Green and clean process to obtain low degree of polymerisation xylooligosaccharides from almond shell

Journal of Cleaner Production 241 (2019) 118237

Contents lists available at ScienceDirect

Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro

Green and clean process to obtain low degree of polymerisation xylooligosaccharides from almond shell Ramkrishna D. Singh a, b, c, Cresha Gracy Nadar c, Jane Muir a, b, Amit Arora a, c, * a

Indian Institute of Technology Bombay-Monash Research Academy, Indian Institute of Technology Bombay, Maharashtra, 400076, India Department of Gastroenterology, Central Clinical School, Monash University, Melbourne, Victoria, 3004, Australia c Bio- Processing Laboratory, Centre for Technology Alternatives for Rural Areas, Indian Institute of Technology Bombay, Maharashtra, 400076, India b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 2 May 2019 Received in revised form 27 August 2019 Accepted 29 August 2019 Available online 6 September 2019

This work presents a green and chemical free sequential process consisting of autohydrolysis, enzymatic treatment, and membrane assisted refining for the valorization of almond shell into the low degree of polymerisation xylooligosaccharides. For autohydrolysis, the temperatures (180, 200, and 220
C) with different reaction times were evaluated. Further, enzymatic treatment of the autohydrolysate was performed to increase the concentration of low degree of polymerisation xylooligosaccharides. For enzymatic treatment, three different doses (5, 10, and 15 U) was used, and the optimum dose estimated using statistical analysis. Finally, the XOS rich enzyme liquor was subjected to membrane assisted refining using 1 kDa and 250 Da membranes to obtain XOS concentrate. Under the optimal condition (200
C, 5 min) of autohydrolysis, about 54.5% of xylan could be obtained as oligosaccharides. However, the autohydrolysate was composed of 3.5% (w/w of biomass) of low degree of polymerisation xylooligosaccharides (xylobiose and xylotriose). The enzymatic treatment using 10 U of the enzyme could increase the concentration of low degree of polymerisation xylooligosaccharides to 8.2% (w/w of biomass). Finally, the membrane-assisted refining could recover 69.1 ± 0.1% (w/w) of produced xylooligosaccharides. © 2019 Elsevier Ltd. All rights reserved.

Handling Editor: Panos Seferlis Keywords: Autohydrolysis Enzymatic hydrolysis Membrane purification Xylooligosaccharides

1. Introduction Annually, abundant lignocellulosic residues are generated from agricultural and allied practices. These residues are incinerated either as fuel, to recover their calorific value, or to eliminate them. Such practices lead to environmental pollution and also represent a loss of valuable biomolecules. To add value, efforts such as biogas production from catering and crop residues (Anjum et al., 2016), bioethanol from pineapple leaf (Chintagunta et al., 2017) or sunflower accessions (Brazil et al., 2019), vanillin from lignin of bamboo (Harshvardhan et al., 2017), or bio-manure from food waste (Du et al., 2018) has been reported. Recently, purposing of lignocellulosic biomass for cellulosic ethanol production has seen a rise (Pourbafrani et al., 2014). However, in such plants, a hemicellulosic sugar side stream is currently discarded as waste (Zhang et al., 2014). The life-cycle and environmental impact assessment

* Corresponding author. Bio- Processing laboratory, Centre for Technology Alternatives for Rural Areas, Indian Institute of Technology Bombay, Maharashtra, 400076, India. E-mail address: [email protected] (A. Arora). https://doi.org/10.1016/j.jclepro.2019.118237 0959-6526/© 2019 Elsevier Ltd. All rights reserved.

for obtaining value-added products from waste streams of sugar beet pulp industry or woody biomass support valorization of such
lez-García et al., 2016; Gonzalez-Garcia et al., 2018). streams (Gonza Similarly, for cellulosic ethanol to achieve the desired commercial success and generate a bio-based economy, the hemicellulose fraction should be utilized to obtain bio-based products (Wang et al., 2017). The application of hemicellulosic sugar can include the production of biofuel or sweetener and as a starting material for xylooligosaccharides (XOS), a prebiotic. Thus, in comparison to other prebiotics (fructooligosaccharides, inulin, galactooligosaccharides) which are either obtained from limited natural sources or synthesized chemically, XOS is obtained from abundantly available lignocellulosic residues. The XOS can be produced from biomass either via alkaline pretreatment or autohydrolysis. The alkaline pretreatment process consumes multiple chemicals and generates a liquor containing dissolved lignin, phe€nsson and Martín, 2016). The liquor is difficult nolics, and acids (Jo to handle and requires downstream purification or leads to waste generation. On the other hand, autohydrolysis is a green process for XOS production; wherein water acts as the reactant. The

2

R.D. Singh et al. / Journal of Cleaner Production 241 (2019) 118237

autoionization of water at a higher temperature generates hydronium ion which catalyzes the cleavage of glycosidic linkages resulting in depolymerisation of xylan backbone. Also, the deacetylation of xylan backbone liberates acetic acid, which contributes to making the reaction medium acidic (Surek and Buyukkileci, 2017). This is also known as autohydrolysis. The autohydrolysis process has been used to obtain XOS from substrates such as Miscanthus x giganteus (Chen et al., 2014), vine pruning (Jesus et al., 2017), almond shell (Nabarlatz et al., 2005), and palm empty fruit bunches (Ho et al., 2014). A difference in XOS yield could be observed due to the diversity in feedstock composition and experimental conditions. However, in most of the work, a wide degree of polymerisation (DP) XOS product was reported. For XOS, the prebiotic activity is thought to be dependent on its DP. The XOS with DP 2e4 has been reported to be preferentially fermented by the beneficial bacterium. Also, the fermentability is
n et al., 2008a; observed to decrease with an increase in DP (Gullo Moura et al., 2007; Wang et al., 2010). In addition, it has been reported that for XOS product of wide DP range, the bacterium utilizes DP 2e4 more preferentially (Chen et al., 2016). Thus, available literature suggests that low DP range XOS have better prebiotic potential. As narrow or low DP XOS production is challenging with autohydrolysis alone, researchers have suggested an increase in autohydrolysis reaction time. However, this approach can cause degradation of oligosaccharides into monosaccharide and other impurities (Surek and Buyukkileci, 2017). Also, upon enzymatic treatment of the autohydrolysate of palm (Chapla et al., 2012) and softwood (Deloule et al., 2017), the authors reported higher levels of monosaccharides. In this work, we have aimed at developing a chemical-free process to produce low DP XOS (for this work considered as the sum of xylobiose and xylotriose). The autohydrolysis was conducted in the severity range of 3.36e4.23, to achieve maximum xylan hydrolysis while preventing/limiting the formation of degradation products. Further, endoxylanase was used for enzymatic treatment to increase the concentration of low DP XOS and prevent excess monosaccharide formation. Finally, membrane assisted bio-separation was conducted for refining of produced XOS liquor. Thus, the work presents a process to produce an industrially and health-relevant prebiotic from the almond shell generated as residue from the nut industry. Annually, 1.4 million metric tonnes of the shell are generated (www.fas.usda.gov), which in the absence of valorization may contribute to environmental concerns. Such waste can be obtained from point sources such as nut processing industry. Though their valorization to cellulosic ethanol may not be economically viable due to limited scale, alternatives such as production of prebiotic xylooligosaccharides may hold incentives.

2. Material and methods

2.2. Autohydrolysis pretreatment of shell powder The autohydrolysis pretreatment was conducted in 1L SS 316 pressure reactor with provision for water circulation, and the temperature was controlled using a proportional integral derivative (PID) controller (Trident Labotek, India). For each treatment, the material was loaded into the reactor at 1:10 w/v ratio (AS: water) and heated to reach the desired temperature and held at that temperature for the required time. After the heating period, cold water was circulated through the mixture to decrease the temperature and halt the reaction. The mixture was then filtered to collect the autohydrolysate. The autohydrolysate was centrifuged at 5000 rpm for 10 min to obtain clear supernatant and stored at 4
C until further use. An aliquot was analysed by HPLC to estimate the percentage of xylobiose, xylotriose, and monosaccharide content. Another aliquot was subjected to post-hydrolysis using sulphuric acid to estimate the oligosaccharide content (Sluiter et al., 2008). The parameter under study includes 180
C (10, 20, 40 min); 200
C (5, 10, 15 min) and 220
C (2.5 and 5 min). 2.3. Enzymatic hydrolysis of pretreated liquor The clear autohydrolysate was treated with endoxylanase from Thermomyces lanuginosus (expressed in Aspergillus oryzae) at 50 ± 2
C (temperature decided based on preliminary experiments) after adjusting pH to 5.5 ± 0.2. For enzymatic hydrolysis, the autohydrolysate was incubated with the desired enzyme dose (5U, 10U and 15U) at 70 rpm. At regular intervals (4, 8, 12, 24, 36, 48 h) an aliquot was withdrawn and heated in a boiling water bath for 5 min to inactivate the enzyme. The aliquot was then filtered using a 0.2 m filter and analysed for low DP XOS using HPLC. The outcome of enzymatic treatment was subjected to Analysis of Variance (ANOVA) test. 2.4. Membrane assisted refining of XOS liquor The enzymatically treated liquor was refined using membranes having molecular weight cut-off (MWCO) of 1 kDa and 250 Da membrane (47 mm) in an HP 4750 Stirred Cell (Sterlitech Corporation, USA). The membranes before use were preconditioned using HPLC grade water at the pressure designated for the membrane. For filtration, 90 ml of liquor was added to the vessel provided with a Teflon coated magnetic bar to prevent concentration build-up. The first step of filtration was performed using 1 kDa membrane at 689.5 kPa (using nitrogen gas) until about 96e97% of feed could permeate. For the next step, the liquor was passed through 250 Da membrane at 1034.2 kPa, until about 65% feed was permeated. The ml of permeate collected was recorded at regular interval and used to determine the flux. The volume concentration ratio (VCR) to estimate a concentration level was calculated using the following formula (equation (1))

2.1. Raw materials and chemical composition analysis VCR ¼ VF/VR The powdered almond shell (AS) was purchased from the local market, Maharashtra, India. The powder was dried to constant moisture content and packed in airtight containers. All other chemicals were procured from Merck (USA) and used without further purification. Double distilled water was used for all experiments. The membranes were purchased from Permionics Membrane Pvt Ltd. India. The extractives (%) of the shell was determined by sequential extraction using hexane, ethanol, and water using a Soxhlet extractor. The percentage glucan, xylan, and acid insoluble lignin was determined by following the National Renewable Energy Laboratory (NREL) protocol (Sluiter et al., 2012).

(1)

where, VF is the initial feed volume, and VR is the retentate volume. An aliquot of the permeate and retentate was analysed by HPLC for its composition and compared to feed. The composition of feed and permeate was then used to determine the rejection (R) by using the following formula (equation (2)): R ¼ [(Cf-Cp)/Cf]  100

(2)

where Cf is the concentration of the constituent in feed, Cp is the concentration of the component in the permeate.

R.D. Singh et al. / Journal of Cleaner Production 241 (2019) 118237

2.5. Analytical methods The HPLC (Infinity 1260, Agilent, USA) analysis was performed using Hi-Plex H column (300*7 mm) maintained at 65
C and attached to the autosampler. The sugars, oligosaccharides and acetic acid were eluted using 5 mM sulphuric acid in water at 0.7 ml/min and detected using RID (50
C). The concentration of xylobiose, xylotriose, monosaccharides, and acetic acid was estimated by comparison with the area of pure standards. 2.6. Statistical analysis The statistical analysis of data was performed using Minitab Statistical Software version 16 (Pennsylvania State University, USA). Two-way ANOVA and Tukey's post hoc analysis was used to determine the significance of the results of enzymatic treatment. All experiments were performed in triplicate and values expressed as average ± s.d. 3. Results and discussion 3.1. Almond shell lignocellulose composition The lignocellulosic composition of the almond shell was 34.3 ± 0.4% glucan, 20.2 ± 0.2% xylan, and 28.8 ± 2.1% acid insoluble lignin. The almond shell was found to be low in extractives, with about 0.8 ± 0.1% n-hexane soluble extractives, 1.0 ± 0.1% alcohol soluble, and 5.48 ± 0.1% water soluble extractives. 3.2. Autohydrolysis pretreatment 3.2.1. Effect on xylan hydrolysis The parameters evaluated for autohydrolysis of almond shell includes temperature and reaction time. The average time required to reach the desired temperature was 16 min, 20 min, and 24 min for 180
C, 200
C, and 220
C, respectively. The pressure attained during the reaction was 8e9 kg/cm2 for 180
C, 12e14 kg/cm2 for 200
C and 18e20 kg/cm2 for 220
C. At the end of the reaction, the mixture was cooled to 60
C by cold water circulation (within 5e7 min) to halt the hydrolysis. The percentage xylan depolymerized and resulting in oligosaccharides is as shown in Fig. 1a. At 180
C, the produce oligosaccharide was observed to be 32.6 ± 0.8% of initial xylan for 10 min of reaction time. As observed from Fig. 1a, upon a further increase in the reaction time at 180
C, causes no significant increase in the oligosaccharides production. A low solubility of xylan under certain autohydrolysis condition can result in low oligosaccharide yield (Surek and Buyukkileci, 2017). This can be the reason for low yield at 180
C even when reaction time was increased. The highest xylan hydrolysis (54.5 ± 1.0%) was observed at 200
C and 5 min. However, a further increase in reaction time at the same temperature resulted in a decrease in oligosaccharide percentage. The percent xylan hydrolysed was 50.1 ± 1.1% and 41.5 ± 1.2% for 10 and 15 min, respectively. A further increase in temperature to 220
C also results in a decrease in the oligosaccharides percent, with 39.0 ± 1.7% at 2.5 min and 23.1 ± 2.6% at 5 min of reaction time. For autohydrolysis pretreatment, severity factor has been commonly used to observe the effect of experimental conditions on the outcome. The severity factor (Ro), presents an equation which combines the effect of temperature (oC) and time (min) and is expressed as shown in equation (3): Ro ¼ t*exp((T-Tref) / u)

(3)

where, T is the set temperature, Tref is the reference temperature

3

(373.15 K) and u is the empirical parameter representing activation energy which is required for hemicellulose solubilization. The activation energy (u) is calculated using the following formula (equation (4))

u ¼ RTref / Ea

(4)

where, R is ideal gas constant (8.314 J/mol K), Ea is the activation energy of hemicellulose (J/mol). The literature suggests a value of 14.75 K for hemicellulose solubilization (Jesus et al., 2017). The log severity factor in this work was in the range of 3.3e4.2. As can be observed from Fig. 1b, the percentage of xylan hydrolysis into oligosaccharides increases with an increase in severity factor up to 3.9, after which the value decreases. The data presented in Fig. 1a and b suggests that reaction temperature and time, which can be translated into severity factor, are crucial to achieving higher oligosaccharide yields. At a constant temperature, with an increase in the reaction time, the extent of xylan hydrolysis may increase. However, as suggested by the data, there is an optimum temperature and time beyond which the xylan hydrolysis may increase but oligosaccharides concentration decrease. This finding can be corroborated by analysis of data presented in Fig. 1b. For similar Log Ro values of 3.94, 3.93, and 3.96, the highest oligosaccharide concentration was observed when the reaction temperature was 200
C (Log Ro ¼ 3.94). Thus, for almond shell, the optimum temperature can be considered as 200
C. The data also suggest an optimum reaction time of 5 min as a further increase in reaction time at 200
C causes a decrease in oligosaccharides concentration. This decrease can be attributed to further hydrolysis of oligosaccharides into monosaccharide and degradation product as the reaction proceeds. The data for 220
C suggest that a temperature above 200
C is not favourable even when the reaction time is reduced. Thus, in this work, the optimum severity factor can be considered as 3.64, corresponding to 200
C and 5 min. Under this condition, about 54.5% of xylan could be hydrolysed to produce 10.97% oligomers (based on dry biomass). As autohydrolysis results in cleavage of hemicellulose sugar backbone, the concentration and composition of released soluble oligosaccharides depend on the reaction temperature and time. The total XOS (sum of xylobiose and xylotriose) and content of xylobiose and xylotriose produced upon autohydrolysis is as presented in Fig. 2. At 180
C, the liquor was composed of 1.9 ± 0.1% xylobiose at 10 min, which increases to 2.2 ± 0.1% at 40 min. The concentration of xylotriose was found to be in the range of 1.2e1.3% in the same time range. When the reaction temperature was 200
C, the yield of xylobiose was 2.1 ± 0.1% at 5 min, which increases to 3.1 ± 0.1% at 15 min of reaction time. The xylotriose yield was in the range of 1.1e1.6%, which was almost similar at each reaction time. The yield of xylobiose at 220
C, 2.5 min was observed to be 2.6 ± 0.1% and 2.5 ± 0.2% at 5 min reaction time, whereas the xylotriose yield was 1.3e1.4%. 3.2.2. Composition of autohydrolysis liquor Upon autohydrolysis, the cleavage of susceptible bonds on xylan backbone can release monosaccharides, acids, and their degradation products. Hence, the autohydrolysate was composed of sugars such as xylose, glucose, and arabinose resulting from the backbone of hemicellulose. The severity factor and corresponding sugar content are as shown in Table 1. The xylose content was found to increase with an increase in the severity factor and was observed to be in the range of 2.4e3.4%. Similarly, glucose could be detected in the autohydrolysate at 2.5e3.9%. The presence of glucose in liquor can be ascribed to the possible glucose substituent on xylan backbone and/or hydrolysis of the amorphous region of cellulose. However, the low concentration of glucose at all severity factor

4

R.D. Singh et al. / Journal of Cleaner Production 241 (2019) 118237

Fig. 1. The effect of reaction condition on biomass (a) effect of temperature (180, 200 and 220
C) and holding time on XOS yield, (b) influence of severity factor (log Ro) upon xylan hydrolysis and generation of oligosaccharides (% of xylan) where blue triangle represents data of 180
C, black square represents 200
C and red circle represents 220
C. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

indicates minimal glucan hydrolysis, which is advantageous when the objective is to produce XOS. The content of arabinose was found to be nearly similar for all studied concentration and might be originating from hemicellulose structure. Also, as seen in Table 1, the amount of acetic acid produced was dependent on the reaction severity and increased with increase in reaction temperature and time. The amount of acid produced at different reaction condition is reflected in the variation of liquor pH (Table 1). The lignin content of biomass is considered to be resistant to autohydrolysis pretreatment and thus undergo limited to none degradation or hydrolysis. However, certain susceptible groups such as phenolics may be cleaved upon autohydrolysis. The total phenolics content of the autohydrolysate liquor was found to be in the range of 0.15e0.28 mg/100 g biomass, thus indicating a negligible breakdown of lignin. The concentration of furfural in the

autohydrolysate liquor was also found to be very low in concentration (below 1 mg/ml). Under the autohydrolysis condition, the oligosaccharides can be hydrolysed into monosaccharides and ultimately degrades to furfural and hydroxymethyl-furfural (Ruiz et al., 2013). However, the data in this work suggest that the experimental conditions were not severe, and hence, low levels of monosaccharides and degradation products were produced. 3.3. Enzymatic hydrolysis As mentioned, upon autohydrolysis, about 10.97 g of xylan was solubilized into XOS. However, the concentration of low DP XOS (xylobiose and xylotriose) was 3.3 ± 0.1 g/100 g biomass. Thus, to maximize the concentration of low DP XOS, the autohydrolysate was treated with the endoxylanase. During preliminary runs, it

R.D. Singh et al. / Journal of Cleaner Production 241 (2019) 118237

5

Fig. 2. The percentage of xylobiose and xylotriose and corresponding XOS (sum of xylobiose and xylotriose) produced upon autohydrolysis at different experimental conditions.

Table 1 The composition of monosaccharides, acetic acid and corresponding pH of autohydrolysate upon autohydrolysis pretreatment. Experimental Condition Temperature (oC)

Time (min)

180 180 180 200 200 200 220 220

10 20 40 5 10 15 2.5 5

Severity factor (log Ro)

Xylose

Glucose

Arabinose

Acetic acid

pH

3.36 3.66 3.96 3.64 3.94 4.12 3.93 4.23

2.5 ± 0.1 2.6 ± 0.2 2.8 ± 0.1 2.5 ± 0.1 2.6 ± 0.1 3.0 ± 0.1 3.3 ± 0.1 3.4 ± 0.1

2.5 ± 0.2 2.6 ± 0.1 3.0 ± 0.1 3.7 ± 0.1 3.8 ± 0.2 3.9 ± 0.2 3.8 ± 0.1 3.9 ± 0.2

2.7 ± 0.1 2.8 ± 0.1 2.5 ± .01 2.9 ± 0.1 2.6 ± 0.3 2.4 ± 0.2 2.5 ± 0.2 2.5 ± 0.1

0.4 ± 0.1 0.6 ± 0.1 1.2 ± 0.1 0.6 ± 0.1 0.8 ± 0.2 1.4 ± 0.1 1.1 ± 0.2 1.6 ± 0.1

4.31 4.01 3.72 3.84 3.56 3.42 3.35 3.21

Monosaccharide and acetic acid are in g/100 g biomass.

could be seen that the enzyme shows decreased activity at the pH of autohydrolysate (pH below 4). As the optimum pH of the enzyme is 5.5e5.8, the pH of autohydrolysate was adjusted to 5.5 using 2 M NaOH. As observed from Fig. 3, the concentration of low DP XOS increases upon enzymatic treatment. The maximum XOS yield (8.3 ± 0.1%) was observed at an enzyme dose of 15 U and 36 h of reaction. However, the yield was not significantly different (p < 0.05) from that obtained at 10 U (8.2 ± 0.1%). Thus, 10 U of the enzyme can be considered optimal for producing maximum low DP concentration. Under the optimal conditions, the liquor was composed of 5.3 ± 0.1% xylobiose and 3.0 ± 0.1% xylotriose as compared to 2.2% xylobiose and 1.4% xylotriose present in the autohydrolysate. The enzymatic liquor was also composed of 3.3 ± 0.2% xylose and 0.9% acetic acid, majorly arising due to autohydrolysis treatment. The Two-way ANOVA test presents an Rsq value of 0.99 and adjusted R-sq of 0.98, indicating a good agreement for the obtained results. Table 2 presents the results of Two-way ANOVA for XOS yield at different enzyme dose. As can be observed, the individual factor (enzyme dose and reaction time) had a significant influence on XOS yield. Among the two independent factors, enzyme dose has a higher effect on the XOS yield as depicted by higher F value when compared to reaction time. However, the effect of the interaction term was not significant (at p < 0.05). The ANOVA test for XOS yield at each sampling time was performed to determine the reaction period from where a significant difference in XOS yield appears for different enzyme dose. The test reveals that similar XOS yields were observed at 4 h for all three

enzyme doses. However, from 8 h, it was observed that the highest XOS yield was obtained with 15 U, though it was not significantly different from that obtained at 10 U. Also, a significantly lower yield was obtained with 5 U of the enzyme. Thus, the enzyme dose of 10 and 15 U give similar XOS yield after 8 h of hydrolysis. The main and interaction effects plots are as shown in Fig. 4a and b, respectively. Fig. 4a shows that as the enzyme dose increases from 5 U to 10 U, the yield of XOS increases with time, reaching a maximum at about 36 h of the enzymatic reaction. Fig. 4b supports the results of the ANOVA test indicating no interaction among the independent factors. The contour plot as shown in Fig. 4c helps to visualize the effect of a change in enzyme dose and reaction time on XOS yield. The plot suggests that a reaction time above 35 h and enzyme dose above 7.5 U enables to produce above 8 g/100 g XOS from biomass. Based on the data, it can be concluded that under the optimal condition for enzymatic hydrolysis (10 U, 36 h) the concentration of the low DP XOS could increase from 3.5 ± 0.1% (in autohydrolysate) to 8.2 ± 0.1%. Importantly, the obtained liquor has a narrow spectrum of DP. Thus, in this work, 10.97 g of almond shell xylan could be hydrolysed containing 31.9% low DP XOS, however, this could be increased to 76.8% upon enzymatic treatment. 3.4. Membrane separation of enzymatic XOS liquor The enzyme-treated liquor was subjected to sequential membrane-assisted refining to concentrate the low DP

6

R.D. Singh et al. / Journal of Cleaner Production 241 (2019) 118237

Fig. 3. Enzymatic treatment of autohydrolysate to increase concentration of low DP XOS (sum of xylobiose and xylotriose) at different enzyme doses (5 U, 10 U and 15 U).

oligosaccharides. The XOS liquor was purified using ultrafiltration and nanofiltration with membranes having MWCO of 1 kDa and 250 Da, respectively. The properties of the membranes are as described in Table 3. An aliquot of the feed, permeate and retentate for each membrane was analysed by HPLC to estimate the composition. The schematic of the sequence of membrane filtration and the flux of the pure water and liquor using 1 kDa and 250 Da membrane is as shown in Fig. 5. The first step was to reject high molecular weight components if any, including enzyme, using 1 kDa membrane while allowing low DP XOS, monomeric sugars, and acetic acid to permeate. The resulting permeate (P1) from 1 kDa membrane was subsequently filtered through 250 Da membrane to concentrate low DP XOS in the retentate. The pure water flux for 1 kDa membrane (Javg ¼ 93 LMH) was determined by measuring the cumulative permeate volume using distilled water as a feed. The permeate flux of liquor declined from 55.4 to 30.1 LMH in 100 min of operation with the 1 kDa membrane. This decline could be due to concentration polarization, and membrane fouling in the stirred cell. Since enzymatic hydrolysate was centrifuged, most of the suspended solids were removed which could have helped in reduced fouling during membrane separation. The selection of the membrane MWCO was based on the molecular weight of the XOS and monosaccharides. We hypothesized that since low DP XOS (DP2, DP3) molecular weights are above 282 g/mol, 250 Da membrane should be able to retain these constituents while allowing passage of most of the monosaccharides and acetic acid (molecular weight less than 180 g/ mol) in permeate stream (P2). Thus, subsequent fractionation of P1 stream using tighter cut-off (250 Da) membrane could allow the concentration of low DP XOS in the retentate. For 250 Da, the permeate flux decreased gradually from 38.1 to 27.1 LMH in 60 min Table 2 Analysis of Variance test for experimental response upon enzymatic hydrolysis. Source

DF

Seq SS

Adj SS

Adj MS

F

P

Enzyme dose Time Enzyme*Time Error Total R-Sq R-Sq (adj)

2 5 10 18 35 99.01% 98.08%

1.4042 22.0464 0.0966 0.2347 23.7819

1.4042 22.0464 0.0966 0.2347

0.7021 4.4093 0.0097 0.0130

53.86 338.23 0.74

0.000 0.000 0.679

*p value < 0.05 indicates significance.

of operation, during which 31e33% of feed was retained as retentate (R2). Table 4 represents the mass balance of filtration process, indicating the percentage of xylobiose, xylotriose, xylose, glucose, arabinose, and acetic acid in the initial feed and permeate or retentate. In the first step, about 93e95% of xylobiose, xylotriose, and monosaccharides (xylose, glucose, arabinose) along with 87% acetic acid could be recovered in the permeate fraction. Thus, the rejection (R) of constituents on 1 kDa membrane was as follows: xylobiose: 7%; xylotriose: 5.3%; xylose: 6.4%; glucose: 5.9%; arabinose: 5.9% and acetic acid: 13%. Hence, the collection of most of the feed as permeate allows for maximum recovery of smaller molecular weight components. Next, when the P1 was passed through 250 Da membrane, the concentration of xylobiose increased from 4.9 mg/ml (feed) to 13.7 mg/ml in 250 Da retentate (R2) with a VCR ~ 3.0 and 85% recovery. Similarly, the concentration of xylotriose changed from 2.8 mg/ml to 3.8 mg/ml, with a recovery of about 42%. It was observed that negligible xylobiose and xylotriose were detected in 250 Da permeate (P2). Based on mass balance data, it was speculated that the loss of some portion of xylobiose and xylotriose may be due to thin layer formation of solids on the membrane. This was subsequently confirmed through SEM (Scanning electron microscopy) analysis (data not shown). The rejection (%) for 250 Da membrane was 54% for xylose, 38.7% for glucose, 40.8% for arabinose and 42.9% for acetic acid. In other words, the percent removal of xylose was 46% along with 61.3% glucose, 59.2% arabinose, and 57.1% acetic acid. The permeate (P2) volume represents about 62% of feed volume containing monosaccharides and acetic acid. Since enzymatic treatment of autohydrolysate was performed at low solid-liquid ratio, the solids and monomers concentration in the P2 stream was low to meet the requirement of economical fermentation of sugars for any value addition. The total solids content based on mass balance was less than 0.3%. Therefore, it would be prudent to recycle P2 in the process to meet the water requirement of autohydrolysis step, thereby potentially reducing the fresh water requirement by 50%. However, further studies are warranted using this approach to determine the actual number of P2 stream recycling, as the sugar and acid concentration will increase with each cycle. Thus, upon fractionation of feed using sequential membrane filtration process, a concentrated low DP XOS (xylobiose and xylotriose) source containing 69.1 ± 0.1% of the initial XOS (after

R.D. Singh et al. / Journal of Cleaner Production 241 (2019) 118237

7

Fig. 4. Statistical analysis of enzymatic treatment of autohydrolysate on XOS yield (%) of biomass): (a) main plot (b) interaction plot and (c) contour plot.

Table 3 Configuration of membranes used in this work. Properties

HPA 1 kDA

HPA 250 Da

Form Membrane material Membrane area (cm2) Molecular weight Cut-Off (Da) Temperature range (oC) pH range Operation pressure (MPa)

Flat sheet Hydrophilized Polyamide 17.34 1000 Ambient to 50 4e9 1e1.37

Flat sheet Hydrophilized Polyamide 17.34 250 Ambient to 50 4e9 1e1.37

enzymatic hydrolysis) could be obtained. Recently, XOS refining and purification has been reported for Miscanthus XOS with alcohol: water mixture as the eluting solvent (Chen et al., 2016). This process will require an additional energy-intensive unit operation for alcohol recycling. The membrane separation process developed in the current work provides a greener approach for obtaining XOS syrup rich in the low DP XOS. For further purification of low DP XOS, dia-filtration and hybrid membrane-resin based systems could be used, but these would require detailed analysis and process optimization. The final decision needs to be taken based on the techno-economic feasibility of the process. 4. Conclusion In this work, the green and chemical free sequential process comprising of autohydrolysis, enzymatic reaction, and membrane separation is proposed for the valorization of the almond shell into low DP XOS. Under the optimal conditions, 10.97 g oligosaccharides

containing 3.5 ± 0.1% low DP XOS was obtained. The concentration could be further increased to 8.2 ± 0.1 g/100 g biomass with enzyme treatment. Finally, the filtration step could recover 69.1 ± 0.1% of produced XOS containing 85% of xylobiose and 42% of xylotriose while removing most of the monosaccharides and acetic acid. Also, upon filtration, a low DP XOS syrup with 57% purity was obtained with 2.8  (w/v) and 1.38  (w/v) increase in concentration of xylobiose and xylotriose, respectively. As compared to single step autohydrolysis process, the two-step process described here allows for production of a narrow and low DP XOS syrup. The process also simplifies the downstream processing for refining of liquor. Thus, the work presented here demonstrated the feasibility of a sequential green process for efficient production of low DP XOS from almond shell residue. Declaration of interest None.

8

R.D. Singh et al. / Journal of Cleaner Production 241 (2019) 118237

Fig. 5. The figure presents (a) the schematic of membrane assisted filtration process and data recording and (b) the pictorial representation of the sequence of membrane used for filtration to obtain XOS concentrate where the observed Flux (J ¼ L/hm2) of water and XOS liquor for 1 kDa and 250 Da membrane are indicated as graphs. P1: Permeate from 1 kDa, P2: Permeate from 250 Da, R1: Retentate from 1 kDa and R2: Retentate from 250 Da.

R.D. Singh et al. / Journal of Cleaner Production 241 (2019) 118237 Table 4 Mass balance of components upon membrane assisted refining of XOS liquor. Component

Xylobiose Xylotriose Xylose Glucose Arabinose Acetic acid

Feed

441.1 ± 5.5 248.7 ± 2.2 299.8 ± 5.7 317.2 ± 0.6 265.3 ± 3.3 7.7 ± 0.2

1 kDa

250 Da

Permeate

Retentate

Permeate

Retentate

397.8 ± 5.0 238.0 ± 2.1 277.1 ± 5.3 306.3 ± 0.6 253.3 ± 3.2 6.7 ± 0.2

18.5 ± 0.2 7.8 ± 0.1 10.5 ± 0.2 10.6 ± 0.1 8.8 ± 0.1 0.2 ± 0.0

N.D N.D 137.9 ± 2.6 194.5 ± 0.4 157.1 ± 2.0 4.4 ± 0.1

375.0 ± 4.7 104.7 ± 0.9 149.8 ± 2.9 103.6 ± 0.2 104.1 ± 1.3 2.5 ± 0.1

Values are in mg, mean ± s.d. Volume of feed: 90 ml, Volume of 1 kDa permeate: 87 ml, Volume of 1 kDa retentate: 3.3 ml, Volume of 250 Da permeate: 55.96 ml and Volume of 250 Da retentate: 27.34 ml.

Acknowledgment The authors thank Indian Institute of Technology BombayMonash Research Academy to provide financial support (IMURA0444) to RDS for his doctoral study. This work did not receive any specific funding from either public, private, or nonprofit organizations or institutions. References Anjum, M., Khalid, A., Mahmood, T., Aziz, I., 2016. Anaerobic co-digestion of catering waste with partially pretreated lignocellulosic crop residues. J. Clean. Prod. https://doi.org/10.1016/j.jclepro.2015.11.061.
Brazil, O.A.V., Vilanova-Neta, J.L., Silva, N.O., Vieira, I.M.M., Lima, A.S., Ruzene, D.S., Silva, D.P., Figueiredo, R.T., 2019. Integral use of lignocellulosic residues from different sunflower accessions: analysis of the production potential for biofuels. J. Clean. Prod. https://doi.org/10.1016/j.jclepro.2019.02.274. Chapla, D., Pandit, P., Shah, A., 2012. Bioresource Technology Production of xylooligosaccharides from corncob xylan by fungal xylanase and their utilization by probiotics. Bioresour. Technol. 115, 215e221. https://doi.org/10.1016/j.biortech. 2011.10.083. Chen, M.H., Bowman, M.J., Dien, B.S., Rausch, K.D., Tumbleson, M.E., Singh, V., 2014. Autohydrolysis of Miscanthus x giganteus for the production of xylooligosaccharides (XOS): kinetics, characterization and recovery. Bioresour. Technol. 155, 359e365. https://doi.org/10.1016/j.biortech.2013.12.050. Chen, M.H., Bowman, M.J., Cotta, M.A., Dien, B.S., Iten, L.B., Whitehead, T.R., Rausch, K.D., Tumbleson, M.E., Singh, V., 2016. Miscanthus  giganteus xylooligosaccharides: purification and fermentation. Carbohydr. Polym. 140, 96e103. https://doi.org/10.1016/j.carbpol.2015.12.052. Chintagunta, A.D., Ray, S., Banerjee, R., 2017. An integrated bioprocess for bioethanol and biomanure production from pineapple leaf waste. J. Clean. Prod. https://doi. org/10.1016/j.jclepro.2017.07.179. Deloule, V., Chirat, C., Boisset, C., Toussaint, B., Chroboczek, J., 2017. Production of hemicellulose oligomers from softwood chips using autohydrolysis followed by an enzymatic post-hydrolysis. Holzforschung. https://doi.org/10.1515/hf-20160181. Du, C., Abdullah, J.J., Greetham, D., Fu, D., Yu, M., Ren, L., Li, S., Lu, D., 2018. Valorization of food waste into biofertiliser and its field application. J. Clean. Prod.

9

https://doi.org/10.1016/j.jclepro.2018.03.211.
lez-García, S., Gullo
n, B., Rivas, S., Feijoo, G., Moreira, M.T., 2016. EnvironGonza mental performance of biomass refining into high-added value compounds. J. Clean. Prod. https://doi.org/10.1016/j.jclepro.2016.02.015.
n, B., Moreira, M.T., 2018. Environmental assessment of Gonzalez-Garcia, S., Gullo biorefinery processes for the valorization of lignocellulosic wastes into oligosaccharides. J. Clean. Prod. https://doi.org/10.1016/j.jclepro.2017.02.164.
n, P., Moura, P., Esteves, M.P., Girio, F.M., Domínguez, H., Parajo
, J.C., 2008. Gullo Assessment on the fermentability of xylooligosaccharides from rice husks by probiotic bacteria. J. Agric. Food Chem. 56, 7482e7487. https://doi.org/10.1021/ jf800715b. Harshvardhan, K., Suri, M., Goswami, A., Goswami, T., 2017. Biological approach for the production of vanillin from lignocellulosic biomass (Bambusa tulda). J. Clean. Prod. https://doi.org/10.1016/j.jclepro.2017.02.125. Ho, A.L., Carvalheiro, F., Duarte, L.C., Roseiro, L.B., Charalampopoulos, D., Rastall, R.A., 2014. Production and purification of xylooligosaccharides from oil palm empty fruit bunch fibre by a non-isothermal process. Bioresour. Technol. 152, 526e529. https://doi.org/10.1016/j.biortech.2013.10.114. Jesus, M.S., Romaní, A., Genisheva, Z., Teixeira, J.A., Domingues, L., 2017. Integral valorization of vine pruning residue by sequential autohydrolysis stages. J. Clean. Prod. https://doi.org/10.1016/j.jclepro.2017.08.230. €nsson, L.J., Martín, C., 2016. Pretreatment of lignocellulose: formation of inhibitory Jo by-products and strategies for minimizing their effects. Bioresour. Technol. https://doi.org/10.1016/j.biortech.2015.10.009. Moura, P., Barata, R., Carvalheiro, F., Gírio, F., Loureiro-Dias, M.C., Esteves, M.P., 2007. In vitro fermentation of xylo-oligosaccharides from corn cobs autohydrolysis by Bifidobacterium and Lactobacillus strains. LWT - Food Sci. Technol. (Lebensmittel-Wissenschaft -Technol.) 40, 963e972. https://doi.org/10.1016/j.lwt. 2006.07.013.
, D., 2005. Autohydrolysis of almond shells for the Nabarlatz, D., Farriol, X., Montane production of xylo-oligosaccharides: product characteristics and reaction kinetics. Ind. Eng. Chem. Res. 44, 7746e7755. https://doi.org/10.1021/ie050664n. Pourbafrani, M., McKechnie, J., Shen, T., Saville, B.A., Maclean, H.L., 2014. Impacts of pre-treatment technologies and co-products on greenhouse gas emissions and energy use of lignocellulosic ethanol production. J. Clean. Prod. https://doi.org/ 10.1016/j.jclepro.2014.04.050. Ruiz, H.A., Rodríguez-Jasso, R.M., Fernandes, B.D., Vicente, A.A., Teixeira, J.A., 2013. Hydrothermal processing, as an alternative for upgrading agriculture residues and marine biomass according to the biorefinery concept: a review. Renew. Sustain. Energy Rev. https://doi.org/10.1016/j.rser.2012.11.069. Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., 2008. Determination of Sugars, Byproducts, and Degradation Products in Liquid Fraction Process Samples. National Renewable Energy Laboratory. Sluiter, a., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., Crocker, D., 2012. NREL/TP-510-42618 analytical procedure - determination of structural carbohydrates and lignin in Biomass. Lab. Anal. Proc. 17. Surek, E., Buyukkileci, A.O., 2017. Production of xylooligosaccharides by autohydrolysis of hazelnut (Corylus avellana L.) shell. Carbohydr. Polym. 174, 565e571. https://doi.org/10.1016/j.carbpol.2017.06.109. Wang, J., Sun, B., Cao, Y., Wang, C., 2010. In vitro fermentation of xylooligosaccharides from wheat bran insoluble dietary fiber by Bifidobacteria. Carbohydr. Polym. 82, 419e423. https://doi.org/10.1016/j.carbpol.2010.04.082. Wang, Y.S., Zhu, X., Jin, L.Q., Zheng, Y., Liao, C.J., Zheng, Y.G., 2017. Post-hydrolysis of the prehydrolysate from eucalyptus pulping with xylanase. J. Clean. Prod. https://doi.org/10.1016/j.jclepro.2016.10.183. Zhang, L., You, T., Zhang, L., Li, M., Xu, F., 2014. Comprehensive utilization of waste hemicelluloses during ethanol production to increase lactic acid yield: from pretreatment to fermentation. Biotechnol. Biofuels 7, 494. https://doi.org/10. 1186/s13068-014-0189-4.