Pilot scale process for polysaccharide extraction and fractionation from cereal by-products

(1)

INOM

EXAMENSARBETE BIOTEKNIK, AVANCERAD NIVÅ, 30 HP

STOCKHOLM SVERIGE 2017,

Pilot scale process for

polysaccharide extraction and fractionation from cereal by- products

RESKANDI CHASTELIA RUDJITO

KTH

SKOLAN FÖR BIOTEKNOLOGI

(2)

Master Thesis

Pilot scale process for polysaccharide extraction and fractionation from cereal by-products

Reskandi Chastelia Rudjito

May 26, 2017

Supervisors: Assoc. Prof. Francisco Vilaplana and Dr. Andrea C. Ruthes Division of Glycoscience

School of Biotechnology Royal Institute of Technology

(3)

ABSTRACT

Wheat bran is a low-valued agricultural by-product that has the potential to be valorised into commercial goods through the optimisation of biorefinery processes. Wheat bran is rich in hemicelluloses, of which feruloylated arabinoxylans (F-AX) are the most abundant.

The presence of ferulic acid provides the arabinoxylans (AX) with antioxidant properties hence generating great interest for the development of bio-based materials. It has been shown that F-AXs can be extracted with high yields by means of subcritical water extraction at 160 °C pH 7 using water or formate buffer. In this study, the effect of pretreatment (consisting of not treating, defatting, destarching as well as defatting and destarching the wheat bran) towards the extraction of F-AX was evaluated at small scale.

An optimised pretreatment condition was then validated at pilot scale while varying the solid to liquid ratio, of wheat bran to solvent. The extraction of both small and pilot scale was carried out sequentially summing up 5, 10, 15 and 30 minutes. The results revealed the destarched wheat bran extracted with tap water (pH 7.1) generated the highest total solid yield of 46.7 % DW, AX content of 83.3 %, high molar mass distribution of 10⁵ Da and maintained presence of ferulic acid at 0.17 – 0.68 mg/g DW. This condition was later validated at pilot scale, nonetheless a difference in performance was observed. In the pilot scale, the total solid yields decreased to 27.9 % DW while the AX content dropped to 57.9

% for the destarched sample of the same solid to liquid ratio (1:17). An increase in ratio to 1:10 did not show any statistical significance. The small scale extraction appeared to have run more efficiently as a majority of the glucose content was extracted in the first 5 minutes followed by optimum extraction of AXs between 10 to 30 minutes. While in the pilot scale, the extraction seems to be delayed as it required 30 minutes in order for the AX population to be enriched. Interestingly however, the molecular characteristic of the AXs from the pilot scale exhibited the same molar mass distribution (10⁵ Da) with the presence of ferulic acid at 0.63 mg/g DW as in the small scale. Hence, further optimization of the extraction process at pilot scale can be performed to aim for obtaining larger amounts of feruloylated arabinoxylans that can be explored into various industrial applications.

Keywords: wheat bran, arabinoxylans, ferulic acid, subcritical water extraction

(4)

SAMMANFATTNING

Vetekli är en lågvärdig biprodukt från spannmålsproduktion som har stor potential för att utnyttjas i avancerade produkter genom optimering av bioraffinaderi processer. Vetekli är rik på hemicellulosor, mest feruloylerade arabinoxylaner (F-AX). Feruloylsyra ger arabinoxylaner (AX) antioxidantegenskaper, vilket erbjuder stort intresse för utvecklingen av biobaserade material. F-AX kan extraheras med höga utbyten genom användning av subkritisk vattenextraktion vid 160°C pH 7 i laboratorieskala. I denna studie utvärderade vi effekten av förbehandling mot extraktion av F-AX i laboratorieskala.

Ett optimerat förbehandlingsförhållande validerades efteråt i pilotskala genom att variera förhållandet mellan fast kli och vatten. Utvinningen av både laboratoriet och pilotskalor utfördes under 5, 10, 15 och 30 minuter. Resultaten avslöjade att behandlade vetekli utom stärkelse extraherad med kranvatten (pH 7,1) erbjöd högsta totalt utbyte och AX-halt, med hög molekylvikt (över 10⁵ Da) och bibehållna feruloylsyra. Extraktionsförhållande validerades senare i pilotskala. Vi observerade dock en skillnad i prestanda, med en minskning av det totala utbytet och AX-halten för samma kli:vatten förhållande (1:17). En ökning av förhållandet mellan fast kli och vatten till 1:10 visade ingen statistisk relevans.

Extraktion vid pilotskala verkar försenade i jämförelse med laboratorieskala, för vid 30 minuter observeras ett extrakt med hög AX halt och med samma molekylära egenskaper som i liten skala (liknande molekylvikt över 10⁵ Da med närvaron av feruloylsyra).

Genom att finjustera pilotextraktionen kommer stora mängder feruloylerade arabinoxylaner förhoppningsvis att extraheras och vidareutvecklas i framtiden.

Nyckelord: vetekli, arabinoxylan, feruloylsyra, subkritisk vattenextraktion

(5)

I INTRODUCTION

1.1 Background

The global effort to implement sustainable solutions in replacing the use of fossil resources has driven the development of biorefinery systems. In particular, second generation biorefinery systems aim to valorise low-valued agricultural waste or by- products into valuable goods by means of environmentally friendly processes ^[1]. In retro respect, wheat bran is a by-product that is continuously produced during the milling of white flour ^[2] More than a hundred million tons of wheat bran is produced annually worldwide ^[3] thus making it an interesting feedstock for biorefinery.

Wheat bran makes up almost a quarter of the whole kernel structure and is composed of approximately 60% carbohydrates ^[4]. Among the carbohydrates, the main hemicellulose that comprise wheat bran polysaccharides is arabinoxylan (AX). Arabinoxylans originating from wheat bran are often substituted with ferulic acid ^[5] and this gives feruloylated AXs bioactive properties, hence generating avenues for novel bio-based materials.

Henceforth, in order to valorise arabinoxylans into potential commercial products, these polysaccharides must first be extracted from the wheat bran and fractionated. A series of processes can be employed to obtain AXs with differing degrees and patterns of substitution and molecular weights [6]. Maintaining a high molecular weight of the AX polymer and the ferulic acid bound to arabinose residues has been a great focus at the Glycoscience division at KTH. Ruthes et al (2017) has recently developed an extraction and fractionation process to obtain AXs with such properties. The process utilizes water at subcritical conditions combined with ultrafiltration and a post-enzymatic treatment.

The extraction step using water at subcritical conditions will be up-scaled and is the main focus of this project.

1.2 Objectives

1. Design and implement an upscale process for the extraction and fractionation of wheat bran into potentially valuable polysaccharides and oligomers

2. Evaluate the performance of the biorefinery process in lab scale and pilot scale 3. Determine the operating parameters of the biorefinery process in pilot scale that

produces the most desirable outcome.

1.3 Scope

The emphasis of this project lies heavily on the validation and optimization of the subcritical water extraction system at pilot scale. This step in the process is the basis of AXs extraction and thus requires thorough investigation from bench scale to pilot scale.

(8)

1.4 Hypothesis

The biorefinery process at pilot scale should produce a comparable outcome to that of the bench scale study (high AX yield and content with maintained ferulic acid linked to the AXs).

(9)

II LITERATURE

2.1 The biorefinery concept

In definitive terms, biorefinery is the means of producing valued goods such as biopolymers, platform chemicals, transportation fuels and energy from renewable biomass ^[7]. Renewable biomass in this case can originate from agricultural harvest, lignocellulose residue or feedstocks of algae of which are categorized into first, second and third generation biorefinery, respectively ^[1]. The main advantage of biorefinery systems, which are analogous to the conventional petrochemical refinery systems, is that the feedstock used can be regenerated continuously, hence labelled as renewable.

Biorefinery systems with its sustainable goal to produce less to zero waste aims to gradually replace the use of fossil resources which are constantly under depletion ^[8]. Still, several challenges remain in the implementation of biorefinery systems. One being the variability and complexity that constitute the structure of renewable biomass. An array of technologies confined into an optimized chain of processes are often required to produce the desired product ^[2]. This compilation of processes must compete in cost and efficiency with today’s established chemical processes for it to completely replace the use of fossil resources.

2.2Wheat bran as a biorefinery feedstock 2.2.1 Wheat bran production

In the process of milling flour, the three major components of wheat kernel namely the germ, endosperm and bran are separated in a series of rolling procedures ^[2]. After separation, the endosperm is milled into white flour whereas the germ and bran are often discarded as by-products. The bran which makes up the outer shell of the kernel, can occasionally be incorporated into making whole grain flour ^[9], otherwise it is mostly used as a low-cost ingredient in animal feed ^[10].

Wheat bran as a by-product is commercially available in great quantities ^[10]. In respect to the total worldwide production of wheat which is estimated to reach more than 650 million tons ^[4], an approximate of 100 million tons of wheat bran is produced annually ^[8]. In Sweden, the amount of wheat harvested accounts for the largest cereal produced in comparison to barley and rapeseed. According to the Sveriges Officiella Statistik^[11], the production of wheat in Sweden almost reached 2.5 million tonnes in 2016. This would account for an estimated half a ton of wheat bran to be produced that year. Therefore, with an abundance in availability, wheat bran poses as a cheap feedstock for biorefinery

[9].

2.2.2 Structure and composition of wheat bran

The wheat bran accounts for 14 – 19 % of the whole kernel structure ^[4]. Several different layers with distinctive histological characteristics constitute the wheat bran and these layers include the pericarp, testa, hyaline and aleurone ^[9]. The pericarp or the outer most layer of wheat bran consists of thick cell walls containing cellulose, cuticle, hemicelluloses mainly in the form of highly substituted arabinoxylans, phenolic compounds and lignin ^[9]. The phenolic compounds such as ferulic acid are often cross-linked to other

(10)

biomacromolecules ^[6] and makes the pericarp a challenging fraction to exploit ^[12]. In contrast, the testa is merely composed of lignin and alkylresorcinols. This layer which is known as the seed coat protects the seed from lipid peroxidation. Hence, the high amount of alkylresorcinols act as a protective antioxidant layer ^{[9, 12]}. The hyaline and aleurone layers which are considered part of the endosperm ^[3] are composed of living cells rich in proteins and bioactive compounds. Together with the pericarp, an approximate of 98 % of the total ferulic acids are believed to be concentrated in the aleurone. Thin cell walls containing arabinoxylans ^[13] with a low A/X ratio are also found in this layer ^[9].

Figure 2.1. Cut-through of a wheat kernel with its different compartments and layers (Adapted from Javed et al, 2012)

The overall composition of the different components in wheat bran are listed in Table 2.1.

In comparison to other tissues in the wheat plant, the composition of wheat bran is distinctive in that it contains a large fraction of hemicelluloses. The cellulose and lignin content in wheat bran is relatively low and this has been linked to the need for rapid degradation of seed coat following germination. Moreover, the seed does not perform load- bearing functions, thus does not require high amounts of cellulose and lignin for propagation ^[14].

Table 2.1. Composition of wheat bran ^[3-5, 12]

Component Amount [%]

Water 12.1 – 22.0

Protein 13.2 – 19.7

Fat 3.5 – 3.9

Total Carbohydrates 56.8 – 60.0

Starch 13.8 – 24.9

Cellulose 11.0 – 19

Arabinoxylan 10.9 – 30.0

Beta-glucan 2.1 – 6.0

Phenolic acids 1.1

Ferulic acids 0.02 – 1.5

Lignin 5.0 – 17.8

Ash 3.4 – 8.1

2.2.3 Arabinoxylan

As a character of cereals, the main hemicellulose that comprises wheat bran is arabinoxylan (AX). Arabinoxylans make up most of the non-starched polysaccharides in wheat bran ^[5], as shown in Table 2.1. These polysaccharides are composed of a 1,4-linked- β-D-xylopyranosyl backbone to which monomeric α-L-arabinofuranose residues are

(11)

substituted at the C(O)-3 and/or the C(O)-2 positions. More complex substitutions in the form of D-galactopyranose, α-D-glucuronic acid or its derivative 4-O-methyl-α-D- glucuronic acid units can also be present depending on the tissue origin ^[6]. The physical and structural properties of AXs are notably determined by the character of their substitution pattern ^[5]. The AXs can be further ester bonded to ferulic acid at C(O)-5 position of the arabinose, where these phenolic compounds form cross links with other biomacromolecules. The cross-linking of AXs is mainly found as the formation of diferulic acid bridges between neighbouring AX chains^[6]. Figure 2.2 illustrates a feruloylated arabinoxylan chain.

Figure 2.2. Feruloylated arabinoxylans. Black: xylopyranosyl backbone, blue: arabinose substitutions and orange: ferulic acid ester bonded to arabinose.

2.2.4 Ferulic acid

The wheat bran is rich in hydroxycinnamic acids of which ferulic acid (4-hydroxy-3- methoxy cinnamic acid) is the most dominant ^[9]. Aside from their singular substitution, around 39% of the ferulic acid is present in their dehydrodiferulic form ^[3], most commonly as a diferulic bridge. In general, ferulic acids or FA are commonly derived in plants from the metabolism of phenylalanine and tyrosine ^[15].

Ferulic acid is regarded as a bioactive compound since it has shown to exhibit antioxidant, antimicrobial, anti-inflammatory and anti-thrombosis activities ^{[3, 9]}. The antioxidant capability of ferulic acid is due to the nature of the molecule, which consists of several electron-rich parts. The benzene ring of ferulic acid contains both a hydroxyl and a methoxy group. As an acid, it also possesses a carboxyl group adjacent to an unsaturated double carbon bond. Together these electron-rich parts enable the formation of a phenoxy radical when FA is in contact with a reactive compound. The phenoxy radical is resonance-stabilized and can delocalize the unpaired electron throughout its entire molecule. Aside from its electron-donating capability, FA has also been shown to chelate metals ^[15].

2.3 Sequential processing of AXs from wheat bran 2.3.1 Extraction and fractionation of AXs

As mentioned above, AXs are crosslinked and create a complex organisational network that is mostly insoluble in aqueous solutions. Nonetheless, several methods have been

(12)

explored to extract and fractionate AXs. Some of these methods include extraction using alkali solution, acid solution, ultrasound treatment and microwave treatment. The use of water to extract AXs has grown interest over its benefit of being more environmentally friendly. However, water alone is incapable of breaking down the recalcitrant structure of the cross-linked AXs ^[6]. Hence, the use of water is often manipulated by means of thermophysical or subcritical conditions. This in turns alters the physical properties of water and improve extraction of the insoluble target ^{[5] [16]}. A recent study developed at the Glycoscience division at KTH has shown that subcritical water extraction (SWE) coupled with enzymatic treatment allows for an optimum extraction of AXs from wheat bran ^[12].

A general scheme of the bioprocess is illustrated in Figure 2.4. In brief the process begins with a pretreatment step, involving the defatting and/or destarching of wheat bran. Pre- treated wheat bran is then subjected to subcritical water extraction (SWE) at pressurised conditions to obtain AXs with phenolic compounds still bound to the polymeric fractions.

The extract is then filtered through an ultrafiltration system to separate the polymeric AXs from the arabinoxylan oligosaccharides (AXOs). Meanwhile the left-over residue can be further processed through an enzymatic treatment followed by another round of SWE

[12]. In this study, only the subcritical extraction will be analysed in both the small and pilot scale.

Figure 2.4 Cascade process for extraction and fractionation of AXs (adapted from Ruthes, et al 2017)

2.3.2 Pretreatment

The main objective of performing a pretreatment step is to improve the extractability of AXs as well as increase the content in the extracts obtained. Hence, an investigation on a defatting and destarching step was done to see whether the two pretreatments could considerably affect the extraction process of AXs in subcritical conditions. The presence of fat in wheat bran is relatively low, 3.5 – 3.9 % DW ^[3]. Nonetheless, it is undetermined whether this low presence would have a significant effect on the extractability of AXs.

With regards to the starch granules, the majority of their presence originate from the remainders of the endosperm that adheres to the wheat bran^{[3] [9]}.The content of starch

(13)

can reach up to 16.9 % DW ^[12] and removal of its granular structures has shown to improve yield and content of AXs ^[5]. Starch has a non-recalcitrant structure and can easily be degraded with commercial hydrolytic enzymes, such as α-amylase ^[17]. Alpha-amylase an endo-acting hydrolase, acts by cleaving the 1,4-linked-α-D-glucan chains of amylose and amylopectin. As a result, linear and branched oligosaccharides with varying lengths are released during starch removal ^[18].

2.3.3 Subcritical water extraction

The use of water in subcritical conditions has recently been studied to develop an ecological method for the extraction of hemicelluloses from plant and food sources. As recalled by Ruthes et al (2017), subcritical water extraction (SWE) allows for the isolation of hemicelluloses with high molecular weight and functional compounds still bound to the polysaccharides. The main advantage of SWE is the use of water, which is not only environmentally-sustainable but can also be recycled continuously ^[19].

Subcritical water is defined as water in a high temperature above its boiling point, at atmospheric pressure and below the critical point (Tc 374 oC and Pc 22 MPa) ^[20]. At these conditions water undergoes several physical changes while remaining at liquid state ^[12]. Water is more compressible close to the critical point due to a decrease in density ^[20]. The reduction in density consequently alters the dielectric strength of water enabling it to dissolve less polar compounds. As a result of high temperature, the viscosity of water is also reduced improving the penetration and diffusivity of water ^[19]. With lowered viscosity and enhanced diffusivity, the mass transfer capacity of the extraction process can be improved^[12]. Figure 2.5 illustrates the different phases of water at different temperatures and pressure.

Figure 2.5. Pressure-temperature phase diagram of water (Modified from Peterson et al, 2008)

2.4 Valorization of AXs

As mentioned above, the valorization of AXs is dependent on the polymeric degree of the hemicellulose. One interesting application of AXs is the development of bio-based materials ^[21]. At high molar masses, the long AXs chains can be assembled into materials for packaging, films, coatings and hydrogels. Hemicelluloses have oxygen permeability properties which are considered important in food packaging ^[22]. The presence of hydroxyl groups along the xylan backbone allows for chemical functionalization to modify

(14)

the crystallinity, solubility and hydrophobicity of the fibers. In order to produce barrier films such as moisture barrier films ^[23], surface modifications are often performed on the AX films. The presence of ferulic acid and other phenolic compounds adds value to the AXs by giving it antioxidant properties. Hence, feruloylated AXs place as an interesting precursor for material applications in both the food and health sector ^[12].

On the other hand, the lower molar mass AXOs have a great potential to be applied as prebiotics and dietary fibers. Prebiotics are non-digestible consumed products that allow for beneficial modifications of the gastrointestinal microflora. Consumed AXOs as dietary fiber contributes to an increase in the gut population of Bifidobacterium and/or Lactobacillus species. These microbes are capable of fermenting prebiotics such as AXOS to short chain fatty acids (SCFA) such as propionate and butyrate that subsequently improves the colon absorptive capacity and inhibition of harmful bacteria ^[24]. AXOS have also been linked to antioxidant and anti-inflammatory activities leading to a decreased oxidation of human low-density protein ^[25]. Arabinoxylans also possess water-holding capabilities that enables them to be used as food additives ^[6]. Further processing of AXs can lead to numerous valorised products such as xylitol from xylose and vanillin from ferulic acid. The protein content that is not removed in the extraction process of AXs is also a source of nutritional ingredient for food or special feeds^[3].

(15)

III. MATERIALS AND METHODS

3.1 Materials

Fine and medium sized wheat bran were provided by Lantmännen, (Stockholm, Sweden).

The substances and reagents used in both the bioprocess and analysis were of analytical grade.

3.2 Methods in the bioprocess design 3.2.1 Defatting

Samples of wheat bran were weighed and placed in a glass bottle with a lid. The samples were suspended in methanol:chloroform (1:3; v/v) at a solid:liquid ratio of 1:10 (w/v)., under stirring at 60 ^°C for 3h. Defatted wheat bran was then filtered using Milchfilter 190 Ø and left to dry in the fume hood overnight. Both the defatted sample and fat content were analyzed gravimetrically.

3.2.2 Destarching

The wheat bran was destarched using r-amylase. Weighed samples of wheat bran were suspended in sodium phosphate buffer 50 mmol, pH 7.0, 1:10 (w/v) solid:liquid ratio. The samples were gelatinized at 85 ^°C for 5 minutes and destarched with 16 U/g of r-amylase at 37 ^°C overnight (17 h) under constant stirring. Destarched wheat bran was then centrifuged (8000 rpm, 30 minutes, 4 ^°C) and washed with cold absolute ethanol (95 %) twice with centrifugation in between, before being dried in the oven of 60 ^°C overnight.

3.2.3 Small scale subcritical water extraction

Subcritical water extraction was done using a bench scale accelerated solvent extractor Dionex^TM ASE^TM 350^[12]. Samples of wheat bran were weighed to 3 g and sequentially extracted using 0.1 M sodium formate buffer (pH 7.0) or tap water (pH 7.1) at 160 °C for 5, 10, 15 and 30 minutes. The solid to liquid ratio used was 1:17 (w/v) with a pressure of 10-11 MPa. Both the extracts and residues that underwent extraction using buffer were dialyzed using 0.1-0.5 kDa and 6-8 kDa MWCO membranes for 72 h, respectively, prior to freeze-drying. Samples extracted with water were freeze-dried directly. The yields from the extraction were determined gravimetrically.

3.2.4 Pilot scale subcritical water extraction

The pilot scale subcritical water extraction was performed using a rotary autoclave apparatus submerged in a bath of PEG. As much as 100 g (duplicates) of wheat bran samples were placed inside the autoclave cells and vacuumed for 20-30 minutes. The wheat bran was then suspended in tap water (pH 7.1) at a solid to liquid ratio of 1:17 and 1:10 (w/v). Cells containing the sample mixture were then placed into the pre-heated rotary apparatus and extraction was run sequentially for 5, 10, 15 and 30 minutes. In order for the cells to reach the running temperature of 160 °C, a ramping time of 10 minutes was employed to heat up the cells. In between the runs, the extracts were

(16)

removed manually and filtered through a nylon cloth membrane. Aliquots of each extract and residue were freeze dried and analyzed gravimetrically.

3.3 Methods of analysis 3.3.1 Yield determination

The yields of the extraction process were determined gravimetrically and calculated based on the dry weight of each wheat bran sample, in accordance to the equation below.

To obtain the dry weight, the humidity of each sample, previously determined as the loss of mass after freeze-drying for 72 h. The dry weight is equivalent to the original weight minus the humidity content. As much as 5 mg (triplicates) of each sample were used to determine the humidity content.

stuvw_xyz{|}z = utÄℎÇ_xyz{|}z[Ä]

utÄℎÇ_ÉÑ[Ä] − (utÄℎÇ_ÉÑ[Ä] ×áàâtwtÇä [%]

100 )

3.3.2 Total carbohydrate content

The total carbohydrate content was determined using the phenol-sulfuric method ^[26]. Wheat bran samples were weighed and dissolved in H2O MilliQ to 50-100 µg/ml. As much as 500 µl of the solution was carefully mixed with 500 µl of 5% phenol solution (w/v) and 2.5 ml of concentrated H2SO4 (98 %) in Falcon tubes. The samples were boiled for 5 minutes and cooled to room temperature before being measured using a spectrophotometer at 490 nm. To quantify the samples, a standard curve using xylose was made with a range of 10 – 100 ng/µl.

To determine the polysaccharide content of the pulled extracts (5, 10, 15 and 30 minutes) in mg per gram of starting material, the following equation was used. The abbreviation PC and CC designate polysaccharide content and carbohydrate content, respectively.

çéÇèv êéväëèííℎèìtwu [âÄ Ä^îï] =

ñó [%]

100 × óó [âÄ Ä^îï]

1000 × stuvw [Ä]

òÇèìÇtôÄ öutÄℎÇ [Ä]

3.3.3 Monosaccharide composition

The monosaccharide composition was determined by TFA hydrolysis followed by GC-MS analysis ^[27]. Samples were weighed to 1 mg (duplicates) and hydrolysed using 1 ml trifluoroacetic acid (TFA) 2M at 120 ^°C for 3 h. As much as 200 µl of the hydrolysed samples were transferred to new vials, dried and washed with methanol. Derivatization of the samples was initiated with a reduction step using NaBH4 at 60 ^°C for 1h. Samples were neutralized and dried with acetic acid 10% in methanol (v/v) 3 times. Reduced sugars were then acetylated using acetic anhydride and pyridine as a catalyst (1:1 v/v of 100 µl) at 100 ^°C for 1h. The alditol acetates were then partitioned in H2O MilliQ and ethyl acetate (2:1 v/v), dried and resuspended with 300 µl of ethyl acetate before being transferred to GC vials. The analysis was done using a HP-6890 gas chromatographer and a HP-5973

(17)

electron-impact mass spectrometer. A solution containing 0.1 g/L of acetylated sugars in ethyl acetate (fucose, rhamnose, xylose, arabinose, glucose, mannose, galactose and inositol) was used as a standard. The alditol acetates were identified by their typical retention times and electron impact profiles.

3.3.4 Protein content

The total protein content was determined using the Bradford method ^[28]. This method is only applicable on soluble samples and therefore only done on the extracts. As much as 2 mg of each sample was weighed and diluted to 2-6 mg/ml with H2O MilliQ. In a separate Eppendorf tube, 10 µl of the sample was mixed with 790 µl H2O MilliQ and 200 µl of Bradford dye. The mixture was incubated at room temperature for 5 minutes before being measured at 595 nm. To quantify the samples, a standard curve using bovine serum albumin (BSA) was made with a range of 0.125 to 1 mg/ml.

3.3.5 Molecular weight distribution

The molecular weight distribution was determined using size exclusion chromatography (SEC) coupled to a refractive index detector. As much as 3 mg of samples were dissolved in dimethyl sulfoxide (DMSO) with 0.5 % w/w LiBr. Dissolved samples were filtered through a nylon 0.2 µm filter before being analyzed at 60 ^°C at a flow rate of 0.5 ml/min through a column set comprising of a GRAM PreColumn, 30 and 10000 analytical columns.

3.3.6 Total starch content

The total starch content was determined enzymatically using the Total Starch amyloglucosidase/r-amylase kit provided by Megazyme with modifications. As much as 10 mg of samples were weighed in glass tubes and wetted with ethanol 80 % (v/v). The samples were suspended in 3 mL of MOPS buffer (50mM, pH 7.0) and boiled for 6 minutes. The gelatinized samples were then equilibrated to 37 ôC and destarched using 1U/mg r-amylase for 1 h. Following treatment with r-amylase, the samples were further treated with 0.2U/mg amyloglucosidase. As much as 4 ml of sodium acetate buffer (200 mM, pH 4.5) were added to the samples along with the amyloglucosidase and the mixture was incubated at 50 ôC for 30 minutes. The destarched samples were then aliquoted and centrifuged at 3000 rpm for 10 minutes. As much as 0.1 mL of the supernatant was transferred to a new glass tube and 3.0 mL of GOPOD reagent was added and incubated at 50 ôC for 20 minutes. A solution of D-Glucose (1 mg/ml) and maize starch were used as a standard and control, respectively. In principle, the GOPOD reagent works by first oxidizing the released glucose into gluconate and H2O2. Hydrogen peroxide is then quantified by means of colorimetry as a quinoneimine dye that is measured using a spectrophotometer at 510 nm.

3.3.7 Hydroxycinnamic acids content

The analysis was done by saponification of the phenolic compounds followed by GC-MS analysis ^[29]. As much as 10 mg of the dried extracts were weighed in 1.5 ml tubes and 300

(18)

µl of 2M NaOH was added. The mixture was flushed with N2 and left overnight at 30 ^oC in the dark, under stirring. Samples were then acidified using 12M HCl to pH 3 and extracted by partition using 1 ml ethyl acetate (3 times). Pulled extracted samples were dried under a N2 stream and silylated using 50 µl N-O-bis (trimethylsyl acetamide) at 100 ^oC for 5 minutes, then resuspended using acetone. The samples were analyzed using a HP-6890 gas chromatographer and a HP-5973 electron-impact mass spectrometer. The sylilated derivatives were identified by their typical retention times and electron impact profiles compared to standards (ferulic acid, p-coumaric acid, t-cinnamic acid, synaptic acid and caffeic acid).

(19)

IV. RESULTS AND DISCUSSION

4.1 Preliminary Study of Starting Material

A preliminary analysis of the chemical composition of the different pretreated fine wheat bran samples is tabulated in Table 4.1. The protein and hydroxycinnamic acid content were not determined directly, but according to a previous study done by Ruthes et al (2017), the values for the two substances are 202.0 mg/g DW and 2.2 mg/g DW, respectively.

Table 4.1. Chemical composition of the different pretreated wheat brans

Untreated Defatted Destarched Defatted and

Destarched Moisture content [% w/w]^a 27.0 (1.6) 21.0 (0.3) 5.8 (0.0) 25.5 (2.1) Total carbohydrate [mg g^-1 DW]^b 700.1 (90.9) 762.5 (193) 700.3 (0.9) 762.0 (108)

AX [% of Total carbohydrate]^c 30.6 (0.6) 34.8 (2.2) 55.3 (5.6) 50.0 (0.2)

A/X ratio^c 0.63 0.59 0.24 0.56

Glc [% of Total carbohydrate]^c 66.8 (0.3) 62.5 (2.2) 29.8 (9.0) 48.0 (0.4) Starch [mg g^-1 DW]^b 164.4 (1.7) n.d.^e 46.8 (0.5) 28.9 (2.9)^f Numbers in brackets refer to the standard deviation of duplicates

a Moisture determined by gravimetric analysis after freeze-drying

b Carbohydrate content was determined by the phenol sulfuric method

c Monosaccharide composition was determined using GC-MS analysis of alditol acetates

d Starch content was determined enzymatically using r-amylase and amyloglucosidase

e Not determined

f Adapted from Ruthes et al (2017)

4.2 Small Scale Lab: Effects of Pretreatment

In the small-scale lab, the effect of pretreating the wheat bran prior to subcritical water extraction was evaluated. A majority of the wheat arabinoxylans (AXs) are classified as being water un-extractable ^{[6, 30]} meaning that pretreatment steps are often needed to assist the release of AXs from its crossed-liked form, hence making it more water- extractable. The primary motive of performing the small-scale lab was to see whether selected pretreatment steps were necessary in the upscale, bearing in mind that at pilot or industrial scale, the process should be optimized to improve the cost and efficiency.

Four different pretreatment conditions were tested and these comprised of not treating, defatting, destarching as well as defatting and destarching the wheat bran. To simplify the discussion, only the fine-sized wheat bran will be discussed as this type was later chosen to be implemented on to the pilot scale. The results of the small-scale lab are divided into the total solid yields and the polysaccharide composition of combined extracts from the same treated sample (extracts of 5, 10, 15 and 30 minutes were pulled together); the effect of time towards the yield and polysaccharide composition; and the molecular characteristics of the polysaccharides extracted.

(20)

4.2.1 Total solid yields and polysaccharide composition from the small scale process

Figure 4.1 (A) shows that the pretreatments resulted in a general increase of the total extract yields (except for destarched wheat bran with buffer), implying the importance of performing a pretreatment step prior to extraction. Among the different pretreatment conditions, the destarched sample extracted with water (W) gave the highest total extract yield of 46.7 % DW, followed by defatting (44.0 % DW), defatting and destarching (39.7

%) and destarching with buffer (26.0 % DW).

(A) (B)

Figure 4.1. (A) Total solid yields of the extracts (E) and residues (R) of different pre-treated fine wheat bran samples in % DW. (B) Total amount of polysaccharides extracted in mg of carbohydrate per gram of starting wheat bran. The destarched sample extracted with water is

denoted with (W).

In the defatted sample of the fine wheat bran, an increase in the yield of extract was observed compared to the untreated samples (Figure 4.1A). This increase suggests that the removal of fat improved the extraction of polysaccharides from wheat bran. The wheat bran retains different populations of hydrophobic compounds such as pigments and alkylresorcinols that are concentrated in the testa layer ^[31, 32]and fatty acids mainly in the form of linoleic and palmitic acids ^[33] that are likely derived from the cellular membranes. As these hydrophobic compounds are removed during defatting, it is likely that the rigid histological layers making up the wheat bran are partly disintegrated, hence improving the extractability of the polysaccharides. In respect to the composition of the total polysaccharides extracted (Figure 4.1B), both the untreated and defatted samples displayed proportional amounts, whereby the glucose amount was higher than that of the AXs, indicating that defatting did not considerably change the polysaccharide composition of the extracts. Here, the presence of glucose is most likely derived from the remaining starch and mixed linkage õ-D-glucans that are inherently present in the wheat bran ^[14].

In the case of the destarched sample, the total solid yield of extracts was found to decrease compared to the untreated control (Figure 4.1A). The presence of starch in wheat bran originates from remaining endosperm tissues that adheres to the aleurone layer, where it makes up around 15 – 20 % of the total wheat bran ^[30]. The starch content of the untreated wheat bran was measured at 16.4 % (w/w), which is in good agreement with the aforementioned literature. The aleurone layer in the wheat bran does not contain

E

R R E

R

0 10 20 30 40 50 60 70 80

Untreatted Defatted Destarched Defatted and Destarched Destarched

(W)

Total Solid Yield [% DW]

0 100 200 300 400

Untreated Defatted Destarched Defatted and Destarched Destarched

(W) Total Extracted Polysaccharide [mg/g] ^AXs ^Glc ^Other

(21)

starch ^[14]. Destarching with α-amylase resulted in a reduction of starch content to 4.7 % (w/w). Consequently, this implies that a majority of the starch has been removed to leave the aleurone layer more exposed for extraction of other polysaccharides. Hence, the lower yield could simply translate to obtaining less glucose while acquiring the same amount or more of the arabinoxylans. This was the phenomenon observed. The total extracted AX increased from 78.0 mg/g in the untreated sample to 105.5 mg/g in the destarched sample extracted with buffer. The destarching resulted in a 60.8 % (w/w) drop of total extracted glucose and a 35.3 % (w/w) increase in the total AXs content of the whole extract. In conclusion even though the yields were lower than the untreated, the total resulting amount of AX obtained in mg from the destarched sample was still higher.

Interestingly, as for the water-extracted destarched wheat bran, the yield was found to be highest among all the pretreatment conditions. The water-extracted destarched sample was however not dialyzed and the higher yield could possibly be due to the accumulation of small compounds that would not have been present if dialyzed. Also, as a result of not performing a dialysis, both the arabinoxylans (AXs) and glucose content in the destarched (W) sample (181.8 mg/g DW and 164.2 mg/g DW, respectively) were higher than that of the destarched sample extracted with buffer (105.5 mg/g DW and 58.8 mg/g DW, respectively). The higher content of AXs and glucose are most likely due to the accumulation oligosaccharides that was released during the extraction. In terms of the AX content as a percentage of the total polysaccharides, the values (83.3 – 91.9 %) were comparable to that of the buffer-extracted destarched samples (Appendix B.1), specifying that the change in solvent had no substantial effect in the polysaccharide composition of the extracts.

The other pretreatment condition was to combine both the defatting and destarching process sequentially. In this sample, an increase in the total solid yields was observed compared to the untreated sample. As expected, since it underwent a destarching process, the total extracted glucose content in mg/g of the starting material declined by 19.1 % (w/w) compared to the untreated (Figure 4.2 B). This reduction in glucose was however not as substantial as when performing the destarching treatment alone. The higher content of glucose could possibly be related to the difference in activity of the α-amylase in untreated and defatted wheat bran.

4.2.2 Effect of extraction time in the small scale process

In order to see the effect of time towards the fractionation of arabinoxylans (AXs), the extraction procedure was done sequentially, summing up 5, 10, 15 and 30 minutes for each of the samples. Based on Table 4.2, it can be concluded that as the extraction time lengthens, the amount of total solid yield generally decreases. This is expected as the extraction process is multi-staged ^[34] and therefore the amount of available material to be extracted will be reduced every time.

(22)

Table 4.2. Total solid yields per time of extraction of fine wheat bran samples Total Solid

Yields [% DW]

Untreated Defatted Destarched Defatted and Destarched

Destarched (W)

5’ 10.1 23.1 11.7 10.6 22.9

10’ 10.7 13.6 8.4 15.8 9.8

15’ 4.5 5.0 3.4 5.4 6.4

30’ 3.7 2.3 2.6 7.8 7.6

Total Extract 29.1 44.0 26.0 39.7 46.7

Residue 45.2 43.0 57.4 69.0 53.8

As previously mentioned by Ruthes et al (2017), the subcritical water extraction (SWE) seems to preferentially extract AXs from the aleurone and intermediate layers. The arabinoxylans found in the outer pericarp have a higher A/X ratio, causing them to be cross-linked to other biomacromolecules in an intricate matrix that makes them more difficult to extort with water alone ^[12, 32]. With this premise in mind, the extraction process can be considered to proceed from the aleurone layer, where leftover starchy endosperm parts may remain, and outwards onto the intermediate and pericarp layers.

(A) (B) (C)

(D) (E)

Figure 4.2. Polysaccharide content of the different extracts and residues from various pre-treated fine wheat bran samples. Extractions were done with 0.1 M sodium formate buffer (pH 7.0) or water (pH 7.1) at 160 ^oC. Only samples extracted with buffer were dialyzed before analysis. The

graphs represent the amount of polysaccharide in mg per g DW of extract.

0 200 400 600 800 1000

5' 10' 15' 30' R

Polysaccharide content [mg /g DW]

Untreated

AXs Glc Other

5' 10' 15' 30' R

Defatted

AX Glc Other

5' 10' 15' 30' R

Destarched

AX Glc Other

0.0 200.0 400.0 600.0 800.0 1000.0

5' 10' 15' 30' R

Polysaccharide content [mg/g DW]

Defatted and Destarched

AX Glc Other

5' 10' 15' 30' R

Destarched (W)

AX Glc Other

(23)

As shown in Figure 4.2, for samples that were not destarched (untreated and defatted), a high amount of glucose (710.5 g/mg and 710.1 mg/g, respectively) was extracted in the first 5 minutes that was probably derived from the starchy remainder of the endosperm.

The amount of glucose obtained in the destarched (420 mg/g) as well as in the defatted and destarched (320.9 mg/g) samples was much lower. Another contributing source of glucose, especially in the destarched samples are the mixed linkage õ-D-glucans that make up approximately 28 % of the polysaccharides in the aleurone layer ^[14]. These hemicelluloses are found in the inner seam of the aleurone tissue ^[14], thereby serving as a frontier structure to be penetrated by water during the SWE.

Moreover, in the destarched sample extracted with both buffer and water (Figure 4.2 C and D), the glucose content considerably dropped after 5 minutes, resulting in AX-rich extracts with an AX content > 89 %. Here, it can be assumed that the õ-D-glucans have mostly diminished and the outer aleurone layer consisting of the arabinoxylans ^[14] is being mainly extracted. The AX content remained fairly high after 30 minutes (89.3 % and 83.3 % for destarched wheat bran extracted with buffer and water, respectively) and this high value was sustained in the residue (81.0 % and 91.7 % for destarched wheat bran extracted with buffer and water, respectively) implying the possibility to further optimize the extraction conditions.

In all the wheat bran samples, it is observed that the A/X ratio generally rose with longer extraction time, where the highest ratios are at 30 minutes (Table 4.3 and Appendix B.1).

The A/X ratios are even higher in the residues (up to 0.66 for the untreated and defatted samples), suggesting that the increase of time allowed for the extraction of more branched arabinoxylans that are most likely to originate from the deeper layers (intermediate and pericarp) of the wheat bran.

Therefore, it can be implied that different populations of polysaccharides are obtained on the different times of extraction. In theory, the time of extraction will have an impact on the efficiency of the SWE ^[35]. Sufficient extraction time is required to allow for diffusion, dissolution and elution^[35] of the solute, in this case the polysaccharides, from the wheat bran into the subcritical water. With regards to the experimental scheme employed, the destarched wheat bran extracted with water showed the best overall yield with high AX content after 10, 15 and 30-minute extractions. Hence, this condition was later chosen to be continued on to a pilot scale. The implementation of the multi-staged solid-liquid extraction will only be validated at this point, but would most likely require further optimization if it were to be applied at a commercial or industrial scale. The multi-stage extraction allows for improved recovery of the solute, however it has the consequence of resulting in relatively dilute extracts ^[34].

4.2.3 Characterization of extracted polysaccharides from the small scale process

A detailed summary of the constituents of the extracts obtained from the different pretreatment conditions are presented in Table 4.3 (also in Appendix B.1). The protein detected in the extracts range from 21.3 to 85.6 mg/g DW. This value is relatively low compared to the total protein content in the starting wheat bran, which is approximated at 202 mg/g DW ^[12]. The Bradford method ^[28] was employed to determine the protein

(24)

content in the extracts and it is only suitable for soluble proteins. Thereby, insoluble proteins and proteins that remained in the residue were not measured. This could have accounted for the lower values. For a majority of the treatments (except the defatted and destarched), there is a general increase in protein content between 5 to 15 minutes, followed by a decline after 30 minutes. This decline is likely due to the proteins being denatured and degraded after a prolonged exposure to high temperature. The applicative benefit of retaining the proteins in the extracts is still to be investigated. But, a possible use is to incorporate the protein-containing AXs as nutritional ingredients for food or special feeds ^[3].

The total extractive content was measured on the destarched wheat bran extracted with water (Table 4.3). This analysis was done to see where the fat and other hydrophobic compounds had been retained after extraction, that is whether it was extracted along with the polysaccharides or remained in the residue. According to the results obtained, it was apparent that extractives are present in both the extracts (168.8 – 203.2 mg/g) and the residue (263.7 mg/g). This implies that the extraction conditions along with changes in the physical property of the subcritical water may have aided the abstraction of hydrophobic compounds in the extracts. Nonetheless, the amount of extractives present in the extracts remain relatively low compared to the carbohydrate content (602.8 – 844.2 mg/g).

4.2.3.1 Molar mass distribution of polysaccharides from the small scale process

The molar mass distributions of the polysaccharides extracted from the destarched wheat bran with water were analyzed using size exclusion chromatography (SEC) (Figure 4.3).

As mentioned by Ruthes et al (2017), the molar mass characteristics of the AXs are somewhat correlated to the pH of the solvent, whereby lowered pH may result in acid- catalyzed hydrolysis of the polymeric chains. Meanwhile, the use of tap water as solvent resulted in a slight drop of pH from 7.1 to 6.1 in the extracts. Hence, to comprehend the effect of this pH drop towards the molar mass of the polysaccharides, a SEC analysis was performed.

Figure 4.3. Molecular mass distribution of polysaccharides extracted from destarched fine wheat bran using tap water. The legend indicates the different extraction times of 5, 10, 15 and 30

minutes

(25)

The SEC analysis revealed a binary molar mass distribution comprising of a low molar mass fraction (10³Da – 10⁴ Da) and a high molar mass fraction (10⁵Da) in all the samples.

This is consistent with the findings observed previously by Ruthes et al (2017) and also validates that a change in solvent did not pointedly alter the molar mass of the extracted polysaccharides. In the 5-minute extract, there is a higher abundance of small molar mass compounds (10²Da – 10³ Da), most likely derived from the inherent oligomers and partially degraded polysaccharides that were obtained in the first extraction step.

Interestingly, the molar mass of the 10 and 15-minute extracts overlapped seamlessly indicating that similar molar mass distribution polysaccharide populations were extracted during this time. In the 30-minute extract however, there is an increase of the low molar mass fraction (10³Da – 10⁴ Da), a decrease of the high molar mass fraction (10⁵ Da) and a general graphical shift to the left. This possibly implies that the polysaccharides begin to degrade after 30 minutes under high temperature and pressure.

4.2.3.2 Hydroxycinnamic acid content of polysaccharides from the small scale process The hydroxycinnamic acid content was determined for the destarched wheat bran extracted with water. A range of phenolic acids (ferulic acid, p-coumaric acid, t-cinnamic acid, synaptic acid and caffeic acid) were used as standards, though only ferulic acid was detected in the samples. As illustrated in Figure 4.4, all the samples extracted at the different times contained ferulic acid (0.17 – 0.68 mg/g DW). This implies that the AXs obtained from the extraction preserved the ferulic acids attached to the polysaccharides.

Ferulic acids are most often ester bonded to the C(O)-5 position of the arabinose ^[6]. The amount of ferulic acid rose considerably after 5 minutes, followed by a fairly constant quantity throughout 10 to 30 minutes. This tendency correlates well to the AX content of the extracts (Figure 4.2 E), where the AX content also rose after 5 minutes and was kept at a rather constant fraction between 10 to 30 minutes. The presence of ferulic acid is greatly desired as it adds value to the extracts by providing the AXs with antioxidant properties ^[12].

Figure 4.4. Ferulic acid content of extracts from destarched wheat bran extracted with water. The x-axis denotes the times of extraction in minutes

0 0.2 0.4 0.6 0.8 1

5' 10' 15' 30'

Ferulic Acid Content [mg/g DW]

Destarched (W)

(26)

Table 4.3 Characterization of polysaccharides from untreated and destarched wheat bran of the small scale

Untreated Wheat Bran Destarched Wheat Bran (W)

5’ 10’ 15’ 30’ Residue 5’ 10’ 15’ 30’ Residue

Yields [%

DW]^a ^10.2 10.7 4.6 3.7 45.2 22.6 9.7 6.3 7.5 53.1

Carbohydrate [mg g^-1 DW]^b

873.9 (21.3)

791.2 (263.6)

649.4 (57.7)

774.2 (22.1)

853.7 (81.5)

836.3 (20.7)

844.2 (18.3)

619.8 (15.6)

602.8 (3.1)

526.2 (45.9) AX [%]^c 16.6 (2.2) 35.0 (2.5) 50.6 (2.1) 64.8 (0.8) 81.3 (1.7) 16.9 (0.6) 91.9 (5.8) 88.9 (1.0) 83.3 (5.4) 91.7 (2.3)

A/X^c 0.46 0.41 0.39 0.46 0.66 0.20 0.22 0.22 0.37 0.34

Glc [%]^c 81.3 (1.9) 63.9 (2.7) 47.0 (1.8) 33.8 (0.8) 15.0 (0.3) 76.3 (0.5) 8.1 (0.4) 11.13 (0.6) 14.7 (4.2) 8.03 (0.4) Other [%]^c 1.9 (0.1) 1.1 (0.2) 2.4 (0.3) 1.4 (0.03) 3.6 (0.2) 6.8 (0.3) - - 1.93 (0.6) - Protein

[mg g^-1 DW]^d

81.73 (4.0)

84.44 (8.6)

75.73 (5.7)

61.36

(0.4) ^n.d.^f 41.3 (0.7) 72.4 (12.0) 85.6 (12.7) 57.1 (5.7) n.d.^f Phenolics

[mg g^-1 DW]^e ^n.d.^f n.d.^f n.d.^f n.d.^f n.d.^f

0.170 0.598 0.681 0.673 n.d.^f

Extractives

[mg g^-1 DW]^g n.d.^f n.d.^f n.d.^f n.d.^f n.d.^f 178.6 (55.1)

203.2 (8.8)

196.9 (0.0)

168.8 (8.1)

263.7 (11.5)

Numbers in brackets refer to the standard deviation of duplicates

a Yields were determined by gravimetric analysis after dialysis and freeze-drying

b Carbohydrate content was determined by the phenol sulfuric method

c Monosaccharide composition was determined using GC-MS analysis of alditol acetates

d Soluble protein content was determined using Bradford

e Total phenolic content was determined using GC-MS

f Not determine

g Extractives were obtained using CHCl3 : MeOH (3:1) and analyzed gravimetrically

(27)

4.2.4 Summary of the small scale lab

Henceforth, based on the total yields, AX content and molecular characteristics acquired from the different pretreatment conditions, it was evident that the destarched wheat bran extracted with water delivered the most promising outcome. This condition was later chosen to be validated at pilot scale. From both an economical and environmental stance, the removal of a defatting step is advantageous. Chloroform which is the main solvent used in the defatting does not bioaccumulate nor does it pose as a significant risk to the environment ^[36]. Nevertheless, removal of its complete use would ensure a more environmentally-sustained process. The use of water compared to buffer is clearly more economical as no additional chemical substances are needed, while the process of recycling the water itself will also be easier. Additionally, the use of enzymes in the pretreatment of wheat bran is becoming more and more common as enzymes are considered more green than chemicals. Currently, thermostable α-amylases are widely explored to be implemented in the pretreatments of wheat bran as it can withstand gelatinization conditions, therefore enabling the destarching process to proceed more quickly ^[10].

4.3 Pilot Scale Lab: Validation

To validate the process in pilot scale, as much as 100 g (scale factor of 33 times) of wheat bran samples were processed through the selected pretreatment and extraction procedure. In the upscale scheme, aside from the effect of pretreatment and extraction time, the effect of the solid to liquid ratio of (1:17) and (1:10) were also evaluated.

4.3.1 Total solid yields and polysaccharide content from the pilot scale process

As demonstrated in Figure 4.5 A, the total solid yields of the extracts from the untreated samples were higher in the pilot scale (35.2 % DW and 40.4 % DW for solid to liquid ratios of 1:17 and 1:10, respectively) as compared to the small scale (29.1 % DW, shown in Figure 4.1 A). The two data sets are however incomparable since in the pilot scale, the extracts were not dialyzed and the higher yield could possibly be due to the accumulation of small molecules. Therefore, the untreated samples in the pilot scale served as a mere control for the destarched samples.

Pilot scale process for polysaccharide extraction and fractionation from cereal by-products

Pilot scale process for

polysaccharide extraction and fractionation from cereal by- products

RESKANDI CHASTELIA RUDJITO

Master Thesis

Pilot scale process for polysaccharide extraction and fractionation from cereal by-products

Reskandi Chastelia Rudjito

ABSTRACT

SAMMANFATTNING

TABLE OF CONTENTS

I INTRODUCTION

II LITERATURE

III. MATERIALS AND METHODS

IV. RESULTS AND DISCUSSION