Control of Powder Properties for Food Materials by Fine Grinding with Various Pulverizers

BYDr. Satoshi Akiyama

ManagerNisshin Seifun Group Inc.

Dr. Carl K. IshitoGeneral Manager

Nisshin Engineering Inc.

* Powder & Bulk Solids 2006 (29th) Technical Session S21

* Presented date: May 11, 2006 (Thu) 9:00-10:15AM

* Place: Donald E. Stephens Convention Center, IL USA

Content

• Background and objectives

• A new approach to optimize blade geometry of a mechanical mill using a computer simulation

• Effects of types of mill on powder properties

• How to control powder properties including pasting property, color and flavor by fine grinding

BackgroundStress mechanisms in particle size reduction

a) Pressure between two surfaces (compressive)b) Cutting between two surfacesc) Shear stressd) Impact with a solid surfacee) Impact with another particle f) Attrition (friction)

(a) Pressure (b) Cutting & shear (c) Impact (d) Attrition

0.1 1 10 100 1000

Shear(attrition)

Freq

uenc

yPressure(compressive)

Impact

Diameter　[µm]

Typical particle distributions of powder ground by three kind of forces

BackgroundStress mechanisms in particle size reduction

a) Pressure between two surfacesb) Cutting between two surfacesc) Shear stressd) Impact with a solid surfacee) Impact with another particle f) Attrition (friction)

Product quality for food materials:

dissolution, absorption, stability etc.

Changes in powder properties:particle size (surface area), crystal structure etc.

Objectives

1 Development of a new approach to optimize blade geometry of a mechanical impact mill for control of powder properties.

2 Control of powder properties for food materials by fine grinding

A new approach to optimize blade geometry of a mechanical impact mill

Numerical simulation: Particle trajectories in a mechanical impact mill

Impact energyExperiment: Calcium carbonate particles ground in a mechanical mill with various rotor blades

Particle size distribution, crystallinity changes

Grinding mechanisms of a mechanical impact mill

Blade

Rot

atio

nal

dire

ctio

nStator

Grinding zone

a) collision against walls (blades and stator)

b) collision with another particle

c) attrition on walls

d) cutting by blades

e) shear in a narrow gap between blade tip and stator

Mechanical mill

Screw feederBlower

Bag filter

Tachometer

Manometer

Orifice

Experimental apparatus

D1=150mm, D2=136mmW1=20mm, T=6mmc=2mm, H1=40mm

Table Dimensions

θΒ>0

Air flow

ω

Blade

Definition of blade angle

θΒ>0

Air flow

ω

BladeθΒ>0

Air flow

ω

Blade

Definition of blade angle

Blade

Inlet

Stator(cylinder)

Outlet(to Bag filter)

D1D2

H1T

c

RotorW1

zr

Blade

Inlet

Stator(cylinder)

Outlet(to Bag filter)

D1D2

H1T

c

RotorW1

zr

zr

Cross-sectional diagram of a mechanical impact mill

Flow fieldRealizable k-ε turbulent model Conditions:

Rotation speed ω: 14,000 min-1

Inlet velocity uz: 12 m/s

Particles TrajectoriesLagrangian dispersed phase modelConditions and assumptions:

Particle diameter Dp: 100 µmParticle density ρp: 1,400 kg/m3

Coefficient of restitution: 1 (perfectly elastic)Particle-particle interaction : negligibleEffect of particles on air properties: negligibleChange in particle size: negligibleNumber of particles: 100 (mono-sized )Injected velocity v: 0 m/s

Numerical simulations using Fluent

Calculation region

Blade (uθ= r ω)

Stator wall(uz=0,ur=0,uθ=0)

Periodic

Rear side

Front side

Blade

uInletuz=12m/s

Outlet

Rotor disc

zθ

0 0 0 =∂∂=∂∂=∂∂ θ zu,zu,zu zr

Rotor disc

r

Inlet

Pressure

-0.5

×10 4 Pa

0.0

0.5

1.0

z

r θ

(a) Velocity vectors andpressure contours

(b) Typical particle trajectories (Dp=100µm)

Impact position with stator wall

ω =14,000 min-1

Calculated results for blade angle θB=0°

vvz

Rebounding velocity in z-direction

Impact positions with stator wall

(a) θB=0o (b) θB=15o (c) θB=30o (d) θB=45o

Typical trajectories of a particle with a diameter of 100µm for various blade angles

160

120

80

40

0

Num

ber o

f im

pact

s [-]

50403020100Blade angle θB [deg.]

with blade with stator total

100

80

60

40

20

0

Impa

ct v

eloc

ity [m

/s]

50403020100Blade angle θB [deg.]

vtvn

θ

ω

u

θ

ω

u

B

Number of impacts and impact velocityas a function of blade angle

The definition of impact energies per particle mass

101

102

103

Impa

ct E

nerg

y [k

J/kg

]

50403020100

Blade angle θB [deg.]

Dp=100µm En Et

∑∑= =

⎟⎠⎞

⎜⎝⎛=

t iN

i

n

jjin

tn v

NE

1 1

2

,211

∑∑= =

⎟⎠⎞

⎜⎝⎛=

t iN

i

n

jjit

tt v

NE

1 1

2

,211

whereNt=total number of injected particles[-]ni=total number of impacts for an ith particle　[-]vn=normal component of impact velocity　[m/s]vt=tangential component of impact velocity　[m/s]

(3)

(2)

vn

vt

vα

Normal impact energy En and tangential impact energy Et by the computer simulation

Inte

nsity

[a.u

.]

504030202θ [deg.]

Feed

0deg.

15deg.

30deg.

45deg.

(104)

I0

I

Calcium carbonate

0 RatioIntensity II=

Estimation of change in crystallinity:

X-ray diffraction (XRD) patterns for calcium carbonate particles ground in the impact mill with various blade angles

15

10

5

0Med

ian

diam

eter

Dp5

0 [µm

]

50403020100Blade angle θB [deg.]

0.4

0.3

0.2

0.1

0.0Inte

nsity

ratio

I/I 0 [

-]

50403020100Blade angle θB [deg.]

(a) Particle size (b) Crystallinity

Changes in particle size and crystallinity of calcium carbonate particles ground in the impact mill with various blade angles

Another application of the optimizing approach-effects of blade angle on properties of potato starch particles--

Material: Potato starchMill: the mechanical impact mill

with different blade angles Measurement

Particle sizeCrystallinity

X-ray diffraction (XRD)Pasting properties:

Rapid Visco Analyser (RVA)

RVA (Rapid Visco Analyser, Newport Scientific Pty. Ltd.)

60

40

201600

100

80

Time　[s]

10000

8000

6000

4000

2000

0

Vis

cosi

ty　

[cP]

Temperature

Peak viscosity

12008004000

Tem

pera

ture

[o C]Pasting Temp.

waterSwollen Broken

Pasting state in the RVA

Typical RVA pasting curve

Heating/cooling

Viscosity measurement

Paddle

Canister

Parameters for pasting propertiesPasting temperature:

Temperature where viscosity first increases

Peak viscosity:Maximum viscosity

25µm

Potato starch particles ground in the impact millwith various blade angles and a ball mill

(iii) θB=0°

(v) θB=30°

(iv) θB=15°

(vi) θB=45°

(a) Particle size distributions (b) SEM Photos

(i) Feed (ii) Ball mill 25µm

25µm

25µm

100

80

60

40

20

0

Cum

ulat

ive

unde

rsiz

e [%

]

10 100Particle diameter Dp [µm]

D50

Feed (37.0µm) 0º (31.5µm) 15º (27.6µm) 30º (25.5µm) 45º (32.0µm) Ball mill (59.3µm)

25µm

25µm

X-ray diffraction (XRD) patterns and the definition of crystalline index for potato starch particles

Inte

nsity

[a.u

.]

403020102θ [deg.]

IC

Ia

Ball-milled

Feed material

Inte

nsity

[a.u

.]

40302010

2θ [deg.]

Feed

0º

15º

30º

θB=45º

Ball-milled

(a) X-ray diffraction patterns

θθ

22

⋅∫ −⋅∫ −

=dIIdII

Cac

as

Crystalline index (Wakelin et al.)

where Ic: intensity of feed particles (crystalline)Ia: intensity of ball-milled particles (amorphous)Is: intensity of pulverized particles

(1)

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Cry

stal

line

inde

x [-

]

50403020100

Blade angle θB [deg.]

θ

ω

u

θ

ω

u

B

Changes in crystallinity for potato starch particles ground in the impact mill with various blade angles

100

80

60

40

20

Tem

pera

ture

[o C]

12008004000Time [s]

10000

8000

6000

4000

2000

0

Vis

cosi

ty [c

P] 0o

15o

30o

45o

8000

6000

4000

Peak

vis

cosi

ty [c

P]

50403020100

Blade angle θB [deg.]

(b) Peak viscosity(a) RVA pasting curve

Pasting properties of potato starch particles ground in the impact mill with various blade angles

40

35

30

25

20Med

ian

diam

eter

Dp5

0 [µm

]

50403020100Blade angle θB [deg.]

60

58

56

54

52

50

Past

ing

tem

pera

ture

[o C]

50403020100

Blade angle θB [deg.]

(b) Particle size(a) Pasting temperature

Changes in properties of potato starch particles ground in the impact mill with various blade angles

Effects of types of mill on powder properties

Material: Potato starchMills: a mechanical impact mill, a jet mill, a ball millMeasurement:Particle size, crystallinity, pasting properties

Mechanical mill:Blade Mill Jet mill:Super Jet Mill

Chilling jacket

Stator

Rotor

Raw material

Pulverized powder

Blade

Rot

atio

nal

dire

ctio

nStator

Pulverizing zone

Structure of a mechanical impact mill (Blade Mill)

10

8

6

4

2

0

Freq

uenc

y [w

t%]

0.1 1 10 100 1000Particle diameter Dp [µm]

Wheat bran

ConventionalD50=35.2µmD100=352µm

OptimizedD50=9.6µmD100=124µm

Comparison of particle size distributions for wheat branComparison of particle size distributions for wheat branground in optimized and conventional mechanical millsground in optimized and conventional mechanical mills

Optimized mill

100µm 100µm

Conventional mill

SEM photos of wheat bran powderSEM photos of wheat bran powderground in optimized and ground in optimized and

conventional mechanical millsconventional mechanical mills

0.1

1

10

100

Spec

ific

ener

gy c

onsu

mpt

ion

w [k

Wh/

kg]

10 100Median particle diameter Dp50 [µm]

Optimized Conventional Jet mill

Wheat bran

80%

Comparison of specific energy consumption in the optimized Comparison of specific energy consumption in the optimized mechanical mill, a conventional mechanical mill and a jet millmechanical mill, a conventional mechanical mill and a jet mill

Outlet

Grinding nozzle

Inlet

Comp.air

Grinding zone(1st stage classifying)

Classifying zone(2st stage classifying)Classifying ring

Structure of a jet mill (Super Jet Mill)

Classifying ring

Inner wall structure (calculation region)

(a) Velocity vectors (b) Typical particle trajectories (Dp=10µm)

Classifying ring Classifying ring

Calculated results in the Super jet mill

Conventional jet millConventional jet mill

Relation between particle size distribution of product and mixture ratio ( color toner)

Super Jet MillSuper Jet Mill

30

25

20

15

10

5

0

Freq

uenc

y [v

ol.%

]

1 10 100 1000Particle diameter Dp [µm]

Feed (36.5µm) Jet mill (15.3µm) Mechanical mill (15.1µm) Ball mill (59.3µm)

Particle size distributions of potato starch particles ground in various mills

(a) Feed (b) Jet mill

(c) Mechanical mill (d) Ball mill

SEM photos of potato starch particles ground in various mills

5 10 15 20 25 30 352θ [deg.]

Feed

Jet mill

Mechanical mill

Ball mill

Inte

nsity

[a.u

.]

X-ray diffraction (XRD) patterns for potato starch particles ground in various mills

0 400 800 12000

20

40

60

80

100

Feed

Ball mill

0

2000

4000

6000

20

Jet mill

Mechanical mill

Temp.

T0=61.1oC，∆H=13.3J/g

Tem

pera

ture

[o C]

Vis

cosi

ty [c

P]

T0=50.7oC，∆H=4.1J/g

T0=56.1oC，∆H=11.2J/g

T0=49.2oC，∆H=3.5J/g

Time [s]

RVA (Rapid Visco Analyser） profile of potato starch particles ground in various mills

How to control color of powder by grindingMaterial: Green teaMethod:Grinding temperature and particle sizeType of mills: mechanical mill, jet mill, disc attrition mill

Measurement(color):A tristimulus colorimeter CR-310 (Minolta)

Raw material(leaves of green tea)

Ground product

L*a*

b*-60

-60+60

+60

22 *)(*)(* baC +=Chroma:

White (100)White (100)

Black (0)Black (0)

GreenGreen RedRed

BlueBlue

YellowYellow

CIE(International Committee on Illumination)L* a* b* color coordinates

L* : Lightness of a color( ranges from black at 0 to white at 100 )

C* : Chroma ( color saturation or purity )

34

32

30

28

26

24

22Ch

rom

a C*

806040200-20Outlet temperature To [

oC]

D50 8µm 12µm 15µm

76

72

68

64

60

Ligh

tnes

s L*

806040200-20Outlet temperature To [

oC]

D50 8µm 12µm 15µm

Effect of grinding temperature on color of green tea powder for various particle sizes

25

20

15

10

5

0

Freq

uenc

y [v

ol.%

]

0.1 1 10 100 1000Particle diameter Dp [µm]

Jet mill (3.6µm) Mechanical mill (7.9µm) Disc attrition mill (19.8µm)

Particle size distributions of green tea powder ground in different types of mill

20µm 20µm 20µm

(b) Mechanical mill(a) Jet mill (c) Disc attrition mill

SEM photos of green tea powder ground in various mills

72

70

68

66

64

62

60

Ligh

tnes

s L*

1 10 100Median particle diameter D50 [µm]

Jet mill Mechanical mill Disc attrition mill

Effect of particle size on color of green tea powder

How to control flavor of powder by grindingMaterial: Buckwheat, green teaMethod:

Grinding temperature and particle sizeType of mills;mechanical mill, jet mill, disc attrition mill

Measurement (flavor):GC-MS(Gas Chromatograph-Mass Spectrometer) system: GC-17A&QP-5000 (Shimazu, Japan)

Buckwheat(a) Flowers (b) Husks (c) Grains (d) Flour

4 6 8 10 12 14 16Time [min]

Abu

ndan

ce

Mechanical mill

To = 50 oC

To = 10 oC

Buckwheat flourDp50=30µm

Example of Gas chromatogram for buckwheat flour ground in the mechanical mill

0 0.5 1.0 1.5Flavor intensity ratio

Stone mill(Disc attrition mill)

Mechanical mill(To = 50oC)

Mechanical mill(To = 10oC)

BuckwheatDp50=30µm

Effect of grinding temperature on flavor of buckwheat flour ground in the mechanical mill

Effect of grinding temperature and particle size on flavor intensity of green tea powder

10

8

6

4

2

0

Flav

or in

tens

ity

6040200

Outlet temperature T0 [oC]

8 µm12µm

Stone mill(Dp50=13µm)

Mechanical mill(Dp50=15µm, To = 45oC)

Mechanical mill(Dp50=15µm, To = 20oC)

Jet mill(Dp50=11µm, To = 16oC)

Flavor intensity ratio0 0.4 1.2 1.41.00.80.60.2

Green tea

Effect of type of mill on flavor of green tea powder

ConclusionsConclusions(1) The simulated impact energies are useful parameters

for determining the optimum blade angle of the mechanical mill.

(2) The optimum blade angle for efficient changes in crystallinity of particles and pasting properties is different from that for size reduction.

(3) The particle size has a significant effect on color of powder. The grinding temperature and particle size influence flavor of powder.

(4) Types of mill strongly influence pasting properties of potato starch.