- Email: [email protected]
Starch–protein separation from chickpea flour using a hydrocyclone
Journal of Food Engineering 82 (2007) 460–465 www.elsevier.com/locate/jfoodeng
Starch–protein separation from chickpea flour using a hydrocyclone Shahram Emami a,1, Lope G. Tabil a,*, Robert T. Tyler b,2, William J. Crerar a,1 a
Department of Agricultural and Bioresource Engineering, University of Saskatchewan, 57 Campus Drive, Saskatoon, SK, Canada S7N 5A9 Department of Applied Microbiology and Food Science, University of Saskatchewan, 51 Campus Drive, Saskatoon, SK, Canada S7N 5A8
b
Received 17 January 2006; received in revised form 27 February 2007; accepted 3 March 2007 Available online 12 March 2007
Abstract The effect of defatting and suspension pH on starch–protein separation from chickpea flour using a hydrocyclone was evaluated. Suspensions were made from whole and defatted chickpea flours in distilled water at pH 6.6 (initial pH of suspension) and pH 9.0. The suspensions were subjected to double-pass hydrocyclone process. The use of defatted flour resulted in higher total solids in the underflows. Defatting of the whole chickpea flour and high suspension pH (pH 9.0) improved starch and protein separation efficiencies. In terms of starch separation, the use of defatted flour at pH 9.0 resulted in the highest starch content in the underflow with separation efficiency of 99.8%. Defatting chickpea flour and using high pH resulted in 2.0 times enrichment of the starch fraction of the second-pass underflow (U2) and 3.4 times enrichment of the protein fraction in the sediment of the second-pass overflow (O2) compared to the original chickpea flour which contained 48.0% d.b. starch and 26.3% d.b. protein. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Fractionation; Liquid cyclone separation; Chickpea; Flour; Separation efficiency
1. Introduction Legumes, such as chickpea, are good sources of starch and protein. Researchers have tried different methods to separate starch and protein from legumes and improve starch and protein purification. The protein and starch fractions of legumes can be used as ingredients in food (Neves & Lourenco, 1995; Tian, Kyle, & Small, 1999) including human food, pet food, animal feed, as well as in processing of non-food products (Sa´nchez-Vioque, Clemente, Vioque, Bautista, & Milla´n, 1999), and in some industries such as mining, paper, cosmetic, and pharmaceutical industries (International Starch Institute, 2003). The isolation of starch fraction from legume seeds is difficult because of the presence of insoluble flocculent protein
*
Corresponding author. Tel.: +1 306 966 5317; fax: +1 306 966 5334. E-mail addresses: [email protected] (S. Emami), lope.tabil@ usask.ca (L.G. Tabil), [email protected] (R.T. Tyler), bill.crerar@ usask.ca (W.J. Crerar). 1 Fax: +1 306 966 5334. 2 Fax: +1 306 966 8894. 0260-8774/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2007.03.002
and fine fiber which diminishes sedimentation, thus co-settling with the starch fraction and resulting in a brownish deposit (Hoover & Sosulski, 1991; Ratnayake, Hoover, & Warkentin, 2002). In addition, the relatively high fat content of chickpea seeds prevents high separation efficiency and results in low yield of fractions (Sosulski, Walker, Fedec, & Tyler, 1987). Research work on starch separation from legume seeds which involved the separation of starch from protein using isoelectric method, has been undertaken by Anderson and Romo (1976), and Colonna, Gueguen, and Mercier (1981). The advantage of this method is the low cost of chemicals used for processing (Sa´nchez-Vioque et al., 1999). Colonna et al. (1981) reported a study where the protein of smooth pea and broad bean was dissolved using an aqueous medium at pH 9.0; then the starch fraction was obtained by sieving and washing. Tian et al. (1999) used an aqueous medium at pH 9.0 to dissolve protein of field pea flour followed by decantation and filtration to remove solid residuals; to recover protein, the protein extract was adjusted to pH 4.5 and centrifuged. According to Anderson and Romo (1976), starch–protein separation pH ranging between 5.5 and 7.5 is the likely cause of
S. Emami et al. / Journal of Food Engineering 82 (2007) 460–465
contamination of lentil and field pea starch fraction with protein. The effect of high pH values, i.e. up to pH 9.5, on the protein contamination of starch was minimal, although higher pH increased protein solubility. However, starch yield was not affected by pH of the medium (Anderson & Romo, 1976). Vose (1980) also applied the isoelectric method in combination with a hydrocyclone to separate the starch and protein from field pea and horsebean. Hydrocyclone uses centrifugal force to separate two immiscible liquids, solid from liquid, or gas from liquid. The advantages of hydrocyclone are: (a) flexibility of application such as clarifying, thickening, degassing, and sorting, (b) applicability to different degree of separation, (c) small plant area required for installation, and (d) low installation, operation, and maintenance costs (Cheremisinoff, 1995; Svarovsky, 1984). Vose (1980) reported that a feed concentration of 25% solids resulted in overflows containing 3–8% solids with 60–70% protein. Protein recovery was performed by reducing pH to isoelectric point followed by centrifugation at 1500 g; the cake contained 85% protein (Vose, 1980). Dehulled chickpea grains contain 3.1–6.9% lipid, 20.5– 30.5% protein, and 63.0–65.0% carbohydrates. Starch is the major component of chickpea grain and constitutes 55.3–58.1% of the dehulled seed (Chavan, Kadam, & Salunkhe, 1986). Chickpea grain has higher lipid content than many pulse grains; it makes fractionation of starch and protein difficult and diminishes separation efficiencies. Previous studies confirmed that air classification of chickpea flour was not performed as successfully as other legumes (Sosulski et al., 1987). With regards to wet processing, studies have shown that there was no investigation conducted on starch–protein separation from chickpea flour using high pH (at which nitrogen solubility index is maximum) in combination with hydrocyclone. The overall objective of this study was to improve the starch–protein separation for chickpea flour. Specifically, the effects of pH and defatting on starch–protein separation from chickpea flour using a hydrocyclone were evaluated.
461
by exposing it under the fume hood. The whole chickpea flour and defatted chickpea flour were used in the trials. 2.2. Processing To determine the pH in which chickpea flour protein had the highest and lowest solubility, the nitrogen solubility index (NSI) was measured. The NSI of chickpea flour was determined according to AACC approved method 46–23 (AACC, 2000) in duplicate at pH range of 2.0–10.0 with 1-unit increment. The nitrogen content was measured using LECO Model FP-528 Nitrogen/Protein Determinator (LECO Corporation, St. Joseph, MI). Fig. 1 presents the NSI of chickpea flour. The graph demonstrates that at pH values of 4.3 and 9.0, the sample had the lowest and highest nitrogen solubility, respectively. The NSI trend obtained was similar to those reported by other researchers (Bencini, 1986; Sa´nchez-Vioque et al., 1999). Therefore, a suspension pH of 9.0 was tested aside from pH 6.6 to disperse protein particles in the aqueous medium. The hydrocyclone system (Fig. 2) included a feed tank and a 10-mm hydrocyclone (Dorrclone, GL& V Canada Inc., Orillia, ON) connected to a positive displacement
Fig. 1. Nitrogen solubility profile of chickpea flour.
2. Materials and methods
Pressure gage
Valve Overflow
2.1. Sample preparation Pressure gage
Feeding tank
Commercial dehulled split desi chickpea (dhal) from the crop harvested in fall of 2003 were obtained from Canadian Select Grains of Eston, Saskatchewan, Canada. The split chickpea grain was stored in a walk-in cooler maintained at a temperature of 2 ± 2 °C. It was then milled using a pin mill (GM 280/S-D, Condux werk, Hannau, Wolfgang, Germany). The pin mill had two discs: the first disc had 86 pins rotating at 8034 rpm; and the second disc had 108 pins and was stationary. A portion of the whole chickpea flour was defatted by ACS-grade isopropyl alcohol (EMD, Gibbstown, NJ) using Soxhlet extraction system. The remaining alcohol in the sample was dried
Valve Pressure gage
Hydrocyclone Valve
Pump
Pressure gage
Valve Underflow
Fig. 2. Schematic of the hydrocyclone system.
462
S. Emami et al. / Journal of Food Engineering 82 (2007) 460–465
pump (Model 4100C, Hypro Inc., St. Paul, MN) running at 1950 rpm. The Dorrclone unit consisted of four 10-mm hydrocyclones operating in parallel. Three of these hydrocyclones were removed and the vacant ports were plugged using rubber stoppers. Using only one hydrocyclone reduced the feed requirement and increased the operating pressure, making the unit suitable for lab-scale trials. A by-pass valve, located between the pump and feed tank, was employed to circulate the suspension, keep flour particles in suspension form, and help control pressure. A valve was located before the cyclone to control inlet pressure. The overflow and underflow valves were kept fully-opened during the test. Suspension was made using flour (whole chickpea flour or defatted chickpea flour) and deionized water at a concentration of 1.5% (w/w). The measured pH of the suspension was 6.6. To obtain a suspension pH of 9.0, a 10 M NaOH solution was used. The suspension, which was left overnight, was stirred for one hour using a mixer (OMINI-MIXER 17105, Sorvall Inc., Newtown, CT) before feeding it to the hydrocyclone. The suspension was fed to the hydrocyclone at two pH values of 6.6 and 9.0. The separation process in the hydrocyclone is shown schematically in Fig. 3. The suspension was passed through the hydrocyclone at an inlet pressure of 690 kPa resulting in first-pass overflow (O1) and underflow (U1). O1 and U1 were separately passed through the hydrocyclone to obtain second-pass overflow (O2) and underflow (U2) products. O2 was left overnight to obtain sediment (SE) and supernatant (SU). SU was poured to another container and the protein contents of both SE and SU were determined.
used for legumes by other researchers (Han & Khan, 1990; Reichert, 1982; Reichert & Youngs, 1978; Sa´nchezVioque et al., 1999; Sosulski, Hoover, Tyler, Murray, & Arntfield, 1985). Starch content was assessed using the method described by Holm, Bjorck, and Asp (1986). This measurement involves the gelatinization of starch in boiling water bath in the presence of a-amylase, followed by 30 min incubation of the subsample in the presence of amyloglucosidase, followed by addition of glucose oxidase/peroxidase reagent, and determination of glucose content using a spectrophotometer at a wavelength of 510 nm. The moisture content was measured using the method described by Egan, Ronald, and Ronald (1990) for determining the moisture content of syrup and condensed milk. For this purpose, the vacuum oven was set at 70 °C, and a vacuum gage pressure of 3.33 kPa was applied. All measurements were conducted in triplicates. The yield of fraction was measured based on recovery percentage. Starch and protein separation efficiencies were calculated according to Eq. (1) (Tyler, Youngs, & Sosulski, 1981) with some modification for dilute materials (Eq. (1)). S:E: ¼
TSP C P 100 TSF C F
ð1Þ
where S.E. is separation efficiency (%), TSP is total solid of product (kg), CP is component content in the product (% d.b.), TSF is total solid of feed material (kg), and CF is component content in the feed material (% d.b.). The values were calculated with respect to the starch and protein contents of the feed material into the hydrocyclone. 2.4. Statistical analysis
2.3. Analytical and physical measurements Protein content was measured by AACC method 46–30 (AACC, 2000) using LECO Model FP-528. A factor of 6.25 was used to convert nitrogen to protein content as
The Statistical Analysis System (SAS, 2003) was used to perform analysis of variance and mean values were compared using Duncan’s multiple range test and t-test, where appropriate. Overflow (O2) Supernatant (SU) Sediment (SE)
First-pass Overflow (O1) Cyclone
Inlet
Cyclone Underflow (UO1) Overflow (OU1)
First-pass Underflow (U1)
Supernatant (SU) Sediment (SE) Supernatant (SU) Sediment (SE)
Cyclone
Underflow (U2)
Supernatant (SU) Sediment (SE)
Fig. 3. Schematic of the extraction process in the hydrocyclone.
S. Emami et al. / Journal of Food Engineering 82 (2007) 460–465
3. Results and discussion Table 1 shows the fraction yield and total solids content of the overflows and the underflows. Because some material (suspension) remained in the hydrocyclone, the sum of the overflow and underflow in some runs were less than 100%. The total solids content of the underflow products was more than that of the overflow products showing that large and dense solid particles, i.e. starch granules, were collected in the underflow. This result is in agreement with that of Svarovsky (1990) who explained that coarse particles in the feed material settle on the hydrocyclone wall and collect in the underflow. 3.1. Starch separation
Table 2 Starch contenta (% d.b.) of the overflow and underflow using different feed materials and pH values Fraction Whole flourb Overflow
Underflow
Defatted flourc Overflow
Underflow
The starch content of the overflows and the underflows is presented in Table 2. The effect of the double-pass hydrocyclone process on starch content of the underflows (U1, U2, and UO1 (underflow of the second-pass of O1)) was significant for both defatted and whole chickpea flour at two pH levels. U2 had significantly higher (a = 0.05) starch content (1.2 to 1.4 times) than U1. This result clearly shows that the double-pass hydrocyclone process was better than the single-pass. The starch content of U2 from whole flour was 86.6% and 90.9% at pH 6.6 and 9.0, respectively. Statistical analysis (t-test) between U2 at two pH levels showed that starch content was significantly higher (a = 0.05) at pH 9.0 than at pH 6.6. Since at pH 9.0 (maximum NSI) the protein materials were better dispersed in the aqueous media than at pH 6.6, more of the protein went into the overflow; this phenomenon resulted in less protein contamination in the underflow. The small amount of starch in the overflows (O1, O2, and OU1 (overflow of
463
pH 6.6
pH 9.0
O1 O2 OU1
3.4 b 1.9 c 5.2 a
3.5 b 1.9 c 5.5 a
U1 U2 UO1
63.2 b 86.6 a 5.8 c
67.1 b 90.9 a 3.6 c
O1 O2 OU1
7.6 a 1.1 b 1.1 b
0.7 a 0.6 a 1.1 a
U1 U2 UO1
77.8 b 93.1 a 14.7 c
79.9 b 99.7 a 0.84 c
O1 = Overflow of the first-pass, O2 = Overflow of the second-pass of O1, U1 = Underflow of the first-pass, U2 = Underflow of the second-pass of U1, OU1 = Overflow of the second-pass of U1, UO1 = Underflow of the second-pass of O1. a–c values in the same column in each group of flour followed by common letter are not significantly different at 5% level. a Values are average of three determinations. b Whole flour starch content = 48.0% d.b. c Defatted flour starch content = 51.0% d.b.
the second-pass of U1)) is likely associated with damaged starch granules resulting from pin milling (Tyler et al., 1981). For the defatted flour, the starch content of U2 at pH 9.0 (99.7%) was significantly higher (by t-test at a = 0.05) than at pH 6.6 (93.1%). Among all applied conditions, the double-pass hydrocyclone process using defatted flour at pH 9.0 resulted in the highest starch content (99.7%
Table 1 Fraction yield and total solida of the overflow and underflow using different feed materials and pH values Fraction
Whole flour Overflow
Underflow
Defatted flour Overflow
Underflow
pH 6.6
pH 9.0
Fraction yield (%)
Total solids (% w.b.)
Fraction yieldb (%)
Total solids (% w.b.)
O1 O2 OU1
50.2 48.4 47.9
0.5 0.5 0.5
49.3 46.1 47.7
0.6 0.6 0.6
U1 U2 UO1
49.8 52.1 51.7
1.9 2.7 0.5
47.1 52.8 53.9
2.0 2.7 0.8
O1 O2 OU1
49.0 49.0 46.5
0.6 0.4 0.7
47.3 47.4 46.3
0.6 0.5 0.5
U1 U2 UO1
49.0 53.5 51.0
2.0 3.0 0.6
49.1 53.7 52.7
1.9 2.8 0.5
O1 = Overflow of the first-pass, O2 = Overflow of the second-pass of O1, U1 = Underflow of the first-pass, U2 = Underflow of the second-pass of U1, OU1 = Overflow of the second-pass of U1, UO1 = Underflow of the second-pass of O1. a Values are average of three determinations. b Because of unavoidable residual material in the system, the sum of the overflow and underflow yield is less than 100%.
464
S. Emami et al. / Journal of Food Engineering 82 (2007) 460–465
starch). The starch contents obtained in the current study were similar to those reported by Vose (1980) for field pea and horsebean employing a high pH (pH 8.5) and a series of three hydrocyclones, where the starch fraction obtained had 99% starch. Employing high pH (pH 10.2 for field pea and pH 9.6 for fababean) but using centrifugation, Sosulski and Sosulski (1986) obtained starch fraction with a starch content of as high as 94% from field pea and fababean, which was lower than value obtained in this study. Table 3 shows the starch separation efficiency values in the underflows. The starch separation efficiencies of U1 and U2 ranged between 93.8% and 99.8% and these values were similar to those reported by Tyler et al. (1981) for air classification of legumes. However, values from this study were higher than those reported by Sosulski and Sosulski (1986) who employed high pH and centrifugation in separating starch from field pea and fababean (76–79%). In this study, the lowest starch separation efficiency was obtained from UO1. 3.2. Protein separation
Defatted flour Underflow
Fraction Whole flourb Overflow
pH 6.6
Defatted flourc Overflow
Underflow
pH 6.6
pH 9.0
SE
SU
SE
SU
O1 O2 OU1
66.9 a 68.5 a 64.0 b
69.6 a 64.1 a 72.1 a
68.4 b 70.3 a 65.7 c
67.0 a 68.3 a 65.7 a
U1 U2 UO1
8.0 b 7.0 b 52.3 a
34.1 b 31.5 c 67.1 a
7.0 b 5.4 c 53.2 a
22.4 b 20.1 c 55.3 a
O1 O2 OU1
72.8 c 97.8 a 88.3 b
68.1 a 82.4 a 72.7 a
92.5 b 98.9 a 88.3 c
63.5 ab 66.1 a 59.1 b
U1 U2 UO1
18.5 b 12.0 c 58.1 a
67.1 a 65.6 a 69.6 a
16.6 b 11.1 c 90.1 a
61.1 a 60.4 a 60.5 a
SE = Sediment, SU = Supernatant, O1 = Overflow of the first-pass, O2 = Overflow of the second-pass of O1, U1 = Underflow of the firstpass, U2 = Underflow of the second-pass of U1, OU1 = Overflow of the second-pass of U1, UO1 = Underflow of the second-pass of O1. a–c values in the same column in each group of flour followed by common letter are not significantly different at 5% level. a Values are average of three determinations. b Whole flour protein content = 26.3% d.b. c Defatted flour protein content = 27.9% d.b.
Table 5 Protein separation efficiencya (%) achieved in the overflows
Table 3 Starch separation efficiencya (%) achieved in the underflows
Whole flour Underflow
Table 4 Protein contenta (% d.b.) of sediment and supernatant of the overflow and underflow using different feed materials and pH values
Underflow
Table 4 shows the protein content of the overflows and the underflows. For the whole flour and defatted flour, O1 had higher protein content than U1 at both pH levels. SEs of the overflow and the underflow had very high difference in protein contents at both pH levels. Although high pH improved the protein contents of the overflow SEs as well as SUs, statistically (t-test) it did not have any significant effect (a = 0.05). The protein contents of SEs from O2 were 2.5–2.7 times that of the whole flour, whereas they were 3.5 times that of the defatted flour. UO1 had still high protein content which may be due to the agglomeration of starch granules and proteinaceous material. The effect of defatting on the protein content of the overflow SEs was significant (a = 0.05) which is desirable since protein is concentrated in the SEs. However, it was not significant (a = 0.05) for the overflow SUs. Using the double-pass hydrocyclone
Feed material
process, the protein contents of the overflow products (O2 and OU1), where they are expected to accumulate, did not increase because the protein particles were still dispersed in the media. Adjusting the pH of the overflow to isoelectric point (pH 4.3) could have improved protein separation and increased protein content. Table 5 shows the protein separation efficiency values. For both whole flour and defatted flour, the high pH improved the protein separation efficiencies due to the higher solubility of the chickpea protein at pH 9.0 than pH 6.6. The values in this study were similar to those
pH 9.0
U1 U2 UO1
93.8 ± 2.4 99.4 ± 0.9 73.7 ± 1.3
97.2 ± 0.6 96.6 ± 1.2 72.2 ± 0.6
U1 U2 UO1
96.8 ± 0.4 99.8 ± 0.3 98.3 ± 0.3
96.7 ± 1.6 99.8 ± 0.2 61.9 ± 0.4
O1 = Overflow of the first-pass, O2 = Overflow of the second-pass of O1, U1 = Underflow of the first-pass, U2 = Underflow of the second-pass of U1, OU1 = Overflow of the second-pass of U1, UO1 = Underflow of the second-pass of O1. a Values are average of three determinations. Value following the mean is standard error.
Feed material Whole flour Overflow
Defatted flour Overflow
pH 6.6
pH 9.0
O1 O2 OU1
65.7 ± 1.2 48.5 ± 0.7 51.2 ± 0.6
70.1 ± 1.1 50.1 ± 0.9 55.7 ± 0.8
O1 O2 OU1
67.5 ± 0.8 74.3 ± 0.4 62.5 ± 0.3
83.1 ± 0.9 51.9 ± 0.8 67.4 ± 0.8
O1 = Overflow of the first-pass, O2 = Overflow of the second-pass of O1, U1 = Underflow of the first-pass, U2 = Underflow of the second-pass of U1, OU1 = Overflow of the second-pass of U1, UO1 = Underflow of the second-pass of O1. a Values are average of three determinations. Value following the mean is standard error.
S. Emami et al. / Journal of Food Engineering 82 (2007) 460–465
reported by Sosulski and Sosulski (1986) (72–74%) for field pea and fababean using high pH and centrifugation. 4. Conclusion The effect of the double-pass hydrocyclone process was significant on starch content of the underflow products (U1 and U2) at both pH 6.6 and 9.0. The use of pH 9.0 and defatting of chickpea flour significantly increased the starch content of U2. However, pH 9.0 did not significantly affect the protein content of the SEs and SUs of the overflows. Defatting of chickpea flour significantly increased the protein content of overflow SEs; whereas, it did not have any significant effect on the protein content of the overflow SUs. U2 (underflow product resulting from the second-pass process) had high starch content ranging from 86.6% to 99.7% with starch separation efficiencies ranging from 99.4% to 99.8%. Overall, using pH 9.0, defatting, and double-pass hydrocyclone process improved the starch–protein separation from chickpea flour. Acknowledgements The authors would like to acknowledge the following: (1) the technical assistance of Douglas Hassard and Connie Perron of the Crop Development Centre (College of Agriculture, University of Saskatchewan) and Ms. Hong Qi of the Centre of Agri-Industrial Technology (Alberta Agriculture and Rural Development), (2) the Natural Science Engineering Research Council (NSERC) and Agriculture Development Fund (ADF) for financial support, (3) Canada-Saskatchewan Agri-Food Innovation Fund for the new Bioprocess Engineering Research Lab, and (4) Pilot Plant Crop (POS) for loaning us the hydrocyclone unit. References AACC (2000). Approved Methods of the American Association of Cereal Chemists. Method 46-23 and Method 46-30 (10th ed.). St. Paul, MN: American Association of Cereal Chemists. Anderson, C. G., & Romo, G. R. (1976). Influence of extraction medium pH on the protein content of some legume starches. Journal of Food Technology, 11(6), 647–654. Bencini, M. C. (1986). Functional properties of drum-died chickpea (Cicer arietinum L.) flours. Journal of Food Science, 51(6), 1518–1526. Chavan, J. K., Kadam, S. S., & Salunkhe, D. K. (1986). Biotechnology and technology of chickpea (Cicer arietinum L.) seeds. CRC Critical Reviews in Food Science and Nutrition, 25(2), 107–158. Cheremisinoff, P. N. (1995). Processing Engineering Handbook SeriesSolid/liquid. Lancaster, PA: Technomic Publishing Co. Inc.
465
Colonna, P., Gueguen, J., & Mercier, C. (1981). Pilot scale preparation of starch and cell wall material from Pisum sativum and Vicia faba peas, broadbeans, flours, composition. Sciences des Aliments, 1(3), 415–426. Egan, H., Ronald, S. K., & Ronald, S. (1990). Pearson’s Chemical Analysis of Foods (8th ed.). London, UK: Longman Scientific and Technical. Han, J., & Khan, K. (1990). Physicochemical studies of pin-milled and airclassified dry edible bean fractions. Cereal Chemistry, 67(4), 384–390. Holm, J., Bjorck, A. D., & Asp, N. G. (1986). A rapid method for the analysis of starch. Starch, 38(7), 224–226. Hoover, R., & Sosulski, F. W. (1991). Composition, and chemical modification of legume starches: A review. Canadian Journal of Physiology and Pharmacology, 69(1), 79–92. International Starch Institute. (2003). Tapioca Starch Applications. Internet document, (accessed 19.03.2006). Neves, V. A., & Lourenco, E. J. (1995). Isolation and in vitro hydrolysis of globulin G1 from lentils (Lens culinaris, Medik). Journal of Food Biochemistry, 19(2), 109–120. Ratnayake, W. S., Hoover, R., & Warkentin, T. (2002). Pea starch: Composition, structure and properties – a review. Starch, 54(6), 217–234. Reichert, R. D. (1982). Air classification of peas (Pisum sativum) varying widely in protein content. Journal of Food Science, 47(4), 1263–1267, 1271. Reichert, R. D., & Youngs, C. G. (1978). Nature of residual protein associated with starch fractions from air-classified field peas. Cereal Chemistry, 55(4), 469–480. Sa´nchez-Vioque, R., Clemente, A., Vioque, J., Bautista, J., & Milla´n, F. (1999). Protein isolate from chickpea (Cicer arietinum L.): Chemical composition, functional properties and protein characterization. Food Chemistry, 64(2), 237–243. SAS. (2003). SAS User’s Guide: Statistics. Ver. 9.1. Cary, NC: Statistical Analysis System Inc. Sosulski, F. W., Hoover, R., Tyler, R. T., Murray, E. D., & Arntfield, S. D. (1985). Differential scanning calorimetry of air-classified starch and protein fraction from eight legume species. Starch, 37(8), 257–262. Sosulski, F.W., & Sosulski, K. (1986). Composition and functionality of protein, starch, and fiber from wet and dry processing of grain legumes. In American Chemical Society Symposium (pp. 176–189), Washington, DC. Sosulski, F. W., Walker, A. F., Fedec, P., & Tyler, R. T. (1987). Comparison of air classification for separation of protein and starch pin-milled legume flours. Food Science and Technology, 20(5), 221–225. Svarovsky, L. (1984). Hydrocyclones. London, UK: Technomic Publishing Co. Inc.. Svarovsky, L. (1990). Solid-Liquid Separation (3rd ed.). London, UK: Butterworth. Tian, S., Kyle, S. A. W., & Small, D. M. (1999). Pilot scale isolation of protein from field peas (Pisum sativum L.) for use as food ingredients. International Journal of Food Science and Technology, 34(1), 33–39. Tyler, R. T., Youngs, C. G., & Sosulski, F. W. (1981). Air classification of legumes. I. separation efficiency, yield, and composition of the starch and protein fractions. Cereal Chemistry, 58(2), 144–147. Vose, J. R. (1980). Production and functionality of starch and protein isolates from legume seeds (field peas and horsebeans). Cereal Chemistry, 57(5), 406–410.
Starch–protein separation from chickpea flour using a hydrocyclone Shahram Emami a,1, Lope G. Tabil a,*, Robert T. Tyler b,2, William J. Crerar a,1 a
Department of Agricultural and Bioresource Engineering, University of Saskatchewan, 57 Campus Drive, Saskatoon, SK, Canada S7N 5A9 Department of Applied Microbiology and Food Science, University of Saskatchewan, 51 Campus Drive, Saskatoon, SK, Canada S7N 5A8
b
Received 17 January 2006; received in revised form 27 February 2007; accepted 3 March 2007 Available online 12 March 2007
Abstract The effect of defatting and suspension pH on starch–protein separation from chickpea flour using a hydrocyclone was evaluated. Suspensions were made from whole and defatted chickpea flours in distilled water at pH 6.6 (initial pH of suspension) and pH 9.0. The suspensions were subjected to double-pass hydrocyclone process. The use of defatted flour resulted in higher total solids in the underflows. Defatting of the whole chickpea flour and high suspension pH (pH 9.0) improved starch and protein separation efficiencies. In terms of starch separation, the use of defatted flour at pH 9.0 resulted in the highest starch content in the underflow with separation efficiency of 99.8%. Defatting chickpea flour and using high pH resulted in 2.0 times enrichment of the starch fraction of the second-pass underflow (U2) and 3.4 times enrichment of the protein fraction in the sediment of the second-pass overflow (O2) compared to the original chickpea flour which contained 48.0% d.b. starch and 26.3% d.b. protein. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Fractionation; Liquid cyclone separation; Chickpea; Flour; Separation efficiency
1. Introduction Legumes, such as chickpea, are good sources of starch and protein. Researchers have tried different methods to separate starch and protein from legumes and improve starch and protein purification. The protein and starch fractions of legumes can be used as ingredients in food (Neves & Lourenco, 1995; Tian, Kyle, & Small, 1999) including human food, pet food, animal feed, as well as in processing of non-food products (Sa´nchez-Vioque, Clemente, Vioque, Bautista, & Milla´n, 1999), and in some industries such as mining, paper, cosmetic, and pharmaceutical industries (International Starch Institute, 2003). The isolation of starch fraction from legume seeds is difficult because of the presence of insoluble flocculent protein
*
Corresponding author. Tel.: +1 306 966 5317; fax: +1 306 966 5334. E-mail addresses: [email protected] (S. Emami), lope.tabil@ usask.ca (L.G. Tabil), [email protected] (R.T. Tyler), bill.crerar@ usask.ca (W.J. Crerar). 1 Fax: +1 306 966 5334. 2 Fax: +1 306 966 8894. 0260-8774/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2007.03.002
and fine fiber which diminishes sedimentation, thus co-settling with the starch fraction and resulting in a brownish deposit (Hoover & Sosulski, 1991; Ratnayake, Hoover, & Warkentin, 2002). In addition, the relatively high fat content of chickpea seeds prevents high separation efficiency and results in low yield of fractions (Sosulski, Walker, Fedec, & Tyler, 1987). Research work on starch separation from legume seeds which involved the separation of starch from protein using isoelectric method, has been undertaken by Anderson and Romo (1976), and Colonna, Gueguen, and Mercier (1981). The advantage of this method is the low cost of chemicals used for processing (Sa´nchez-Vioque et al., 1999). Colonna et al. (1981) reported a study where the protein of smooth pea and broad bean was dissolved using an aqueous medium at pH 9.0; then the starch fraction was obtained by sieving and washing. Tian et al. (1999) used an aqueous medium at pH 9.0 to dissolve protein of field pea flour followed by decantation and filtration to remove solid residuals; to recover protein, the protein extract was adjusted to pH 4.5 and centrifuged. According to Anderson and Romo (1976), starch–protein separation pH ranging between 5.5 and 7.5 is the likely cause of
S. Emami et al. / Journal of Food Engineering 82 (2007) 460–465
contamination of lentil and field pea starch fraction with protein. The effect of high pH values, i.e. up to pH 9.5, on the protein contamination of starch was minimal, although higher pH increased protein solubility. However, starch yield was not affected by pH of the medium (Anderson & Romo, 1976). Vose (1980) also applied the isoelectric method in combination with a hydrocyclone to separate the starch and protein from field pea and horsebean. Hydrocyclone uses centrifugal force to separate two immiscible liquids, solid from liquid, or gas from liquid. The advantages of hydrocyclone are: (a) flexibility of application such as clarifying, thickening, degassing, and sorting, (b) applicability to different degree of separation, (c) small plant area required for installation, and (d) low installation, operation, and maintenance costs (Cheremisinoff, 1995; Svarovsky, 1984). Vose (1980) reported that a feed concentration of 25% solids resulted in overflows containing 3–8% solids with 60–70% protein. Protein recovery was performed by reducing pH to isoelectric point followed by centrifugation at 1500 g; the cake contained 85% protein (Vose, 1980). Dehulled chickpea grains contain 3.1–6.9% lipid, 20.5– 30.5% protein, and 63.0–65.0% carbohydrates. Starch is the major component of chickpea grain and constitutes 55.3–58.1% of the dehulled seed (Chavan, Kadam, & Salunkhe, 1986). Chickpea grain has higher lipid content than many pulse grains; it makes fractionation of starch and protein difficult and diminishes separation efficiencies. Previous studies confirmed that air classification of chickpea flour was not performed as successfully as other legumes (Sosulski et al., 1987). With regards to wet processing, studies have shown that there was no investigation conducted on starch–protein separation from chickpea flour using high pH (at which nitrogen solubility index is maximum) in combination with hydrocyclone. The overall objective of this study was to improve the starch–protein separation for chickpea flour. Specifically, the effects of pH and defatting on starch–protein separation from chickpea flour using a hydrocyclone were evaluated.
461
by exposing it under the fume hood. The whole chickpea flour and defatted chickpea flour were used in the trials. 2.2. Processing To determine the pH in which chickpea flour protein had the highest and lowest solubility, the nitrogen solubility index (NSI) was measured. The NSI of chickpea flour was determined according to AACC approved method 46–23 (AACC, 2000) in duplicate at pH range of 2.0–10.0 with 1-unit increment. The nitrogen content was measured using LECO Model FP-528 Nitrogen/Protein Determinator (LECO Corporation, St. Joseph, MI). Fig. 1 presents the NSI of chickpea flour. The graph demonstrates that at pH values of 4.3 and 9.0, the sample had the lowest and highest nitrogen solubility, respectively. The NSI trend obtained was similar to those reported by other researchers (Bencini, 1986; Sa´nchez-Vioque et al., 1999). Therefore, a suspension pH of 9.0 was tested aside from pH 6.6 to disperse protein particles in the aqueous medium. The hydrocyclone system (Fig. 2) included a feed tank and a 10-mm hydrocyclone (Dorrclone, GL& V Canada Inc., Orillia, ON) connected to a positive displacement
Fig. 1. Nitrogen solubility profile of chickpea flour.
2. Materials and methods
Pressure gage
Valve Overflow
2.1. Sample preparation Pressure gage
Feeding tank
Commercial dehulled split desi chickpea (dhal) from the crop harvested in fall of 2003 were obtained from Canadian Select Grains of Eston, Saskatchewan, Canada. The split chickpea grain was stored in a walk-in cooler maintained at a temperature of 2 ± 2 °C. It was then milled using a pin mill (GM 280/S-D, Condux werk, Hannau, Wolfgang, Germany). The pin mill had two discs: the first disc had 86 pins rotating at 8034 rpm; and the second disc had 108 pins and was stationary. A portion of the whole chickpea flour was defatted by ACS-grade isopropyl alcohol (EMD, Gibbstown, NJ) using Soxhlet extraction system. The remaining alcohol in the sample was dried
Valve Pressure gage
Hydrocyclone Valve
Pump
Pressure gage
Valve Underflow
Fig. 2. Schematic of the hydrocyclone system.
462
S. Emami et al. / Journal of Food Engineering 82 (2007) 460–465
pump (Model 4100C, Hypro Inc., St. Paul, MN) running at 1950 rpm. The Dorrclone unit consisted of four 10-mm hydrocyclones operating in parallel. Three of these hydrocyclones were removed and the vacant ports were plugged using rubber stoppers. Using only one hydrocyclone reduced the feed requirement and increased the operating pressure, making the unit suitable for lab-scale trials. A by-pass valve, located between the pump and feed tank, was employed to circulate the suspension, keep flour particles in suspension form, and help control pressure. A valve was located before the cyclone to control inlet pressure. The overflow and underflow valves were kept fully-opened during the test. Suspension was made using flour (whole chickpea flour or defatted chickpea flour) and deionized water at a concentration of 1.5% (w/w). The measured pH of the suspension was 6.6. To obtain a suspension pH of 9.0, a 10 M NaOH solution was used. The suspension, which was left overnight, was stirred for one hour using a mixer (OMINI-MIXER 17105, Sorvall Inc., Newtown, CT) before feeding it to the hydrocyclone. The suspension was fed to the hydrocyclone at two pH values of 6.6 and 9.0. The separation process in the hydrocyclone is shown schematically in Fig. 3. The suspension was passed through the hydrocyclone at an inlet pressure of 690 kPa resulting in first-pass overflow (O1) and underflow (U1). O1 and U1 were separately passed through the hydrocyclone to obtain second-pass overflow (O2) and underflow (U2) products. O2 was left overnight to obtain sediment (SE) and supernatant (SU). SU was poured to another container and the protein contents of both SE and SU were determined.
used for legumes by other researchers (Han & Khan, 1990; Reichert, 1982; Reichert & Youngs, 1978; Sa´nchezVioque et al., 1999; Sosulski, Hoover, Tyler, Murray, & Arntfield, 1985). Starch content was assessed using the method described by Holm, Bjorck, and Asp (1986). This measurement involves the gelatinization of starch in boiling water bath in the presence of a-amylase, followed by 30 min incubation of the subsample in the presence of amyloglucosidase, followed by addition of glucose oxidase/peroxidase reagent, and determination of glucose content using a spectrophotometer at a wavelength of 510 nm. The moisture content was measured using the method described by Egan, Ronald, and Ronald (1990) for determining the moisture content of syrup and condensed milk. For this purpose, the vacuum oven was set at 70 °C, and a vacuum gage pressure of 3.33 kPa was applied. All measurements were conducted in triplicates. The yield of fraction was measured based on recovery percentage. Starch and protein separation efficiencies were calculated according to Eq. (1) (Tyler, Youngs, & Sosulski, 1981) with some modification for dilute materials (Eq. (1)). S:E: ¼
TSP C P 100 TSF C F
ð1Þ
where S.E. is separation efficiency (%), TSP is total solid of product (kg), CP is component content in the product (% d.b.), TSF is total solid of feed material (kg), and CF is component content in the feed material (% d.b.). The values were calculated with respect to the starch and protein contents of the feed material into the hydrocyclone. 2.4. Statistical analysis
2.3. Analytical and physical measurements Protein content was measured by AACC method 46–30 (AACC, 2000) using LECO Model FP-528. A factor of 6.25 was used to convert nitrogen to protein content as
The Statistical Analysis System (SAS, 2003) was used to perform analysis of variance and mean values were compared using Duncan’s multiple range test and t-test, where appropriate. Overflow (O2) Supernatant (SU) Sediment (SE)
First-pass Overflow (O1) Cyclone
Inlet
Cyclone Underflow (UO1) Overflow (OU1)
First-pass Underflow (U1)
Supernatant (SU) Sediment (SE) Supernatant (SU) Sediment (SE)
Cyclone
Underflow (U2)
Supernatant (SU) Sediment (SE)
Fig. 3. Schematic of the extraction process in the hydrocyclone.
S. Emami et al. / Journal of Food Engineering 82 (2007) 460–465
3. Results and discussion Table 1 shows the fraction yield and total solids content of the overflows and the underflows. Because some material (suspension) remained in the hydrocyclone, the sum of the overflow and underflow in some runs were less than 100%. The total solids content of the underflow products was more than that of the overflow products showing that large and dense solid particles, i.e. starch granules, were collected in the underflow. This result is in agreement with that of Svarovsky (1990) who explained that coarse particles in the feed material settle on the hydrocyclone wall and collect in the underflow. 3.1. Starch separation
Table 2 Starch contenta (% d.b.) of the overflow and underflow using different feed materials and pH values Fraction Whole flourb Overflow
Underflow
Defatted flourc Overflow
Underflow
The starch content of the overflows and the underflows is presented in Table 2. The effect of the double-pass hydrocyclone process on starch content of the underflows (U1, U2, and UO1 (underflow of the second-pass of O1)) was significant for both defatted and whole chickpea flour at two pH levels. U2 had significantly higher (a = 0.05) starch content (1.2 to 1.4 times) than U1. This result clearly shows that the double-pass hydrocyclone process was better than the single-pass. The starch content of U2 from whole flour was 86.6% and 90.9% at pH 6.6 and 9.0, respectively. Statistical analysis (t-test) between U2 at two pH levels showed that starch content was significantly higher (a = 0.05) at pH 9.0 than at pH 6.6. Since at pH 9.0 (maximum NSI) the protein materials were better dispersed in the aqueous media than at pH 6.6, more of the protein went into the overflow; this phenomenon resulted in less protein contamination in the underflow. The small amount of starch in the overflows (O1, O2, and OU1 (overflow of
463
pH 6.6
pH 9.0
O1 O2 OU1
3.4 b 1.9 c 5.2 a
3.5 b 1.9 c 5.5 a
U1 U2 UO1
63.2 b 86.6 a 5.8 c
67.1 b 90.9 a 3.6 c
O1 O2 OU1
7.6 a 1.1 b 1.1 b
0.7 a 0.6 a 1.1 a
U1 U2 UO1
77.8 b 93.1 a 14.7 c
79.9 b 99.7 a 0.84 c
O1 = Overflow of the first-pass, O2 = Overflow of the second-pass of O1, U1 = Underflow of the first-pass, U2 = Underflow of the second-pass of U1, OU1 = Overflow of the second-pass of U1, UO1 = Underflow of the second-pass of O1. a–c values in the same column in each group of flour followed by common letter are not significantly different at 5% level. a Values are average of three determinations. b Whole flour starch content = 48.0% d.b. c Defatted flour starch content = 51.0% d.b.
the second-pass of U1)) is likely associated with damaged starch granules resulting from pin milling (Tyler et al., 1981). For the defatted flour, the starch content of U2 at pH 9.0 (99.7%) was significantly higher (by t-test at a = 0.05) than at pH 6.6 (93.1%). Among all applied conditions, the double-pass hydrocyclone process using defatted flour at pH 9.0 resulted in the highest starch content (99.7%
Table 1 Fraction yield and total solida of the overflow and underflow using different feed materials and pH values Fraction
Whole flour Overflow
Underflow
Defatted flour Overflow
Underflow
pH 6.6
pH 9.0
Fraction yield (%)
Total solids (% w.b.)
Fraction yieldb (%)
Total solids (% w.b.)
O1 O2 OU1
50.2 48.4 47.9
0.5 0.5 0.5
49.3 46.1 47.7
0.6 0.6 0.6
U1 U2 UO1
49.8 52.1 51.7
1.9 2.7 0.5
47.1 52.8 53.9
2.0 2.7 0.8
O1 O2 OU1
49.0 49.0 46.5
0.6 0.4 0.7
47.3 47.4 46.3
0.6 0.5 0.5
U1 U2 UO1
49.0 53.5 51.0
2.0 3.0 0.6
49.1 53.7 52.7
1.9 2.8 0.5
O1 = Overflow of the first-pass, O2 = Overflow of the second-pass of O1, U1 = Underflow of the first-pass, U2 = Underflow of the second-pass of U1, OU1 = Overflow of the second-pass of U1, UO1 = Underflow of the second-pass of O1. a Values are average of three determinations. b Because of unavoidable residual material in the system, the sum of the overflow and underflow yield is less than 100%.
464
S. Emami et al. / Journal of Food Engineering 82 (2007) 460–465
starch). The starch contents obtained in the current study were similar to those reported by Vose (1980) for field pea and horsebean employing a high pH (pH 8.5) and a series of three hydrocyclones, where the starch fraction obtained had 99% starch. Employing high pH (pH 10.2 for field pea and pH 9.6 for fababean) but using centrifugation, Sosulski and Sosulski (1986) obtained starch fraction with a starch content of as high as 94% from field pea and fababean, which was lower than value obtained in this study. Table 3 shows the starch separation efficiency values in the underflows. The starch separation efficiencies of U1 and U2 ranged between 93.8% and 99.8% and these values were similar to those reported by Tyler et al. (1981) for air classification of legumes. However, values from this study were higher than those reported by Sosulski and Sosulski (1986) who employed high pH and centrifugation in separating starch from field pea and fababean (76–79%). In this study, the lowest starch separation efficiency was obtained from UO1. 3.2. Protein separation
Defatted flour Underflow
Fraction Whole flourb Overflow
pH 6.6
Defatted flourc Overflow
Underflow
pH 6.6
pH 9.0
SE
SU
SE
SU
O1 O2 OU1
66.9 a 68.5 a 64.0 b
69.6 a 64.1 a 72.1 a
68.4 b 70.3 a 65.7 c
67.0 a 68.3 a 65.7 a
U1 U2 UO1
8.0 b 7.0 b 52.3 a
34.1 b 31.5 c 67.1 a
7.0 b 5.4 c 53.2 a
22.4 b 20.1 c 55.3 a
O1 O2 OU1
72.8 c 97.8 a 88.3 b
68.1 a 82.4 a 72.7 a
92.5 b 98.9 a 88.3 c
63.5 ab 66.1 a 59.1 b
U1 U2 UO1
18.5 b 12.0 c 58.1 a
67.1 a 65.6 a 69.6 a
16.6 b 11.1 c 90.1 a
61.1 a 60.4 a 60.5 a
SE = Sediment, SU = Supernatant, O1 = Overflow of the first-pass, O2 = Overflow of the second-pass of O1, U1 = Underflow of the firstpass, U2 = Underflow of the second-pass of U1, OU1 = Overflow of the second-pass of U1, UO1 = Underflow of the second-pass of O1. a–c values in the same column in each group of flour followed by common letter are not significantly different at 5% level. a Values are average of three determinations. b Whole flour protein content = 26.3% d.b. c Defatted flour protein content = 27.9% d.b.
Table 5 Protein separation efficiencya (%) achieved in the overflows
Table 3 Starch separation efficiencya (%) achieved in the underflows
Whole flour Underflow
Table 4 Protein contenta (% d.b.) of sediment and supernatant of the overflow and underflow using different feed materials and pH values
Underflow
Table 4 shows the protein content of the overflows and the underflows. For the whole flour and defatted flour, O1 had higher protein content than U1 at both pH levels. SEs of the overflow and the underflow had very high difference in protein contents at both pH levels. Although high pH improved the protein contents of the overflow SEs as well as SUs, statistically (t-test) it did not have any significant effect (a = 0.05). The protein contents of SEs from O2 were 2.5–2.7 times that of the whole flour, whereas they were 3.5 times that of the defatted flour. UO1 had still high protein content which may be due to the agglomeration of starch granules and proteinaceous material. The effect of defatting on the protein content of the overflow SEs was significant (a = 0.05) which is desirable since protein is concentrated in the SEs. However, it was not significant (a = 0.05) for the overflow SUs. Using the double-pass hydrocyclone
Feed material
process, the protein contents of the overflow products (O2 and OU1), where they are expected to accumulate, did not increase because the protein particles were still dispersed in the media. Adjusting the pH of the overflow to isoelectric point (pH 4.3) could have improved protein separation and increased protein content. Table 5 shows the protein separation efficiency values. For both whole flour and defatted flour, the high pH improved the protein separation efficiencies due to the higher solubility of the chickpea protein at pH 9.0 than pH 6.6. The values in this study were similar to those
pH 9.0
U1 U2 UO1
93.8 ± 2.4 99.4 ± 0.9 73.7 ± 1.3
97.2 ± 0.6 96.6 ± 1.2 72.2 ± 0.6
U1 U2 UO1
96.8 ± 0.4 99.8 ± 0.3 98.3 ± 0.3
96.7 ± 1.6 99.8 ± 0.2 61.9 ± 0.4
O1 = Overflow of the first-pass, O2 = Overflow of the second-pass of O1, U1 = Underflow of the first-pass, U2 = Underflow of the second-pass of U1, OU1 = Overflow of the second-pass of U1, UO1 = Underflow of the second-pass of O1. a Values are average of three determinations. Value following the mean is standard error.
Feed material Whole flour Overflow
Defatted flour Overflow
pH 6.6
pH 9.0
O1 O2 OU1
65.7 ± 1.2 48.5 ± 0.7 51.2 ± 0.6
70.1 ± 1.1 50.1 ± 0.9 55.7 ± 0.8
O1 O2 OU1
67.5 ± 0.8 74.3 ± 0.4 62.5 ± 0.3
83.1 ± 0.9 51.9 ± 0.8 67.4 ± 0.8
O1 = Overflow of the first-pass, O2 = Overflow of the second-pass of O1, U1 = Underflow of the first-pass, U2 = Underflow of the second-pass of U1, OU1 = Overflow of the second-pass of U1, UO1 = Underflow of the second-pass of O1. a Values are average of three determinations. Value following the mean is standard error.
S. Emami et al. / Journal of Food Engineering 82 (2007) 460–465
reported by Sosulski and Sosulski (1986) (72–74%) for field pea and fababean using high pH and centrifugation. 4. Conclusion The effect of the double-pass hydrocyclone process was significant on starch content of the underflow products (U1 and U2) at both pH 6.6 and 9.0. The use of pH 9.0 and defatting of chickpea flour significantly increased the starch content of U2. However, pH 9.0 did not significantly affect the protein content of the SEs and SUs of the overflows. Defatting of chickpea flour significantly increased the protein content of overflow SEs; whereas, it did not have any significant effect on the protein content of the overflow SUs. U2 (underflow product resulting from the second-pass process) had high starch content ranging from 86.6% to 99.7% with starch separation efficiencies ranging from 99.4% to 99.8%. Overall, using pH 9.0, defatting, and double-pass hydrocyclone process improved the starch–protein separation from chickpea flour. Acknowledgements The authors would like to acknowledge the following: (1) the technical assistance of Douglas Hassard and Connie Perron of the Crop Development Centre (College of Agriculture, University of Saskatchewan) and Ms. Hong Qi of the Centre of Agri-Industrial Technology (Alberta Agriculture and Rural Development), (2) the Natural Science Engineering Research Council (NSERC) and Agriculture Development Fund (ADF) for financial support, (3) Canada-Saskatchewan Agri-Food Innovation Fund for the new Bioprocess Engineering Research Lab, and (4) Pilot Plant Crop (POS) for loaning us the hydrocyclone unit. References AACC (2000). Approved Methods of the American Association of Cereal Chemists. Method 46-23 and Method 46-30 (10th ed.). St. Paul, MN: American Association of Cereal Chemists. Anderson, C. G., & Romo, G. R. (1976). Influence of extraction medium pH on the protein content of some legume starches. Journal of Food Technology, 11(6), 647–654. Bencini, M. C. (1986). Functional properties of drum-died chickpea (Cicer arietinum L.) flours. Journal of Food Science, 51(6), 1518–1526. Chavan, J. K., Kadam, S. S., & Salunkhe, D. K. (1986). Biotechnology and technology of chickpea (Cicer arietinum L.) seeds. CRC Critical Reviews in Food Science and Nutrition, 25(2), 107–158. Cheremisinoff, P. N. (1995). Processing Engineering Handbook SeriesSolid/liquid. Lancaster, PA: Technomic Publishing Co. Inc.
465
Colonna, P., Gueguen, J., & Mercier, C. (1981). Pilot scale preparation of starch and cell wall material from Pisum sativum and Vicia faba peas, broadbeans, flours, composition. Sciences des Aliments, 1(3), 415–426. Egan, H., Ronald, S. K., & Ronald, S. (1990). Pearson’s Chemical Analysis of Foods (8th ed.). London, UK: Longman Scientific and Technical. Han, J., & Khan, K. (1990). Physicochemical studies of pin-milled and airclassified dry edible bean fractions. Cereal Chemistry, 67(4), 384–390. Holm, J., Bjorck, A. D., & Asp, N. G. (1986). A rapid method for the analysis of starch. Starch, 38(7), 224–226. Hoover, R., & Sosulski, F. W. (1991). Composition, and chemical modification of legume starches: A review. Canadian Journal of Physiology and Pharmacology, 69(1), 79–92. International Starch Institute. (2003). Tapioca Starch Applications. Internet document,