An Evaluation of Sonication Pretreatment for Enhancing Saccharification of Brewers’ Spent Grain

This paper deals with the investigation of ultrasound (US) pretreatment of brewer’s spent grain (BSG) as a means of releasing fermentable sugars, and the subsequent production of ethanol from this lignocellulosic biomass. Using response surface methodology (RSM), the influence of US power, time, temperature and biomass loading on fermentable sugar yield from BSG was studied. The optimal conditions were found to be 20% US power, 60 min, 26.3 C, and 17.3% w/v of biomass in water. Under these conditions, an approximate 2.1-fold increase in reducing sugar yield (325 ± 6 mg/g of biomass) was achieved, relative to untreated BSG (151.1 ± 10 mg/g of biomass). In contrast to acid or alkaline pretreatment approaches, the use of water obviated the need for neutralization for the recovery of sugars. The characterization of native and pretreated BSG was performed by HPLC, FTIR, SEM and DSC. Fermentation studies using S. cerevisiae growing on pretreated BSG resulted in a conversion of 66% of the total sugar content ininto ethanol with an ethanol content of 17.73 ± 2 g/ 100 g of pretreated BSG. These results suggest that ultrasound pretreatment is a promising technology for increased valorization of BSG as a feedstock for production of bioethanol, and points ton the need for further work in this area. 2020 The Authors. Published by Elsevier Ltd. This is an open access article under theCCBY license (http:// creativecommons.org/licenses/by/4.0/).


Introduction
According to a Brewers of Europe report published in 2017, Europe is the second largest beer producer in the world after China, with over 8490 breweries producing over 400,168 million hectoliters of beer in 2016 (BOE, 2017). Brewers' spent grain (BSG) is the most plentiful agro-industrial waste generated from the beerbrewing process, and about 3.4 million tonnes are produced annually in the EU (McCarthy et al., 2013). Spent grain is the insoluble part of the barley grain, separated during the mashing process before fermentation of the soluble liquid wort (Lynch et al., 2016). BSG represents about 85% of the total by-products, with about 2 Kg generated per 10 L of beer. BSG composition offers considerable options for developing value-added products, but the complex structure of lignin. and the high crystallinity of cellulose, reduces its accessibility to hydrolytic enzymes. This structure necessitates a pretreatment step prior to enzymatic hydrolysis.
Significantly, the EU aims to replace 10% of all transport fuel with biofuel by 2020. Such targets are an impetus to the investigation of the lignocellulosic waste utilization in order to achieve sustainable and profitable production of bioethanol. The current biofuel market is sensitive to sugar prices and supply shortages.
Previous studies have employed BSG for production of ethanol (sake) at 37 g/L over a 10 day fermentation using S. cerevisiae and A. oryzae (Wilkinson et al., 2017). Acid pre-treatment of BSG is a common approach in the literature. Plaza et al. (2017) pretreated BSG with sulfuric acid (15% w/w BSG, pH 1.0, 121°C), producing 75 g butanol/kg BSG. Rojas-Chamorro et al. (2018) optimized the phosphoric acid pretreatment for BSG, and reported recovery of 92% of total sugars from this substrate under harsh of 155°C and 2.0% H 3 PO 4 ; the authors also reported conversion of 69% of the total sugars in the BSG to ethanol using Escherichia coli. Such conventional methods for pretreatment have major drawbacks, including high energy consumption, the necessity to use harsh chemicals, production of inhibitors or low efficiency in large scale production (Hassan et al., 2018a).
The application of ultrasound technology for pretreatment of lignocellulosic biomass holds some potential for large scale processes (Hassan et al., 2018a) media results in cavitation phenomena, whereby microbubbles are formed which grow and then violently collapse at a critical size, converting sonic into mechanical energy. Such cavitation effects generate high temperature, pressure and violent shear forces, which lead to the formation of the 'hot-spot' effect in the liquid, and the generation of free radicals (Bundhoo and Mohee, 2018). Such effects may be postulated to aid in disruption of the lignocellulosic complex structure, decreasing the degree of crystallinity of the cellulose, and increasing the accessible surface area to improve enzymatic hydrolysis (Suresh et al., 2014).
Recently, chemical pretreatment methods combined with US technique have received some attention (Wang et al., 2018). However, very few reports discuss the ultrasound pretreatment of lignocellulosic material using water as solvent, and optimization of such processes has not been attempted. He et al. (2017) compared the ultrasound pretreatment (300 W, frequency of 28 kHz) of wood in soda solution, acetic solution, and distilled water. The authors reported an increase in sample crystallinity up to 35.3% in the case of soda solution, and up to 35.5% in the case of distilled water or acetic solution, which was due to physico-chemical-induced structural changes. However, the use of only water as a matrix for treatment, as opposed to conventional acid and alkaline approaches would eliminate the use of harsh chemicals, and the consequent problems with waste treatment.
The critical parameters for optimization of an ultrasoundmediated process for biomass pretreatment include US power (Song et al., 2015), biomass loading (Sasmal et al., 2012), treatment time, (Rehman et al., 2014;Sasmal et al., 2012;Song et al., 2015) and temperature (Rehman et al., 2014). Optimization of the pretreatment process by one-factor at a time in a multifactorial process is time-consuming and does not consider the possible interactive effects between the variables. Such limitation can be overcome by using Response Surface Methodology (RSM) as a means of statistical modeling and optimization of the multiple variables of pretreatment processes (Flores-Gómez et al., 2018).
The aim of the present study is to investigate the effect of US pretreatment on brewers' spent grain (BSG) in an aqueous medium with the goal of achieving maximum cellulose (glucose) and hemicellulose (xylose) recovery for bioethanol production, and maximum lignin recovery for use in a sugar-lignin platform biorefinery (Hassan et al., 2018b(Hassan et al., , 2019.

Materials
Brewer's spent grain (BSG) was generously donated by a local brewery in Dublin, Ireland. The BSG was dried at 60°C for 48 h followed by miling and sorting using a 350 mm sieve. The BSG was then stored in a dry place at ambient temperature for further experiments. Cellulase from Trichoderma reesei (aqueous solution, !700 units/g), and hemicellulase from Aspergillus niger (powder, 0.3-3.0 unit/mg solid) were purchased from Sigma Aldrich (Ireland), and were of analytical grade. The ethanol producing yeast, Saccharomyces cerevisiae Meyen ex E.C. Hansen (ATCC Ò 9763 TM ), was obtained from the TU Dublin microbiology repository. The culture was grown in slants at 30°C for 24 h in Yeast Peptone Dextrose (YPD) medium

Compositional analysis
Compositional analysis of BSG, before and after pretreatment, was carried out according to the National Renewable Energy Laboratory (NREL) protocol (Sluiter et al., 2008). In this method, BSG samples are subjected to acid hydrolysis using 72% (v/v) H 2 SO 4 at 30°C for 60 min, followed by a second step of acid hydrolysis by 4% (v/v) H 2 SO 4 , and then autoclaving at 121°C for 60 min. Post hydrolysis, the mixture was separated using a filtration crucible and dried at 105°C until constant weight was achieved. The dried solids were then ashed at 595°C for 24 h in a blast furnace to calculate the acid insoluble lignin (AIL) content, while the acid soluble lignin (ASL) fraction in the liquid was determined using UV-Vis absorbance at 205 nm. Total lignin content in each sample was calculated by combining the value of AIL and ASL. Furthermore, the reducing sugar concentration in the liquid was estimated by the dinitrosalicylic acid (DNS) method (Miller, 1959), while the detection and quantification of monosaccharides were performed by an Alliance Ò HPLC system (Waters, e2695 Separation module) using a Rezex ROA-Organic Acid H + (8%) column, (350 Â 7.8 mm; Phenomenex, UK) with 5 mM H 2 SO 4 as the mobile phase at 65°C, maintaining a flow rate of 0.6 mL/min. The HPLC system was equipped with a refractive index detector for the detection of monosaccharides.

Experimental design
For experimental design and optimization of ultrasound pretreatment of BSG, the Response Surface Methodology (RSM) based on Central Composite Design (CCD) was employed (Manorach et al., 2015). The effects of four pretreatment variables, each at five levels, on total sugar yield from BSG after saccharification were examined: biomass loading, temperature, ultrasound power and time. (Table 1), This used a CCD with five replicates at the center points, requiring 30 experiments. Experimental runs were randomized in order and carried out in triplicate. These pretreatment variables were further optimized using RSM. Experimental design was carried out using STATGRAPHICS Centurion XV software (StatPoint Technologies Inc. Warrenton, VA, USA).

Pretreatments
A total of 30 experiments as given by the experimental design were carried out to investigate the effect of varying conditions of ultrasound pretreatment on the enzymic hydrolysis of BSG (Fisher Scientific TI-H-10 Ultrasonic bath, output 750 W). After pretreatment, the biomass was collected, air-dried, and stored for further enzymatic hydrolysis, compositional analysis and characterization.

Enzymatic hydrolysis
To perform the saccharification, 1 g of BSG, 158.76 mL of cellulase (77 FPU/mL), and 153.3 mL hemicellulase (72 U/ mL) were mixed in sodium citrate buffer (0.05 M) and distilled water at pH 5.4 (total volume 10 mL) and heated to 50°C for 120 h (Ravindran et al., 2018). After treatment, the hydrolysate liquors were collected and stored at 4°C for further compositional analysis.

Model development and optimization
After conducting the experiments, a second-order polynomial regression model, as given by Eq. (1), was generated and analysed by the statistical software STATGRAPHICS Centurion XV software (StatPoint Technologies Inc. Warrenton, VA, USA) to define the response in terms of the independent variables.
where Y, b 0 , b i , b j , b ij , X represent process response (total sugar yield), linear coefficients, quadratic coefficients, interaction coefficients and coded independent variables (biomass loading, temperature, ultrasound power, and time), respectively. The regressors (b0, bi, bj, and bij) provide a quantitative measure of the significance of linear effects, quadratic of factors and interactions between factors.
The significant differences between each pretreatment with respect to the components of BSG were analyzed by performing analysis of variance (ANOVA) and multiple comparisons (Fischer's least significant difference test); Values of p < 0.05 was considered as significant.
2.2.6. Fermentation BSG hydrolysate (supplemented with all yeast extract peptone dextrosebroth medium ingredients except dextrose) was used for bioethanol fermentation as described by Kumar et al., 2011. A 24 h cell suspension of Saccharomyces cerevisiae (10 6 cfu/ml) was grown in shake flask culture (1.0 L, 50% working volume) at 30°C for 72 hrs. Gas Chromatography (Bruker Scion 456-GC coupled with flame ionization detector) was used to quantify the ethanol produced. Temperature of the injector, the detector and column oven were 220°C, 200°C and 160°C, respectively. The column oven initial temperature was 80°C and was set to increase temperature by 40°C/min and held for 7 min until the target temperature was achieved (Ravindran et al., 2017b).
2.2.7. Characterization 2.2.7.1. Scanning electron microscopy. The morphological structure of BSG before and after pretreatment was observed by performing field emission-scanning electron microscopy FE-SEM (Hitachi SU-70). Dried samples of the untreated and pretreated BSG were subjected to FE-SEM at an electron beam energy of 0.5 keV (Raghavi et al., 2016).
2.2.7.2. FTIR analysis. FTIR analysis was performed to observe any alteration in the composition of the BSG by assessing possible changes in functional groups after pretreatment and compared with native BSG. A Perkin Elmer Spectrum GX FTIR (UATR) Microscope (USA) was used for this part of the work (Raghavi et al., 2016). The FTIR spectra were recorded from 4000 to 400 cm À1 , with 32 scans at a resolution of 0.3 cm À1 in transmission mode.
2.2.7.3. Thermal behavior. The thermal behavior of BSG before and after pretreatment was studied and compared using differential scanning calorimetry (DSC). The thermal analysis instrument (Shimadzu DSC-60) was controlled by TA-60WS software. To perform the analysis, 55 mg of each BSG sample was placed in an aluminum pan, with an empty pan serving as a reference. The measurements were carried out between 25°C and 500°C, with a linear increase of 10°C per min (Ballesteros et al., 2014).

Compositional analysis
Polysaccharides were found to represent up to 53% of the dry weight of BSG, comprising xylan (38%), glucan (15%), and lignin (10%). This is largely in agreement with the values previously reported in the literature for this biomass type (Lynch et al., 2016). The hemicellulose and cellulose content of BSG as described in the literature varies between 19 and 40 % and 9-25% per dry matter, respectively (Xiros et al., 2008). This notwithstanding, the chemical composition of BSG is known to vary with barley cultivars, cultivation conditions, harvest time, malting practices, mashing regime, the presence of hops or adjuncts, and brewing conditions associated with lager and ale fermentations (Mussatto, 2014).

Modeling and optimization of ultrasound pretreatment
The total sugar yield from enzymatic hydrolysis was used in response surface methodology (RSM) to optimize the ultrasoundassisted pretreatment. The model adequacy was confirmed based on the coefficient of determination (R 2 ) and adjusted coefficient of determination (R 2 -adj). Accuracy range for R 2 is from 0 to 1, considering that a value closer to 1 means that the model is more accurate. A value of 98.28% R 2 was observed in the present work, while R 2 -adj was 96.68%, illustrating that the model adequately fits the data. The polynomial equation (2) describing the total sugar yield behavior is: Reducing sugar ðmg=mlÞ ¼ 2 :18129 À 0:1184 Â A À 0:009875 Â B À 0:0658167 The significance of the coefficients of the model was determined by ANOVA. The ANOVA table showed that 12 effects have P-values less than 0.05, indicating that they are significantly different from zero at the 95.0% confidence level, and implying a considerable effect of these coefficients on reducing sugar yield. The predicted levels of sugar yield in pretreated BSG using the Eq. (2) are given in Table 2, along with experimental data.
Additionally, response surface plots were generated as a function of two-factors-at-a-time in order to understand the principle interaction effects of different variables, and to determine the optimal level of each variable for maximum response (Fig. 1). A maximum reducing yield was observed with low ultrasound power (20%) and temperature (26°C), as well as high biomass loading (17 g/100 mL) and long pretreatment time (60 min). Under optimum conditions, the model predicted the maximum sugar yield to be 388 mg/mL of reaction volume.
After the optimized pretreatment of the native BSG, and following saccharification, 74% of sugars in BSG were recovered, while no degradation of lignin was observed (remaining as 10.6 g per 100 g   of raw BSG). Improvement in saccharification of BSG without lignin degradation may be attributed to the type of solvent used in this pretreatment step (water, with no chemicals), and the mechanoacoustic (physical) effects of ultrasound that increase surface erosion and pore size, and therefore the accessibility of biomass to hydrolytic enzymes (Bussemaker and Zhang, 2013). Obtaining high saccharification yields from lignocellulose, while maintaining a high amount of lignin after the pretreatment available for further valorization, can maximize the utilization of lignocellulose.
For the validation of the model, a confirmation experiment was conducted using the optimized parameters, and the experimentally obtained values of total reducing sugar amounted to 325 ± 1.0 mg/mL. Thus, since the experimentally obtained values of total reducing sugar from native BSG amounted to 151 ± 0.6 m g/mL, this model provided a 2.1-fold higher reducing sugar yield.
Ultrasonic frequency has a significant effect on ultrasound pretreatment, due to its influence on the critical size of the cavitation bubble. Lower frequency ultrasound (20-100 kHz) can produce more violent cavitation, resulting in higher localized temperatures and pressures at the cavitation site, as well as more effective shock waves (Terán Hilares et al., 2018); water molecules dissociate into oxidative radicals that oxidize and degrade organic molecules. Therefore, low frequency ultrasound is commonly used in biomass processing for intense physical effects, such as cell disruption (Tang and Sivakumar, 2015). Using similar low frequency ultrasound for pretreatment of biomass was reported by other researchers (Rehman et al., 2014) who used ultrasonication at an operating frequency of 20 kHz (power 750 W) and employing 20% of amplitude for enhancing sulfuric acid pretreatment of rice straw. Moreover, the results were consistent with reported data in the literature (Subhedar and Gogate, 2013) for newspaper pretreated with ultrasound (at frequency of 20 KHz) combined with alkali (NaOH) solution (concentration of 1 N). Eblaghi et al. (2016) also employed ultrasonic irradiation at a frequency of 35 kHz combined with alkaline solution (3% NaOH concentration) for pretreatment of sugarcane bagasse.
The ultrasonic treatment is also known to be influenced by solvent properties (Bussemaker and Zhang, 2013): water can produce more favorable conditions for cavitation than solvents with higher viscosity, surface tension or density. According to the literature, increases in biomass loading, temperature, sonication time, or the US power applied to the reaction mixture results in an increased rate of the reaction only up to a certain point; further increases in these parameters do not result in greater disruption effects (Bussemaker and Zhang, 2013;Karimi et al., 2014).

Fermentation
The hydrolysate obtained after enzymatic digestion of the ultrasound pretreated BSG (under the optimum conditions) was used for bioethanol production using S. cerevisiae, and gas chromatography was used to quantify the ethanol produced. An ethanol content of 17.73 ± 2 g/ 100 g of pretreated BSG was obtained after fermentation, compared to 11.53 ± 1.1 g/100 g for native BSG. The ethanol yield from this study was 88% of the theoretical yield (0.51 g ethanol/g sugars). Table 3 summarizes the findings of various authors regarding the production of bioethanol from BSG using different pretreatments. Ethanol production from ultrasound-pretreated BSG in this study is higher than that reported in earlier work using acid-pretreated BSG (15 g and 10.3 g / 100 g BSG; as reported by Rojas-Chamorro et al., 2017, andRaftopoulou, 2019, respectively.), while being similar to that obtained by recent research on BSG (17.9-18 g of ethanol per 100 g of BSG) using optimized chemical pretreatment (Rojas-Chamorro et al., 2018;   et al., 2019) and 18.8 g/100 g BSG using the recombinant Zymomonas mobilis strain 8b (Raftopoulou, 2019). It is also worth pointing out that using simultaneous saccharification fermentation (SSF) instead of the separated hydrolysis fermentation (SHF) may reduce the total time required for ethanol production from BSG as the time for enzymatic hydrolysis is not required (Rojas-Chamorro et al., 2017). However, authors reported that SHF gave highest value of ethanol productivity (0.94 g/L/h after 72 h enzymatic hydrolysis + 24 h fermentation) as compared to SSF (0.595 g/L/h without xylanases after 72 h fermentation, and 0.829 g/L/h with xylanases after 30 h fermentation). This preliminary finding suggests the potential of ultrasound pretreatment in a water solvent system for BSG degradation, thereby dispensing with the need for acid-alkali additions.

Characterization
3.4.1. Scanning electron microscopy SEM analysis of the raw (control) and pretreated BSG revealed surface modifications in response to pretreatment (Fig. 2). The untreated BSG had an undulating and crumbled surface as a result of the brewing process. The ultrasound pretreated BSG showed a porous surface of uneven and non-uniform cavities. This may be due to the modifications in the external fibers arising from the effect of the cavitation due to ultrasound waves (Ravindran et al., 2017a). Gabhane et al. (2014) compared alkali-microwave and alkali-sonication pretreatment of lignocelluosic biomass, and reported that the latter resulted only in fibrillation of the cell wall of the pretreated biomass; however, microwave pretreatment led to complete tissue collapse. Ultrasound pretreatment was found to increase the total surface area of BSG that was exposed to enzyme activity, and enhanced the enzyme accessibility.

FTIR analysis
The chemical changes in the functional groups of the BSG were studied based on FTIR analysis (Fig. 3). The pretreated BSG displayed significant decreases in band intensities at characteristic peaks for cellulose (3290 Cm À1 ), hemicellulose (1030 Cm À1 ) and lignin (1240 Cm À1 ) (Table 4). Similar reduction in intensity of peaks representing functional groups of BSG after ultrasoundassisted pretreatment of lignocellulose were reported by Ravindran et al. (2017a). These observations indicate that ultrasound treatment disturbed the lignocellulosic structure of BSG, increasing the yield of the reducing sugar.

Differential scanning calorimetry
Differential scanning calorimetry determines the difference in the heat flow associated with heating, cooling or isothermal conditions of the sample, compared with a reference, and as a function of temperature. The DSC thermogram (Fig. 4) represented the thermal characteristics of the native and the pretreated BSG between 20°C and 500°C, which were obtained at a heating rate of 10°C/min.  Both native and pretreated BSG exhibited a similar trend in their thermo-gram profile, suggesting that they were approximately similar in their composition. An exothermic event was evident between a temperature range of 20-300°C. This temperature range was associated with several processes which gave rise to compounds such as carbon monoxide, carbon dioxide and other pyrolysis products. In addition, an endothermic event was observed between a temperature range of 300-500°C. Both native and pretreated BSG exhibited a thermal behaviour that included a crystallization peak at the temperature of 370°C.

Conclusion
The optimal conditions for ultrasound-assisted pretreatment of BSG in aqueous medium were found to be 20% US power, 60 min, 26.3°C, and 17.3% w/v of biomass. BSG pretreatment under the optimal ultrasonication conditions resulted in a 2.1-fold higher reducing sugar yield, relative to native BSG. These results suggest that ultrasound pretreatment is a viable technology for increased valorization of BSG, and further work is needed in this area.
E-supplementary data of this work can be found in online version of the paper.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.