaction of phenoxyl radicals dimerization can have near zero activation energy, but oxidation of the same radicals into the aldehydes has a significantly higher activation barrier [74,75]. Therefore, a higher temperature has a stronger accelerating effect on the reaction that leads to the intended products, and the side reaction of radical dimerization is thereby suppressed. And speaking of lignin depolymerization at high temperatures, it is worth noting that this reaction is known to occur in alkaline delignification, i.e., when temperature and alkalinity are similar to those for oxidizing lignins into the aldehydes. Formation of free radicals with delocalization of uncoupled electrons was detected by electron spin resonance spectroscopy during alkaline delignification Int. J. Mol. Sci. 2017, 18, 2421 9 of 29 by Kleinert [76], probably phenoxyl radicals from homolytic cleavage of ether bonds between phenylpropane units. As will be discussed in Section8, phenoxyl radicals are very important for generation of the aldehydes. In this context, one coincidence should be noted: in alkaline delignification, a certain portion of lignin (approx. 30%) is dissolved quickly compared to the rest of this polymer [41]. Approximately the same amount of lignin is typically converted to the aldehydes in oxidative processes [42]. This suggests that the high-temperature homolytic cleavage of lignin molecules may play the central role in the formation of the fragments suitable for oxidation into vanillin and syringaldehyde. We should also mention two kinetic considerations that may have the opposite effect: increase of the aldehyde yield with lowering temperature. First, decreasing temperature biases the concurrence between the non-catalytic and the catalytic reaction pathways in favor of the latter because catalytic reactions generally have lower activation energy; and since the catalytic one is more selective, better aldehyde yields can be expected. Second, as explained in the previous Section, at lower temperatures the heterogeneous process is limited by strongly activated chemical reactions as opposed to easily activated diffusion, and chemical control is necessary for high yield of the aldehydes. While both these trends should increase the process selectivity in aldehydes at lower temperatures, in most works they do not manifest. Still, the oxidation of Kraft lignin [77] in the temperature range 110–154 ◦C yields the highest vanillin amount (10.8 wt. % based on lignin) at 133 ◦C. The minimum temperature of 100 ◦C at which high yields of the aldehydes and their aceto-derivatives (17–19 wt. %) were still attained was reported for oxygen-alkali bleaching of technical-grade cellulose [78].

7. Kinetic Trends of Oxidation of Lignins Understanding mechanisms and kinetics of chemical processes is crucial to the development of their technology. This Section discusses the kinetic trends of lignin oxidation, the next one focuses on the mechanisms behind these reactions, then the technological prospects of these processes will be discussed. There are works that studied the influence of oxygen pressure and hydroxide ion concentration on the rate of oxygen consumption in the oxidation of powdered aspen and fir wood under chemical control conditions in the temperature range 110–160 ◦C[44,67,68]. It was shown that at different reaction media alkalinity levels the process obeys the kinetics corresponding to different mechanisms of radical chain reactions. At low alkalinity (pH 7–9 [79]), chain initiation is caused via chain branching (hydroperoxide cleavage) and by oxidation of phenolate ions by oxygen (reaction order in OH− is Int. J. Mol. Sci. 2017, 118, 2421 10 of 29 within the range 0– 2 ). The chain branching mechanism also occurs in oxidation of non-extracted fir woodnon-extracted [68]. At higher fir alkalinitywood [68]. (pH At 10–12.5),higher alkalinity chain initiation (pH 10–12.5), proceeds chain via phenolateinitiation proceeds ion oxidation via by oxygenphenolate and the ion process oxidation is governed by oxygen by non-branchingand the process long-chain is governed reaction by non-branching kinetics (reaction long-chain order in 1 O2 nearreaction2 , see kinetics Figure (reaction2). order in O2 near ½, see Figure 2).

-5 1 O2

Ln W -6

-2 -1 -1 0 1 1 Ln PO2

FigureFigure 2. Logarithmic 2. Logarithmic plots plots of oxygen of oxygen consumption consumption rate rate versus versus oxygen oxygen pressure pressure in in oxidation oxidation of of aspen aspen wood◦ at 110 °C.− (1) [OH−] = 0.003 M, reaction order in O2 0.56 ± 0.04. (2) [OH− −] = 2 M, reaction wood at 110 C. (1) [OH ] = 0.003 M, reaction order in O2 0.56 ± 0.04. (2) [OH ] = 2 M, reaction order order in O2 0.99 ± 0.10 (see references [44,67]). in O2 0.99 ± 0.10 (see references [44,67]). Further increase of NaOH concentration to 2 mol/L leads to the transition towards short-chain reaction kinetics with chain length as short as υ = 1–1.2, and the reaction order in oxygen increases to 1 [44,67]. This outcome is fully consistent with a general pattern of radical chain reactions [69,79]: chain length decreases as chain initiation rate rises (in this case, the rate of lignin’s phenolate anions oxidation). The results of the reviewed publications [44,67,68,79] demonstrate that despite wood’s complex composition and irregular structure, basic trends of its oxidation can be analyzed and appear to concur with those in simple well-studies systems [80]. As for the connection between the discussed oxygen-related kinetic aspects of the process and the formation of vanillin, the following can be pointed out: wood oxidation enters the non-chain reaction mode due to the high rate of initiation via lignin’s phenolate ions oxidation at high alkali concentrations of 1–2 mol/L. Coincidentally, this strong alkalinity is necessary for selective oxidation of lignins into the aromatic aldehydes; the reasons for this coincidence will be analyzed in Section 8. Other interesting patterns can be derived from studying the kinetics of lignosulfonate oxidation at the high temperatures typical for the technological conditions of vanillin synthesis. Analyzing the rate of vanillin accumulation with respect to the rate of oxygen consumption [43] reveals some trends concerning the influence of the process conditions and its selectivity towards vanillin. At 160 °C in the presence of catalyst, vanillin concentration in the reaction solution reaches the maximum (corresponding to 12 wt. % yield based on lignin) and then decreases upon further oxidation (Figure 3). This yield figure is close to the yield from nitrobenzene oxidation of lignosulfonates (16%, Table 1). Without a catalyst, the peak vanillin concentration is two times lower, but the oxygen consumption rate is practically independent of the catalyst amount (Figure 3, Table 4 rows 1, 4 and 5). Similar catalyst influence on the aromatic aldehydes yield was found with different lignin-containing feedstocks [42,43,51]. At 110 °C, catalysts (oxides of copper and silver) result in five-fold higher oxygen consumption rate and 1.4-fold higher vanillin yield (Table 4 rows 10–13). At 160 °C the oxygen consumption rate increase caused by the catalyst does not manifest, probably due to limitations by diffusion and the experimental accuracy: oxygen consumption (Figure 3) relative to vanillin yield in the highlighted experiments attains 13 mol/mol, while the reaction stoichiometry requires an order of magnitude smaller oxygen amount (see Section 8). A temperature decrease from 160 to 110 °C leads to an almost two-fold decrease of the vanillin yield in the catalytic process (Table 4, rows 4 and 11), and the pertinent explanations were given in Section 6.

Int. J. Mol. Sci. 2017, 18, 2421 10 of 29

Further increase of NaOH concentration to 2 mol/L leads to the transition towards short-chain reaction kinetics with chain length as short as υ = 1–1.2, and the reaction order in oxygen increases to 1 [44,67]. This outcome is fully consistent with a general pattern of radical chain reactions [69,79]: chain length decreases as chain initiation rate rises (in this case, the rate of lignin’s phenolate anions oxidation). The results of the reviewed publications [44,67,68,79] demonstrate that despite wood’s complex composition and irregular structure, basic trends of its oxidation can be analyzed and appear to concur with those in simple well-studies systems [80]. As for the connection between the discussed oxygen-related kinetic aspects of the process and the formation of vanillin, the following can be pointed out: wood oxidation enters the non-chain reaction mode due to the high rate of initiation via lignin’s phenolate ions oxidation at high alkali concentrations of 1–2 mol/L. Coincidentally, this strong alkalinity is necessary for selective oxidation of lignins into the aromatic aldehydes; the reasons for this coincidence will be analyzed in Section8. Other interesting patterns can be derived from studying the kinetics of lignosulfonate oxidation at the high temperatures typical for the technological conditions of vanillin synthesis. Analyzing the rate of vanillin accumulation with respect to the rate of oxygen consumption [43] reveals some trends concerning the influence of the process conditions and its selectivity towards vanillin. At 160 ◦C in the presence of catalyst, vanillin concentration in the reaction solution reaches the maximum (corresponding to 12 wt. % yield based on lignin) and then decreases upon further oxidation (Figure3). This yield figure is close to the yield from nitrobenzene oxidation of lignosulfonates (16%, Table1). Without a catalyst, the peak vanillin concentration is two times lower, but the oxygen consumption rate is practically independent of the catalyst amount (Figure3, Table4 rows 1, 4 and 5). Similar catalyst influence on the aromatic aldehydes yield was found with different lignin-containing feedstocks [42,43,51]. At 110 ◦C, catalysts (oxides of copper and silver) result in five-fold higher oxygen consumption rate and 1.4-fold higher vanillin yield (Table4 rows 10–13). At 160 ◦C the oxygen consumption rate increase caused by the catalyst does not manifest, probably due to limitations by diffusion and the experimental accuracy: oxygen consumption (Figure3) relative to vanillin yield in the highlighted experiments attains 13 mol/mol, while the reaction stoichiometry requires an order of magnitude smaller oxygen amount (see Section8). A temperature decrease from 160 to 110 ◦C leads to an almost two-fold decrease of the vanillin yield in the catalytic process (Table4, rows 4 and 11), and the pertinent explanations were given in Section6. At 160 ◦C, a decrease of pH from 11 to 10 leads to almost complete suppression of vanillin formation, but does not have a significant effect on the rate of oxygen consumption (Figure3, Table4, curves and rows 7 and 8). The latter drops to zero only at veraciously lower pH 9–9.5 [ 43]. The difference between the minimum pH values at which oxidation of lignins per se takes place, and at which specifically vanillin formation occurs is the most important detail about these processes, something that was not known prior to that publication [43]. A similar result was observed in oxidation of Kraft lignin where vanillin stops forming when the reaction solution pH decreases from 14 to 10 [77]. This trend is very important for the formulation of the process mechanism as presented in the next Section because it demonstrates that vanillin formation requires higher alkalinity than it is necessary for the dissociation of lignin’s phenolic hydroxyls. In this Section’s conclusion, we should again mention very high consumption of oxygen (and consequentially, of alkali) in the oxidation of lignosulfonates, around 13 mol per mol of vanillin. The situation is similar with other lignins: around 5% of the consumed oxygen goes into vanillin formation, and the rest of it is wasted to oxidation of some 20% of reaction solution’s organic compounds into carbon dioxide and carboxylic acids that bind alkali and decrease the system pH. The primary purpose of utilizing catalysts in oxidation of lignins is increasing the yield of the aromatic aldehydes to the level observed in oxidation by nitrobenzene, and in many cases it is quite successfully fulfilled (Tables1 and2). However, there appears to be no research focused on catalysts that would considerably lower the consumption of oxygen and alkali. Int. J. Mol. Sci. 2017, 18, 2421 11 of 29

At 160 °C, a decrease of pH from 11 to 10 leads to almost complete suppression of vanillin formation, but does not have a significant effect on the rate of oxygen consumption (Figure 3, Table 4, curves and rows 7 and 8). The latter drops to zero only at veraciously lower pH 9–9.5 [43]. The difference between the minimum pH values at which oxidation of lignins per se takes place, and at which specifically vanillin formation occurs is the most important detail about these processes, something that was not known prior to that publication [43]. A similar result was observed in oxidation of Kraft lignin where vanillin stops forming when the reaction solution pH decreases from 14 to 10 [77]. This trend is very important for the formulation of the process mechanism as presented in the next Section because it demonstrates that vanillin formation requires higher alkalinity than it Int. J. Mol.is necessary Sci. 2017, 18 for, 2421 the dissociation of lignin’s phenolic hydroxyls. 11 of 29

1 3 12 1 2 3 8 2 2 4

1 Vanillin,g/L 4 3

4 Consumption, g

0 0 0 204060 Reaction time, min. 0 20406080100 Reaction time, min. (a) (b)

Figure 3. The influence of alkali concentration and pH on vanillin yield (a) and oxygen consumption Figure 3. The influence of alkali concentration and pH on vanillin yield (a) and oxygen consumption (b) in lignosulfonate oxidation at 160 °C. Curve numbers correspond to the row numbers in Table 4 (b) in lignosulfonate oxidation at 160 ◦C. Curve numbers correspond to the row numbers in Table4 (1–5: 120 g/L NaOH; 2–6: 80 g/L NaOH; 3–7: 40 g/L NaOH + 300 g/L K2CO3, pH 10.5–10.8; 4–8: 300 g/L (1–5: 120 g/L NaOH; 2–6: 80 g/L NaOH; 3–7: 40 g/L NaOH + 300 g/L K2CO3, pH 10.5–10.8; 4–8: K2CO3, pH 10.3–10.0) (see reference [43]). 300 g/L K2CO3, pH 10.3–10.0) (see reference [43]). Table 4. The influence of catalysts and pH (measured at 20 °C) on initial oxygen consumption rate; Table 4.andThe the influence parameters of catalystscorresponding and pH to (measuredpeak vanillin at 20concentrations.◦C) on initial Oxidation oxygen consumption of lignosulfonates. rate; and the parametersUnless otherwise corresponding noted, the to content peak vanillin of the dry concentrations. sulfite liquor components Oxidation ofin reaction lignosulfonates. solutions Unless180 otherwiseg/L, noted,oxygen thepartial content pressure of the 0.2 dryMPa, sulfite reaction liquor solution components volume 60 in mL reaction (see reference solutions [43]). 180 g/L, oxygen partial pressure 0.2 MPa, reaction solution volume 60 mL (see reference [43]). Peak Vanillin Initial Initial O2 Consumption Concentration Parameters No. Catalyst, g/L Base, g/L T, °C pH Rate, g/min [V] *, Time, Attained Peak Vanillin Initial O g/L min pH Initial 2 Concentration Parameters No. 1 Catalyst, g/L- Base, NaOH, g/L 120 T,160◦C >13 Consumption0.094 6.5 >60 9.7 pH 2 Cu(OH)2, 3.24 NaOH, 120 160 >13 Rate, 0.110 g/min [V] 8.7 *, Time, >60 Attained 10.4 3 Cu(OH)2, 6.51 NaOH, 120 160 >13 0.136 g/L 11.3 min 40 9.8pH 14 Cu(OH) -2, 9.75 NaOH, NaOH, 120 120 160 160 >13 >13 0.080 0.094 6.5 13.9 >60 40 9.9 9.7 25 Cu(OH)Cu(OH)2, 3.242, 16.3 NaOH, NaOH, 120 120 160 160 >13 >13 0.103 0.110 8.7 14.7 >60 40 10.9 10.4 36 Cu(OH)Cu(OH)2, 6.512, 26.0 NaOH, NaOH, 120 120 160 160 >13 >13 0.137 0.136 11.3 11.4 20 40 10.6 9.8 47 Cu(OH)Cu(OH)2, 9.752, 9.75 NaOH, NaOH, 120 80 160 160 >13 >13 0.115 0.080 13.9 10.7 20 40 11.0 9.9 58 ** Cu(OH) Cu(OH)2, 16.32, 9.75 NaOH, NaOH, 40 + 120K2CO3, 300 160 160 10.8 >13 0.056 0.103 14.7 3.3 25 40 10.5 10.9 69 ** Cu(OH) Cu(OH)2, 26.02, 9.75 NaOH, K2CO 1203, 300 160 160 10.3 >13 0.071 0.137 11.4 0.9 20 20 10.0 10.6 710 Cu(OH)2, 9.75- NaOH, NaOH, 80 120 160110 >13 >13 0.018 0.115 10.75.4 70 20 - 11.0 8 **11 Cu(OH)Cu(OH)2, 9.752, 9.75 NaOH, 40 NaOH, + K2CO 1203 , 300 160 110 >13 10.8 0.050 0.056 3.3 7.7 70 25 9.9 10.5 9 **12 Cu(OH) AgOH,2, 9.75 12.5 K2 NaOH,CO3, 300 120 160 110 >13 10.3 0.042 0.071 0.9 5.7 70 20 9.5 10.0 1013 Cu(OH)2 -, 9.75 + AgOH, 12.5 NaOH, NaOH, 120 120 110 110 >13 >13 0.092 0.018 5.4 7.0 >90 70 10.0 - 11 Cu(OH)2, 9.75 NaOH, 120 110 >13 0.050 7.7 70 9.9 12* Vanillin AgOH, 12.5concentration; ** content NaOH, 120of the dry sulfite 110 liquor >13 components 0.042 in reaction 5.7 solution 70120 g/L. 9.5 13 Cu(OH)2, 9.75 + AgOH, 12.5 NaOH, 120 110 >13 0.092 7.0 >90 10.0 *In Vanillin this Section’s concentration; conclusion, ** content we of should the dry sulfiteagain liquormention components very high in reactionconsumption solution of 120 oxygen g/L. (and consequentially, of alkali) in the oxidation of lignosulfonates, around 13 mol per mol of vanillin. The 8. Thesituation Mechanism is similar of Oxidation with other of lignins: Lignins around 5% of the consumed oxygen goes into vanillin formation, and the rest of it is wasted to oxidation of some 20% of reaction solution’s organic Therecompounds is a considerable into carbon array dioxide of literatureand carboxylic devoted acids to that oxidation bind alkali of phenols; and decrease and forthe an system introduction pH. into aThe discussion primary onpurpose the mechanism of utilizing catalysts of lignin in oxidation, oxidation of we lignins will lookis increasing into catalytic the yield oxidation of the of p-cresol—the simplest alkyl-substituted phenol—into p-hydroxybenzaldehyde. The reaction of hydroxyl radicals with cresol under radiolysis leads to phenoxyl radicals, and proceeds two orders of magnitude faster in alkaline media than in neutral [81]. Selective catalytic oxidation of p-cresol takes place only in alkaline media. To attain near 90% selectivity, superstoichiometric amounts of alkali are necessary (alkali: cresol = 2–3 mol/mol); in neutral media no oxidation occurs at all [82]. These trends of oxidation are quite general for any phenols, and have been known for a long time [73,74,83]. Various published mechanisms can differ in details, but there are universally recognized principal stages of the process: formation of phenoxyl radicals, their transformation into quinone methides, and solvolysis of the latter into alcohols or aldehydes depending on the oxidation depth. In Scheme3, Sheldon’s oxidation mechanism [83] is presented; it includes the listed intermediate species, and also—as should be pointed out in the interest of the later discussion—the CH-acid dissociation of a phenoxyl radical (V) into an anion-radical (VI). Int. J. Mol. Sci. 2017, 18, 2421 12 of 29

aromatic aldehydes to the level observed in oxidation by nitrobenzene, and in many cases it is quite successfully fulfilled (Tables 1 and 2). However, there appears to be no research focused on catalysts that would considerably lower the consumption of oxygen and alkali.

8. The Mechanism of Oxidation of Lignins There is a considerable array of literature devoted to oxidation of phenols; and for an introduction into a discussion on the mechanism of lignin oxidation, we will look into catalytic oxidation of p-cresol—the simplest alkyl-substituted phenol—into p-hydroxybenzaldehyde. The reaction of hydroxyl radicals with cresol under radiolysis leads to phenoxyl radicals, and proceeds two orders of magnitude faster in alkaline media than in neutral [81]. Selective catalytic oxidation of p-cresol takes place only in alkaline media. To attain near 90% selectivity, superstoichiometric amounts of alkali are necessary (alkali: cresol = 2–3 mol/mol); in neutral media no oxidation occurs at all [82]. These trends of oxidation are quite general for any phenols, and have been known for a long time [73,74,83]. Various published mechanisms can differ in details, but there are universally recognized principal stages of the process: formation of phenoxyl radicals, their transformation into quinone methides, and solvolysis of the latter into alcohols or aldehydes depending on the oxidation depth. In Scheme 3, Sheldon’s oxidation mechanism [83] is presented; it includes the listed intermediate Int. J. Mol. Sci.species,2017 ,and18, 2421also—as should be pointed out in the interest of the later discussion—the CH-acid 12 of 29 dissociation of a phenoxyl radical (V) into an anion-radical (VI).

O O Co(III) Co(II)

Co(III)

CH3 CH3 Co(II) (IV) (V)

O O O O Co(III) Co(II) O2/Co(II)

O Co(III) C CH O CH2 2 CH2 H H (VI)

-HOCo(III) CH3OH ROH -Co(III)OOH O O

O2/Co(II) R=H, CH3

CHO CH2OR Scheme 3. The mechanism of p-cresol oxidation according to Sheldon [83]. Curved arrows show the Scheme 3. The mechanism of p-cresol oxidation according to Sheldon [83]. Curved arrows show the transformations happening to the oxidant. transformations happening to the oxidant. The reactions of oxidation of alkyl phenols are obviously not completely selective. The principal Theside reactions reaction of pathway oxidation is the of transformation alkyl phenols of phenoxyl are obviously radicals: not their completely dimerization selective. (3) followed The by principal oligomerization [74,84]. side reaction pathway is the transformation of phenoxyl radicals: their dimerization (3) followed by oligomerization [74,84]. Int. J. Mol. Sci. 2017, 18, 2421 13 of 29

O O OH

O CH3 (3) (3)

CH3 CH3 CH3 Phenoxyl radicals and quinone methides are widely discussed when describing the Phenoxylmechanisms radicals of andlignin quinone oxidation methides in cellulose are widelybleaching discussed processes when[85], as describing well as in the aromatic mechanisms of ligninaldehydes oxidation synthesis in cellulose [2,9,86–88]. bleaching Since the processes 1980’s, several [85], hypo astheses well asconcerning in the the aromatic mechanism aldehydes of synthesis [lignins2,9,86 –oxidation88]. Since into the the aromatic 1980’s, aldehydes several were hypotheses suggested. concerning Before discussing the mechanism their credibility, of we lignins oxidationwill into summarize the aromatic the aldehydeshereinabove werediscussed suggested. and well-known Before discussing process trends their that credibility, a theorized we will mechanism of lignin oxidation into the aromatic aldehydes in alkaline media needs to be able to summarize the hereinabove discussed and well-known process trends that a theorized mechanism explain or not contradict: of lignin oxidation into the aromatic aldehydes in alkaline media needs to be able to explain or 1. Oxidation of lignins into vanillin and syringaldehyde proceeds with oxidants of a different not contradict: nature (nitrobenzene, copper oxide [30], oxygen with and without [71] catalysts) with high and 1. Oxidationsimilar of lignins selectivities into (over vanillin 40 wt. and % ofsyringaldehyde aldehydes). It appears proceeds unlikely with that the oxidants same outcome of a different can transpire through different oxidant-dependent mechanisms in such a chemically complicated nature (nitrobenzene, copper oxide [30], oxygen with and without [71] catalysts) with high and system. Therefore, a mechanism hypothesis should be universal with respect to oxidant nature. similar2. selectivitiesSelective oxidation (over 40of lignins wt. % ofrequires aldehydes). higher reaction It appears medium unlikely alkalinity that than the sameis necessary outcome for can transpirethe through dissociation different of phenolic oxidant-dependent hydroxyls [43,77]. mechanisms Thus, a vanillin in such formation a chemically mechanism complicated may system. Therefore,include acidic a mechanism dissociation hypothesisof an intermediate should bethat universal is less acidic with respectthan lignin’s to oxidant phenolic nature. 2. Selectivehydroxyls. oxidation of lignins requires higher reaction medium alkalinity than is necessary for the 3. Production of vanillin and syringaldehyde from lignins is accompanied by the formation of the dissociation of phenolic hydroxyls [43,77]. Thus, a vanillin formation mechanism may include corresponding aceto-derivatives as side products. acidic4. dissociationVanillin, syringaldehyde, of an intermediate and their that aceto-derivatives is less acidic are than produced lignin’s in phenolicsmaller amounts hydroxyls. in lignin 3. Productionalkaline of vanillin hydrolysis and without syringaldehyde oxidants. This from indicates lignins that is accompaniedlignin oxidation by and the lignin formation alkaline of the correspondinghydrolysis aceto-derivatives may have common as stages, side products. and these should probably be the final stages. The first known hypothesis about the mechanism of vanillin formation does not relate to oxidation reactions. It is known that alkaline hydrolysis of lignosulfonates produces vanillin (up to 5.9 wt. % based on lignin) without oxidants [1,24]. In 1936, Hibbert [26] suggested that this vanillin formation is caused by alkali-catalyzed retroaldol reaction of the α-hydroxy-γ-carbonyl structure of a phenylpropane lignin structural unit:

OH- Ar CH(OH) C CHO Ar CHO + CH CHO (4)

where Ar is 3-methoxy-4-phenoxy anion. This mechanism is good at describing oxidantless alkaline hydrolysis of lignin. First, it can explain the formation of acetovanillone and acetosyringone as the byproducts of lignin alkaline hydrolysis [24,26] by similar cleavage of an α-carbonyl structure:

Ar C(O) CH2CR2OH Ar C(O)CH3 + R2CO (5) Second, Reactions (4) and (5) require strong alkalinity because formation of the enolate anion necessary for aldol condensation and cleavage reactions [89] is caused by further dissociation of a phenolate anion of a lignin PPU:

CH3O OH CH3O OH O O + (6) O CH CH2 C O CH CH C +H R R

Int. J. Mol. Sci. 2017, 18, 2421 13 of 29 Int. J. Mol. Sci. 2017, 18, 2421 13 of 29

O O OH O O OH Int. J. Mol. Sci. 2017, 18, 2421 13 of 29 O CH3 (3) O CH3 (3) O O OH CH3 CH3 CH3 CH3 CH3 CH3 O CH3 (3) PhenoxylPhenoxyl radicals radicals andand quinonequinone methidesmethides aree widelywidely discusseddiscussed whenwhen describing describing the the mechanismsmechanisms of oflignin lignin oxidation oxidation inin cellulosecellulose bleaching processesprocesses [85], [85], as as well well as as in inthe the aromatic aromatic CH3 CH3 aldehydesaldehydes synthesis synthesis [2,9,8 [2,9,86–88].6–88]. Since Since thethe 1980’s,1980’s,CH several3 hypohypothesestheses concerning concerning the the mechanism mechanism of of lignins oxidation into the aromatic aldehydes were suggested. Before discussing their credibility, we ligninsPhenoxyl oxidation radicals into the andaromatic quinone aldehydes methides were suggested.are widely Before discussed discussing when their describing credibility, the we will summarize the hereinabove discussed and well-known process trends that a theorized mechanismswill summarize of lignin the hereinaboveoxidation in discussedcellulose bleaching and well-known processes process [85], as trendswell as that in thea theorizedaromatic mechanismaldehydesmechanism ofsynthesis oflignin lignin oxidation[2,9,8 oxidation6–88]. into intoSince thethe the aromaticaromatic 1980’s, several aldehydes hypo inintheses alkalinealkaline concerning media media needs theneeds mechanism to tobe beable able ofto to explainligninsexplain or oxidation ornot not contradict: contradict: into the aromatic aldehydes were suggested. Before discussing their credibility, we 1. will1. Oxidation summarizeOxidation of of ligninsthe lignins hereinabove into into vanillin vanillin discussed andand syringaldehydesyringaldehyde and well-known proceedsproceeds process with with trends oxidants oxidants that ofa ofatheorized differenta different mechanismnaturenature (nitrobenzene, of(nitrobenzene, lignin oxidation copper copper into oxide oxide the [30],aromatic[30], oxygen aldehydes withth andand in without withoutalkaline [71] media[71] catalysts) catalysts) needs withto with be high able high and to and explainsimilarsimilar or selectivitiesnot selectivities contradict: (over (over 40 40 wt. wt. % % of of aldehydes)aldehydes).. ItIt appearsappears unlikely unlikely that that the the same same outcome outcome can can 1. transpireOxidationtranspire through throughof lignins different different into vanillin oxidant-dependent oxidant-dependent and syringaldehyde memechanismschanisms proceeds in in such suchwith a achoxidants chemicallyemically of complicated a complicateddifferent system.naturesystem. Therefore, (nitrobenzene, Therefore, a amechanism mechanism copper oxide hypothesis hypothesis [30], oxygen should with bebe and universaluniversal without with with [71] respect catalysts)respect to to oxidant with oxidant high nature. nature.and 2. Selective oxidation of lignins requires higher reaction medium alkalinity than is necessary for 2. Selectivesimilar selectivitiesoxidation of (over lignins 40 wt. requires % of aldehydes) higher reaction. It appears medium unlikely alkalinity that the thansame isoutcome necessary can for Int. J. Mol. Sci. 2017the, 18 ,dissociation 2421 of phenolic hydroxyls [43,77]. Thus, a vanillin formation mechanism may13 of 29 thetranspire dissociation through of differentphenolic oxidant-dependent hydroxyls [43,77]. me Thus,chanisms a vanillin in such formationa chemically mechanism complicated may include acidic dissociation of an intermediate that is less acidic than lignin’s phenolic includesystem. acidic Therefore, dissociation a mechanism of anhypothesis intermediate should bethat universal is less with acidic respect than to oxidantlignin’s nature.phenolic hydroxyls. 2. hydroxyls.Selective oxidation of lignins requires higher reaction medium alkalinity than is necessary for 4. Vanillin,3. syringaldehyde,Production of vanillin and and their syringaldehyde aceto-derivatives from lignins are produced is accompanied in smaller by the formation amounts of in the lignin 3. Productionthe dissociation of vanillin of phenolic and syringaldehyde hydroxyls [43,77]. from ligninsThus, a is vanillin accompanied formation by themechanism formation may of the alkaline hydrolysisincludecorresponding acidic without aceto-derivativesdissociation oxidants. of anas This sideintermediate products. indicates that that is ligninless acidic oxidation than lignin’s and lignin phenolic alkaline corresponding aceto-derivatives as side products. hydrolysis4. hydroxyls.Vanillin, may have syringaldehyde, common stages, and their and aceto-derivatives these should are probably produced be in the smaller ﬁnal amounts stages. in lignin 4. Vanillin, syringaldehyde, and their aceto-derivatives are produced in smaller amounts in lignin 3. Productionalkaline hydrolysis of vanillin without and syringaldehyde oxidants. This from indicates lignins that is accompanied lignin oxidation by the and formation lignin alkaline of the alkaline hydrolysis without oxidants. This indicates that lignin oxidation and lignin alkaline The ﬁrstcorresponding knownhydrolysis hypothesis may aceto-derivatives have common about stages, the as side mechanism and products. these should of vanillin probably formationbe the final stages. does not relate to hydrolysis may have common stages, and these should probably be the final stages. oxidation reactions.4. Vanillin,The first It syringaldehyde, isknown known hypothesis that and alkaline theirabout aceto-derivatives hydrolysisthe mechanism of lignosulfonatesareof producedvanillin formation in smaller produces does amounts not vanillin inrelate lignin (upto to 5.9 wt. % basedoxidationThealkaline onfirst lignin)reactions. knownhydrolysis without Ithypothesis is withoutknown oxidants that oxidants.about alkaline [ 1the, 24This ].mechanismhydr Inindicatesolysis 1936, of Hibbert thatof lignosulfonates vanillin lignin [26 oxidation ]formation suggested produces and does lignin thatvanillin not thisalkaline relate(up vanillin to to formationoxidation is5.9 caused wt.hydrolysis %reactions. based by alkali-catalyzed may on It lignin) haveis known common without that retroaldol stages,oxidants alkaline and [1,24]. reactionhydr thesolysis eIn should 1936, of of the Hibbertprobablylignosulfonatesα-hydroxy- [26] be thesuggested γfinal produces-carbonyl stages. that vanillin this structure vanillin (up ofto a phenylpropane5.9formation wt. The% lignin based first is caused structural knownon lignin) by hypothesis alkali-catalyzed unit:without aboutoxidants retroaldol the [1,24].mechanism reaction In 1936, of of Hibbertvanillin the α-hydroxy- [26]formation suggestedγ-carbonyl does that not structure thisrelate vanillin toof formationoxidationa phenylpropane is reactions. caused lignin by It alkali-catalyzedis structuralknown that unit: alkaline retroaldol hydr olysisreaction of lignosulfonatesof the α-hydroxy- producesγ-carbonyl vanillin structure (up to of a phenylpropane5.9 wt. % based ligninon lignin) structural without unit: oxidants [1,24]. In 1936, Hibbert [26] suggested that this vanillin OH- formation isAr caused CH(OH) by alkali-catalyzed C CHO retroaldol reactionAr CHO of the + α -hydroxy- CH CHO γ -carbonyl structure(4) of (4) a phenylpropane lignin structural unit: OH- Ar CH(OH) C CHO Ar CHO + CH CHO (4) where Ar is 3-methoxy-4-phenoxy anion. This mechanism is good at describing oxidantless alkaline OH- where Ar ishydrolysis 3-methoxy-4-phenoxy Arof lignin. CH(OH) First, it anion. can C explain ThisCHO the mechanism formationAr of isCHO acetovanillone good + at CH describing and CHO acetosyringone oxidantless as(4) alkaline the where Ar is 3-methoxy-4-phenoxy anion. This mechanism is good at describing oxidantless alkaline hydrolysisbyproducts of lignin. of First, lignin it alkaline can explain hydrolysis the formation[24,26] by similar of acetovanillone cleavage of an α and-carbonyl acetosyringone structure: as the hydrolysis of lignin. First, it can explain the formation of acetovanillone and acetosyringone as the byproductswhere of lignin Ar is alkaline3-methoxy-4-phenoxy hydrolysis anion. [24,26 This] by mechanism similar cleavage is good at of describing an α-carbonyl oxidantless structure: alkaline byproducts of ligninAr C(O)alkaline CHhydrolysis2CR2OH [24,26] by similarAr C(O)CHcleavage3 of + an R α2CO-carbonyl structure:(5) hydrolysis of lignin. First, it can explain the formation of acetovanillone and acetosyringone as the Second, Reactions (4) and (5) require strong alkalinity because formation of the enolate anion byproductsAr of lignin C(O) alkaline CH hydrolysis2CR2OH [24,26] by similarAr cleavage C(O)CH of3 an + α -carbonylR2CO structure: (5) (5) necessary for aldol condensation and cleavage reactions [89] is caused by further dissociation of a phenolateSecond, anionReactionsAr of C(O)a lignin (4) and PPU:CH (5)2CR require2OH strong alArkalinity C(O)CH because3 +formation R2CO of the enolate(5) anion Second,necessary Reactions for aldol (4) condensation and (5) require and strongcleavage alkalinity reactions because[89] is caused formation by further of the dissociation enolate anionof a necessary forCH aldolSecond,3O condensation ReactionsOH (4) and and (5) cleavage require strong reactionsCH al3kalinityO [89] because is causedOH formation by further of the dissociationenolate anion of a phenolate anion of a lignin PPU: O O phenolate anionnecessary of afor lignin aldol PPU:condensation and cleavage reactions [89] is caused by further dissociation of a +H+ (6) CHphenolate3O anion ofOH aCH lignin CH PPU:2 C CHO3O CHOH CH C OR RO CH3O OH CH3O OH + (6) O CH CH2 C O O CH CH C O +H R+H+ (6) (6) O CH CH2 C R O CH CH C R R

Int. J. Mol. Sci. 2017, 18, 2421 14 of 29 pK values of model lignin compounds attain 9–10 and more [90,91]. Obviously, the dissociation constant ofp K Reaction values of (6)model is severallignin compounds orders of attain magnitude 9–10 and less more than [90,91]. the Obviously, ﬁgure for the the dissociation PPU phenolic hydroxyl.constant Therefore, of Reaction effective (6) is conduction several orders of of the magnitude Reactions less(4) than and the (5)figure is onlyfor the possible PPU phenolic in media with considerablyhydroxyl. higherTherefore, alkalinity effective conduction than what of is the necessary Reactions for(4) and acidic (5) is dissociation only possible of inlignin media with phenolic hydroxylsconsiderably (7), and for higher their fastalkalinity oxidation than whenwhat is oxidants necessary are for present. acidic dissociation of lignin phenolic hydroxyls (7), and for their fast oxidation when oxidants are present.

CH3O OH CH3O OH

O O OH + (7) (7) CH CH2 C O CH CH2 C +H R R Vanillin yield without oxidants is modest which agrees with relatively low content of carbonyl Vanillingroups yield in withoutlignins [92]. oxidants So, Hibbert’s is modest suggested which pathway agrees of withthe lignin relatively alkaline low hydrolysis content is in of good carbonyl groups inagreement lignins [92with]. experimental So, Hibbert’s facts suggested about oxidantless pathway vanillin of formation, the lignin and alkaline with general hydrolysis patterns is in good agreementof retroaldol with reaction. experimental Hibbert’s facts hypothesis about oxidantlessagrees with three vanillin out of formation, four hereinabove and with defined general patterns ofrequirements retroaldolreaction. for a mechanism Hibbert’s of hypothesislignin oxidation agrees into with the threearomatic out aldehydes. of four hereinabove It is naturally defined requirementstempting for a mechanismto extrapolateof this lignin productive oxidation hypothes intois the towards aromatic oxidative aldehydes. lignin cleavage. It is naturally This would tempting mean limiting the role of an oxidant to formation of the carbonyl group necessary for the retroaldol to extrapolate this productive hypothesis towards oxidative lignin cleavage. This would mean limiting reaction which finalizes the aldehyde production process. However, for 60 years there were no the role ofattempts an oxidant at using to formation the Hibbert’s of thehypothesis carbonyl as th groupe basis necessaryfor elucidating for thethe mechanism retroaldol of reaction oxidative which finalizes thelignin aldehyde cleavage. production process. However, for 60 years there were no attempts at using the Hibbert’s hypothesisNow we as will the briefly basis consider for elucidating hypotheses the (H1 mechanism–H3) aboutof the oxidative mechanism lignin of lignin cleavage. oxidation Nowinto we willthe aromatic briefly consider aldehydes hypotheses that were formulated (H1–H3) aboutin 1980’s the and mechanism are frequently of lignin cited oxidation in many into reviews to the present day. the aromatic aldehydes that were formulated in 1980’s and are frequently cited in many reviews to the present day.Hypothesis 1 (H1). Schultz’s hypothesis [93,94] on the mechanism of lignin oxidation by nitrobenzene via abstraction of an electron and then of a proton from the benzylic hydroxyl of a lignin structural unit, instead of an attack on the phenolic hydroxyl. β-cleavage decay of the remaining benzoxyl radical leads to vanillin:

R CH OH R CH O O CH

-e -H+ + R (8)

CH3O CH3O CH3O OH OH OH Schultz’s hypothesis does not agree with the mechanism requirements 3 and 4.

Hypothesis 2 (H2). A hypothesis about the mechanism of lignin oxidation by oxygen via formation of 1,2-dioxetane ring and its subsequent cleavage leading to vanillin and a carboxylic acid [9]:

O2 CH CR O CH CR O CH CHR O -H2O -O2 ' ' ' OR OR OH OR CH O CH3O CH3O 3 OOH - HO2 HO O CH CR O CH CR -H2O ' OR' OR (9) CH3O CH3O

O O HO- ' O CH CHR O CHO + RCOOR -H2O OH OR' CH3O CH3O This mechanism is adopted from discussions on chemistry of cellulose oxygen bleaching [95–97] in order to explain vanillin formation [9], and is cited in many reviews. Its principal

Int. J. Mol. Sci. 2017, 18, 2421 14 of 29

pK values of model lignin compounds attain 9–10 and more [90,91]. Obviously, the dissociation Int.constant J. Mol. Sci. of 2017Reaction, 18, 2421 (6) is several orders of magnitude less than the figure for the PPU phenolic14 of 29 hydroxyl. Therefore, effective conduction of the Reactions (4) and (5) is only possible in media with pconsiderablyK values of highermodel ligninalkalinity compounds than what attain is nece9–10ssary and formore acidic [90,91]. dissociation Obviously, of thelignin dissociation phenolic constanthydroxyls of (7), Reaction and for (6) their is sevefast oxidationral orders when of magnitude oxidants areless present. than the figure for the PPU phenolic hydroxyl. Therefore, effective conduction of the Reactions (4) and (5) is only possible in media with CH3O OH CH3O OH

considerably higher alkalinity than whatO is necessary for acidic dissociationO of lignin phenolic hydroxylsOH (7), and for their fast oxidation when oxidants are present. + (7) CH CH2 C O CH CH2 C +H R CH3O OH R CH3O OH

O O VanillinOH yield without oxidants is modest which agrees with relatively low content+ of carbonyl(7) CH CH2 C O CH CH2 C +H groups in lignins [92]. So, Hibbert’s suggested pathway of the lignin alkaline hydrolysisR is in good R agreement with experimental facts about oxidantless vanillin formation, and with general patterns of retroaldolVanillin yieldreaction. without Hibbert’s oxidants hypothesis is modest agrewhiches agreeswith three with outrelatively of four low hereinabove content of carbonyl defined groupsrequirements in lignins for [92].a mechanism So, Hibbert’s of ligninsuggested oxidation pathway into of thethe ligninaromatic alkaline aldehydes. hydrolysis It is isnaturally in good agreementtempting to with extrapolate experimental this productive facts about hypothes oxidantlessis towards vanillin oxidative formation, lignin and withcleavage. general This patterns would ofmean retroaldol limiting reaction. the role ofHibbert’s an oxidant hypothesis to formation agre ofes the with carbonyl three groupout of necessaryfour hereinabove for the retroaldol defined requirementsreaction which for finalizes a mechanism the aldehyde of lignin production oxidation process. into the However, aromatic for aldehydes. 60 years Itthere is naturallywere no temptingattempts atto usingextrapolate the Hibbert’s this productive hypothesis hypothes as the isbasis towards for elucidating oxidative ligninthe mechanism cleavage. ofThis oxidative would meanlignin limitingcleavage. the role of an oxidant to formation of the carbonyl group necessary for the retroaldol reactionNow which we will finalizes briefly the consider aldehyde hypotheses production (H1 process.–H3) about However, the mechanism for 60 years of lignin there oxidationwere no Int. J. Mol. Sci. attemptsinto2017 ,the18, 2421aromaticat using thealdehydes Hibbert’s that hypothesis were formulated as the basis in for1980’s elucidating and are the frequently mechanism cited of oxidativein many 14 of 29 ligninreviews cleavage. to the present day. Now we will briefly consider hypotheses (H1–H3) about the mechanism of lignin oxidation HypothesisintoHypothesis 1 (H1). the aromaticSchultz’s 1 (H1). aldehydes hypothesisSchultz’s that hypothesis [93were,94 ]formulated on [93,94] the mechanism onin 1980’sthe mechanism and of ligninare frequently of oxidation lignin citedoxidation by in nitrobenzene many by via abstractionreviewsnitrobenzene ofan to the electron via present abstraction andday. then of an of electron a proton and from then theof a benzylicproton from hydroxyl the benzylic of ahydroxyl lignin structuralof a unit, insteadlignin of an structural attack on unit, the phenolicinstead of hydroxyl.an attack onβ-cleavage the phenolic decay hydroxyl. of the β remaining-cleavage decay benzoxyl of the radical Hypothesisremaining benzoxyl 1 (H1). radical Schultz’s leads hypothesis to vanillin: [93,94] on the mechanism of lignin oxidation by leads to vanillin:nitrobenzene via abstraction of an electron and then of a proton from the benzylic hydroxyl of a CH lignin structural unit, insteadR CH of OHan attackR onCH theO phenolicO hydroxyl. β-cleavage decay of the remaining benzoxyl radical leads to-e vanillin: -H+ + R (8) (8) R CH OH R CH O O CH CH3O CH3O CH3O OH OH OH -e -H+ + R (8) Schultz’s hypothesis does not agree with the mechanism requirements 3 and 4. CH3O CH3O CH3O OH OH OH 3 4 Schultz’sHypothesis hypothesis 2 (H2). does A hypothesis not agree about with the the mechanism mechanism of lignin requirements oxidation by oxygenand .via formation of 1,2-dioxetaneSchultz’s hypothesis ring and itsdoes subsequent not agree cleavagewith the leadingmechanism to vanillin requirements and a carboxylic 3 and 4. acid [9]: Hypothesis 2 (H2). A hypothesis about the mechanism of lignin oxidation by oxygen via formation of O2 Hypothesis 2 (H2). A hypothesis about the mechanism CH CR of ligninO oxidation CHby oxygen CR via formation 1,2-dioxetane ring andO its subsequent CH CHR cleavageO leading to vanillin and a carboxylic acid [9]: -H2O -O2 of 1,2-dioxetane ring and its subsequent cleavage leading' to vanillin and a carboxylic' acid [9]: ' OR OR OH OR CH O CH3O CH3O 3 O2 CH CR O CH CR O CH CHR O OOH -H2O -O2 - HO ' HO ' ' 2 OR CR OR O OH CHOR CR O CH CH O CH3O CH3O -H2O 3 ' ' (9) OR OOHOR CH3O CH3O - HO2 HO O CH CR O CH CR -H2O O O ' OR' OR (9) (9) - HOCH3O CH3O ' O CH CHR O CHO + RCOOR -H2O OH OR' O O CH3O CH3O HO- ' O CH CHR O CHO + RCOOR This-H mechanismO is adopted from discussions on chemistry of cellulose oxygen bleaching 2 ' [95–97] in order to explainOH vanillinOR formation [9], and is cited in many reviews. Its principal CH3O CH3O This mechanism is adopted from discussions on chemistry of cellulose oxygen bleaching This mechanism[95–97] in order is adopted to explain from vanillin discussions formation on [9], chemistry and is cited of cellulose in many oxygenreviews. bleachingIts principal [95 –97] in order to explain vanillin formation [9], and is cited in many reviews. Its principal drawback is that it cannot be extrapolatedInt. J. Mol. Sci. 2017 towards, 18, 2421 two more selective oxidants: nitrobenzene or copper oxide.15 of 29 Int. J. Mol. Sci. 2017, 18, 2421 15 of 29 Hypothesis 3drawback (H3) Ais that hypothesis it cannot be about extrapolated the mechanism towards two more of lignin selective oxidation oxidants: nitrobenzene that involves or a drawbackcopper oxide. is that it cannot be extrapolated towards two more selective oxidants: nitrobenzene or retroaldol reactioncopper oxide. step. Based on various experimental results [86–88,98] and discussions in literature [85,Hypothesis95,99–101] ,3 a(H3) mechanism A hypothesis hypothesis about the for mechanism formation of of lignin vanillin oxidation in lignin that oxidationinvolves a was developedHypothesis inretroaldol 1997–2004. reaction3 (H3) It unitesAstep. hypothesis Based the universallyon about various the experimental mechanism recognized resuof lignin ltslignin [86–88,98] chemistryoxidation and thatdiscussions principles involves in about a phenoxyl radicalsretroaldolliterature and reaction[85,95,99–101], quinone step. methide aBased mechanism on intermediates various hypothesis experimental for with formation theresu conceptoflts vanillin [86–88,98] ofin retroaldollignin and oxidationdiscussions reaction was in as developed in 1997–2004. It unites the universally recognized lignin chemistry principles about the ﬁnal processliterature step. [85,95,99–101], To keep the a presentationmechanism hypothesis of this hypothesis for formation [86 of– vanillin88,98] brief,in lignin it will oxidation be rendered was developedphenoxyl radicalsin 1997–2004. and quinone It unites methide the universallintermediatesy recognized with the concept lignin ofchemistry retroaldol principles reaction as about the as oxidationphenoxyl offinal eugenol process radicals asstep. an and To example quinonekeep the modelpresentationmethide compound. intermediates of this hypothesis with the [86–88,98] concept brief,of retroaldol it will be reaction rendered as asthe finaloxidation process of step. eugenol To keep as an the example presentation model compound. of this hypothesis [86–88,98] brief, it will be rendered as The oxidationoxidation begins of eugenol with as abstractionan example model of an compound. electron from a phenolate anion: The oxidation begins with abstraction of an electron from a phenolate anion: The oxidation begins with abstraction of an electron from a phenolate anion: -e O CH2 CH CH2 O CH2 CH CH2 (10) -e (VIII) (10) O (VII) CH2 CH CH2 O CH2 CH CH2 CH3O CH3O (10) (VII) (VIII) DisproportionationCH3O of these primary radicalsCH (VIII3O ) leads to a quinone methide (IX): Disproportionation of these primary radicals (VIII) leads to a quinone methide (IX): Disproportionation of these primary radicals (VIII) leads to a quinone methide (IX): 2 O CH2 CH CH2 (VIII)

2 OCH3O CH2 CH CH2 (VIII) (11)

CH3O HO CH2 CH CH2 + O CH CH CH2 (11) (11) (IX) CH CH O HOCH3O 2 CH CH2 + O 3 CH CH CH2 (IX) and nucleophilic addition of a hydroxide ion to theCH latterO creates the coniferyl alcohol structure (X). CH3O 3 - and nucleophilic addition of a hydroxide ionOH to the latter creates the coniferyl alcohol structure (X). O CH CH CH2 O CH CH CH2OH (12) (IX) - CH O OH (X) O 3 CH CH CH2 CH O 3O CH CH CH2OH (12) It should be noted(IX) that formation of the quinone methide(X) (IX) out of the phenoxyl radical CH3O CH O (VIII) is also possible via acidic dissociation of3 the latter with the following single-electron oxidationIt should of thebe resultingnoted that radical-anion. formation ofThis the possibility quinone wasmethide discussed (IX) outby Sheldonof the phenoxyl(Scheme 3radical [83]) (VIIIand) inis our also works possible [86,87,102]. via acidic dissociation of the latter with the following single-electron oxidationSubsequent of the resulting oxidation radical-anion. of coniferyl alcohol This possibility (X) is analogous was discussed to Reactions by Sheldon(10)–(12) (Scheme and results 3 [83]) in andthe in γ our-carbonyl works group [86,87,102]. (XI). Retroaldol reaction of the α-unsaturated aldehyde (XI) yields vanillin (XIIISubsequent): oxidation of coniferyl alcohol (X) is analogous to Reactions (10)–(12) and results in the γ-carbonyl group (XI). Retroaldol reaction of the α-unsaturated aldehyde (XI) yields vanillin H O CH CH CHO 2 (XIII): O O CH CH2 CHO (13) (XI) OH (XII) CH O H O 3 CH CH CHO 2 CH3O O O CH CH2 CHO (13)

(XI) - OH (XII) OH CH3O CH3O O CH CH2 CHO O CHO + CH2 CHO -H2O (14) OH (XII) (XIII) - CH3O CH O OH 3 O CH CH2 CHO O CHO + CH2 CHO -H2O Step (14) proceeds through C-H acidic dissociation of the phenoxyl anion (XII) in a fashion(14) OH (XII) (XIII) similar to Equation (6). This requires strong alkalinity, and if the medium alkalinity is not sufficient CH3O CH3O for this, then retroaldol reaction will fail, and instead oxidation of (XII) into byproducts (perhaps, intoStep ferulic (14) acid) proceeds will take through place. C-HThe latteracidic takes dissociation place, for of example, the phenoxyl in oxidation anion of (XII lignosulfonates) in a fashion similarat pH to 10.0–10.8 Equation (Figure (6). This 3). requires The assumed strong ferulicalkalinity, acid and formation if the medium produces alkalinity new carboxylic is not sufficient acidic for this, then retroaldol reaction will fail, and instead oxidation of (XII) into byproducts (perhaps,

into ferulic acid) will take place. The latter takes place, for example, in oxidation of lignosulfonates at pH 10.0–10.8 (Figure 3). The assumed ferulic acid formation produces new carboxylic acidic

Int. J. Mol. Sci. 2017, 18, 2421 15 of 29 Int. J. Mol. Sci. 2017, 18, 2421 15 of 29 Int. J. Mol. Sci. 2017, 18, 2421 15 of 29 drawback is that it cannot be extrapolated towards two more selective oxidants: nitrobenzene or copperdrawback oxide. is that it cannot be extrapolated towards two more selective oxidants: nitrobenzene or copper oxide. Hypothesis 3 (H3) A hypothesis about the mechanism of lignin oxidation that involves a retroaldolHypothesis reaction 3 (H3) step.A hypothesis Based on aboutvarious the experimental mechanism resuof ligninlts [86–88,98] oxidation and that discussions involves ina literatureretroaldol [85,95,99–101], reaction step. a Basedmechanism on various hypothesis experimental for formation resu oflts vanillin[86–88,98] in ligninand discussionsoxidation was in developedliterature [85,95,99–101], in 1997–2004. a Itmechanism unites the hypothesis universall yfo recognizedr formation ligninof vanillin chemistry in lignin principles oxidation about was phenoxyldeveloped radicals in 1997–2004. and quinone It unites methide the intermediatesuniversally recognized with the concept lignin ofchemistry retroaldol principles reaction asabout the finalphenoxyl process radicals step. Toand keep quinone the presentation methide intermediates of this hypothesis with the [86–88,98] concept ofbrief, retroaldol it will bereaction rendered as the as oxidationfinal process of eugenol step. To as keep an examplethe presentation model compound. of this hypothesis [86–88,98] brief, it will be rendered as oxidation of eugenol as an example model compound. The oxidation begins with abstraction of an electron from a phenolate anion: The oxidation begins with abstraction of an electron from a phenolate anion: -e O CH2 CH CH2 O CH2 CH CH2 -e O CH CH CH CH (10) O CH2 CH CH2 -e O CH2 CH CH2 (VII) 2 2 O (VIII) 2 CH CH2 (10) CH O (VII) CH O (VIII) (10) 3 (VII) 3 (VIII) CH O CH O CH3O CH3O Disproportionation3 of these primary radicals (VIII3 ) leads to a quinone methide (IX): Disproportionation of these primary radicals (VIII) leads to a quinone methide (IX):

2 O CH2 CH CH2

2 O (VIII) CH2 CH CH2 2 O CH2 CH CH2 CH O (VIII) 3 (VIII) CH O (11) CH3O 3 (11) Int. J. Mol. Sci. 2017, 18, 2421 HO CH2 CH CH2 + O CH CH CH2 (11)15 of 29

HO CH2 CH CH2 + O CH CH CH2 HO CH2 CH CH2 + O (IX)CH CH CH2 CH O CH3O 3 (IX) CH O (IX) CH3O CH3O and nucleophilicand nucleophilic addition addition ofCH a3O hydroxide of a hydroxide ion to ion the to latter the latter3 creates creates the the coniferyl coniferyl alcohol alcohol structure structure ((X). and nucleophilic addition of a hydroxide ion to the latter creates the coniferyl alcohol structure (X). - CH OH O CH CH2 - O CH CH CH2OH OH- O CH CH CH2 OH O CH CH CH2OH (12)(12) O (IX)CH CH CH2 O CH CH CH2OH (X) (12) CH3O (IX) CH3O (12) (IX) (X) CH3O CH O (X) CH3O CH3O It should be noted that formation of the quinone3 methide (IX) out of the phenoxyl radical It should be noted that formation of the quinone methide (IX) out of the phenoxyl radical (VIII) (VIIIIt) isshould also bepossible noted viathat acidicformation dissociation of the quinone of the lattermethide with (IX )the out following of the phenoxyl single-electron radical is also possible via acidic dissociation of the latter with the following single-electron oxidation of oxidation(VIII) is ofalso the possible resulting via radical-anion. acidic dissociation This possibility of the waslatter discussed with the by following Sheldon (Schemesingle-electron 3 [83]) the resultingandoxidation in radical-anion. our of works the resulting [86,87,102]. This radical-anion. possibility was This discussed possibility bywas Sheldon discussed (Scheme by Sheldon3[ 83 (Scheme]) and 3 in [83]) our works [86and,87, 102inSubsequent our]. works oxidation [86,87,102]. of coniferyl alcohol (X) is analogous to Reactions (10)–(12) and results in Subsequentthe γSubsequent-carbonyl oxidation group oxidation of coniferyl(XI). ofRetroaldol coniferyl alcohol reactionalcohol (X) is analogous( Xof) isthe analogous α-unsaturated to Reactions to Reactions aldehyde (10)–(12) (10)–(12) (XI and) yields and results results vanillin in the in γ-carbonyl(theXIII group γ):-carbonyl (XI). Retroaldolgroup (XI). reactionRetroaldol of reaction the α-unsaturated of the α-unsaturated aldehyde aldehyde (XI) yields (XI vanillin) yields vanillin (XIII): (XIII): H O CH CH CHO 2 O O CH CH2 CHO H2O O CH CH CHO H2O (13) O CH CH CHO O CH CH2 CHO (XI) O OHCH CH(XII)2 CHO (13) CH O (13)(13) 3 (XI) CH3O OH (XII) CH O (XI) OH (XII) CH3O CH3O 3 CH3O - OH O CH CH2 CHO - O CHO + CH2 CHO -HOHO- O CH CH CHO OH2 O CHO + CH (14) O CH CH2 CHO O (XIII) CHO + CH2 CHO OH (XII)2 -H2O 2 CHO (14) CH O -H2O (XIII) (14) 3 OH (XII) CH3O (XIII) (14) CH O OH (XII) CH O CH3O CH3O Step (14)3 proceeds through C-H acidic dissociation3 of the phenoxyl anion (XII) in a fashion Step (14) proceeds through C-H acidic dissociation of the phenoxyl anion (XII) in a fashion Stepsimilar (14)Step proceeds to Equation(14) proceeds through (6). This C-Hthrough requires acidic C-H dissociationstrong acidic alkalinity, dissociation of the and phenoxyl of if thethe mediumphenoxyl anion (alkalinity XIIanion) in ( aXII is fashion )not in sufficienta fashion similar forsimilar this, to then Equation retroaldol (6). Thisreaction requires will strongfail, and alkalinity, instead oxidationand if the ofmedium (XII) into alkalinity byproducts is not (perhaps,sufficient to Equationsimilar (6). Thisto Equation requires (6). strong This requires alkalinity, strong and alkalinity, if the medium and if the alkalinity medium alkalinity is not sufﬁcient is not sufficient for this, intofor this, ferulic then acid) retroaldol will take reaction place. Thewill latterfail, and takes instead place, oxidationfor example, of ( XIIin oxidation) into byproducts of lignosulfonates (perhaps, then retroaldol reaction will fail, and instead oxidation of (XII) into byproducts (perhaps, into ferulic atinto pH ferulic 10.0–10.8 acid) (Figurewill take 3). place. The Theassumed latter ferulictakes place, acid forformation example, produces in oxidation new ofcarboxylic lignosulfonates acidic acid) willat take pH place. 10.0–10.8 The latter(Figure takes 3). The place, assumed for example, ferulic inacid oxidation formation of lignosulfonatesproduces new carboxylic at pH 10.0–10.8 acidic at pH 10.0–10.8 (Figure 3). The assumed ferulic acid formation produces new carboxylic acidic (Figure3). The assumed ferulic acid formation produces new carboxylic acidic protons that neutralize Int. J. Mol. Sci. 2017, 18, 2421 16 of 29 the alkali and decrease the pH. Speaking of alkali consumption, other unintended oxidation reactions (e.g., causedprotons by the that excessively neutralize reactivethe alkali intermediate and decrease products the pH. ofSpeaking molecular of oxygenalkali consumption, reduction, moreother about themunintended later) will oxidation also eventually reactions generate (e.g., caused acidic by byproducts the excessively (other reactive carboxylic intermediate acids and products CO2) that of bind alkali.molecular oxygen reduction, more about them later) will also eventually generate acidic byproducts The suggested(other carboxylic mechanism acids allowsand CO explaining2) that bind alkali. the formation of the aceto-derivatives of the intended aldehydes by assumingThe suggested the addition mechanism of a hydroxideallows explaining ion into th thee formationα-position of of the the aceto-derivatives quinone methide of (IXthe) intended aldehydes by assuming the addition of a hydroxide ion into the α-position of the quinone with its subsequent oxidation and eventual retroaldol reaction of the resulting α-oxo-β-unsaturated methide (IX) with its subsequent oxidation and eventual retroaldol reaction of the resulting structure (XVII) (15): α-oxo-β-unsaturated structure (XVII) (15):

- - OH - e O CH CH CH2 O CH CH CH2 (IX) OH CH3O CH3O

+ - - e- - H - e O CH CH CH2 O C CH CH2

OH OH (XIV) CH O (XV) CH3O 3 (15)(15) - - - - e OH OH ,H2O O C CH CH2 O C CH CH2 OH (XVI) O (XVII) CH3O CH3O

- OH ,H2O O C CH3 + CH2O

O (XVIII) CH3O Reaction (15) demonstrates formation of the quinone methide (XVI) by acidic dissociation of the phenoxyl radical (XIV) with the subsequent oxidation of the anion-radical (XV), instead of the disproportionation (11) of (XIV). Hydroxide anion addition into the α-position of a phenylpropane unit is less thermodynamically favorable than addition into the γ position since the former disrupts the π conjugated chain [85,102]. Thus, under thermodynamic reaction control, aceto-derivatives will be a minor byproduct. According to literature data [85], nucleophilic addition to quinone methides is controlled by thermodynamics. However, one exception is known to the pattern concerning the high ratio of the aldehydes to their aceto-derivatives in the process products: In oxidation of birch phloem (inner bark), peak yields of syringaldehyde and acetosyringone are almost equal [103]. We should point out one important aspect of the mechanism (10)–(15) as well as of many other radical reactions involved in lignin transformations: Interconversion between a phenolate anion and a phenoxyl radical can be fast and reversible. This kind of electron transfer between semiquinone radicals and hydroquinone anions proceeds with rates close to that of diffusion [74,104]. This means that the ratio between the molecular species along the oxidation chain—eugenol, coniferyl alcohol, coniferyl aldehyde, vanillin—can be controlled to a significant degree by thermodynamics of the electron exchange in various combinations of their phenolate and phenoxyl radical forms. These reactions may determine high selectivity of vanillin formation (more specifically, low rate of vanillin oxidation). For example, in a Steelink’s study, in oxidation of syringyl alcohol, oxidation of syringaldehyde is only observed when the alcohol conversion of over 90% is reached [105]. Redox potentials of single-electron oxidation of eugenol, isoeugenol, coniferyl alcohol, and other phenols were derived in [106]; it shows that donor substituents decrease the potential, whereas electron-withdrawing substituents increase it. This means that in oxidation under thermodynamical control, vanillin is a more stable species than eugenol and coniferyl alcohol. Vanillin cannot be oxidized by the phenoxyl radical of eugenol. On the other hand, an otherwise (16) generated phenoxyl radical of vanillin can revert (17) to vanilloxide anion by oxidizing a phenolate ion of the

Int. J. Mol. Sci. 2017, 18, 2421 16 of 29

Reaction (15) demonstrates formation of the quinone methide (XVI) by acidic dissociation of the phenoxyl radical (XIV) with the subsequent oxidation of the anion-radical (XV), instead of the disproportionation (11) of (XIV). Hydroxide anion addition into the α-position of a phenylpropane unit is less thermodynamically favorable than addition into the γ position since the former disrupts the π conjugated chain [85,102]. Thus, under thermodynamic reaction control, aceto-derivatives will be a minor byproduct. According to literature data [85], nucleophilic addition to quinone methides is controlled by thermodynamics. However, one exception is known to the pattern concerning the high ratio of the aldehydes to their aceto-derivatives in the process products: In oxidation of birch phloem (inner bark), peak yields of syringaldehyde and acetosyringone are almost equal [103]. We should point out one important aspect of the mechanism (10)–(15) as well as of many other radical reactions involved in lignin transformations: Interconversion between a phenolate anion and a phenoxyl radical can be fast and reversible. This kind of electron transfer between semiquinone radicals and hydroquinone anions proceeds with rates close to that of diffusion [74,104]. This means that the ratio between the molecular species along the oxidation chain—eugenol, coniferyl alcohol, coniferyl aldehyde, vanillin—can be controlled to a signiﬁcant degree by thermodynamics of the electron exchange in various combinations of their phenolate and phenoxyl radical forms. These reactions may determine high selectivity of vanillin formation (more speciﬁcally, low rate of vanillin oxidation). For example, in a Steelink’s study, in oxidation of syringyl alcohol, oxidation of syringaldehyde is only observed when the alcohol conversion of over 90% is reached [105]. Redox potentials of single-electron oxidation of eugenol, isoeugenol, coniferyl alcohol, and other phenols were derived in [106]; it shows that donor substituents decrease the potential, whereas electron-withdrawing substituents increase it. This means that in oxidation under thermodynamical control, vanillin is a more stable species than eugenol and coniferyl alcohol. Vanillin cannot be oxidized by the phenoxyl Int. J. Mol. Sci. 2017, 18, 2421 17 of 29 radicalInt.of J. Mol. eugenol. Sci. 2017 On, 18, the 2421 other hand, an otherwise (16) generated phenoxyl radical of vanillin can17 of 29 revert (17) to vanilloxide anion by oxidizing a phenolate ion of the predecessor compounds (lignin, predecessor compounds (lignin, eugenol, coniferyl alcohol, etc.) into a corresponding phenoxyl eugenol,predecessor coniferyl compounds alcohol, etc.) (lignin, into a corresponding eugenol, coniferyl phenoxyl alcohol, radical etc.) of theinto latter: a corresponding phenoxyl radicalradical ofof thethe latter:latter:

- - - e O OO CHOCHO - e O CHOCHO (16)(16)(16) CH3O CH3O CH3O CH3O

O CHO O CHO + O CH2 CH CH2 + O CH2 CH CH2 CH O CH3O CH3O 3 CH3O (17)(17)(17)

O CHO + O CH2 CH CH2 O CHO + O CH2 CH CH2

CH3O CH3O CH3O CH3O On one hand, the presence of the electron-withdrawing carbonyl group in the para-position relative On oneOn hand,one hand, the the presence presence of of the the electron-withdrawing electron-withdrawing carbonyl group group in in the the parapara-position-position relative to the phenolic hydroxyl in the aromatic aldehydes contributes to their high yields in oxidation of relativeto tothe the phenolic phenolic hydroxyl hydroxyl in in the the ar aromaticomatic aldehydes contributes toto theirtheir highhigh yields yields in in oxidation oxidation of lignins (by increasing the redox potential of the former). On the other hand, the same withdrawing of ligninslignins (by (by increasing increasing the the redox redox potential potential of theof former).the former). On theOn otherthe other hand, ha thend, same the same withdrawing withdrawing substituent causes a negative effect related to increased phenolic acidity: the pKa of vanillin is 7.2 [107] substituentsubstituent causes causes a negative a negative effect effect related related to to increased increased phenolicphenolic acidity: acidity: the the pK paK ofa vanillinof vanillin is 7.2 is [107] which is lower than the phenolic pKa of lignins by approximately 3 [91]. So, in weakly alkaline media 7.2 [107which] which is lower is lower than the than phenolic the phenolic pKa of pligninsKa of ligninsby approximately by approximately 3 [91]. So, 3 in [91 weakly]. So, inalkaline weakly media (pH 8–9) vanillin is dissociated, whereas lignin is not. In this situation, vanillin—the intended alkaline(pH media 8–9) (pHvanillin 8–9) is vanillin dissociated, is dissociated, whereas whereaslignin is lignin not. isIn not. this In situation, this situation, vanillin—th vanillin—thee intended product—will be oxidized faster than lignin and therefore will be unstable. Thus, an alkali excess is intendedproduct—will product—will be oxidized be oxidized faster faster than thanlignin lignin and andtherefore therefore will willbe unstable. be unstable. Thus, Thus, an alkali an alkali excess is necessary for the product stabilization that is achieved via lignin phenolic dissociation. The similar excessnecessary is necessary for the for product the product stabilization stabilization that is that achi iseved achieved via lignin via ligninphenolic phenolic dissociation. dissociation. The similar differencedifference betweenbetween thethe aciditiesacidities ofof pp-hydroxybenzaldehyde-hydroxybenzaldehyde andand pp-cresol-cresol [90][90] explainsexplains whywhy superstoichiometricsuperstoichiometric amountsamounts ofof alkalialkali areare necessarynecessary forfor selectiveselective oxidationoxidation ofof thethe latterlatter [83].[83]. TheThe conductedconducted analysisanalysis showsshows thatthat thethe mechanismechanismm wewe presentedpresented forfor vanillinvanillin andand syringaldehydesyringaldehyde formationformation inin ligninlignin oxidationoxidation inin alkalinealkaline aqueousaqueous mediamedia hashas certaincertain advantagesadvantages overover otherother knownknown hypotheses.hypotheses. ItIt allowsallows explainingexplaining allall ofof thethe hereinhereinaboveabove outlinedoutlined patternspatterns ofof thethe process,process, andand alsoalso aa numbernumber ofof experimentalexperimental resultsresults ((R1R1––R4R4)) obtainedobtained inin testingtesting thisthis mechanism’smechanism’s validityvalidity [86,88,98]:[86,88,98]: ResultResult 11 (R1).(R1). InfluenceInfluence ofof pHpH onon thethe kineticskinetics ofof vanillinvanillin formationformation showsshows thethe samesame trendstrends inin thethe oxidationoxidation ofof vanillideneacetone,vanillideneacetone, eugenol,eugenol, andand lignin;lignin; thisthis supportssupports thethe rolerole ofof retroaldolretroaldol reactionreaction inin thethe testtest processes.processes. ResultResult 22 (R2).(R2). ConiferylConiferyl alcoholalcohol waswas detecteddetected asas anan intermedintermediateiate inin thethe processprocess ofof eugenoleugenol oxidation.oxidation. ResultResult 33 (R3).(R3). TheThe compositionscompositions ofof thethe aromaticaromatic productsproducts fromfrom oxidationoxidation ofof guaiacylpropanolguaiacylpropanol andand guaiacylethanolguaiacylethanol werewere fundamentallyfundamentally different;different; thisthis factfact isis consistentconsistent withwith thethe proposedproposed mechanismmechanism andand cannotcannot bebe explainedexplained inin termsterms ofof otherother well-knownwell-known hypotheses.hypotheses. ResultResult 44 (R4).(R4). TheThe oldold industrialindustrial processprocess ofof eugenoleugenol toto vanillinvanillin conversionconversion includedincluded aa separateseparate andand veryvery slowslow stepstep ofof eugenoleugenol toto isoeugenolisoeugenol isomerisomerizationization followedfollowed byby isoeugenolisoeugenol oxidationoxidation [8].[8]. WeWe showedshowed thatthat selectiveselective catalyticcatalytic oxidationoxidation ofof eugenoleugenol toto vanillinvanillin cancan bebe carriedcarried outout directlydirectly withoutwithout thethe isomerizationisomerization step.step. ThisThis factfact supportssupports thethe suggsuggestedested mechanism,mechanism, becausebecause thethe stepssteps (11)(11) andand (12)(12) leadlead toto thethe doubledouble bondbond migrationmigration fromfrom αα-- toto ββ-position-position duringduring thethe oxidationoxidation insteadinstead ofof thethe separateseparate eugenoleugenol isomerizationisomerization step.step. BasedBased onon thethe mechanismmechanism ofof thethe aromaticaromatic aldehydesaldehydes formationformation (10)–(15),(10)–(15), severalseveral conclusionsconclusions ((C1C1––C4C4)) andand recommendationsrecommendations aboutabout thethe choicechoice ofof catalystscatalysts andand thethe processprocess conditionsconditions forfor ligninlignin oxidationoxidation byby oxygenoxygen cancan bebe made.made. ConclusionConclusion 11 (C1).(C1). PhenolatePhenolate anionsanions areare amongamong thethe mostmost easilyeasily oxidizedoxidized functionalfunctional groupsgroups inin organicorganic chemistry,chemistry, andand oxidationoxidation ofof ligninslignins inin alkalinealkaline mediamedia proceedsproceeds withwith quitequite highhigh ratesrates eveneven withoutwithout catalysts.catalysts. ForFor thisthis reason,reason, thethe principalprincipal rolerole ofof catacatalystslysts inin thisthis processprocess isis increasingincreasing itsits selectivityselectivity insteadinstead ofof thethe oxidationoxidation rate.rate. ToTo thisthis end,end, assumingassuming stepwisestepwise interactioninteraction ofof thethe catalystcatalyst withwith phenolatephenolate anionanion andand thenthen withwith oxygen,oxygen, aa catalyst’scatalyst’s redoxredox potepotentialntial needsneeds toto bebe asas lowlow asas possiblepossible whilewhile beingbeing

Int. J. Mol. Sci. 2017, 18, 2421 17 of 29

The similar difference between the acidities of p-hydroxybenzaldehyde and p-cresol [90] explains why superstoichiometric amounts of alkali are necessary for selective oxidation of the latter [83]. The conducted analysis shows that the mechanism we presented for vanillin and syringaldehyde formation in lignin oxidation in alkaline aqueous media has certain advantages over other known hypotheses. It allows explaining all of the hereinabove outlined patterns of the process, and also a number of experimental results (R1–R4) obtained in testing this mechanism’s validity [86,88,98]: Result 1 (R1). Influence of pH on the kinetics of vanillin formation shows the same trends in the oxidation of vanillideneacetone, eugenol, and lignin; this supports the role of retroaldol reaction in the test processes. Result 2 (R2). Coniferyl alcohol was detected as an intermediate in the process of eugenol oxidation. Result 3 (R3). The compositions of the aromatic products from oxidation of guaiacylpropanol and guaiacylethanol were fundamentally different; this fact is consistent with the proposed mechanism and cannot be explained in terms of other well-known hypotheses. Result 4 (R4). The old industrial process of eugenol to vanillin conversion included a separate and very slow step of eugenol to isoeugenol isomerization followed by isoeugenol oxidation [8]. We showed that selective catalytic oxidation of eugenol to vanillin can be carried out directly without the isomerization step. This fact supports the suggested mechanism, because the steps (11) and (12) lead to the double bond migration from α- to β-position during the oxidation instead of the separate eugenol isomerization step. Based on the mechanism of the aromatic aldehydes formation (10)–(15), several conclusions (C1–C4) and recommendations about the choice of catalysts and the process conditions for lignin oxidation by oxygen can be made. Conclusion 1 (C1). Phenolate anions are among the most easily oxidized functional groups in organic chemistry, and oxidation of lignins in alkaline media proceeds with quite high rates even without catalysts. For this reason, the principal role of catalysts in this process is increasing its selectivity instead of the oxidation rate. To this end, assuming stepwise interaction of the catalyst with phenolate anion and then with oxygen, a catalyst’s redox potential needs to be as low as possible while being able to oxidize the phenolate anion. The optimal redox potential for a catalyst of lignin oxidation can be qualitatively characterized: it needs to be between the redox potentials of single-electron oxidation of vanillin and of, e.g., eugenol. In the context of the mechanism (10)–(15), the oxidized form of an optimal catalyst withdraws an electron from phenolate anions that are predecessors to vanillin (10)–(12), and should have sufficiently low oxidation potential to not affect more deeply oxidized ligneous phenolic species and vanillin. The reduced form of the catalyst should not lead to formation of free active oxygen-containing radicals upon reaction with molecular oxygen. A traditional and quite selective catalyst (also, a stoichiometric oxidant [30]) of the process is copper oxide, it has modest electrode potential (−0.16 V for CuO/Cu2O redox pair at pH 14) [108]. Sliver oxide has higher potential (+0.34 V at pH 14) [108], and using it in lignin oxidation results in higher vanillic acid yield compared to copper oxide [24], i.e., oxidation of vanillin takes place. This comparison shows that an electrode potential +0.34 V in alkaline media is already excessive for a selective catalyst. As far as vanillin stability is concerned, it is worth mentioning that molecular oxygen − in alkaline media has a redox potential of +0.40 V, the figure is about the same for superoxide O2 , while hydroxyl radical HO has an electrode potential +2.0 V in alkaline media [108]. All these figures are too high in terms of vanillin preservation. Therefore, a good catalyst of the process might fulfill the important role: decreasing the impact of oxygen and its radical forms on the process selectivity by consuming them via reactions with reduced catalyst species. However, an important conclusion to the discussion about the suitable catalyst electrode potential should be that the catalytic reaction cycle with any catalyst might occur without formation of unoccupied reduced catalyst sites as intermediates. This type of mechanism would reduce the influence of the catalyst redox potential on its ability Int. J. Mol. Sci. 2017, 18, 2421 18 of 29 to maintain the process selectivity (even mild oxidants may prove to be good catalysts of vanillin oxidation if the mechanism does not actually involve redox cycling of unoccupied catalytic sites). Conclusion 2 (C2). Another way of improving the selectivity is suppression of the phenoxyl radicals dimerization pathway (3) in favor of their oxidation (11) and (15). The most straightforward approach here is to increase temperature (see Section6)[73,104,109]. Conclusion 3 (C3). It is worth noting that the described mechanism of selective lignin oxidation into vanillin involves radicals but is not essentially a chain reaction: the phenoxyl radical (VIII) formed at the initiation stage (10) is transformed into a molecular product without chain propagation. Therefore, the mechanism (10)–(14) and (15) can be considered chain termination. Non-catalytic oxidation by oxygen typically proceeds as chain reactions involving alkoxyl, hydroxyl, and peroxyl radicals that are too active to maintain the process selectivity. Therefore, the optimal catalyst and process conditions need to suppress the oxidation chain propagation. It was noted in Section7 that apparent chain length in the process of powdered aspen wood oxidation decreases to as low as unit at alkali concentration 1–2 mol/L. At this point, we can name three phenomena via which high reaction media alkalinity contributes to better accumulation of the aldehydes: (1) promotion of retroaldol reaction, (2) extension of phenolic dissociation to lignin molecules as opposed to only the intended products, (3) suppression of oxidation chain propagation by increasing the chain initiation rate. Conclusion 4 (C4). Based on the offered mechanism, it is not difficult to estimate the minimum oxygen 1 amount necessary to obtain one mol of vanillin. e.g., in oxidation of coniferyl alcohol, it is 2 mol O2 per mol of vanillin. In oxidation of lignin, this oxygen requirement may be even lower since lignin is formed from coniferyl and sinapyl alcohols via condensation after their single-electron oxidation into the corresponding phenoxyl radicals [22]. As was previously noted, in practice, oxygen consumption is immensely higher, alkali consumption is therefore much higher than theoretically needed too. Decreasing the consumption of these substances is an unresolved and clearly attention-worthy problem. Detailed insight into the nature of this problem lies outside the scope of this review, but at a glance, a few factors contributing to it can be mentioned. The described mechanism depends on a series of single-electron oxidation attacks on phenolate − anions, and when an oxygen molecule performs such attack, a superoxide anion-radical O2 is generated. The latter, being an anion, will experience electrostatic hindrance when trying to attack the next phenolate anion, and is more likely to be instead wasted on some particle irrelevant to the formation of the intended product. i.e., while an oxygen molecule could perform up to four acts of single-electron oxidation before being fully reduced, not all of this capacity is guaranteed to go towards the aldehydes formation, hence low selectivity based on oxygen. Overcoming this problem would require a challenging job of finding a catalyst which ensures not only binding of all intermediate oxygen species (as described above), but also their selective participation in lignin oxidation. Another requirement for high selectivity is confining the action of oxygen to the surface of this catalyst (decreasing the contribution of the unselective non-catalytic process). Finally, certain components of wood (e.g., extractives, hemicelluloses) can be removed prior to the process so that additional oxygen is not spent on their oxidation. However, the latter two factors relate more to the process implementation design, and as such will be discussed more in the following Section.

9. Development Prospects for the Technology of the Aromatic Aldehydes Production from Lignins The concluding Section reviews the data on the results concerning technological aspects of lignin oxidation. Int. J. Mol. Sci. 2017, 18, 2421 19 of 29

9.1. Oxidation ofInt. aJ. Mol. Lignosulfonate Sci. 2017, 18, 2421 Solution or Aspen Wood Slurry in a Flow Reactor System19 of 29

The kineticO2− trendsis generated. of lignosulfonateThe latter, being an anion, oxidation will experience in a shaking electrostatic batch hindrance reactor when [trying43] described to in the attack the next phenolate anion, and is more likely to be instead wasted on some particle irrelevant previous Sectionsto the were formation used of tothe devise intended a product. similar i.e., process while an of oxygen continuous molecule could operation perform [ 50up ,to110 four]. The oxidation ◦ of lignosulfonatesacts of was single-electron carriedout oxidation at 170 beforeC withbeing fully operating reduced, pressurenot all of this around capacity 1is MPaguaranteed using to an installation with two fixed-bedgo towards catalytic the aldehydes reactors formation, connected hence inlowseries selectivity completely based on oxygen. filled Overcoming with cupric this oxide catalyst problem would require a challenging job of finding a catalyst which ensures not only binding of all (Scheme4). Theintermediate reaction oxygen solution species and (as oxygen described were above), fed but into also thetheir reactors selective participation from top toin bottom,lignin achieving falling film flowoxidation. of reactants Another requirement along the for surface high selectivity of the is catalyst confining grains.the action of oxygen to the surface of this catalyst (decreasing the contribution of the unselective non-catalytic process). Finally, certain The rate ofcomponents oxygen of flowwood (e.g., has extractives, a significant hemicellul effectoses) on can vanillinbe removed yield prior to in the oxidation process so that of sulfite liquor: increasing it byadditional the factor oxygen of 2–4is not leads spent on to their three-fold oxidation. improvement However, the latter of two the factors vanillin relate concentration more to the (Figure4). This effect canprocess be explained implementation by assuming design, and as that such the will higher be discussed gas more amount in the following displaces Section. the liquid, thus leading to a thinner liquid9. Development layer on Prospects the outer for the surface Technology of of the the catalystAromatic Aldehydes grains. ThisProduction in turn results in a shorter diffusion pathfrom through Lignins the liquid that oxygen has to travel to reach the catalyst surface and/or deep reactant-rich liquidThe region. concluding Overall, Section thisreviews leads the data to higher on the results oxidation concerning selectivity, technological as was aspects described of in detail lignin oxidation. in Section5. At flow rates above 40 L/h, the yield decrease can be explained by lower liquid residence time inside the9.1. reactors, Oxidation of a which Lignosulfonate is insufficient Solution or Aspen for Wood attaining Slurry in a adequateFlow Reactor System oxidation extent. Maximum aldehydes yield attainsThe kinetic 9.9 trends wt. of % lignosulfonate based on oxidation the lignin, in a shaking and batch the liquidreactor [43] reaction described mass in the hourly space velocity throughprevious the Sections reactors were attained used to devise 1–2h a− similar1. For process comparison, of continuous we operation obtained [50,110]. 12 wt. The % yield of the oxidation of lignosulfonates was carried out at 170 °C with operating pressure around 1 MPa using aldehydes in aan batch installation reactor with using two fixed-bed oxygen, catalytic andoxidation reactors connected of lignosulfonates in series completely by nitrobenzenefilled with yields up to 16 wt. % ofcupric the aldehydes oxide catalyst based (Scheme on 4). theThe reaction lignin. solution and oxygen were fed into the reactors from top to bottom, achieving falling film flow of reactants along the surface of the catalyst grains.

Int. J. Mol. Sci. 2017, 18, 2421 20 of 29

described in detail in Section 5. At flow rates above 40 L/h, the yield decrease can be explained by lower liquid residenceScheme 4. The time flow inside system for the lignosulfonate reactors, oxidation.which is (1) insufficient initial reaction solutionfor attaining tank; (2) plungeradequate oxidation Schemeextent. Maximum 4. Thepump ﬂow aldehydes(1–10 system L/h, 100 for yieldkgf/cm lignosulfonate 2);attains (3) heater 9.9 of wt.liquid oxidation. % (0.25 ba sedL); (4) (1) on reactor initial the (1.0lignin, reaction L); (5) and reactor solution the (0.25 liquid L);tank; (6) reaction (2) plunger mass pumphourly (1–10 space L/h, cooler;velocity 100 (7) separator; kgf/cmthrough (8)2 );the solenoid (3) reactors heater valve; ofattained(9) liquidreaction 1–2mass (0.25 hsampling− L);1. For (4) vessel;comparison, reactor (10) (1.0manometers; L);we (5)obtained (11) reactor 12 (0.25 wt. L);% rotameter; (12) regulating valves. (6)yield cooler; of the (7) aldehydes separator; in (8) a solenoid batch reactor valve; (9)using reaction oxygen, mass and sampling oxidation vessel; of lignosulfonates (10) manometers; by (11)nitrobenzene rotameter; yieldsThe (12) rate regulating up of oxygento 16 wt. flow valves. % has of athe significant aldehydes effect basedon vanillin on theyield lignin. in oxidation of sulfite liquor: increasing it by the factor of 2–4 leads to three-fold improvement of the vanillin concentration (Figure 4). This effect can be explained by assuming that the higher gas amount displaces the liquid, 14 thus leading to a thinner liquid layer on the outer surface of the catalyst grains. This in turn results in a shorter diffusion path through12 the liquid that oxygen has to travel to reach the catalyst surface and/or deep reactant-rich liquid region. Overall, this leads to higher oxidation selectivity, as was 10 1

Vanillin, g/l 4 2 2

0 20406080100120 Gas flow rate, l/h

FigureFigure 4. The 4. influenceThe influence of oxygen of oxygen flow rateflow onrate yield on yiel of vanillind of vanillin (1) and (1) syringaldehyde and syringaldehyde (2) in (2) oxidation in of sulfiteoxidation liquor of in sulfite the flow liquor reactor in the system. flow reactor 170 ◦ system.C, 0.9 MPa, 170 °C, 210 0.9 g/L MPa, of sulfite210 g/L liquor of sulfite dry liquor mass, dry 150 g/L of NaOH,mass, liquid150 g/L flow of NaOH, rate 1.8 liquid L/h flow (see rate reference 1.8 L/h (see [110 reference]). [110]).

It is well known that in catalytic processes that involve gaseous, liquid, and solid phases only the catalyst grain outer surface is primarily active [111,112]. Therefore, the high catalyst surface area necessary for a good yield of the products can be practically achieved only by decreasing the catalyst grain size. This would lead to high hydrodynamic flow resistance which constrains the maximum possible flow system efficiency. When discussing practical implementations of catalytic lignin oxidation, we should point out another factor that restricts utilization of a catalyst’s inner surface in these processes. The aldehydes resulting from oxidation of lignins can be oxidized in lignin-free media (it should be emphasized, in lignin-free media) with rates that are comparable to the rate of their formation [43,54,113]. This means that under flow conditions there needs to be a very tight distribution of catalyst contact time for the reacting molecules. It is quite obvious that the contact time for reactant particles that are carried along the catalyst outer surface is considerably shorter than for those that diffuse inside the catalyst grains. Therefore, the aldehyde molecules that form on the catalyst inner surface are likely to transform into deep oxidation products before diffusing out into the liquid flow. Oxidation of finely powdered (less 0.1 mm) aspen wood slurry in the same flow installation (Scheme 4) was accompanied by complete wood and cellulose dissolution. Yields of the aldehydes up to 37% based on lignin were attained [50]. This corresponds to 80–85% of the selectivity in oxidation by nitrobenzene (Table 1). This aldehyde yield is several times greater than anything attainable with technical lignins (Tables 1 and 2). However, utilization of native lignin oxidation for production of the aldehydes alone is likely not an economically effective endeavor because potentially valuable cellulose and hemicelluloses are lost in the oxidation process. Moreover, oxidation of these carbohydrates requires additional consumption of oxygen (and consequentially, of alkali). Thus, an economically viable approach to produce vanillin and syringaldehyde must be based around comprehensive wood processing that involves production of value-added chemicals from cellulose and hemicelluloses.

Int. J. Mol. Sci. 2017, 18, 2421 20 of 29

It is well known that in catalytic processes that involve gaseous, liquid, and solid phases only the catalyst grain outer surface is primarily active [111,112]. Therefore, the high catalyst surface area necessary for a good yield of the products can be practically achieved only by decreasing the catalyst grain size. This would lead to high hydrodynamic flow resistance which constrains the maximum possible flow system efficiency. When discussing practical implementations of catalytic lignin oxidation, we should point out another factor that restricts utilization of a catalyst’s inner surface in these processes. The aldehydes resulting from oxidation of lignins can be oxidized in lignin-free media (it should be emphasized, in lignin-free media) with rates that are comparable to the rate of their formation [43,54,113]. This means that under flow conditions there needs to be a very tight distribution of catalyst contact time for the reacting molecules. It is quite obvious that the contact time for reactant particles that are carried along the catalyst outer surface is considerably shorter than for those that diffuse inside the catalyst grains. Therefore, the aldehyde molecules that form on the catalyst inner surface are likely to transform into deep oxidation products before diffusing out into the liquid flow. Oxidation of finely powdered (less 0.1 mm) aspen wood slurry in the same flow installation (Scheme4) was accompanied by complete wood and cellulose dissolution. Yields of the aldehydes up to 37% based on lignin were attained [50]. This corresponds to 80–85% of the selectivity in oxidation by nitrobenzene (Table1). This aldehyde yield is several times greater than anything attainable with technical lignins (Tables1 and2). However, utilization of native lignin oxidation for production of the aldehydes alone is likely not an economically effective endeavor because potentially valuable cellulose and hemicelluloses are lost in the oxidation process. Moreover, oxidation of these carbohydrates requires additional consumption of oxygen (and consequentially, of alkali). Thus, an economically viable approach to produce vanillin and syringaldehyde must be based around comprehensive wood processing that involves production of value-added chemicals from cellulose and hemicelluloses.

9.2. The Prospects of Comprehensive Wood Processing into the Aromatic Aldehydes and Valuable Carbohydrate-Derived Products It was mentioned earlier that vanillin production by oxidation of lignosulfonates dominated the industry in 1950–1970’s. With the exception of “Borregaard LignoTech”, this technology was universally supplanted by the guaiacol-glyoxylic synthesis. There are many reasons for this shift, and firstly an environmental one has to be stated: production of one tonne of vanillin from lignins results in over 100 m3 of waste water [8]. On the other hand, waste water volumes largely depend on the technology development level: for example, in the pulping industry this amount used to attain 60–160 m3 per tonne of cellulose [114], but in modern implementations it is near zero. Another considerable drawback of lignin oxidation is high alkali consumption (that attained 7–14 kg per kg of vanillin [43,44,115]). Both these problems stem from low vanillin yield from lignosulfonates, 7–12 wt. %. Considerably higher amounts can be obtained from brown rotted pine wood, 20% (Tables 1 and 2). This 1.5–3 fold gain of the aldehyde yield from enzymatic lignin as opposed to lignosulfonates is a clear advantage of the former. However, the vast natural resources of brown rotted wood are sparsely dispersed in nature, and thus are not a viable industrial feedstock. Enzymatic technologies of biobutanol production are in intense development, and their Int. J. Mol. Sci. 2017, 18, 2421 21 of 29 byproduct is a substance that is quite similar in structure and properties to lignin of brown rotted wood [116–118]. It was shown that the enzymatic lignin from biobutanol production can be a promisingApparently, feedstock the possibilityfor production of obtaining of vanillin the aromaticand syringaldehyde aldehydes and [39]. cellulose together by oxidation of hardwoodEven higher and vanillin softwood yields was can first be described obtained in in patents oxidation [119 of,120 softwood]. Catalytic native oxidation lignins, of 23–28 pine wt. wood % basedinto vanillin on lignin and [29,42]. cellulose However, with subsequent for the enzymaticreasons stated hydrolysis herein of above, the latter production into glucose of vanillin was recently and syringaldehydeexplored in more from detail native [121 ].lignins The maximum can only be attained economically vanillin efficient yield is 18.6if th wt.e carbohydrates % based on the of wood initial arelignin also (Figure processed5). Cellulose into valuable yield variescompounds. between 64% and 93% (based on its initial content in the wood), dependingApparently, on oxidation the possibility conditions. of Theobtaining best compromise the aromatic achieved aldehydes is 17 wt.and % cellulose vanillin (basedtogether on theby oxidationinitial lignin) of hardwood and 84% cellulose and softwood (based was on thefirst initial descri cellulosebed in patents in wood). [119,120]. Catalytic oxidation of pine Thus,wood there into existvanillin certain and conditions cellulose under with whichsubsequent the catalytic enzymati oxidativec hydrolysis production of the of vanillin latter into also glucosefulfills the was role recently of oxidative explored wood in more delignification, detail [121] and. The it ismaximum quite selective attained in this vanillin regard. yield The is cellulose18.6 wt. %obtained based on in the this initial manner lignin may (Figure not be 5). suitable Cellulose for high-qualityyield varies between paper production, 64% and 93% but (based there areon noits initialrestrictions content for in its the further wood), processing depending into on glucose, oxidation levulinic conditions. acid, The 5-hydroxymethylfurfural, best compromise achieved and otheris 17 wt.products. % vanillin These (based results on [ 119the– initial121] reveal lignin) the and possibility 84% cellulose of implementing (based on the a novel initial approach cellulose toin catalyticwood). woodThus, conversion there exist into certain an assortment conditions of fine under chemicals. which the catalytic oxidative production of vanillin also fulfillsEffective the comprehensive role of oxidative wood wood processing delignificat intoion, these and fine it is chemicals quite selective should in in this general regard. include The cellulosethree principal obtained stages: in this (1) prehydrolysismanner may not and be hemicellulose suitable for processing;high-quality (2) paper oxidative production, conversion but ofthere the arelignocellulosic no restrictions residue for its into further vanillin, processing syringaldehyde, into glucose, and cellulose;levulinic (3)acid, processing 5-hydroxymethylfurfural, the cellulose into andglucose, other levulinic products. acid, These 5-hydroxymethylfurfural, results [119–121] reveal andthe possibility other products of implementing (Scheme5). Depending a novel approach on the totype catalytic of wood, wood these conversion stages may into be an preceded assortment by removing of fine chemicals. wood extractives, phenols, etc.

100

60 Yields of cellulose, % cellulose, of Yields

50 0 5 10 15 20 Vanillin yields, % Figure 5. Vanillin and cellulose yields in pine wood catalytic oxidation in alkaline aqueous media at 160 °C◦C[ [121].121].

The stage of hemicellulose removal allows both reducing consumption of the reactants during catalytic oxidation, and obtaining additional quantities of furfural or levulinic acid [122]. We are currently developing other ways of decreasing the oxygen and alkali consumption that will soon be revealed. It should be noted that softwood and hardwood considerably differ not only in structure of lignin but also of hemicelluloses: softwoods contain 5–9% of pentosans as opposed to 16–24% in hardwoods (aspen and birch respectively) [123]. So, softwoods are not particularly suitable for furfural production, and birch is preferable to aspen. Syringaldehyde that forms together with vanillin in oxidation of hardwood lignins should preferentially be obtained from birch wood as well, as it produces better syringaldehyde to vanillin ratio than aspen [50,51]. Technical lignins are not very suitable for syringaldehyde production with the exception of lignins from enzymatic hydrolysis [39]. The problem of separating the product mixture of syringaldehyde and vanillin can be solved, for example, via precipitation of a syringaldehyde aldimine [124–128]. All the mentioned products with the exception of furfural are sparingly volatile and are typically isolated from aqueous solutions by extraction, an overview of the relevant methods can be found in [126,128]. Int. J. Mol. Sci. 2017, 18, 2421 22 of 29

Decreasing the volumes of the waste water from the process may be accomplished together with alkali regeneration: Similarly to procedures used in Kraft pulping [129], organic contaminants are oxidized in the liquid phase or combusted and the remaining soda ash is caustified. Residue from alkaline oxidation of native lignins can be used as pigments for timber and chipboards, as plant growth stimulators in farming, etc. [130]. Unlike similar substances derived from coal or lignosulfonates, these products do not contain carcinogens or sulfur compounds. As a conclusion to this Section, Table5 presents the data on global production and prices of the relevant products. The global vanillin market amasses 15–20 thousand tonnes annually at the price 10–15 USD/kg, and the world’s production capacity outweighs the market demand [131]. Global price of furfural is around 1–2 USD/kg, levulinic acid is worth around 3 USD/kg. Levulinic acid is primarily used in the polymer industry for synthesis of diphenolic acid and porophors [132,133]. Syringaldehyde is potentially the most valuable product among the discussed ones, serving as a precursor to 3,4,5-trimethoxybenzaldehyde the price of which exceeds 25 USD/kg. 3,4,5-trimethoxybenzaldehyde is probably ordinarily synthesized from vanillin [16] and is used in production of pharmaceuticals and other speciality chemicals. Market value estimates of the products from one tonne of birch wood processed according to Int. J. Mol. Sci. 2017, 18, 2421 22 of 29 the suggested approach attains 1800–2400 USD and potentially even more if production of xylitol is implemented.Effective comprehensive For comparison, wood theprocessing value of into the these cellulose fine chemicals alone obtained should from in general that wood include is 250–300three principal USD. stages: (1) prehydrolysis and hemicellulose processing; (2) oxidative conversion of the lignocellulosicSignificant portion residue of the into gross vanillin, product syringaldehyde, value depends and on syringaldehydecellulose; (3) processing and xylitol, the which cellulose are exclusiveinto glucose, to hardwood levulinic acid, processing. 5-hydroxymethylfurfural, Assuming sufficient and demand other forproducts these products,(Scheme 5). the Depending processing on of birchthe type wood of wood, will be these more stages economically may be efficientpreceded than by removing the similar wood processing extractives, of softwoods. phenols, etc. Hardwood

Pentoses Acid-catalyzed Lignocellulose prehydrolysis

Acid-catalyzed CHO CHO conversion O Furfural Catalytic oxidation OCH3 O2 Vanillin OH CHO

CH3O OCH3 SyringaldehideOH Cellulose

HOCH CHO 2 O 5-Hydroxymethylfurfural Acid-catalyzed conversion

COOH

Levulinic acid O

Scheme 5. Comprehensive processing ofof hardwoodhardwood involvinginvolving oxidationoxidation ofof lignin.lignin.

Int. J. Mol. Sci. 2017, 18, 2421 23 of 29

Table 5. Market ﬁgures for the products of birch wood processing.

Price, Global Production, Yield from One Value of the Product from Substance 103 USD/t 103 t/year Tonne of Wood, kg One Tonne of Wood, USD Vanillin 10–15 15–20 20 200–300 3,4,5-trimethoxybenzaldehyde from 25–30 10–20 50 1250–1500 syringaldehyde Levulinic acid 3 2–5 100–120 300–360 Furfural from pentosanes 1–2 90 100–120 100–240 Xylitol from xylose (in pentosanes) 6–8 20–30 130–180 800–1200

10. Conclusions The review of literature data reveals that properly organized processes of catalytic oxidation of various lignins are only insignificantly (10–15%) inferior to oxidation by nitrobenzene in terms of yield and selectivity in vanillin and syringaldehyde. Traditionally applied copper, manganese, and cobalt catalysts insignificantly differ in the selectivity of oxidation, and they do not perform appreciably worse than later developed catalytic systems. Heterogeneous catalysts based on copper oxide were used (and are currently used [115]) in industrial lignosulfonate oxidation, and will probably remain the most suitable for the processes of this kind. These catalysts have numerous advantages in lignin oxidation processes: low cost, resistance to poisoning, high selectivity, ease of separation from the aromatic aldehydes (by extraction) and cellulose (based on density difference). Among the as yet unresolved problems in oxidation of lignins into the aromatic aldehydes, very high consumption of oxygen (and consequentially, of alkali) is highlighted—over 10 mol per mol of obtained vanillin. While the presently known catalysts allow achieving yields of the aldehydes that are very close to the results of oxidation by nitrobenzene, there is still an unexplored challenge of finding catalysts that will also markedly decrease the consumption of oxygen and alkali. Among the discussed hypotheses of the aromatic aldehydes’ formation mechanism, the most convincing one is the pathway that involves formation of the aldehyde molecule by retroaldol reaction of substituted coniferyl aldehyde (or β-oxycarbonyl structure). The latter structure is generated from lignin by sequential acts of single-electron oxidation of phenolate ions. This mechanism does not depend on the oxidant nature, and hence is a more general pathway compared to other hypotheses [9,93]. This mechanism explains the two most important trends of lignin catalytic oxidation by oxygen: formation of the aceto-derivatives of the intended aldehydes as byproducts, as well as the great difference between minimal pH needed for oxygen consumption as such (>9), and formation of vanillin as a product (>11). The possibility of producing both cellulose and the aromatic aldehydes in one-stage catalytic oxidation of wood [119–122] is discussed. These results pave the road towards a novel technology of comprehensive wood processing into vanillin, syringaldehyde, and carbohydrate-derived products like levulinic acid and furfural. Native lignins yield 2–3 times more of the aromatic aldehydes compared to lignosulfonates, and it is a strong argument supporting the technology development of the direct wood conversion into platform chemicals. The carbohydrate components of wood can be processed by various known methods, when hemicelluloses are removed prior to the catalytic oxidation of wood, and when the cellulose is isolated afterwards. Syringaldehyde is a possible precursor to 3,4,5-trimethoxybenzaldehyde and pharmaceuticals derived from the latter, and it is not produced in industry by lignin oxidation. Synthetic approaches to syringaldehyde production are more costly compared to vanillin [16], while production of the former by hardwood lignin oxidation can be cheaper compared to lignin-derived vanillin. This is because of a higher yield of syringaldehyde from hardwood lignins (up to 30 wt. % of lignin) than of vanillin from softwood lignins (up to 20%). Int. J. Mol. Sci. 2017, 18, 2421 24 of 29

Three major industrial technologies of vanillin synthesis were implemented throughout history: oxidation of eugenol (the ﬁrst half of the twentieth century), oxidation of lignosulfonates (1940–1970’s) [8], and the contemporary technology with guaiacol and glyoxylic acid. The glyoxylic technology currently dominates the market with the exception of Borregaard that utilizes both lignin oxidation and the glyoxylic method to manufacture vanillin, and this suggests that both technologies are comparable in economic viability. The discussed results and concepts can serve as the basis for developing a technology for production of the aromatic aldehydes from renewable plant matter that may supersede the current technology of vanillin production from petrochemicals.

Acknowledgments: The reported study was funded by Russian Foundation for Basic Research, Government of Krasnoyarsk Territory, Krasnoyarsk Region Science and Technology Support Fund to the research project No: 16-43-242102. This includes covering the cost of publishing the present review. The authors thank M. A. Smirnova for valuable technical assistance. Author Contributions: Valery E. Tarabanko wrote Sections1–4 and Sections6–10 in the first variant of the review; Nikolay Tarabanko wrote Section5 in the first variant of the review; Valery E. Tarabanko and Nikolay Tarabanko discussed its moot points; Nikolay Tarabanko translated the text from the authors’ native language into English; Valery E. Tarabanko and Nikolay Tarabanko discussed the entire first variant of the review, improved it and wrote the final version of the manuscript. Conflicts of Interest: The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.

Abbreviations

PPU Phenylpropane units NBO Nitrobenzene oxidation

References

1. Brauns, F.E.; Brauns, D.A. The Chemistry of Lignin; Lesnaya Promyshlennost: Moskow, Russia, 1964; p. 583. 2. Zakzeski, J.; Bruijnincx, P.C.A.; Jongerius, A.L.; Weckhuysen, B.M. The Catalytic valorization of lignin for the production of renewable chemicals. Chem. Rev. 2010, 110, 3552–3599. [CrossRef][PubMed] 3. Pandey, M.P.; Kim, C.S. Lignin depolymerization and conversion: A Review of thermochemical methods. Chem. Eng. Technol. 2011, 34, 29–41. [CrossRef] 4. Li, C.Z.; Zhao, X.C.; Wang, A.Q.; Huber, G.W.; Zhang, T. Catalytic transformation of lignin for the production of chemicals and fuels. Chem. Rev. 2015, 115, 11559–11624. [CrossRef][PubMed] 5. Lange, H.; Decina, S.; Crestini, C. Oxidative upgrade of lignin—Recent routes reviewed. Eur. Polym. J. 2013, 49, 1151–1173. [CrossRef] 6. Ma, R.; Xu, Y.; Zhang, X. Catalytic oxidation of bioreﬁnery lignin to value-added chemicals to support sustainable biofuel production. ChemSusChem 2015, 8, 24–51. [CrossRef][PubMed] 7. Behling, R.; Valange, S.; Chatel, G. Heterogeneous catalytic oxidation for lignin valorization into valuable chemicals: What results? What limitations? What trends? Green Chem. 2016, 18, 1839–1854. [CrossRef] 8. Mark, H.V. (Ed.) Encyclopedia of Chemical Technology, 2nd ed.; Interscience Publishers: New York, NY, USA; John Wiley & Sons: London, UK, 1972; Volume 21, pp. 80–196. 9. Dardelet, S.; Froment, P.; Lacoste, N.; Robert, A. Aldehyde syringique. Possibilites de production a partir de bois feuillus. Rev. ATIP 1985, 39, 267–274. 10. Gogotov, A.F. Analysis of raw material base for obtaining of aromatic aldehides. Khimiya Rastitel’nogo Syr’ya 1999, 2, 73–79. 11. Emsley, J. Plant a tree for chemistry. New Sci. 1987, 116, 39–42. 12. Gogotov, A.F.; Babkin, V.A. Lignin—Potential source of value low-molecular substances. Khimiya Rastitel’nogo Syr’ya 1994, 2, 597–604. 13. Mashkovsky, M.D. Medicinal Products; Novaya Volna: Moskow, Russia, 2014. 14. Szallasi, A.; Blumberg, P.M. Mechanisms and terapeutic potentials of vanilloids (capsaicin-like molecules). Adv. Pharmacol. 1993, 24, 123–155. [PubMed] Int. J. Mol. Sci. 2017, 18, 2421 25 of 29

15. MacRae, W.D.; Towers, G.H.N. Biological activities of lignans. Phytochemistry 1984, 23, 1207–1220. [CrossRef] 16. Erofeev, Y.V.; Afanas’eva, V.L.; Glushkov, R.V. Synthetic routes to 3,4,5-trimethoxybenzaldehyde. Pharm. Chem. J. 1990, 24, 501–510. [CrossRef] 17. Heitner, C.; Dimmel, D.R.; Schmidt, J.A. (Eds.) Lignin and Lignans: Advances in Chemistry; CRC Press: London, UK; New York, NY, USA, 2010; p. 651. 18. Nguyen, H.T.H.; Reis, M.H.; Qi, P.X.; Miller, S.A. Polyethylene ferulate (PEF) and congeners: Polystyrene mimics derived from biorenewable aromatics. Green Chem. 2015, 17, 4512–4517. [CrossRef] 19. Fache, M.; Boutevin, B.; Caillol, S. Vanillin production from lignin and its use as a renewable chemical. ACS Sustain. Chem. Eng. 2016, 4, 35–46. [CrossRef] 20. Zhao, S.; Abu-Omar, M.M. Renewable thermoplastics based on lignin-derived polyphenols. Macromolecules 2017, 50, 3573–3581. [CrossRef] 21. Zhou, J.; Zhang, H.; Deng, J.; Wu, Y. High glass-transition temperature acrylate polymers derived from biomasses, syringaldehyde, and vanillin. Macromol. Chem. Phys. 2016, 217, 2402–2408. [CrossRef] 22. Fengel, D.; Wegener, G. Wood (Chemistry, Ultrastructure, Reactions); Walter de Gruyter: Berlin, Germany; New York, NY, USA, 1984. 23. Leopold, B. Studies on lignin. III. Oxidation of wood from Norway spruce with nitrobenzene and alkali. Acta Chem. Scand. 1952, 6, 38–48. [CrossRef] 24. Kurshner, K. The difficulties in the production of vanillin from sulphite liquors. Rus. J. Appl. Chem. 1955, 28, 957–968. 25. Bots, R.H. Process of Manufacturing Vanillin. U.S. Patent 1,643,805, 27 September 1927. 26. Tomlinson, G.H.; Hibbert, H. Studies on lignin and related compounds. XXV. Mechanism of vanillin formation from spruce lignin sulfonic acids in relation to lignin structure. J. Am. Chem. Soc. 1936, 58, 348–353. [CrossRef] 27. Min, D.; Xiang, Z.; Liu, J.; Jameel, H.; Chiang, V.; Jin, Y.; Chang, H. Improved protocol for alkaline nitrobenzene oxidation of woody and non-woody biomass. J. Wood Chem. Technol. 2015, 35, 52–61. [CrossRef] 28. Adaci, S.; Tanimoto, M.; Tanaka, M.; Matsuno, R. Kinetics of the alkaline nitrobenzene oxidation of lignin in rice straw. Chem. Eng. J. 1992, 49, 17–21. [CrossRef] 29. Leopold, B.; Malmstrom, I.L. Studies on lignin. IV. Investigation on nitrobenzene oxidation products of lignin from different woods by paper partition chromatography. Acta Chem. Scand. 1952, 6, 49–54. [CrossRef] 30. Pepper, J.M.; Casselman, B.W.; Karapally, J.C. Lignin oxidation. Preferential use of cupric oxide. Can. J. Chem. 1967, 45, 3009–3012. [CrossRef] 31. Gogotov, A.F.; Makovskaya, T.I.; Babkin, V.A. The effect of various additives on the yield of aromatic aldehydes in the aspen nitrobenzene oxidation. Chem. Sustain. Dev. 1996, 4, 187–192. 32. Simpson, W.G.; Sondhimer, E. Silica acid column chromatography in the alkaline nitrobenzene oxidation of wood. TAPPI 1960, 43, 1025–1026. 33. Ikeda, T.; Holtman, K.; Kadla, J.F.; Chang, H.M.; Jameel, H. Studies on the effect of ball milling on lignin structure using a modified DFRC method. J. Agric. Food Chem. 2002, 50, 129–135. [CrossRef][PubMed] 34. Leopold, B. Studies on lignin. Part VIII. Nitrobenzene oxidation and sulphonation of wood decayed by brown—Rotting fungi. Svensk Kemisk Tidskrift 1951, 63, 260–271. 35. Ikeda, T.; Magara, K. Chemical properties of softwood soda-anthraquinone lignin. J. Wood Chem. Technol. 2015, 35, 167–177. [CrossRef] 36. Moskovtsev, N.G.; Strel’skaya, S.A. Composition and properties of the sediment formed during the conversion of aqueous of wood prehydrolysates. Wood Chem. 1986, 1, 63–67. 37. Jin, Z.; Jin, G.; Shao, S.; Katsumata, K.S. Lignin characteristics of bast fiber and core in kenaf, bark and wood of paper mulberry and mulberry. J. Wood Sci. 2016, 58, 144–152. [CrossRef] 38. Gogotov, A.F. A new hypothesis of the interaction of lignin and an oxidizer in the process of nitrobenzene oxidation. Khimiya Rastitel’nogo Syr’ya 1999, 2, 65–72. 39. Tarabanko, V.E.; Kaigorodov, K.L.; Koropachinskaya, N.V.; Chelbina, Y.V. Preparation of aromatic aldehydes from biobutanol production wastes. Chem. Sustain. 2012, 20, 471–476. 40. Chirkin, G.A.; Tishenko, D.V. Redox-oxidative reactions during alkaline boiling of wood. Rus. J. Appl. Chem. 1962, 35, 153–159. 41. Kleinert, T.N. Effect of temperature on the ratio of the contributions of fast and slowly soluble lignin. TAPPI 1964, 49, 53–57. Int. J. Mol. Sci. 2017, 18, 2421 26 of 29

42. Tarabanko, V.E.; Koropachinskaya, N.V.; Kudryashov, A.V.; Kuznetsov, B.N. Influence of lignin origin on the efficiency of the catalytic oxidation of lignin into vanillin and syringaldehyde. Russ. Chem. Bull. 1995, 2, 367–371. [CrossRef] 43. Tarabanko, V.E.; Fomova, N.A.; Kuznetsov, B.N.; Kudryashev, A.V.; Ivanchenko, N.M. On the mechanism of vanillin formation in catalytic oxidation of lignin with oxygen. React. Kinet. Catal. Lett. 1995, 55, 161–170. [CrossRef] 44. Tarabanko, V.E.; Pervishina, E.P.; Hendogina, Y.V. Kinetics of aspen wood oxidation by oxygen in alkaline media. React. Kinet. Catal. Lett. 2001, 72, 153–162. [CrossRef] 45. Kamaldina, O.D.; Massov, Y.A. Preparation of Vanillin from Lignosulfonates; TsBTI TsINIS: Moscow, Russia, 1959. 46. Epshtein, R.B. Preparation of vanillin from wood. Tr. Ukrainian NII Pischevoy Promishlennosty 1959, 2, 201–213. 47. Villar, J.C.; Caperos, A.; Garcia Ochoa, F. Oxidation of hardwood kraft-lignin to phenolic derivatives with oxygen as oxidant. J. Wood Chem. Technol. 2001, 35, 245–255. [CrossRef] 48. Wu, G.; Heits, M.; Chornet, E. Improved alkaline oxidation process for the production of aldehydes from steam-explosion lignin. Ind. Eng. Chem. Res. 1994, 33, 718–723. [CrossRef] 49. Mathias, A.L.; Rodrigues, A.E. Production of vanillin by oxidation of pine kraft-lignins with oxygen. Holzforschung 1995, 49, 273–278. [CrossRef] 50. Tarabanko, V.E.; Gulbis, G.R.; Ivanchenko, N.M.; Kuznetsov, B.N.; Kudryashev, A.V. Study of wood and lignosulfonates processing into fine products. Chem. Sustain. Dev. 1996, 4, 405–417. 51. Koropachinskaya, N.V.; Tarabanko, V.E.; Chernyak, M.Y. Catalytic oxidation of birch wood (Betula Pendula Roth.) with oxygen in syringaldehyde and vanillin. Khimiya Rastitel’nogo Syr’ya 2003, 2, 9–14. 52. Santos, S.G.; Marcues, A.P.; Lima, D.L.D.; Evtuguin, D.V.; Esteves, V.I. Kinetics of Eucaliptus lignosulfonate oxidation to aromatic aldehydes by oxygen in alkaline media. Ind. Eng. Chem. Res. 2011, 50, 291–298. [CrossRef] 53. Villar, J.C.; Caperos, A.; Garcia Ochoa, F. Oxidation of hardwood kraft-lignin to phenolic derivatives. Nitrobenzene and copper oxide as oxidation. J. Wood Chem. Technol. 1997, 17, 259–285. 54. Pinto, P.C.R.; Borges da Silva, E.A.; Rodrigues, A.E. Insights into oxidative conversion of lignin to high-added-value phenolic aldehydes. Ind. Eng. Chem. Res. 2011, 50, 741–748. [CrossRef] 55. Xiang, Q.; Lee, Y.Y. Production of oxychemicals from precipitated hardwood lignin. Appl. Biochem. Biotechnol. 2001, 91–93, 71–80. [CrossRef] 56. Zhang, J.H.; Deng, H.B.; Lin, L. Wet aerobic oxidation of lignin into aromatic aldehydes catalysed by a

perovskite-type oxide: LaFe1−xCuxO3 (x = 0, 0.1, 0.2). Molecules 2009, 14, 2747–2757. [CrossRef][PubMed] 57. Sales, F.G.; Maranhao, L.C.A.; Filho, N.M.L.; Abreu, C.A.M. Kinetic evaluation and modeling of lignin catalytic wet oxidation to selective production of aromatic aldehydes. Ind. Eng. Chem. Res. 2006, 45, 6627–6631. [CrossRef] 58. Goncalves, A.R.; Schuchardt, U. Oxidation of organosolv Lignins in Acetic Acid. Inﬂuence of oxygen pressure. Appl. Biochem. Biotechnol. 1999, 77, 127–132. [CrossRef] 59. Werhan, H.; Mir, J.M.; Voitl, T.; von Rohr, P.R. Acidic oxidation of kraft lignin into aromatic monomers catalyzed by transition metal salts. Holzforschung 2011, 65, 703–709. [CrossRef] 60. Werhan, H.; Assmann, N.; von Rohr, P.R. Lignin oxidation studies in a continuous two-phase ﬂow microreactor. Chem. Eng. Process. 2013, 73, 29–37. [CrossRef] 61. Deng, W.P.; Zhang, H.X.; Wu, X.J.; Li, R.S.; Zhang, Q.H.; Wang, Y. Oxidative conversion of lignin and

lignin model compounds catalyzed by CeO2-supported Pd nanoparticles. Green Chem. 2015, 17, 5009–5018. [CrossRef] 62. Partenheimer, W. The Aerobic oxidative cleavage of lignin to produce hydroxyaromatic benzaldehydes and carboxylic acids via metal/bromide catalysts in acetic acid/water mixtures. Adv. Synth. Catal. 2009, 351, 456–466. [CrossRef] 63. Das, L.; Kolar, P.; Osborne, J.A.; Sharma-Shivappa, R.R.; Classen, J.J. Selective oxidation of lignin into aromatic aldehydes using niobium oxalate. Trans. ASABE 2016, 59, 727–735. 64. Schmitt, D.; Regenbrecht, C.; Hartmer, M.; Stecker, F.; Waldvogel, S.R. Highly selective generation of vanillin by anodic degradation of lignin: A combined approach of electrochemistry and product isolation by adsorption. Beilstein J. Org. Chem. 2015, 11, 473–480. [CrossRef][PubMed] 65. Gong, J.Y.; Imbault, A.; Farnood, R. The promoting role of bismuth for the enhanced photocatalytic oxidation

of lignin on Pt-TiO2 under solar light illumination. Appl. Catal. B 2017, 204, 296–303. [CrossRef] Int. J. Mol. Sci. 2017, 18, 2421 27 of 29

66. Chatel, G.; Rogers, R.D. Review: Oxidation of lignin using ionic liquids-an innovative strategy to produce renewable chemicals. ACS Sustain. Chem. Eng. 2014, 2, 322–339. [CrossRef] 67. Tarabanko, V.E.; Pervishina, E.P.; Nevkrytova, T.A.; Pen, R.Z. Investigation of the oxidation kinetics of aspen wood by oxygen in an alkaline medium. Khimiya Rastitel’nogo Syr’ya 1999, 4, 53–59. 68. Tarabanko, V.E.; Pervishina, E.P.; Tarabanko, N.V.; Chernyak, M.Y.; Kaygorodov, K.L.; Chelbina, Y.V.; Boyarchuk, D.V. Kinetics of fir wood oxidation by oxygen in aqueous alkaline media. Khimiya Rastitel’nogo Syr’ya 2016, 4, 57–63. 69. Denisov, E.T. The Kinetics of Homogeneous Chemical Reactions; Vysshaya Shkola: Moskow, Russia, 1988; p. 391. 70. Tarabanko, V.E.; Koropachinskaya, N.V.; Kydryashov, A.V.; Kyznetsov, B.N.; Polyakov, S.V.; Zolotukhin, V.N. Investigation of the process of processing wheat straw into aromatic aldehydes and levulinic acid. Khimiya Rastitel’nogo Syr’ya 1998, 3, 59–64. 71. Schoefel, E.W. Manufacture of Vanillin. U.S. Patent 2,598,311, 27 May 1952. 72. Tarabanko, V.E.; Pervishina, E.P.; Kuznetsov, B.N. Method of Processing of Small-Leaved Wood. Russian Patent 2,178,405, 20 January 2002. 73. Kaigorodov, K.L.; Tarabanko, V.E.; Chernyak, M.Y.; Chelbina, Y.V.; Tarabanko, N.V.; Smirnova, M.A. Kinetics of low-temperature oxidation of enzime pine lignin (Pinus silvestris) in water-alkaline. Khimiya Rastitel’nogo Syr’ya 2017, 3, 63–70. 74. Nonhebel, D.C.; Walton, J.C. Free-Radical Chemistry: Structure and Mechanism; Cambridge University Press: New York, NY, USA, 1974. 75. Denisov, E.T. The Rate Constants of Homolytic Liquid-Phase Reactions; Khimia: Moskow, Russia, 1971; p. 712. 76. Kleinert, T.N. Mechanisms of Alkaline Delignification: II. Free Radical Reactions. TAPPI 1966, 49, 126–130. 77. Fargues, C.; Mathias, A.; Rodrigues, A. Kinetics of vanillin production from kraft lignin oxidation. Ind. Eng. Chem. Res. 1996, 35, 28–36. [CrossRef] 78. Wong, Z.J.; Chen, K.F.; Li, J. Formation of vanillin and syringaldehyde in an oxygen delignification process. BioResources 2010, 5, 1509–1516. 79. Hermer, E.I. Intensification of oxygen-alkaline delignification of lignocellulosic materials by the O-phenanthroline. Wood Chem. 1992, 4–5, 46–55. 80. Denisov, E.T.; Mitzkevich, N.I.; Agabekov, V.E. The Mechanisms of Liquid-Phase Oxidation of Oxygen-Containing Compounds; Nauka i Tekhnika: Minsk, Belarus, 1975; p. 240. 81. Wojnarovits, L.; Foldiak, G.; D’Angelantonio, M.; Emmi, S.S. Mechanism of OH radical-induced oxidation of p-cresol to p-methylphenoxyl radical. Res. Chem. Intermed. 2002, 28, 373–386. [CrossRef] 82. Maa, Q.; Lianga, C.; Chenb, K.; Liuc, K.; Maod, J.; Chena, Z.; Li, H. Role of alkali in catalytic oxidation of p-cresols. J. Mol. Catal. A 2016, 420, 45–49. [CrossRef] 83. Sheldon, R.A.; de Heij, N. Aromatic aldehydes via catalytic oxidation. In Role of Oxygen in Chemistry and Biochemistry; Ando, W., Moro-Oka, Y., Eds.; Elsevier: Amsterdam, The Netherlands, 1988; pp. 243–256. 84. Barton, B.; Logie, G.C.; Schoonees, B.M.; Zeelie, B. Practical process for the air oxidation of cresols: Part A. mechanistic investigations. Org. Proc. Res. Dev. 2005, 9, 62–69. [CrossRef] 85. Shevchenko, S.M.; Apushkinsky, A.G. Quinone methides in the chemistry of wood. Russ. Chem. Rev. 1992, 61, 105–131. [CrossRef] 86. Tarabanko, V.E.; Petukhov, D.V.; Selutin, G.E. New Mechanism for the Catalytic Oxidation of Lignin to Vanillin. Kinet. Catal. 2004, 45, 569–577. [CrossRef] 87. Tarabanko, V.E.; Il’ina, I.I.; Petukhov, D.V.; Pervishina, E.P. On the mechanism of oxidative cleavage of the carbon-carbon bond of lignins at the alkaline medium. Khimiya Rastitel’nogo Syr’ya 1997, 3, 51–58. 88. Tarabanko, V.E.; Hendogina, Y.V.; Petukhov, D.V.; Pervishina, E.P. On the role of retroaldol reaction in the process of lignin oxidation into vanillin. Kinetics of the vanillylidene acetone cleavage in the alkaline media. React. Kinet. Catal. Lett. 2000, 69, 361–368. [CrossRef] 89. Guthrie, J.P.; Cooper, K.J.; Cossar, J.; Dawson, B.A.; Taylor, K.F. The retroaldol reaction of cinnamaldehyde. Can. J. Chem. 1984, 62, 1441–1445. [CrossRef] 90. Palm, V.A. (Ed.) Tables of Rate and Equilibrium Constants of Heterolytic Organic Reactions; VINITI: Moskow, Russia, 1975; Volume 1, p. 82. 91. Ermakova, M.I.; Kirushina, M.F.; Zarubin, M.Y. Comparison of the OH-acidity of lignin-related phenols in water, alcohols and aqueous-alcoholic media. Wood Chem. 1984, 5, 23–29. 92. Zakis, G.F. Functional Analysis of Lignins and Their Derivatives; Zinatne: Riga, Latvia, 1987; p. 230. Int. J. Mol. Sci. 2017, 18, 2421 28 of 29

93. Schultz, T.P.; Templeton, M.C. Proposed mechanism for nitrobenzene oxidation of lignin. Holzforschung 1986, 40, 93–97. [CrossRef] 94. Schultz, T.P.; Hubbard, T.V.; Fisher, T.H. Oxidation of β-O-4 lignin models: Reaction difference between Ce(IV) ammonium nitrate and lignin peroxidase. Holzforschung 1995, 49, 528–532. [CrossRef] 95. Gierer, J. The chemistry of delignification. A general concept. Part II. Holzforschung 1982, 36, 55–64. [CrossRef] 96. Gierer, J.; Nilvebrant, N.O. Studies on the degradation by oxygen in aciditic media. Holzforschung 1994, 48, 51–58. [CrossRef] 97. Nilvebrant, N.O. Stilbenes and Their Role in Bleaching. Synthesis and Behaviour in Autoxidative Reactions. Ph.D. Thesis, The Royal Institute of Technology, Stokholm, Sweden, 1993. 98. Tarabanko, V.E.; Petukhov, D.V. Investigation of the mechanism and improvement of the process of oxidative cleavage of lignins into aromatic aldehydes. Chem. Sustain. Dev. 2003, 11, 645–657. 99. Gierer, J.; Nelson, C.R. Mechanisms of aryl group migration in the formation of stilbens from 1.1-bis (p-hydroxyaryl)-ethane 2-Oaryl ethers. J. Org. Chem. 1978, 43, 4028–4032. [CrossRef] 100. Gierer, J.; Imsgard, F. The reactions of lignins with oxygen and hydrogen peroxide in alkaline media. Svensk Paperstidn 1977, 80, 510–518. 101. Wong, D.F.; Leary, G.; Arct, G. The reaction of lignin model stilbenes with alkali & oxygen & its relevance to mechanical pulp bleaching. In Proceedings of the 8th International Symposium on Wood and Pulping Chemistry, Helsinki, Finland, 6–9 June 1995; Volume 1, pp. 209–216. 102. Tarabanko, V.E.; Petukhov, D.V.; Avramov, P.V. Study of the acidity of radical intermediates of lignin oxidation by quantum chemical methods. Khimiya Rastitel’nogo Syr’ya 1998, 3, 101–109. 103. Koropachinskaya, N.V.; Tarabanko, V.E.; Levdanskyi, V.A. Catalytic oxidation of birch bark with oxygen into syringaldehyde and vanillin. Khimiya Rastitel’nogo Syr’ya 2004, 1, 27–30. 104. Khudyakov, I.V.; Levin, P.P.; Kuz’min, V.A. Reversible Recombination of Radicals. Russ. Chem. Rev. 1980, 49, 982–1002. [CrossRef] 105. Caldwell, E.S.; Steelink, C. Phenoxy radical intermediates in the enzimatic degradation of lignin model compounds. Biochim. Biophys. Acta 1969, 184, 420–431. [CrossRef] 106. Hapiot, P.; Pinson, J. Oxidative dimerization of phenolic aldehydes related to lignin formation. J. Phys. Chem. 1994, 98, 2641–2645. [CrossRef] 107. Ragnar, M.; Lindgren, C.T.; Nilvebrant, N.O. pKa-values of guaiacyl and syringyl phenols related to lignins. J. Wood Chem. Technol. 2000, 20, 277–305. [CrossRef] 108. Dobos, D. Electrochemical Data. A Handbook for Electrochemists in Industry and Universities; Akademiai Kiado: Budapest, Hungary, 1978; p. 320. 109. Shibaeva, L.V.; Metelitsa, D.I.; Denisoiv, E.T. Oxidation of phenol by molecular oxygen in aqueous solutions. 1. Kinetics of oxidation of phenol with oxygen. Kinet. Catal. 1969, 10, 1020–1025. 110. Tarabanko, V.E.; Kudryashev, A.V.; Kuznetsov, B.N.; Gulbis, G.R.; Koropachinskaya, N.V.; Ivanchenko, N.M. Investigation of the process of catalytic oxidation of lignosulfonates to vanillin in a flow reactor. J. Appl. Chem. 1996, 69, 620–624. 111. Ramachadran, P.A.; Chaudhari, R.V. Three-Phase Catalytic Reactors; Nauka: Novosibirsk, Russia, 1992; Volume 2, p. 396. 112. Ioffe, I.I.; Reshetov, V.A.; Dobrotvorsyi, A.M. Heteroheneous Catalysis; Khimia: Leningrad, Russia, 1985; p. 224. 113. Dardelet, S.; Froment, P.; Lacoste, N.; Robert, A. Vanilline et aldehyde syringique. Stabilite a l’oxydation en milieu alcalin par l’oxygene. Rev. ATIP 1985, 39, 369–376. 114. ITS 1–2015. Manufacture of Pulp, Paper and Board; Information and Technical Reference Book; Rosstandart, NDT Bureau: Moscow, Russia, 2016; p. 216. 115. Bjørsvik, H.-R. Fine Chemicals from Lignosulfonates. 1. Synthesis of Vanillin by Oxidation of Lignosulfonates. Org. Process Res. Dev. 1999, 3, 330–340. [CrossRef] 116. Kalunyants, K.A.; Shanenko, E.F.; Zai’tseva, L.V. Modern methods of enzymatic hydrolysis of cellulosic materials. In Itogi Nauki e Tehkniki. Kmimia e Tehknologiya Pish’evyh Productov; VINITI: Moskow, Russia, 1988; Volume 1, p. 187. 117. Ripachek, V. Biology of Wood-Destroing Fungi; Lesnaya Promyshlennost: Moskow, Russia, 1976; p. 245. 118. Garcia, V.; Pakkila, J.; Ojamo, H.; Muurinen, E.; Keiski, R.L. Challenges in biobutanol production: How to improve the efficiency? Renew. Sustain. Energy Rev. 2011, 15, 964–980. [CrossRef] Int. J. Mol. Sci. 2017, 18, 2421 29 of 29

119. Tarabanko, V.E.; Pervishina, E.P.; Kuznetsov, B.N. Method of Wood Processing into Products of Fine Organic Synthesis. Russian Patent 2,119,427, 27 September 1998. 120. Tarabanko, V.E.; Kozlov, I.A.; Pervishina, E.P.; Chernyak, M.Y.; Kuznetsov, B.N. Chemical Wood Processing Method. Russian Patent 2,158,192, 27 October 2000. 121. Tarabanko, V.E.; Kaygorodov, K.L.; Skiba, E.A.; Tarabanko, N.; Chelbina, Y.V.; Baybakova, O.V.; Kuznetsov, B.N.; Djakovitch, L. Processing pine wood into vanillin and glucose by sequential catalytic oxidation and enzymatic hydrolysis. J. Wood Chem. Technol. 2017, 37, 43–51. [CrossRef] 122. Tarabanko, V.E.; Shambazov, V.K.; Kuznetsova, S.A.; Kuznetsov, B.N. Method of Hardwood Processing into Valuable Organic Products. Russian Patent 2,219,048, 20 December 2003. 123. Sharkov, V.I.; Kyibina, N.I. Chemistry of Hemicelluloses; Lesnaya Promyshlennost: Moskow, Russia, 1972; p. 440. 124. Tarabanko, V.E.; Kudryashov, A.V.; Pervishina, E.P.; Kuznetsov, B.N.; Petukhov, D.V.; Hendogina, Y.V. New methods of separation of vanillin and syringcaldehyde. Khimiya Rastitel’nogo Syr’ya 1998, 3, 95–100. 125. Tarabanko, V.E.; Kudryashov, A.V.; Gulbis, G.R.; Kuznetsov, B.N. Method of Separation of Vanillin and Syringcaldehyde. Russian Patent 2,059,600, 10 May 1996. 126. Tarabanko, V.E.; Chelbina, Y.V.; Sokolenko, V.A.; Tarabanko, N.V. The study of vanillin extraction by octylamine. Solvent Extr. Ion Exch. 2007, 25, 99–107. [CrossRef] 127. Tarabanko, V.E.; Chelbina, Y.V.; Kudryashev, A.V.; Tarabanko, N.V. Separation of vanillin and syringaldehyde produced from lignins. Sep. Sci. Technol. 2013, 48, 127–132. [CrossRef] 128. Mota, M.I.F.; Rodrigue, P.P.C.; Loureiro, J.M.; Rodrigues, A.E. Recovery of vanillin and syringaldehyde from lignin oxidation: A review of separation and purification processes. Sep. Purif. Rev. 2016, 45, 227–259. [CrossRef] 129. Nepenin, Y.A. Cellulose Technology; Lesnaya Promyshlennost: Moskow, Russia, 1990; Volume 2, p. 597. 130. Chudakov, M.I. Industrial Use of Lignin; Lesnaya Promyshlennost: Moskow, Russia, 1983; p. 42. 131. Study into the Establishment of An Aroma and Fragrance Fine Chemicals Value Chain in South Africa (Tender Number T79/07/03). Available online: https://www.thedti.gov.za/industrial_development/docs/ fridge/Aroma_Part3.pdf (accessed on 15 October 2017). 132. Corma, A.; Iborra, S.; Velty, A. Chemical routes for the transformation of biomass into chemicals. Chem. Rev. 2007, 107, 2411–2502. [CrossRef][PubMed] 133. Pileidis, F.D.; Titirici, M.M.; Filoklis, D. Levulinic acid biorefineries: New challenges for efficient utilization of biomass. ChemSusChem 2016, 9, 562–582. [CrossRef][PubMed]

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).