Abstract
The purpose of this work was to determine the main factors influencing the phenolic content of bark during acid hydrolysis. The optimization of polysaccharides hydrolysis was done by response surface methodology. The hydrolysis was performed under atmospheric pressure in an aqueous solution of sulfuric acid. An experimental design was applied to analyze the effects of the reaction time (5 to 24 hours), acid concentration (3 to 20%), and solid/liquid ratio (1/10 to 1/5) on the weight loss, lignin content, holocellulose content, and sugar yield for the hydrolysis. The pretreated bark had a high lignin content of 60% resulting from hemicelluloses hydrolysis and phenolic compound condensation.
Download PDF
Full Article
Optimization of Acid Pretreatment in order to Increase the Phenolic Content of Picea abies Bark by Surface Response Methodology
Xavier Duret,a,b,d,* Emmanuel Fredon,a,b,* Eric Masson,d Lyne Desharnais,a and Philippe Gérardina,c
The purpose of this work was to determine the main factors influencing the phenolic content of bark during acid hydrolysis. The optimization of polysaccharides hydrolysis was done by response surface methodology. The hydrolysis was performed under atmospheric pressure in an aqueous solution of sulfuric acid. An experimental design was applied to analyze the effects of the reaction time (5 to 24 hours), acid concentration (3 to 20%), and solid/liquid ratio (1/10 to 1/5) on the weight loss, lignin content, holocellulose content, and sugar yield for the hydrolysis. The pretreated bark had a high lignin content of 60% resulting from hemicelluloses hydrolysis and phenolic compound condensation.
Keywords: Norway spruce; Bark; Acid hydrolysis; Phenolic content optimization; Response surface methodology
Contact information: a: Université de Lorraine, Laboratoire d’Etudes et de Recherche sur le Matériau Bois (LERMAB), Nancy, EA 4370, F-54506 France; b: ENSTIB, 27 rue Philippe Séguin, BP 1041, 88051 EPINAL Cedex 9, France; c: Faculté des Sciences et Technologies, BP 70239, 54506 Vandœuvre les Nancy, France; d: Centre de ressources pour les industries du bois (CRITT Bois), 27 rue Philippe Séguin, B.P 91067 88051, EPINAL cedex 9, France;
* Corresponding authors: rhodanis@wanadoo.fr; emmanuel.fredon@enstib.uhp-nancy.fr
INTRODUCTION
Phenolic resins for wood products have been derived largely from petroleum, a finite natural resource. Some efforts have been devoted to the development of thermosetting phenol-formaldehyde type adhesives from vegetable origins. The best way is the use of tannins or lignin to replace a part of the phenol (Pizzi and Mittal 2003; Yasaki and Collins 1994a,b; Navarette et al. 2012). Bark is a low-value wood by-product of sawmills and paper mills, which is a readily available and renewable resource, amounting to about 7% of the total weight of a tree and 12% of the total volume of a tree (Florentin 2011). Investigation of alternative higher value added uses of bark could be economically advantageous according to the biorefinery concept. Moreover, it brings a new contribution to the attempt concerning the climate change by incorporating the carbon from a photosynthetic material into a durable product. The naturally higher phenolic content of bark and lower polysaccharides content relative to wood is convenient for the phenolic thermosetting resins synthesis. Nevertheless, the still remaining carbohydrate fraction is known to induce a negative effect on the curing of the resins, mechanical properties, and durability (Christiansen and Gillespie 1986; Pilato 2010; Yang et al.2006). The aim of this study was to increase the phenolic content (mainly tannins and lignin) in Picea abies bark solids remaining after pretreatment with acid hydrolysis. As a consequence, the best way to eliminate the carbohydrate fraction in biomass is to degrade them. The polysaccharides can be hydrolyzed by thermal, chemical, or enzymatic reaction (Girio et al. 2010; Matsushita et al. 2010; Mehrotra et al. 2010). Bark is a source of phenolic compounds that are enzymatic inhibitors in the enzymatic hydrolysis, and the thermal treatment induces important chemical changes (Tejirian and Xu 2011). These are the reasons why the pretreatment needs to be performed under acid hydrolysis. Optimization of the phenolic content of bark for chemical feedstock could be a better way to upgrade the spruce bark than ethanol production. The pretreated bark will be liquefied and used to synthesize thermosetting resins, which can be used as adhesives (Pan 2007), foams (Alma and Shiraishi 1998), films (Kurimoto et al. 2001), composites (Kishi and Fijita 2008), or wood impregnation to improve its durability.
EXPERIMENTAL
Materials
Anhydrous Picea abies bark particles were used. The bark was collected in a sawmill of Vosges Department of France. Samples were dried at 103°C until constant weight. Results are based on the anhydrous weight of bark. For higher scale this step will be not necessary. Nevertheless, low moisture is necessary to prevent the degradation of bark in the stock. Bark was milled to pass through a sieve of 400 µm.
Extractive Content
The extractive content of Picea abies bark was determined by successive extractions. The solvents were petroleum ether, a mixture of toluene/ethanol (80/20), and water. The extraction was carried in soxhlet apparatus with more than four cycles per hour for 9 h. The extractive content is calculated as a percentage based on the weight of anhydrous bark.
Total Phenolic Content of Extracts
Analysis of total phenolic content of bark was performed employing the literature method involving Folin-Ciocalteu reagent and gallic acid as standard (Wilfred and Ralph 2006). The reaction is based on the reduction of phosphomolybdic acid by phenols in aqueous alkali, allowing determination of total soluble phenolic compounds. The sample was diluted in water (10 mg/mL). Two milliliters of freshly prepared 2% (w/v) of sodium carbonate was added to 0.1 mL of dilute sample and mixed on a vortex mixer. After 5 min at 25°C, 0.1 mL of 1:1 dilution of Folin-Ciocalteu reagent was added. After 30 min at 25°C, the absorbance was read at 750 nm. Gallic acid was used to generate a calibration curve. The results were then reported as gallic acid equivalent per gram of the extract.
Holocellulose Content
The holocellulose content was determined by the chlorite method (Rowell 2005). To 1 g of the sample, 32 mL of hot distilled water was added with 200 µL of acetic acid and 0.4 g of sodium chlorite. The mixture was heated in a water bath at 75°C. After each hour, a fresh portion of 200 µL of acetic acid and 0.4 g of sodium chlorite was added with shaking. The addition of acetic acid and sodium chlorite was repeated until total delignification and the mixture was let in the water bath at 75°C overnight without
further addition of acetic acid and sodium chlorite. After the reaction, the sample was cooled, filtrated on a sintered glass number 3, and washed with acetone and water. Holocellulose was dried at 105°C until constant weight. The holocellulose content was based on the weight of bark without treatment.
Cellulose and Hemicelluloses Content
The hemicelluloses and α-cellulose content was determined by treatment of holocellulose with sodium hydroxide solution and then with acetic acid according to the procedure described by Rowell (2005). Cellulose content was directly calculated from dry residue weight. One-half grams of anhydrous holocellulose was placed into a glass beaker. Two and a half milliliters of 17.5% NaOH solution was added to the holocellulose. A glass rod was used to manipulate lightly the holocellulose until it became soaked with NaOH solution. After the first addition of NaOH solution, 1.25 mL of 17.5% NaOH solution was added three times at 5 min intervals. The mixture was kept at 20°C for 30 min. Then 8.25 mL of distilled water was added, and the mixture was kept at 20°C for 1 h. The cellulose was filtered with a fritted glass crucible of medium porosity with suction and washed with 25 mL of 8.3% NaOH solution. Then the residues were washed with distilled water; 3.75 mL of 10% acetic acid solution was added into the crucible without suction for 3 min. Then, the suction was applied to filter the residues. The residues were washed for the last time with distilled water until the cellulose was free of acid as indicated by pH paper. The residues were dried at 103°C until constant weight. The cellulose content was expressed as the ratio of residues weight on the initial mass of anhydrous bark (with extractive content). The hemicellulose content was deduced from the cellulose content and holocellulose content of the bark.
Klason Lignin Content
Klason lignin is acid-insoluble lignin (ASTM standard D1106 – 96, 2007). A first hydrolysis was achieved in 72% sulfuric acid at 30°C for 60 min. For 0.1 g of the sample, 1.5 mL of 72% sulfuric acid solution was added. Then the reaction mixture was diluted to obtain 3% sulfuric acid and heated at 120°C in an autoclave for 1 h. The reaction mixture was filtered in sintered glass filter number 3 and washed with hot water. The solid residues were dried at 105°C overnight and weighted. The yield of acid insoluble lignin was based on oven dried original bark (%).
Bark Hydrolysis
Bark was hydrolyzed in aqueous solution of sulfuric acid. The unwashed bark was mixed with an acid solution and heated in oil bath at 100°C, under reflux, with a magnetic stirring. The reaction time, acid concentration, and solid/liquid ratio are set by the experimental design (Table 1).
Table 1. Experimental Domain of Central Composite Design
At the end of the reaction time the mixture was cooled in an ice bath to stop the reaction. The mixture was filtrated under reduced pressure on a glass fiber filter. The bark was rinsed with distilled water. The hydrolyzed bark was dried to a constant weight at 105°C. The hydrolysate was completed to 1 L with distilled water and was kept refrigerated for further analysis. The mass loss was expressed as the ratio of the mass of bark before and after hydrolysis.
HPAEC-PAD Analyses
Separation and quantification of neutral sugars were performed using a Dionex ICS-3000 system consisting of a SP gradient pump, an AS auto sampler, an ED electrochemical detector with a gold working electrode, an Ag/AgCl reference electrode and Chromeleon version 6.8 (Dionex Corp., USA). A Carbopac PA1 (4 x 250 mm, Dionex) column with a guard column (4 x 50 mm, Dionex) was used as a stationary phase using isocratic conditions with 1 mM sodium hydroxide as the eluent. Eluents were prepared by dilution of a 46 to 48% NaOH solution (PA S/4930/05 Fisher Scientific) in ultra-pure water. All eluents were degassed before use by flushing helium through for 20 min; subsequently they were kept under a constant helium pressure (eluent degassing module, Dionex). After each run, the column was washed for 10 min with 200 mM NaOH solution and reequilibrated for 15 min with 1 mM sodium hydroxide. Samples were injected through a 25 µL full loop and separations were performed at 25°C at a rate of 1 mL/min. The pulse sequence for pulsed amperometric detection consisted of a potential of +100 mV (0-200 ms), +100 mV integration (200-400 ms), -2000 mV (410-420 ms), +600 mV (430 ms), and -100 mV (440-500 ms). The results were expressed as a percentage of glucose and non-glucosidic sugars based on the initial mass of anhydrous bark.
Response Surface Methodology
In order to determine the optimal conditions to obtain a maximum phenolic content in the bark, an experimental design with response surface methodology was applied. This experimental design allows studying a phenomenon that can be explained by a quadratic model (Equation 1). β0 is a constant coefficient, βi is the coefficient of the factors i, Xi is the level of the factor i, Xj is the level of the factor j, βi,j is the coefficient of the first order interaction between the factor i and j, and βi,i is the coefficient of the quadratic term of the coefficient i. Three factors determined from a previous experimental design were studied, and their effects on the weight loss, holocellulose content, and Klason lignin yield were determined. For this experimental design, the experimental domain is presented in the Table 1. The experimental design was a central
composite design, and it is showed in Table 4. The results were analyzed using the statistic software R (R Development Core Team 2011). The optimal region were calculated through the first derivate of the mathematical function. Thus, the optimal point was determined by solving the first grade system formed by first derivate equations equal to zero.
(1)
RESULTS AND DISCUSSION
Characterization of Chemical Bark Composition
The chemical composition of Norway spruce bark is shown in Table 2. Values were examined in the objective of enrichment of phenolic part and minimization of polysaccharides. For this purpose, fresh bark was separated following inner and outer bark. Each of them was analyzed for its main components, cellulose, hemicelluloses, lignin, and total phenol content of extracts. The holocellulose content of inner bark and outer bark were quite similar, with 51.4 and 48.9%, respectively. But outer bark also contained more lignin (32%) than the inner bark (12%). Nevertheless, the Klason lignin of bark is composed of lignin and condensed tannins which are not extracted by the common extraction process.
Table 2. Chemical Composition of Picea abies Bark
Moreover, bark extracts are known to be constituted of non-cellulosic polysac-charides, stilbene glucosides, low molecular weight phenolic compounds, and condensed tannins (Krogell et al. 2011). The inner bark extractives content was twice that of outer bark. Nevertheless, inner bark had lower total phenolic content in water extracts than outer bark with 276 mg/g and 436 mg/g, respectively (Table 3). Phenolic compounds are a predominant fraction in bark extracts: The yield of total phenols in the toluene/ethanol and water bark extracts were 773 mg/g (gallic acid equivalent) and 401 mg/g, respec-tively (Table 3). In toluene ethanol extracts, inner bark contained as much phenolic compounds as outer bark. In water extracts, inner bark contains less than outer bark. Krogell et al. (2011) indeed found, in inner bark water extracts, predominantly phenolic extractives such as stilbene glucosides and non-phenolic extractives such as saccharides. Most of the condensed tannins, which have a great importance in resin formulation, are found in the inner bark.
Table 3. Total Phenolic Content of Picea abies Bark extracts
Finally, the difference in global phenolic and carbohydrate parts between outer and inner bark was not important enough to discard one of the fractions. Then, the whole pretreated bark will be used for the thermosetting resins formulation. Nevertheless, it is advantageous to remove the carbohydrate fraction in order to enhance the thermosetting ability.
Response Surface Experimental Design Analysis
Response surface methodology was used to optimize the phenolic content of spruce bark. This method allows finding the quadratic model describing the responses with the factors that were investigated. This kind of experimental design allows for the study of second order phenomena. The response surfaces are presented with only the significant effects. The results of central composite experimental design are presented in Table 4. The analysis of variance and coefficients of response surface are presented in Table 5. The R2 adjusted was acceptable for studied surfaces responses. Nevertheless, caution should be taken with the surface responses of holocellulose due to a low R2 adjusted.
Table 4. Factors, Coded Levels, and Responses of Central Composite Experimental Design
The F test value for variance analysis was good with a p-value lower than 0.01. The F test value for the lack of fit was good with a p-value lower than 0.02, except for Klason lignin content and holocellulose content. A p-value lower than 0.05 is usually taken as a limit, but a p-value of 0.1 is still significant. Holocellulose content and Klason lignin content could be fitted better with another model (third order), (Table 4).
Table 5. Analysis of Central Composite Experimental Design for Quadratic Model
Mass loss analysis
Figure 1 shows the response surface of the mass loss. The mass loss is represen-ted as a function of reaction time and acid concentration. One response surface is presented for each solid/liquid ratio levels of -1.668, 0, and 1.668. The most significant effects were the solid/liquid ratio (β3) and the quadratic effect of the reaction time (β11). The acid concentration (β2) and the interaction between reaction time and solid/liquid ratio were also significant (β13). The reaction time (β1) and the quadratic effect of the acid concentration (β22) had low effect on the mass loss. The interactions β12, β23, and the quadratic effect of solid/liquid ratio (β33) were negligible (Table 5). The mass loss as a function of reaction time reached a maximum. The optimal reaction time did not change greatly with the acid concentration, but the mass loss was higher when the acid concentration increased. The optimal reaction time and the mass loss were greatly influenced by the solid/liquid ratio. The optimal reaction time was longer and the mass loss was higher when the solid/liquid ratio was lower. Then, the interaction between reaction time and solid/liquid ratio was significant and was the only significant interaction. The non-negligible quadratic effect of the reaction time on the mass loss could be explained by condensation reactions resulting from degradation products of carbohydrate as furfural (Nguyen et al. 1999; Sakostschikoff et al. 1934; Stanciu and Ciurea 2008) and condensed tannins polymerization under acidic conditions (Evelyn et al. 1960).
Fig. 1. Response surface of mass loss
Klason lignin content
The Klason lignin of hydrolyzed barks is composed of lignin and condensed tannin. Figure 2 shows the response surface of the Klason lignin content, represented as a function of reaction time and acid concentration.
Fig. 2. Response surface of lignin content
Response surfaces are presented for the following solid/liquid ratio levels: -1.668, 0, and 1.668. The levels -1.668 and 1.668 are represented by the same surface. The Klason lignin content was most significantly influenced by the main effect of acid concentration (β1), then by the main effect of reaction time (β2) and all the quadratic effects (β11, β22, β33). The interactions and the solid/liquid ratio were negligible (Table 5). Klason lignin content increased with high acid concentration and reached a maximum of 60% at optimal reaction time.
Holocellulose content
Figure 3 shows the response surface of the holocellulose content. The holocellulose content is represented as a function of reaction time and acid concentration. The holocellulose content was mainly influenced by the main effect of reaction time (β1) and acid concentration (β2), and also the quadratic effect of the reaction time (β11). The effects of solid/liquid ratio (β3), the interactions (β12, β13, β23), the quadratic effects of the acid concentration (β22), and the solid/liquid ratio (β33) were negligible. The acid concentration (β2) is the most significant factor (Table 5). Holocellulose content clearly decreased with increasing of reaction time and acid concentration due to hydrolysis of polysaccharides.
Fig. 3. Response surface of holocellulose content
Total sugar recovery
Figure 4 shows the response surface of sugar recovery. The sugar recovery in hydrolysate is represented as a function of reaction time and acid concentration. The sugar recovery was mainly influenced by the acid concentration, the interaction between reaction time and acid concentration (β12), and the quadratic effect of acid concentration (β22). The other coefficients were negligible (Table 5). For a low acid concentration, the sugar recovery increased with increasing of the reaction time. For a high acid concentration, the sugar recovery decreases with increasing of the reaction time. That is the reason why the interaction between reaction time and acid concentration is so important. The decreasing of sugar recovery is due to furfural formation from sugar degradation reaction.
Fig. 4. Response surface of sugar recovery
Glucose and non-glucosidic sugars recovery
Figures 5 and 6 show the response surface of glucose and non-glucosidic sugars from hemicelluloses, respectively. The glucose and other sugars are represented as a function of reaction time and acid concentration. The glucose content of hydrolysate after bark hydrolysis was influenced by the reaction time, the acid concentration, the solid/liquid ratio, the interaction between the reaction time and the acid concentration, and also the quadratic effect of the reaction time. The other effects were negligible. Variations of the glucose content were low. Cellulose is difficult to hydrolyze. The effect of the acid concentration was higher for a short reaction time. The effect of reaction time was also higher for a low acid concentration. This means that the interaction between reaction time and acid concentration was important.