Radar Tool for Lignin Classification on the ... - ACS Publications

Jul 17, 2015 - ABSTRACT: A tool for lignin classification was developed based on radar plots using six descriptors identified as key characteristics f...
0 downloads 0 Views 6MB Size
Article pubs.acs.org/IECR

Radar Tool for Lignin Classification on the Perspective of Its Valorization Carina A. Esteves Costa, Paula Cristina Rodrigues Pinto,* and Alírio Egídio Rodrigues Laboratory of Separation and Reaction Engineering - LSRE, Associate Laboratory LSRE/LCM, Department of Chemical Engineering, Faculty of Engineering, University of Porto, Rua Dr. Roberto Frias s/n, 4200-465 Porto, Portugal S Supporting Information *

ABSTRACT: A tool for lignin classification was developed based on radar plots using six descriptors identified as key characteristics for vanillin (V) and syringaldehyde (Sy) production by oxidation in alkaline medium: content on β-O-4 structures, noncondensed structures, syringyl and guaiacyl units, and yield of Sy and V by nitrobenzene oxidation (NO). A set of lignins was classified according to the radar information, simplifying the evaluation and discussion of the impact of the delignification process, wood species, and morphologic part on lignin. Lignin from tobacco stalks was one of the targets, reporting 13C NMR and NO characterization data to ascertain the influence of delignification process. Structural data on lignins from different hardwoods (eucalyptus, mimosa, and willow), several parts of the same species (bole, bark, branches, sawdust), and different delignification processes were also used as a basis for the developed methodology. The radar plots of tobacco lignins allow classifying the lignin produced by organosolv process with ethanol as that with the higher aptitude for V and Sy production with O2. This classification was confirmed by batch oxidation of this lignin as compared with that produced by organosolv process with butanol. In the same way, among the processed hardwood lignins, the one produced by organosolv of eucalyptus bole wood showed the highest intensity in all descriptors, being classified as a privileged source of Sy in comparison to Kraft lignins. The reasons behind the differences on descriptors that gave rise to lignins classification are discussed. The radar classification can be used as a predictive tool for product and process design, for both lignin production and application. The requisite for this is the previous knowledge of the relevant structural parameters. This is a key step to demystify the lignin complexity in key descriptors and consolidation of valorization routes in flexible processing units.

1. INTRODUCTION As a byproduct of the pulp and paper industry and biorefineries, lignin is most often used as fuel for the recovery of its energy content. However, due to the large generated quantities and the successful prospective scientific and preindustrial work on valorization, the interest in lignin has increased in the last years. The aromatic structure of lignin makes it an attractive source of value-added compounds such as vanillin (V) and syringaldehyde (Sy),1 although many other applications are also valuable and already established.2,3 Aromatic aldehydes are produced by oxidative depolymerization in alkaline4,5 or acidic6,7 medium, and different ranges of production yields have been reported (V and Sy yields between 3.1% and 14.2% w/wlignin).4−6 V and Sy yields depend first on lignin type: guaiacyl lignins (G-type, typical of softwoods) yield V under oxidative depolymerization, whereas p-hydroxyphenyl/guaiacyl/syringyl lignins (H:G:Stype lignins, typical of annual plants and hardwoods) are able to produce Sy and p-hydroxybenzaldehyde (Hy) in addition to V. Usually, annual plants contain a higher frequency of H units than hardwoods. Therefore, one should expect lower V and Sy in these lignins compared with lignins composed exclusively by G and S units. On the other hand, among hardwoods, there is a high variety of proportions between G and S units as detailed in the literature.8 For that reason, the first factor to take into account is the lignin nature, an unalterable factor. However, there are other implications of the H:G:S ratio. The most relevant is the higher likelihood for condensation (lignin moieties linked via one C position of the aromatic ring)2 for H© XXXX American Chemical Society

and G-rich lignins, in both in situ lignin (in the lignification of the plant) and in the delignification process. Moreover, lower reactivity of G-lignins usually implies harder conditions for delignification which promotes, by this way, reactions leading to additional condensation. Lower reactivity of G units in lignins could also imply lower yields on oxidation (in V) than for S units (in Sy). In short, H:G:S ratio could also impart some effect on frequency of C−C linkages and on lignin reactivity, which can be designated as a secondary effect. However, degree of condensation also depends on plant genetics and environmental factors, whatever H:G:S is. Previous studies have indicated that the frequency of alkyl−aryl ether linkages, in particular β-O-4, is a determinant factor for the application of lignin as a source of V and Sy.9,10 This is the most frequent linkage in lignin, and a high yield is usually related with the preservation of its native structure, but it also differs among species. Delignification processes promote the cleavage of these linkages, leading to modifications of the lateral chain of the phenylpropane unit (ppu), most of them unfavorable to the production of V and Sy by oxidation. In lignin characterization, several complementary methods are usually applied in order to determine the major structural features. Extensive data are generated; however, it is difficult to Received: May 19, 2015 Revised: July 17, 2015 Accepted: July 17, 2015

A

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

(LAEgbark), and branches (LAEgbranch), was performed with dioxane/water containing 2 M HCl, following the procedure and conditions as described before.9,11 2.2. Inorganics and Carbohydrate Contents. The inorganic content was determined by gravimetric quantification after incineration of lignin samples at 600 °C during 6 h.5 Determination of the carbohydrate content in lignins was performed following the method previously described,9 briefly: about 15 mg of lignin was suspended in 2 mL of 2 M HCl methanolic solution and submitted to acid methanolysis at 100 °C for 4 h. After cooling, pyridine and an internal standard solution of sorbitol were added and the mixture was carefully evaporated under reduced pressure. Then, the dried methanolysates were converted to trimethylsilylated derivatives. The products were identified by GC−MS and quantified by GCFID as previously described.9,12 The content of inorganics and carbohydrates of tobacco lignins is depicted in Table S2 (Supporting Information) as component weight per 100 g of lignin. The maximum amount of contaminants measured (inorganics and carbohydrates) in lignins from tobacco stalks is 5.4% for LTobObut, which is in agreement with the degree of contamination found for other lignins from herbaceous plants.13,14 Compared to wood species (softwood and hardwood), lignins from herbaceous plants usually contain higher levels of carbohydrates, due to the presence of lignin−carbohydrate complexes.13 Additionally, compared to values found in our previous works on acidolysis of lignins9 and pulping liquor lignins from hardwoods,10,12 the overall results are similar. 2.3. Nitrobenzene Oxidation (NO). Lignins were submitted to alkaline NO as already described in the literature for reaction and products analysis.15 Briefly, 30 mg of a lignin sample was dissolved in NaOH aqueous solution and, after adding nitrobenzene, heated to 170 °C for 4 h. After a first extraction with chloroform, the aqueous phase was acidified with H2SO4, and then extracted again with chloroform. The organic phase was evaporated under reduced pressure, redissolved in methanol, and made up to 10.00 mL with methanol. The products were analyzed in a Shimadzu HPLC (Prominence model) equipped with a UV−vis detector SPDM20A (operating at 280 nm wavelength), using the column, eluents, and gradient already described.15 2.4. 13C NMR Analysis. Quantitative 13C NMR spectra were recorded using a Bruker AVANCE III 400 spectrometer operating at 400 MHz, with a temperature of 45 °C over 72 h. About 170 mg of dried lignin was dissolved in 0.5 mL of deuterated dimethyl sulfoxide (DMSO-d6), and quantitative conditions for 13C NMR measurements were used: simple 1D pulse sequence, relaxation delay of 12 s, 1400 scans, and 1D sequence with power gated coupling using 90° flip angle. 2.5. Alkaline Oxidation with O2. Oxidations with O2 in alkaline medium of lignins LTobObut and LTobOethan were performed in a Büchi AG laboratory autoclave with a capacity of 1 L (model BEP280 type II, Switzerland). Detailed information about the equipment and the reaction conditions were described in previous publications.5,12 Briefly, a solution of 60 g/L of lignin is prepared in 2 M NaOH, introduced into the reactor, heated to 120 °C, and pressurized. The reaction begins with the admission of O2. Samples from the reaction mixture were collected at regular time intervals. The analysis of the products was performed by the procedure and equipment/ conditions as previously published in the literature.5,15

identify the individual linkages due to the lignin complexity, with multiples overlapping, making it difficult to extract focused and “ready to use” information. The evaluation of a lignin relative to its suitability as a source of these compounds should combine the assessment of key characteristics such as H:G:S ratio, condensation, and β-O-4 content, and allow a qualitative prediction of the yield expected by oxidative depolymerization under the same range of conditions. The need of defining structural parameters for the assessment of lignin potential for V and Sy production was previously identified by the authors.5 The comparison of nitrobenzene oxidation (NO) yields with those from oxidation with O2 of different lignins led to the conclusion that the differences found for the results of oxidation with O2 are not properly described by NO.5 This work aims to establish a classification technique for lignins based on the above-mentioned major characteristics, as a tool for lignins selection in view of V and/or Sy production. These characteristics are the descriptors used to build the radar plot for each lignin, reducing the unavoidable complexity of lignin structure to its key aspects, while maintaining the scientific basis of the data sets with quantitative information. In this work, lignins obtained from an annual plant (Nicotiana tabacum) and lignins from different hardwoods (Eucalyptus globulus, Acacia dealbata, and Salix spp.) and from different parts of the same species (E. globulus), produced by different delignification processes, provide the information sources used to build the lignin radar plots. These lignins were analyzed by chemical and spectroscopic methods in order to have a pool of information for this goal. Also, in this work, tobacco stalks lignin resulting from different delignification processes are published for the first time, for evaluation of the impact of these processes on lignin structure.

2. MATERIAL AND METHODS 2.1. Lignins: Species and Processes. E. globulus (eucalyptus bole, bark, branches, and sawdust), A. dealbata (mimosa bole), Salix spp. (willow bole), and N. tabacum L. (tobacco stalks) were the sources of lignins for this study. Lignins from different delignification processes were considered (Table S1, Supporting Information) and are described as follows: lignin from industrial Kraft liquor of eucalyptus bole (LKEgbole), lignin from laboratory Kraft pulping liquor of bark (LKEgbark), branches (LKEgbranch), and sawdust (LKEgsawd); lignins isolated from laboratory Kraft pulping liquor of mimosa (LKmimosa) and willow (LKwillow).10 The general Kraft pulping conditions were as follows: active alkali 22−26%, temperature 160−170 °C, and total pulping time 180−210 min. LOEgbole was produced from eucalyptus bole by ethanol organosolv process.9 This lignin was prepared using an autocatalyzed ethanol/water process under Lignol’s proprietary conditions. Typical processing conditions use temperatures in the range of 180−200 °C and a residence time of 1−3 h. From tobacco stalks, four lignins were considered: lignin produced by organosolv process using butanol (LTobObut) and ethanol (LTobOethan), lignin produced by a steam explosion process (LTobSE), and another one isolated by mild acidolysis (LTobA). LTobObut was produced by pulping with butanol− water 50/50 v/v, at 178 °C, 30−60 min (with H2SO4); LTobOethan was produced in standard autocatalyzed conditions at 195 °C for 90 min; LTobSE was produced at 205 °C, 10 min residence time, with 15% caustic loading. The mild acidolysis process to obtain the fourth lignin from tobacco stalks LTobA, as well as lignin from eucalyptus bole (LAEgbole), bark B

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research Table 1. Yields of Monomeric Phenolic Products and S:G:H Ratio Obtained by NO of Lignins from Tobacco Stalks products, % w/wlignina lignin LTobObut LTobOethan LTobSE LTobA a

Hy 0.42 0.33 0.37 0.35

± ± ± ±

VA 0.01 0.01 0.01 0.01

0.88 0.63 0.77 0.99

± ± ± ±

SA 0.03 0.02 0.02 0.03

0.79 0.47 0.26 0.64

± ± ± ±

V 0.02 0.02 0.01 0.02

2.8 7.2 5.2 13.3

± ± ± ±

Sy 0.1 0.4 0.3 0.8

2.5 4.8 3.2 8.3

± ± ± ±

0.1 0.2 0.1 0.2

total yield

S:G:H

± ± ± ±

44:49:07 39:58:03 36:59:04 38:61:02

7.4 13.5 9.8 23.2

0.2 0.4 0.3 0.9

Reported to nonvolatile solids weight after deducting ashes and carbohydrates.

Table 2. Assignments and Quantification (Number/C6) of the Structures/Linkages and Functional Groups Identified by 13C NMR for Tobacco Lignins amount (number/C6) assignments (spectroscopic range) Cβ in β-5 and β−β structures (δ 51.0−53.8 ppm) aromatic OCH3 (δ 54.3−57.3 ppm) Cγ in β-O-4 structures without CαO (δ 59.3−60.8 ppm) Cγ in β-5 and β-O-4 structures with CαO; Cγ in β-1 (δ 62.5−63.8 ppm) Cα in β-O-4 structures; Cγ in pinoresinol/syringaresinol and β−β structures (δ 70.0−76.0 ppm) Cβ in β-O-4 structures; Cα in β-5 and β−β structures (δ 80.0−90.0 ppm) aromatic CAr−H (δ 103.0−123.0 ppm) aromatic CAr−C (δ 123.0−137.0 ppm) C4 in H units (δ 157.0−162.0 ppm) CHO in benzaldehyde structures (δ 191.0−192.0 ppm) CHO in cinnamaldehyde structures (δ 193.5−194.5 ppm) CO in aldehydes and ketones (δ 195.0−210.0 ppm)

3. RESULTS AND DISCUSSION

LTobObut

LTobOethan

LTobSE

LTobA

0.29 0.91 0.31 0.20 0.53 0.45 1.93 1.67 0.16 0.05 0.05 0.94

0.19 0.97 0.26 0.11 0.44 0.51 2.07 1.74 0.13 0.06 0.07 1.08

0.23 0.84 0.19 0.11 0.53 0.65 1.99 1.82 0.18 0.06 0.07 1.21

0.23 1.28 0.43 0.19 0.79 0.90 2.07 1.39 0.17 0.05 0.05 0.70

in particular, in those resulting from extensive delignification processes due to the cleavage of the ester linkages. β-Aryl linkages of lignin, denoted as β-O-4, are the main basis of the lignin network. The frequency of these structures in native lignin can reach to near 0.8 per ppu.9 For tobacco stalks lignins, no information about the content of this structure has been reported in the literature. In the delignification process resulting from organosolv and steam explosion, this ether linkage is cleaved, liberating phenolic and aliphatic OH groups, improving the lignin solubility in the medium. The extent of the cleavage depends on the process and conditions. The net content of β-O-4 was calculated following a reliable estimation method published in literature:9,22,25 subtraction of the amount of β-5 and β−β structures (δ 51.0−53.8 ppm) from the integral of δ 80.0−90.0 ppm. On the basis of this estimation, the values for β-O-4 structures are 0.16, 0.32, 0.42, and 0.67 for LTobObut, LTobOethan, LTobSE, and LTobA, respectively (results summarized in Table 3). The considerably lower amount of β-O-4 structures in lignins resulting from organosolv and steam explosion processes compared to lignin from mild acidolysis is the result of the extensive depolymerization in the formers. The β-O-4 cleavage was also reported in the literature26,27 as result of other processes, such as Kraft delignification of perennial

3.1. Structural Data of Lignins from Tobacco Stalks: Evaluation of Processing Impact. 3.1.1. Analysis by NO. V, Sy, Hy, vanillic acid (VA), and syringic acid (SA) are the main monomeric phenolics that result from the oxidation of the noncondensed fraction of lignin by NO. The quantification is reported in Table 1. V and Sy are the main products, accounting for 72−93% of the total phenolic monomers identified. Minor contents of their corresponding aromatic acids (SA and VA) and Hy were also obtained. The NO yield of lignins follows the sequence LTobA > LTobOethan > LTobSE > LTobObut, indicating that the delignification process, organosolv (in particular with butanol) and steam explosion, have induced an increase in the number of condensed linkages. NO has been one of the approaches for the evaluation of the potential for lignin to produce V and Sy by oxidative depolymerization. In the literature it is suggested that the yield from oxidation with O2 can reach 50% of the NO yield,16 while in other studies with diverse lignins lower relative yields (20−37%) were obtained.5,12 3.1.2. Analysis by 13C NMR. Quantitative 13C NMR has been widely used for the evaluation of the main structural features of hardwoods,11,17 softwoods,18,19 and also annual plant20,21 lignins. In this work, structures, linkages, and functional groups of tobacco lignins were assessed by 13C NMR based on acquisition conditions and assignments referred in the literature.9,18,22 The assignments and the corresponding chemical shifts, as well as the resulting content (number/C6) are presented in Table 2. In addition to H unit, p-coumaric and ferulic esters can be found in lignins of annual plants. However, these phenolic moieties should not represent a noteworthy interference on 13C NMR spectra23 due to the usually low percentage in lignins,24

Table 3. Main Structural Characteristics of the Tobacco Lignins Assessed by 13C NMR parameter β-O-4 structures (number/100C6) NCS (%)a S:G:H

LTobObut

LTobSE

LTobA

16

LTobOethan 32

42

67

45 38:48:18

58 38:49:13

49 31:50:19

61 37:46:17

a

NCS, frequency of noncondensed structures; see ref 9 for calculation of degree of condensation (DC), NCS = 100 − DC.

C

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

Figure 1. Radar classification for tobacco lignins produced by different processes.

grasses.10 Organosolv process with butanol of tobacco stalks has led to lignin with the lowest content of β-O-4 structures (Table 3). Noncondensed structures (NCS) are lignin moieties not involved in CAr−C and CAr−O linkages. This is an important structural feature of lignins resulting from the processing of lignocellulosic materials. Native lignins naturally contain condensed structures; however, in the course of delignification, new condensed structures are produced,25,28,29 decreasing the NSC content. This decrease depends on the impact of delignification process on a particular lignin. For tobacco stalks lignin, the lowest value of NCS was found for LTobObut (Table 3), followed by LTobSE, and finally LTobOethan, which presented a similar value to that of mild acidolysis lignin. This observation has led to the conclusion that organosolv process with ethanol was the industrial process imparting fewer modifications to tobacco stalks lignin. Other frequent carbon−carbon linkages, such as β-5, found in phenylcoumarans, and β−β, found in resinols, depicted in Table 2, were also found in high amounts in tobacco lignins, compared to lignins from other species.9,10 Herbaceous plants usually present higher contents of H units than woody materials,10 which is in accordance with the S:G:H ratio found for these lignins (Table 3). Approximately 20% of the

total units in tobacco lignin is H-type. The content of V and Sy precursors in herbaceous lignins, and in tobacco lignins in particular, are naturally lower in comparison to woody biomasses. Moreover, the low content of Hy found in tobacco lignins by NO indicates that H units are more involved in condensed structures than are the S and G units. Another observation can be drawn from the NMR data: processing induces changes in methoxyl content (decrease from LTobA to LTobObut, being more accentuated for LTobSE) mainly due to demethoxylation reactions but also due to degradation of syringyl units and/or topochemical effects of delignification. The lower methoxyl content found in processed lignins (organosolv and steam explosion) due to the demethoxylation reactions was already noticed by other authors.21,30,31 This could be pointed out as one of the reasons for the lower value of methoxyl found by 13C NMR (δ 54.3−57.3 ppm) than the methoxyl content calculated from the S:G:H ratio. 13C NMR data summarized in Table 3 allows drawing the main structural features of the tobacco lignins that will be further used as normalized values for lignin classification by the radar tool. 3.2. Lignin Radar as a Tool for Lignin Classification. One important factor in the evaluation of the potential of any lignin as a source of V and Sy by oxidative process involving O2 D

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

Figure 2. Profiles of monomeric products (Sy, SA, SO, V, VA, VO, and Hy) during the oxidation of LTobObut (A and B) and LTobOethan (C and D). General conditions: 60 g/L of lignin, pO2 = 3 bar, Ptotal = 9.8 bar, Tinitial = 120 °C.

in alkaline medium is the content of β-O-4 structures, in particular, those without carbonyl groups at α position of ppu (briefly β-O-4 without CαO). In alkaline medium, in an early stage of oxidation, β-O-4 structures with free phenolic groups (nonetherified) and carrying a hydroxyl (or alkoxyl or phenoxyl) at Cα are first converted to the enol ether,28,29,32 a very reactive structure in the presence of O2. This structure is the precursor of the phenolic aldehydes,32 resulting in V (from G units) or Sy (from S units) when the aromatic ring is not involved in a condensed structure. In etherified β-O-4 structures, the ether linkage is cleaved producing an vicinal diol28,29 and generating new phenolic groups (in the second ppu) that can become the reactive conjugated structure (enol ether structure). These new structures undergo oxidation producing V and Sy if they are not condensed. Therefore, these two coexistent features are important for V and Sy production by oxidation by O2: low frequency of condensed linkages and high β-O-4 structures without CαO. Structures carrying C O at Cα lead preferentially to the respective carboxylic acids,32 VA and SA. These structures are always produced in the oxidation process with O2 in alkaline medium, but it is preferred that they be at low levels in the initial lignin. However, the estimation of this particular structural feature in lignin (β-O-4 structures without CαO) by 13C NMR is based on Cγ in a very narrow chemical shift interval (Table 2), with a high probability of interference of signals from adjacent chemical shifts (for example, Cγ in cinnamyl alcohol units).33 Therefore, for the purpose of lignin evaluation, total β-O-4

estimation was taken as a reference for evaluation of reactive structures in oxidation. Considering these statements, β-O-4 structures, NCS, S and G proportion (drawn from S:G:H as total), and NO yield on Sy and V (as a measurement of the reactivity of noncondensed fractions of lignin) were the parameters selected as descriptors for the radar plots for lignin evaluation. 3.2.1. Radar Classification of Tobacco Lignins. The values of all the parameters for tobacco lignins are presented as radar plots in Figure 1. This representation allows a direct classification of various lignins by comparison of the key descriptors. The area defined by β-O-4 and NCS in each radar plot provides the first illustration of the amount of lignin degradation effected by processing as compared with lignin isolated by mild acidolysis. The radar plots demonstrate that LTobObut is the lignin with the lowest area of the triangle defined by these two descriptors (hereafter designated as triangle Δ) and, as such, is the most transformed lignin. These observations and the low values of Sy and V (produced by NO) suggest that this lignin is the one with the lowest potential as a source of V and Sy by alkaline oxidation by O2. The representation of G and S as axes of radar plots is particularly relevant when comparing lignins from different origins due to natural variation among species (higher variation than for the same lignin produced by different processes). In these cases, it is more difficult to classify and evaluate the lignins. The descriptors G and S reveal two levels of impact on lignin: (1) E

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

Figure 3. Radar classification for Kraft lignins from bole wood of different species.

described in section 2. The product profile for each lignin is presented in Figure 2, disclosing the maximum yields obtained for V and Sy. LTobOethan lignin produces 1.2% of V and 0.94% of Sy, while for LTobObut lower maximum values were obtained (0.74% of V and 0.34% of Sy), which is in accordance with the prediction provided by the radar classification using the descriptors. Other phenolic compounds were also identified in the oxidation mixture in lower percentages: Hy, SA, VA, acetosyringone (SO), and acetovanillone (VO), Figure 2. Considering G-derivatives, the yield of VA and VO represent 49% to 55% and 16% to 20% of the maximum yield obtained for V, respectively. The proportion of SA and SO relative to the maximum yield of Sy was between 43% and 45% for SA and 16% and 20% for SO. The different reactivity of S and G units influences the time to maximum yield of Sy and V, and is lower for S-derivatives. These observations are in accordance with product profiles obtained by alkaline oxidation with O2 of other lignins from different species and delignification processes.5,12 3.2.2. Radar Classification of Hardwood Lignins. The proposed methodology for radar classification was applied to a larger assortment of lignins, from different species and

Different reactivity of S and G could impart different ability for Sy and V production for the same triangle Δ. (2) β-O-4 and NCS could be unevenly distributed between G and S units. This is the particular case of G units which are more prone to condensation, and therefore, V production is unfavorable in comparison to Sy. As a result, the information imparted by triangle Δ should be complemented by the S and G content and the Sy and V data resulting from NO. Among the tobacco lignins, S and G did not show noteworthy differences. However, Sy and V descriptors showed that LTobObut is the least reactive lignin, followed by LTobSE. In this study, the ranking of lignins was performed using constant weighting factors for each descriptor. As such, and based on the classification, the ascending order of lignins according to the prospective yield for V and Sy by oxidation with O2 in alkaline medium under the same conditions (pH, temperature, O2 partial pressure) is LTobObut < LTobSE < LTobOethan< LTobA. Experimental validation of the radar classification of tobacco lignins was performed using two tobacco lignins. Batch oxidation in alkaline medium with O2 of lignins LTobObut and LTobOethan was performed under the same conditions, as F

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

Figure 4. Radar classification for eucalyptus bole lignins produced by different processes.

classification approach, the nonrandom distribution of S and G units among β-O-4 and NCS leads to different reactivity by oxidation, which is quite evident in this case. As such, LKEgbole has earned a better classification as a source of Sy, although for V the potential is similar for the three lignins. 3.2.3. Different Delignification Processes. The driving force of this comparison is to state the effect of different delignification processes on the same hardwood species (eucalyptus) and to qualify the lignins produced. In a wide scope, this approach allows evaluation of the consequences of the chosen delignification process or conditions for the same process and, in a last instance, tailoring the process taking into account the lignin produced and the route of valorization. As shown in the radar plots presented in Figure 4, lignin from mild acidolysis has the highest values for all descriptors. The best ranking of this lignin as a source of phenolic aldehydes is a consequence of the low impact that the mild acidolysis process has on the native lignin structure. Although this process is limited to laboratory scale because of solvent and conditions used, it is widely used for detailed characterization purposes. The lower area of triangle Δ found for LKEgbole is caused by the Kraft process which promotes a higher depolymerization of

delignification processes. The lignins from eucalyptus, mimosa, and willow were formerly analyzed in our laboratory, using the same methods and techniques used for tobacco lignins. These results are summarized in Table S3 (Supporting Information). Lignins were classified based on descriptors identified as key factors for the production of V and Sy by oxidation with O2, as described for tobacco lignins: β-O-4, NCS, S and G, V and Sy from NO. Figure 3 depicts the radar plots of Kraft lignin from the bole wood pulping (at similar process conditions) of different species. Willow Kraft lignin shows the lowest area defined by triangle Δ denoting that this lignin is the most modified among the three. The highest S/G ratio in LKEgbole is related with the high proportion of S units in this lignin, also quite evidenced by radar plots in Figure 3. In accordance with previous studies on the effect of the S/G ratio on Kraft pulping of hardwoods, this is an important parameter affecting pulp yield under different pulping conditions required for the same delignification degree.8,34−36 From the point of view of lignin’s aptitude for aldehyde production, LKEgbole benefits from its high content of S and Sy descriptors, when compared with LKwillow and LKmimosa. As referred to in the first explanation of the radar G

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

Figure 5. Radar classification for acidolysis lignins obtained from bole, bark, and branches of eucalyptus.

components in the raw material and, in particular, the structure of the fibrous material (which defines also the accessibility of chemicals and the ability for diffusion of products). Therefore, the descriptors of these lignins take into account also the effect of process on a particular material, making this tool even more comprehensive and useful for the classification of lignins for a specific application. Among all, bark lignin presents the lowest radar area indicating a high degree of condensation, intense depolymerization of original aryl−ether linkages, and the lowest Sy and V. In spite of the similarities of bole and bark acidolysis lignins (Figure 5), Kraft lignins are rather different (Figure 6): LKEgbark presents lower NCS as well as lower β-O-4, Sy, and V. The radar areas of LKEgbranch and LKEgsawd (Figure 6) showed that these two processed lignins have similar potential for phenolic aldehydes production and are not very different from LKEgbole. This last one is produced by the pulp and paper industries using eucalyptus bole as a raw material, and for this reason, it would draw more attention in the short term. However, Kraft lignins from the other eucalyptus-derived materials could also be explored for biomaterials. Using the radar areas, the ranking of Kraft lignins as source of phenolic

lignin in bulk and in liquid phase than pulping by organosolv. The differences in the radar plots for lignins from organosolv and Kraft processes clearly indicate that the former is a more preserved lignin. The organosolv process stands out as a better choice for wood delignification from the perspective of lignin valorization by means of oxidation to produce Sy and V. 3.2.4. Different Morphologic Parts and Sawdust of the Same Woody Species. Mild acidolysis lignins from wood and bark of eucalyptus trees were characterized in a previous publication,9 concluding that the lignins of these two morphologic parts showed few differences. In accordance, Figure 5 shows only slightly lower intensity of descriptors for βO-4 and S in the case of LAEgbark, as reflected by the radar area. In spite of these differences, radar plots allow ranking these lignins as having equivalent potential for oxidation. LAEgbranch radar is also similar to bole and bark of eucalyptus, but indicating lower S. The Sy descriptor, comparing bark and bole, denotes a less reactive lignin toward oxidation, which suggests that it would require different conditions in oxidation with O2. Kraft lignins of bole, bark, branches, and sawdust of eucalyptus were produced using tailored conditions for each material to reach the same pulp delignification level. The required pulping conditions are related to the content of other H

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

Figure 6. Radar classification for Kraft lignins obtained from bole, bark, branches, and sawdust of eucalyptus.

aldehydes under identical conditions is LKEgbark < LKEgsawd ∼ LKEgbranch < LKEgbole. One of the challenges associated with exploiting lignin is the variability resulting from the type of plant and species, the delignification process, and the subsequent processing, all of which modifies its structure, making a constant and uniform lignin difficult to obtain. This is one aspect to which lignocellulosic processors must pay attention, today and in the future. Scientific knowledge about the impact of processing on lignin should be a key consideration when deciding both (i) lignin production and recovery processes used in biorefinery operations and (ii) process conditions employed in the further valorization process. Similar considerations have been utilized for many years for wood in the pulp and paper industry and for different lignocellulosic materials in flexible biorefineries. A similar approach which considers lignin does not exist today, but as lignin becomes a useful byproduct it may be necessary. This paper intends to present one approach to provide useful data for assessing lignin characteristics with the aim of maximizing lignin valorization.

4. CONCLUSIONS The radar classification technique presented in this paper allows the screening of lignins resulting from industrial or preindustrial processes for their potential as sources of Sy and V. For the lignins evaluated in this study the ranking is as follows: LTobObut < LTobSE ∼ LKEgbark < LKwillow < LTobOethan < LKmimosa < LKEgsawd ∼ LKEgbranch < LKEgbole < LOEgbole. Lignin of tobacco stalks was isolated by acidolysis (to preserve as much as possible their native structure) and then characterized extensively. Data on this lignin, reported for the first time, and data from the literature on lignins isolated by the same process from eucalyptus bole wood, bark, and branches allowed the lignin of each material to be classified and compared. Among all lignins, LAEgbole is the one with highest radar area due to the highest intensity in all descriptors. This lignin is an excellent source of Sy due to its particularly high content in S units and favorable structure. However, the classification of the processed lignin is quite different and as denoted above is dependent on the delignification process. The potential for LAEgbark is slightly higher than LAEgbranch; however, the required conditions for processing bark by Kraft pulping led to a pronounced increase in degradation reactions and drastically reduced its potential. I

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research Tobacco stalks showed a similar potential for V and for Sy production with an overall favorable classification. Additionally, the effect of processing is once again demonstrated as evidenced by the drastic changes induced in the lignin structure by organosolv pulping with butanol, ranking it at the bottom of the above list. Oxidation with O2 in alkaline medium of LTobObut and LTobOethan under the same conditions has confirmed the qualitative differences of yields predicted by the radar classification. The radar classification of lignins can be adapted to include different or additional descriptors according to the application planned for the lignin being a useful predictive tool for product and process design. Knowledge of the relevant structural parameters for each lignin and the intended application is required to utilize this tool. Radar plots developed from key descriptors may demystify the complexity of lignin and direct process variable selection to achieve maximum valorization of lignin.



ASSOCIATED CONTENT

S Supporting Information *

Additional tables: Table S1, species, process delignification, and denomination of each sample; Table S2, inorganic compounds and carbohydrate contents of lignins; Table S3, main structural characteristics of lignins assessed by 13C NMR and NO. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.iecr.5b01859.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: 351 22 041 3606. Fax: 351 22 508 1449. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Carina Costa thanks FCT, Fundaçaõ para a Ciência e Tecnologia (Portugal,) for Ph.D. Grant SFRH/BD/89570/ 2012. Tobacco lignins were provided by R. J. Reynolds Tobacco Company, Winston Salem, NC, U.S.A. This work was carried out under the Project No. 33969 Conception of biobased products from renewable lignocellulosic sources as precursors for bioindustry of chemical synthesis and biomaterials, funded by FEDER through the Operational Programme for Competitiveness Factors of the National Strategic Reference Framework (QREN). This work was also cofinanced by FCT/MEC, FEDER under Programe PT2020 (Project UID/EQU/50020/2013), by FCT and FEDER under Programe COMPETE (Project PEst-C/EQB/LA0020/2013), and by QREN, ON2, and FEDER under Programe COMPETE (Project NORTE-07-0162-FEDER-000050).



LAEgbole = lignin isolated by mild acidolysis from eucalyptus bole LAEgbranch = lignin isolated by mild acidolysis from eucalyptus branches LKEgbark = lignin from laboratory Kraft pulping liquor of eucalyptus bark LKEgbole = lignin from industrial Kraft liquor of eucalyptus bole LKEgbranch = lignin from laboratory Kraft pulping liquor of eucalyptus branches LKEgsawd = lignin from laboratory Kraft pulping liquor of eucalyptus sawdust LKmimosa = lignin isolated from laboratory Kraft pulping liquor of mimosa LKwillow = lignin isolated from laboratory Kraft pulping liquor of willow LOEgbole = lignin produced by ethanol organosolv process of eucalyptus bole LTobA = lignin isolated by mild acidolysis from tobacco stalks LTobObut = lignin produced by butanol organosolv process of tobacco stalks LTobOethan = lignin produced by ethanol organosolv process of tobacco stalks LTobSE = lignin produced by steam explosion process of tobacco stalks NCS = noncondensed structures NMR = nuclear magnetic resonance NO = nitrobenzene oxidation ppu = phenylpropane unit S = syringyl SA = syringic acid SO = acetosyringone Sy = syringaldehyde V = vanillin VA = vanillic acid VO = acetovanillone

REFERENCES

(1) Pinto, P. C. R.; Borges da Silva, E.; Rodrigues, A. Lignin as source of fine chemicals: vanillin and syringaldehyde. In Biomass Conversion; Baskar, C., Baskar, S., Dhillon, R. S., Eds.; Springer: Berlin, Germany, 2012. (2) Berlin, A.; Balakshin, M. Industrial lignins: Analysis, properties, and applications. In Bioenergy Research: Advances and Applications; Gupta, V. K., Tuohy, M. G., Kubicek, C. P., Saddler, J., Xu, F., Eds.; Elsevier: Amsterdam, The Netherlands, 2014. (3) Lora, J. H.; Glasser, W. G. Recent Industrial Applications of Lignin: A Sustainable Alternative to Nonrenewable Materials. J. Polym. Environ. 2002, 10, 39−48. (4) Wu, G.; Heitz, M.; Chornet, E. Improved alkaline oxidation process for the production of aldehydes (vanillin and syringaldehyde) from steam-explosion hardwood lignin. Ind. Eng. Chem. Res. 1994, 33, 718−723. (5) Pinto, P. C.; 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. (6) Voitl, T.; Rohr, P.R.v. Demonstration of a process for the conversion of kraft lignin into vanillin and methyl vanillate by acidic oxidation in aqueous methanol. Ind. Eng. Chem. Res. 2010, 49, 520− 525. (7) Labat, G.; Gonçalves, A. Oxidation in acidic medium of lignins from agricultural residues. Appl. Biochem. Biotechnol. 2008, 148, 151− 161.



ABBREVIATIONS DC = degree of condensation DMSO-d6 = deuterated dimethyl sulfoxide G = guaiacyl GC-FID = gas chromatography with flame ionization detector GC−MS = gas chromatography−mass spectrometry H = p-hydroxyphenyl Hy = p-hydroxybenzaldehyde LAEgbark = lignin isolated by mild acidolysis from eucalyptus bark J

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

(29) Sjöström, E. Wood Chemistry: Fundamentals and Applications; Academic Press: London, 1993. (30) Wen, J. L.; Sun, S. L.; Yuan, T. Q.; Xu, F.; Sun, R. C. Structural elucidation of lignin polymers of Eucalyptus chips during organosolv pretreatment and extended delignification. J. Agric. Food Chem. 2013, 61, 11067−11075. (31) Li, J.; Gellerstedt, G.; Toven, K. Steam explosion lignins; their extraction, structure and potential as feedstock for biodiesel and chemicals. Bioresour. Technol. 2009, 100, 2556−2561. (32) Kuitunen, S.; Kalliola, A.; Tarvo, V.; Tamminen, T.; Rovio, S.; Liitiä, T.; Ohra-aho, T.; Lehtimaa, T.; Vuorinen, T.; Alopaeus, V. Lignin oxidation mechanisms under oxygen delignification conditions. Part 3. Reaction pathways and modeling. Holzforschung 2011, 65, 587−599. (33) Lin, S. Y., Dence, C. W., Eds. Methods in Lignin Chemistry; Springer Series in Wood Science; Springer-Verlag: Berlin, Heidelberg, Germany, 1992. (34) del Río, J. C.; Gutiérrez, A.; Hernando, M.; Landín, P.; Romero, J.; Martínez, Á .T. Determining the influence of eucalypt lignin composition in paper pulp yield using Py-GC/MS. J. Anal. Appl. Pyrolysis 2005, 74, 110−115. (35) Pinto, P. C.; Evtuguin, D. V.; Neto, C. P. Effect of structural features of wood biopolymers on hardwood pulping and bleaching performance. Ind. Eng. Chem. Res. 2005, 44, 9777−9784. (36) González-Vila, F. J.; Almendros, G.; del Río, J. C.; Martın, F.; Gutiérrez, A.; Romero, J. Ease of delignification assessment of wood from different Eucalyptus species by pyrolysis (TMAH)-GC/MS and CP/MAS 13C-NMR spectrometry. J. Anal. Appl. Pyrolysis 1999, 49, 295−305.

(8) Santos, R. B.; Capanema, E. A.; Balakshin, M. Y.; Chang, H.-M.; Jameel, H. Effect of hardwoods characteristics on kraft pulping process: emphasis on lignin structure. BioResources 2011, 6, 3623− 3637. (9) Costa, C. A. E.; Pinto, P. C. R.; Rodrigues, A. E. Evaluation of chemical processing impact on E. globulus wood lignin and comparison with bark lignin. Ind. Crops Prod. 2014, 61, 479−491. (10) Pinto, P. C. R.; Oliveira, C.; Costa, C. A.; Gaspar, A.; Faria, T.; Ataíde, J.; Rodrigues, A. E. Kraft delignification of energy crops in view of pulp production and lignin valorization. Ind. Crops Prod. 2015, 71, 153−162. (11) Evtuguin, D. V.; Neto, C. P.; Silva, A. M. S.; Domingues, P. M.; Amado, F. M. L.; Robert, D.; Faix, O. Comprehensive study on the chemical structure of dioxane lignin from plantation Eucalyptus globulus wood. J. Agric. Food Chem. 2001, 49, 4252−4261. (12) Pinto, P. C. R.; Costa, C. E.; Rodrigues, A. E. Oxidation of lignin from Eucalyptus globulus pulping liquors to produce syringaldehyde and vanillin. Ind. Eng. Chem. Res. 2013, 52, 4421−4428. (13) Buranov, A. U.; Mazza, G. Lignin in straw of herbaceous crops. Ind. Crops Prod. 2008, 28, 237−259. (14) Ghaffar, S. H.; Fan, M. Structural analysis for lignin characteristics in biomass straw. Biomass Bioenergy 2013, 57, 264−279. (15) Pinto, P. C.; Borges da Silva, E. A.; Rodrigues, A. E. Comparative study of SPE and liquid-liquid extraction of lignin oxidation products for HPLC-UV quantification. Ind. Eng. Chem. Res. 2010, 49, 12311−12318. (16) Tarabanko, V. E.; Koropatchinskaya, N. V.; Kudryashev, 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, 44, 367−371. (17) Fernández-Costas, C.; Gouveia, S.; Sanromán, M. A.; Moldes, D. Structural characterization of Kraft lignins from different spent cooking liquors by 1D and 2D Nuclear Magnetic Resonance spectroscopy. Biomass Bioenergy 2014, 63, 156−166. (18) Capanema, E. A.; Balakshin, M. Y.; Kadla, J. F. A comprehensive approach for quantitative lignin characterization by NMR spectroscopy. J. Agric. Food Chem. 2004, 52, 1850−1860. (19) Nimz, H. H.; Robert, D.; Faix, O.; Nemr, M. Carbon-13 NMR spectra of lignins, 8. Structural differences between lignins of hardwoods, softwoods, grasses and compression wood. Holzforschung 1981, 35, 16−26. (20) Xiao, B.; Sun, X. F.; Sun, R. Chemical, structural, and thermal characterizations of alkali-soluble lignins and hemicelluloses, and cellulose from maize stems, rye straw, and rice straw. Polym. Degrad. Stab. 2001, 74, 307−319. (21) Sun, X. F.; Xu, F.; Sun, R. C.; Wang, Y. X.; Fowler, P.; Baird, M. S. Characteristics of degraded lignins obtained from steam exploded wheat straw. Polym. Degrad. Stab. 2004, 86, 245−256. (22) Capanema, E. A.; Balakshin, M. Y.; Kadla, J. F. Quantitative characterization of a hardwood milled wood lignin by nuclear magnetic resonance spectroscopy. J. Agric. Food Chem. 2005, 53, 9639−9649. (23) Heitner, C., Dimmel, D., Schmidt, J., Eds. Lignin and Lignans: Advances in Chemistry; Taylor & Francis: Boca Raton, FL, 2010. (24) Rencoret, J.; Prinsen, P.; Gutiérrez, A.; Martínez, Á .T.; del Río, J. C. Isolation and structural characterization of the milled wood lignin, dioxane lignin, and cellulolytic lignin preparations from brewer’s spent grain. J. Agric. Food Chem. 2015, 63, 603−613. (25) Balakshin, M. Y.; Capanema, E. A.; Santos, R. B.; Chang, H.; Jameel, H. Structural analysis of hardwood native lignins by quantitative 13 C NMR spectroscopy. Holzforschung 2015, DOI: 10.1515/hf-2014-0328. (26) Xu, F.; Sun, J.-X.; Sun, R.; Fowler, P.; Baird, M. S. Comparative study of organosolv lignins from wheat straw. Ind. Crops Prod. 2006, 23, 180−193. (27) Agrupis, S.; Maekawa, E.; Suzuki, K. Industrial utilization of tobacco stalks II: preparation and characterization of tobacco pulp by steam explosion pulping. J. Wood Sci. 2000, 46, 222−229. (28) Gierer, J. Chemistry of Delignification. Part 1: General Concept and Reactions during Pulping. Wood Sci. Technol. 1985, 19, 289−312. K

DOI: 10.1021/acs.iecr.5b01859 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX