Triticum spelta straw hydrothermal pretreatment for the production of glucose syrups via enzymatic hydrolysis

Biochemical Engineering Journal 151 (2019) 107340

Contents lists available at ScienceDirect

Biochemical Engineering Journal journal homepage: www.elsevier.com/locate/bej

Regular article

Triticum spelta straw hydrothermal pretreatment for the production of glucose syrups via enzymatic hydrolysis

T

⁎

Rodrigo da Rocha Olivieri de Barrosa, , Paolo Becarellib, Raul Alves de Oliveiraa, Leonardo Tognottic, Elba Pinto da Silva Bona a

Bioethanol Laboratory, Department of Biochemistry, Chemistry Institute - Federal University of Rio de Janeiro, CEP 21941-596, Rio de Janeiro, RJ, Brazil Department of Engineering of Energy, Systems, Territory and Construction (DESTeC), University of Pisa, Largo Lucio Lazzarino 1, 56122, Pisa, Italy c Department of Chemical Engineering, University of Pisa, 56100, Pisa, Italy b

H I GH L IG H T S

spelta straw was used for the production of glucose syrups. • Triticum experimental design was successfully used for hydrothermal pretreatment parameters. • Statistical hydrolysis of the solid stream yielded a 56.33 g L glucose syrup. • Enzymatic pretreatment liquid stream presented 16 g L of hemicellulose derived sugars. • The • The pretreatment liquid stream presented low concentration of inhibitors. −1

−1

A R T I C LE I N FO

A B S T R A C T

Keywords: Glucose sugar syrups Triticum spelta straw Wheat straw Hydrothermal pretreatment Statistical experimental design Enzymatic hydrolysis

The global climate changes related to the use fossil fuels and the oil price volatility have been feeding the interest for renewable feedstock such as the by-products of the agroindustry. This is a challenging scenario as the diversity of plant materials worldwide calls for customized pretreatment processes to achieve a significant cellulose enzymatic saccharification coupled to the preservation of the biomass cellulose content as well as to a low degradation of biomass sugars into inhibitory compounds that is a recurrent problem in biomass processing. This work reports a statistical experimental design for hydrothermal pretreatment of spelt straw (Triticum spelta) aiming to maximize the straw enzymatic conversion in glucose syrups while minimizing the formation of undesirable biomass derived inhibitors. In the best pretreatment conditions of 180 °C for 10 min it was observed a cellulose conversion yield of 52.16% that resulted in a glucose syrup with 56.33 g L−1 in a reaction mixture presenting 200 g L-1 of pretreated straw. The liquid current from the pretreatment, rich in hemicellulose derived sugars, presented 1.63 g L−1 acetic acid, 0.67 g L−1 furfural and 0.06 g L−1 HMF. The cellulose content of the pretreated straw and the corresponding glucose concentration in the hydrolysate were mathematically modeled with relative deviation smaller than 5 % as a function of pretreatment parameters temperature and pretreatment time.

1. Introduction The serious global climate changes related to the use fossil fuels and the oil price volatility have been feeding the interest for renewable alternative feedstock with wider geographic distribution and lower cost such as the by-products of the agroindustry [1,2]. The lignocellulosic biomass is a rich and chemically complex material whose structure is represented by the physical-chemical interaction of cellulose a linear polymer of glucose, hemicellulose a highly

⁎

branched mostly C5 sugars polysaccharide of variable composition, and lignin a very complex macromolecule rich in aromatic subunits [3–5]. The conversion of lignocellulosic materials into sugar syrups has been currently done via enzymatic hydrolysis. For that the lignocellulosic materials must undergo a pretreatment step to alter a structure that has evolved to resist the action of physical, chemical, and biological external agents [3,6]. The pretreatment is intended to expose, as much as possible, the chain of the cellulose to the multi-enzymatic process that convert this polysaccharide into glucose molecules [7]. Biomass

Corresponding author at: Avenida Pedro Calmon s/n, Bloco P, P4. Cidade Universitária, CEP 21941-596, Ilha do Fundão, Rio de Janeiro, RJ, Brazil. E-mail address: [email protected] (R.d.R.O. de Barros).

https://doi.org/10.1016/j.bej.2019.107340 Received 11 April 2019; Received in revised form 22 July 2019; Accepted 13 August 2019 Available online 14 August 2019 1369-703X/ © 2019 Elsevier B.V. All rights reserved.

Biochemical Engineering Journal 151 (2019) 107340

R.d.R.O. de Barros, et al.

processed in a cutting mill using a Retsch (Haan, Germany) machine model SM-300 equipped with a sieve with a cut-off of 2 mm. This procedure selected a standardized sample size to avoid a wide variability in the particle size distribution that statistically hinders valid and reproducible results [8]; another reason for choosing this procedure was to homogenize as much as possible the variability in the composition of this plant material. The hydrothermal pretreatment experiments were carried out in a reactor Parr Instruments (Parr Instruments, Illinois, USA) model 4520 with the capacity of 1 L. A total of 30 g (dry weight) of biomass was placed in the reactor followed by the addition of 300 mL (around 30% of the reactor capacity) of distilled water. After a careful mixing, the pH of the biomass-water suspension was measured to verify a pH in the range of 5 to 6 that minimize the formation of biomass-derived microbial inhibitors. The reactor was afterward sealed by a safety lock, the electric heater was put into position, and the control thermocouple was connected. The reactor was flushed with nitrogen up to a pressure of 5 barg and left to drain slowly from the exhaust valve. This operation was repeated five times in succession to ensure an inert atmosphere inside the reactor avoiding oxidation of the material at high temperatures. Finally, the reactor was pressurized with nitrogen at a working pressure of 20 barg; the temperature was set to a specific set point, and the stirrer was set to 100 RPM and heater startup. In the measurement of the pretreatment time, it was assumed for the onset of the hydrothermal process the time when the temperature reached its set point. After the predetermined pretreatment time, the heater was turned off, and the reactor was cooled down with ice and water. Upon reaching a temperature of 40 °C, the valve was opened to drain the nitrogen until the atmospheric pressure (0 barg); the vessel was then open to retrieve the pretreated material, which was composed of the solid cellulignin and the liquid hemicellulose-derived stream that were separated by filtration using a vacuum pump. A sample of the liquid fraction was stored and frozen for the analysis of the inhibitors concentration, total xylose and mass balance as well. The solid fraction was washed with 600 mL of distilled water to remove impregnated sugars and inhibitors.

pretreatment studies have been at the center of scientific and industrial research aiming the enzymatic production of sugar syrups with high concentrations and high yields coupled to a low degradation of biomass sugars from both the hemicellulose and the cellulose, into inhibitory compounds, a recurrent problem in biomass processing. Moreover, the biomass pretreatment process must be cost attractive and technically effective. The hydrothermal pretreatment, also known as liquid hot water (LHW) or hot compressed water (HCW), are based on the use of pressurized water to maintain the liquid state at high temperatures (160 to 240 °C). The process alters the native structure of the biomass through the partial removal of the hemicellulose fraction, combined to changes in the lignin structure, rendering the cellulose more accessible to the subsequent enzymatic hydrolysis step [6]. Pending on the process conditions of temperature and pretreatment time the hemicellulose is solubilized in different proportions of C5 sugars monomers and oligosaccharides. The process conditions also dose the extension of the C5 and C6 sugars respective dehydration into 5-furfural (FF) and 5-hydroximethylfurfural (HMF), the extension of the removal of acetic acid from the hemicellulose chain and the removal of phenolic compounds from the lignin. The FF, HMF, acetic acid, and phenols released to the pretreatment liquid stream are microbial inhibitors that have been hindering the biotechnological use of this sugar-rich liquid stream [8]. Thus, an optimized hydrothermal pretreatment condition must balance the highest possible enzymatic hydrolysis of the cellulose present in the solid fraction to acceptable levels of inhibitors in the pretreament liquid stream rich in hemicellulose derived sugars. This work evaluated the production of glucose sugar syrups from spelt straw (Triticum spelta) through hydrothermal pretreatment followed by enzymatic saccharification. Statistical experimental design was used to define the better pretreatment condition for the subsequent enzymatic hydrolysis, taking into account the extent of hemicellulose removal and deacetylation and the formation of the inhibitory compounds FF and HMF. Triticum spelta that is closely related to the common wheat Triticum aestivum [9] is gaining the interest of farmers as it requires less fertilizer and is more resistant to drought and fungus pathogens than common wheat. Moreover, its high nutritional value increases its importance in agriculture, especially in organic farming systems [10,11]. After the grains are processed for food, the spelt biomass (the chaff, glumes and stems) lags as a by-product or waste that can be used as raw material for biomass combustion for sustainable heat production [12,13]. The spelt straw is widely available in Italy, particularly in the region of Garfagnana where, despite its use as animal feed, the vast majority is discarded as waste. It is estimated that 8 metric tons ha−1 of Triticum spelta straw, a potential biorefinery feedstock, are produced yearly, calling for customized pretreatment studies as only the chemical composition of this material has been so far addressed [14,15].

2.3. Use of experimental design to optimize the hydrothermal pretreatment of spelt straw biomass This work used Central Composite Design (CCD) that allows the analysis of the interactions and the optimization of two factors previously shown to be the most significant. It consists of a central point that will be executed in multiple replicas giving an internal estimate of pure error and axial points that will determine the quadratic terms [16,17]. The conditions of central points were chosen considering the possibility of experimental errors. For the pretreatment experiments it was necessary to associate the coded values to the real ones. The choice of the extremes of the interval was based on experiments already done for the similar material sugar cane straw allowing confidence in the operation of the equipment. The temperature range was from 120 °C to 240 °C with the pretreatment time of 10 min to 120 min. By placing the ends of the two intervals with the coded values of ± 1.41, with linear interpolation, it was possible to obtain the real value matrix of the CCD. The coded matrix of experiments and the real values for temperature and pretreatment time considering two input variables are shown in Table 1. Each pretreatment condition was repeated the necessary number of times to obtain 30 g of pretreated spelt straw to perform triplicates of the enzymatic hydrolysis experiments. This phase allowed the evaluation of the response variables Y to validate the model. As outputs six different responses were evaluated: (a) the cellulose content of the HCW-treated straw, (b) the glucose concentration of the sugar syrup, (c) the glucose yield after 48 h of enzymatic hydrolysis and the concentrations of (d) the acetic acid, (e) the furfural, and (f) HMF in the liquid fraction from the pretreatment. The software Statistica 7.0 (StatSoft, USA) was used to gather all data of the complete CCD

2. Materials and methodology 2.1. Materials A total amount of 6 kg of Triticum spelta was collected in the region of Garfagnana, Northern Tuscany, Italy, and sent by mail to the Bioethanol Laboratory of the Federal University of Rio de Janeiro (www.bioetanol-ufrj.com.br) where all the experiments were conducted. The material was first subjected to moisture determination using a moisture analyzer Gehaka (São Paulo, Brazil) model IV-3000 that records the decrease in weight and estimates the instantaneous moisture until the humidity value stabilizes with a variation less than 0.1% min−1. The equipment was set at 105 °C in auto-dry mode. 2.2. Hydrothermal pretreatment Before the hydrothermal pretreatment, the spelt straw was 2

Biochemical Engineering Journal 151 (2019) 107340

R.d.R.O. de Barros, et al.

the moisture content. The concentration of furfural, HMF, and acetic acid in the liquid fraction from HCW pretreatment was carried out using HPLC for all the 11 different pretreatment conditions.

Table 1 Coded and real value matrix for complete Central Composite Design, 22. Experiment

Temperature (X1)

Pretreatment time (X2)

1 2 3 4 5 6 7 8 9 10 11

−1 (137 °C) 1 (223 °C) −1 (137 °C) 1 (223 °C) −1.41 (120 °C) 1.41 (240 °C) 0 (180 °C) 0 (180 °C) 0 (180 °C) 0 (180 °C) 0 (180 °C)

−1 (26 min.) −1 (26 min.) 1 (104 min.) 1 (104 min.) 0 (65 min.) 0 (65 min.) −1.41 (10 min.) 1.41 (120 min.) 0 (65 min.) 0 (65 min.) 0 (65 min.)

2.5. Enzymatic hydrolysis All pretreated materials from the eleven CCD pretreatment conditions were submitted to enzymatic hydrolysis to comparatively evaluate the efficiency of each pretreatment. For comparison, the enzymatic hydrolysis of the untreated material was also carried out. The enzymatic hydrolysis of cellulose is catalyzed by the cellulase enzymes endoglucanase EC 3.2.1.4 and exoglucanase EC 3.2.1.91 that act synergistically with the enzyme β-glucosidase EC 3.2.1.21 and performs the complete hydrolysis of cellulose to glucose. Total cellulase activity is usually expressed as filter paper unit (FPU), defined by Ghose (1987) [21] and measured by the procedure described by the National Renewable Energy Laboratory [22]. The enzymatic hydrolysis experiments were done using a commercial powder cellulase preparation Power Cell (Prozyn, Brazil) that presents an FPase activity of 412.98 FPU g−1 of powder. The hydrolysis experiments were performed in 100 mL reaction mixtures presenting 0.2421 g of the enzyme powder (10 FPU g−1 of dry biomass), 10 g of dry pretreated or raw biomass, 50 mL of 100 mM sodium citrate buffer solution (pH 4.8) and the necessary amount of distilled water to a final biomass concentration of 100 g L−1. To validate the best pretreatment condition, the enzymatic hydrolysis experiments were performed using the Power Cell preparation and also a laboratory-made enzyme cocktail composed of the concentrated culture supernatants from Trichoderma reesei RUT C-30 and Aspergillus awamori [23,24] The validation tests were performed using biomass concentrations of 100 g L−1 and 200 g L−1 and enzyme load of 10 FPU g−1. The hydrolysis were carried out in flasks placed in a rotatory incubator shaker (New Brunswick Scientific model Innova 4230, New Brunswick, New Jersey, USA) at 50 °C and 200 min−1 that allowed the solid substrate to remain in suspension. Samples of 1.5 mL were taken at 24-, 48-, and 72 -hs and boiled for 10 min for enzyme inactivation before they were centrifuged at 7000 min.−1 for 10 min to separate the liquid glucose syrup from the solid lignin-rich fraction. Glucose concentration was measured in a YSI 2700 Biochemistry Analyzer (YSI, Ohio, USA) and the values prior to the enzymatic hydrolysis was measured and discounted from the final results. Xylose, galactose and arabinose concentrations were measured by HPLC.

experimental design to estimate the effect of each variable, the analysis of variance (ANOVA), and the response surfaces as well as the possibility of a mathematical model to be validated for each response. 2.4. Chemical characterization of spelt straw and the solid fraction from the HCW pretreatments Lignocellulosic materials are largely formed by the structural components cellulose, hemicellulose, and lignin besides nitrogen compounds, proteins, chlorophyll, waxes, and minerals [18]; these are usually quantified in both the raw and pretreated materials according to well-established procedures [18–20]. For the chemical characterization the spelt straw was firstly extracted by water and alcohol to determine the biomass extractives, using a Soxhlet. For that, 6 g of spelt straw was added to cellulose cartridges, and 190 mL of distilled water was added to the collecting flask. The electric heater was adjusted for about 4–5 extraction cycles per hour during 24 h. After the extraction the water with dissolved extractives was removed and replaced with 190 mL of alcohol for the second extraction step. At the end of this process, the cellulose cartridge containing the extracted straw was dried in a vacuum oven at 60 °C. The extractives content of the straw corresponded to the weight difference of the cartridge that was measured before and after the extractions, taking into account the moisture content. The extracted straw was submitted in triplicate to acid hydrolysis for the total hydrolysis of the cellulose and the hemicellulose. Approximately 0.3 g of dry material was placed in the hydrolysis flasks, added of 3 mL of 72% sulfuric acid and placed in a water bath at 30 °C for one hour under agitation with a magnetic stirrer. After this time interval, 84 mL of distilled water was added to reduce the acid concentration to 4% and the flasks were transferred to an autoclave, saturated with water vapor at 121 °C, for one hour. After cooling, the mixture was filtered with a glass fiber filter previously dried and weighed, and the solid lignin-rich fraction (IL) was dried and weighed. The liquid fraction was used for the spectrophotometric measurement of soluble lignin (SL). The total lignin corresponded to the sum of IL and SL. The liquid fraction was analyzed by HPLC to determine the concentrations of the biomass-derived monosaccharides. For the ash content determination, 1 g of spelt straw was placed in a previously dried and weighed ceramic crucible and transferred to a muffle furnace model Lindberg/Blue M (Thermo Scientific, Waltham Massachusetts, USA) at a temperature of 575 °C for six hours. After this period the crucibles were cooled and weighed to quantify the ash content based on the initial mass considering the moisture content. All the experiments were performed in triplicate. The same procedure was followed for the determination of the ash content in the IL residue to quantify the content of ash insoluble in acid and, by difference, estimate the insoluble lignin content. The final normalized chemical composition of the spelt straw was calculated considering the concentration of each monomeric sugar, the total lignin (IL + SL) content, the total extractives, and aches, based on the initial spelt straw mass considering

3. Results and discussion 3.1. Chemical composition of raw spelt straw and of the HCW pretreated spelt straw in different conditions Table 2 presents the chemical composition in terms of cellulose, hemicellulose, lignin, and ash for the raw spelt straw and of this material differently pretreated by HCW. The values represent the mean value and the standard deviation for three independent experiments. The spelt straw biomass with 14.9% moisture presented 38.25 ± 0.67% of cellulose, 24.28 ± 0.42% of hemicellulose, 14.77 ± 1.17% of lignin, 5.71 ± 0.08% of ash, and 7.14 ± 1.30% of extractives. The results show that for treatments at 120 °C for 65 min and 137 °C for 26 or 104 min the hemicellulose and the cellulose content of the pretreated material was comparable to that of the untreated material confirming that these are too mild pretreatment conditions. The temperature rise from 137 °C to 180 °C increased significantly hemicellulose removal so that the pretreated material yielded 8.02% to 1.85% residual hemicellulose in response to the treatment time from 10 to 120 min. The normalized cellulose content, in the range of 57.74% to 3

Biochemical Engineering Journal 151 (2019) 107340

R.d.R.O. de Barros, et al.

Table 2 Chemical composition of raw spelt straw and that of the pretreated materials in different HCW pretreatment conditions. HCW Condition

Raw spelt straw 1 (137 °C / 26 min.) 2 (223 °C / 26 min.) 3 (137 °C / 104 min.) 4 (223 °C / 104 min.) 5 (120 °C / 65 min.) 6 (240 °C / 65 min.) 7 (180 °C / 10 min.) 8 (180 °C / 120 min.) 9 (180 °C / 65 min.) 10 (180 °C / 65 min.) 11 (180 °C / 65 min.)

Cellulose (%)

Hemicellulose (%)

Lignin (%)

Std dev.

Mean.

Std dev.

Mean.

Std dev.

Mean.

Std dev.

Mean.

Std dev.

38.25 45.34 47.82 41.47 35.53 40.23 24.47 57.74 50.27 52.66 54.03 55.47

0.67 0.26 0.56 1.05 0.43 0.14 0.51 1.07 0.42 0.81 1.06 1.36

24.28 27.11 0.00 25.52 0.00 25.58 0.00 8.02 1.85 2.87 3.09 2.85

0.42 1.13 0.00 0.21 0.00 0.07 0.00 0.18 0.03 0.04 0.23 0.09

14.77 14.77 50.04 24.21 62.53 26.25 80.95 35.83 49.95 40.94 37.39 41.19

1.17 0.36 0.78 0.46 1.20 0.82 0.10 0.95 0.70 1.03 0.23 0.59

5.71 2.79 3.74 3.56 3.07 3.64 4.16 3.76 2.89 2.97 4.08 4.49

0.08 0.25 0.12 0.08 0.19 0.32 0.32 0.56 0.22 0.12 0.30 0.41

7.14 – – – – – – – – – – –

1.30 – – – – – – – – – – –

3.2. Enzymatic hydrolysis All the eleven samples of treated straw, as well as the raw spelt straw, were subjected to enzymatic hydrolysis and samples were taken for glucose concentration measurement at 24, 48 and 72 h. As results showed that in all cases glucose concentrations plateaued within 48 h the corresponding glucose concentrations were considered for the determination of the hydrolysis yields. Table 4 presents the mean value and the standard deviation for glucose concentration and glucose yield that was calculated based on the following equations.

Glucose yield (%) =

Glucose yield (%) =

Xylan (%)

Galactan (%)

Arabinan (%)

Mean.

Std dev.

Mean.

Std dev.

Mean.

Std dev.

20.92 23.63 0.00 22.69 0.00 21.04 0.00 8.02 1.85 2.87 3.09 2.85

0.45 0.31 0.00 2.04 0.00 2.21 0.00 0.18 0.03 0.04 0.23 0.09

0.74 0.76 0.00 0.00 0.00 0.69 0.00 0.00 0.00 0.00 0.00 0.00

0.02 0.20 0.00 0.00 0.00 0.06 0.00 0.00 0.00 0.00 0.00 0.00

2.62 2.47 0.00 1.56 0.00 2.40 0.00 0.00 0.00 0.00 0.00 0.00

0.03 0.19 0.00 0.18 0.00 0.26 0.00 0.00 0.00 0.00 0.00 0.00

mass of glucose released (g ) * 100 mass of total glucose available (g ) [Glucose released]* V 180

Cellulose content *( 162 )*[Biomass] * V

* 100

Where cellulose content corresponds to the absolute value presented on Table 2 for raw biomass and any pretreatment condition, i.e., the cellulose content of raw biomass was 0.3825. Biomass concentration was 100 g L−1, and the volume of enzymatic hydrolysis (V) was 0.1 L. The results in Table 4 show that the highest glucose concentrations in the range of 28 and 30 g L−1 and glucose yields from 46.11 to 50.31%, corresponded to pretreatments carried out at 180 °C from 10 to 120 min. The highest glucose yield of 60% was observed for the highest studied temperature of 240 °C for 65 min. However, in this condition, the biomass cellulose content dropped to 24.47% resulting in glucose syrup with 16.33 g L−1, which was 55% lower than the average concentration observed for the pretreatments carried out at 180 °C. Although, to our knowledge, there is no report on hydrothermal pretreatment of spelt straw, the hydrothermal pretreatment of wheat straw

Table 3 Chemical composition, as xylan, galactan and arabinan, of the hemicellulose of the raw spelt straw and that of the pretreated materials in different HCW pretreatment conditions.

Raw biomass 1 (137 °C / 26 min.) 2 (223 °C / 26 min.) 3 (137 °C / 104 min.) 4 (223 °C / 104 min.) 5 (120 °C / 65 min.) 6 (240 °C / 65 min.) 7 (180 °C / 10 min.) 8 (180 °C / 120 min.) 9 (180 °C / 65 min.) 10 (180 °C / 65 min.) 11 (180 °C / 65 min.)

Extractives (%)

Mean.

50.27%, indicated overall cellulose preservation at 180 °C. The normalized lignin content also increased in the range of 35.83% to 49.95% in response to the removal of hemicellulose. A further temperature increase to 223 °C for 26 min and 104 min resulted in the complete removal of the hemicellulose, nevertheless these conditions were detrimental to the normalized cellulose content that decrease to 47.82% for the 26-minutes treatment and 35.53% for the 104-minutes treatment. Regarding the treatment at 240 °C for 65 min, besides the expected complete hemicellulose removal, it was observed an even higher cellulose degradation whose normalized amount decreased to 24.47% alongside the normalized increase in the lignin content to 80.95%. As the preservation of the cellulose content is of paramount importance, the pretreatment conditions around the central points, i.e., 180 °C from 10 min to 120 min, favored a higher recovery of cellulose alongside a quite efficient removal of hemicellulose. These results indicated a higher influence of the pretreatment temperature in comparison to the pretreatment time. Nevertheless, these results were rather encouraging the full picture of the treatment effectiveness would depend on the glucose concentrations and yields upon the subsequent enzymatic hydrolysis step as well as on the inhibitors concentration in the pretreatment liquid stream. Table 3 presents the hemicellulose chemical composition, as xylan, galactan and arabinan for the raw spelt straw and that for the pretreated materials from the eleven HCW pretreatment conditions evaluated in CCD phase. Results show that the sugars galactan and arabinan were preserved at 120 °C for 65 min and 137 °C for 26 min, being fully degraded in all other studied conditions, indicating its higher lability in comparison to xylan.

HCW Condition

Ash (%)

Table 4 Glucose concentration (g L−1) and glucose yield (%) of raw and pretreated spelt straw within 48 h of enzymatic hydrolysis. Condition

Glucose 48h (g L−1) Mean.

Raw biomass 1 (137 ºC / 26 min.) 2 (223 ºC / 26 min.) 3 (137 ºC / 104 min.) 4 (223 ºC / 104 min.) 5 (120 ºC / 65 min.) 6 (240 ºC / 65 min.) 7 (180 ºC / 10 min.) 8 (180 ºC / 120 min.) 9 (180 ºC / 65 min.) 10 (180 ºC / 65 min.) 11 (180 ºC / 65 min.)

4

8.01 13.61 24.2 15.05 21.67 10.24 16.33 29.58 28.1 28.53 28.89 30.32

Std dev. 0.05 0.22 0.72 1.54 1.26 0.32 0.22 3.1 0.76 0.34 1.24 0.38

Glucose yield (%) Mean. 18.85 27.01 45.54 32.67 54.89 22.91 60.07 46.11 50.31 48.76 48.12 49.19

Biochemical Engineering Journal 151 (2019) 107340

R.d.R.O. de Barros, et al.

has been reported. Thus, a fixed temperature of 185 °C for 10 min, similar to a condition evaluated in the present work, was used to evaluate the effect of the pretreatment on lignin relocation on the pretreated material and its effect on the enzymatic hydrolysis [25]. In another study when optimization was performed considering response variables jointly, optimal conditions were 188 °C and 40 min., leading to hemicellulose recovery yield of 43.6% and cellulose enzymatic hydrolysis yield of 79.8%. Nevertheless these optimized conditions were not too harsh in comparison to our best conditions of 180 °C from 10 to 120 min, the study did report on the concentration of the glucose syrups that corresponded to the 79.8% cellulose hydrolysis yield and on the presence and concentration of inhibitors in the liquid current from the hydrothermal pretreatment considering the low recovery of this fraction, of 43.6%. According to the same study, when response variables were analysed separately, 71.2% of hemicellulose derived sugars recovery was obtained at 184 °C and 24 min., whereas, conditions of 214 °C and 2.7 min led to a maximum enzymatic hydrolysis yield of 90.6% of theoretical [26]. According to the foregoing and comparing our results to reports from the literature, xylan derived sugars from the wheat straw are possibly destroyed in temperatures higher than 180 °C, pending on the process time. However, as this condition does not have the necessary effect on the cellulose structure to support enzymatic hydrolysis with high glucose concentrations and yields, it is of paramount importance to identify an hydrothermal pretreatment temperature and time process condition that conciliates acceptable levels of hemicellulose degradation to an efficient cellulose enzymatic hydrolysis so that these valuable by products can be economically used at an industrial scale. Considering a scenario scenario where one third of the spelt straw would be used for the production of sugar syrups or ethanol, using the pretreatment and hydrolysis conditions from the present study, it would be possible to produce approximately 518 kg ha-1 y-1 of glucose or 33.5 L ha-1 y-1 of ethanol [13,14].

the conditions that associated hemicellulose removal with cellulose preservation (180 °C and 10 min) of 1.63 g L−1 and conditions 8 to 11 (180 °C, 65 min and 120 min) in the range of 2.62 to 3.38 g L−1 were all bellow inhibitory levels. For furfural concentration, the highest level corresponds to conditions 9, 10 and 11 which represents the central points of CCD reaching values between 2.38 and 3.98 g L−1, which are inhibitory for the yeast Saccharomyces cerevisiae that is severely affected in the presence of this compound at 2 g L-1 or higher [28]. However, the furfural concentration in condition 7 (180 °C, 10 min) was 0.67 g L−1 and therefore not inhibitory. The inhibition by HMF is not as severe as that for furfural, as HMF significantly inhibits the yeast growth and decreases ethanol fermentation above 5 g L−1, reaching full inhibition at 15 g L-1 [29]. In this study, the maximum HMF concentration observed corresponded to 0.96 g L−1 (condition 2) HMF, which individually would be harmful to a C5 fermentation process. Additionally, in condition 6, which is one of the most severe ones, a decrease in HMF concentration was observed, suggesting that this compound may have been degraded into the organic acids such as formic acid and levulinic acid [30]. 3.4. Statistical analysis All the mathematical models of DOE pretreatment experiments are second-order models since they consist of two independent input variables: (a) temperature (X1) and (b) pretreatment time (X2). The response variables (Yk) are: Y1 = Cellulose content in pretreated solid fraction (%); Y2 = Glucose concentration after 48 h of enzymatic hydrolysis (g L−1); Y3 = Glucose yield after 48 h of enzymatic hydrolysis (%); Y4 = Acetic acid concentration in pretreated liquid fraction (g L−1); Y5 = Furfural concentration in pretreated liquid fraction (g L−1); and Y6 = HMF concentration in pretreated liquid fraction (g L−1). In this way, it is possible to have a model composed of six terms for each output variable:

3.3. Microbial inhibitors concentration in HCW liquid fractions Table 5 shows the concentration of furfural resulting from the dehydration of C5 sugars, HMF resulting from the dehydration of C6 sugars, and acetic acid released from the hemicellulose chain in the liquid stream of spelt straw differently treated by HCW. It can be noticed that the concentration of the acetic acid varied around tenfold (from 0.44 to 4.27 g L−1) in tandem with the increase of the temperature from 137 °C to 240 °C. Although the pretreatment time, at the same temperature, increased the acetic acid release from the hemicellulose structure, temperature showed a more prominent effect on acetic acid release. In all cases, the acetic acid concentration did not reach the threshold inhibitory concentration of 6.00 g L−1 for yeast Saccharomyces cerevisiae, widely used for ethanol fermentation [27]. This is an exciting finding, as it would benefit 2 G ethanol production from both the C6 and the C5 sugars using a GMO yeast. The acid acetic concentration for

Yk = ak * X12 + bk * X22 + c k * X1 + dk * X2 + ek * X1 * X2 + fk The f term is fixed and comes from the center point, where both X1 and X2 assume 0 as value. The first step involves verifying if all of the other five terms significantly influence the solution or can be ignored through the p-value test using Pareto’s Diagram from Statistica 7.0. Given that the pretreatment was an experiment without an optimal reproducibility (mainly due to the limited instability of the temperature control of the reactor), we adopted a confidence interval of 95%. A confidence interval estimates how the variation of variables (in this case, parameters a, b, c, d, e, and f) are really due to statistical significance rather than experimental error or random variability; hence, in adopting a too high confidence interval, although the results extracted can be more accurate, there is the risk of cutting out variables that have a minor, albeit not a zero, significance for our experiment. In particular, for deciding which of the above terms should be present in the mathematical model, the software Statistica 7.0 runs the p-value test; this test gives an estimate between 0 and 1 of how probable it is that the variability between the 11 samples will be random or due to significance. Thus, if the p-value is lower than 0.05, we assume the term to be in the mathematical equation. After discarding the non-significant terms of each mathematical model, it is possible to obtain the following equations:

Table 5 Acetic acid concentration (g L−1) furfural concentration (g L−1) and HMF concentration (g L−1) in liquid fraction from spelt straw biomass pretreated by HCW in different conditions. Condition Raw biomass 1 (137 ºC / 26 min.) 2 (223 ºC / 26 min.) 3 (137 ºC / 104 min.) 4 (223 ºC / 104 min.) 5 (120 ºC / 65 min.) 6 (240 ºC / 65 min.) 7 (180 ºC / 10 min.) 8 (180 ºC / 120 min.) 9 (180 ºC / 65 min.) 10 (180 ºC / 65 min.)

Acetic acid (g L−1)

Furfural (g L−1)

HMF (g L−1)

– 0.44 3.81 0.69 3.98 0.25 4.27 1.63 3.38 2.62 3.01

– 0.01 1.9 0.03 0.5 0 0.2 0.67 3.96 3.38 3.98

– 0.01 0.96 0.01 0.89 0 0.85 0.06 0.58 0.26 0.33

Y1 = −11.0133 * X12 − 3.2200 * X1 − 3.3473 * X2 + 53.8980

Y2 = −8.39669 * X12 + 3.23423 * X1 + 28.49943 Y3 = − 4.83835 * X12 − 1.45855 * X22 + 11.67891 * X1 + 2.62276 * X2 + 48.70956 5

Biochemical Engineering Journal 151 (2019) 107340

R.d.R.O. de Barros, et al.

Table 6 Validity of mathematical models according to ANOVA and F-test. Triticum spelta HCW 180 ºC / 10 min.

Biomass concentration −1

100 g L 200 g L

−1

Enzyme Power cell T.reesei RUT C-30 + A. awamori Power cell T.reesei RUT C-30 + A. awamori

Glucose Conc. (g L−1) Yield (%) 29.77 ± 0.47 49.56 ± 0.79 31.33 ± 1.55 52.16 ± 2.59 56.33 ± 0.12 46.89 ± 0.10 55.90 ± 1.76 46.53 ± 1.46

Xylose Conc. (g L−1) Yield (%) 2.01 ± 0.12 60.28 ± 3.56 2.35 ± 0.32 70.39 ± 8.50 3.81 ± 0.31 56.96 ± 4.57 4.55 ± 0.36 68.13 ± 5.44

Y4 = 1.544996 * X1 + 0.363982 * X2 + 2.439455 Y5 = −1.55012 * X12 + 2.66855

Y6 = 0.079528 * X12 + 0.380798 * X1 + 0.084298 * X2 + 0.328607 In the equations above, the values used for X1 and X2 should be used in coded form (i.e., the values between -1.41 and +1.41). Besides the construction of the mathematical models for each of the output variables, the validity of each of the models was also verified based on analysis of variance (ANOVA) through the F-test. The null hypothesis is rejected if the F calculated (F-calc.) from the data is higher than the critical value of the F-distribution (F-dist.) for some desired false-rejection probability (e.g., 0.05). Table 6 below shows the validity of models according to 5 % of significance level. The data in Table 6 indicate that all six mathematical models can be counted as valid since the value of F was higher than the F value of the distribution for all output variables at a significance level of 5 %. Also, except for acetic acid concentration, all the mathematical models yielded a coefficient of determination (R2) greater than 88%. Since all the models can be considered valid, drawing on the previously discussed topics (e.g., the microbial inhibitors concentrations, glucose concentration in the enzymatic hydrolysis, and hydrolysis yields), some good options can be estimated for the pretreatment conditions in terms of temperature and pretreatment time due to the relevance of these parameters and the exclusion of unfavorable conditions.

Fig. 1. 3-D fitted surface of glucose concentration after 48 h of enzymatic hydrolysis as a function of temperature (ºC) and pretreatment time (min.).

activity in the yeast strain at concentrations higher than 2.0 g L-1 [27]. Given these observations, it can be considered that one of the possible best pretreatment conditions is located at a temperature of 180 °C and 10 min of pretreatment time. This condition corresponds precisely to the DOE experimental condition number 7 and conforms to the restriction conditions related to the maximum concentration of microbial inhibitors; it also shows a high concentration of glucose released in the enzymatic hydrolysis step as a result of high cellulose content in the pretreated solid fraction. In order to validate the best pretreatment conditions, additional enzymatic hydrolysis assays were performed using two different enzyme preparations and biomass concentrations of 100 g L−1 and 200 g L−1. The glucose and xylose concentrations and their yields after 48 h of enzymatic hydrolysis are shown in Table 7. Results of Table 7 show that the two fold increase on biomass concentration resulted in a proportional increase in the concentrations of glucose and xylose without any significant effect in the yield parameter. Moreover, it was observed equivalent concentrations and yields

3.5. Evaluation of HCW pretreatment best conditions Among all the output variables evaluated by the DOE, particular emphasis can be given to the glucose concentration from the enzymatic hydrolysis, since this parameter strongly depends on the cellulose content of the pretreated material, as a rich cellulose pretreated material will increase the possibility of a high glucose concentration released during the enzymatic hydrolysis step. On the other hand, considering the microbial inhibitors concentrations, it can be said that they favor the definition of restriction conditions. Indeed, if a specific pretreatment condition releases a high concentration of an inhibitory compound, the operating region in this condition may not be recommended. As an irrelevant parameter for the definition of an optimum pretreatment condition, it is possible to mention the glucose yield, since even conditions where both the cellulose content and the glucose are low can result in a higher glucose yield. Fig. 1 below shows the glucose concentration after 48 h of enzymatic hydrolysis as a function of temperature and pretreatment time. Fig. 1 illustrates that the highest glucose concentration occurs at temperatures close to 180 °C, and the pretreatment time does not exert a significant influence on this output variable and can be used in its minimum DOE value, which is equivalent to 10 min. The temperature value of 180 °C refers precisely to the value at the central point of this input variable in DOE; however, for the pretreatment time, the central value was 65 min. As the triplicate of the central point has already shown (experiments 9, 10, and 11), 65 min of pretreatment time at 180 °C generate high concentrations of inhibitory compounds, mainly furfural, which begins to exert inhibitory activity in the production of CO2 in values higher than 1.5 g L−1 and inhibitory

Table 7 Glucose and xylose concentrations and yields after 48 h of enzymatic hydrolysis of hydrothermal pretreated spelt straw at 180 °C and 10 min using different biomass concentrations and different enzyme preparations. Output variable Cellulose (%) Glucose concentration (g L−1) Glucose yield (%) Acetic acid concentration (g L−1) Furfural concentration (g L−1) HMF concentration (g L−1)

6

Significance (%)

F-calc.

F-dist.

R2 (%)

Validity

5 5

19.86 44.4

4.347 4.459

92.81 94.29

OK OK

5 5

16.26 6.78

4.534 5.117

94.25 55.22

OK OK

5

13.14

4.347

88.97

OK

5

39.82

4.459

93.68

OK

Biochemical Engineering Journal 151 (2019) 107340

R.d.R.O. de Barros, et al.

Table 8 Evaluation of output variables tested at 180 °C and 10 min of pretreatment time. Output variable

Model predicted value

Observed experimental value

Relative deviation

57.25 28.50 49.87 2.08 2.67 0.24

57.74 29.58 46.11 1.63 0.67 0.06

0.9 % 3.8 % −7.5 % −21.5 % −74.9 % −75.4 %

Cellulose content (%) Glucose concentration (g L−1) Glucose yield (%) Acetic acid concentration (g L−1) Furfural concentration (g L−1) HMF concentration (g L−1)

recovery. By doing that it was possible to fulfill a knowledge gap regarding a more complete biomass response towards different process conditions. Under the best pretreatment conditions, it was obtained a glucose syrup with 31.33 g L−1 in a reaction mixture presenting 100 g L−1 pretreated biomass, which corresponded to 52.16% glucan hydrolysis yield while for a reaction mixture presenting 200 g L−1 it was obtained a glucose sugar syrup with 56.33 g L−1, equivalent to a glucan hydrolysis yield of 46.89% The concentrations of acetic acid, furfural, and HMF in the pretreatment liquid fraction were 1.63 g L−1, 0.67 g L−1, and 0.06 g L−1, respectively, which were not inhibitory for the use of the biomass xylose syrup via biotechnological routes, e.g., for Saccharomyces cerevisiae ethanol fermentation. The mass balance under the best pretreatment condition, of 180 °C for 10 min, showed that 60% of the initial biomass was recovered as pretreated solids and around 20% as soluble solids that were distributed in the liquid fraction of the pretreatment as hemicellulose derived sugars (14.67%) and the pretreatment washing water (4.37%). The acidic hydrolysis of the liquid fraction showed that it presented 16 g L−1 of xylose whereas the washing water presented a xylose concentration of 4.6 g L−1. The best hydrothermal pretreatment conditions corresponded to 180 °C for 10 min. The use of statistical design of experiments for different pretreatment conditions enabled modeling mathematically two of the six output variables with relative deviation smaller than 5 %, as observed for the cellulose content in the solid pretreated straw as well as the glucose concentration after 48 h of enzymatic hydrolysis.

of glucose and xylose when the two enzyme preparations, the industrial Power Cell and the laboratory-made enzyme cocktail composed of the concentrated culture supernatants from Trichoderma reesei RUT C-30 and Aspergillus awamori were used in these validation tests. However the lab-made preparation showed a slightly better behavior for the biomass concentration of 100 g L−1, reaching 31.33 g L−1 of glucose, equivalent to 52.16% of glucose yield. At the higher biomass concentration of 200 g L−1 the commercial Power Cell, presented a slightly better performance allowing 56.33 g L−1 of glucose, equivalent to a glucose yield of 46.89%. The increase in the biomass concentration from 100 g L-1 to 200 g L−1 allowed a proportional increase in the glucose syrup concentration showing that the higher biomass concentration did not impose further mass transfer limitations. Although it is difficult to infer on the causes for the overall cellulose hydrolysis yields around 50% it is likely that the cellulose crystallinity that may not be significantly altered in the pretreated material was an important obstacle for a more extensive hydrolysis. Further studies aiming to compare the crystallinity of the treated and untreated spelt straw will be carried out as well as studies using higher enzyme loads and different cellulases preparations. The mass balance under the best pretreatment condition, of 180 °C for 10 min, showed that 60% of the initial biomass was recovered as pretreated solids and around 20% as soluble solids that were distributed in the liquid fraction of the pretreatment as hemicellulose derived sugars (14.67%) and the pretreatment washing water (4.37%). The acidic hydrolysis of the liquid fraction showed that it presented 16 g L−1 of xylose whereas the washing water presented a xylose concentration of 4.6 g L−1.

Acknowledgement This work received financial support from the brazilian Studies and Projects Financing Agency (FINEP), Grant ID 1421-08.

3.6. Mathematical model performance on best condition

References

For evaluating the functionality of the mathematical models obtained by the DOE in this session, the observed behavior and the predicted behavior of the six models were evaluated concerning one of the most favorable pretreatment conditions, which occurred at 180 °C for 10 min of pretreatment time. Table 8 presents the behavior of the models for six output variables under these conditions. Results presented in Table 8 showed that only two of the six mathematical models yielded deviations lower than 5 % in comparison to the value observed experimentally; after 48 h of enzymatic hydrolysis, the cellulose content and glucose concentration showed deviations of 0.9% and 3.8%, respectively. The mathematical models related to the concentration of furfural and HMF yielded deviations equivalent to 75% of the value observed experimentally.

[1] F. Perera, Pollution from fossil-fuel combustion is the leading environmental threat to global pediatric health and equity: solutions exist, Int. J. Environ. Res. Public Health 15 (2017) 1. [2] L. Wagner, I. Ross, J. Foster, B. Hankamer, Trading off global fuel supply, CO2 emissions and sustainable development, PLoS One 11 (2016) e0149406. [3] D. Kiš, N. Jovičić, A. Matin, S. Kalambura, S. Vila, S. Guberac, Energy value of agricultural spelt residue (Triticum spelta L.) forgotten cultures, Teh. Vjesn. - Tech. Gaz. 24 (Supplement 2) (2017). [4] B. Hames, R. Ruiz, C. Scarlata, A. Sluiter, J. Sluiter, D. Templeton, Preparation of Samples for Compositional Analysis: Laboratory Analytical Procedure (LAP) Technical Report NREL/TP-510-42620, (2008) (accessed 10 April 2019), https:// www.nrel.gov/docs/gen/fy08/42620.pdf. [5] H.V. Lee, S.B.A. Hamid, S.K. Zain, Conversion of lignocellulosic biomass to nanocellulose: structure and chemical process, The Scientific World Journal (2014) 631013. [6] A.S. da Silva, R.S. Sobral Teixeira, R.D. Oliveira, V. Santana, M. Antonieta Barros RdRO de, et al., Sugarcane and woody biomass pretreatments for ethanol production, in: A.K. Chandel, S.S.D. Silva (Eds.), Sustainable Degradation of Lignocellulosic Biomass: Techniques, Applications and Commercialization, InTech, Rijeka, Cr, 2013, pp. 47–88. [7] N. Mosier, C. Wyman, B. Dale, R. Elander, Y.Y. Lee, M. Holtzapple, et al., Features of promising technologies for pretreatment of lignocellulosic biomass, Bioresour. Technol. 96 (6) (2005) 673–686. [8] A.K. Kumar, S. Sharma, Recent updates on different methods of pretreatment of lignocellulosic feedstocks: a review, Bioresour. Bioprocess. 4 (7) (2017) 1. [9] K. Hammer, A.A. Filatenko, K. Pistrick, Taxonomic remarks on Triticum l. And ×Triticosecale wittm, Genet. Resour. Crop Evol. 58 (2011) 3–10. [10] G. Galterio, P. Codianni, A.M. Giusti, B. Pezzarossa, C. Cannella, Assessment of the

4. Conclusions The straw of the agroindustrial crop Spelt (Triticum spelta) was used for the production of biomass sugar syrups via hydrothermal pretreatment followed by enzymatic hydrolysis. The study was done using statistical experimental design that allowed a better understanding of the effect of the process conditions, temperature and time, on the biomass hemicellulose removal and deacetylation, on sugars degradation and the formation of 5-furfural and 5-hydroximethylfurfural, on cellulose preservation and response to enzymatic hydrolysis as well as lignin 7

Biochemical Engineering Journal 151 (2019) 107340

R.d.R.O. de Barros, et al.

[11]

[12]

[13] [14]

[15]

[16] [17]

[18]

[19]

[20]

fy08/42622.pdf. [21] T.K. Ghose, Measurement of cellulase activities, Pure Appl. Chem. 59 (2) (1987) 257–268. [22] B. Adney, J. Baker, Measurement of Cellulase Activities: Laboratory Analytical Procedure (LAP), (2008) (accessed 10 April 2019), https://www.nrel.gov/docs/ gen/fy08/42628.pdf. [23] L.M.F. Gottschalk, R.A. Oliveira, E.Pd.S. Bon, Cellulases, xylanases, β-glucosidase and ferulic acid esterase produced by Trichoderma and Aspergillus act synergistically in the hydrolysis of sugarcane bagasse, Biochem. Eng. J. 51 (2010) 72–78. [24] Enzyme Composition From Trichoderma reesei and Aspergillus awamori, (2008) (accessed 11 July 2019), https://patentimages.storage.googleapis.com/ac/28/f9/ 67301411fc6ea2/WO2009065199A1.pdf. [25] M.A.T. Hansen, J.B. Kristensen, C. Felby, H. Jørgensen, Pretreatment and enzymatic hydrolysis of wheat straw (Triticum aestivum L.) – the impact of lignin relocation and plant tissues on enzymatic accessibility, Bioresour. Technol. 102 (2011) 2804–2811. [26] J.A. Pérez, I. Ballesteros, M. Ballesteros, F. Sáez, M.J. Negro, P. Manzanares, Optimizing liquid hot water pretreatment conditions to enhance sugar recovery from wheat straw for fuel-ethanol production, Fuel 87 (2008) 3640–3647. [27] S. Larsson, E. Palmqvist, B. Hahn-Hägerdal, C. Tengborg, K. Stenberg, G. Zacchi, et al., The generation of fermentation inhibitors during dilute acid hydrolysis of softwood, Enzyme Microb. Technol. 24 (3–4) (1999) 151–159. [28] B. Sanchez, J. Bautista, Effects of furfural and 5-hydroxymethylfurfural on the fermentation of Saccharomyces cerevisiae and biomass production from Candida guilliermondii, Enzyme Microb. Technol. 10 (5) (1988) 315–318. [29] C.V.R. Baptista, Efeitos Inibitórios Do Hidroximetilfurfural Na Fermentação Alcoólica: Avaliação Por Citometria De Fluxo: Dissertação Para Obtenção Do Grau De Mestre Em Engenharia Alimentar, (2013) (accessed 10 April 2019), https:// www.repository.utl.pt/bitstream/10400.5/6697/1/Tese%20Versao%20Definitiva. pdf. [30] C.K. Nitsos, K.A. Matis, K.S. Triantafyllidis, Optimization of hydrothermal pretreatment of lignocellulosic biomass in the bioethanol production process, ChemSusChem 6 (1) (2013) 110–122.

agronomic and technological characteristics of Triticum turgidum ssp. dicoccum Schrank and T. spelta L, Nahrung 47 (2003) 54–59. M. Wiwart, M. Bytner, Ł. Graban, W. Lajszner, E. Suchowilska, Spelt (Triticum spelta) and Emmer (T. dicoccon) Chaff Used as a Renewable Source of Energy, BioResources 12 (2017) Nº 2. P.E. Marriott, L.D. Gómez, S.J. McQueen-Mason, Unlocking the potential of lignocellulosic biomass through plant science, New Phytol. 209 (4) (2016) 1366–1381. Markus Buerli, Farro in Italy - a Desk-study, (2007) http://www.underutilizedspecies.org/Documents/PUBLICATIONS/farro.pdf. B. Godin, S. Lamaudière, R. Agneessens, T. Schmit, J.-P. Goffart, D. Stilmant, P.A. Gerin, J. Delcarte, Chemical characteristics and biofuel potential of several vegetal biomasses grown under a wide range of environmental conditions, Ind. Crops Prod. 48 (2013) 1–12. World Food Situation - FAO Cereal Supply and Demand Brief, Food and Agriculture Organization of the United Nations, 2019 (accessed 10 July 2019), http://www.fao. org/worldfoodsituation/csdb/en/. G.E.P. Box, W.G. Hunter, J.S. Hunter, Statistics for Experimenters: an Introduction to Design, Data Analysis, and Model Building, Willey, New York, 1978. M.I. Rodrigues, A.F. Iemma (Eds.), Planejamento de experimentos e otimização de processos, 3rd ed., Casa do Espírito Amigo Fraternidade Fé e Amor, Campinas, 2014. A. Sluiter, B. Hames, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton, D. Crocker, Determination of Structural Carbohydrates and Lignin in Biomass: Laboratory Analytical Procedure (LAP) Technical Report NREL/TP-510-42618, (2012) (accessed 10 April 2019), https://www.nrel.gov/docs/gen/fy13/42618.pdf. A. Sluiter, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton, Determination of Extractives in Biomass: Laboratory Analytical Procedure (LAP) Technical Report NREL/TP-510-42619, (2008) (accessed 10 April 2019), https://www.nrel.gov/ docs/gen/fy08/42619.pdf. A. Sluiter, B. Hames, R. Ruiz, C. Scarlata, J. Sluiter, D. Templeton, Determination of Ash in Biomass: Laboratory Analytical Procedure (LAP) Technical Report NREL/TP510-42622, (2008) (accessed 10 April 2019), https://www.nrel.gov/docs/gen/

8