Op

Ta

b

a

ARR2A

KENSTCF

1

tfiFtcffi1a

t(mmttts

fi

0h

Industrial Crops and Products 45 (2013) 100– 105

Contents lists available at SciVerse ScienceDirect

Industrial Crops and Products

journa l h o me pag e: www.elsev ier .com/ locate / indcrop

perating air velocities for fiber separation from corn flour using the Elusieverocess

ejas S. Pandyaa, Radhakrishnan Srinivasana,∗, Jason K. Johnsonb

Department of Agricultural and Biological Engineering, 130 Creelman, Mississippi State, MS 39762, United StatesUSDA, ARS, Poultry Research, 606 Spring Street, Starkville, MS 39759, United States

r t i c l e i n f o

rticle history:eceived 21 September 2012eceived in revised form3 November 2012ccepted 24 November 2012

a b s t r a c t

Fiber separation from corn flour could increase ethanol productivity and increase energy value as feedfor non-ruminants (swine and poultry). Elusieve process, a combination of sieving and air classification,has been found to be effective in separating fiber. The objectives of this study were to determine theoperating air velocities for corn particles and to compare physical properties of corn particles with thatof DDGS particles from an earlier study. The operating air velocities for large, medium and small corn

eywords:lusieveDFeparationerminal velocityorniber

size fractions were 2.9–3.8, 2.8–3.0 and 2.5–2.6 m/s, respectively. Densities of nonfiber particles for cornflour were higher than for DDGS (earlier study). Compared to DDGS, the difference between fiber andnonfiber particle terminal velocities was higher for corn, which signifies relative ease of operability forfiber separation from corn flour.

© 2012 Elsevier B.V. All rights reserved.

. Introduction

Corn is widely used as animal feed and for fuel ethanol produc-ion. The fiber present in corn does not convert to ethanol. Also, thisber is not digested well by non-ruminants (swine and poultry).iber separation from corn flour could increase ethanol produc-ivity and increase energy value as feed for non-ruminants. Fiberan also be used as combustion fuel, cattle feed, and as a feedstockor producing valuable products such as cellulosic ethanol, cornber gum, oligosaccharides, phytosterols, and polyols (Dien et al.,997; Crittenden and Playne, 1996; Moreau et al., 1996; Buhnernd Agblevor, 2004).

Elusieve process, a combination of sieving and air classifica-ion, was found to be effective in separating fiber from corn flourSrinivasan and Singh, 2008a; Pandya and Srinivasan, 2011). Ham-

er millled corn flour is sieved into different size fractions: large,edium, small, fines and the pan, pan being the smallest size frac-

ion. The large, medium, small and fines fractions are air classifiedo separate fiber. The process of air classification takes advantage ofhe difference in physical properties of fiber and nonfiber particles

uch as shape, weight and density of particles.

Air classification is carried out at air velocity higher thanber terminal velocity and lower than nonfiber terminal velocity.

∗ Corresponding author. Tel.: +1 662 325 8536; fax: +1 662 325 3853.E-mail address: rs634@msstate.edu (R. Srinivasan).

926-6690/$ – see front matter © 2012 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.indcrop.2012.11.045

Terminal velocity is the constant velocity attained by a falling par-ticle when the upward drag on the particle and the buoyancy forcebalance the downward force of gravity (DRI, 2005). The gravita-tional force on the particle is higher than the buoyancy force, hencethe downward acceleration of the particle. For a sphere shapedparticle, the upward drag force on the particle is proportional tothe square of the velocity of the particle in laminar flow. As thedownward falling particle is accelerating, the upward drag forceincreases. At some velocity, the gravitational force acting down-ward due to weight of the particle balances the upward drag forceand the buoyancy force. This velocity is the terminal velocity of theparticle, at which the acceleration of the particle becomes zero andthere is no further change in velocity. Air velocity higher than ter-minal velocity of fiber particles and lower than terminal velocity ofnonfiber particles results in effective fiber separation. If air velocityis higher than the terminal velocity of nonfiber particles, nonfiberparticles would get carried along with the fiber particles, resultingin poor quality separation.

Srinivasan and Singh (2008b) experimentally determined theterminal velocity of DDGS particles. No work has determined theterminal velocities for corn flour particles, till now. Presently, the airvelocities are adjusted by trial and error for separation of fiberfrom corn flour. Experimental determination of terminal velocities

for fiber and nonfiber particles will enable precise adjustment ofoperating air velocities in aspirators for effective fiber separation.In this study, terminal velocities of fiber and nonfiber particles incorn flour were obtained experimentally using a 152 mm (6-inch)

T.S. Pandya et al. / Industrial Crops and Products 45 (2013) 100– 105 101

Blower

Material Feed

Heavier Fraction

Light er Fracti onFlow Me asure ment

using hot-wire anemometer

MotorSurge Box

Collecti onBox

Ø 152 mm

1575 mm length

200 mm length

585 mm length

Air Flow Control

ehtmetp

2

2

cMapbaet((tt80Iptwtcfic

2011). Studies were conducted using three replicates by dividingeach sample into three batches of 25 kg each (Fig. 3). Each batch ofcorn flour was sieved into five size categories using SWECO sifter

Fig. 1. Schematic of elutriation column set up.

lutriation column. Measurement of particle size and densitieselps us understand the relative differences in physical propertieshat govern separation. The objectives of this study were to deter-

ine air velocities that can be used to operate pilot and commerciallusieve set-ups for effective fiber separation from corn flour ando compare size, weight and densities of corn fiber and nonfiberarticles with that of DDGS particles from previous study.

. Materials and methods

.1. Elutriation column

To measure the terminal velocities of fiber and nonfiber parti-les, an elutriation column was constructed at the Pace Seed lab,ississippi State University, similar to the one used by Srinivasan

nd Singh (2008b) (Figs. 1 and 2). It mainly consists of a 6-inchipe, a blower and motor assembly to generate air flow, a surgeox for controlling the air flow and a fiber collection box. To attain

fully developed flow in the pipe, the distance of the top of thelutriation column from the material inlet (1575 mm) was main-ained more than six times the column internal diameter (155 mm)ASHRAE Standard 41.8-1989). A 2440 mm (8 feet) long, 152 mm6 inch) diameter clear rigid Schedule 40 PVC pipe was used. Theransparent pipe provided good visibility of the fiber separationaking place in the elutriation column. The blower (Dayton blower

15/16′ ′, model 2C820, Lake Forest, IL, USA) was powered by a.4 kW (0.5 hp) motor (Dayton motor model# 6K482, Lake Forest,L, USA). The surge box was built out of plywood sheet and wasrovided with a vent and a sliding cover to control the air flowhrough the column. An inlet was provided in the column through

hich the sieved corn flour was fed at a rate of 50 g/min. To facili-

ate the collection and to reduce loss of lighter fraction material, aollection box made of transparent acrylic sheet was used. To arrestne particles that might escape with the exiting air, the top of theollection box was fitted with a very fine mesh screen.

Fig. 2. Photograph of elutriation column set-up for measuring particle terminalvelocities.

2.2. Elusieving for separation of fiber and nonfiber

Yellow dent corn (75 kg) was procured from the local feed store(Oktibbeha County Cooperative, Starkville, MS). The initial mois-ture content of the corn was 11.5%. The corn kernels were milledusing a hammer mill (Bliss Industries, Ponca, OK). The 3.2 mm(8/64-inch) retainer screen opening was chosen in the hammermill as it results in best fiber separation (Pandya and Srinivasan,

Fig. 3. Schematic of sieving and sequential air classification of large size fraction(based on Srinivasan and Singh, 2008b). The medium, small and fines size fractionswere also air classified in similar way. Pan size fraction was not subjected to airclassification.

1 ops an

(S3to

bttataa

2

SwwToelfttrtiptdiltmeianS

tfEcttn

vfvoingto

fhoes

02 T.S. Pandya et al. / Industrial Cr

model ZS30-S6666, SWECO Vibro-Energy Separator, Florence, KY).ieving with mesh 12 (1532 �m), 16 (1130 �m), 24 (704 �m) and5 (447 �m) resulted in large, medium, small, fines and pan frac-ions with wt% of 29.7, 16.9, 17.8, 24.1 and 11.5%, and NDF contentf 9.3, 11.0, 10.0, 6.3 and 3.3%, respectively (Fig. 3).

The large, medium, small and fines fractions were elutriated oney one. Each size fraction was divided into three batches for statis-ical purpose. The sliding gate on the surge box was used to controlhe air velocity inside the column. An initial velocity was set forbout 5% lighter fraction yield, that is, 5% of the sample materialo be collected at the top of the column as lighter fraction. Their velocity was measured using a calibrated hand-held hot wirenemometer (Extech Instruments, Model 407123, Waltham, MA).

.3. Experimental determination of terminal velocities

The experimental procedure used was similar to Srinivasan andingh (2008b). The lighter fraction is the material that gets carriedith the airflow and comes out at the top of the elutriation column,hile the heavier fraction drops down at the bottom of the column.

o obtain best quality fiber, that is the fiber with negligible amountf nonfiber, material for each of the four largest size categories waslutriated at about 5% lighter fraction yield at air velocity V1. A lowighter fraction yield at V1 ensures high fiber purity in the lighterraction and hence, is taken as representative sample for fiber forhe particular size fraction and used to determine physical proper-ies of fiber. “For each air velocity, the lighter fraction material wasetained for sampling, while the heavier fraction remaining afterhe first pass at velocity V1 was elutriated at an incremental veloc-ty of about 10% to V1 (Fig. 3). The heavier fraction remaining fromrevious air velocity was used, instead of using whole sample, athe next incremental velocity in order to increase the sensitivity ofetecting the onset of carryover of nonfiber. The incremental veloc-

ty steps were about 10% or higher, the minimum increment beingarge enough to allow collection of lighter fraction material in quan-ity that is sufficient for analysis. To ensure that most of the fiber

aterial was removed, 20–30% of lighter fraction was separated forach size fraction. The heavier fraction material left after elutriat-ng for four to six incremental velocities was the one with negligiblemounts of fiber, and was used to determine physical properties ofonfiber material for the respective size fraction.” (Srinivasan andingh, 2008b).

Nonfiber material tends to get carried with fiber during elu-riation. The amount of nonfiber fraction being carried in lighterraction is dependent on terminal velocity of nonfiber particles.ach size fraction consists of a range of fiber and nonfiber parti-le sizes with different terminal velocities. The velocity at whichhe NDF content in the lighter fraction becomes less than 50%, washe criterion set to determine the velocity at which the onset ofonfiber particles takes place for a size fraction.

For each size fraction, elutriation was done at four different airelocities for large and fine fractions, and five different air velocitiesor medium and small fractions, resulting in lighter fraction for eachelocity. The number of air velocity steps was based on the amountf lighter fraction material that was being air lifted. While conduct-ng the experiment, if it was found that the amount of material wasot enough for compositional analysis, the air flow velocity wasradually increased to result in higher quantity of material separa-ion. The highest air velocity for a size fraction was the one at whichnly a negligible amount of material was air lifted.

For each batch of a given size fraction of material, one lighterraction sample was collected per velocity, while one sample of

eavier fraction was collected per size. For example, for a batchf large fraction flour, four lighter fraction samples were collected,ach for velocities V1, V2, V3 and V4, while one heavier fractionample was collected from the material retained at the end. Thus,

d Products 45 (2013) 100– 105

there were a total of five elutriation samples per batch for largefraction. Large and fines size fractions were elutriated at four dif-ferent air velocities, while medium and small size fractions wereelutriated at five different air velocities. Thus, 22 samples were col-lected from elutriation of large, medium, small and fines fractionsfor every batch. The samples of original corn flour and pan size werealso analyzed, making a total of 24 samples per batch of corn flour.A total of 72 samples were available from three batches. The mois-ture content of the lighter fractions was 11.4–12.5%, while that ofheavier fraction was 11.6–12.8%.

2.4. Determination of particle densities

Corn fiber and nonfiber particle densities were measured using ahelium pycnometer at Quantachrome Inc.’s commercial laboratory(Boynton Beach, FL). Three replicates of each sample were analyzed.Lighter fraction at velocity V1 was used as a representative samplefor fiber particles for each size fraction. The heavier fraction mate-rial left after elutriating for four to six incremental velocities wasnonfiber material and was used to measure physical properties ofnonfiber.

2.5. Determination of equivalent spherical diameters

Equivalent spherical diameters (dsph) need to be determined forestimating terminal velocities. The dsph of a particle is the diameterof sphere of equivalent volume. Accurately measured 3.000 g sam-ples of fiber and nonfiber particles were taken. Mean particle mass(mp) was calculated by dividing the weight of the sample by num-ber of particles. The number of particles was counted manually byspreading on a sheet of paper. The dsph was calculated as:

dsph =(

6�

× mp

�p

)1/3

where dsph = equivalent spherical diameter, mmp= mean particle mass, kg�p = density of particle, kg/m3

Each size fraction of corn flour has a range of size of fiber andnonfiber particles. The heavier fraction material is the one retainedafter elutriating at four to six incremental velocities. Since the par-ticles from heavier fraction were used for measuring the dsph ofnonfiber particles, they represent the largest size of nonfiber parti-cles. In a given size fraction, the dsph of smallest sized nonfiber wastaken as proportional to the ratio of sieve size range. For mediumfraction nonfiber, the range of sieve is from 1532 to 1130 �m, thusthe dsph of smallest nonfiber medium fraction particle is propor-tional to 0.74 (1130/1532). The dsph of largest nonfiber for mediumfraction was 1689 �m, thus dsph of smallest nonfiber particle ofmedium fraction is 0.74 times 1235 �m, that is 914 �m. Simi-larly, for small fraction nonfiber, the range of sieve is from 1130to 704 �m, and dsph of largest nonfiber particle of small fractionis 832 �m, thus dsph of smallest nonfiber of small fraction is 0.62times 832 �m i.e. 516 �m.

2.6. Theoretical estimation of terminal velocities

The terminal velocities were theoretically determined using cor-relations developed by Becker (1959) and were compared withexperimentally determined terminal velocities

Rep = dsphut�f

�

[24Rep + CI(Rep)2]�2

(8�f)= (dsph)3g(�s − �f)

6

ops and Products 45 (2013) 100– 105 103

w

scswmSaSwnt(v

2

stpmfiS2t2tmr

3

3

phDdsdp(tssctwdhe

D

Fig. 4. %NDF in lighter fraction v/s air velocity, m/s. The ‘dashed’ line indicates that

carried with fiber particles in lighter fraction material. The veloc-ity at which the NDF content in the lighter fraction drops to50% of total lighter fraction weight was considered as minimum

T.S. Pandya et al. / Industrial Cr

hereCI = 5.31ϕ − 4.88(ϕ)2 for ϕ > 0.67 (nonfiber)CI = 2.25(5.5)0.34

√ϕ/2 × (Rep) − 0.34

√ϕ/2 for ϕ < 0.67 (fiber)

ϕ = particle sphericityut = terminal velocity of particle in fluid, m/s�f = density of fluid, kg/m3

�s = density of particle, kg/m3

g = acceleration due to gravity, = 9.81 m/s2

� = viscosity of fluid, kg/(m.s)dsph = equivalent spherical diameter, mParticle sphericity (ϕ) is the ratio of surface area of sphere having

ame volume as the particle to the actual surface area of the parti-le. Fiber and nonfiber particles of corn were found to be similar inhape as DDGS particles. For this study, sphericity of corn particlesere assumed to be equivalent to mean ϕ values of equivalent large,edium and small DDGS fiber and nonfiber particles reported by

rinivasan and Singh (2008b). Sphericity values for large, mediumnd small fiber particles were 0.23, 0.20 and 0.46, respectively.phericity values for large, medium and small nonfiber particlesere 0.80, 0.80 and 0.81, respectively. The sphericity value of largeonfiber particle for corn was taken same as medium nonfiber par-icle (0.80), since it was not determined by Srinivasan and Singh2008b). The ϕ values were used to estimate theoretical terminalelocities using correlation developed by Becker (1959).

.7. Sample analyses

Compositions of fractions were obtained by collecting threeamples from each of the classifications. The samples were groundo a fine powder using a coffee grinder prior to analysis to avoidarticle segregation. Analyses of samples were carried out at a com-ercial laboratory (Midwest Labs, Omaha, NE). Neutral detergent

ber (NDF) content was determined using the procedure of Vanoest et al. (1991). Samples were analyzed for total nitrogen (AOAC,003, Method 990.03). Crude protein content was calculated asotal N × 6.25. Samples were also analyzed for crude fat (AOAC,003, Method 920.39) and ash (AOAC, 2003, Method 942.05). Mois-ure content was determined using the two-stage convection oven

ethod (AACC International, 2000, Method 44 18). Compositionalesults for fractions are reported and discussed in dry basis.

. Results and discussion

.1. Physical properties of fiber and nonfiber corn particles

The density of fiber particles was higher than that of nonfiberarticles (Table 2). Srinivasan and Singh (2008b) also reportedigher density of fiber particles relative to nonfiber particles forDGS (Table 3). The dsph for fiber and nonfiber corn particles wereetermined for estimating terminal velocities (Table 2). Within theame size fraction, the dsph of nonfiber particles were higher thansph of fiber particles. Similar observation of higher dsph of nonfiberarticles compared to fiber particles has been reported for DDGSSrinivasan and Singh (2008b); Table 3). Higher dsph of nonfiber par-icles compared to fiber particles is attributed to the near-sphericalhape of nonfiber compared to the flat shape of fiber particles. Thehapes of particles were visually observed. The shapes are not per-eivable from photographic or microscopic images because of thehree-dimensional nature of the described shapes. Fiber particlesere preferentially carried by air, despite having higher particleensity than nonfiber particles, because flat-shaped fiber particles

ad lower dsph than near-spherical shaped nonfiber particles withinach size fraction.

Densities of fiber particles were similar for DDGS and corn flour.ensities of nonfiber particles for corn flour (1408–1419 kg/m3)

the NDF content in lighter fraction has decreased to 50%, wherein onset of non-fiber phase takes place. The air velocity at this point is considered as the minimumterminal velocity of nonfiber particles.

were higher than for DDGS (1312–1333 kg/m3) perhaps because ofhigher starch content in nonfiber particles from corn flour. DDGScontains negligible starch content because the starch in the corngets converted into ethanol in the production plant. The differencebetween densities of fiber and nonfiber particles was higher forDDGS than for corn flour.

3.2. Experimental terminal velocities of fiber

Minimum terminal velocities for large, medium and small cornsize fraction fiber particles was found to be 2.9, 2.8, 2.5 and 1.6 m/s,respectively, which resulted in lighter fraction yields of 5.0, 6,0,5.7 and 3.7%, respectively (Table 1). The NDF in lighter fraction atminimum terminal velocities for large, medium, small and finessize fractions was found to be 73.0, 65.7, 58.5 and 5.4% respec-tively. Air classification was ineffective for the fines size fraction asobserved from the low NDF (5.4%) of its lighter fraction. The trendof decreasing NDF with decreasing fraction size was observed. Adecreasing trend in minimum terminal velocities of fiber has beenreported for DDGS also (Srinivasan and Singh, 2008b).

3.3. Experimental terminal velocities of nonfiber and operatingair velocities

Figs. 4 and 5 show variation in NDF and starch content inlighter fraction with respect to air velocity. Nonfiber particles get

Fig. 5. %Starch in lighter fraction v/s air velocity m/s.

104 T.S. Pandya et al. / Industrial Crops and Products 45 (2013) 100– 105

Table 1%yield, NDF and starch contents for corn fractions. Rows for heavier fractions are italicized.

Material fraction Air velocity Yield% Cumulative yield% NDF % % Starch

Original 8.29 63.33Large 9.32 61.96LF at V1 = 2.9 2.9 5.03 5.03 72.97 8.90LF at V2 = 3.2 3.2 2.00 7.03 60.67 15.94LF at V3 = 4.8 4.8 2.87 9.90 32.27 34.07LF at V4 = 6 6.0 11.97 21.87 11.55 45.86HF-Large 5.78 67.52Medium 11.04 57.65LF at V1 = 2.8 2.8 6.00 6.00 65.73 17.23LF at V2 = 3.1 3.1 2.77 8.77 32.27 37.98LF at V3 = 3.8 3.8 3.93 12.70 16.50 51.10LF at V4 = 4.5 4.5 2.27 14.97 13.47 51.90LF at V5 = 5.2 5.2 8.33 23.30 7.30 58.54HF-Medium 2.62 69.85Small 9.97 62.27LF at V1 = 2.5 2.5 5.68 5.68 58.47 20.65LF at V2 = 2.7 2.7 5.03 10.71 41.80 29.03LF at V3 = 3.0 3.0 3.53 14.24 28.00 40.15LF at V4 = 3.6 3.6 7.25 21.49 12.77 53.55LF at V5 = 3.8 3.8 7.90 29.39 7.21 60.86HF-Small 5.08 65.82Fines 6.27 67.51LF at V1 = 1.6 1.6 3.70 3.70 5.41 70.37LF at V2 = 1.8 1.8 4.60 8.30 6.88 69.30LF at V3 = 2.2 2.2 9.89 18.19 7.49 67.38LF at V4 = 2.5 2.5 8.16 26.35 8.16 65.81HF-Fines 5.69 67.84

L

tttt(Sfc

atnthwSi

3

lcmoTse

3

sSfra

With a decrease in particle size of size fractions, the differencebetween fiber and nonfiber particle terminal velocities decreased,signifying a decrease in operating velocity range. This narrow-ing of operating velocity range can be attributed to the change in

Pan

F – lighter fraction; HF – heavier fraction; NDF – neutral detergent fiber.

erminal velocity of nonfiber particles of the respective size frac-ion. For large size fraction, linear interpolation method showedhat NDF in lighter fraction dropped to 50% at air velocity of 3.8 m/s;hus, this velocity is taken as nonfiber minimum terminal velocityFig. 4). The corresponding starch content was found to be 23%.imilarly, minimum terminal velocities for medium and small sizeraction, were 2.9 m/s and 2.6 m/sec respectively. Starch contentarried over (at the onset of nonfiber) ranged from 23 to 28%.

Different minimum terminal velocities for fiber and nonfiberst different particle sizes justifies size fractionation before elutria-ion. If whole corn flour was elutriated for fiber removal, air velocityeeded will be atleast 2.9 m/s, which is minimum terminal veloci-ies of large fiber fraction. At 2.9 m/s, the small sized nonfiber thatave minimum terminal velocity of 2.6 m/s, will get carried alongith fiber, thus decreasing the effectiveness of fiber separation.

ingh et al. (2002) carried out elutriation of whole DDGS and foundt to be ineffective.

.4. Estimated terminal velocities of fiber and nonfiber

The minimum terminal velocities for corn fiber particles forarge, medium and small size fractions were also estimated usingorrelations developed by Becker (1959) (Tables 2 and 3), Esti-ated terminal velocities for corn fiber particles had a variation

f 5.3–8.6% from the experimentally determined values (Table 2).he estimated minimum terminal velocities for large, medium andmall nonfiber corn particles had a variation of 24.2–36.3% from thexperimentally determined values (Table 2).

.5. Operating air velocities

The operating air velocities for large, medium and small cornize fractions were 2.9–3.8, 2.8–3.0 and 2.5–2.6 m/s, respectively.

rinivasan and Singh (2008b) reported the operating air velocitiesor DDGS fractions to be 1.77–2.01, 1.46–2.01 and 1.27–1.63 m/s,espectively (Fig. 6). The operating air velocities for corn fiber sep-ration were found to be higher than for DDGS fiber separation.

3.26 72.91

Compared to DDGS, the difference between fiber and nonfiberparticle terminal velocities was higher for corn. For example, thedifference between fiber and nonfiber particle terminal velocitiesfor large sized corn fraction was 0.9 m/s, while the maximum dif-ference for DDGS was 0.55 m/s. This signifies better separation andease of operability for corn compared to DDGS. Higher differencebetween fiber and nonfiber particle terminal velocities for corncan be attributed to higher density of nonfiber particles in corn(1408–1419 kg/m3) compared to DDGS (1312–1333 kg/m3). In airclassification, nonfiber particles do not get carried by air, despitetheir lower density, because of their higher particle size (dsph) thatcauses them to have higher weight compared to fiber particles.Thus, higher particle density of nonfiber particles (as observed forcorn) is beneficial because it aids in preferential carry over of fiberparticles.

Fig. 6. Operating velocity regions for effective fiber separation from corn flour andDDGS (Values of DDGS reported by Srinivasan and Singh, 2008b).

T.S. Pandya et al. / Industrial Crops and Products 45 (2013) 100– 105 105

Table 2Equivalent spherical diameter (dsph), density and minimum terminal velocities for corn.

Sizecategory

Opening(�m)

Equivalent spherical diameter(dsph), �m

Particle density,kg/m3

Corn minimum terminal velocity, m/s Operating airvelocities

Fiber Largestnonfiber

Smallestnonfiber

Fiber Nonfiber Fiberestimated

Fiberexperimentallydetermined

Nonfiberestimated

Nonfiberexperimentallydetermined

Large 1532 968 (50) 1689 (35) NA 1455 (6) 1408 (2) 3.15 2.9 4.81 3.8 2.9–3.8Medium 1130 925 (53) 1235 (6) 914 1444 (2) 1419 (1) 3.04 2.8 4.07–4.09 3 2.8–3.0Small 704 749 (14) 832 (1) 516 1458 (1) 1419 (2) 2.63 2.5 3.08–3.23 2.6 2.5–2.6Fines 447 NA NA NA NA NA NA 1.6 NA NA NA

Values in parenthesis represent the standard deviation. NA – not applicable.

Table 3Equivalent spherical diameter (dsph), density and minimum terminal velocities for DDGS (Values reported by Srinivasan and Singh, 2008b).

Sizecategory

Opening(�m)

Equivalentspherical diameter(dsph), �m

Particle density,kg/m3

Minimum terminal velocity, m/s Operating airvelocities

Fiber Nonfiber Fiber Nonfiber Fiber(experimentallydetermined)

Nonfiber(experimentallydetermined)

Fiber(estimated)

Nonfiber(estimated)

Large 869 595 1061 1438 1318 1.77 1.77–2.01 2.19 NA 1.77–2.01Medium 582 400 716 1446 1312 1.46 1.83–2.01 1.63 2.08–2.23 1.46–2.01Small 389 337 497 1405 1333 1.27 1.27–1.63 1.45 1.56–1.87 1.27–1.63

N

pspnw

4

wpsleflhciaficfmcosc

A

E

Fines 295 NA NA NA NA NA

A – not applicable.

article shape from flat disc-shaped to spherical from large to finesize fraction. As the size fraction becomes smaller, even flat shapedarticles are similar to near spherical particles. A similar trend ofarrowing of terminal velocity range with decrease in particle sizeas observed for DDGS (Srinivasan and Singh, 2008b).

. Conclusions

Particle density of corn fiber particles was 1444–1458 kg/m3,hile that of nonfiber was 1408–1419 kg/m3. Fiber particles werereferentially carried by air, despite having higher particle den-ity than nonfiber particles, because flat-shaped fiber particles hadower dsph than near-spherical shaped nonfiber particles withinach size fraction. Densities of fiber particles were similar for cornour and DDGS. Densities of nonfiber particles for corn flour wereigher than for DDGS. Higher particle density of nonfiber parti-les in corn, compared to DDGS, was beneficial because it helpedn preferential carry over of fiber particles by air. This resulted in

higher difference between terminal velocities of fiber and non-ber particles in corn, indicating relative ease of operability fororn compared to DDGS. The operating air velocities determinedor corn can be used for scaled up system or for industrial imple-

entation of Elusieve process for fiber separation. Factors such asorn variety, moisture content and particle size could have effect onperating air velocities. Thus, it is recommended that a procedureimilar to described above be carried out for determining accuratease-specific operating air velocities.

cknowledgments

Thanks to Dr. Scott Branton, Dr. Joseph Purswell and Williamlliot of USDA ARS for their technical inputs in construction

NA NA NA NA

of elutriation column. Thanks to Courtney Paige Thompson andWilliam Fuller for technical assistance. Thanks to SustainableEnergy Research Centre (SERC) for partial funding towards thiswork.

References

AACC International, 2000. Approved Methods of the AACC, 10th ed. The AmericanAssociation of Cereal Chemists International, St Paul, MN.

AOAC, 2003. Official Methods of the AOAC, 17th ed. The Association of Official Ana-lytical Chemists, Gaithersburg, MD.

ASHRAE, 1989. Standard 41. 8-1989 – Standard Methods of Measurement of Flow ofLiquids in Pipes Using Orifice Flowmeters. American Society of Heating Refrig-erating and Air-conditioning Engineers.

Becker, H.A., 1959. The effects of shape and Reynolds number on drag in the motionof a freely oriented body in an infinite fluid. Can. J. Chem. Eng. 37, 85–91.

Buhner, J., Agblevor, F.A., 2004. Effect of detoxification of dilute acid corn fiberhydrolysate on xylitol production. Appl. Biochem. Biotechnol. 119, 13–30.

Crittenden, R.G., Playne, M.J., 1996. Production, properties and applications of foodgrade oligosaccharides. Trend Food Sci. Technol. 7, 353–361.

Dien, B.S., Hespell, R.B., Ingram, L.O., Bothast, R.J., 1997. Conversion of corn millingfibrous coproducts into ethanol by recombinant Escherichia coli strains K011 andSL40. World J. Microbiol. Biotechnol. 13, 619–625.

DRI, 2005. Climate and weather terms glossary. Desert Research Institute, Reno, NV.October 2005. www.wrcc.dri.edu/ams/glossary.html

Moreau, R.A., Powell, M.J., Hicks, K.B., 1996. Extraction and quantitative analysis ofoil from commercial corn fiber. J. Agric. Food Chem. 44, 2149–2154.

Pandya, T.S., Srinivasan, R., 2011. Effect of hammermill retention screensize on fiberseparation from corn flour using the Elusieve process. Ind. Crops Prod. 35, 37–43.

Srinivasan, R., Singh, V., 2008a. Pericarp fiber separation from corn flour using sievingand air classification. Cereal Chem. 85, 27–30.

Srinivasan, R., Singh, V., 2008b. Physical properties that govern fiber separation fromdistillers dried grains with solubles (DDGS) using sieving and air classification.Sep. Purif. Technol. 61, 461–468.

Van Soest, P.J., Robertson, J.B., Lewis, B.A., 1991. Methods for dietary fiber, neutral

detergent fiber and non-starch polysaccharides in relation to animal nutrition.J. Dairy Sci. 74, 3583–3597.

Singh, V., Moreau, R.A., Hicks, K.B., Belyea, R.L., Staff, C.H., 2002. Removal of fiberfrom distillers dried grains with solubles (DDGS) to increase value. Trans. ASAE45, 389–392.