Solvent-free production of protein-enriched fractions from navy bean ...

Journal of Food Engineering 174 (2016) 21e28

Contents lists available at ScienceDirect

Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng

Solvent-free production of protein-enriched fractions from navy bean flour using a triboelectrification-based approach Solmaz Tabtabaei a, Mousa Jafari a, c, Amin Reza Rajabzadeh b, c, *, Raymond L. Legge a a

Department of Chemical Engineering, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada School of Engineering Technology, McMaster University, Hamilton, Ontario, L8S 0A3, Canada c Advanced CERT Canada Inc. Waterloo, Ontario, N2L 6R5, Canada b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 10 June 2015 Received in revised form 20 September 2015 Accepted 12 November 2015 Available online 19 November 2015

Conventional plant protein production methods involve the use of solvents, concentrated acids and alkali that can result in protein denaturation and loss of solubility, thereby reducing the quality and functionality of protein ingredients. A novel solvent-free tribo-electrostatic separation approach was developed for the production of high quality plant protein powders using navy bean flour as a model system. The process consisted of a fluidized bed flour reservoir, tribo-charging unit, and a separating device. Using this approach the protein content of the original navy bean flour was increased from 25.4% to 47.0% in the final protein-enriched fraction collected from the bottom section of the plate in the separator. A wide variety of tribo-charging tubes with different sizes, shapes, and materials were tested to evaluate their efficiency. Protein recovery by this approach was higher than other solvent-free separation approaches and could be improved by further optimization of the flow rate and design configuration of the separator. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Electrostatic separation Triboelectrification Protein- and carbohydrate-enriched fractions Protein purification Navy bean flour Solvent-free

1. Introduction Many plant proteins are low-cost and excellent sources of nonanimal proteins containing sufficient amounts of essential amino acids for human nutritional needs (Boye et al., 2010; Day, 2013). Plant proteins are utilized alone as protein supplements or as additives in various food processing systems to provide structural integrity for food products through processes such as emulsification, foaming, gelation and dough formation (Day, 2013). The quality and nutritional value of the protein extracted from plant resources are significantly dependent on the processing technology used to produce them. Therefore, the development of a process technology that can separate and preserve plant proteins with high nutritional values and functional properties is highly desirable. The production of protein-enriched ingredients from pulses, grains, and oilseeds still relies on wet fractionation techniques that involve the use of solvents, alkali, and/or concentrated acids. Most of the traditional wet extraction processes focus on the production

* Corresponding author. School of Engineering Technology, McMaster University, 1280 Main Street West, Hamilton, Ontario, L8S 0A3, Canada.. E-mail address: [email protected] (A.R. Rajabzadeh). http://dx.doi.org/10.1016/j.jfoodeng.2015.11.010 0260-8774/© 2015 Elsevier Ltd. All rights reserved.

of soy protein isolates from defatted soy flakes/flours through a series of unit operations including aqueous extraction, centrifugation, isoelectric precipitation, washing, neutralization, and freeze/ spray drying (Lusas and Rhee, 1995; Mondor et al., 2012). While such processing conditions can produce isolates containing over 90% protein (Lusas and Riaz, 1995), they can result in protein denaturation and loss of solubility (Brooks and Morr, 1985; Omosaiye and Cheryan, 1979; Petruccelli and Anon, 1994), reducing the quality and functionality of the protein products. These processes also generate large volumes of whey-like acid effluents with relatively high quantities of proteinaceous material which can lead to serious water quality issues (Mondor et al., 2012). The application of combined electroacidification and ultrafiltration processes in recent years was able to improve the solubility characteristics of the resulting soy protein isolates/concentrates (Mondor et al., 2004; Skorepova and Moresoli, 2007); however, these technologies suffer from membrane fouling and solute build up in the vicinity of the membrane surface which decreases the performance of the process by lowering the permeate flux (Rajabzadeh et al., 2010). The application of wet fractionation processes is necessary for the production of protein isolates with high protein content (>90%); however, most food processing systems require protein

22

S. Tabtabaei et al. / Journal of Food Engineering 174 (2016) 21e28

enriched fractions with moderate protein concentrations to achieve acceptable structural formation of the different food products. Therefore, a sustainable solvent-free separation approach for the production of protein-enriched fractions from plant resources is very attractive. Dry milling and air classification are the most commonly used technologies for the production of protein- and starch-enriched fractions from cereals and grain legumes. During air classification, seeds are first ground into very fine flour, followed by classifying in a spiral air stream. In this approach the separation of proteinenriched particles (light fine fraction) from starch enriched particles (heavy coarse fraction) is achieved according to their aerodynamic properties, a combination of density and particle size (Boye et al., 2010; Day, 2013; Schutyser and van der Goot, 2011). The separation of fine and coarse fractions are usually carried out in air classifiers where the centrifugal force is supplied by the rotor or classifier wheel. One of the major factors that must be controlled during air classification is the cut-point of the separation. The optimum cut point for protein-starch separation is ~10 mm, just below the size of most starch granules (Day, 2013; Dijkink et al., 2007; Schutyser and van der Goot, 2011). Starch and protein enriched fractions have been produced by pin milling and air classifying for several dehulled legume grains including mung beans, green lentils, Great Northern beans, faba beans, field peas, navy beans, baby lima beans and cowpeas (Tyler et al., 1981). Using another processing system compromised of micronization and air classification stages, a protein-rich fraction containing 38.2% protein and 2.4% fiber was produced from sunflower meal with an original content of 34% protein and 14.5% fiber (Laudadio et al., 2013). Air classification technology of flour requires large and expensive equipment and provides poor classification efficiency due to the adhesion of small protein particles to the larger starch granules. Therefore, air classification processes are usually repeated several times to improve their separation efficiency (Boye et al., 2010; Dijkink et al., 2007; Tyler et al., 1981). Clearly, efficient and cost-effective techniques for fractionation of plant materials are desired. Another solvent-free separation technology suitable for separation of particles from powdery or granular materials is electrostatic separation. Electrostatic separation is a process that imparts a positive or negative charge to the particles, enabling the separation of particles based on differences in charge characteristics under the influence of an external electric field (Dwari and Hanumantha Rao, 2007; Pandya et al., 2013). The principles of electrostatic separation have been successfully applied in plastic waste recycling (Park et al., 2008; Wu et al., 2013) and also in the mining industry to separate various materials (Dwari and Hanumantha Rao, 2007, 2008, 2009). For commercial applications of electrostatic separation, conductive induction, corona bombardment and triboelectrification have been used to impart surface charge (Dwari and Hanumantha Rao, 2007; Mayr and Barringer, 2006; Mazumder et al., 2006). Induction charging is based on the polarity and conductivity of the particles. The principle of corona charging is based on ion bombardment where a flow of ionized gas generated between a pair of high voltage electrodes can significantly impart charge to all particles entering the corona field. The charged particles are then introduced to a grounded drum where the conductive particles lose their charge to the drum, permitting the separation of the conductive and non-conductive particles (Dwari and Hanumantha Rao, 2007). An electrostatic approach using corona discharge has been successfully applied for cleaning and upgrading agricultural seeds (Kovalyshyn et al., 2013). Electrostatic separation of peelings and gluten from ground wheat grains was recently developed based on conductive induction charging of powders using a grounded metallic belt conveyor and a

rotating roll electrode connected to a negative high voltage supply (Remadnia et al., 2014). The tribo-electrostatic separation process involves tribocharging of particles by physical contact against each other and/or friction with other dissimilar materials followed by the separation of the charged particles under the influence of an electric field. The tribocharging can occur by either electron or ion transfer mechanisms (Kamiyama et al., 1994; Nemeth et al., 2003) through the movement of the particles in a tribo-charger media such as a fluidized bed, a vibrating bed, or a pneumatically conveyed stream of particles (Dwari and Hanumantha Rao, 2007, 2008, 2009). The transfer of the electrons through an electron transfer mechanism is achieved by a difference between the work functions of the two particles or contact surfaces. The work function is the surface parameter of the particles defined as the minimum energy required to remove an electron from the surface. It is a measure of the relative affinity for electrons and reflects how tightly electrons are bound to a particle. Upon physical contact, the particles with higher work functions are charged negatively, whereas the particles with lower work functions are charged positively (Dwari and Hanumantha Rao, 2007; Li et al., 1999). There is limited information in the literature regarding the fractionation of protein from legume flour using tribo-electrostatic processes. The closest reported technology explored the triboelectrification properties of different food powders for electrostatic coating (Mayr and Barringer, 2006). The authors examined the tribo-electric chargeability of fish collagen hydrolyzate, protein soy powder, protein whey powder, and native potato starch using either Teflon or nylon tribo-chargers. Using Teflon, the protein powders were characterized by the highest charge-to-mass values, whereas the potato starch was characterized by the lowest chargeto-mass value. Recent research on triboelectrification indicates that particle surfaces containing ionizable functional groups are charged more effectively (Kamiyama et al., 1994; Mayr and Barringer, 2006; Nemeth et al., 2003). Proteins have many ionizable side chains in addition to their amino- and carboxy-terminal groups. Therefore, proteins can be charged upon physical contact with the tribocharging media due to the transfer of electrons or ions between the surface of the tribo-charger and the ionizable groups exposed at the surface of the protein particles. The direction and amount of electron or ion transfer between the surface of the protein particles and tribo-charging media depends on the chemical composition and isoelectric point of the proteins as well as the contact material of the tribo-charger (Mayr and Barringer, 2006). In contrast, carbohydrates are characterized by low proton affinity and ionizability (Ahn and Yoo, 2001); thus, carbohydrate particles cannot be charged effectively using triboelectricity. In a recent study, an electrostatic-based separation approach was applied as a posttreatment to improve the protein content of air-classified and unclassified pea and lupine flour (Pelgrom et al., 2015). It was shown that electrostatic separation increased the protein content of lupine fractions from 35% to 59% while no significant improvement in the protein content of pea flour was observed. The results indicate that electrostatic separation could be potentially used as a purification process for lupine flour. There are large gaps in our understanding of tribo-electrostatic separation of protein from plant materials. This provided the motivation to examine the feasibility of applying this approach at lab scale. The present study was undertaken with the aim of classifying navy bean flour by use of tribo-electrostatic separation as a solvent-free approach to produce high-quality protein- and carbohydrate-enriched fractions. The objectives of the this study were: (1) to investigate the effect of different contact materials for tribo-charging the proteins and carbohydrates in navy bean flour to different levels; (2) to examine effects of protein and carbohydrate


concentrations on the resulting charge density; (3) to explore the possibility of classifying navy bean flour for production of proteinand carbohydrate-enriched fractions using of a triboelectric separator system along with investigating the effects of various factors on protein extractabilities including the size and configuration of the tribo-charger as well as the voltage of the plate-based separator. 2. Materials and methods 2.1. Materials Pin-milled navy bean flour was kindly provided by the Canadian International Grains Institute (CIGI, Winnipeg, MB, Canada) and stored at 4 C. Soy protein concentrate powders with different protein contents of 66.7% and 41.6% were purchased from Purely Bulk (Guelph, ON, Canada) and North Coast Naturals (Port Coquitlam, BC, Canada), respectively. Corn starch flour containing 0% protein and 90% carbohydrates was purchased from Bulk Barn (Waterloo, ON, Canada). Potassium sulfate (99.0þ%) and selenium oxychloride (97%) were from Sigma Aldrich (Oakville, ON, Canada) and used during the digestion step of the Kjeldahl analysis for protein determination. Nessler reagent was from Ricca Chemical Company (Arlington, TX, USA). 2.2. Contact materials for tribo-electrification A variety of materials were screened to determine the most effective for use in tribocharging of the bean flour. Four small tubes (20 cm length 1 cm ID) made of different contact materials of polytetrafluoroethylene (PTFE), polyvinylchloride (PVC), nylon and copper were filled with 1 g of navy bean flour. After shaking the tubes for 2 min at room temperature using a vibrator, the flour was dispensed directly into a Faraday cup (Monroe Electronics Inc., Lyndonville, NY, USA) for charge measurement. The charge of the non-treated navy bean flour was also measured to serve as a comparison point. The charge measurement experiments were repeated six times. 2.3. Comparison of protein and carbohydrate chargeability The classification of navy bean flour into protein- and carbohydrate-enriched fractions under the influence of an electric field requires that the proteins and carbohydrates have different chargeability characteristics. Therefore, evaluating the effect of protein and carbohydrate content on the charge density is important. Four types of flour including corn starch flour (0% protein), navy bean flour (25.4% protein) and two different soy protein concentrates containing 41.6% and 66.7% protein were transferred into the PTFE tubes (20 cm length 1 cm ID) followed by vibration for 2 min. The samples were dispensed into a Faraday cup for charge measurement. Charge densities measurements were performed in triplicate at room temperature. 2.4. Triboelectric separator and tribocharger tests The laboratory-scale tribo-electrostatic separator was designed by Advanced CERT Canada Inc. (Waterloo, ON, Canada) and contained seven different compartments through which navy bean flour was sequentially passed and charged for separation and fractionation of protein- and carbohydrate-enriched fractions. Fig. 1 provides the experimental set-up and a schematic diagram of the feed, tribo-charging and separation units. The electrostatic separator consisted of a fluidized bed feeding system, a straight- or spiral-type tribo-charger tube, one vertical

23

copper plate electrode placed in a rectangular chamber, high DC voltage supply source (ES813 electrostatic/high voltage generator, ESDEMC Technology, Rolla, MO, USA), high-pressure air cylinder and air dryer column coupled with flow controller system. The batch fluidized bed feeding system consisted of a PVC cylindrical vessel, 7 cm in diameter and 12 cm in length, with a PTFE-coated magnetic stir bar to provide constant stirring to the bed material. The top and bottom ends of the fluidized bed were flat and completely sealed utilizing self-sealing screw cap closures. The pressurized air stream (~200 kPa) was first passed through a 94 cm long drying column, with the air introduced vertically at a constant air flow rate into the fluidized bed through a nozzle installed on top of the bed. The fluidization of the flour facilitated suspension of the bed particles to allow interparticle contact and contact with the inner surface of the bed. The suspended particles were fed into either a straight or spiral tube that served as the main triboelectric charger. During pneumatic transport, particle-wall and particleeparticle contact resulted in differential particle charging depending on the particle surface chemistry. Once the protein-rich and carbohydrate-rich particles in the flour were charged, they were fed into the separation chamber and exposed to an electric field which deflected the particles according to the magnitude and type of charge. The separating chamber consisted of a plate-type separator in which one electrode in the form of a copper plate (66 cm height 26 cm width) with adjustable angle was placed. The plate is made of copper as a conductive material, and charged negatively with a voltage of 1 kV to 3 kV through the use of a high DC voltage supply. The fractions were collected in four different bins at the bottom of the separation chamber. Thirty g of pin milled navy bean flour was dried in a drying oven for 12 h at 70 C to reduce moisture content prior to being transferred to the fluidized bed. The flour was mechanically agitated and a dry air stream introduced at a constant flow rate of 7 LPM to suspend the particles in the air. The suspended particles were then pneumatically conveyed to the separation chamber through the tribo-charger tube to charge the particles. To evaluate the effect of acquired charge density on separation efficiency, four different straight or spiral tribo-charger tubes were constructed of PTFE the details of which are provided in Table 1. The charge densities of the particles after leaving each tribo-charger were measured in triplicate using a Faraday cup. After charging, the particles were directed to the separation chamber and separation studies performed at the voltages of 1 to 3 kV, and the plate angle of less than 20 . After complete electrostatic separation, the particles on the plate were collected and classified as the protein-enriched fraction (Bplate); particles accumulated in the three bins at the bottom of the separator (B1, B2 and B3) were collected and classified as the carbohydrate-enriched fractions. The weights of the collected protein- and carbohydrateenriched fractions were recorded. All of the fractions were analyzed for protein content. The yields (%) and percentages of total protein (%) were calculated based on the weight and protein content of each specific fraction using Eqs. (1) and (2), respectively.

Yieldð%Þ ¼ ðmassðgÞ of each fraction =massðgÞ of the starting navy bean flourÞ 100 (1)

24


Fig. 1. Process flow diagram of the tribo-electrostatic separator consisting of a fluidized bed (feed), tribocharging tube and separation chamber. The total height of the plate was 66 cm. The plate was divided to three sections of Ptop with the height of 15 cm, Pmiddle with the height of 15 cm, and Pbottom with the height of 36 cm.

Percentage of total proteinð%Þ ¼ ðproteinðgÞ in each fraction =proteinðgÞ in starting navy bean flourÞ 100 (2)

2.5. Analytical methods The moisture content of pin-milled navy bean flour was measured gravimetrically by drying in a forced-air oven at 110 C for 12 h. The ash content of the original navy bean flour was measured in triplicate based on AOCS Method Ba5a-49. The oil content of the original flour was determined in triplicate using the Mojonnier method (AOAC Method 922.06). The crude protein contents (N 6.25) of the starting navy bean flour and all of the protein and carbohydrate-enriched fractions produced by electrostatic separation were determined in triplicate by microdetermination of Kjeldahl nitrogen (Lang, 1958). The method consists of three main steps of acid digestion, Nesslerization and measurement of the resultant color at 420 nm using a spectrophotometer (Bio Tek Instruments, USA). 2.6. Statistical analysis The Least Significant Difference (LSD) analysis was conducted

using SAS (version 9.4, SAS Institute Inc., Cary, NC, USA) at a 5% significance level. 3. Results and discussion 3.1. Navy bean flour composition The composition of the original navy bean flour is given in Table 2. It contained 25.4% protein, 3.9% ash, 2.5% oil, and 6.6% moisture. The remaining 68.2% was attributed by difference to carbohydrates including fiber and other polysaccharides. 3.2. Effect of contact materials on tribocharging of navy bean flour The charge-to-mass ratios of the navy bean flour treated with different contact materials are presented in Fig. 2. The navy bean flour charged positively for all of the contact materials and amongst the materials tested, PTFE showed the highest triboelectrification efficiency followed by PVC and nylon. Copper was not a suitable material as the charge obtained was in the same range as that for non-treated flour (P < 0.05). The higher charge level obtained for navy bean flour contacted with PTFE can be attributed to the effective work function of PTFE (5.75 eV), which is higher than the work functions reported for PVC (5.13 eV), nylon (4.54 eV) and copper (4.38 eV) (Mazumder et al., 2006; Seanor, 1972; Trigwell et al., 2003). The charge transfer improved as the work function of the contacting materials

Table 1 Length and diameter of the PTFE tribo-charger tubes. Tribo-charger number

Tribo-charger type

Tribo-charger diameter (inch)

Tribo-charger length (cm)

Number of coils

Diameter of coils (cm)

I II III IV

spiral straight straight spiral

1/8 1/8 3/16 3/16

120.0 60.0 60.0 250.0

5 0 0 5

5 e e 15


25

Table 2 Composition of the pin-milled navy bean flour (dry-basis). Analysis

Units

Compositions

Protein content Oil content Moisture content Ash content Carbohydrate contenta

% % % % %

25.4 ± 1.7 2.5 ± 0.3 6.6 ± 0.5 3.9 ± 0.6 68.2

a The carbohydrate content was calculated by difference, and thus includes insoluble fiber.

increased (Fig. 3). A direct correlation between the acquired charge on the flour and the work function of the contact materials was observed. Consequently, PTFE was selected as the best contact material for use in the tribo-charger tube employed in this study.

Fig. 3. Charge (nC/g) acquired by navy bean flour based on the work function of the contact materials; copper (4.38 eV), nylon 66 (4.54 eV), PVC (5.13 eV), and PTFE (5.75 eV).

3.3. Effect of protein content on the charge density Table 3 presents the chargeability of various plant-derived flour sources ranging in protein and starch content. Soybean protein concentrates and navy bean flour acquired a net positive charge upon frictional contact with PTFE, whereas corn starch acquired a negative charge. Flours containing a higher protein content acquired more charge compared to those with a lower protein and higher carbohydrate content (Table 3). This observation is consistent with the literature (Kamiyama et al., 1994), indicating that the presence of easily ionizable functional groups or polar molecules is necessary to effectively charge the particles as the ion or electron transfer is the dominant tribo-charging mechanism (Kamiyama et al., 1994; Nemeth et al., 2003). Proteins should tribocharge effectively as they contain a variety of ionizable functional groups including the N-terminus amino group, C-terminus carboxyl group and the various ionizable functional groups on the side chains. In contrast, carbohydrates are characterized by low proton affinity and ionizability and should not tribocharge significantly. In this study, navy bean proteins have more negatively charged carboxyl and functional groups than the positively charged amino groups (with isoelectric point below 6), thus carrying a net negative chemical charge. We speculate that upon frictional contact with PTFE, the high work-function PTFE withdraws unbonded electrons on the carboxyl or other functional groups of the proteins, thus enhancing the triboelectrification efficiency by increasing the charge-to-mass ratios (Table 3). This difference in the tribocharging behavior of proteins and carbohydrates should enable the application of the triboelectric-based separation to possibly classify navy bean flour into protein- and

Fig. 2. Effect of contact materials on navy bean flour tribocharging. All values represent the charge-to-mass ratio means (n ¼ 6) ± standard deviation. Means sharing the same letters are not significantly different at P < 0.05.

carbohydrate-enriched fractions. 3.4. Effects of tribo-charger and plate's voltage on protein separation efficiency Given the ability to differentially tribocharge the bean flour constituents, a laboratory-scale electrostatic separator consisting of a fluidized bed, PTFE tribo-charger and the plate-type separation chamber was constructed (Fig. 1). During separation, the positively charged particles attached to the plate. As the proteins are positively charged after tribocharging with PTFE, the particles collected in the bin beneath the plate (BPlate) are protein-enriched particles. However, the uncharged or weakly charged particles are mostly carbohydrates and they were collected in three different bins distal to the negatively charged plate (B1, B2 and B3). The diameter, shape and length of the tribo-charger play an important role in transferring the required amounts of charge to the surface of the flour particles. The charge-to-mass ratio acquired by the flour particles should significantly affect the subsequent separation of the proteins and carbohydrates. To assure optimal tribocharging and subsequent separation, four different PTFE tribochargers were designed and tested (Table 1) to evaluate the effects of shape, diameter and length on the separation. The protein content, yield, and percentage of total protein of the carbohydrate- and protein-enriched fractions as well as the charge transferred were determined using each tribocharger (Table 4). To fully understand the effect of charge-to-mass ratio on the efficiency of the separation process, the plate of the separation chamber was divided into three parts (Ptop, Pmiddle and Pbottom) as shown in Fig. 1, and the protein content, yield, and percentage of total protein were determined (Table 5). The spiral-type tribo-charger (I) resulted in overcharging of the navy bean particles as indicated by the high charge-to-mass ratio. The configuration used in this charger likely provided a large number of particleeparticle and/or particle-wall collisions over a short period of time. The protein-enriched fraction (BPlate) for this charger had a protein content of 28.6% and accounted for 93.8% of the flour weight and 90.2% of the total protein content (Table 4). Overcharging of the particles prevented the effective separation of the protein- and carbohydrate-enriched fractions as only 6.2% of the flour was collected as the carbohydrate-enriched fraction (B1, B2 & B3). No significant difference was observed between the protein content of different fractions (BPlate, B1, B2 & B3) produced after electrostatic separation (Table 4). It should be noted that ~77% of the flour that entered the separator attached to the top and middle sections of the plate and contained only 24.5% protein (Table 5), comparable to the protein content of the navy bean flour feed

26


Table 3 Chargeability of various plant-derived flours using PTFE as a contact material for tribocharging. Materials

Protein content (wt%)

Carbohydrate content (wt%)

Charge (nC/g)

Soy protein concentrate Soy protein concentrate Navy bean flour Corn starch

66.7 41.6 25.4 0.0

22.2 29.2 68.2 90.0

23.7 16.5 5.5 2.1

± ± ± ±

2.9A 0.7B 0.4C 0.7D

All values represent the charge-to-mass ratio means (n ¼ 3) ± standard deviation. Means sharing the same letters are not significantly different at P < 0.05.

Table 4 Results of navy bean flour separation using different charger designs and separator conditions. Chargea (nC/g)

Fractions

Tribo-charger I V ¼ 1 KV

8440.0

Protein-rich fraction Carbohydrate-rich fractions

Tribo-charger II V ¼ 1 KV

854.8


Tribo-charger III V ¼ 3 KV

259.0


Tribo-charger IV V ¼ 3 KV

439.3


Yield (%) 93.8 1.6 1.7 2.9 88.7 2.7 3.0 5.6 33.4 14.5 18.5 33.5 25.6 17.7 17.1 39.6

Bplate B1 B2 B3 Bplate B1 B2 B3 Bplate B1 B2 B3 Bplate B1 B2 B3

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

Protein content (%)

1.5 0.2 0.2 1.1 4.4 0.9 0.7 2.7 9.4 2.9 1.7 4.8 8.9 3.0 0.6 10.6

28.6 26.3 23.7 25.3 30.9 23.1 22.4 23.8 40.3 25.3 24.7 22.8 40.8 22.6 20.9 20.2

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

5.6BC 1.2CD 1.6DEF 2.8DE 1.7B 3.8EFG 3.7EFG 1.7DEF 0.4A 1.7DE 2.8DE 1.8EFG 3.1A 1.1EFG 0.1FG 0.0G

Percentage of total protein 90.2 1.4 1.4 2.4 92.8 2.1 2.2 4.5 43.5 12.6 15.5 26.1 39.6 15.9 14.2 31.8

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

8.4A 0.0E 0.1E 0.5E 1.6A 0.6E 0.3E 2.3E 8.0B 3.2D 1.2D 4.3C 9.9B 4.5D 1.4D 10.5C

All of the electrostatic separations were performed at flow rate of 7 LPM. All the measurements were performed in duplicates. Means sharing the same letters in each column are not significantly different at P < 0.05. Details for the different tribocharger tubes are given in Table 1. The yields and percentages of total protein were calculated using Eqs. (1) and (2), respectively. The protein content of the feed flour was 25.4% as indicated in Table 2. a The charge-to-mass ratios (nc/g) represent the amounts of charge acquired using the combination of PVC fluidized bed reservoir and PTFE tribo-charger. The charge density acquired exclusively in the fluidized bed was found to be negligible at ~7.5 nc/g.

(25.4%) while the flour collected from the bottom section of the plate contained 47.0% protein. The collection of the carbohydraterich particles in the top and middle sections of the plate can be attributed to the structure of pin-milled navy bean flour and overcharging of the protein particles. Pin-milled navy bean flour is composed of free starch granules as well as starch granules surrounded by smaller protein particles (Tyler et al., 1981). The starch granules surrounded by overcharged protein particles may quickly attract to the high voltage plate right after entering the separator, thereby reducing the protein content in the fractions collected from the top and middle sections of the plate. Tribo-charger II had the same diameter as I, but was shorter in

length and contained no coils. The charge-to-mass ratio of the navy bean flour leaving this charger was 854.8 nC/g, significantly lower than the charge density obtained using charger I (8440.0 nC/g). This charger also resulted in overcharging of the particles, preventing effective separation into protein and carbohydrate-enriched fractions. Charger II in combination with separation at 1 kV resulted in the production of a protein-enriched fraction containing 30.9% protein, slightly higher than the protein content of the feed flour. Only 11.3% of the flour was collected in the bins located at the bottom of the separator and 88.7% was attached to the plate. 42% of the fraction collected at the top section of the plate had 21.5% protein, lower than the protein content of the feed flour (Table 5).

Table 5 Protein content of the particles attached to the top, middle, and bottom of the plate after electrostatic separation.

Tribo-charger I V ¼ 1 KV Tribo-charger II V ¼ 1 KV Tribo-charger III V ¼ 3 KV Tribo-charger IV V ¼ 3 KV

Protein-rich fractions (plate Parts)

Yield (%)

Ptop & middle Pbottom Ptop Pmiddle Pbottom Ptop Pmiddle Pbottom Ptop Pmiddle Pbottom

76.7 17.1 42.0 19.4 27.2 4.5 5.9 23.0 5.0 5.8 14.8

± ± ± ± ± ± ± ± ± ± ±

5.0 3.5 4.1 5.4 3.1 0.5 0.9 8.9 2.4 3.2 8.6

Protein content (%) 24.5 47.0 21.5 34.5 43.7 43.8 39.5 37.8 44.1 42.7 38.9

± ± ± ± ± ± ± ± ± ± ±

4.7F 4.6A 3.7F 2.6E 1.9AB 4.3AB 3.2BCD 2.1DE 0.1AB 0.5ABC 5.6CDE

All measurements were conducted in duplicate. Means sharing the same letters in each column are not significantly different at P < 0.05. The yields and percentages of total protein were calculated using Eqs. (1) and (2), respectively. The protein content of the feed flour was 25.4% as indicated in Table 2. Different portions of the plate are shown in Fig. 1.

Percentage of total protein 63.1 27.1 30.3 22.3 40.2 6.9 7.9 28.7 8.6 9.7 21.4

± ± ± ± ± ± ± ± ± ± ±

2.6A 5.8CD 0.3C 2.6D 0.7B 1.7E 1.2E 8.4C 3.6E 4.8E 8.5D


The fraction collected from the bottom section of the plate contained 43.7% protein accounting for 40.2% of the total protein content in the feed. In order to avoid overcharging and improve the protein separation, the charger diameter was increased to 3/16”. The charge-tomass ratios acquired by the flour particles by use of chargers III and IV resulted in improved separation of protein- and carbohydrateenriched particles. To compensate for the low charge-to-mass ratio and to ensure that particles deflect towards the plate electrode, the voltage of the plate was increased from 1 kV to 3 kV. The protein-enriched fraction produced using charger III at 3 kV had a protein content of 40.3% and accounted for 33.4% of the flour weight and 43.5% of the total protein content. All of the particles attached to the plate had a high protein content (37.8e43.8%, Table 5), and the particles collected at the bottom of the separator had a low protein content of 22.8e25.3% (Table 4), lower than the protein content of the starting flour. A similar separation behavior was also observed using charger IV. Charging the navy bean particles by the use of tribo-chargers III and IV followed by separation at a voltage 3 kV resulted in a recovery of over 40% of the protein into protein-rich fractions containing ~40% protein content. All of the resulting carbohydrate-rich fractions had similar (P < 0.05) protein contents of ~20.2e25.3%. Amongst the various combinations of tribo-charger and separation voltage, tribo-charger III with a separation voltage of 3 kV would appear optimal as it resulted in the highest yield of a fraction with a protein content of approximately 40%. 4. Conclusion A novel solvent-free electrostatic-based separation approach for the production of high quality protein powders from plant materials is described. The proposed approach provides an efficient, inexpensive, and environmentally friendly alternative to conventional protein separation processes in a solvent-free environment. The approach preserves the native bio-functionality of the protein and averts the likelihood of toxic microbial contamination common for currently used wet processes. A bench-scale solvent-free triboelectrostatic separator was developed and successfully applied in the fractionation of navy bean flour. PTFE was found to be the most suitable material for the tribo-charger. Optimum operating condition resulted in the production of protein-rich fractions with a protein content of 40.3% accounting for 43.5% of the total protein. Under different operating conditions, a fraction with a protein content of 47.0% was collected from the bottom section of the plate with a lower percentage of total protein (27.1%). This approach is superior to less-efficient sieving and expensive air classification systems, and could be further improved by optimizing the plate voltage, plate angle, flow rate of the separator and tribo-charger length and/or by using a multi-stage tribo-electrostatic approach. Acknowledgments This project was funded in part through Growing Forward 2 (GF2), a federal-provincial-territorial initiative, MITACS Canada, Ontario Centres of Excellence (OCE), Natural Sciences and Engineering Research Council (NSERC), and the industrial partner Advanced CERT Canada. The Agricultural Adaptation Council assists in the delivery of GF2 in Ontario. The authors would like to thank Margaret Carter from Lambton College for her assistance with the ash content measurement. References Ahn, Y.H., Yoo, J.S., 2001. Piperidine as an efficient organic catalyst of derivatization

27

of oligosaccharides with malononitrile for high-sensitivity electrospray ionization mass analysis. Anal. Sci. 17 (7), 893e895. Boye, J., Zare, F., Pletch, A., 2010. Pulse proteins: processing, characterization, functional properties and applications in food and feed. Food Res. Int. 43 (2), 414e431. Brooks, J.R., Morr, C.V., 1985. Effect of phytate removal treatments upon the molecular weight and subunit composition of major soy protein fractions. J. Agric. Food Chem. 33 (6), 1128e1132. Day, L., 2013. Proteins from land plants e potential resources for human nutrition and food security. Trends Food Sci. Technol. 32 (1), 25e42. Dijkink, B.H., Speranza, L., Paltsidis, D., Vereijken, J.M., 2007. Air dispersion of starch-protein mixtures: a predictive tool for air classification performance. Powder Technol. 172 (2), 113e119. Dwari, R.K., Hanumantha Rao, K., 2007. Dry beneficiation of coal e a review. Mineral Process. Extr. Metall. Rev. 28 (3), 177e234. Dwari, R.K., Hanumantha Rao, K., 2008. Non-coking coal preparation by novel triboelectrostatic method. Fuel 87 (17e18), 3562e3571. Dwari, R.K., Hanumantha Rao, K., 2009. Fine coal preparation using novel triboelectrostatic separator. Miner. Eng. 22 (2), 119e127. Kamiyama, M., Maeda, M., Okutani, H., Koyama, K., Matsuda, H., Sano, Y., 1994. Effect of functional groups on the triboelectric charging property of polymer particle. J. Appl. Polym. Sci. 51 (9), 1667e1671. Kovalyshyn, S.J., Shvets, O.P., Grundas, S., Tys, J., 2013. Use of the electro-separation method for improvement of the utility value of winter rapeseeds. Int. Agrophys. 27 (4), 419e424. Lang, C.A., 1958. Simple microdetermination of Kjeldahl nitrogen in biological materials. Anal. Chem. 30 (10), 1692e1694. Laudadio, V., Bastoni, E., Introna, M., Tufarelli, V., 2013. Production of low-fiber sunflower (Helianthus annuus L.) meal by micronization and air classification processes. Cyta J. Food 11 (4), 398e403. Li, T.X., Ban, H., Hower, J.C., Stencel, J.M., Saito, K., 1999. Dry triboelectrostatic separation of mineral particles: a potential application in space exploration. J. Electrost. 47 (3), 133e142. Lusas, E.W., Rhee, K.C., 1995. Soy protein processing and utilization. In: Erickson, D.R. (Ed.), Practical Handbook of Soybean Processing and Utilization. AOCS Press, Champaign, IL, pp. 117e160. Lusas, E.W., Riaz, M.N., 1995. Soy protein products: processing and use. J. Nutr. 125 (3 Suppl.), 573Se580S. Mayr, M.B., Barringer, S.A., 2006. Corona compared with triboelectric charging for electronic powder coating. J. Food Sci. 71 (4), E171eE177. Mazumder, M.K., Sims, R.A., Biris, A.S., Srirama, P.K., Saini, D., Yurteri, C.U., Trigwell, S., De, S., Sharma, R., 2006. Twenty-first century research needs in electrostatic processes applied to industry and medicine. Chem. Eng. Sci. 61 (7), 2192e2211. Mondor, M., Ippersiel, D., Lamarche, F., 2012. Electrodialysis in food processing. In: Boye, J.I., Arcand, Y. (Eds.), Green technologies in Food Production and Processing. Springer, Boston, MA, pp. 295e326. Mondor, M., Ippersiel, D., Lamarche, F., Boye, J.I., 2004. Production of soy protein concentrates using a combination of electroacidification and ultrafiltration. J. Agric. Food Chem. 52 (23), 6991e6996. Nemeth, E., Albrecht, V., Schubert, G., Simon, F., 2003. Polymer tribo-electric charging: dependence on thermodynamic surface properties and relative humidity. J. Electrost. 58 (1e2), 3e16. Omosaiye, O., Cheryan, M., 1979. Low-phytate, full-fat soy protein product by ultrafiltration of aqueous extracts of whole soybeans. Cereal Chem. 56 (2), 58e62. Pandya, T.S., Srinivasan, R., Thompson, C.P., 2013. Fiber separation from ground corn flour using an electrostatic method. Cereal Chem. 90 (6), 535e539. Park, C.H., Jeon, H.S., Yu, H.S., Han, O.H., Park, J.K., 2008. Application of electrostatic separation to the recycling of plastic wastes: separation of PVC, PET, and ABS. Environ. Sci. Technol. 42 (1), 249e255. Pelgrom, P.J.M., Wang, J., Boom, R.M., Schutyser, M.A.I., 2015. Pre- and posttreatment enhance the protein enrichment from milling and air classification of legumes. J. Food Eng. 155, 53e61. Petruccelli, S., Anon, M.C., 1994. Relationship between the method of obtention and the structural and functional properties of soy protein isolates. 1. Structural and hydration properties. J. Agric. Food Chem. 42 (10), 2161e2169. Rajabzadeh, A.R., Moresoli, C., Marcos, B., 2010. Fouling behavior of electroacidified soy protein extracts during cross-flow ultrafiltration using dynamic reversibleirreversible fouling resistances and CFD modeling. J. Membr. Sci. 361 (1e2), 191e205. Remadnia, M., Kachi, M., Messal, S., Oprean, A., Rouau, X., Dascalescu, L., 2014. Electrostatic separation of peeling and gluten from finely ground wheat grains. Part. Sci. Technol. 32 (6), 608e615. Schutyser, M.A.I., van der Goot, A.J., 2011. The potential of dry fractionation processes for sustainable plant protein production. Trends Food Sci. Technol. 22 (4), 154e164. Seanor, D.A., 1972. Electrical properties of polymers. In: Jenkins, A.D. (Ed.), Polymer Science, A Materials Science Handbook. North-Holland Pub. Co., Amsterdam, pp. 1267e1268. Skorepova, J., Moresoli, C., 2007. Carbohydrate and mineral removal during the production of low-phytate soy protein isolate by combined electroacidification and high shear tangential flow ultrafiltration. J. Agric. Food Chem. 55 (14), 5645e5652. Trigwell, S., Grable, N., Yurteri, C.U., Sharma, R., Mazumder, M.K., 2003. Effects of surface properties on the tribocharging characteristics of polymer powder as

28


applied to industrial processes. IEEE Trans. Ind. Appl. 39 (1), 79e86. Tyler, R.T., Youngs, C.G., Sosulski, F.W., 1981. Air classification of legumes .1. Separation efficiency, yield, and composition of the starch and protein-fractions.

Cereal Chem. 58 (2), 144e148. Wu, G., Li, J., Xu, Z., 2013. Triboelectrostatic separation for granular plastic waste recycling: a review. Waste Manag. 33 (3), 585e597.