BIOBASED PRODUCTS FROM STARCH USING EXTRUSION PROCESSING
AND CHEMICAL MODIFICATIONS
By
Zhiguan Yang

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Chemical Engineering – Doctor of Philosophy
2013

ABSTRACT
BIOBASED PRODUCTS FROM STARCH USING EXTRUSION PROCESSING
AND CHEMICAL MODIFICATIONS
By
Zhiguan Yang
Biobased products from starch have shown great potential to replace petroleum
based products and they are also environmentally friendly and biodegradable. Starch
foams are one of the product forms obtained from starch and are a possible replacement
for petroleum-based foams. Such foams have been prepared in the past but their
sensitivity to humidity has remained a problem. Extensive research over the last few
years has been focused on resolving the collapse of the foam at high humidity, improving
the cushioning protection and developing an economical foaming process. However,
these issues are still problematic and need to be resolved before such biobased foams can
gain widespread entry into the marketplace.
Initially, the effects of starch feed rate, the addition of a nucleating agent, and the
extruder screw configuration were studied. It was demonstrated that screw configuration
plays an important role in the extrusion process and the nucleating agent talc is an
effective component with which to control the cell size. An annular die was used to
extrude tubular starch foams which were then sliced to yield foam sheets suitable for
cushioning protection and insulation of shipping containers. A Box-Behnken statistical
design of experiment (DOE) was used to optimize the properties of the foams extruded
with various additives. It was found that the density, cell structure and water sensitivity of
these foams were affected by the feed rates (e.g. foam composition) of the water, talc and

polyhydroxy ether (PHE). The use of PHE was found to be extremely effective in
minimizing water sensitivity.
Because PHE contains Bisphenol A, a suspected endocrine disrupter, another
alternative polymer additive – Polyvinyl buytral (PVB) was studied. It was shown that
PVB minimizes the moisture sensitivity of the foam and provided a more hydrophobic
character. A statistical design experiment was used again to identify the composition and
process parameters affecting the physical properties of the foams and to optimize the feed
rates of the various additives. The foam extrusion process to make foam sheet was scaled
down to a lab scale extruder in order to provide a more convenient technique for
exploring different raw materials and formula variations.
The performance (moisture adsorption, cushion curve, and thermal insulation) of
the starch foam sheets under different processing conditions and compositions were
evaluated. The results indicate that these foams are classified in the moderate fragile level,
implying that these foams are suitable as protective packaging of network hardware
equipment, personal computers and medical diagnostic apparatus. The starch sheets have
similar thermal resistance, R value, compared with polystyrene foams.
Finally, starch phosphate was synthesized using sodium trimetaphosphate and
glycerol phosphoric acid by an extrusion method to examine its use as a possible
electrorheological fluid or as a flame retardant additive.
The main accomplishment of this research is a starch foam sheet with a
significantly high moisture resistance that can be prepared by extrusion using readily
available and affordable resin additive. The process and the composition have been used
to scale up the manufacturing of these foam sheets for commercial production.

ACKNOWLEDGMENTS
I would like to thank my advisor Dr. Ramani Narayan, for financial support, for
the opportunity to work on this interesting project, and for the guidance during my PhD.
Also, thank Dr. Dan Graiver for his tremendous help to my research and thank Dr. Rafael
Auras, Dr. Dan Graiver, Dr. K. Jayaraman and Dr. Andre Lee for serving on my PhD
guidance committee. I would thank Ken Farminer for help with the manuscripts. I would
like to thank the members of Biobased Materials Research Group for helping me in the
many ways they did. Lastly my thanks to KTM Industries, Inc, especially, Tim
Colonnese and Rachelle Padgett, for the commercial trials and help with the lab foam
extrusions.

iv

TABLE OF CONTENTS
LIST OF TABLES .......................................................................................................... vii
LIST OF FIGURES ......................................................................................................... ix
Chapter 1 Introduction and background ....................................................................... 1
1.1 Introduction............................................................................................................. 1
1.1.1 Objectives.......................................................................................................... 2
1.1.2 Organization of this Thesis ............................................................................. 3
1.2 Background ............................................................................................................. 3
1.2.1 Introduction to starch...................................................................................... 3
1.2.2 Introduction to extrusion ................................................................................ 5
1.2.3 Starch foam extrusion (physical process) ...................................................... 8
Chapter 2 Materials, Methods and Characterizations................................................ 12
2.1 Materials ................................................................................................................ 12
2.1.1 Starch and Talc .............................................................................................. 12
2.1.2 Polymer modifiers – PVOH, PHAE, PHE, and PVB.................................. 13
2.2 Experimental ......................................................................................................... 15
2.2.1 Century ZSK-30 co-rotating twin screw extruder...................................... 15
2.2.2 Wenger TX-80 extruder ................................................................................ 20
2.3 Characterization methods .................................................................................... 21
2.4 Statistics and design of experiments methods .................................................... 25
Chapter 3 Foam extrusion results and discussion ....................................................... 36
3.1 Screw configuration effect.................................................................................... 36
3.2 Starch feed rate effect........................................................................................... 38
3.3 Effect of talc content ............................................................................................. 43
3.4 Different modifiers................................................................................................ 46
3.4.1 Aqueous PVOH .............................................................................................. 46
3.4.2 Box-Behnken DOE with PHE and annular die........................................... 48
3.4.3 Box-Behnken DOE with PVB ....................................................................... 61
3.4.4 Starch foam homogeneity with PVB ............................................................ 77
3.5 Scale down starch foam extrusion from Wenger TX-80 to Century ZSK-30 . 79
Chapter 4 Performance characterizations of starch foam sheets............................... 81
4.1 Moisture adsorption and humidity sensitivity ................................................... 81
4.2 Thermal insulation................................................................................................ 89
4.3 Cushion curves ...................................................................................................... 93
4.4 Dynamic Stress-Strain Curve ............................................................................ 100
Chapter 5 Starch foam extrusion mechanism ............................................................ 104
5.1 Extensional viscosity and strain hardening ...................................................... 104
5.2 Starch foaming steps........................................................................................... 109
5.3 Die design............................................................................................................. 109
v

5.3.1 Strand die...................................................................................................... 110
5.3.1 Annular die ................................................................................................... 113
Chapter 6 Chemically modified starch ....................................................................... 116
6.1 Starch extrusion with soy based ester ............................................................... 116
6.2 Starch phosphate................................................................................................. 123
6.2.1 Starch phosphate for ER fluid .................................................................... 123
6.2.2 Starch phosphate with glycerol phosphoric acid ...................................... 131
6.3 Dihydroxyl starch................................................................................................ 134
Chapter 7 Conclusions, accomplishments and recommendations for future work 141
7.1 Conclusions.......................................................................................................... 141
7.2 Main accomplishment......................................................................................... 141
7.3 Future recommendations ................................................................................... 142
REFERENCES.............................................................................................................. 144

vi

LIST OF TABLES
Table 1 Commercial foam products’ property.................................................................... 9
Table 2 A summary of starch foam research .................................................................... 10
Table 3 Screw configuration of Century ZSK-30 extruder (L/D=42) .............................. 17
Table 4 Screw configuration of Century ZSK-30 extruder (L/D=20.4) ........................... 18
Table 5 Screw configurations for Wenger TX-80 ............................................................ 20
Table 6 Different salt solutions for relative humidity....................................................... 22
Table 7 Calculation for ANOVA...................................................................................... 29
Table 8 Cylindrical foam extrusion conditions and diameter of the foams...................... 39
Table 9 cylindrical foam characterizations ....................................................................... 42
Table 10 Box-Behnken DOE with PHE and annular die.................................................. 49
Table 11 Polynomials used to evaluate the foam density and cell diameter .................... 50
Table 12 Design space using polyvinyl butyral................................................................ 61
Table 13 Box-Behnken DOE for starch foam with PVB.................................................. 62
Table 14 Box-Behnken DOE experimental results for starch foam with PVB ................ 63
Table 15 Polynomials used to evaluate the foam diameter, density and wear ................. 64
Table 16 Table for 3D plots (noodle diameter and density as responses) ........................ 66
Table 17 Experimental conditions and results for starch foams prepared by the Century
ZSK-30 and Wenger TX-80 extruders.............................................................................. 79
Table 18 Formulation for lab starch foam tube ................................................................ 80
Table 19 Lab starch foam tube characterization ............................................................... 80
Table 20 k1 and k2 values for starch powder..................................................................... 83
Table 21 k1 and k2 values for PELEG model (85% RH) .................................................. 83

vii

Table 22 flexibility of the foam sheets under different relative humidities at room
temperature ....................................................................................................................... 88
Table 23 Soxhlet extraction with different solvents ....................................................... 118
Table 24 Products from ozonolysis................................................................................. 119
Table 25 Extrusion conditions for thermal plastic starch with esters ............................. 120
Table 26 Suspensions for electrorheological fluid.......................................................... 124
Table 27 Extrusion formulation using sodium trimetaphosphate and sodium
tripolyphosphate.............................................................................................................. 133
Table 28 Phosphorylate starch using glycerol phosphate disodium salt......................... 133
Table 29 Phosphorylate starch using glycerol phosphoric acid...................................... 133
Table 30 Phosphorylate starch using glycerol phosphoric acid and glycerol phosphate
monohydrogen ................................................................................................................ 134
Table 31 Intrinsic viscosity comparison ......................................................................... 134
Table 32 weights for reactants during DHS synthesis .................................................... 136

viii

LIST OF FIGURES
Figure 1 Carbon cycle......................................................................................................... 2
Figure 2 Chemical structures of starch and cellulose ......................................................... 4
Figure 3 Chemical structure of amylose and amylopectin.................................................. 5
Figure 4 Extruder system .................................................................................................... 6
Figure 5 Kneading elements, covey elements and barrels.................................................. 6
Figure 6 Distribution and dispersive mixing ...................................................................... 7
Figure 7 Starch plasticization.............................................................................................. 8
Figure 8 Foam extrusion process ........................................................................................ 8
Figure 9 Hydroxypropylate reaction................................................................................. 12
Figure 10 Polyhydroxy aminoether synthesis................................................................... 13
Figure 11 Chemical structure of polyvinyl alcohol .......................................................... 14
Figure 12 Polyhydroxyl ether synthesis............................................................................ 14
Figure 13 Chemical structure of polyvinyl butyral........................................................... 15
Figure 14 The barrels arrangement for the screw configuration (L/D=20.4) ................... 15
Figure 15 The barrels arrangement for the screw configuration (L/D=42) ...................... 16
Figure 16 Screw configuration of Century ZSK-30 extruder (L/D=20)........................... 19
Figure 17 Annular die ....................................................................................................... 21
Figure 18 Screw configuration for Wenger TX-80........................................................... 21
Figure 19 Thermal conductivity test ................................................................................. 24
Figure 20 Sample size from bulk ...................................................................................... 26
Figure 21 Operating Characteristic Curve (α=β=0.05)..................................................... 26
Figure 22 t and F distribution curves ................................................................................ 27
ix

Figure 23 A: Function vs. polynomial (square terms); B: Function vs. Function vs.
polynomial (cubic terms) .................................................................................................. 33
Figure 24 Experimental points for Box-Behnken design ................................................. 35
Figure 25 DOE for 4 variables.......................................................................................... 35
Figure 26 Specific mechanical energy (SME) comparison for two screw configuration. 36
Figure 27 Regular starch foam sheets (left) vs. sheets with braiding problem (right two)38
Figure 28 Four and six holes strand die ............................................................................ 38
Figure 29 Cells in the starch foam .................................................................................... 39
Figure 30 Cell size distribution for cylindrical foams (A: noodle extruded at 460 lbs/hr, B:
noodle extruded at 285 lbs/hr) .......................................................................................... 41
Figure 31 Characterization of cylindrical foam prepared from a die with six holes ........ 42
Figure 32 Cell size distributions for starch foam with PVOH, different talc levels: A:
0.5wt%, B: 1.3wt%, C: 2.7wt% ........................................................................................ 43
Figure 33 Cell size distributions for starch foam with PHAE, different talc levels: A:
0.5wt%, B: 1.3wt%, C: 2.7wt% ........................................................................................ 45
Figure 34 starch foam densities vs. PVOH solution concentration .................................. 47
Figure 35 Starch foam resiliencies vs. PVOH solution concentration.............................. 48
Figure 36 Goodness of fit of the calculated polynomials to the experimental values. ..... 51
Figure 37 Effects of resin and talc feed rates on the foam density at low (A) and high (B)
water feed rates. ................................................................................................................ 53
Figure 38 Effects of water and resin feed rates on the foam density at low (A) and high (B)
talc feed rates. ................................................................................................................... 55
Figure 39 Effects of water and talc feed rates on the average cell diameter at low (A) and
high (B) resin feed rates.................................................................................................... 57
Figure 40 Effects of talc and resin feed rates on the average cell diameter at low (A) and
high (B) water feed rates................................................................................................... 59
Figure 41 Goodness of fit of the calculated polynomials to the experimental values. A:
Foam diameter B: Density C: Wear.................................................................................. 66
x

Figure 42 3D plots, noodle diameter as response ............................................................. 68
Figure 43 3D plots, density as response............................................................................ 71
Figure 44 Changes in the wear as a function of PVB and water feed rates with no talc and
extruding through a small pinhole die (diameter = 10 mm) ............................................. 73
Figure 45 Changes in the wear as a function of PVB and water feed rates with talc
coextruded at 0.0155 Kg/hr through a large pinhole die (diameter = 20 mm) ................. 74
Figure 46 Changes in the wear as a function of PVB and talc feed rates with water
coextruded at 0.80 Kg/hr through a small pinhole die (diameter 10 mm) ........................ 75
Figure 47 Changes in the wear as a function of PVB and talc feed rates with water
coextruded at 0.50 Kg/hr extruding through a pinhole die with a diameter = 16 mm...... 76
Figure 48 Changes in the wear as a function of PVB and pinhole diameter with water
coextruded at 0.80 Kg/hr and no talc ................................................................................ 77
Figure 49 TGA analysis on starch foam with PVB .......................................................... 78
Figure 50 Hydroxypropyl high amylose starch powder moisture content vs. time at 23°C
........................................................................................................................................... 81
Figure 51 Starch foam (no polymer modifier) moisture content vs. time at 23°C ........... 83
Figure 52 Starch foam (PVOH as polymer modifier) moisture content vs. time at 23°C 84
Figure 53 Starch foam (PHAE as polymer modifier) moisture content vs. Time at 23°C 84
Figure 54 Starch foam (PHE as polymer modifier) moisture content vs. time at 23°C ... 85
Figure 55 Starch foam with PVB as polymer modifier moisture content vs. time at 23°C
........................................................................................................................................... 85
Figure 56 Moisture content at equilibrium under different RH (starch foam with PVOH
as polymer modifier)......................................................................................................... 86
Figure 57 Water penetration time for foam sheets (A) foam with different additives (B)
foam with different additive content................................................................................. 87
Figure 58 Starch foam sheets extruded after 5 hours in contact with water. (A) PVOH (B)
PHE. .................................................................................................................................. 88

xi

Figure 59 Thermal insulation property comparison (A: different additives B: different talc
concentration) ................................................................................................................... 91
Figure 60 Starch foam without polymer modifiers cushion curves.................................. 95
Figure 61 Starch foam with PHAE cushion curves .......................................................... 95
Figure 62 Starch foam with PVOH cushion curves.......................................................... 96
Figure 63 Starch foam with PHE cushion curves ............................................................. 96
Figure 64 Starch foam with PVB cushion curves............................................................. 97
Figure 65 Polyethylene foam cushion curves ................................................................... 97
Figure 66 Cushion curves comparison (First drop) .......................................................... 98
Figure 67 Cushion curves comparison (Fifth drop).......................................................... 98
Figure 68 Starch foam with PVOH at 75% RH and RT ................................................... 99
Figure 69 Starch foam with PVOH and 1.3% talc............................................................ 99
Figure 70 Starch foam with PVOH and 2.7% talc.......................................................... 100
Figure 71 Stress-strain curve .......................................................................................... 102
Figure 72 First drop cushion curves for 1 inch and 2 inch starch foam.......................... 103
Figure 73 The sh/t v.s. (a/G+1)s curve ........................................................................... 103
Figure 74 Strain hardening.............................................................................................. 104
Figure 75 Uniaxial pull of polymer block....................................................................... 104
Figure 76 strain hardening effects on foam and film process (1: force applied to a block, 2:
foaming process, 3: film process; a: original state, b: strain weakening, c: stain hardening)
......................................................................................................................................... 108
Figure 77 Strand die........................................................................................................ 110
Figure 78 Annular die ..................................................................................................... 113
Figure 79 Die pressure (Pa) vs. die gap (m), curve: calculated from the model, two dots:
experimental values ........................................................................................................ 115

xii

Figure 80 Reactions in MTPS........................................................................................ 116
Figure 81 Ozonolysis ...................................................................................................... 119
Figure 82 Transesterification reaction ............................................................................ 120
Figure 83 GC-FID before and after ozonation using Dean-stark setup (The retention time
for dimethyl azelate is 20.74 minutes)............................................................................ 121
Figure 84 FTIR results for the extrudates after soxhletion with acetone (A) MTPS (B)
MTPS with two phases esters (C) MTPS with one phase ester...................................... 122
Figure 85 Soxhletion results (acetone as solvent) for the extrudate ............................... 122
Figure 86 Alignment of suspended particles .................................................................. 124
Figure 87 Starch phosphorylation using sodium trimetaphosphate................................ 126
Figure 88 Feed position for starch phosphorylation ....................................................... 127
Figure 89 Rheology measurement setup......................................................................... 128
Figure 90 Dynamic light scattering for starch nanoparticles .......................................... 129
Figure 91 starch nanoparticles stabilized with phosphate............................................... 129
Figure 92 Viscosity v.s. shear rate for ER fluid.............................................................. 130
Figure 93 Starch phosphorylation reaction ..................................................................... 132
Figure 94 Reactions to make dihdyroxyl starch ............................................................. 135
Figure 95 Setup for measuring aldehyde group.............................................................. 137
Figure 96 FTIR for starch, DAS and DHS, A: high amylose, B: waxy, C: regular. ...... 138
Figure 97 Viscosity vs. shear rate for DHS .................................................................... 139
Figure 98 Dialdehyde content in DAS and DHS ............................................................ 140
Figure 99 annular die (Pin) structure .............................................................................. 142
Figure 100 Corrugation in the foam sheet (left: extruded with 10 holes pin, middle:
extruded with 6 holes pin, right: extruded with 10 holes pin and small gap) ................. 143

xiii

Chapter 1 Introduction and background
1.1 Introduction
Advances in petroleum-based polymers have benefited mankind in numerous
ways. Most general plastics are petroleum based and hence are derived from nonrenewable resources.
The critical disadvantage of using petroleum-based resources can be understood
based on the carbon cycle. As shown in the Figure 1, there are four elements (#1 fossil
resources, #2 polymer products, #3 carbon dioxide, and #4 bio-mass) in the cycle. Crude
oil is pumped out of the ground and is then converted into products --- polymers,
chemicals and fuels ---through chemical processing. These products become carbon
dioxide after usage (end-of-life). The carbon dioxide in the atmosphere is fixed through
photosynthesis by biomass and agricultural crops. After biomass plants die, it takes
millions of years to become fossils. The problem in this carbon cycle is that there is one
limiting step, from biomass to fossil resources, which takes too long. If the cycle
continues without modification, in time all the carbon will accumulate in biomass and
agricultural crops and no petroleum will remain for the production of the products on
which we currently rely.

1

Figure 1 Carbon cycle
It is critical, therefore, to avoid the #1 fossil resource step in this carbon cycle in
order to balance the carbon flow rate. This is where bio-based products come into play,
by by-passing the fossil resource elements to use technology to directly convert biomass
to polymers, chemicals and fuels. Although bio-based polymers have been studied for
many years, those polymers still have limitations, such as process instability and poor
product performance compared to petroleum- based polymers. Therefore, there is a need
investigate the production process itself in order to improve the processability and
performance of biobased and biodegradable products.
1.1.1 Objectives
This study focuses on using starch to make useful products - starch foam,
thermoplastic starch, etc - and attempts to improve the process stability and performance
of products based on starch foam and explores the potential possibility of chemically
modifying the starch for define applications.

2

1.1.2 Organization of this Thesis
This thesis is divided into seven parts: In chapter 1, the need for biobased
products from starch is addressed. The chapter also includes a review of products based
on starch.
Chapter 2 gives a detailed raw materials information, experimental setup and
experimental procedures/characterization methods. This chapter also includes the basics
of statistical design of experiments used in this study.
In Chapter 3, the effects of different composition and processing conditions
(screw configuration, starch, talc, polymer modifiers) on the properties of the starch foam
are studied.
In Chapter 4, the performance properties of starch foam sheets (moisture
adsorption, cushioning protection and thermal insulation) are investigated.
In Chapter 5, the focus is on starch foam extrusion mechanism.
In Chapter 6, chemical modifications of the starch (maleated thermoplastic starch
with hydrophobic improvement, starch phosphate, dihydroxyl starch) using extrusion are
studied.
In Chapter 7, conclusions are drawn based on finished work and future work is
suggested.
1.2 Background
1.2.1 Introduction to starch
Starch, an abundant, inexpensive and a naturally occurring polymer, provides a
good platform from which to manufacture renewable and biodegradable foams for

3

packaging and insulation applications. Starch and cellulose are two common
carbohydrates. Starch contains alpha-glucose as its monomer, whereas cellulose contains
beta-glucose (Figure 2).
H OH

H

H

O
H

O

O

OH

H

HO

HO

H

OH

H

H

H
O

O

H

H

H

HO

H OH
H

O

OH

H
HO

O

O

O

HO
H

OH

H

H
H

O

H OH

Figure 2 Chemical structures of starch and cellulose
Starch exists in granule structure form and is bimodal and polydispersed, both at
the granular and molecular levels.
Amylose is a linear polymer with molecular weight in the range of 105 to 106
g/mol. Amylopectin is a branched polymer with molecular weight in the range of 107 to
109 g/mol, with branching points (α(1-6) linkage) occurring every 25-30 glucose units
(Figure 3).

4

The ratio of these two components depends on the source of the starch and can
vary from 100% amylopectin to 100% amylose either reported as occurring in nature or
as a result of classical plant breeding.
CH2OH

CH2OH

O

O

H

H

OH

H

H

O

OH

OH

H

H

O

OH

O

O
H
HO

OH H
H

CH2OH
O

O
CH2OH
O

O

O

H

H

H

OH

H

H

O

CH2OH

CH2

OH

O

OH

H

H

OH

O

OH

H

H

OH

O

Figure 3 Chemical structure of amylose and amylopectin
1.2.2 Introduction to extrusion
An extruder system is shown in Figure 4. The main components of a twin screw
extruder are screw elements and barrels (Figure 5 [1]), motor, feeder and die.

5

Figure 4 Extruder system
For interpretation of the references to color in this and all other figures, the reader is
referred to the electronic version of this dissertation.

Figure 5 Kneading elements, covey elements and barrels
Extrusion is typically used for 1) melt blending 2) filler dispersion 3) chemical
modification (branching, functionalization etc) 4) Co-polymerization reaction. The
combinations of high temperature, high shear and high pressure provide excellent mixing.

6

There are two different mixing concepts (distributive mixing and dispersive
mixing) which are shown in Figure 6. When the kneading disk is narrow, the extruded
mixture can easily by pass the top of the disk resulting in more distributive mixing. When
the kneading disk is wide, the extruded mixture is more likely to reach the top of the
disk/paddle (Figure 6, [1]), resulting in more dispersive mixing.

Figure 6 Distribution and dispersive mixing
Starch has poor thermal processing properties; it decomposes before it melts due
to its strong hydrogen bonding association and crystallization. This makes starch
unsuitable for thermoplastic applications. However, using a plasticizer in a twin screw

7

extruder with appropriate screw elements can break up the hydrogen bonding within the
starch, disrupts the crystalline region (Figure 7) and makes it flow like a thermoplastic
material, since water can form hydrogen bonding with starch and release the starch chain
in the granule without significantly reducing the starch molecular weight.

Figure 7 Starch plasticization
1.2.3 Starch foam extrusion (physical process)
Starch loose-fill foam producing steps (Figure 8): (1) Feed starch, additives and
water into a twin screw extruder. Under high pressure, temperature and shear, the starch
granular structure is destroyed. (2) As the mixture exits the die, the sudden drop in
pressure causes the water to turn into steam that acts as a blowing agent. Screw
configuration is an important factor in this process. Commercial foam properties are
summarized in Table 1.

Figure 8 Foam extrusion process
8

3

Trade name [foam peanuts]
Resiliency Density (kg/m )
Pelaspan Pac (EPS based)
79.30%
9.6
Flow-Pak S (EPS based)
82.70%
7.5
Star-Kore
70.20%
20.6
Flo-Pak Bio 8
70.10%
16.8
Envirofil
67.80%
22.1
Renature
67.20%
21
Clean Green
68.80%
21.8
Table 1 Commercial foam products’ property
Many biodegradable polymers and common petroleum based polymers as additives were
extensively studied in order to improve the physical properties of the starch foams (
Table 2) [2-54]. Starch ester was also studied to improve the humidity resistance
of starch foam. Curiously, few researchers focused on the technical aspects (screw
configuration, die design, etc.) of the foam extrusion, each of which is crucial for a
successful process. Also, most of the researchers used strand dies for making cylindrical
foam. This research focuses on the technical/engineering aspects of the starch foam
extrusion, an understanding of the principals involved, investigation of the relationship
between those operational parameters and characterization of starch foam sheets
produced using an annular die.

9

Materials added to starch and water
Polyhydroxy aminoether
Maleated poly(butylene adipate-co-terephthalate)
Epichlorohydrin, acetic anhydride, supercritical carbon dioxide
Wheat flour, wheat bran fibers
Cellulose fiber
Glycerol
Sugarcane bagasse fibers, polyvinyl alcohol
Epichlorohydrin, supercritical carbon dioxide
Supercritical carbon dioxide
Polycaprolactone, supercritical carbon dioxide
α-cellulose; polylactic acid; polystyrene; glycerol; NaCl; talc; Na2CO3;
citric acid
Paper powder, polypropylene
Polylactic acid, Cloisite 10A organoclay
Polylactic acid, Cloisite 30B organoclay
Polylactic acid, Cloisite 10A, 25A, 93A, 15A organoclay
Polylactic acid, Cloisite 30B, Na+, 20A clay
Polystyrene, talc, Azodicarbonamide, citric acid
Polystyrene, polycarbonate, talc
Polylactic acid
Egg shell powder as nucleating agent
Thermoplastic starch, polylactic acid, carbon dioxide as blowing agent
Thermoplastic starch, polystyrene, 1,1,1,2-Tetrafluoroethane
Sorbitol, glycerol, poly(ethylene-co-vinyl alcohol)
Starch acetate, ethanol as blowing agent
Starch acetate, cellulose
Starch acetate, corn cob fiber, ethanol
Starch acetate, natural fiber, ethanol
Starch acetate, cellulose, ethanol
Starch acetate, ethanol, Ethyl acetate
Starch acetate, water or ethanol as blowing agent
Starch acetate, corn stalk fibers
Starch acetate + native corn starch
Yellow dent corn
polyvinyl alcohol, cellulose acetate, polylactic acid, polyhydroxyester ether,
polycaprolactone, polyester amide and poly(hydroxybutyrate-co-valerate)
Poly(ethylene-co-vinyl alcohol)
Mater-Bi®
Eastar Bio Copolymer®
Poly(ethylene-co-vinyl alcohol), polystyrene
Table 2 A summary of starch foam research

10

Reference
2
3
4
5
6
7
8
9
10, 11
12
13
14
15
16
17
18
19
20
21, 22, 23
24
25
26
27
28
29
30
31
32
33
34, 35
36
37
38
39
40
41
42
43

Table 2 (cont’d)
Polystyrene, magnesium silicate, polycarbonate, azodicarbonamide
Polystyrene, polymethyl methacrylate
Polyalkylene glycol, silica
Poly(ethylene-co-acrylic acid), urea, NaHCO3
>45%, ~70% amylose
Starch ester, modified with alkylene oxide
Poly(ethylene-co-acrylic acid), poly(ethylene-vinyl alcohol), Na2CO3

11

44
45
46, 47
48
49,50,51
52
53, 54

Chapter 2 Materials, Methods and Characterizations
2.1 Materials
2.1.1 Starch and Talc
Hydroxypropylated high-amylose (70%) starch (molecular weight: 250K~1000K
g/mol) was purchased from Ingredion Incorporated (Westchester, IL). This starch was
chosen due to its reduced branching (less cross linking) compared with regular starch,
which has a higher amylopectin content. The high amylose starch gives the foam good
resiliency due to its linear chain compared to regular starch (30wt% amylose). The high
amylose starch was treated with propylene oxide (5wt%) to disrupt hydrogen bonding
and thus improve processability (Figure 9).
OH

O
starch

OH +

H2C

CH

starch

CH3

O

CH2

CH

CH3

Figure 9 Hydroxypropylate reaction
In commercial scale extrusions, in order to obtain a satisfactory feed rate, the
hydroxypropylated high amylose starch is chilsonated (compact granulated) and in pellet
form.
Water was used as both a plasticizer and a blowing agent. Talc (hydrated
magnesium silicate) was used as a nucleating agent.

12

2.1.2 Polymer modifiers – PVOH, PHAE, PHE, and PVB
Several polymer modifiers were used to modify the starch foam properties and
improve extrusion process. They are polyhydroxyl aminoether (PHAE), polyvinyl
alcohol (PVOH), polyhydroxy ether (PHE) and polyvinyl butyral (PVB).
Polyhydroxy aminoether (PHAE) was supplied by the Dow Chemical Co.
(Midland, MI), under the trade name of BLOX 110. It is produced by the reaction of
Bisphenol A diglycidyl ether with monoethanol amine using a reactive extrusion process
(Figure 10).

O

O
NH2

HO
O

O

*

O

O

OH

N

*
n

OH
OH

Figure 10 Polyhydroxy aminoether synthesis
Polyvinyl alcohol (Mowiol 40-88) (Figure 11) was obtained from Kuraray
America, Inc. (Houston, TX). It is synthesized by hydrolysis of polyvinyl acetate. The
“40” in the grade title indicates 4% water solution. The viscosity is 40cp and the degree
of hydrolysis is 88%.

13

*

*
O

m

n
OH

O
Figure 11 Chemical structure of polyvinyl alcohol
Polyhydroxyl ether (PHE) was purchased from InChem Corporation (Rock Hill,
SC) under the trade name PKHH. The polyhydroxyl ether is synthesized by condensation
of bisphenol A and epichlorohydrin (Figure 12).

O
HO

OH

*

O

Cl

O

*
OH

n

Figure 12 Polyhydroxyl ether synthesis
Polyvinyl butyral (PVB) was supplied by Kuraray America, Inc. (Houston, TX)
PVB 60HH was used in all the experiments. The “60” in the title grade refers to the
molecular weight, the number is proportional to the molecular weight; “HH” indicates the
residual PVOH content, with PVOH 11-14%, PVAc 1-8% and PVOAcetal 78-88%.

14

*

*
O

p

m
OH

n
O

O

O

Figure 13 Chemical structure of polyvinyl butyral
2.2 Experimental
2.2.1 Century ZSK-30 co-rotating twin screw extruder
The laboratory starch foam extrusion was done with a twin screw extruder (Century
ZSK-30) with an L/D of 42:1 (
Figure 15, Table 3) and 20.4:1 (Figure 14 and Figure 16, Table 4). Two sets of screw
configuration were used. The numbers in the Figure 14 and Figure 15 are the barrel
orders.

Figure 14 The barrels arrangement for the screw configuration (L/D=20.4)

15

Figure 15 The barrels arrangement for the screw configuration (L/D=42)

16

#
L/D = 42:1
1
28/14
2
60/60
3
60/60
4
60/60
5
42/42
6
28/28
7
28/28
8
20/20
9
20/20
10
KB 45/5/14
11
KB 45/5/14
12
KB 45/5/14
13
KB 45/5/20
14
KB 45/5/20
15
60/60
16
42/42
17
28/28
18
28/28
19
20/20
20
KB 45/5/42
21
KB 45/5/42
22
60/60
23
42/42
24
28/28
25
28/28
26
20/20
27
KB 45/5/14
28
KB 45/5/14
29
KB 90/5/28
30
KB 90/5/28
31
60/60
32
42/42
33
42/42
34
28/28
35
28/28
36
28/14
37
20/20
38
KB 90/5/28
39
20/20
40
20/20
41
20/20
Table 3 Screw configuration of Century ZSK-30 extruder (L/D=42)

17

#

Usable L/D
=20.4:1
1
28/14
2
60/60
3
60/60
4
60/60
5
42/42
6
28/28
7
28/28
8
KB 45/5/14
9
KB 45/5/14
10
KB 45/5/20
11
KB 90/5/28
12
42/42
13
28/28
14
28/28
15
KB 45/5/42
16
KB 45/5/42
17
KB 45/5/14
18
42/42
19
28/28
20
28/14
21
60/60
22
60/60
23
60/60
24
42/42
25
42/42
26
28/28
27
20/20
28
20/20
29
KB 45/5/14
30
KB 45/5/14
31
20/20
32
20/20
33
KB 45/5/28
34
KB 45/5/28
35
20/20
36
20/20
37
KB 90/5/28
38
KB 90/5/28
39
20/20
40
20/20
41
20/20
Table 4 Screw configuration of Century ZSK-30 extruder (L/D=20.4)

18

Figure 16 Screw configuration of Century ZSK-30 extruder (L/D=20)
A peristaltic pump was used for injecting water into the extruder and accurate
single-screw feeders were used for feeding starch and additives. A screw speed of
200rpm was used for laboratory experiments for foam noodle extrusion. The strand die
used was 3.1mm in diameter and 12.4 mm in land length. The slit die was 25.4mm in
width, 1.27mm in height, and 12.7mm in land length.
When an annular die (0.38mm gap and 54 mm outside diameter, 3.5mm land
length) (Figure 17) was used to produce the foam tube, the screw speed was set at
400rpm.
Initially, during start-up, water was pumped into the water feed port at 20-30% of
the starch feed rate, and later its flow rate was reduced to about 5-8% of starch. The raw
materials were fed individually or pre-blended with each other in the lab scale extruder.

19

2.2.2 Wenger TX-80 extruder
Commercial scale starch foam extrusion was done in a twin screw extruder
(Wenger TX-80) with an L/D of 16:1. All materials were fed separately using precalibrated feeders or pumps. An annular die was used to produce foam sheet and a strand
die was used to produce foam peanuts (with a high speed cutter). The strand die
employed 4 holes and 6 holes with 2.9 mm and 2.3 mm diameters, with land
length/diameter ratio 1.5:1. The annular die used to produce the foam sheet has 6.25 inch
Do, 6.186 inch Di, and 0.31 inch L (land length), see Figure 17. The raw materials were
fed individually in the commercial scale extruder. The screw configurations used are
shown in Table 5 and Figure 18.
Config. #1

Config. #2
80/120 (single flight, under
cut)
40/120 (single flight)
60/60
60/120
KB 45/3/40
KB 45/3/40
60/120
KB 45/3/40
KB 90/6/80
KB 45/3/30
KB 45/3/40

Config. #3
80/120 (Single flight,
40/120 (single flight)
undercut)
40/120 (single flight)
40/120 (Single flight)
60/60
60/60
40/120
40/120
KB 30/4/40
KB 30/4/40
KB 30/4/40
KB 30/4/40
40/120
40/120
KB 90/8/106.67
KB 45/11/146.67
KB 90/5/50
KB 90/1/10
40/120
40/120
KB 45/4/53.33 (R)
KB 45/3/40
KB 90/7/93.33 (fist disk
KB 45/4/53.33 (R)
60/120
is 45° with previous disk)
KB 90/3/40
60/120
KB 45/1/10
40/120
60/120
40/120
60/115 (conical)
60/115 (conical)
60/115 (conical)
Table 5 Screw configurations for Wenger TX-80

20

Figure 17 Annular die

Figure 18 Screw configuration for Wenger TX-80
2.3 Characterization methods
The foam products were characterized by the following methods:
Precondition samples: The foam samples collected were conditioned as per
ASTM D-4332 at 23±1°C and 50±2% relative humidity for 3 days before testing.
Density, Expansion Ratio: The density of the foam was calculated from the mass
and volume of specimen according to test method ASTM D-3575 (Section 43, Method A).
The dimension of the sample was measured using a Vernier caliper graduated to permit

21

measurements to accuracy of ±0.01 mm. The expansion ratio was calculated as the ratio
of the cross-sectional area of the foam to the area of the opening of the die.
Compressive Strength and Resiliency: Compressive strength was measured on a
United Testing Systems SFM-20 tensile testing machine. The specimens were cut into 1
inch long pieces and then were securely fastened lengthwise and compressed with a steel
probe with a hemispherical end cap (0.25 inch diameter). By lowering the probe to the
foam surface, an initial load of 0.5 N was applied to the specimen for approximately 5
seconds. The probe was lowered further at a rate of 0.5 mm/s for a distance of 3 mm and
then held in that position for 1 minute. Compressive strength was obtained by dividing
the maximum load over the cross-sectional area of the probe. Resiliency was calculated
from the percentage of the compressive force after the 60 seconds holding period divided
by the maximum force required to compress the foam by 3 mm.
Cell size characterization: Environmental scanning electron microscope (ESEM)
and pocket microscope were used to take pictures on the cutting surface of the foam.
Moisture adsorption: The moisture adsorption under different relative humidity
(RH) at room temperature was studied. Different salt saturated solutions were placed
under the plate in closed desiccators to create different RH environment (Table 6).
Salt
RH
Potassium acetate
23%
Magnesium chloride 33%
Potassium carbonate 43%
Magnesium nitrate
52%
Sodium bromide
59%
Potassium iodide
70%
Sodium chloride
75%
Potassium chloride
85%
Table 6 Different salt solutions for relative humidity

22

Cushioning property: The cushioning property of the starch foam sheet was
measured by a cushioning testing machine manufactured by Lansmont Corp. (Monterey,
CA) following ASTM D-1596 [55] test procedure. In this test a guided platen assembly
of known weight was dropped from a known height (0.75m) onto a motionless cushion
sample of known bearing area and thickness (50.8 mm thickness). The dynamic shock
cushioning characteristics of the sample were obtained through the deceleration-time data
that was recorded by an accelerometer.
Insulation property: The thermal conductivities of foam sheet and foam paper
were measured using the ice-melt test and Anter Thermal Conductivity Unit, respectively.
The Ice-melt test includes the following steps: Make cubical cavity with 6 walls (10 inch
* 10 inch) using starch foam sheets. Place ice in a sealed plastic bag. When the ice is wet,
discard any water. Then place the bag into the test package and weigh the water volume
after several days.
Calculation:

k ⋅ 6 ⋅W 2 ⋅
Where

hsf

T1 − T2
⋅ t m = M ⋅ hsf
L

⇒

k=

M ⋅ hsf ⋅ L
6 ⋅ W 2 ⋅ (T1 − T2 ) ⋅ t m

latent heat of ice (J/kg), M is the weight of ice melt (kg),

(2-1)

t m is the melting

time (s), W it the area of one face of the cubic, T1 is the room temperature, T2 is the
melting temperature of ice (0°C). Control experiments are done using empty corrugated
boxes without foams.
The insulation property is also measured by Anter Thermal Conductivity Unit
(Figure 19). The sample used is 1 inch diameter disk.

23

Figure 19 Thermal conductivity test

q
dT
= −k
A
dx

⇒

R S + Rint ΔTS
=
RR
ΔTR

(2-2)

Rint is interfacial thermal resistance, A is the area of the contact surface, q is the heat
transfer rate, k is the conductibility,

dT
is temperature gradient, and Rs is thermal
dx

resistance of sample material, RR is the thermal resistance of reference material.
temperature difference cross the sample,

ΔTR

ΔTS

is

is temperature difference cross the

reference material. Calibration must first be done using known thermal conductivity

ΔTS
materials (reference), RS is proportional to
ΔT R
Water penetration time test (hydrophobic test): The hydrophobic character of the
foams was measured by monitoring the time it took a water stream to penetrate the foam
24

sheets. In these experiments the water flow rate was 60 kg/hr, the tube diameter was 8
mm and the vertical distance between the outlet of the tube and the foam surface was kept
constant at 20 cm.
Surface wetting test (hydrophobic test): The foam sheets were dipped into water,
removed and left undisturbed for five hours. After five hours, photographs were taken of
the foam sheets to determine the degree of disintegration.
Wear (hydrophobic test): The resistance of the foams to moisture was measured
by wet wear using a Sheen machine (Sheen Wet Abrasion Scrub Tester, model 903/PG
manufactured by Sheen Instruments, Ltd, Arlington Heights, IL). The test consists of four
scrubbing heads that rub against the test sample under a load over a predetermined period
of time. In this test the foam samples were placed on panels and the scrubbing heads were
swept back and forth across them. The load was set to 800 grams and the scrubbing
speed was set to 37 strokes per minute for a total of 5 minutes.
2.4 Statistics and design of experiments methods
Sample size selection
The reason for choosing appropriate sample size: Figure 20, the length of the
sticks, if there are much variation in the bulk, more samples are needed in order to
represent the bulk. If the stick is uniform, only one sample is need for characterization.

25

Figure 20 Sample size from bulk
In order to make sure that the selected sample represented the bulk properties,
O.C. curve (Operating Characteristic Curve) (Figure 21) was used for selecting the
correct sample size.

Figure 21 Operating Characteristic Curve (α=β=0.05)
The bulk standard deviation for measuring the diameter is approximately 0.4 mm,
and the mean difference which is wanted to detect is preferably 1mm, According to the
26

curve, the sample number is around 10. In the cylindrical foam diameter measurement, 10
samples from different running times were collected to be taken for measurement.
Variation Analysis
t distribution and F distribution (Figure 22) were used to calculate the confidence
interval (CI) and variation analysis 56.

Figure 22 t and F distribution curves
The confidence interval was calculated by the following equations.

tα / 2 ,n −1 = tinv(α , n - 1)

(2-3)
27

μ = y ± CI

(2-4)

CI = tα / 2,n −1
Where

σ

(2-5)

n

tα / 2,n −1

is the t distribution value at n degree of freedom and α possibility,

tinv is the Excel® function. The σ is the standard deviation.
ANOVA table was used to determine whether a treatment is effective.
i is the factor level (1 ≤ i ≤ a)
j is the replicate (1 ≤ j ≤ n)
N is the total number of experiments (N = a*n)

28

Source of
Variation

Sum of
Squares

Treatments SSTreatment
SSError
Error
SSTotal
TOTAL

Degrees of
Freedom

Mean
Square

a-1
N-a
N-1

MSTreatment
MSError

F0
Ratio of
mean
squares

Table 7 Calculation for ANOVA

29

Fcritical
Use
Excel’s
finv

P value
Use
Excel’s
fdist

Conclusion
Base on
rejection
criteria

n

a

j =1

i =1

y i• = ∑ y ij ; y•• = ∑

SS Treatment

SSTotal

n

∑y
j =1

ij

1 ⎛ a 2 ⎞ y •2•
SSTreatment
= ⎜ ∑ yi• ⎟ −
; MSTreatment =
n ⎝ i =1
a −1
⎠ N

⎛ a
= ⎜∑
⎜
⎝ i =1

⎞ y•2•
SS
y ⎟−
MS Total = Total
∑ ⎟ N;
N −1
j =1
⎠
n

2
ij

SS Error = SSTotal − SSTreatment ; MS Error =

SS Error
N −a

(2-6)

(2-7)

(2-8)

The criteria for effective treatment is to compare the two sources of variation (a
treatment and error) and comparing this value (F0) to the F value, if F0 is located to the
far right side of the F distribution curves, the two variations are significantly different and
the treatment is effective.
Design of Experiments
Every function can be extended by the following Taylor expansion (polynomial)
and this polynomial can be used directly for the prediction after the coefficients were
obtained / regressed through experimental data.
Taylor expansion (n variables)

f ( x1 + Δx1 , x 2 + Δx 2 , L , x n + Δx n ) =
i

∂
∂
∂ ⎞
1⎛
∑ i ! ⎜ Δx1 ∂x + Δx 2 ∂x + L + Δx n ∂x ⎟ f ( x1 , x 2 , L , x n )
⎜
⎟
i =0
1
2
n ⎠
⎝
∞

For functions with two variables:

30

(2-9)

i

∂ ⎞
1⎛
∂
f ( x + Δx , y + Δy ) = ∑ ⎜ Δx + Δy ⎟ f ( x , y )
⎜ ∂x
∂y ⎟
i =0 i ! ⎝
⎠
∞

= f ( x, y ) + Δx

∂f ( x, y )
∂f ( x, y )
+ Δy
∂x
∂y

2
2
1⎡
∂ 2 f ( x, y )
2 ∂ f ( x, y )
2 ∂ f ( x, y ) ⎤
+ (Δy )
+ ⎢(Δx)
+ 2ΔxΔy
⎥
2! ⎣
∂x∂y
∂y 2 ⎦
∂x 2
3
3
∂ 3 f ( x, y )
1⎡
3 ∂ f ( x, y )
2
2 ∂ f ( x, y )
+ ⎢(Δx)
+ 3(Δx) Δy
+ 3Δx(Δy )
3! ⎣
∂x 3
∂x 2 ∂y
∂x∂y 2

∂ 3 f ( x, y ) ⎤
+ (Δy )
⎥ +L
∂y 3 ⎦
3

2

1⎛
1⎛
∂
∂ ⎞
∂
∂ ⎞
= f ( x, y ) + ⎜ Δx + Δy ⎟ f ( x, y ) + ⎜ Δx + Δy ⎟ f ( x, y )
1! ⎜ ∂x
2 ! ⎜ ∂x
∂y ⎟
∂y ⎟
⎠
⎝
⎠
⎝
3

1⎛
∂
∂ ⎞
+ ⎜ Δx + Δy ⎟ f ( x , y ) + L
3! ⎜ ∂x
∂y ⎟
⎝
⎠

(2-10)

For example, function with 3 variables:

1 w 1− w
= +
T T1
T2

⇒

T=

1
w 1− w
+
T1
T2

(2-11)

Where, w, T1 and T2 are three varibles.
The above equation can be mathematically expanded around (w, T1, T2) = (0.5, 273.15,
273.15):

T = T2 − T2 ⋅ w + T1 ⋅ w + 0.003661 ⋅ T1 ⋅ T2 − 0.0018305 ⋅ T22
31

− 0.0018305 ⋅ T12 − 1.63754 × 10 −6 ⋅ T2 ⋅ T12 − 1.63754 × 10 −6 ⋅ T1 ⋅ T22
+ 1.63754 × 10 −6 ⋅ T23 + 1.63754 × 10 −6 ⋅ T13 + ...
When w=0.5, Figure 23 are the response surfaces.

32

Figure 23 A: Function vs. polynomial (square terms); B: Function vs. Function vs.
polynomial (cubic terms)

33

When the function is expanded to higher order terms, the results of the
polynomial are close to the result from the function. So, there are two steps: regression
analysis of experimental data with the polynomial to get the value of those coefficients,
and then use the polynomial to predict and optimize the process to obtain optimum
product properties.
Many of the properties of the starch foams are a function of feed rates of starch,
talc and water during the extrusion process. Since there are significant interactions
between these factors, changes in one factor while holding the other two constant, does
not yield the same response as repeating the same set of experiments and holding the
other two factors at a different level. This is due to the effect of one variable depending
on the level of the others involved in the process. Thus, a better approach is to use a BoxBehnken statistical design experiment (Figure 24) [57] where the levels of all the
variables are changed simultaneously. This approach is also advantageous since it
requires fewer experiments, resulting in lower reagent consumption and considerably less
laboratory work. It also yields mathematical models with information related to the
statistical significance of each factor as well as information on the interaction between
the factors. The mathematical models are obtained by selecting appropriate polynomial
equations that describe the experimental data. Various statistical analyses can then be
used to determine how well the selected model satisfies the data.

34

Figure 24 Experimental points for Box-Behnken design
If there were 4 variables instead of 3 variables (Figure 24), two variables should
stay in the “0” line and the other two variables should in the “-1” or “+1” line, therefore,
the total experimental points except center should be 6×2+4×2+2×2=24, and there are
usually 5 points in the design space center, so there are 29 experiments totally. Thus, this
experiment amount is much less than 81 (34=81, 4 variables at three different levels).

Figure 25 DOE for 4 variables

35

Chapter 3 Foam extrusion results and discussion
3.1 Screw configuration effect

Figure 26 Specific mechanical energy (SME) comparison for two screw configuration
Three different screw configurations (as shown in Table 5) were used in starch
foam sheet commercial scale extrusions. The first screw configuration has the greatest
shear, the second has the lowest shear and the shear generated by the third screw
configuration was at an intermediate level (Table 5). Among the three screw
configurations, the 2nd screw configuration design is failure; the tube foam has
ripples/bumps on the top and bottom. Also, the foam expanded too much, even when
adding talc which normally reduces bubble size.
Table 5 Configuration #1 has more mixing effect than Configuration #3, and thus
needs more mechanical energy. Configuration #3 is currently used for the normal
production of starch foam sheets (commercially sold as Green Cell Foam™).

36

Screw configuration plays an important role in starch foam extrusion. Unlike
plastic extrusion, polymer pellets can be easily fed into and conveyed to the next section.
The starch must be in granulated form instead of powder form in order for the screw
flight to pick them up while rotating. To avoid water/moisture adhering to the starch and
causing a bridging problem, the water feed port should be separated from the starch feed
port and also, the feed zone should be kept cool to avoid vaporization of the water. The
feed zone includes the largest single pitch screw elements in order to convey rapidly. The
pitch of the screw is then reduced in order to force the materials downstream.
The first kneading zone has thin kneading elements with a small stagger angle in
order to rapidly convey the materials to the second kneading zone with light
kneading/mixing. The second kneading zone has less kneading effect and less degree of
fill than the third kneading zone since there are a few 90 degree of kneading disks.
The screw configuration and die design are the two important factors that affect
the process. The screw configuration builds the pressure, and the die should have the
ability to hold the pressure (force balance). Improper screw configuration will cause
problems such as surging, braiding (Figure 27), bumping, or high torque.

37

Figure 27 Regular starch foam sheets (left) vs. sheets with braiding problem (right two)
3.2 Starch feed rate effect
The starch feed rate was varied for cylindrical foam extrusion and
characterizations were done on the products. In order to increase the starch feed rate to
increase the production efficiency, cylindrical foam extrusions with six holes (Figure 28)
were also investigated. Compared with petroleum based foam, the bubble cells have a
wider distribution (Figure 29).

Figure 28 Four and six holes strand die

38

Figure 29 Cells in the starch foam
When the starch feed rate decreased from 460 lbs/hr to 285 lbs/hr, the diameter of
the foam decreased from 20 to 16 mm statistically approved by t test. (Table 8)
Exp.
1
Starch (lbs/hr)
Polyvinyl alcohol
(lbs/hr)
Bentonite with color
(lbs/hr)

Exp.
2

Exp.
3

Exp.
#1
μ0
20
(diameter) mm

Exp.
#2
16
mm

460

285

560

36

36

43

t0

-0.9

1.86

117

80

86

t α/2, ν
|t0|<tα/2,
ν?

2.26

2.26

Talc (lbs/hr)
3.6
1.4
0
Yes
Yes
μ=μ0 μ=μ0
Water (lbs/hr)
61
46
65
Vegetable oil (lbs/hr)
6
6
6
Screw speed (rpm)
286
270
345
Torque (%)
82%
56%
90%
Cutter speed (rpm)
1667 1140
1375
# of holes in the die
face
4
4
6
Table 8 Cylindrical foam extrusion conditions and diameter of the foams

39

As the starch feed rate decreased the diameter of the cylindrical foam and the cell
diameter decreased accordingly (Figure 30 and Table 9). The density and compressibility
increased while resiliency of the foam remained constant.

40

Figure 30 Cell size distribution for cylindrical foams (A: noodle extruded at 460 lbs/hr, B:
noodle extruded at 285 lbs/hr)

41

Exp. 1 [Q] CI
Exp. 2 [0.65Q] CI
P value
Diameter (mm)
↓
19.89
0.29
16.19
0.23 9.52E-15
3
Density (kg/m )
↑
21.08
0.64
31.69
0.73 2.24E-15
Compressibility(Pa) ↑ 1.74E+05 1.26E+04
2.86E+05 1.23E+04 2.25E-11
Resiliency
↓
60.10%
2.30%
57.30%
1.30%
0.03
Cell size (μm)
↑
605.0
55.5
499.9
45.7
0.006
Table 9 cylindrical foam characterizations
As the number of holes increased from 4 to 6, cylindrical foams with inconsistent
lengths were extruded. In addition, the cylindrical foam’s diameters were also not
consistent, which somewhat depended on the position of the holes (Figure 31, P value =
0.021). This is due to the holes location at different levels and therefore different
pressures. The flow rate thus varies with materials flowing more rapidly from the bottom
holes than the top. This can be corrected by adjusting the land lengths of the die.

Figure 31 Characterization of cylindrical foam prepared from a die with six holes

42

3.3 Effect of talc content
The effect of different talc levels on the starch foam with PVOH and PHAE as
polymer modifiers was investigated.
The talc acts as a nucleating agent which creates new bubble sites during the
foaming process. Increasing the talc content (0.5wt%, 1.3wt%, and 2.7wt%) will decrease
the cell size of the foam (Figure 32 and Figure 33).

Figure 32 Cell size distributions for starch foam with PVOH, different talc levels: A:
0.5wt%, B: 1.3wt%, C: 2.7wt%

43

Figure 32 (cont’d)

44

Figure 33 Cell size distributions for starch foam with PHAE, different talc levels: A:
0.5wt%, B: 1.3wt%, C: 2.7wt%

45

Figure 33 (cont’d)

The effect of the talc content with PHE and PVB as polymer modifiers was also
investigated using design of experiments methods (section 3.4.2 and 3.4.3).
3.4 Different modifiers
3.4.1 Aqueous PVOH
PVOH solution was studied as a polymer modifier in the starch foam extrusion.
The PVOH used for preparing the solution was from Mowiol and the grade was 28-99.
This indicates that 4% solution of PVOH in water has 28 cps viscosity. Due to the
internal hydroxyl bonding it is not easily dissolved in water. The dissolving technique is
to suspend the PVOH into cold water and with constant stirring to avoid agglomeration or
settling. Low pressure steam is then introduced to heat up the suspension rapidly. When
the solution becomes transparent, the steam is discontinued and the solution cooled with
stirring to less than 50°C. It can then be diluted to desired concentrations.
46

The foam extrusion with PVOH solution was done using lab extruder Century
ZSK-30. The PVOH water and starch ratio was kept at 0.06:1.

Figure 34 starch foam densities vs. PVOH solution concentration

47

Figure 35 Starch foam resiliencies vs. PVOH solution concentration
Different PVOH solution concentrations have little effect on the density and
resiliency of the cylindrical foam (Figure 34 and Figure 35), this is probably due to the
low addition of PVOH, which has little effect on the bulk foam properties.
3.4.2 Box-Behnken DOE with PHE and annular die
It is apparent that many biodegradable or non-biodegradable additives have been
used in attempts to control the structure of the foam and improve the stability toward
humidity. Only limited success has been achieved in these studies to date.
Table 10 is the operational conditions and product characterization for starch
foam with PHE. The design space picked was: 18.1 kg/hr < PHE < 54.4 kg/hr; 0.23 kg/hr
< Talc < 2.27 kg/hr; 20.4 kg/hr < Water < 29.5 kg/hr [58].

48

PHE
(kg/hr)

Talc
(kg/hr)

36.3
54.4
36.3
54.4
36.3
18.1
18.1
36.3
36.3
18.1
18.1
36.3
36.3
54.4
54.4

1.25
1.25
1.25
1.25
0.23
0.23
2.27
1.25
2.27
1.25
1.25
2.27
0.23
2.27
0.23

Density
3
[kg/m ]

Water
(kg/hr)

Thickness
[mm]

Width
[mm]

Cell
diameter
24.9
20.2
27.0
1.14
575
29.5
17.7
28.4
1.19
580
24.9
19.5
26.7
1.30
580
20.4
25.7
19.4
0.70
520
29.5
11.2
31.9
3.55
730
24.9
16.3
30.8
3.44
24.9
18.9
25.5
1.15
600
24.9
18.1
24.2
1.18
640
20.4
21.9
22.4
0.81
600
20.4
20.0
24.3
1.24
620
29.5
18.6
25.3
1.46
650
29.5
21.7
21.0
0.88
620
20.4
17.5
23.2
1.41
730
24.9
17.0
26.2
1.37
710
24.9
21.8
20.0
0.62
590
Table 10 Box-Behnken DOE with PHE and annular die

Water
penetration
time [s]
77.4
73.4
187.8
63.0
14.4
11.6
49.8
59.8
10.2
13.2
59.4
66.8
195.0
187.8

A plot of the experimental values as a function of the calculated values (Figure 36)
should yield a perfect fit when the data are arranged on a diagonal line with an intercept
at the origin. Any deviation from this ideal line as well as any outliers can be determined
from the scatter of the data points around the diagonal line. It was found empirically that
quadratic polynomials with a few higher terms provided adequate fit to the data:
2

2

2

2

Y=A0+A1R+A2T+A3W+A4R +A5T +A6W +A7R*T+A8R*W+A9T*W+A10T*W
where R, T and W denote the feed rates of PHE resin, talc and water (in Kg/hr),
respectively and the value of the polynomials coefficients are listed in Table 11.

49

(3-1)

Constant

Density
[Kg/m3]

Cell
diameter
[mm]

A0

-128.5

-0.16343

A1

0.5000

-0.08672

A2

125.4

-0.4314

A3

11.57

0.2407

A4

0.00252

-0.00016

A5

-1.533

0.4578

A6

-0.2344

-0.00129

A7

-0.09994

0.04106

A8

0.02005

0.00082

A9

-9.832

-0.1118

A10

0.2037
Table 11 Polynomials used to evaluate the foam density and cell diameter

50

Figure 36 Goodness of fit of the calculated polynomials to the experimental values.
A: Foam Density, B: Average Foam Cell Size.
The foam density depends on the combination of the feed as shown in Figure 37.
The lowest foam density is obtained at the lowest resin feed rate and lowest talc feed rate
whether low water concentrations (Figure 37A) or high water concentrations are used
(Figure 37B). It is apparent that the density remains almost constant as the resin feed rate

51

is increased while keeping low talc concentrations in the feed. However, maximum
density is obtained at the lowest resin feed rate and maximum talc feed rate. This density
maximum is more pronounced when the water concentration in the feed is high (Figure
37B). The data further indicate that at low water feed rates only relatively small changes
in the density are observed compared to the foams prepared with high water content
where either increasing the resin feed rate or decreasing the talc feed rate lead to higher
density foams. It is also observed that the changes in the talc concentrations have a more
significant effect on the density than either the resin or the water feed rates.

52

Figure 37 Effects of resin and talc feed rates on the foam density at low (A) and high (B)
water feed rates.
53

The effect of water concentration in the feed rate is more complex as shown in
Figure 38. It is apparent that at the low talc feed rate (Figure 38A) the foam density is
inversely proportional to the concentration of the resin and this effect is more pronounced
at high water feed rates. In fact, the resin content in the feed has almost no effect when
the water feed rate exceeds about 22 Kg/hr. The situation is somewhat different when the
talc content is increased. Under these conditions, the foam density appears to decrease
through a shallow minimum and then increases again as the water content in the feed is
increased (Figure 38B). Under these conditions the resin feed rate has only a minor effect
on the foam density leading to slightly denser foams when the resin content in the feed is
increased.

54

Figure 38 Effects of water and resin feed rates on the foam density at low (A) and high (B)
talc feed rates.

55

The effect of water on the density is easy to understand since the water acts as a
blowing agent. It is intimately mixed with the extrudate under high pressure and
temperature and then released as steam immediately beyond the die upon exit to
atmospheric pressure. Thus, low water concentrations inherently lead to low expansion
3

and high foam density. Similarly, the high density of talc (2.7-2.8 g/cm ) greatly impacts
the overall density of the foam which is directly proportional to its concentration in the
feed. It is important to note that the situation is more complex since it is not sufficient to
have only high expansion at the die exit in order to obtain a low density foam. Low
density foams can be obtained only if the expanded foams are stable and do not shrink
beyond the die. As a result of this shrinkage, a high foam density is obtained at high
concentrations of water in the feed. At low water feed rates, insufficient water is available
to expand the foam resulting in high density foams. However, at intermediate
concentrations sufficient water as a blowing agent is available and no shrinkage occurs
resulting in a shallow minimum and low density foams (Figure 38B).

56

Figure 39 Effects of water and talc feed rates on the average cell diameter at low (A) and
high (B) resin feed rates.
57

The dimensional stability and the overall structure of the foams as a function of
the feed are shown in Figure 39 and Figure 40. The average cell diameter of the foam at
low resin and high talc feed rates is very small (about 1 mm) and essentially independent
of the talc content in the feed (Figure 39A). However, decreasing the talc content in the
feed drastically increases the average cell diameter up to about 5 mm. This increase in the
average cell size is more pronounced at high water concentration in the feed compared to
a similar feed having low water content where the average cell diameter increases to only
3 mm. It is also observed that at these low resin feed rates the talc has a more pronounced
effect on the foam cell diameter than the water content in the feed. Increasing the resin
content in the feed leads to a similar response plot (Figure 39B) but the effect of the talc
and the water are less obvious and the overall surface is flatter with the majority of the
cells being closed cells.

58

Figure 40 Effects of talc and resin feed rates on the average cell diameter at low (A) and
high (B) water feed rates.
59

The changes in the foam structure as a function of the talc and resin contents in
the feed are best observed in Figure 40. At low water content (Figure 40A) the largest
average cell diameter is observed at low resin and low talc concentrations. Increasing
either of these variables leads to finer morphology and foams with smaller cell sizes.
Moreover, the same phenomenon is observed at high water content (Figure 40B) but the
effect of these variables on the structure of the foams is more noticeable. Under these
conditions the average cell diameter decreases from about 5 mm to about 0.2 mm by
increasing the talc content in the feed to 2.5 Kg/hr or to about 2 mm by increasing the
resin content in the feed to 55 Kg/hr. The finest cell structure is obtained when the water,
resin and talc feed rates are at their maximums.
The structure of the foam as a function of these variables is related to expansion
of the extrudate as it exits the die. At this point water vapors can either diffuse into an
existing cell (bubble) to further expand it leading to a larger cell diameter or create a new
cell (bubble) by overcoming the surface tension barrier. Here the talc acts as a nucleating
agent, hence increasing its concentration will decrease the surface tension barrier which
will then lead to the formation of new small bubbles. In the absence of talc or at low talc
concentrations, water vapor tends to diffuse into an existing cell resulting in the
expansion of an existing bubble and produce foam with a larger average cell size.
Furthermore, strain hardening is a necessary requirement for the foaming process. In the
early stages of the foaming process, the elongation viscosity of the mixture must be low
in order to expand the bubbles. However, in the later stages of the foaming process, high
elongation viscosity is preferred in order to stabilize the bubbles. The actual composition
of the feed has a strain hardening effect on the foam processing since during the foaming

60

water evaporates and the water vapors gradually diffuse into the bubble. Consequently,
less water is used as a plasticizer resulting in a higher resin viscosity.
3.4.3 Box-Behnken DOE with PVB
In the absence of PVB, the foam is very hydrophilic and dissolves rapidly in
water. At high PVB feed rates miscibility becomes an issue and results in “chunky”
extrudates. Higher talc loadings produce foam with too high a density. Low water feed
rates causes high torque on the extruder with possible blockage. High water feed rates
result in a large bubble size in the foam or a low drop in die pressure resulting in no
blowing at all. Smaller pin-hole will result in lower density foam and larger pin hole
diameter in dense or no foam. In consideration of these restrictions, the following design
parameters were selected.
Lowest value Highest value
PVB feed rate
0.75 Kg/hr
2.5 Kg/hr
Talc feed rate
0
0.031Kg/hr
Water feed rate
0.5 Kg/hr
0.8 Kg/hr
Pin hole diameter
10 mm
20 mm
Table 12 Design space using polyvinyl butyral
A Box-Behnken statistical design experiment (DOE) was used and the detailed
experimental conditions based on the design parameters are listed in Table 13.

61

Pin hole
Sample PVB
Talc
Water diameter
label # (Kg/hr) (Kg/hr) (Kg/hr) (mm)
1
1.63
0
0.5
15
2
0.75
0
0.65
15
3
1.63 0.0155
0.65
15
4
2.5
0
0.65
15
5
1.63
0
0.8
15
6
1.63 0.0155
0.65
15
7
0.75 0.0155
0.5
15
8
2.5 0.0155
0.5
15
9
1.63 0.0155
0.65
15
10
0.75 0.0155
0.8
15
11
2.5 0.0155
0.8
15
12
1.63 0.0155
0.65
15
13
1.63 0.031
0.5
15
14
0.75 0.031
0.65
15
15
1.63 0.0155
0.65
15
16
2.5 0.031
0.65
15
17
1.63 0.031
0.8
15
18
1.63 0.0155
0.5
20
19
1.63
0
0.65
20
20
0.75 0.0155
0.65
20
21
2.5 0.0155
0.65
20
22
1.63 0.031
0.65
20
23
1.63 0.0155
0.8
20
24
1.63 0.0155
0.5
10
25
1.63
0
0.65
10
26
0.75 0.0155
0.65
10
27
2.5 0.0155
0.65
10
28
1.63 0.031
0.65
10
29
1.63 0.0155
0.8
10
Table 13 Box-Behnken DOE for starch foam with PVB
There are four independent variables (PVB [Kg/hr], Talc [Kg/hr], Water [Kg/hr]
and Pin hole diameter [mm]). The center of the design space was tested for 5 times, so
the total experiments were 2*(6+4+2)+5 = 29.

62

Noodle
Sample diameter Density Wear
label # (mm)
(Kg/m3) (mm)
1
43
65.9
4.35
2
49.2
40.1 25.63
3
40.9
71.9
3.68
4
45.4
54.7
5.53
5
46.2
44.4 17.86
6
41.9
65.8
6.09
7
36.5
82.2 25.05
8
36.4
108.3
2.07
9
41.5
68.9
4.94
10
45.6
48.4 23.98
11
39.7
65.5
3.95
12
41
65.9
5.12
13
33.6
103.4
2.84
14
37.8
70.4 20.40
15
38.4
72.2 13.82
16
36.2
91.4
3.13
17
38
68.7
4.88
18
35.9
150.6
4.01
19
39.8
129.2
1.59
20
40.6
85.6 14.04
21
39.9
116
4.21
22
38.8
111.7
2.80
23
40.6
113.1
9.10
24
31.8
105.8
3.01
25
55.7
29.2 14.33
26
41
43.5 25.88
27
34.5
75.3
6.74
28
28.8
108.7
3.17
29
36.9
54.1 18.21
Table 14 Box-Behnken DOE experimental results for starch foam with PVB
Box-Behnken DOE results are listed in Table 14. The mathematical models to
regress the Box-Behnken DOE responses are obtained by selecting appropriate
polynomials. The criteria are to pick polynomials with less higher order terms whose
calculated values are close to the experimental value. Thus, a plot of the experimental
values vs. calculated values should be in a diagonal line with origin intercept if the
63

regressions yield a perfect fit. The scatter plots will show the goodness of fit. It was
found that quadratic polynomials provided adequate fit to the data empirically:

Y = β 0 + β1 X 1 + β 2 X 2 + β 3 X 3 + β 4 X 4 + β1 2 X 1 X 2 + β13 X 1 X 3
+ + β 23 X 2 X 3 + β 24 X 2 X 4 + β 34 X 3 X 4 + β 11 X 12 β 1 4 X 1 X 4
2
2
+ β 22 X 2 + β 33 X 32 + β 44 X 4

(3-2)

where X1, X2 and X3 and X4 denote the feed rates of PVB resin, talc and water (in Kg/hr),
respectively and the value of the polynomials coefficients are listed in Table 15

β0

Noodle
diameter
(mm)
0.3966

β1

Density
3
(Kg/m )
359.8

Wear (mm)
49.51

-1.446

50.00

-50.01

β2

-1958

6658

-9.706

β3

150.4

-717.1

7.472

β4

0.7670

-22.03

0.4260

β12

39.57

118.0

51.84

β13

-11.13

-17.20

5.734

β14

0.3333

-0.07566

0.5331

β23

140.1

-1414

-1235

β24

83.34

-312.8

39.90

β34

-0.1488

4.733

-3.373

β11

0.3962

-8.024

8.331

β22

6328

-6897

-1389

β33

-88.98

440.6

56.38

β44

-0.07954

0.9584
-0.01043
Table 15 Polynomials used to evaluate the foam diameter, density and wear

64

(A) R2=0.9312

(B) R2=0.9643

65

(C) R2=0.9133

Figure 41 Goodness of fit of the calculated polynomials to the experimental values. A:
Foam diameter B: Density C: Wear
The eight 3D plots are based on the Table 16. There are four variables in the
polynomials; the 3D plots are drawn with setting two variables constant (low and high
levels) and the other two variables as X-axle and Y-axle respectively.
Response
Noodle diameter
(mm)
[4 figures]
Density (Kg/m3)
[4 figures]

Talc (Kg/hr)
Low/High

Pin hole
diameter (mm)
Small/Big

Water (Kg/hr)
X-axle

PVB (Kg/hr)
Y-axle

X-axle

Small/Big

Low/High

Y-axle

Table 16 Table for 3D plots (noodle diameter and density as responses)

66

Figure 42 is the noodle diameter response dependent on the feed rates. When
using small pin holes, increasing talc feed rate greatly decreases the noodle diameter
(Figure 42A vs. Figure 42B). The effect of talc feed rate is much less pronounced when a
large pin hole is used; the noodle diameter is maintained relatively constant when the talc
feed rate is varied (Figure 42C vs. Figure 42D). At low talc feed rate, increasing pin hole
diameter decreases noodle diameter (Figure 42A vs. Figure 42C), This may be due to a
drop in expansion ratio. However, at high talc feed rate, increasing pin hole diameter
increases noodle diameter (Figure 42B vs. Figure 42D). Figure 42A shows that increasing
water feed rate increases noodle diameter at low PVB feed rate, however, water feed rate
does not affect the noodle diameter when PVB feed rate is high; the effect of PVB feed
rate is pronounced when water feed rate is high.

67

(A) Low talc content and small pin hole diameter

(B) High talc content and small pin hole diameter
Figure 42 3D plots, noodle diameter as response

68

Figure 42 (cont’d)

(C) Low talc content and big pin hole diameter

(D) High talc content and big hole diameter

69

The foam density as a function of the feed rate and pin hole diameter is shown in
Figure 43. Generally, increasing the water feed rate decreases the foam density (Figure
43A vs. Figure 43B and Figure 43C vs. Figure 43D). At low talc content, increasing the
pin hole diameter significantly increases the foam density (Figure 43B vs. Figure 43D),
this is caused by the pressure drop’s significant decreased with increasing in pin hole
diameter. However, at high talc content, the foam density is not significantly changed by
the pin hole diameter (Figure 43A vs. Figure 43C). When using small pin holes,
increasing the talc feed rate greatly increases the foam density (Figure 43A and Figure
43B). However, when using a large pin hole, the talc effect is much less pronounced
(Figure 43C vs. Figure 43A and Figure 43D vs. Figure 43B). It is observed that a
consistently lower density is obtained at lowest talc and PVB feed rates, smallest pin hole,
and highest water feed rate.

70

(A) Low water content and small pin hole diameter

(B) High water content and small pin hole diameter
Figure 43 3D plots, density as response

71

Figure 43 (cont’d)

(C) Low water content and big pin hole diameter

(D) High water content and big hole diameter

72

The changes in the hydrophobicity (wear in mm) as a function of the PVB feed
rate and the water feed rate of the foam is shown in Figure 44. These data were obtained
for the foams that were extruded with no talc using the smallest pinhole die. It is apparent
from the data that highest wear (e.g. maximum hydrophilic characteristics) occurs when
the starch foam is extruded at the lowest PVB feed rate and highest water feed rate (wear
= 41 mm). Increasing the feed rate of the PVB or decreasing the feed rate of water
significantly increases the hydrophobicity of the foams. However, changing the PVB feed
rate appears to be more significant than changes in the water feed rate. Lowest wear (e.g.
high hydrophobicity) under these conditions is obtained at the highest PVB feed rate and
at the lowest water feed rate.

Figure 44 Changes in the wear as a function of PVB and water feed rates with no talc and
extruding through a small pinhole die (diameter = 10 mm)
73

The effects of PVB and water feed rates on the wear when talc is coextruded (at
0.0155 Kg/hr) using a larger diameter die (20 mm) is shown in Figure 45. Under these
conditions the most hydrophobic foam is still obtained at the highest PVB feed rate and
lowest water feed rate. However, the changes in the wear are less pronounced than before
when no talc was used and the foam was extruded through a small diameter die.
Apparently, adding talc contributes to the hydrophobicity of the foam simply because talc
is inherently hydrophobic. It is also important to note that under these conditions the
water feed rate has very little effect on the wear and the major changes are directly
related to the extrusion rate of the PVB resin.

Figure 45 Changes in the wear as a function of PVB and water feed rates with talc
coextruded at 0.0155 Kg/hr through a large pinhole die (diameter = 20 mm)

74

The effect of the talc on the wear is shown in more detail in Figure 46. Foams
with best hydrophobic characters (lowest wear) are obtained when the starch is coextruded with highest levels of PVB and talc as both of these additives are inherently
hydrophobic and contribute to the wear as long as they are well dispersed in the foam.
Conversely, the most hydrophilic foam is obtained when no talc and the lowest PVB feed
rate are used (maximum wear in Figure 46). It is further apparent that the PVB has a
greater effect on the wear than the talc. Thus, the wear decreases from 42 mm to 23 mm
when the talc feed rate increases to 0.03 Kg/hr at the lowest PVB feed rate but it
decreased from 42 mm to 16 mm when the PVB feed rate increases from 0.75 Kg/hr to
2.5 Kg/hr.

Figure 46 Changes in the wear as a function of PVB and talc feed rates with water
coextruded at 0.80 Kg/hr through a small pinhole die (diameter 10 mm)
75

The difference in the impact of talc and PVB feed rates on the wear is best
observed under slightly different extrusion conditions. When the water feed rate is
reduced to 0.5 Kg/hr and the die diameter is increased to 16 mm, the talc has negligible
effect on the hydrophobicity and only the PVB impacts the wear (Figure 47). Thus, under
these conditions it is possible to obtain hydrophobic starch foams even when no talc is
used.

Figure 47 Changes in the wear as a function of PVB and talc feed rates with water
coextruded at 0.50 Kg/hr extruding through a pinhole die with a diameter = 16 mm.
Unlike changing the composition of the foam, it is conceivable that the pinhole
diameter should have no effect on the wear. However, this is not the case as evident from
Figure 48 where the wear is inversely proportional to the die diameter. Furthermore, this
trend is observed at all PVB concentrations but the lowest wear (maximum
76

hydrophobicity) is observed at the highest PVB feed rate and largest die diameter. The
die diameter affects the density and the extent of expansion of the foam which, in turn are
related to the moisture sensitivity.

Figure 48 Changes in the wear as a function of PVB and pinhole diameter with water
coextruded at 0.80 Kg/hr and no talc
3.4.4 Starch foam homogeneity with PVB
PVB is commercially produced by reacting polyvinyl alcohol with n-butyl
aldehyde. The hydrophobicity of the polymer can be adjusted by controlling the amount
of n-butyl aldehyde used to react with the alcohols. Thus, a high degree of reaction leads
to few hydroxyl groups in the polymer and a hydrophobic resin. In preliminary
experiments it was noted that the compatibility of PVB with the starch foam depends on
77

the grade of the PVB. Specifically, increasing the molecular weight of the PVB and
decreasing the concentration of the hydroxyl groups decreased the miscibility of the
polymer in the starch. Based on these preliminary experiments it was decided to use
grade Mowital B 60HH, as it should provide good miscibility and a high degree of
hydrophobicity. However, in order to ensure compatibility of this grade with minimum
gross phase separation that could lead to heterogeneous foam morphology, the
homogeneity of foams was evaluated. This was done by taking small samples from 9
different locations of the foam and determining the amount of PVB by TGA. Since the
decomposition temperature of the unmodified starch foam occurs below 300 °C and PVB
decomposes above 350 °C it is possible to establish the relative concentration of PVB in
the starch foam. No apparent difference in the concentration of PVB throughout the
sample was observed thus indicating that PVB was evenly distributed within the starch
foam with no apparent gross phase separation (Figure 49).

Figure 49 TGA analysis on starch foam with PVB
78

3.5 Scale down starch foam extrusion from Wenger TX-80 to Century ZSK-30
The Wenger TX-80 extruder has a large throughput ~1000 lbs/hr, and is not
suitable for R&D. In order to mimic the process, the feed zone of the lab scale Century
ZSK-30 extruder was moved to the middle zone to reduce the L/D ratio (
Figure 15) and a modified screw configuration (Figure 16) was chosen in order to mimic
Wenger TX-80 screw configuration more closely. The following equation was used to
correlate the feed rates between the different extruders.

QL
D
= ( L )3
QS
DS

(3-3)

Century ZSK-30

Wenger TX-80

Feed rates
Starch (Kg/hr)
21.5
408
PVB
3.1
59
Talc
0.18
3.4
Water
0.85
41.3
Processing parameters
Extruder RPM
400
400
Extruder Torque
60%
90%
Temperature profile (°C)
30, 90, 130, 140
55, 75, 110, 140
5
5
Specific Mechanical Energy (J/kg)
5.1×10
7.7×10
Foam sheet properties
Foam width (mm)
203
610
Foam thickness (mm)
8.0
14.8
3
33.0
27.7
Foam density (kg/m )
Table 17 Experimental conditions and results for starch foams prepared by the Century
ZSK-30 and Wenger TX-80 extruders
It has been tested that with starch powder and starch pellets a 50:50 ratio is the
highest amount of powder that can be fed to the extruder without feed problems due to
bridging in the feed hopper.

79

The possibility of using oxidized starch to replace the hydroxypropylated high
amylose starch was evaluated using the small scale extruder.
Experimental conditions:
Premixed starches, PVB and talc, total feed rate was kept at 23.2 kg/hr
Water feed rate was kept at 1.3 kg/hr (5.6wt% based on starch).
Sample #
0
1
2

chilsonated
Oxidized
PVB
starch from
starch
Based on
KTM
starch
100%
0%
15wt%
80%
20%
15wt%
60%
40%
15wt%
Table 18 Formulation for lab starch foam tube

Talc
Based on
starch
1wt%
1wt%
1wt%

Screw speed: 400 rpm. Die used: annular die
Temperature profile: 30, 90, 120, 140 °C
Torque: 48%, die pressure: 450 psi, slight decrease in torque and die pressure was
observed when oxidized starch was added.
Characterization:
Sample #
0
1
2

Thickness
Density
(mm), eight different locations (Kg/m3)
7.4 ± 0.7
32.8
7.6 ± 0.5
32.6
7.6 ± 0.6
32.1
Table 19 Lab starch foam tube characterization

When the foam tube was extruded, the relative humidity in the room was 17%,
and all the foam tubes were observed to be brittle and the foam tube with oxidized starch
disintegrated. Preconditioning of the foam tubes at 52% RH, sample #0 and sample #1
were flexible, sample #2 was brittle but not cracked into pieces. The conclusion from
these experiments is starch foam made from oxidized starch has decreased flexibility.

80

Chapter 4 Performance characterizations of starch foam sheets
4.1 Moisture adsorption and humidity sensitivity
Due to the hydroxyl group in the starch structure, moisture can be absorbed from
the atmosphere due to hydrogen bonding. Figure 50 shows the moisture adsorption vs.
time on the hydroxypropyl high amylose starch powder. It is apparent from the shape of
the curves that the water absorption was initially fast but the rate of absorption gradually
decreased with time and equilibrium was established after about 70 hours.

Figure 50 Hydroxypropyl high amylose starch powder moisture content vs. time at 23°C
It is apparent from the absorption data that the absorbed water was directly
proportional to the RH as described by the PELEG model:

81

m(t ) = m0 +

t
k1 + k 2 t

(4-1)

Where m0 is weight of the initial adsorbed moisture (at t=0), t is time, m(t) is the
weight of adsorbed moisture at t. k1 and k2 are two constants. In this model, k1 and k2
indicate the absorption rate and the absorption capacity, respectively. Thus, high k1
values indicate low absorption rate and high k2 values indicate low saturated moisture
contents.
Raw starch has a very high k1 value and a low k2 value whereas the starch foam
sheet extruded with no polymer modifiers has significantly lower k1 and k2 values (Table
20). This significant increase in the adsorption rate and the saturated moisture content of
the starch foam are a direct consequence of the degradation of the hierarchal structure of
native starch granules under high temperature, high pressure and the presence of water
during the extrusion. The data in Table 21 further indicate that adding PVOH does not
improve the resistance of the foams to moisture. However, noticeable improvements are
observed when PHAE and PHE are used as additives in the process (Table 21). Moisture
adsorption of the starch foam products demonstrates a type II adsorption isotherm (this
type of adsorption isotherm shows an indefinite multi-layer formation after completion of
the monolayer).
The starch foam retains moisture sensitivity (Figure 51). However the foam
absorbed moisture faster than the starch powder. Using PHAE, PHE and PVB as polymer
modifiers will have lower moisture content under high RH than using PVOH as polymer
modifiers (Figure 53-Figure 55). Moisture adsorption on the starch foam products
82

belongs to a type II adsorption isotherm, Surface area (accessible to moisture) got
2

according to BET equation is 130m /g.
RH
k1
k2

23%
65.82
11.83

33%
43%
52%
59%
70%
52.48
49.32
39.91
36.95
32.95
9.47
7.62
6.39
5.93
5.06
Table 20 k1 and k2 values for starch powder

75%
29.13
4.58

85%
28.54
3.75

k1
k2
Raw starch
28.5 3.7
Starch foam without additive
9.8 3.9
Starch foam with PVOH
9.1 4.0
Starch foam with PHAE
10.3 4.3
Starch foam with PHE
12.8 4.5
Table 21 k1 and k2 values for PELEG model (85% RH)

Figure 51 Starch foam (no polymer modifier) moisture content vs. time at 23°C

83

Figure 52 Starch foam (PVOH as polymer modifier) moisture content vs. time at 23°C

Figure 53 Starch foam (PHAE as polymer modifier) moisture content vs. Time at 23°C
84

Figure 54 Starch foam (PHE as polymer modifier) moisture content vs. time at 23°C

Figure 55 Starch foam with PVB as polymer modifier moisture content vs. time at 23°C
85

Figure 56 Moisture content at equilibrium under different RH (starch foam with PVOH
as polymer modifier)
The use of PHAE, PHE and PVB as additives to minimize the water sensitivity
was found to be extremely effective. These hydrophobic ethers significantly decreased
the water penetration time from 4 seconds to 85 seconds when it was added during
extrusion (Figure 57). For comparison, PVOH did not improve the hydrophobic nature
of the starch foam sheets and the water penetration time did not change significantly
compared to the control sheet foams that were extruded with no additives. Although all
the additives that were used contain hydroxyl groups to provide compatibility with the
starch, the hydrophobic backbone of the hydroxyl ethers was much more effective as it
provided better protection against water and improved the humidity resistance of the
foam sheets. This protection against high humidity is clearly observed in Figure 58 where

86

wet foams that were extruded with PVOH disintegrated after five hours while similar
foams that were extruded with PHE remained intact.

(A)

Figure 57 Water penetration time for foam sheets (A) foam with different additives (B)
foam with different additive content
87

Figure 58 Starch foam sheets extruded after 5 hours in contact with water. (A) PVOH (B)
PHE.
Additive
type
NONE
PVOH
PHAE
PHE

23%
RH
-

33%
RH
-

43%
RH
+
+
+

52%
RH
+
+
+
+

59%
RH
+
+
+
+

70%
RH
+
+
+
+

75%
RH
+
+
+
+

85%
RH
shrink
shrink
shrink
+
slightly
PVB
+
+
+
+
+
+
shrink
Table 22 flexibility of the foam sheets under different relative humidities at room
temperature
The moisture absorption curves of a typical starch foam sheet that was prepared
with PHE are shown in Figure 54 at room temperature (RT) under different relative
humidity conditions. It is apparent from the shape of the curves that the water absorption
was initially fast but the rate of absorption gradually decreased with time.

88

In this set of experiments the foam sheets retained their shape and flexibility when
the RH was between 52% and 75%. Below this lower limit of the RH the foam sheets
became brittle but regained their flexibility when the RH was increased back above 50%.
It appears that these changes are reversible with no visible loss or degradation in the
mechanical properties. However, when the RH was set at or above 85% noticeable
shrinkage occurred with visual collapse of the foam cells. Although the foam sheets
remained flexible at this high RH, the changes in the cell structures were irreversible and
once the foam cell was destroyed, it did not recover even when the foam was placed back
at lower RH. Some improvements were noted when additives were incorporated during
the extrusion and the foams remained flexible down to 43% RH (Table 22). Only the
addition of PHE suppressed the cell collapse up to 85% RH.
4.2 Thermal insulation
The basic heat transfer mechanisms of foam insulation materials are conduction
and convection. Heat transfer by conduction occurs when energy is transferred through
the foam material itself and is related to the composition of the foam itself. Convective
heat transfer occurs as a result of energy transfer by the gas molecules in (or through) the
foam and is dependent on the type of gas, the density of the foam and the pore structure
(e.g. open or closed cells). In general, heat flow through a typical air space is primarily by
convection; thus, closed cell foams with large number of small cells and thin walls are
most desirable.
The thermal conductivity of foams is a combination of the foam material
conduction and convection of air trapped in and is expressed as 1/R-value (thermal
resistance). We found that the cell structure of the foam is greatly affected by the amount
89

of resin, talc and water in the feed composition during the extrusion process (as described
previously). However, the presence of additives in the feed had only a minor effect on the
cell structure and the overall morphology of the foams. Although the thermal resistance
of starch foam sheets containing PVOH and PHAE were similar to each other and to the
2

thermal resistance of expanded polystyrene foam (R=3.6°F*ft *hr/(Btu*in)), the thermal
resistance of the starch foam sheet containing PHE was 10% better than the other foams
(Figure 59A).
The thermal resistance of the starch foams prepared with PVB is better than that
of similar starch foams prepared with polyhydroxy aminoether (PHAE) as well as
expanded polystyrene foams. As the nucleating agent content increased in the starch
foam sheet, the thermal resistance of starch foam decreased since the density of the foam
increased (Figure 59).

90

Figure 59 Thermal insulation property comparison (A: different additives B: different talc concentration)

91

Figure 59 (cont’d)

92

4.3 Cushion curves
Cushioning properties are generally determined based on the ability to provide the
necessary protection under shock (impact) and vibration. Shock refers to brief but intense
forces, such as dropping, while vibration refers to the less intense but longer duration
forces that occur during transportation. Foams intended for packaging are usually
classified by determining their ability to absorb shock by dynamic cushioning testing
(also commonly known as a cushion drop testing) which indicates how the foam can
absorb impacts efficiently and recover its physical shape quickly.
Typical examples of the cushion curves of the control starch foam sheets and
foam sheets extruded with different polymer modifiers are shown in Figure 60-Figure 64.
It is apparent that the cushion curves shift upward upon multiple drops. The contact
surface/bearing area required between the foam sheets and the product can be calculated
from the force over the maximum static load and was found to be 0.2 psi. The fragility
index, defined as the maximum G force that an item to be packaged can withstand
without sustaining damage, can therefore be determined from these tests. Thus, starch
foam sheets, 50.8 mm thick, extruded with different polymer modifiers provide sufficient
multiple cushion protections for moderately delicate fragile level (65-80Gs) packaging
products. Products in this category include floppy disc drives, networking hardware,
personal computers, medical diagnostic apparatus, etc. These data indicate that the starch
foam sheets can absorb impacts efficiently and recover their physical shape quickly,
making the product ideal for packaging fragile items within a wide range of weights and
sizes.

93

Unlike polyethylene foam (Figure 65), the peak acceleration of starch foam at
high static loading increases with repeated drop times, this significant increase is mainly
due to the corrugation. Therefore, the fifth drop (Figure 67) instead of first drop (Figure
66) cushion curves should be considered to calculate the contact surface for multiple
cushion protection. The highest static loading for 2 inch thickness green cell foam and 30
inch drop is 0.4 psi (from fifth drop cushion curve) instead of 1 psi (from first drop
cushion curve) assuming the target fragile level is 80G.
It is known that, when the environmental relative humidity reaches 85%RH, the
foam sheet shrinks. Figure 68 shows at 75%RH, the cushion performance does not
dramatically change.
When the talc level increased (Figure 69, Figure 70), at first drop the peak
deceleration decreases and then increases, the lowest peak deceleration shifts from 0.2 psi
to 0.6 psi.

94

Figure 60 Starch foam without polymer modifiers cushion curves

Figure 61 Starch foam with PHAE cushion curves

95

Figure 62 Starch foam with PVOH cushion curves

Figure 63 Starch foam with PHE cushion curves

96

Figure 64 Starch foam with PVB cushion curves

Figure 65 Polyethylene foam cushion curves

97

Figure 66 Cushion curves comparison (First drop)

Figure 67 Cushion curves comparison (Fifth drop)

98

Figure 68 Starch foam with PVOH at 75% RH and RT

Figure 69 Starch foam with PVOH and 1.3% talc

99

Figure 70 Starch foam with PVOH and 2.7% talc
4.4 Dynamic Stress-Strain Curve
Young’s equation:

σ = Eε

(4-2)

Energy absorbed by cushion:
ε

ε2

0

2

U = ∫ EεA ⋅ tdε = EAt

=

σ 2 At
2E

(4-3)

t= cushion thickness
Potential energy

U = Wf h

Peak force:

F = σA =

2 EW f hA

(4-4)

t
100

2 EW f hA
t
a=

−Wf = W *a

2 EGhA
2 EhA
2 Eh
2 Eh
− G = G(
− 1) = G (
− 1) = G (
− 1)
Wf
tW
tW f
ts
t
A
(4-5)

E value is not constant in reality; it will change with the strain.
Energy balance
ε

W f h = ∫ σ ⋅ dε ⋅ t ⋅ A
0

εm
sh
= ∫ σ ⋅ dε
0
t

(4-6)

Force balance at ε m :

σ m A − W f = W ⋅ a ⇒ σ m A − sA = W ⋅ a
⇒σm =

Wf a
AG

+s=(

a
+ 1) s
G

(4-7)

In the stress-strain curve (Figure 71), the area under the curves shows sh/t, and the
maxium stress value is (a/G+1)s. Theoretically, from this sh/t vs. (a/G+1)s curve, the
peak deceleration for different drop height and foam thickness can be calculated.
sh/t vs. (a/G+1)s curve method can predict polystyrene foam and polyethylene
foam cushion performance well. However, with starch foam with different thicknesses,
sh/t vs. (a/G+1)s curves derived from cushion curves (Figure 72) are different for
different thickness. This is due to the corrugation of the foam plank. sh/t vs. (a/G+1)s
101

curve of 2 inch thickness is higher than that of 1 inch thickness (Figure 73), because the
stress strain curves are different, which leads to a different area under the curve.
Therefore, this method cannot predict the cushion performance on different thickness
foam.

Figure 71 Stress-strain curve

102

Figure 72 First drop cushion curves for 1 inch and 2 inch starch foam

Figure 73 The sh/t v.s. (a/G+1)s curve

103

Chapter 5 Starch foam extrusion mechanism
This chapter describes qualitatively the starch foam extrusion process.
5.1 Extensional viscosity and strain hardening
Strain hardening plays an important role in the foam extrusion process, which can
be described as extensional viscosity increase when strain increases at a certain shear rate
(Figure 74).

Figure 74 Strain hardening
The relationship between the extensional viscosity and shear viscosity can be
derived as follows:

Figure 75 Uniaxial pull of polymer block
Uniaxial pull (x1 direction), (Figure 75)
The definition of True Strain:
104

dε =

dl
l

(5-1)

Taylor expansion:

dl
l
1
1
1
= ln
= (l − l0 ) + (− 2 )(l − l0 ) 2 + ...
l0 l
l0
l0
2! l0

ε =∫

l

(5-2)

If higher order terms are omitted, then it is engineering strain

ε=

l − l0
l0

(5-3)

The strain rate is defined by

&
ε=

dε
1 dl dl 1 v
= ⋅ = ⋅ =
dt dt l
dt l l

(5-4)

The derivation of relationship between the extensional viscosity and shear viscosity under
one dimensional pull:
Mass conservation over the boundary (Figure 75):
.

.

.

.

in − out + gen = accum

v3 ⋅ 2 ⋅ ( x1 ⋅ 2 x 2 ) + v 2 ⋅ 2 ⋅ ( x1 ⋅ 2 x3 ) + v1 ⋅ 4 x 2 x3 = 0
&
&
&
⇒ ε 33 x3 ⋅ x1 x 2 + ε 22 x 2 ⋅ x1 x3 + ε 11 x1 ⋅ x 2 x3 = 0
Set

&
&
ε 11 = ε

&
&
⇒ ε 33 = ε 22 = −

&
ε

(5-5)

2

105

Internal stress has two components, the fluid pressure p and the stress due to viscous:

τ ij = η (

∂vi ∂v j
+
)
∂x j ∂xi

Force balance on

x1 direction

&
σ = σ 11 = ( p − p atm ) + 2ηε
Force balance on

x2

x3

and

(5-6)

direction

0 = σ 22 = σ 33 = ( p − p atm ) − 2η

&
ε

2

(5-7)

σ is the external stress applied.
&
Solve the above two equation, ⇒ σ = 3ηε (uniaxial pull)

(5-8)

-1

&
One dimensional pulling, the true strain rate needs to be constant, such as ε = 0.1 s , so

the length changing rate is not constant
t

&
&
ε (t ) = ∫ εdt = ε ⋅ t
0

So:

and

ε (t ) = ∫

l

l0

dl
l (t )
= ln
l
l0

&
l (t ) = l 0 exp(ε ⋅ t )

(5-9)

The cross section area will also change:
Conservation of mass (volume):

A(t ) ⋅ l (t ) = A0 ⋅ l 0

&
A(t ) = A0 exp(−ε ⋅ t )
The stress will be: σ (t ) =

(5-10)

F (t )
A(t )
106

So, the transient extensional viscosity at time t:
+
η E (t ) =

F (t )
σ (t )
=
&
&
ε (t ) ε ⋅ t ⋅ A0 exp(−ε ⋅ t )

at the transient length

&
l (t ) = l 0 exp(ε ⋅ t )

If biaxial pull (x1 and x2 directions):

&
&
&
ε 11 = ε 22 = ε

&
From mass conservation: ε 33

&
= −2ε

&
σ 11 = σ 22 = ( p − p atm ) + 2ηε , σ

is the external stress applied

&
0 = σ 33 = ( p − p atm ) − 2η ⋅ 2ε
&
⇒ σ 11 = 6ηε
So, η E

+

= 6η

(5-11)

(biaxial pull)

To non-Newtonian fluid, η E

+

= 3χη

(uniaxial pull)

(5-12)

When a force is applied to melt a block/sheet, strain weakening will selfaccelerate the deformation, since higher strain will cause lower viscosity, which cause the
polymer melt deform even faster. Strain hardening will self-regulate to slow down the
deformation, and as the strain increases the viscosity also increases correspondingly.
Therefore, the strain will be slowed down when a fixed force is applied (Figure 76 (1)).
Strain hardening has a self-healing mechanism.
During the foaming process, if the polymer melt is strain weakening, the gas in
the polymer will expand to form a large bubble - Since the viscosity is low due to the
107

high strain this size difference will be pronounced. If the polymer melt exhibits strain
hardening the gas in the polymer will expand the small bubble and since the viscosity
around the large bubble is high due to the high strain, the bubble size difference will be
compensated to make the bubble sizes more uniform (Figure 76 (2)).
During blowing film, strain hardening will help stabilize the film as any uneven
deformation rate will be offset by increased viscosity (Figure 76 (3)).

Figure 76 strain hardening effects on foam and film process (1: force applied to a block, 2:
foaming process, 3: film process; a: original state, b: strain weakening, c: stain hardening)
The shear viscosity of the starch and water extrudate can be described as follows:

η = Kγ& n −1

(5-13)

k
K = k 0 ⋅ exp( 1 ) ⋅ exp(−k 2 ⋅ Mc)
T

(5-14)

108

n = exp(

− k3
) ⋅ exp(−k 4 ⋅ Mc)
T

(5-15)

Mc- moisture content, K, n - parameters in power law model, T – temperature, k1, k2, k3,
and k4 - constants
As moisture content decreases, the viscosity exponentially decreases.
Although the starch and water mixture are known not to exhibit strain hardening
behavior with a fixed plasticizer (water) content as the extrudate exits the die, it loses
plasticizer (water) during the foaming process, thus the mixture has built-in strain
hardening behavior.
5.2 Starch foaming steps

The starch foam extrusion process can be understood as follows:
1) Extruder is treated as a mixing machine; the rheology of the final extrudate before the
die can be described using the power law model.
2) When the extrudate passes through the die, it experiences significant pressure drop,
which can be calculated using die geometry and extrudate viscosity. The high rate of
pressure drop causes the nucleation sites.
3) After the extrudate exits the die, the water continues to vaporize and diffuse into and
expand the existing bubble.
5.3 Die design

The extruder die geometry is important to the foam extrusion process, as it
controls the pressure build at the die and thus the pressure drop rate which causes
nucleation under the same operating conditions.
109

5.3.1 Strand die

When strand die is used for the extrusion, the shear stress and shear rate can be
calculated using equations:

τ=

Rh ΔP
ΔL

(5-16)

γ& =

2Q ( a + bn)
Rh An

(5-17)

τ = Kγ& n

(5-18)

Rh: hydraulic radius, flow area / perimeter
Q: volumetric flow rate
a, b: a=0.25; b=0.75 for strand die
The above equations could be derived in easier and more understandable
following way:

Figure 77 Strand die

110

Figure 77 (cont’d)

When the extrusion reach steady state, which means no acceleration
Force balance:

ΔP ⋅ πr 2 = 2πrL ⋅ τ
⇒τ =

ΔP ⋅ r
and
2L

ΔP ⋅ R
2L

τw =

(5-19)

Flow rate:
R

R

0

0

R

u ( r =0 )

Q = ∫ 2πrudr = π ∫ udr 2 = π ⋅ [u ( r ) ⋅ r 2 − ∫ r 2 du ]
0

0

R

= π ⋅ −∫ r 2
0

R
du
dr = π ⋅ ∫ r 2γ&dr
0
dr

Boundary condition:

u ( R ) = 0; τ ( R ) = τ w

r
τ
R
= ⇒r=
⋅τ
τw R
τw
τw

Q =π ⋅∫

0

(5-20)

and

dr =

R

τw

⋅ dτ

πR 3
( τ ) ⋅ γ ⋅ ⋅ dτ = 3
τw
τw
τw
R

Power law model τ

2

R

= Kγ n

111

τw

∫

0

τ 2 γ ⋅ dτ

(5-21)

Q=

πR
3
τw

3

τw

∫

0

1
1
τ
πR
τ 2 ( ) dτ = 3 ⋅ ( ) ⋅
⋅τ
1
K
τw K
3+
1
n

3

1
n

3+

1 τw
n
0

(5-22)

n

And plug in τ w =

ΔP ⋅ R
2L
1

ΔP n
n
⇒ Q = π ⋅(
⋅R
) ⋅
2 LK
3n + 1
dP
⎛ Q ⎞
−
= 2⋅ K ⋅⎜ 3 ⎟
dt
⎝ πR ⎠

n +1

3 n +1
n

⎛ 3n + 1 ⎞
⋅⎜
⎟
⎝ n ⎠

(5-23)

n

(5-24)

As shown in the above equation, when the diameter of the strand die is small and
throughput is large, it helps to increase the die pressure and thus pressure drop rate. If the
viscosity of the polymer mixture prior to exiting the die increases (K increases) e.g. by
reducing water content, the die pressure increases.
From the relationship between Q, L, K, n, R and ∆P it is useful to predict the die
pressure before extrusion trials to avoid pressure overloading.
When running foam peanuts using a strand die, L should be as short as possible.
As can be seen from the above equation, the pressure drop rate is independent of land
length. If the land length is too long in order to avoid die pressure overloading, the water
content is increased so the pressure drop rate is low and bad foaming. This explanation
fits experimental observation. As the land length increases foaming is poor since further
reduction in the water content is not possible since the die pressure and machine torque
are already high.

112

Reducing the total throughput is an effective way to reduce the pressure drop rate.
As seen in previous chapter this will reduce foaming. When the starch feed rate was
reduced from 460 lbs/hr to 285 lbs/hr, the diameter of the foam decreased from 20mm to
16mm.
5.3.1 Annular die

When the gap of the annular die is very small, (Ro-Ri)/L <10, it can be treated as
two parallel plates.

Figure 78 Annular die
Force balance:
ΔP ⋅ w ⋅ y = τ ⋅ w ⋅ L

τ=

ΔP ⋅ y
L

On the wall, y =

τw =

⇒

(5-25)

h
2

ΔP ⋅ h
2L

τ
2y
=
τw
h

(5-26)

and

dτ =

2τ w
dy
h

Flow rate calculation (include the power law model):

113

Q = 2∫

h/2

0

= 2 w∫

h/2

0

w ⋅ dy ⋅ u = 2 w∫

h/2

0

y(

τ

u ⋅ dy = 2 w(uy 0 − ∫
h/2

h/2

0

y⋅

du
⋅ dy )
dy

1/ n

K

) dy
τw

⇒ Q = 2 w∫ (
0

h
h
τ
) ⋅τ ⋅ ( )1 / n ⋅ (
) ⋅ dτ
K
2τ w
2τ w
1

1

2+
h 2 1 n 1
) ⋅( ) ⋅
= 2 w(
⋅τ w n
1
2τ w
K
2+
n

1
⇒ Q = w⋅h
2

2 n +1
n

(5-27)

1

ΔP n
n
⋅
⋅(
)
2n + 1 2 LK

dP
⎛ 2Q ⎞
−
= K ⋅⎜ 2 ⎟
dt
⎝ wh ⎠

n +1

⎛ 2n + 1 ⎞
⋅⎜
⎟
⎝ n ⎠

(5-28)

n

(5-29)

From the above relationship between die geometry, polymer viscosity and die
pressure it is also possible to predict die pressure before experiments in order to avoid
pressure overloading. For reasons similar to using a strand die the land length of the die
should be kept as short as possible to obtain good foaming.
When the gap of the annular die is decreased, the die pressure goes up. This trend
fits experimental observation. In lab extrusions with an annular die under the same
extrusion conditions decreasing the gap from 1mm to 0.38mm, increased the die pressure
from 100psi to 450 psi.

114

In order to maintain the high pressure drop rate, the pressure needed to be reduced
at the very front tip for this reason a large die space (space between die PIN and die
sleeve) is preferred for an annular die.
Figure 79 is the model fitting using experimental data, where in the annular die
3

3

design model, Q=1.98E-06 m /s (the density of the melt is 1.2 kg/m ).

Figure 79 Die pressure (Pa) vs. die gap (m), curve: calculated from the model, two dots:
experimental values

115

Chapter 6 Chemically modified starch
6.1 Starch extrusion with soy based ester

Maleated thermoplastic starch (MTPS) (Figure 80) has significant advantage over
traditional thermoplastic starch due to low melt viscosity, and ease of blending and
processing with other polymers. However, it is moisture sensitive because of the
hydrogen bonding with water vapor. In these experiments, maleated thermoplastic starch
was extruded with soy based ester and the extrusion process was studied. The reaction
mechanism is shown in Figure 80.

OH

OH

O
O

OH

H2O, H+

O
O

OH

OH

OH

OH

Figure 80 Reactions in MTPS

116

Figure 80 (cont’d)
OH

OH
OH

O
OH

O

H+

OH

OH

O
OH

O

OH

OH

OH

O

OH
OH

O

O

OH
CH2O
O
O

OH

C

C
H

O
OH

O
O

O

OH

OH

O

OH

OH

117

C
H

C

OH

To prove glycerol is chemically or strongly physically bonded into the starch
backbone, soxhlet extraction was done with different solvents (Table 23). The results
showed that the soxhlet extraction result was dependent on the solvent used. It shows that
the glycerol extracted varies when different solvents are used. This is due to the solvent glycerol interaction: Acetone < Ethanol < Methanol. When the solvent-glycerol
interaction is stronger, more glycerol will be extracted, the acetyl reaction is an
equilibrium reaction, and the reaction will favor the production of glycerol if glycerol is
continuously removed from the system. Starch with low molecular weight was also
extracted by methanol since it has a lower recovery yield. Acetone was selected for the
soxhlet extraction for extruded MTPS with ester.
#
1
2
3
Sample
MTPS
MTPS MTPS
Solvent
Acetone Ethanol Methanol
Weight before (g)
10.058 10.063
10.092
Weight after (g)
9.502
8.078
7.349
Recovery yield (unextracted percentage)
94.5% 80.3%
72.8%
If theoretically only glycerol extracted, the
recovery yield
80.4% 80.4%
80.4%
Table 23 Soxhlet extraction with different solvents
The unsaturated biodiesel was ozonated in the presence of methanol. The
ozonolysis reaction is shown in Figure 81. Basically, the ozone cleaves the carbon-carbon
double bond to form the carbonyl group. After reaction and purification by vacuum
distillation, the ester mixture (the components are shown in Table 24) has two forms, if
water was distillated completely; the ester mixture is one phase. If water was not totally
removed, the ester mixture is two phase, the top is oil phase and the bottom is aqueous
phase. The two forms of oil mixture were extruded with starch and glycerol with maleic
anhydride. The transesterification reaction is shown in Figure 82.

118

R1

R3

R3

R1

O3

O
R2

R4

R4

R2

Figure 81 Ozonolysis
Mixture of the esters

Structure

Dimethyl azelate

Methyl hexanoate

Methyl nonanoate

Methyl palmitate

Methyl stearate

Table 24 Products from ozonolysis

119

O

The extrusion running conditions are shown in Table 25.
Sample

Starch (dried) (g)

Ester (g)

Glycerol (g)

MTPS

800

0

200

Maleic
anhydride
(g)
20

MTPS + two phases 700
100
200
20
ester mixture
MTPS + one phase
700
100
200
20
ester mixture
Table 25 Extrusion conditions for thermal plastic starch with esters
The two phase ester mixture is very effective for starch and ester blending
extrusion. Both of the residue percentage and FT-IR results show transesterification
occurred (Figure 84 and Figure 85). The extrudate with low percent of residue (one phase
ester mixture) is unsuitable for use due to phase separation.

O

O
HO
R1

OH

R2

O

R1

Figure 82 Transesterification reaction

120

O

R2

uV(x10,000)

2.0

1.5

1.0

0.5

0.0

-0.5
21.0

21.5

22.0

22.5

23.0min

Figure 83 GC-FID before and after ozonation using Dean-stark setup (The retention time for dimethyl azelate is 20.74 minutes)

121

Figure 84 FTIR results for the extrudates after soxhletion with acetone (A) MTPS (B)
MTPS with two phases esters (C) MTPS with one phase ester

Figure 85 Soxhletion results (acetone as solvent) for the extrudate
The FTIR result shows that the ester has grafted onto the backbone of starch with
carbonyl group 1720cm-1 after sample Soxhlet extraction using acetone (Figure 84 B).

122

The extrudate (MTPS with two phase esters) has a high residual percentage after acetone
Soxhlet extraction which also demonstrated ester grafting (Figure 85).
6.2 Starch phosphate

The object of this work is to synthesize starch phosphate using sodium
trimetaphosphate and glycerol phosphoric acid by an extrusion method and examine its
use as a possible electrorheological fluid or as a flame retardant additive.
6.2.1 Starch phosphate for ER fluid

Introduction
Electrorheological fluids (also known as smart fluids) are suspensions of fine,
non-conducting particles in an electrically insulating oil that undergos a dramatic and
reversible change in their apparent viscosity (in the order of 100,000 times) in response to
an external electric field.
The mechanism of the ER fluid is simple: when the electric field is applied, the
suspended particles will line up and block the flow of liquid. The principle is shown
schematically in Figure 86.

123

Figure 86 Alignment of suspended particles
A large number of suspensions were studied as show below:

Dispersant

Oil
Additive
chlorinated oil,
liquid paraffin,
water
Cellulose
hydraulic oil,
various electrolytes
Chlorotoluene,
silicone oil
transformer oil,
none
Gelatin
olive oil,
mineral oil
water
Na carboxymethyl- paraffin oil,
silicone oil,
or water and sorbitan
cellulose
chlorinated solvents
mineral oil,
Water and/or sorbitan
olive oil,
Starch (flour)
oleate or laureate
petroleum spirit,
Vaseline oil
Cellulose
silicone oil
none
phosphate
kaolinite carboxy- silicone oil
none
methyl starch,
Table 26 Suspensions for electrorheological fluid
124

Ref.

[59],[60]

[61]
[62], [63],
[64]
[65],[66]

[67]
[68]

A major problem with ER fluids is caused by suspension instability. Over time
there is a tendency to settle due to different density of the solid and liquid. In this
research starch phosphate nanoparticles were synthesized in an attempt to overcome this
obstacle.
Experimental
Starch phosphorylation was done using extruder (Figure 88). The reaction (Figure
87) was under base condition.

125

NaO

O

O

P
O
O

O

P

P

+ St

St

OH

P

O

ONa

O

P

ONa

O
ONa

O

O
O

ONa

P

OH

ONa

O
O
St

O

P

O
O

O

P

O

O

P

ONa

St

O

ONa

O

O

ONa

O

When PH<7:

St

OH

P

O
O

O

P

OH

O

P

O

P

ONa

O

P

ONa

ONa

O

-

+

O

P

ONa

OH

ONa
O

O
St

St

ONa

O

St

+ HO

P
ONa

ONa

O
O

P

O

-

+

ONa

When PH>8:
Figure 87 Starch phosphorylation using sodium trimetaphosphate

126

H

+ P

+

O

Figure 88 Feed position for starch phosphorylation
Experimental procedure:
1: Starch 5 kg/h, Maleic Acid 0.12kg/h, water 1kg/h fed at the beginning of the extruder.
2: Sodium trimetaphosphate (SMTP) 0.22kg/h, water 0.78kg/h, sodium carbonate
0.15kg/h fed at the middle of the extruder.
The extruder temperature is set at 90°C.
Screw speed: 200rpm
Water acts as plasticizer. SMTP is cross-link agent.
After extrusion, the sample was cut into pellets, washed 3 times and dispersed into water
at 55°C for 8 hours.
After 4 hours, the upper suspension was collected and particle size was analyzed using
Dynamic Light Scattering:
Phosphorus characterization method:
This method is based on ISO 3946-1982 [69]. Briefly, there are four steps:
1) Concentrated sulfuric acid and nitric acid to digest sample
2) Ammonium molybdate as reaction agent. The following reaction occurs
3
2
PO4 − + 12 MoO4 − + 27 H + → H 2 [ P ( Mo2O7 ) 6 ] + 10 H 2O

(Yellow color)

127

+6

3) Add reducing agent Ascorbic acid to partially reduce Mo

+5

to Mo

(the solution is

blue)
4) Measure the light absorption percentage at wavelength 825 nm with visual spectra
photometer
Rheology measurement under high voltage:
The viscosity versus shear rate was measured using a plate-plate rheometer
(Figure 89). The shear stress and shear rate can be obtained from the torque applied M
and rotating speed ω respectively. The 1100v and 1500v voltage were applied for
2200v/mm and 3000v/mm electric field.

Figure 89 Rheology measurement setup
Particle size measurement:
90Plus Particle Size Analyzer (Brookhaven Instruments Corporation, Holtsville,
NY) was used to characterize the particle size.
Ionic exchange for sodium glycerophosphate:
Glycerol phosphate disodium hydrate was ion-exchanged (ion-exchange resin:
Dowex® HCR-W2 hydrogen form, acidic sulphonated polystyrene cation exchange resin)
to designed monohydrogen and dihydrogen form.

128

Results:
The number average diameter of starch nano particles is 26.6nm. The particle size
distribution is given in Figure 90:

Figure 90 Dynamic light scattering for starch nanoparticles

Figure 91 starch nanoparticles stabilized with phosphate
In the starch nano particles suspension, inside particle, starch molecules form
hydrogen bonding with themselves. They can not be approached by water molecules,
therefore they form a relative “hydrophobic core”. On the particle / water interface, there
is interaction (hydrogen bonding) between the starch (unbounded –OH group) and water
which acts like a surfactant, to keep particles apart. However, the interaction between
129

starch and water will be finally replaced by interaction between the starch molecules.
This will cause agglomeration and precipitation.
To avoid these effects the hydroxyl groups on the starch particles were
phosphorylated to better stabilize the particle. This should strengthen the interaction
between the solute and water and thus avoid retro gradation (Figure 91).

120
No electric field
2200 v/mm
3000 v/mm

Viscosity (cp)

100
80
60
40
20
0

0

50

100

150

200

250

300

Shear rate (1/s)
Figure 92 Viscosity v.s. shear rate for ER fluid
It is apparent that the effect of an electric field on the fluid (Figure 92) was not
significant, possibly due to a combination of too small particle size, insufficient degree of
phosphorylation, and a suspension having too low solid content.

130

6.2.2 Starch phosphate with glycerol phosphoric acid

The starch phosphate reaction is shown in Figure 93, basically, the reducing end
(Hemiacetal) of the starch reacts with the hydroxyl group in the glycerol phosphoric acid
to form an acetal linkage.

131

OH

OH
OH

OH
+
H

O
O

OH

OH

O
O

OH

O

OH

O
OH

OH
HO

P

O

OH

O
HO

P
OH

Figure 93 Starch phosphorylation reaction

132

O

Extrusion formula
P wt% after purification
Starch + 30 wt% water + 10wt% sodium trimetaphosphate + 1 0.72 wt%
wt% sodium carbonate
Starch + Maleic Anhydride 2.5 wt% + glycerol 20 wt%, after 0.28 wt%
extrusion, extrude it with sodium trimetaphosphate (10wt%
based on starch)
800 g starch + 20 g Maleic anhydride + glycerol 200g + 80 g
0.06 wt%
sodium trimetaphosphate
Starch + 20 wt% water + 20 wt% sodium tripolyphosphate
0.03 wt%
Table 27 Extrusion formulation using sodium trimetaphosphate and sodium
tripolyphosphate
Modifying starch with sodium trimetaphospahte under basic condition give the
highest phosphorus percentage grafted when water is the plasticizer (Table 27).
Formula
P wt% after purification
800 g starch + 175 g glycerol + 25 g glycerol phosphate salt
0.13 wt%
+20 g maleic anhydride
Table 28 Phosphorylate starch using glycerol phosphate disodium salt
When a higher amount of glycerol phosphate salt was used, the process was
unstable, resulted in a high torque and the products were dark in color due to the high
percentage of salt. With low glycerol phosphate disodium salt addition levels, the
phosphorus content was reduced to 0.13% (Table 28). Phosphorus contents of different
formulation (different combination of glycerol phosphate monoester monohydrogen and
dihydrogen form) are shown in Table 29.
Formula
Sample #
1
2
3
4

P wt% after
purification

Glycerol
Glycerol
Dry
Maleic
(g)
phosphoric
Starch (g) Anhydride
acid (g)
(g)
800
20
175
25
0.26 wt%
800
20
150
50
0.33 wt%
800
20
125
75
0.56 wt%
800
0
150
50
0.56 wt%
Table 29 Phosphorylate starch using glycerol phosphoric acid

133

Residue percentage after soxhelt
Glycerol phosphate
using acetone
monoester (g)
Dihydrogen form
Monohydrogen form
1 800
100
30
70
98.5%
2 800
100
20
80
NA, Very dark extrudate
3 800
100
10
90
NA, Very dark extrudate
Table 30 Phosphorylate starch using glycerol phosphoric acid and glycerol phosphate
monohydrogen
# Regular
Starch
(g)

Glycerol
(g)

Materials

Intrinsic viscosity [η]
(dL/g)
Cargill starch SMP 1100
1.26
Maleated thermoplastic starch
0.099
Starch glycerol phosphate sample
0.091
Table 31 Intrinsic viscosity comparison
The above results show that the phosphate has been grafted into the starch
backbone (Table 30). The decrease in intrinsic viscosity shows that the starch was
hydrolyzed and potentially easily processable (Table 31). Future efforts will be to
increase the phosphorus content while not dramatically decreasing the starch molecular
weight.
6.3 Dihydroxyl starch

The dihydroxyl starch (DHS) synthesis consists of two steps: (1) Add sodium
periodate to oxidize the starch to produce dialdehyde starch (DAS) and (2) Add sodium
borohydride to reduce the DAS to produce the corresponding hydroxyls (Figure 94).

134

OH

OH
O

NaIO4

O

OH
O

NaBH4

O
O

HO

O
O
HO

O

OH

OH
O

O

Figure 94 Reactions to make dihdyroxyl starch

135

O

6.3.1 Experimental and characterizations

DHS synthesis steps:
(1) Starch and sodium periodate is put into a flask according to the amount shown in
Table 32. Cover the flask with black plastic bags to avoid light.
(2) Add 500ml distilled water, stir it and let it react at room temperature for 6 hours.
(3) After 6 hours, filter and put it into a flask. Add 500ml distilled water. Add sodium
borohydride, agitate and allow reacting at room temperature for 2 hours.
(4) After reaction, dialysis the solution against distilled water, till the conductivity of the
solution inside of the dialysis tube is less than 10.
(5) Freeze dry at low temperature and low pressure to vaporize water.
40H
10 g

100H
10 g

40W

100W

High amylose starch
Waxy starch
10 g
10 g
4.65 g
13.94 g
4.65 g
13.94 g
NaIO4
1.23 g
3.08 g
1.23 g
3.08 g
NaBH4
Table 32 weights for reactants during DHS synthesis
Characterizations
Determination of aldehyde groups by titration:
The aldehyde groups present in the DAS and DHS were determined by
quantitative reduction using sodium borohydride. This method [70] was found to be
effective and rapid for determination of aldehyde groups in starch. Briefly, the sodium
borohydride reacts with water to evolve hydrogen gas under acid conditions. From the
difference of the hydrogen evolved between control sample without aldehyde group and
sample with aldehyde group, the amount of sodium borohydride reacted with aldehyde
can be determined. The amount of hydrogen was measured by volumetric glass tube [
Figure 95].
136

Figure 95 Setup for measuring aldehyde group
FTIR:
The samples were characterized by Perkins Elmer FTIR. The samples were compressed
with KBr and ran for 64 scans to achieve a high resolution. The wavelength range:
-1

-1

4000cm to 450 cm .
6.3.2 Results and discussions

Three different grades of starch (High Amylose, Waxy, and Regular) were used
-1

for the reactions. The FTIR results are shown in Figure 96, the peak around 1720cm is
the carboxyl C=O peak; DAS samples have stronger carbonyl peak than DHS.

137

Figure 96 FTIR for starch, DAS and DHS, A: high amylose, B: waxy, C: regular.

138

Figure 96 (cont’d)

Figure 97 Viscosity vs. shear rate for DHS

139

Figure 98 Dialdehyde content in DAS and DHS
The results show that dihydroxyl starch was successfully synthesized. The
solution exhibit shears thinning behavior Figure 97. The Figure 98 shows that amylose is
more difficult to oxidize and is easier to reduce once it has been oxidized.

140

Chapter 7 Conclusions, accomplishments and recommendations for future work
7.1 Conclusions

It has been shown that using polyhydroxyl ether (PHE) and polyvinyl butyral
(PVB) are effective polymer modifiers which improve the hydrophobicity (humidity
resistance) of starch foam sheets without sacrificing performance properties (cushion
protection and thermal insulation). With both a lab scale extruder (Century ZSK-30) and
a commercial scale extruder (Wenger TX-80), the foam extrusion processing parameters
(functional aid, talc, water, die geometry) were investigated using black box strategy
(Design of Experiments). These starch foam sheets can provide the design flexibility
required to create a cost effective protective packaging. They are bio-based and
completely biodegradable, provide excellent insulation, and are naturally anti-static.
Chemically modified starch was synthesized by extrusion. The extrusion method
effectively breaks down the hydrogen bonding within starch to provide plastic like flow
in a more efficient technique than conventional solution methods thus reducing
manufacturing costs.
Starch phosphate was synthesized using sodium trimetaphosphate and glycerol
phosphoric acid as phosphorylation reagent, 1.1 wt% phosphorus was obtained through
the phosphorylation. Starch was also extruded with ester monomer (ozonation product
from unsaturated oil) to improve the hydrophobicity of the maleated thermoplastic starch.
7.2 Main accomplishment

The moisture sensitivity of the starch foam sheet poses a limitation for shipping
coolers applications since the moisture will condense onto the sheet surface and gradually
141

dissolve the foam. The main accomplishment of this research is to significantly increase
the moisture resistance of starch foam sheet by using a commercially available,
economically acceptable resin as additive during the extrusion process. This process is in
commercial production.
7.3 Future recommendations

1) The foam sheets are corrugated. While this does not affect thermal insulation
properties, it does cause the foam to exhibit low static loading for multiple cushion
protection (weakness support in the curved location). It will also require more contact
surface thus increasing the cost for heavy product shipment packaging. Investigation of
the need for corrugation and its potential elimination would significantly increase
effectiveness and reduce cost.
Experiments show that when the annular gap is relatively large, the number of
holes (Figure 99 and Figure 17) in the annular die is equal to the number of corrugations
(Figure 100 left and middle picture). However, when the gap of the annular die narrowed,
the corrugation increased (Figure 100 right picture).

Figure 99 annular die (Pin) structure

142

Figure 100 Corrugation in the foam sheet (left: extruded with 10 holes pin, middle:
extruded with 6 holes pin, right: extruded with 10 holes pin and small gap)
The die shape (circle outlet shape from the extruder smooth transition to rectangle
shape) should be investigated. After the extrusion the foam should be in plank shape
(smooth surface) which can be glued together to make larger planks with no corrugations.
The slit instead of pinholes in the PIN part also could be tried in foam extrusion to avoid
the corrugation.
2) The humidity resistance of thermoplastic starch could be improved by reacting
it with butyraldehyde.
3) Polyvinyl alcohol could be reacted with aldehyde derived from ozonation of
unsaturated vegetable oil to increase the biobased content.
4) Testing biodegradability of starch foam with PVB (>10wt %)
5) Soy protein concentrate or high gluten content corn flour could be investigated
to replace the hydroxypropylated high amylose starch using the modified lab scale
extruder.
6) At high polyvinyl alcohol and starch ratio, the strand foam shrinks when
coming out of the die, adding stabilizer to offset the permeability difference between
water vapor and air may correct the problem.
7) Explore different application for starch foam sheet/paper, such as absorbent
pad of meat tray.
143

REFERENCES

144

REFERENCES

1 Harold, F. G.; Eldridge, M. M.; John R. W.; Part II in Extrusion: The Definitive
Processing Guide and Handbook (2005), by Brent Beckley / William Andrew, Inc
2 Nabar, Yogaraj; Narayan, Ramani; Schindler, Melvin; Twin-screw extrusion
production and characterization of starch foam products for use in cushioning and
insulation applications, Polymer Engineering and Science (2006), 46(4), 438-451.
3 Nabar, Yogaraj; Raquez, Jean Marie; Dubois, Philippe; Narayan, Ramani; Production
of starch foams by twin-screw extrusion: Effect of maleated poly(butylene adipate-coterephthalate) as a compatibilizerBiomacromolecules, (2005), 6(2), 807-817.
4 Ayoub, Ali; Rizvi, Syed S. H.; Reactive supercritical fluid extrusion for development of
moisture resistant starch-based foams, Journal of Applied Polymer Science (2011),
120(4), 2242-2250.
5 Robin, Frederic; Theoduloz, Christine; Gianfrancesco, Alessandro; Pineau, Nicolas;
Schuchmann, Heike P.; Palzer, Stefan; Starch transformation in bran-enriched extruded
wheat flour, Carbohydrate Polymers (2011), 85(1), 65-74.
6 Fascio, Carlo; Biodegradable packaging of starch and fiber made by extrusion, PCT Int.
Appl. (2010), WO 2010148489.
7 Sjoeqvist, Mia; Boldizar, Antal; Rigdahl, Mikael; Processing and water absorption
behavior of foamed potato starch, Journal of Cellular Plastics (2010), 46(6), 497-517.
8 Mali, Suzana; Debiagi, Flavia; Grossmann, Maria V. E.; Yamashita, Fabio; Starch,
sugarcane bagasse fibers, and poly(vinyl alcohol) effects on extruded foam properties: A
mixture design approach, Industrial Crops and Products (2010), 32(3), 353-359.
9 Patel, Sameerkumar; Venditti, Richard A.; Pawlak, Joel J.; Ayoub, Ali; Rizvi, Syed S.
H.; Development of crosslinked starch microcellular foam by solvent exchange and
reactive supercritical fluid extrusion, Journal of Applied Polymer Science (2009), 111(6),
2917-2929.
10 Mariam, Irfana; Cho, Ki Yul; Rizvi, Syed S. H.; Thermal properties of starch-based
biodegradable foams produced using supercritical fluid extrusion (SCFX), International
Journal of Food Properties (2008), 11(2), 415-426.
11 Alavi, Sajid; Rizvi, Syed S. H.; Strategies for enhancing expansion in starch-based
microcellular foams produced by supercritical fluid extrusion, International Journal of
Food Properties (2005), 8(1), 23-34.

145

12 Ogunsona, Emmanuel; D'Souza, Nandika A.; Supercritical CO2 foamed
biodegradable polymer blends of polycaprolactone and starch based material, Annual
Technical Conference - Society of Plastics Engineers (2008), 66th, 2261-2264.
13 Lee, Siew Yoong; Eskridge, Kent M.; Koh, Woon Yuen; Hanna, Milford A.;
Evaluation of ingredient effects on extruded starch-based foams using a supersaturated
split-plot design, Industrial Crops and Products (2009), 29(2-3), 427-436.
14 Kitamura, Shushi; Manufacturing method of foamed heat-insulating material utilizing
paper and starch, U.S. Pat. Appl. Publ. (2009), US 20090127731.
15 Lee, S. Y.; Eskridge, K. M.; Hanna, M. A.; Effects of extrusion variables on
organoclay intercalation and properties of tapioca starch-poly(lactic acid) nanocomposite
foams, International Polymer Processing (2009), 24(1), 59-66.
16 Lee, Siew Yoong; Hanna, Milford A.; Tapioca starch-poly(lactic acid)-Cloisite 30B
nanocomposite foams, Polymer Composites (2009), 30(5), 665-672.
17 Lee, Siew Yoong; Chen, Han; Hanna, Milford A.; Preparation and characterization of
tapioca starch-poly(lactic acid) nanocomposite foams by melt intercalation based on
clay type, Industrial Crops and Products (2008), 28(1), 95-106.
18 Lee, S.-Y.; Xu, Y. X.; Hanna, M. A.; Tapioca starch-poly(lactic acid)-based
nanocomposite foams as affected by type of nanoclay, International Polymer Processing
(2007), 22(5), 429-435.
19 Pushpadass, Heartwin A.; Weber, Robert W.; Hanna, Milford A.; Expansion,
Morphological, and Mechanical Properties of Starch-Polystyrene Foams Containing
Various Additives, Industrial & Engineering Chemistry Research (2008), 47(14), 47364742.
20 Pushpadass, Heartwin Amaladhas; Babu, Govindarajan Suresh; Weber, Robert W.;
Hanna, Milford A.; Extrusion of starch-based loose-fill packaging foams: Effects of
temperature, moisture and talc on physical properties, Packaging Technology & Science
(2008), 21(3), 171-183.
21 Zhang, Jian-Feng; Sun, Xiuzhi; Biodegradable foams of poly(lactic acid)/starch. I.
Extrusion condition and cellular size distribution, Journal of Applied Polymer Science
(2007), 106(2), 857-862.
22 Zhang, Jian-Feng; Sun, Xiuzhi; Biodegradable foams of poly(lactic acid)/starch. II.
Cellular structure and water resistance, Journal of Applied Polymer Science (2007),
106(5), 3058-3062.

146

23 Qi Fang, Milford A. Hanna, Functional Properties of Polylactic Acid Starch-Based
Loose-Fill Packaging Foams, Cereal Chemistry (2000), 77(6), 779-783.
24 Xu, Yixiang; Hanna, Milford; Effect of eggshell powder as nucleating agent on the
structure, morphology and functional properties of normal corn starch foams, Packaging
Technology & Science (2007), 20(3), 165-172.
25 Mihai, Mihaela; Huneault, Michel A.; Favis, Basil D.; Li, Hongbo ; Extrusion
foaming of semi-crystalline PLA and PLA/thermoplastic starch blends, Macromolecular
Bioscience (2007), 7(7), 907-920.
26 Mihai, Mihaela; Huneault, Michel A.; Favis, Basil D.; Foaming of
polystyrene/thermoplastic starch blends, Journal of Cellular Plastics (2007), 43(3), 215236.
27 Glenn, Gregory M.; Klamczynski, Artur K.; Holtman, Kevin M.; Shey, Justin; Chiou,
Bor-Sen; Berrios, Jose; Wood, Delilah; Orts, William J.; Imam, Syed H.; Heat expanded
starch-based compositions, Journal of Agricultural and Food Chemistry (2007), 55(10),
3936-3943.
28 Guan, Junjie; Hanna, Milford A.; Physical, Mechanical, and Macromolecular
Properties of Starch Acetate during Extrusion Foaming Transformation, Industrial &
Engineering Chemistry Research (2006), 45(11), 3991-4000.
29 Guan, Junjie; Eskridge, Kent M.; Hanna, Milford A.; Functional Properties of
Extruded Acetylated Starch-Cellulose Foams, Journal of Polymers and the Environment
(2004), 12(3), 113-121.
30 Guan, Junjie; Hanna, Milford A.; Functional properties of extruded foam composites
of starch acetate and corn cob fiber, Industrial Crops and Products (2004), 19(3), 255-269.
31 Guan, J.; Fang, Q.; Hanna, M. A.; Selected functional properties of extruded starch
acetate and natural fibers foams, Cereal Chemistry (2004), 81(2), 199-206.
32 Guan, Junjie; Hanna, Milford A.; Selected morphological and functional properties of
extruded acetylated starch-cellulose foams, Bioresource Technology (2006), 97(14),
1716-1726.
33 Zhou, Jiahua; Hanna, Milford A.; Effects of the properties of blowing agents on the
processing and performance of extruded starch acetate, Journal of Applied Polymer
Science (2005), 97(5), 1880-1890.

147

34 Xu, Yixiang; Hanna, Milford A.; Physical, Mechanical, and Morphological
Characteristics of Extruded Starch Acetate Foams, Journal of Polymers and the
Environment (2005), 13(3), 221-230.
35 Xu, Y. X.; Dzenis, Y.; Hanna, M. A.; Water solubility, thermal characteristics and
biodegradability of extruded starch acetate foams, Industrial Crops and Products (2005),
21(3), 361-368.
36 Ganjyal, G. M.; Reddy, N.; Yang, Y. Q.; Hanna, M. A.; Biodegradable packaging
foams of starch acetate blended with corn stalk fibers, Journal of Applied Polymer
Science (2004), 93(6), 2627-2633.
37 Guan, Junjie; Hanna, Milford A.; Extruding Foams from Corn Starch Acetate and
Native Corn Starch, Biomacromolecules (2004), 5(6), 2329-2339.
38 Lui, Wai-Bun; Peng, Jinchyau; Effects of operating conditions on the cellular structure
and the thermal properties of extruded foams made from corn, Journal of Cellular Plastics
(2003), 39(6), 439-450.
39 Willett, J. L.; Shogren, R. L.; Processing and properties of extruded starch/polymer
foams, Polymer (2002), 43(22), 5935-5947.
40 Nobes, Geoffrey A. R.; Orts, William J.; Glenn, Gregory M.; Gray, Gregory M.;
Harper, Michael V.; Blends of starch with poly(vinyl alcohol)/ethylene copolymers for
use in foam containers; Annual Technical Conference - Society of Plastics Engineers
(2001), 59th(Vol. 2), 1865-1869.
41 Fang, Qi; Hanna, Milford A.; Characteristics of biodegradable Mater-Bi-starch based
foams as affected by ingredient formulations, Industrial Crops and Products (2001), 13(3),
219-227.
42 Fang, Qi; Hanna, Milford A.; Preparation and characterization of biodegradable
copolyester-starch based foams, Bioresource Technology (2001), 78(2), 115-122.
43 Cha, J. Y.; Chung, D. S.; Seib, P. A.; Effects of extrusion temperature and moisture
content on mechanical properties of starch-based foams, Transactions of the ASAE
(1999), 42(6), 1765-1770.
44 Bhatnagar, S.; Hanna, Milford A.; Starch-based plastic foams various starch sources,
Cereal Chemistry (1996), 73(5), 601-604.
45 Bhatnagar, Sandeep; Hanna, Milford A.; Properties of extruded starch-based plastic
foam, Industrial Crops and Products (1995), 4(2), 71-77.

148

46 Neumann, Paul E.; Seib, Paul A.; Manufacture of starch-based biodegradable cellular
packing fillers and the like, US patent (1993), 5,208,267.
47 Neumann, P. E.; Seib, P. A.; Starch-based, biodegradable packing filler and method of
preparing same, US Patent (1993), 5,185,382.
48 Bastioli, Catia; Bellotti, Vittorio; Del Giudice, Luciano; Lombi, Roberto; Rallis,
Angelos; Biodegradable plastic foams containing starch, PCT Int. Appl. (1991), WO
9,102,023.
49 Norman L. Lacourse, Paul A. Altieri, Biodegradable packaging material and the
method of preparation thereof, US Patent (1989), 4,863,655.
50 Norman L. Lacourse, Paul A. Altieri, Biodegradable shaped products and the method
of preparation thereof, US Patent (1991), 5,043,196.
51 Lacourse, Normal L.; Altieri, Paul A.; Starch-based biodegradable cellular molded
products as containers and packaging materials, Eur. Pat. Appl. (1990), EP 376,201.
52 Paul A. Altieri, Martin M. Tessler, Water humidity resistant starch foams, US patent
(1996) 5,554,660.
53 Catia Bastioli, Vittorio Bellotti, Luciano Del Giudice, Roberto Lombi, Angelos Rallis,
Expanded articles of biodegradable plastics materials and a method for their production,
US patent (1994) 5,288,765.
54 Catia Bastioli, Vittorio Bellotti, Luciano Del Giudice, Roberto Lombi, Angelos Rallis,
Expanded articles of biodegradable plastic materials, US patent (1994), 5,360,830.
55 ASTM Standard D1596 - 97(2011) “Standard Test Method for Dynamic Shock
Cushioning Characteristics of Packaging Material”, ASTM International, West
Conshohocken, PA, DOI: 10.1520/D1596-97R11.
56 Montgomery, Douglas C.; Design and Analysis of Experiments (2012), published by
Wiley.
57 Box, G.E.P.; Behnken, D.W.; Some new three level designs for the study of
quantitative variables, Technometrics (1960), 4, 455-475
58 Yang, Zhiguan; Graiver, Daniel; Narayan, Ramani; Extrusion of humidity-resistant
starch foam sheets (2013), 53(4) 857-867
59 Uejima H.; Dielectric mechanism and rheological properties of electro-fluids, Japan J.
Appl. Phys. (1972), 11, 319-326
149

60 Takeo K.; Omura, Y.; Hydraulic oil composition, US Patent (1976) 3,984,339.
61 Winslow, W. M.; Method and means for translating electrical impulses into
mechanical force, US Patent (1947) 2,417,850.
62 Qin, A.; Physico-chemical studies of macromolecular systems in solution, PhD Thesis,
University of Liverpool (1982)
63 Kelly, J. P.; An investigation into the properties of electrorheological fluids
(polyacene quinine radical, phthalic anhydride quinine radical), MSc Thesis, University
of Liverpool (1990)
64 Strangroom, J. E., UK Patent Specification (1978) 1,501,635.
65 Westhaver, J.W.; Electroviscous fluids, US Patent (1976), 3,970,573.
66 Deynega, Yu F., Popko, K. K., and Kovganich N. Ya, Heat Transfer-Sou. Res. (1978),
10, 50.
67 Kim, S. G.; Choi, H. G.; Myung S. J.; Preparation and Characterization of Phosphate
Cellulose-Based Electrorheological Fluids, Macromol. Chem. Phys. (2001), 202, 521–
526.
68 Wang, B. X.; Zhao, X. P.; Electrorheological effect coordinated by kaolinite
carboxymethyl starch hybrid materials, J Mater Chem (2002), 12, 2869.
69 Starches and derived products -- Determination of total phosphorus content -Spectrophotometric method, ISO 3946:1982.
70 Lindberg, B.; Misiorny, A.; Quantitative analysis of reducible carbohydrates by means
of sodium borohydride. Svensk. Papperstidn (1952), 55, 13-14.

150