. ymfimfiu 55%
, . 7 , 5“,.“me
r , , A us view... M.

. . ., A . . . , , . 4;.

ta

3v. 1.x
?s!czs. 0:
(tux. ta 814-57.,
iuiwfiaxtprufldr

 

Agfi. x...

r
I

 

ilkfiadicfifi
a Sluts-:1 I)
z...

i
u.
a .. u.
..... Swan...
4 h... “51$?
. 5,. ..
.. A , iflécflrmfifi In
. s. 1
V . . . .4. i3. ..
. .l
.. ' JEAN/{tun}
. . 11:5. in. man: .
. 3...“...duel. me)“: u
7 . «1. .

 

. «u I I. r
.l I I c. 1.... be
‘t !

 

-4

2...}. .5. ..

nqummwamm.rfifi¢

....h.)firu "flaw £4.13...”

sudfihurfikfluu?
«t

 

‘21}. wk
xii .0:)..Lr..n:u "Ell! ”F1.
ill-1"“! , I IN}.' ‘
qiiduvifluwnflrn «at? 31.1.1...”

Hr.

 

 

 

metre? *’

.

 

m
m
..

z'.

u xvi

1.3.. |~
a: 13.1!

1.8.5.1...
1: [.1‘4?

.4... 3.1, n. ,
37...}: x7“?!

 

 

 

 

 

 

 

EL"

 

 

 

2052 URRARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

ADVANCEMENT OF BIOBASED PRODUCTS THROUGH
DESIGN, SYNTHESIS AND ENGINEERING OF
BIOPOLYESTERS

presented by

CHISA KANDYDA BROOKES

has been accepted towards fulfillment
of the requirements for the

DOCTORAL degree in CHEMICAL ENGINEERING

 

 

‘ I
CW Our: {\I CLkfi-‘faw

 

\ Major Professor’s SignaflJre
N N :2 7, an? 2L

Date

 

MSU is an affirmative-action, equal-opportunity employer

- A—-—I-'-O-O-O-l-O-O-U-I-o-l-O-I-C-l-‘-'-O-I-'-O-O-I-O-O-I-I-I-I-D-I-I-I-l-i-l-o-

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DAIEDUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/C|RC/DateDue.indd-p.1

ADVANCEMENT OF BIOBASED PRODUCTS THROUGH DESIGN.
SYNTHESIS AND ENGINEERING OF BIOPOLYESTERS

By

CHISA KANDYDA BROOKES

‘A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemical Engineering and Materials Science

2007

ABSTRACT

ADVANCEMENT OF BIOBASED PRODUCTS THROUGH DESIGN, SYNTHESIS AND
ENGINEERING OF BIOPOLYESTERS .

By

Chisa Kandyda Brookes

As crude oil approaches $100/barrel, the need for sustainable products continues to
grow. While development of biobased products is necessary and crucial to
sustainability, integration of biobased materials into biodegradable synthetic-based
materials is another viable route to advancing biobased, biodegradable products. In this
work, glycerol, cellulose acetate (CA) and a glycerol-based pre-polymer are the
materials selected to incorporate biocontent into the synthetic monomer, dimethyl
terephthalate (DMT), and the biodegradable synthetic polymer poly(butylene adipate-co—
terephthalate) (PBAT or Ecoflex). Synthesis, via transesterification of glycerol with DMT,
and reactive extrusion are the processing methods employed. Synthesis products were
characterized by hydroxyl values, molecular weights and thermal properties. Molecular
weight and thermal analysis results were also used for the characterization of extrusion
products. In addition, extrusion products were characterized using dynamic mechanical
analysis and microscopy. The results reveal that there is some interaction between CA
and PBAT, and the two can form a compatible blend allowing for the addition of
biocontent. Plasticization and catalytic effects of these blends were studied.
Incorporation of biocontent via transesterification produced a novel bio pre-polymer that
was a transparent solid. Results indicated that using a dibutyltin (IV) oxide catalyst at
0.1 wt.% of DMT yielded optimal performance. Blends of glycerol-based materials and

PBAT confirmed the reactivity of the new material with PBAT based on residual catalyst.

This one is for my island,
St. Croix USVI

ACKNOWLEDGEMENTS
Glory goes to God for completing another phase of His work within me. Only He
truly understands the experiences We had throughout this process. Thanks to
Dr. Narayanfor taking me on as a student when he had no plans of accepting
any students. I’d also like to thank Dr. Narayan for the freedom with which he
allowed me to think and work. I am even more confident in my abilities, largely
because of the latitude given me by Dr. Narayan. Thanks to Dr. Dale, Dr.
Jayaraman and Dr. Selke for sitting on my committee and providing good
discussion around my work. Thanks to the BMRG (formerly SINAS) group for
the experiences we’ve shared. Special thanks to Sloan, RoyaVDutch Shell,
GEM, MSU Graduate School and MSU College of Engineering for their financial
support during this process. Thanks to the Edwards family for being my closest
friends in MI. I’ll miss leaving Ml only because of you guys. Thanks to my family
and close loved ones (my confidante Ohm and best friend La’Vern), because of
whom I never lacked support. I truly enjoyed my red carpet treatment during this
unique journey. There is no better family or loved ones in the world for me; I am
sure of it! I will forever thank my parents, Edred and Inez Brookes, for believing
in me and in my ability to do great things. Mom and Dad you are indispensable!
Thanks to my brother and sister-in-law, Nigel and Elaine Brookes, my sister and
brother-in-law, Patrice and Kevin Hyatt, and my sisters Bethlyn and Kim Brookes.
You guys are always straight with me, never failing to be there when I need you;
you are all priceless. I love you. Unlimited thanks goes to all who have played a

role in making me stronger and wiser. Finally, thank God I am done!

Table of Contents

List of Tables ...................................................................................... vi
List of Figures ................................................................................... viii
Chapter1 ....; ........................................................................................................ 1
1.1 Motivation .......... . ........................................................................................ 1
1.2 Goals and Objectives ................................................................................. 3
Chapter 2 .................................................................................................... 5
2.1 Literature Review ....................................................................................... 5
2.2 Theory of Polymer Mixtures ..................................................................... 14
2.3 Polyester Synthesis ................................................................................. 33
Chapter 3 ........................................................................................................... 45
3.1 Materials .................................................................................................. 45
3.2 Processing and Characterization Methods .............................................. 53
Chapter 4 ........................................................................................................... 62
4.1 Overview .................................................................................................. 62
4.2 Materials & Methods ................................................................................ 65
4.3 Results and Discussion ............................................................................ 66
4.4 Summary ................................................................................................. 95
Chapter 5 ........................................................................................................... 98
5.1 Overview .................................................................................................. 98
5.2 Materials and Methods ........................................................................... 100
5.3 Results and Discussion .......................................................................... 100
5.4 Summary ............................................................................................... 115
Chapter 6 ......................................................................................................... 117
6.1 Overview ................................................................................................ 117
6.2 Materials & Methods .............................................................................. 117
6.3 Results and Discussion .......................................................................... 119
6.4 Summary ............................................................................................... 127
Chapter 7 ......................................................................................................... 128
7.1 Conclusions ........................................................................................... 128
7.2 Recommendations for Future Work ....................................................... 131
References ....................................................................................................... 1 33

Table 1:
Table 2:
Table 3:
Table 4:
Table 5:,
Table 6:

Table 7:

Table 8:

Table 9:

Table 10:

Table 1 1:

Table 12:

Table 13:

Table 14:

Table 15:

Table 16:

List of Tables

Eastman Cellulose Acetate Properties .......................................... 47
BASF PBAT (also known as Ecoflex) Properties ............................ 49
Properties of Glycerol ............................................................... 50

Calculated solubilities compared to literature using Hoy parameters...64
Extruded Blends of CA and PBAT without modification ..................... 67
Preliminary CAPBAT degradation values as retrieved by TGA ............ 68

Summary of DSC thermal properties for preliminary CAPBAT
blends .................................................................................... 71

Molecular weight data for “unplasticized” and plasticized CA-PBAT
samples showing actual and theoretical Mnmix and Mwmix values and
their % difference values ........................................................... 74

Degradation temperatures for "unplasticized" and plasticized
blends .................................................................................... 75

DSC results for ”unplasticized" and plasticized blends with varying
plasticized blend compositions; * indicates values from
DMA .................................................................................... 78

GPC average molecular weight data of uncatalyzed and catalyzed
blends .................................................................................. 85

TGA degradation temperatures of uncatalyzed and catalyzed
blends ................................................................................. 86

DSC transitions for uncatalyzed and catalyzed blends; * indicates
values from DMA ................................................................... 89

Comparison of percent conversion and Xn based on hydroxyl values
versus methanol formation ...................................................... 104

Comparison of percent conversion and Xn based on hydroxyl values
versus methanol formation after adjustments for trapped
methanol ............................................................................ 108

GPC Molecular weights and polydispersity index for all
reactions ............................................................................. 110

vi

Table 17:
Table 18:
Table 19:
Table 20:

Table 21:

TGA degradation temperature values for both catalyst sets .......... 111
DSC glass transition temperatures of all reaction samples ............ 114
Summary of Runs for GDMT-PBAT blends ............................... 119
TGA degradation temperature for GDMT-PBAT samples ............. 120

DSC thermal transitions for blends of GDMT—PBAT; * indicates values
obtained from
DMA............; ...................................................................... 122

vii

Figure 1:
Figure 2:

Figure 3:

Figure 4:

Figure 5:

Figure 6:

Figure 7:

Figure 8:

Figure 9:

Figure 10:

Figure 11:

Figure 12:

Figure 13:

Figure 14:

Figure 15:

Figure 16:

List of Figures

Cellulose as polymer of B-D-glucose ............................................. 6
Portion of a Cellulose Diacetate Chain .......................................... 7
Graphical depiction of Equation 9 in regards to the Property of a
Polymer Blend ........................................................................ 21
Free Energy of Mixing for Polymers A & B with the same molecular
weight .................................................................................. 25
Phase and Transition Behaviors in Polymer Blends ....................... 27
Phase Diagrams for Polymer Mixtures showing the binodal (solid) and
spinodal (dashed) curves along with critical points ......................... 29
Esterification reactions using Maleic Acid and Maleic Anhydride as
examples and Transesterification reaction using Dimethyl
terephthalate ......................................................................... 41
Structure of BASF Po|y(butylene adipate-co-terephthalate) ............. 48
Biodegradation of Cellulose Compared to that of PBAT .................. 48
Structure of Glycerol .............................................................. 49
Biodiesel production chemical reaction showing Glycerol as a by-
product .............................................................................. 50
Important External Plasticizers for Cellulose Acetate .................... 64
Weight loss and derivative weight loss (represented by two maximas)
curve of CAPBAT sample to illustrate the two component blend of CA
and PBAT ............................................................................ 69
Weight loss curves of preliminary CAPBAT samples ..................... 70
DSC crystallization temperatures of preliminary CAPBAT
blends ................................................................................. 72
Chromatogram of samples affected by varying contents of plasticizer

and cellulose acetate: CP37 (0% plasticizer, CTP37 (9% plasticizer),
CTP55 (15%
plasticizer) ........................................................................... 74

viii

Figure 17:

Figure 18:

Figure 19:

Figure 20:

Figure 21:

Figure 22:

Figure 23:
Figure 24:

Figure 25:

Figure 26:

Figure 27:
Figure 28:

Figure 29:

Figure 30

Figure 31

Figure 32

TGA weight loss curves for "unplasticized" and plasticized
samples .............................................................................. 76

TGA derivative weight loss curves comparing plasticized
compositions ....................................................................... 77

DMA storage modulus of “unplasticized" and plasticized blends along
with varying compositions of plasticized blends ............................ 80

DMA loss modulus of "unplasticized” and plasticized blends along
with varying compositions of plasticized blends ........................... 81

DMA tan delta of "unplasticized" and plasticized blends along with
varying compositions of plasticized blends ................................. 82

ESEM images of CP37, CTP37 and CTP55 by column from left to
right, at 1000x, 2000x, 5000x by row from top to bottom. Values
displayed at the bottom represent particle sizes for each

sample ................................................................................ 84
GPC chromatogram of uncatalyzed and catalyzed blends .............. 86
TGA weight loss curve of uncatalyzed and catalyzed blends .......... 87

TGA derivative weight loss curve comparing compositions of
catalyzed blends .................................................................. 88

DMA storage modulus curves for uncatalyzed and catalyzed
blends ................................................................................. 91

DMA loss modulus curves for uncatalyzed and catalyzed blends....92
DMA tan delta curves for uncatalyzed and catalyzed blends...........93
ESEM images of CTP37, CTPD37, CTP55 and CTPDSS by column
from left to right, at 1000x, 2000x, 5000x by row from top to bottom.

Values displayed at the bottom represent particle sizes for each
sample ............................................................................... 95

: Degree of polymerization and gel point for a 1:1 and 2:3 mixture of

glycerolzdimethyl terephthalate ................................................. 99
: a. Reaction conditions b. Reaction Matrix .................................. 101
: Conversion over time based on recorded methanol collection ........ 102

Figure 33:

Figure 34:

Figure 35:

Figure 36:

Figure 37:

Figure 38:

Figure 39:

Figure 40:

Figure 41:

Figure 42:

Figure 43:

Figure 44:

Figure 45:

Degree of polymerization as calculated by Carothers’ Theory for the
theoretical curve and by experimental data for all other curves ...... 103

TGA derivative weight loss curve for NC samples showing % trapped
methanol in the samples ........................................................ 106

TGA derivative weight loss curve for DBTO samples showing %
trapped methanol in the samples .............................................. 107

Catalyst perfomance based on MEOH for different types of catalysts
at different concentrations ........................................................ 108

Catalyst perfomance based on hydroxyl # for different types of
catalysts at different concentrations ......................................... 109

Chromatogram for all reactions with different catalysts and different
catalyst amounts ................................................................... 1 10

Sodium Carbonate percent and derivative weight loss curves based
on different catalyst amounts .................................................. 112

Dibutyltin oxide percent and derivative weight loss curves based on
different catalyst amountsl13

DSC glass transition temperature results for reaction products with
varying amounts of a. Sodium Carbonate and b. Dibutyltin

oxide .................................................................................. 114
TGA weight loss curve for GDMT-PBAT samples ........................ 121

DSC crystallization transition for blends of GDMT-PBAT ............... 124
Storage modulus of GDMT-PBAT blends ......................... . .......... 125
DMA tan dela curve for GDMT-PBAT blends .............................. 126

Chapter 1

 

Introduction

 

1.1 Motivation

Sustainability is no longer just a “buzz” word or a vision of the future.
Critical events, with the 2006 State of the Union Address being one, have
contributed to moving sustainability from abstract to more of a growing reality.
lnstabilities in areas of the world where oil is present have been a driving force
for the increasing oil prices over the past few years. This dwindling access to
foreign oil is what has motivated government and industry to support the
advancement of bio-based research [1], aiding in reducing our reliance on these
sources of energy and raw materials. Bio-based research has managed to take
center-stage, being the topic of discussion for major political debates reviewed in
articles of respected trade journals. One example is a Government and Policy
article by Jeff Johnson in Chemical and Engineering News’ first issue of the year
2007 titled “Ethanol—Is it Worth it?”[2]. While the article highlights pro and con
positions on the viability of corn as a source of fuel, using views from leading
researchers like Dale and Pimental, it manages to end with the fact that the
discussion on corn-based fuel is “likely” to be the start of a bio-fuels market, and
not its end [2].

Not only is it a start for the bio-fuels market, but it is a start for the bio-
based plastics market as well. Starch has been the primary feedstock [3], and

one of the most successful and promising naturally occurring materials in the bio-

based industry. Researchers at Michigan State University (MSU) along with
KTM Industries have developed a one step, reactive extrusion process using
water as a plasticizer and blowing agent for starch, to manufacture bio-based
(non-toxic), biodegradable foam for packaging [4]. More recently from a USDA
News release, electro-active biopolymers made from cornstarch have been
developed for potential use in charging lithium batteries [5].

Besides political drivers there are other fundamental motivations for
“green” development. Environmental and social drivers, while less accepted, an
important the motivation for advancement of biobased materials. The Kyoto
Protocol [6], a world attempt to address concerns about the environment, has
committed countries to decreasing greenhouse gases that contribute to climate
change within the environment. This world effort to protect the environment has
also led to the development of biomaterials research. In a paper published by
the Journal of Industrial Ecology, Duncan mentions the enactment of Title III of
the Agricultural Risk Protection Act (Public Law 106-224), the Biomass Research
and Development Act of 2000 [7]. This act is in place to increase coordination
across departments in the federal government in relation to biomass research
and development [7]. This effort supports the conversion of biomass into bio-
based products. Growth of a biobased industry can then affect the socio-
economic conditions of farmers, and aid in the development of rural areas.

Amidst all the motivators for and excitement around bio-based products,
there exists the need for a certain quality of products—those that excel in

performance while being economically feasible, as is characteristic of synthetic

polymers. Narayan fairly describes the several benefits of oil-based polymeric
materials after which he rightly identifies that “it is these very attributes of
strength and indestructibility that cause problems when these materials enter the
waste stream” [8]. Thus, in addition to performance and economics, the type of
product for which there is a need is one that is not only bio-based, but one that
also has the ability to be bio-degraded after its intended use. Narayan’s ideas
identify with the overall goal of this study in that he recognizes that there is great
value, as well as large areas of improvement, for existing products as shown in
the LCA of polyethylene and starch foams [9]. The successful integration of bio-
based products to the market requires support of continuous improvement on

new and existing products.

1.2 Goals and Objectives

The intent of this research is to incorporate the biobased materials,
glycerol and cellulose acetate, into a synthetic based monomer and polymer,
dimethyl terephthalate (DMT) and biodegradable poly(butylene adipate-co-
terephthalate) (PBAT). Incorporating biocontent in this fashion will advance
bioproduct development and broaden the utility of biomaterials. Biobased
content will be incorporated using two routes, synthesis and reactive extrusion.
Three objectives were set to achieve the goals of this research.

1. Investigate effects of composition, plasticizer and catalyst on reactive

extruded blends of CA and PBAT

2. Determine catalyst performance in the synthesis of glycerol-based
biopolyester (GDMT)
3. Study residual catalyst performance on blends of GDMT with PBAT
The benefit from each of these objectives is the infusion of bio-based
materials into synthetic base products for the development of new-age products
to address current problems. This research is timely in that the growing demand
for biodiesel has brought along tremendous potential opportunities for these new

glycerol-based esters, since glycerol is a waste of biodiesel production.

Chapter 2

 

Background

 

The aim of this chapter is to summarize the literature related to the
overarching goals of this study and discuss the theory behind polymer blends
and polyester synthesis with a brief overview of the chemistry related to polyester

synthesis.

2.1 Literature Review

An overview of CA’s origin and production is given, followed by a
discussion of various CA blends documented in the literature. Cellulose acetate
and its blends are the main focus of this literature review with minimal discussion
of PBAT, given the small amount of work published regarding this relatively new
polymer. The CA blends discussion centers on three categories, which include
(1) miscible and immiscible blends, (2) effects of external and internal plasticizers

on blends, and (3) biodegradation of blends.

2.1.1 Cellulose Acetate Background

Cellulose acetate (CA) is a derivative of the natural polysaccharide
cellulose. Cellulose, whose structure is shown in Figure 1, is mainly found in
plant cell walls and is the most abundant carbohydrate found in plants, followed
by starch [11]. While Cellulose has been an important starting material in plastics

for some time, cellulose-based plastics have lost much of their market to cheaper

plastics like polyethylene and polypropylene [12]. As oil resources have

continued to decrease with increasing

0” OH .
. HO O"
O OH

Figure 1. Cellulose as polymer of B-D-glucose
[10]

 

 

 

 

 

 

 

 

potential to become extremely expensive, cellulose has once again gained
interest as a renewable, abundant alternative raw material source for plastics.
One major class of cellulose derivatives is cellulose esters. Esterification
of cellulose can be achieved with acids, with anhydrides, or with acid chlorides
[13]. The most widely used method to convert cellulose to cellulose esters is via
anhydrides [11]. To get to cellulose acetate (the material of interest in this
study), cellulose goes through esterification to a cellulose triester via acetic
anhydride. Esterification is then followed by hydrolysis with water or a diluted
acetic acid solution to regain some of the OH-groups that were esterified.
Separation then occurs via filtration, precipitation, washing, dewatering, drying
and screening, yielding cellulose acetate of a certain degree of substitution (DS)

[12]. Performing hydrolysis with a water-acid mix that incorporates the acid of

choice (ex. propionic anhydride or butyric anhydride) produces other cellulose
acetate derivatives like cellulose acetate butyrate or cellulose acetate propionate.

The structure of CA, as shown in Figure 2, discloses important
characteristics about its properties. Cellulose acetate has varying degrees of
substitution from zero to three. The degree of substitution is based on the
average number of acetyl groups that replace the OH-groups on an

anyhdroglucose unit [11]. The degree of substitution is

 

CHQOH
CHzoAC

0
0A6 oAc 0 Ac ofio
OH OH 0
OAc

 

 

 

 

Figure 2. Portion of a Cellulose Diacetate Chain [14]

 

 

 

often measured by the % acetyl content. The acetyl content or degree of
substitution often dictates the solubility characteristics of cellulose acetate [13].
With increasing acetyl content, the permeability rate increases while the tensile
strength and elongation characteristics of cellulose acetate decrease[13]. This is
because the strong hydrogen bonding originally occurring in cellulose is
disrupted, and with the greater acetyl content the chains are not likely to have as
strong an interaction (hydrogen bonding and van der Waals forces) as in
cellulose. For situations in which the other acids or anhydrides are used to

produce cellulose acetate derivatives such as cellulose acetate propionate and

cellulose acetate butyrate, the changes to the overall structure of cellulose
acetate causes even more reduction in tensile and thermal properties as

compared to cellulose acetate.

2.1.2 Cellulose Acetate Blends

Cellulose acetate blends provide a mechanism to achieve certain desired
characteristics. K. J. Edgar and others did a review of the literature on blends of
cellulose esters with other polymers [15]. Many of their findings for cellulose

ester blends are included within this literature search.

Misci_b|e anflnmiscibleflngg

Much of the work reviewed by Edgar and others regarding cellulose
acetate blends, as well as additional studies, has been done with cellulose
acetate propionate and cellulose acetate butyrate, probably due to easier
processing as compared to cellulose acetate. Nonetheless, work done by
Schauber et al. and Light at al. showed blends of CAP and CAB, respectively,
capable of forming miscible and immiscible blends with certain copolymers [15,
16, 17]. Specifically, Schauber’s composition, which yielded a miscible blend of
CAP and a copolymer of propylene and carbon monoxide, decreased the
toughness when compared to CAP alone. On the other hand, Light’s
composition which yielded miscible blends of CAB and copolymers of methyl and
butyl acrylate, increased toughness as compared to CAB alone. Both

Schauber’s and Light’s immiscible blend compositions of CAP and CAB with their

respective copolymers yielded significant improvement in toughness as
compared to CAP or CAB alone [15, 16, 17]. Nabar et al. attempted to blend CA
and polyethylene terephthalate in order to improve the crystallization rate of PET
for better moldability, but the blend was immiscible and hard to process [18]. The
difficulty of processing was attributed to the low processing temperatures used to

avoid CA degradation.

External Plasticizers for CA and its Blends

Due to the high processing temperature of cellulose esters, plasticizers
often must be added to lower the processing temperature [11]. In 1995, Ghiya
and others processed CA with triethyl citrate and acetyl triethyl citrate and
reported the miscibility of these plasticizers with CA along with the effects on
strength and elongation [19]. The addition of these citrate esters increased the
biodegradation rate of CA. More recently, Mohanty and others used
triethylcitrate (TEC) as a plasticizer for cellulose acetate biocomposite
applications, and studied the processing methods of (1) compression molding,
(2) extrusion followed by compression molding and, finally, (3) extrusion followed
by injection molding [20]. They concluded that high shear in extrusion and
injection molding processes caused better mixing or may have enhanced
crosslinking, resulting in increased tensile properties. They propose that 30%
plasticizer is a “suitable balance for biocomposite materials” [20]. Wibowo et al
[21] and Park et al [22] also used TEC as a plasticizer in a blend of CA with

layered silicate nanocomposites, showing increased tensile and flexural strength

in both. Rosa’s addition of polyethylene glycol to a CA/polycaprolactone blend
resulted in increased elongation for CA/PCL blends, and for blends with more

than 60% CA increased tensile strength [23].

Internal Plastgzers for CA and its Bflldj

Others have also attempted to get the plasticizing effect without actually
using a liquid plasticizer. Nie and Narayan grafted styrene maleic anhydride
(SMA) to cellulose acetate with the help of the common plasticizer diethyl
phthalate and they reported that an increase in the maleic anhydride content of
SMA helped to increase the grafting of the SMA to CA, thus reducing the
immiscibility between the two, resulting in improved dimensional stability for the
blend [24]. Light and others as well as Lee et al. attempted to get a plasticizing
effect from CAB and CAP by using other polymers via polymeric plasticization
[15, 17, 25]. In Light’s case, there was an improvement in the impact strength
and softening temperature for both CAB and CAP as compared to other normally
plasticized cellulose esters when grafted polymers were blended with CAB and
CAP [15, 17]. For Lee and others, the addition of biodegradable
polyestercarbonates (PEC) did not improve the tensile strength [15, 25].
Crystallization of PEG from the blend above 40% composition was said to be the
cause of poor tensile properties. However, below 40%, PEG proved to enhance
the elongation and the melt flow of the blend. Edgar highlights that this work

supports the idea of biodegradable polymeric plasticizers [15].

10

Biodegradable Blends of Cellgjose Acetate

In addition to improving properties, CA blends have also been formulated
to biodegrade. While Maheras and others have blended CA and starch acetate
for cost purposes [15, 26], cellulose and starch blends are also attractive
because of the biodegradation potential of each [27]. Starch is cheaper than
cellulose. Also, like cellulose, it is very abundant and it comes from renewable
resources. While starch acetate does decrease the tensile properties, and is not
miscible at a 25% composition with cellulose acetate, it remains as a good
addition in the mixture to help reduce cost and enhance biodegradation [26].
Biodegradation of cellulose acetate is strongly dependent on the degree of
substitution of the cellulose acetate. As the degree of substitution (DS)
increases, the biodegradation, in terms of 002 and CH4 formation, decreases
[15]. Cellulose acetate with a 08 of 2.5 is proven to be biodegradable by
exposing the residues of the acetyls on cellulose acetate to cellulose enzymes
[27]. Guruprasad and Shashidhara have confirmed more recently the
degradation of a 90/10, CA/starch blend. Weight loss was over 35% in three
months, with a little over half of the weight loss due to starch [28].

Other researchers focused specifically on biodegradable blends with
internal plasticizers for CA. In a series of publications, Teramoto and Nishio
attempted to use polylactic acid (PLA) as an internal plasticizer for CA with a D8
of 2.15 [29,30,31]. This form of PLA acted as an internal plasticizer and did
reduce the T9 of the blend. It is important to mention that to get this plasticization

of CA required an extensive 3-stage batch process to graft PLA onto CA for

11

composition-controlled grafting. Braganca and Rosa also used an internal
plasticizer [32]. In their CA/PCL blends they noted the poor access to CA
hydroxyl groups as compared to the hydroxyl groups on CAB. Only one blend of
PCUCA (20/80) showed some compatibility and increased tensile strength. With
this increase in tensile strength came a 99% drop in elongation at yield as
compared to PCL alone. Prior to Teramoto’s work, Yoshioka, and Shiraishi had
studied the plasticization of CA for several years [33,34,35]. They attempted to
graft caprolactone and lactide onto CA using a stannous octoate selective
catalyst and moderate temperatures. With just the right reaction conditions, they
were able to add strength as well as elasticity to CA. They noted that the lactide
component grafted to CA more readily than did the caprolactone, which resulted
in brittle blends after short times (from lactide) and more flexible blends after
longer times (from caprolactone) [33].

Their studies continued and led to the grafting of varying molar
substitution combinations of caprolactone and poly(lactic acid) (PLA) CA [34].
This eliminated the lactide intermediate sensitive to moisture and resulted in
better moldability of CA, attributed to preferential grafting of the more flexible
caprolactone. Yoshioka, Shirashi and others also undertook grafting of dibasic
anhydrides and mono epoxides to CA [35]. In this work sodium carbonate was
used as an esterification catalyst that aided in increasing the grafting of
oligomers to CA. Maleic anhydride and phenyl glycidyl ether proved to be the

best pair due to the high reactivity of each in their respective classes. This pair

12

also took longer to biodegrade than succinic anhydride and glycidyl methacrylate,
but biodegraded nonetheless.

Building from the successful plasticization and biodegradation of CA via
maleic anhydride and epoxides, Lee and Shiraishi studied plasticization of CA
using maleic anhydride, glycerol and citrate esters [36]. While the maleic
anhydride glycerol mixtures with CA via a kneading process exhibited brittleness,
addition of a small amount of citrate ester effectively plasticized CA, reducing the
T9 of CA by 80-100 “’0. Lee and Shiraishi claim the mechanical properties
obtained were “comparable to those of general synthetic polymers or
commercially plasticized cellulose acetate”. These results are in line with
Suvorova and Demchick’s finding that ester plasticizers with shorter chains and

polar groups mix well with CA [37].

2.1.3 Poly(butylene adipate-co-terephthalate) (PBA T) Blends

Blend literature on PBAT is not as extensive as CA blend literature. Over
the past few years there have been some studies on the polymer blends of
PBAT. Someya et al. have looked at adding nanocomposites to blends of PBAT
with montmorillonite clay. The addition of the clay to PBAT was reported to slow
the crystallization of PBAT [38]. Nabar et al. utilized PBAT as a functional aid to
improve the properties of starch foams [39]. His later work included an extensive
study on the maleation of PBAT for use as a compatibilizer in starch foams [40].
Results showed improved resilience and hydrophobic properties of starch foams.

Most recently a study of PBAT-PLA blends reported significant increases in

13

elongation and toughness of PLA with the addition of 5 — 20% PBAT, but at the

expense of tensile strength and modulus [41].

2.2 Theory of Polymer Mixtures

This segment focuses on the thermodynamics of polymer mixtures. It begins
by covering the mixing of regular solutions and continues with the mixing of
polymers. Other topics related to polymer blends are also covered. Specifically,
miscibility, compatibility, phase separation and adhesion/interfacial tension are

discussed briefly for a more fundamental understanding of blends.

2.2.1 Thermodynamics of Mixtures Overview

An understanding of polymer blends requires some knowledge of the
thermodynamics of mixtures. Thermodynamic theories center around the idea
that blends of two components are governed by the free energy equation given

below.

AGmix = AHmix - TASmix Equation (1)

In this equation, AGmix is the free energy of mixing, AH...“x is the enthalpy of
mixing, ASmix is the entropy of mixing and T is the absolute temperature of the
system. In order for the components to mix, a necessary condition is that AGmix

must be negative (<0) [42]. In order for this to occur the disorder of the system

during mixing, represented by the entropy term, or the temperature of the system

14

should be relatively high. Another way to enhance mixing is to minimize the heat
related to intermolecular interactions, represented by the enthalpy term.

For ideal solutions and for athermal mixing, the interactions amongst
molecules are equal, yielding no change in volume and no heat exchanged
(AHm,x = 0) [42]. Thus the entropic portion of the free energy equation governs
the behavior of AGmix under these conditions. To determine the entropy of such
a two-component system, statistical methods are employed to relate to the mole
fractions of the species shown by equation 2 where xi is the mole fraction of a
particular species and n, is the moles of a particular species and R is the gas

constant [42].

ASmix = —R[n1 ln(x1) + n21n(x2)] Equation (2)

Thus the change in free energy (AGm) would be equal to equation 3 where all

variables are defined as before.

AGmix = RT[n1 ln(x1) + n21n(x2)] Equation (3)

Regular solutions where the change in enthalpy is not zero have to take
into account intermolecular interactions [42]. For interactions that are not strong,

or intermolecular forces such as van der waaI forces, AHm,x can be defined by the

15

equation below. The total volume of the system is V, c) is the volume fraction of

each component and 5 is the solubility parameter for each component. All other

variables are as defined earlier.
AH mix / V = (<01)(¢2)[51 - 5212 Equation (4)

The solubility parameter is the square root of the cohesive energy density.
The cohesive energy density measures the strength of secondary bonds, thereby
giving insight on specific interactions related to the enthalpy term [43]. The
cohesive energy density is given by equation 5, where AEV is the energy of

vaporization and Vmola, is the molar volume of the specific liquid.

CED = “5% Equation (5)
molar

Thus the solubility parameter incorporates all the internal molecular energies
associated with the contacts between the polymer and solvent. There are
different methods for estimating solubility parameters. Hoy, Van Krevelen and
others have developed group contribution methods by which a non-polar
polymer’s solubility can be approximated within an acceptable range [44]. Using
the group contribution method, the solubility of a polymer with mostly non-polar
species in the repeat unit can be estimated by equation 6. In equation 6, p is the

density, F, is the sum of the molar attraction

l6

6 = p(ZF,-)/M Equation (6)

constants and M is the molecular weight of the polymer repeat unit. While such
solubility parameters are simple and convenient because they can be used to
predict the mixing of two components, they do not take into account strong polar
or hydrogen bonding, causing inaccuracy [44]. This led to the development of

the Flory-Huggins theory.

For polymer-solution systems, the ideal solution often is not a good
estimation because of the difference in sizes of the molecules. In addition, there
is solvent-solvent interaction, polymer-polymer interaction and solvent-polymer
interaction. The interaction between the polymer and the solvent requires some
breaking of solvent-solvent and polymer-polymer contacts. These breaks and
formations of contacts affect the enthalpy portion of the free energy equation.
Thus athermal mixing does not occur. Flory and Huggins developed a theory to
predict the change in free energy for polymer solutions. The relationship
developed by Flory and Huggins includes lattice interaction parameters and the
Gibbs free energy change between the polymer and solvent. Due to the
inaccessibility of these parameters, the Flory-Huggins interaction parameter, x,
was defined to give the following enthalpic and free energy relationship for

polymer-solvent mixtures (Equations 7 and 8).

AHmix = RTZAB(¢2)"1xl Equation (7)

l7

A0me = RTIn1ln(¢1)+ "2 1“((42) + rum/l3] Equation (3)

All variables are the same except that x1 is representative of the degree of
polymerization for polymeric solvents and is equal to one for monomeric
solvents. In the second equation, the first two terms are the entropic portion of
the free energy equation and the last term represents the enthalpic portion.
While the relationship developed by Flory-Huggins has its drawbacks, dilute

polymer solution mixes can be estimated adequately [44].

Flory-Huggins polymer-solution theory and the regular solution theory
which incorporates solubility parameters are different in that Flory-Huggins uses
an interaction parameter, x, that relates to the interaction of components in a
binary mixture to each other and themselves [44]. The solubility parameter has
to be calculated for each component. In addition, one solubility parameter does
not account for all interactions between both components. Also, regular solution
theory uses mole fractions to determine the changing entropy of the mixture,
whereas the Flory-Huggins polymer-solution model uses volume fractions to
estimate the changing entropy of the mixture. Although the latter model proved
satisfactory for polymer-solvent mixtures, polymer-polymer mixtures are more
affected by molecular weight. Neither model completely satisfies and accurately
describes all polymer-polymer systems, but they are the principles that govern
the system and should be understood before tackling other models that may
better predict the system at hand. This understanding of the thermodynamics

involved is crucial to understanding the polymer blend system.

18

2.2.2 Polymer-Polymer Miscibility/Compatibility

In the discussion of polymer blends, miscibility and compatibility must be
addressed. A blend being immiscible does not reflect the compatibility of the
blend. A compatible polymer blend is a blend that exhibits macroscopically
uniform physical properties throughout its whole volume [42]. Miscibility is the
ability of a mixture to form a single phase over specific ranges of temperature,
pressure and composition [42]. Complete miscibility requires the proper type of
interactions, the proper number of interactions and the proper spacing of
interactions [42]. In more practical terms, miscibility is often indicated by 1)
transparent films displaying homogeneity even under magnification, 2) a single
glass transition temperature (T 9) between the Tg values of the blend components,
and 3) good mechanical integrity, especially in tensile strength [42]. Often
associated with miscible blends is the synergistic increase in tensile strength,
which is attributed to strong specific interactions leading to better packing on a

molecular level [42].

Unfortunately, some incompatible blends exhibit these properties.
Incompatible blends form transparent films due to similar refractive indices or due
to the formation of two separate transparent films layered one on the other [42].
Another way that incompatible blends form transparent films is by the
impediment of phase separation due to potential entanglement and/or the high
viscosity of the polymer-polymer matrix [42]. If this small change in free energy
were to occur it would lead to separation of the polymer-polymer matrix and films

would not be transparent. Other issues that can cause incompatible blends to

19

demonstrate miscible behavior may be via an experimental technique. For
example, it is well known that differential scanning calorimetry results may show
one Tg (due to the other one not being as visible), while dynamic mechanical
analysis results may show two because it is more sensitive to secondary
transitions. These techniques help to identify immiscibility more so than
miscibility [42]. Compatibility as defined in this work, on the other hand, relies

more heavily on the properties achieved.

Regardless of the drawbacks, miscibility and compatibility is often
assumed if a blend forms a film with optical clarity and has a single glass
transition temperature. Equations have been developed relating the composition
of the blend and T9 values of the homopolymers to predict the T9 of the blend as
shown in Equation 9, Fox’s Relationship. In this equation, T9 is the glass

transition temperature and w, is the weight fraction [46].

i = i + 3 Equation (9)
T3 T81 T82

While this is characteristic of a miscible mixture, thermodynamics gives a
necessary condition that for a two-component mixture with a homogenous single
phase the change in the Gibbs free energy of mixing should be negative. Thus
additional analysis of the resulting product (ex. mechanical testing) would be

necessary in confirming, at the very least, the compatibility of the blend.

20

Kienzle [47] gives a mathematical description of polymer mixture

properties. In Equation 10, P1 is a property value and C1 is the concentration of

P = RC1 + P2C2 + IRPZ Equation (10)

component 1, while I is the interaction coefficient. Figure 3 gives a graphical
representation of the relationship being described in Equation 10. If I < 1, the
property value is at its lowest at all blend compositions and the blend is

considered incompatible. If I = 0, the property value varies linearly with

 

 

Alloy (l>1)

  
      

.\\\

\ Additive result (l=0)

  

Property

Incompatible blend
(I<1)

 

 

l

0% Concentration of Polymer 1 100%

 

 

 

 

Figure 3. Graphical depiction of Equation 9 in
regards to the Property of a Polymer Blend [47]

 

 

 

composition and exhibits higher property values than the incompatible blend. If I

> 0, the property value is at its highest throughout the composition range. This

21

relationship provides another way to distinguish incompatible blends from

compatible blends from compatible, potentially miscible blends.

Again, a miscible mixture is one in which the free energy of mixing is
negative (<0) [42]. Fortunately, Flory-Huggins development inspired Scott and
Tompa to apply the model to polymer-polymer systems. The free energy of a
polymer-polymer system is given by Equation 11, where Vr is the reference
volume close to the molar volume of the smallest repeat unit in a polymer, and x,
is again defined as the degree of polymerization for each polymer. All other

variables are defined as before. The AG...“x being less than

 

A0"... = <va ){(¢—A)In ¢A + (3%) In ¢B + IAN/1193)} Equation (11)

r xA

zero is known as a necessary, but not sufficient criteria of miscibility because
some incompatible systems also show a negative free energy [42]. A necessary
and sufficient criteria of miscibility is that the second and third derivatives of free
energy with respect to composition are equal to zero, as shown in Equation 12

[42].

2 . 3 ,
w : 199M 2 0 Equation (12)

3(2)}, 349.34

22

Setting both the second and third derivatives equal to zero results in the

derivation of the critical conditions found in Equations 13-15. In these equations

 

l l l .
(ZAB)cr = '" (Ti) + (7,7) Equat'on (13)
2 “A x3
.4” .
(¢i)cr = W Equatlon (14)
(Z ) = ~1- 1 + 1 Equation (15)
”B s” 2 mum, x3(¢3)sp

all variables are as defined earlier, with xi continuing to be the degree of
polymerization for component i. From the critical conditions it is apparent that for
very large degrees of polymerization, the critical interaction parameter chi would
be very small [42]. This makes sense because for polymers with large

segments, more polymer A — polymer B interactions will take place more readily
versus if there were fewer segments. This minimizes the enthalpy of interaction
and thus decreases the critical interaction parameter. Also, for extremely large
degrees of polymerization the critical compositions will be very small [42]. This
predicts that even at very low compositions, very large polymers are incompatible

and separate.

Molecular weight of the individual polymer components can affect the

separation of the polymer blend [42]. Assuming the same molecular weight for

23

each component in order to simplify the relationship, AGmix for a polymer-polymer
mixture would be given by Equation 16. In Equation 16, Mcr is the critical
molecular weight at which the second derivative of free energy with respect to
composition is no longer greater than zero for all compositions, p is the density,

M; is the molecular weight of component i. The first

AGmix = p13“, {A137 I¢A1n ¢A + (1- ¢A)1n(l - ¢A>I+ 2¢A(1- (0,4)} Equation (15)
CT

 

two terms, then the last, are the entropic and enthalpic portions, respectively. It
can be seen that for M less than M0,, the entropic term dominates and Equation
16 is fulfilled, whereas for M greater than Mcr, the enthalpic term begins to
dominate in comparison to the entropic term. A graphical representation of AGmix
versus volume fraction can be seen in Figure 4 for equal molecular weight
polymer blends. As can be seen, free energy becomes more positive with
decreasing M as compared to the Mcr. What is also interesting is that as free
energy becomes more positive, polymer mixtures with large amounts of each
component can phase separate even if the free energy is still negative. The

mixture only maintains one phase at extreme compositions.

24

 

 

 

 

 

 

 

AGmlx (arbitrary units/mole)

 

 

 

 

 

Volume Fraction

 

 

 

Figure 4. Free Energy of Mixing for Polymers A & B
with the same molecular weight [37]

 

 

 

One way to enhance polymer-polymer miscibility at all compositions is to
decrease the enthalpy. For endothermic systems, the lowest value of the
enthalpic term is zero. For exothermic systems, however, enthalpy can be
negative if certain specific interactions occur between polar groups within the
mixture. In these systems, AG...“x will be negative despite the low entropy term

caused by the large molecular weights of polymers. It is important to note that

25

even though there are specific interactions that support a negative enthalpy due
to polarity, parts of the polymer chain will not be polar and will therefore cause a

dispersive component of enthalpy to arise as seen in Equation

AHmix : Aandijfpersive) + AI_I’(,fi~pjrecificint eractions) Equation (17)

17. A moderate number of strong, well-dispersed interactions can help to

maintain a negative enthalpy and avoid phase separation.

To help in predicting the behavior of polymer blend mixtures, molecular
weight averages can be calculated for the blends using Equation 18 and 19

below. In these equations, w, is the weight fraction of component I, (Mn)mix

(XII—"L,“ = -—1— Equation (18)
Elm; I

MW . = wi* Hw- Equation (19)

I I»... Z I I.

corresponds to the Mn of the blend mixture and (Mn): corresponds to the Mn of the

corresponding component. The same is true for (M,,,,)mix and (Mw)i in equation 19.

26

2.2.3 Phase Separation

While the intent is not to develop a phase diagram, understanding phase
transitions and separations is a key to identifying immiscible or incompatible
blends, and to understanding phase behavior in relation to temperature and
changes in compositions. Phase diagrams for polymeric systems usually focus
on the solid-liquid transition as shown in Figure 5 where the polymer goes from
glassy to rubbery to the melt phase in the solid-liquid transition diagram [42]. In

the liquid-liquid phase diagram shown, the effect of temperature on the blend at

 

Liquid-Liquid ‘ Solid-Liquid
Phase Diagram Transition Behavior

LCST

 

 

 

 

 

 

(D V
it
‘3 h .
8. 1 p ase melt + crystal
E
a)
|— UCST T9

2 Phases \

glass + crystal
(a) (b)
(I (ll

 

 

 

 

Figure 5. Phase and Transition Behaviors in Polymer
Blends [42]

 

 

 

various composition causes the blend to either combine as one phase or

separate. The liquid-liquid transition shows lower critical solution temperature

27

(LCST) above which liquid mixtures separate into two phases according to Figure
5. It also shows the upper critical solution temperature (UCST) below which the
liquid mixture will again separate. Between these two points phase separation

doesn’t occur and one phase exists.

Phase separations within the amorphous region of polymer-polymer
blends do cause mechanical failures and most likely can affect the attractiveness
of the blend. It is important to note that miscibility in polymer blends is related to
the amorphous phases of the blend as well [42]. There are cases where one
polymer in a blend may start to crystallize. This may seem like it is phase
separating when it may not be phase separating. While there is a solubility limit
of one polymer in another, there is still value in developing partially miscible

blends.

Phase diagrams of polymer blends give a better understanding of phase
separations and provide insight on immiscible blends [42]. Figure 6 shows two

types of phase diagrams for polymeric mixtures. The solid line represents the

binodal curve boundary where stable polymer blends interface with their
metastability. The dotted line represents the spinodal curve boundary where
metastable polymer blends interface with unstability. The binodal and spinodal

curves are also known as the limit of miscibility and metastability, respectively.

In the metastable region, stable polymer blends can quickly separate into the
compositions along the binodal with large fluctuations in composition as shown in
figure 6 at compositions A, B, C. On the other hand, it can remain in the

metastable phase indefinitely due to highly viscous, entangled, incompatible

28

 

 

 

   
     

 

 

 

 

 

 

 

I
\\ I
T ------ Sin Ie
C Phagse \ TWO-Phase I!
\ \ Region
I \ \ I
I ITwo-Phase \ I
T ' Region I \ ’
T \ /
T1 '- -- D —- 3 ‘ _-_.__-_, \ ,/
’ Tc -_..._ _..__- MB ’
i
Single i Phase
‘ l
o I 0 i 1
A Xc B I Xc
wt. fraction of component Il wt. fraction of component II

 

 

Figure 6. Phase Diagrams for Polymer Mixtures showing the binodal (solid)
and spinodal (dashed) curves along with critical points [42]

 

 

 

polymer mixtures. At the same time, any small fluctuation in composition will

immediately cause phase separation if within the spinodal boundary region.

Also, in phase diagrams for polymer blends (Figure 6) the apex of the
curves where the binodal and spinodal curves meet are again the LCST and
UCST points. If the curve is bent downward it exhibits a UCST, above which
exists a single-phase blend. If the curve is bent upward it exhibits a LCST under
which a single-phase blend exists. Endothermic (positive AHmix) and positive
entropy polymer mixtures usually display a UCST, while exothermic (negative

Hmix) and negative entropy polymer mixtures usually display a LCST [42].

29

Molecular weight affects the LCST and UCST as well by increasing the UCST

and decreasing the LCST with increasing molecular weight.

2. 2. 4 Interfacial Tension/A dhesion

While this work will not focus on interfacial and adhesion calculations, it is
important to understand qualitatively how interfacial tension and adhesion plays
an important role in polymer blends. Interfacial energies are thought to greatly
affect dispersion, morphology and adhesion 42]. In turn, these areas (dispersion,
morphology and adhesion) strongly affect the mechanical properties of a blend
and, hence, the performance. The goal is to have low interfacial tension and

high adhesion in polymer blends [42].

Interfacial tension is the energy barrier preventing one liquid from being
immersed in another [48]. It is directly related to the work of adhesion by
Equation 18, where D, is the interfacial tension of component i, Ci] is the

interfacial tension between two components and WA is the work of adhesion.

WA = 71+ 72 - 1’12 Equation (13)

To maximize the work of adhesion, the interfacial tension of the combined
blend must be minimized [42]. There are various methods for calculating the
interfacial tension of the blend, some of which require knowing the surface

tension of the each component of the blend. More important is the behavior of

30

interfacial temperature with different stimuli. For example, increasing
temperature slowly decreases the interfacial tension between two polymers,
whereas stimuli such as small amounts of additives and/or increasing polar
interactions can dramatically decrease the interfacial tension. An increase in

interfacial tension can be seen with regards to increasing molecular weight [42].

On the other hand, adhesion is the bonding or joining of dissimilar bodies
(or cohesion is the joining of identical bodies) [42]. For polymers, the ideal
adhesive strength is given as 15000 psi and it is reported that the practical
adhesive strength is normally around 1500 psi [42]. This dramatic decrease in
the practical adhesive strength shows imperfect molecular contact under
practical situations. Interfacial contact is the first stage of forming an adhesive

bond, followed by interdiffusion with or without chemical bonding.

Various theories of adhesion exist, including the fracture theory, the
wetting-contact theory and the diffusion theory of adhesion. Each theory has
strong points that relate to this work. The fracture theory states the difference
between the ideal and practical strength of adhesion mentioned earlier is that
practical fracturing is not reversible where as ideal fracturing is reversible [42].
From the wetting-contact theory, it is believed that adhesive strength increases
with decreasing interfacial tension also mentioned earlier [42]. The diffusion

theory makes several claims listed below [42].

1) Compatibility is necessary for interdiffusion to occur and increased

compatibility yields increased interfacial thickness.

31

2) A decrease in solubility differences between the two component will
minimize interfacial tension and increase adhesive strength.

3) Interdiffusion also strengthens adhesion.

4) Rate of adhesive bond formation can be expressed by a power law
equation.

5) During bond formation, increasing applied pressure will increase adhesive
strength because increased pressure promotes interfacial contact and
provides larger interfacial areas for diffusion.

While interdiffusion plays a major role in adhesive bond formation,
chemical adhesion is also a way to provide increased adhesion. It has been
shown that through chemical adhesion, adhesive strength increases
substantially. Adhesive strength gain from chemical adhesion is known to be 35
times greater than dispersive forces alone [42]. This value often happens to be
the same as the ratio of a chemical bond to Van Der Waal bond dissociation
energy. While an increase in functional groups such as carboxyls and hydroxyls
has also been shown to increase adhesion to various substrates, too many
functional groups may lower adhesive strength [42]. Once again, the goal in this

work is to lower interfacial tension and increase adhesive strength.

2.2.5 Summary

The thermodynamics of polymer blends consists of many aspects.
Understanding the fundamental equations, miscibility/compatibility and phase

separation are key to developing a successful blend. Qualitative consideration of

32

interfacial tension and adhesion also increases knowledge about polymer-
polymer blends and it provides insight into increasing the performance of those

blends.

2.3 Polyester Synthesis

This segment will cover the literature review of polyesters from the 1800’s
to present. The literature review will be followed by a basic chemistry overview
of polyesterifications closely associated with this study and a synopsis of

Carothers’ theory as it relates to polyester synthesis.

2.3.1 Synthesis Literature Review

The history of polyester synthesis can be dated back as far as 1847 when
Berzelius reacted tartaric acid and glycerol to form a resin [13, 49]. Other
scientists followed suit years after, reacting glycerol with camphoric and citric
acids [49, 50]. According to Kienle, it was Baekeland’s organized studies on the
reaction between phenol and formaldehyde that sparked interest from General
Electric (GE). This interest then led to Watson Smith’s studies on the synthesis
of a solid, clear resin from glycerol and phthalic anhydride that would gain
significance commercially [50]. Callahan and other scientists at GE studied the
glycerol-phthalic anhydride reactions and disclosed information about the
process for preparing such a resin as well as specific findings that could lead to
commercial applications [51]. One such finding was that using monobasic acids

resulted in more flexible materials as compared to resins made with dibasic

33

anhydrides [49, 50]. These along with other breakthroughs led to the continued
use of alkyd resins as coatings and the new use of alkyd resins commercially as
adhesives [50].

Kienle did a significant amount of work on studying the kinetics and
reported his work in a series of papers on the polyhydric alcohol-polybasic acid
reaction [51 — 54]. First, Kienle set out to provide a systematic study on the
glycerol-phthalic anhydride reactions in hopes that his results would be useful for
other reactions and would disclose correlations with other sciences [51]. Using a
0.520.57 ratio of glycerol to phthalic anhydride, he reports details on the reaction
setup, the process parameters and the isothermal kinetics of the reaction using
acid and saponificaiton numbers [51]. He also reports on the viscosity, density,
color and refractive index [51]. Kienle then repeated his work replacing glycerol
with ethylene glycol and reveals numerous similarities between the reactions
[52]. The major difference that he reports is that “...gelation of the resinous
product does not take place at any temperature...” when ethylene glycol is used.
Kienle continued his work by substituting the phthalic anhydride with phthalic acid
and reacting with glycerol [53,54]. The minor difference found by Kienle was
that it took a slightly longer time for half of the volatile substance to be collected
when compared to using phthalic anhydride. Otherwise he reports comparable
results with the acid or anhydride of phthalate. The last two papers in Kienle’s
series on the polyhydric alcohol-polybasic acid reaction involve the reaction of
glycerol with succinic acid and anhydride, maleic anhydride, adipic and sebacic

acids [55, 56].

34

Slightly after or around the same time, Wallace H. Carothers decided to
study molecular structures in order to identify any similarities with other polymeric
materials. He chose to study aliphatic alcohols and carboxylic acids because
they were commercially significant and relatively well studied scientifically.
Carothers and others carried out esterifications with excess alcohol, and they
were credited with the idea of reducing pressure to increase molecular weight.
These studies eventually led to the Carothers’ equation, which related the extent
of reaction at the gel point to the average functionality of the mixture [57].
Carothers’ expanded his work to include polyamides and was later credited with
founding the very significant fibremforming polyamide, nylon. Carothers’ success
helped lead J. R. Whinfield and J.T. Dickson to the synthetic fiber still of major
importance today, poly(ethylene terephthalate) (PET) [13].

In 1988 Jean Otton and Serge Ratton began publishing a series of papers
that would study the formation of PET based on monofunctional compounds in
order to “start to fill the gap” in regards to data on esterfication reactions at high
temperatures [58 — 61]. Their first two papers study the esterification/alcoholysis
reaction catalyzed solely by the carboxylic acid and, in another case, catalyzed
by metallic derivatives [58, 59]. When catalyzed by carboxylic acids it was found
that the reaction order was 2 and the reaction did not proceed with an order of 1
(or without a catalyst). In regards to metallic catalysts, Otton reports titanium
compounds to be the most effective in both the esterification and alcoholysis
reactions [58]. The activation energies, rate constants and the ester percent

conversion for several metal catalysts are compared [59]. While other metal

35

catalysts show activities during esterification, their activities are low in the
presence of carboxylic acid groups. In the alcoholysis reactions, some metal
catalysts’ activity decreases with decreasing ester content in the mixture,
whereas antimony’s (Sb) activity increases with decreasing ester content in the
mixture [60]. It was also shown that while antimony and titanium metallic
catalysts can both be useful in the formation of PET, antimony—based catalysts
seem to require a higher concentration for the same effect as compared to
titanium catalysts.

The works of many have contributed to the development of polyesters. In
addition to the advances made by Berzelius, Kienle, Carothers, Whinfield, Otton
and those associated with them, others today still contribute to the knowledge of
polyesterifications involving glycerol. Such scientists as Kiyotsukuri, Tsutsumi
and others have done extensive work with synthesizing polyester [62] and
polyamide [63] films using multifunctional aromatic tricarboxylic acids with
glycols, and aliphatic and aromatic diamines with acids, for polyester and
polyamide formation, respectively. With increasing relevance to this current
work, Kiyotsukuri and others have studied polyester films based on melt
condensation of glycerol with aromatic and aliphatic acids in a 2:3 molar ratio to
form cast films for further polymerization [64, 65]. They found that while films
formed by the aliphatic dicarboxylic acids yielded a lower tensile strength than
films fromed by aromatic dicarboxylic acids, they also displayed greater
elongation characteristics. The opposite was true for films formed with aromatic

dicarboxylic acids. Short preparation times were used for the prepolymer

36

formation and further polymerization as cast films continued at temperatures
ranging from 270 — 210 degrees C for 0 to 6 hours. The method of prepolymer
formation followed by film preparation and postpolymerization was continued by
Tsutsumi with Nagata after first esterifying glycerol and sebacic acid to enhance
enzymatic degradation of the polyester films formed [66]. They found that
enzymatic degradation greatly decreased for mixtures of modified glycerol and
terephthalic acid for an amount greater than 50 mol% of terephthalic acid.

The use of glycerol in mixtures has had both promising and less attractive
outcomes. Pramanick and others incorporated glycerol into a mixture with
trimellitic acid and poly(ethylene glycol) (PEG) and transesterified the mixture
using FeCl3 as a catalyst at various temperatures over six hours [67]. They
claimed to have successfully synthesized a series of three dimensional
copolyesters that could potentially be used as “bioerodable matrices for drug
release in a controlled manner" [67]. Again, the addition of glycerol to a mixture
of bisphenol A, phthalates and carbonate was claimed to improve “mechanical
properties for film-substances, which also work well for high aromatic content
systems” [68]. On the other hand unmodified glycerol was added to PET and the
study attributed glycerol’s poor reactivity, as compared to diethylene glycol and
dipropylene glycol, to its inability to solvate PET efficiently.

Based on some studies, polyesterification products using glycerol have
been classified as being hyperbranched polymers—part of the dendrimer
macromolecular family. This group of macromolecular structures is

characterized by their high degree of ordered (dendrimer) or disordered

37

(hyperbranched) branching, multifunctionality and low viscosities [69 — 72].
Studies by Hao and Lin of hyperbranched polyesters of the type X3-Y2 have
shown reasonably low viscosities in the ranges of dendritic and hyperbranched
polyesters [70]. On the other hand, Studies have shown that materials referred
to as ‘hyperbranched’ that are formed from one monomer with three functional
groups (X3) and another with two (Y2), sometimes do have an relatively high
inherent viscosities [69].

Very few of the the X3-Y2 type reactions are carried out in the melt as
suggested by Unal [72]. Carrying out polycondensations in the melt avoid the
use of solvents that could be costly, economically and environmentally. Unal,
Stumbe and Wyatt have carried out successful polycondensations in the melt.
Unal reported that by discontinuing the X3-Y2 type reaction before gelation at
90%, one could obtain highly branched poly(ether esters) with weight average
molecular weights averaging 450000 with a PDI of 13.7 [72 — 74]. More in line
with the current study, Bruchmann reacted glycerol with adipic acid (similar to
Kienle’s work over half a century ago) varying the molar ratio of the reactants
[73]. He found that while ratios above 1 of adipic acid to glycerol tended to gel, a
ratio 1:1 of adipic acid to glycerol did not gel even after 8 hours. Similar to Unal,
he obtained polymers with weight average molecular weights of up to 23,370 by
increasing the ratio of acid to alcohol and stopping the reaction just before
gelation. Most recently a study by the USDA on hyperbranched oligomers
formed by the reaction of glycerol with three different acids (succinic, also studied

earlier by Kienle, azelaic and iminodiacetic acid) claimed that branching is

38

evident via their mass spectra results [74]. The glycerol-based polyester weight
average molecular weights ranged from roughly 1300 to 3000 Daltons with
polydispersity indices around 1.3.

The use of glycerol to form polyester resins have been around since the
mid 1800’s. Extensive work has been done that led to the development of
synthetic polyesters like poly(ethylene terephthalate) and poly(butylenes
terephthalate). Studies continue on developing polyester resins that are useful
and safe for the environment, with much of the published work involving
polyesters based on glycerol coming largely from abroad with a number of

studies from Japan, India and others from a number of European nations.

2.3.2 Chemistry

The chemistry involved in this study includes a combination of
esterification and transesterification reactions, all part of step-growth
polymerizations, although not all step-growth polymerizations involve
condensation reactions. This section will cover the basics of polycondensation to
form polyesters. Forming polyesters is a two step procedure including 1)
Esterification or Transesterification where most of the monomers react and 2)
Polycondensation where higher molecular weight is gained by the removal of low
molecular weight species under vacuum at high temperatures. Classic
polyesters, poly(butylene terephthalate) and poly(ethylene terephthalate), will be

referred to as examples.

39

Esterification and Transesterification

In esterification reactions, mainly carboxylic acids, dicarboxylic acids or
their anhydride derivatives are used in conjunction with other functional groups
such as alcohols (diols, triols or polyols) to form ester linkages [49]. Acid
catalysis enhances the ester reaction by adding electrophilic character to the
carbonyl carbon atom, making it more attractive for the negatively charged
oxygen of the alcohol. Base catalysts are used as well. The reactions in Figure
7 show esterification reactions using maleic acid,
maleic anhydride, and a transesterification reaction using dimethyl terephthalate,
each with glycerol. Using maleic anhydride reduces the amount of water formed
as compared to using maleic acid. The last reaction shows a transesterification
reaction where one ester and one alcohol form a different ester and different
alcohol. Equilibrium exists in these reactions and in order to drive
polycondensation, the reaction must be pushed toward polycondensation by the
removal of by-products. The atoms connected in the ovals represent ester
linkage formations. PBT production transesterification is used and methanol is

distilled off as a by-product.

4O

 

O

OH
0 o 0 HO OH __’ HO 6.
1n )4 ~— .
Maleic Anhydride Glycerol Malelc Anhydrlde-Glycerol Adduct

O

jet.
OH-C icy/OH ’ ”O “"0
+

 

0 HO
Maleic Anhydride Glycerol Malelc Anhydride-Glycerol Adduct Water
OH
/0 0
OH
HO OH
HO ‘— O ‘\ H3C-0H
0 O_ HO O
Dlmethyl phthalate G"°°'°' DMP-Glycerol Adduct "0mm“

 

 

 

 

Figure 7. Esterification reactions using Maleic Acid and Maleic
Anhydride as examples and Transesterification reaction using
dimethyl terephthalate

 

 

 

Polycondensatjm

Polycondensation is where the reaction is driven to form high molecular
weight polyesters. Polycondensation in the melt occurs in a vaccum at high
temperatures. Lower molecular weight species or excess glycol are removed at
these high temperatures and under intense mixing. Even under these conditions,
oligomers may still be present. In fact, PET and PBT contain some oligomers

even on the commercial scale.

41

Polycondensation reactions take time to form high molecular weight species and
they require high conversions. In polycondensation reactions, reactivity is
assumed to be independent of molecular weight under the principle of “equal
reactivity” [49]. Very viscous mixtures form in polycondensation [50]. Hence
flowing a dry, inert gas through the mixture aids in the removal of the appropriate
components from the mixture and in the suppression of oxidative decomposition

due to the presence of oxygen.

2.3.3 Carothers’ Theory
Relation of polymer size, percent conversion and feed molar ratios, groups
in step-growth polymerizations can be attributed to what is known as Carothers’
equation. Carothers’ equation is a guiding principle in polyester synthesis,
helping to design reactions having appropriate yields with respective molecular
weights. The basic assumptions associated with Carothers’ theory are
summarized below:
1) Equal reactivity of functional groups of the same kind or on the same
molecule independent of the size of the molecules reacting
2) Equal reactivity of functional groups even after other functional groups
on the same molecule have reacted
3) Homogenous, single-phase reaction mixture
To arrive at the number-average degree of polymerization, two different
functional groups must be present (A-type and B-type functional groups).

Functionality is defined by Rudin as the number of positions in the monomer

42

available for reaction under specific conditions [43]. For a given monomer, a
certain average functionality exists. Equation 20 defines the average
functionality (fay), where f, is the functionality of the monomer and N, is the
number of moles. This equation is specifically for when two types of functional
groups reacting are present in equal concentrations. When there is an excess of
one functional group, the other type of functional group will be the limiting
functional group and the average functionality can be calculated using Equation

21, where ha is the limiting

 

 

fav 2 22"? Equation (20)
2n A

fav = Equation (21)
EN:

number of equivalents (given by the moles multiplied by the functionality of the
reactant), and N; is the moles of reactant i.

The derivation of Carothers’ equation requires an expression for the extent
of conversion (p). The extent of conversion can be defined as the ratio of the
number of functional groups reacted to the number of functional groups initially
present. This ratio can be seen in Equation 22, where No is the initial total

number of moles and N is the final number of moles. The numerator represents

: 2'(N0—N)

Equation (22)
Nofav

43

the amount of functional groups used in the formation 01 polymer-polymer
anages.

To get the degree of polymerization (Xn) associated with the mixture, the
initial number of monomer unit (No) and the remaining number of moles (N) must
be known. Since the total moles is known, the No is known. To find N, Equation

22 can be arranged to give Equation 23 where all the variables are as defined
N = é) - (2N0 - Nopfav) 5903110" (23)

before. Therefore Xn can be quantified as shown in Equation 24 where Xn is the

N0 2

7g: 2
l _
(5"(2No—Nopfav) 2 ”W

 

Equation (24)

degree of polymerization and all variables are as defined earlier. This is
Carothers’ equation and it can aid in predicting the gel point in step-growth
reactions, choosing an optimal ratio of reactants and more. The degree of
polymerization (X,) will go to infinity when a gel is formed. Thus, by setting the
denominator of Equation 22 equal to zero, the percent conversion at which the
reaction will gel can be found. Since gelation actually occurs when the weight-
average degree of polymerization (Xw) becomes infinite (not when the Xn is
infinite as predicted by Carothers) and XW is greater than or equal to XIn the

mixture will reach its gel point at conversions lower than predicted [43].

44

Chapter 3

 

Materials and Methods

 

This chapter lists all the materials used in the production of CA blends and
glycerol-based polyesters along with the supplier information. It also includes
basic information on materials used in various characterization techniques.

The second half of this chapter briefly discusses the processing
equipment used in blend preparation or in producing test specimens. All
characterization techniques used are summarized in general and, when possible,

general procedures are described.

3.1 Materials
This section divides the materials into the following three categories. The bulk
1) Bulk Materials
2) Catalysts
3) Chemicals
materials include cellulose acetate, triacetin, poly(butylene adipate-co-
terephthalate) and glycerol. The catalysts are sodium carbonate and dibutyltin
oxide and the chemicals are pyridine, acetic anhydride, n-butanol and

tetrahydrofu ran.

45

3. 1.1 Bulk Materials

Cellglose Acetafi

Sleek, vibrant, modern, and shiny are words that all describe some
positive attributes of CA. Cellulose acetate has the ability to produce products
that are sleek and smooth with a shine or a reflection, making something seem
modern. While demand for cellulose acetate-based products has decreased, it
has never been completely eliminated from the market, speaking to its longevity.
Cellulose acetate has still found use in various markets due to aesthetics coupled
with strength. Typical cellulose acetate-based products include or are used in
eyeglass frames, tool handles, cigarette tows, toys, clear adhesive tape, writing
instruments, and face shields.

Clarity, toughness, compatibility with dyes and resistance to grease are among
some of the advantages of using CA and why CA is used in many of these items
[10]. Aside from aesthetics and excellent mechanical properties for various
applications, cellulose acetate is produced from renewable natural resources and
has the ability to biodegrade, making it an ideal raw material source for bio-based
research and development of polymer products. Unfortunately, there are some
drawbacks that come with cellulose acetate including engineering issues related
to processing, along with the miscibility and compatibility of its blends. Table 1
shows some basic properties of CA and Figure 2 gives the structure of CA . The
cellulose acetate, (CA-398-30, D8 of 2.46, 50lb bag) was purchased from

Eastman Chemical Co. (Kingsport, TN).

46

 

 

 

 

 

 

Acetyl Content (%) 39.7
Thermal Decomposition (°C) 304
S ecific Gravity 1.31-1.32
Melting Point (°C) DSC 230-250
Glass Transition (°C) DSC 189

 

 

Table 1. Eastman Cellulose Acetate
Properties

 

 

 

Triacetin

Triacetin or glyceryl triacetate (C9H1406, ~240 kg) was purchased from
Chemical S.P.A in Italy. Triacetin structure can be seen in Chapter 4, Figure 12,
with other plasticizers for cellulose acetate. It has a formula weight of 218.21
g/mol, a boiling point of 258 °C, a melting point of 3.2 °C and a freezing point of —

78 °C.

Poly(bgtvlene adjpate-co-terephth_alate) (PBAT)

Poly(butylene adipate-co-terephthalate) is a biodegradable, aliphatic-
aromatic copolyester introduced by BASF to address the demands of growing
domestic waste issues. The structure of PBAT can be seen in Figure 8. The

copolymer is made

47

 

1
_l

O O . ]

o o
I II n / u n , i :
/ l—(CHzp—o-c—(CH2I4—C—O—/m—(Cqu—o—c00—0— “’1'“

[__

 

 

 

 

 

{ Figure 8. Structure of BASF Poly(butylene "r ‘ :3 t.. ,L‘” ' ‘ )[40] I

 

up of 22.2 mol% terephthalic acid, 27.8 mol% adipic acid and 50 mol%
butanediol as determined by CNMR [40]. Unfortunately, PBAT is not made from
renewable resources. Nevertheless, PBAT boasts a faster degradation rate than

cellulose acetate on composting, as shown in Figure 9.

 

I'll" --_,--c

 
 
 
 

Cellulose a, , - -

Ecoflex

_,, I
“Biodegradation [theor. CO; in %]
C- 1‘) .‘I; :2; 51;; 3,30

Tim. [M

l:-

 

 

 

Figure 9. Biodegradation of
Cellulose Compared to that of
PBAT [41]

 

 

 

It also exhibits properties similar to LDPE, having high molecular weight
and a long chain branched molecular structure [41]. The toughness and flexibility
of PBAT has made it a natural pick for blends with other biodegradable polymers

like starch and PLA [42,43]. Table 2 gives a summary of some properties of

48

PBAT. These properties have allowed the use of PBAT in various areas.
Currently PBAT is used in making films applicable to agricultural, packaging and
composting uses. Some other benefits of PBAT include its good thermostability
(up to 230 °C), and its ability to be easily processed with no predrying of the

pellets. The grade of Ecoflex (~55 lbs) purchased from BASF (Mount Olive, NJ)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

is F BX 7011.
Mass Density (g/cm“) ISO 1183 1.25-1.27
MFR 190°C, 2, 16 kg ISO 1133 2.7-4.9
Melting Point (°C) DSC 110-120
Glass Transition (°C) DSC -3O
Shore D hardness ISO 868 32
Vicat VST A/50 (°C) ISO 306 80

Table 2. BASF PBAT (also known as
Ecoflex) Properties
Glycerol

Glycerol, also known as glycerine, is a hygroscopic, trihydric alcohol, as
shown in Figure 10. It is made naturally as a by-product of soaps and

synthetically from propylene. It is an odorless, colorless, sweet, viscous liquid.

 

HO OH

.014

Figure 10. Structure of
Glycerol

 

 

 

 

 

 

49

Glycerides, the esters of glycerol, are found naturally in animal and/or vegetable
fats and oils. More currently, glycerol has been produced as a by-product in

biodeisel production [44]. Figure 11 shows the associated reaction of the

 

 

9 o
I
CH2—0_C_R1 R’_o_|cl_n1
O 0
ll Catalyst II
CH -O—C— R2 + aa’on ,___——-* R’—O—C—R2
l 1 °
N
CHz—O—c—R3 Fl’—O—C—R3
Glyceride Methanol Methyl Soyate

 

 

 

 

Figure 11. Biodiesel production chemical reaction showing Glycerol as
a by-product [45]

 

 

 

process. A triglyceride reacts with methanol to give methyl esters used in
biodeisel production and glycerol, the by-product. With an increasing interest in
biodiesel production, the development of glycerol esters can play a key role in
the efficiency of biodiesel production, adding giving value to the overall

production. Glycerol properties are summarized in Table 3.

Point 17.8
Point 290

Flash Point Closed 160
1 .261

 

 

Table 3. Properties of Glycerol

 

 

 

50

Some common products that use glycerol in the food and personal care
industries include toothpaste, skin care, hair care, icing, food coloring and more.
In the plastics industry, glycerol is often used as a plasticizer or lubricating agent.
Nabar et al. used glycerol in the production of starch foams via reactive extrusion
[4]. In addition to its softening capabilities, glycerol can increase flexibility and
toughness. The anhydrous glycerol (4L) with a density of 1.257 g/cm3 and a

formula weight of 92.10 g/mol was purchased from J.T. Baker.

3.1.1 Catalysts

The two catalysts used were sodium carbonate and dibutyltin oxide. Both
are useful in transesterification reactions. Sodium carbonate (Na2003), ACS
grade with greater than or equal to 99.5% purity was purchased from Sigma
Aldrich in a 5009 batch. It’s formula weight, melting point and density are 105.99
g/mol, 851 °C and 2.532 g/cm3, respectively. Dibutyltin oxide (Cgl‘lmOSR) with a
purity of 98% was also purchased from Sigma Aldrich in a 1009 batch. Its

formula weight is 248.92 g/mol and it’s melting point is greater than 300 °C.

3.1.3 Chemicals

A number of chemicals were used in the synthesis of glycerol-based pre-
polymeric materials and for the characterization of samples (synthesized or
extruded) depending on the method used. Dimethyl terephthalate (C10H1OO4)

was a main component reacting with glycerol in the esterification process. The

51

chemical was purchased from Sigma Aldrich in 3 kg batches. The formula

weight and melting point of DMT are 194.19 g/mol and 140 °C, respectively.

The use of acetic anhydride, pyridine and n-butanol was necessary for
characterization of polyester prepolymer via the hydroxyl value calculation.
Acetic anhydride ((CH3CO)2O, 500mL) with greater than 97% purity was
purchased from Columbus Chemical Industries meeting ACS specifications. The
formula weight of acetic anhydride is 102.09 glmol. ACS reagent grade pyridine
(CsHsN, 500 mL) was purchased from J.T. Baker. It’s formula weight and boiling
point are 79.10 glmol and 115.3 °C, respectively. The ACS reagent n-butanol
(CH3(CH2)3OH, 500mL) with a formula weight of 74.12 glmol was also

purchased from J.T. Baker.

Tetrahydrofuran was used in titration studies and as the solvent mobile
phase in gel permeation chromatography (GPC). The tetrahydrofuran (C4H30,
500 mL) used in titrations was purchased from J.T. Baker. The ACS reagent
grade tetrahydrofuran has specific gravity of 0.884, formula weight of 72.11 glmol
and a distilling temperature of about 66°C. The tetrahydrofuran (C4H30, 4L) was
inhibitor free, CHROMASOLV Plus specifically for use with GPC experiments.
This tetrahydrofuran was greater than 99.9% pure, and was purchased from
Sigma Aldrich. Its boiling point, melting point and density are 66 °C, -108 °C and

0.889 g/cm3, respectively.

52

3.2 Processing and Characterization Methods

The processing methods discussed include reactive extrusion with a twin-
screw extruder for blending of polymers and extrusion/injection molding with the
DSM Micro extruder and injection molder for test specimens. The
characterization techniques mentioned are categorized by chemical, thermal,
dynamic mechanical and morphological analysis. It is important to note that most
experiments most experiments were repeated at least once. Further repeat

experiments were done based on judgement.

3.2.1 Processing Equipment & Procedures
Centum Extruder

A Century ZSK-30, twin-screw co-rotating extruder was used for all
polymer blending. The extruder has a screw diameter of 30 mm and an length to
diameter (L/D) ratio of 40. The extruder system is equipped with an extruder
driver with a speed control gearbox, 'a main AccuRate, single-screw bulk
metering feeder, another AccuRate single screw side-feeder, a cylindrical
filament die with a diameter of 2.7 mm and a length of 8.1 mm, and a cooling
system for the nine zone heating blocks of the extruder. Cooling of the extruder
barrel can be controlled by adjusting the flow rate of the cooling water supply
manually via nine valves for the nine zone heating blocks. There is a
thermocouple and pressure sensor affixed downstream the extruder to measure
the temperature and pressure of the melt and a venting/vacuum port located

downstream over zone heating block number 8. Other parts of the extrusion

53

system include a water-bath equipped with a hose for draining and filling, a
Scheer Bay Co. pelletizer (Model # BT25) for solids handling and a MasterFLex
pump (Model # 7524-40) from Cole Parmer Instrument Co for addition of liquid
additives to the mixture.

In general, processing of blends included the use of one or two feeders at
certain calibrated speeds for the feeders. The material entered the extruder via a
hopper. The feed for any liquid additives is located right after the main feed
through zone 1 of the extruder. The material passes through a series of
compounding and conveying screws, a venting zone and eventually the die
exiting the extruder. As the polymer strand exited the extruder, it was quenched
in a water bath and continued downstream in the water bath to the pelletizer.

The material was then collected and stored or dried for future use.

DSM Micro Extruder and lniectjon Molcg

The DSM Micro Extruder and Injection Molder is a micro-mixing and -
molding piece of equipment from the Netherlands. The DSM Micro 15 cc
consists of (1) the micro-extruder with twin co-rotating screws for mixing and
conveying, (2) a tiny cylinder attached to a mini feed hopper for feeding material
into the barrel of the extruder, (3) a transfer cylinder for moving the product from
the micro-extruder to be injection molded and (4) the mini-injection molder for
making common molded parts such as tensile bars, rectangular bars of various

sizes and discs.

54

In general about 10 — 15 g of material (depending on polymer pellet or
powder density) was placed in the hopper and plunged directly into the mini-
extruder barrel. The material was allowed to mix until the viscosity began to fall
suggesting good mixing had occurred. The transfer cylinder was then used to
store and transport the polymer melt exiting the micro-extruder. As the transfer
cylinder was being filled, the piston in the cylinder was pushed out. The transfer
cylinder was placed in the mini-injection molder and a high-pressure cylinder
forced the transfer cylinder piston to push the material from the transfer cylinder
to the DSM mould. The mould was removed and the injection moulding process
was repeated until the micro-extruder was empty. At this point the micro-

extruder was refilled and the process continued as before.

3.2.2 Synthesis Procedure

A 1:1 molar ratio of glycerol and DMT was combined in a reaction kettle.
The reaction kettle was equipped with reaction kettle cover containing three
24/40 joint openings and one 29/42 joint in the center. The center joint opening
was used for the glass or metal (recommended) stirrer. For the metal stirrer a
poly(tetrafluoroethylene) (PTFE) bearing was required. The other three joint
openings were used for the (1) D88 trap, (2) nitrogen inlet and (3)
thermometer/temperature controller probe combined. Above the D&S trap was a
condenser and at the top of the condenser was an outlet flowing to a container of
silicon oil to ensure nitrogen flow out of the system and no airflow into the

system. Before the cover was placed over the kettle, silicon grease or an o-ring

55

was placed between the cover and the kettle for a tight seal when clamped
together.

Once the system was contained, nitrogen was allowed to flow through the
system for ten to fifteen minutes. The temperature controller was set to 100 at a
power attenuation of 5 (this means that only 50% of the full power was being sent
to the reaction mantle). The slow heating was used to aid in lower sublimation of
the DMT by allowing uniform heating throughout the vessel. At 100°C a certain
amount of catalyst was added to the mixture and the temperature was increased
after fifteen minutes to 200°C. Once the temperature equilibrated to the set
point, it was held there for four hours to allow for a higher rate of reaction and
more methanol formation. After four hours, the reaction was stopped and the

pre-polymer was collected and cooled.

3.2.3 Characterization Techniques

Characterization of the reaction product and of the blends can be broken
down into three categories. Studying the structure, analyzing the thermal
behavior and assessing dynamic mechanical and morphological equipment will

assess polymer blend characteristics.

Chemical A_nalvsis

The chemical characterization techniques help to identify percent
conversion in synthesis development and average molecular weight data for both

synthesis products and polymer blends.

56

A. Acid/Base Titration

Titration is a method used to quantify an unknown amount of acid in a
solution by using a known concentration of standard solution (also known as the
titrant) that reacts with the unknown. With a few drops of an end-point indicator
(phenolphthalein) in the unknown acidic solution, the end of the reaction can be
identified via a color change. The amount of titrant added to the unknown
solution to bring about a color change can then be related to the unknown
amount of acid on a molar basis via a balanced chemical equation, molarity of
the standard and volume of the unknown solution to give molarity of the reagent
in question. Acid number procedures were determined per the ASTM D1980
Standard [75] with minor changes. About 78 grams of each sample was

dissolved in 100 mL of THF for the acid value titrations.

B. Hydroxyl Value

Hydroxyl value is a technique used to quantify hydroxyl groups in a
sample. In the case of this study, it was being used as a measure of conversion
for polyester synthesis. All hydroxyl values were calculated per the ASTM DI 957
standard [76] using about 0.4 grams of each sample. The samples were heated

for two hours on the water bath.

C. Average Molecular Weights

Gel permeation chromatography measures a polymer’s molecular weight

distribution as compared to a standard. The material is dissolved in a solvent

57

and the solute particles of the specified materials flow through a column of
porous particles in which larger solute particles elute from the column faster than
smaller solute particles because they pass through less of the porous particles
due to their size. The information gathered from this experiment is important
because it provides valuable information that can be indicative of a polymer’s
mechanical properties and processing behavior. In addition, the breadth of
distribution given by the polydispersity index can potentially yield information on
whether there may be significant branching of the polymer sample.

Molecular weight averages and distributions of all samples were carried
out on a 600 multisolvent gel permeation chromatography (GPC). The GPC was
equipped with a 2410 refractive index detector, a 717 auto sampler and the
Breeze software from Waters. Three Waters Styragel Columns ranging from 500
to 500,000 Daltons were used in the GPC. HPLC grade tetrahydrofuran (THF)
was used as the mobile phase and the calibration was performed against narrow
polystyrene standards. Each sample was dissolved in THF at about 2 to 3

mg/mL and filtered into clear vials with snap caps with 0.45 mm PTFE filters.

Thermal Analysis

The thermal analysis techniques used were differential scanning
calorimetry (DSC) and thermogravimetric analysis (TGA). Determination of
transitions helped in developing processing parameters and to explain behaviors

seen in the blends.

A. Thermogravimetric Analysis (TGA)

Thermogravimetric Analysis reveals the thermal stability of a material by
monitoring weight percent and weight percent per degree as a function of
temperature. A high resolution TGA 2950 from TA Instruments was used to
determine the ultimate degradation temperature of all the materials, assessing
the thermal stability.

In general an empty pan was tared in the TGA chamber under nitrogen
and after taring was completed, the sample was added to the pan and enclosed
in the chamber. The confined chamber was purged with nitrogen gas. For all
samples the sample purge of nitrogen was set at 60 ml/min and the balance
purge at 40 mein. All samples were studied by heating to 500 °C at a rate of

20°C/min.

B. Differential Scanning Calorimetry (DSC)

Differential scanning calorimetry provides information on the thermal
transitions within a material. It measures the heat flow to the material and
compares it to a standard, identifying the thermal transitions from the
comparison. It also records values of temperature and time. It has been used to
identify glass transition, melting and crystallization temperatures, and melting and
crystallization enthalpies. Analysis is done in a confined chamber with a nitrogen
purge at 100 mein.

In this study a DSC 2920 from TA Instruments with a cooling unit was

used to identify phase change temperatures (Tg, Tm, To, AHm and AHC , where T9

59

is the glass transition temperature, Tm and TC are melting and crystallization

temperatures, and Hm and Hc are the melting and crystallization enthalpies).

DALnamic Mechanical Behavior

Dynamic Mechanical Analysis assesses the mechanical properties of a
material as it relates to time, temperature, and frequency. DMA also provides a
variety of modes in which to mechanically deform a material. Specifically, a DMA
2980 from TA Instruments was used to measure the storage modulus, loss
modulus and tan delta, to gain insight on the mechanical, glass transition and
viscoelastic behavior, as well as the homo-lheterogeneity of the materials. The
DMA was equipped with a liquid nitrogen tank to allow attaining very cold
temperatures.

The three-point bending mode was used in all tests to facilitate
measurements without grip interference. All DMA tests carried out in this study
utilized rectangular test specimens from the DSM micro-extruder and mini-
injection molder that were cut to have dimensions of about 38 x 12 x 3.6 mm (l x
w x h). The DMA tests were all performed under the same conditions. The
method used for all the samples is listed below.

1) Initial temperature: -100.00 °C
2) Isothermal for 10.00 min

3) Ramp 3.00 °C/min to 85.00 °C

60

Morphology

A Phillips Electroscan 2020 environmental scanning electron microscope
(ESEM) was used to study the surface morphology of pellets. Polymer pellets
were sectioned using a razor blade. The sample was then mounted on an

aluminum stub with graphite tape.

61

Chapter 4

 

Cellulose Acetate Blends with Ecoflex

 

4.1 Overview

The need for bio-based bioplastics is evident. With the opportunity to
design blends with bio-based content comes the ability to be proactive and
integrate biodegradable materials in the design phase. The materials being
introduced in this blend have the potential to meet the need for bio-based,
biodegradable plastics that have the properties necessary to compare to plastics
on the market.

Cellulose acetate is a bio-based, biodegradable raw material that
maintains presence in the market even after being replaced in several areas by
cheap, synthetic plastics [13]. This organic ester’s prime properties include
hardness, good impact resistance, smooth texture, high clarity, and an ability to
incorporate several pigments. Cellulose acetate as a main component in bio-
based plastics is a natural selection.

Also a natural selection for blends with CA is PBAT, a biodegradable,
aliphatic-aromatic copolyester. Both tested for biodegradability in composting
environments, and both containing esters, CA and PBAT would seem good
choices for blending. Good toughness and flexibility, due to the aromatic and
aliphatic components, respectively, are traits of PBAT and should help to
maintain the toughness of CA while offering some flexibility to the blend. Also,

with PBAT having a T9 of —30°C, blends of CA and PBAT could possibly cause a

62

decrease in the blend Tg if the materials are miscible or even partially miscible. It
is important to note that there are many polymer pairs that are not miscible, but
studying these blends (miscible or immiscible) can lead to an understanding of
what is needed to promote compatibilization of the blend.

BASF developed PBAT as an answer to the need for more
environmentally-friendly plastics. They noted that bio-based products could not
perform at the level of synthetic products, and, hence, decided to use synthetic
materials to design their final product [40]. Regarding CA, external plasticizers
have often been used to increase the processing range of CA by decreasing the
T9. Brydson notes that some of the most important plasticizers for CA are
dimethyl phthalate (DMP) and triacetin (TA) [42]. More recently, triethylcitrate
(TEC) has also been recognized as a good plasticizer for CA. The structures for
DMP, TA and TEC are shown in Figure 22 and calculated solubilities using
equation 6 are shown in Table 9. Although these are great external plasticizers
for processing CA, they tend to leach or bleed out, thereby altering the properties

of the blend.

63

 

0

(:3: :91 SW

Dimethyl phthalate YTriacetin / Triethyl citrate

 

 

Figure 12. Important External Plasticizers for Cellulose Acetate

 

 

 

A successful blend of CA and PBAT could potentially address two
fundamental issues in CA and in PBAT. The first issue is the lack of bio-based
content in PBAT and the second issue deals with the small temperature
processing range of CA. Adding CA to the blend can increase the biobased
content of PBAT making it more attractive from a biobased perspective. Also, a
compatible blend of CA with PBAT can broaden the processing range of CA and

eliminate the need for the plasticization of CA with low molecular weight

 

 

 

 

 

plasticizers.

Calculated Hoy Paper by Ljungberg,
Series Solubility Nadia using Hoy

Plasticizers Parameter Series

TA 19.2 19.1

TEC 19.7 19.7

ATEC 18.9 18.9

ATBC 18.0 18.0

 

 

 

 

 

 

Table 4. Calculated solubilities compared to literature
using Hoy parameters

 

 

 

64

4.2 Materials & Methods

The materials used in this study include cellulose acetate, triacetin,
poly(butylene adipate-co-terephthalate), dibutyltin oxide and inhibitor-free
tetrahydrofuran. The methods for processing include a reactive extrusion system
and the DSM for preparation of test specimens. The characterization techniques
employed include gel permeation chromatography, differential scanning
calorimetry, thermogravimetric analysis, dynamic mechanical analysis and

environmental scanning electron microscopy.

4.2.1 Specific Processing Methodology

The CA used in the blends was dried for at least twelve hours to remove
moisture. The Ecoflex used in the blends was dried for the same amount of time
only if the amount in the mixture was less than 70%. Reactive extrusion of CA
and PBAT occurred using the following temperature profile from feed to die:
30/145/200/220/225/230/230/230/225/220 (°C). The melt temperature ranged
from 220-226°C and the motor speed was set at 180 RPM. Blends ranging from
50% to 70% Ecoflex were prepared using the twin-screw extruder under the
given conditions

In the second set of CA-Ecoflex blends, the blend of just 30% CA and
70% Ecoflex was extruded using the following temperature profile from feed to
die: 15/40/80/150/190/220/230/230/230/220 (°C). The melt temperature was
about 227 °C and the motor speed was 250 RPM. CA with 30% triacetin was

extruded with the following temperature profile from feed to die:

65

15/40/100/150/180/210/210/210/210/200 (°C). The melt temperature was 203 °C
and the motor speed was 200 RPM. The blends of plasticized CA (30% and
50%) and Ecoflex (70% and 50%) with and without catalyst were extruded at the
same temperature profile as was used to plasticize CA. The melt temperatures

ranged from 203 - 210 °C and the motor speed was set at 250 RPM.

4.2.2 Specific Characterization Methods

For DSC, the sample was heated to 210 °C at a rate of 20 °C/min followed
by quenching to —55 °C where it was held isothermally for 1 min. The sample
went through the first heat scan to erase the thermal history and allow for the
removal of moisture from the blend. The melting and cooling data were gathered
from the second heat scan when the sample was reheated to 255 °C at a rate of
20 °C/min followed by cooling to 20 °C at 20 °C/min.

The GPC, TGA, DMA and ESEM methods used are as described earlier in
the general characterization section. ESEM Images of the blends were taken at

magnifications of 1000, 2000 and 5000.

4.3 Results and Discussion

4.3.1 Summary of Runs

Various compositions of CA and PBAT were extruded with no additives

simply to study the effect of composition on the blends. These samples are listed

above the darkened line. Blends with higher than 50% CA by weight could not

66

be extruded. Below the darkened line, in the same table, are the samples used
to study the effects of plasticization, catalyst and composition. These samples
were subject to the corresponding processing and characterization methods
discussed in section 4.2. The naming convention for samples below the
darkened line was developed by using the first letter of each material (“C” for CA,
“CT” for plasticized CA and “P” for PBAT), followed by numbers representing the
percentage of each component in the order that the component is listed in the
sample name. These runs were used to help elucidate the effects of composition
and plasticizer on CA-PBAT blends, as well as the effects of catalyst and

composition on plasticized blends of CA-PBAT.

 

 

 

 

 

 

 

 

 

 

Cellulose .
. Acetate Ecoflex DBTO
, Sample. . . (CA) Triacetin (PBAT) (phr)
30% CA 30 70
40% CA 40 60 ---
50% CA 50 50
CP37 30 70
CT73 70 3O
CTP37 21 9 7O
CTP55 35 15 50
CTPDS7 21 9 70 1
CTPDS5 35 15 50 1

 

 

 

 

 

 

 

 

Table 5. Extruded Blends of CA and PBAT without
modification

 

 

 

67

Effect of Composition on Thermal Properties of QA-PB_AT_

The effect of composition on the degradation temperature of the blends
can be seen in Table 6. Note t hat PBAT is directly from the manufacturer and
NPBAT is PBAT that has been extruded one time. Figure 13 shows an example
of the two distinct derivative weight loss curves formed for 30% CA. Samples at
all loadings of CA showed similar curves. This confirms that CA and PBAT do
not form a miscible blend at the weight percentages of CA studied. In fact, a
closer look at Table 6 shows that the CA blend component degradation
temperature shifts towards the degradation temperature of 100% CA, while
PBAT temperatures remain about the same as that of neat PBAT (NPBAT). At a
composition of 30 wt. % CA, however, the blend displays a slightly higher
degradation temperature for the CA component. This suggests that there may

be some compatibility between the two components.

 

 

 

 

 

 

 

Sample CA PBAT
100% CA 344.59 --

50% CA 344.73 374.29
40% CA 349.05 374.63
30% CA 350.49 376.02
100% PBAT 373.68
100% NPBAT 375.24

 

 

 

 

 

 

Table 6. Preliminary CAPBAT degradation
values as retrieved by TGA

 

 

 

68

 

 

100‘

 

344.73°C .

 

80—

Weight (%)

40~

20.1 .—

 

W

 

 

J

374.29°C

0.025

- 0.020

.0
Q
01

Log[Welght] (11°C)

-0.010

"V.

r 0005'

De

~ 0.000

 

 

0...160...260.

 

' 360 '
Temperature (°C)

-0.005

500

WWMTAW

 

 

 

Figure 13. Weight loss and derivative weight loss (represented by two
maximas) curve of CAPBAT sample to illustrate the two component blend of
CA and PBAT

 

Comparison of the samples’ weight loss curve show that the blends first

mimic the weight loss curve of CA (the component with the lower degradation

temperature) and later the PBAT (the component with the higher degradation

temperature) weight loss curve is followed. Weight loss curves shifts toward the

higher degradation temperature weight loss curve (PBAT) with increasing PBAT

content, exhibiting more PBAT-like characteristics and possibly indicating

increased compatibility. These trends can be seen in Figure 14.

69

 

 

 

 

 

 

 

 

 

 

 

120
l
80- o 100% CA
‘ 1:1 50% CA
A ‘ + 40% CA
2;. 60‘ o 30% CA
j I PBAT
g 1 <1» NPBAT
I
401
20-
0 r 1 fi 1 r . 1 . j a . .
100 150 200 250 300 350 400 450 500
Temperature (°C)

 

 

 

 

Figure 14. Weight loss curves of preliminary CAPBAT samples

 

 

 

Thermal transitions in DSC support the claims of non-miscible blends by
exhibiting two glass transitions and two crystallization peaks as shown in Table 7.
Plasticization of CA with the addition of PBAT is evident in the decrease of the
glass transition temperature to 189°C from 207°C and in the elimination of the CA
melting peak. While melting temperatures of the PBAT component remain
somewhat unaffected as compared to NPBAT, the melting endotherm,
crystallization temperature and crystallization endotherms of PBAT decrease with

increasing CA content in the blend.

7O

 

 

Sample TgPeAT(°C) TgCA (°C) Tm (°C) Hm (J/Q) TcPBAT (°C) HcPBAT («l/9) TcCA (°C) HcCA (4’9)

 

100% CA --- 207.00 234.25 9.39 --- --- 202.35 5.48
50% CA -29.62 194.94 122.38 8.08 64.20 9.19 202.28 4.00
40% CA -30.93 192.35 120.97 9.67 69.97 1 1 .41 201.22 3.51
30% CA ~30.97 189.20 121 .72 10.92 68.64 12.25 202.44 3.82
100% PBAT -28.28 --- 1 15.27 18.49 25.88 13.99 --- ---
100% NPBAT -29.85 --- 123.31 13.33 76.55 17.09 ---

 

 

Table 7. Summary of DSC thermal properties for preliminary CAPBAT blends

 

 

The phenomenon occurring begins with CA particles dispersed in the
PBAT. As PBAT begins to melt, CA particles increasingly interfere with the
PBAT crystals and successfully disrupt the crystallinity of some PBAT chains,
lowering the enthalpy of melting. There is no effect on PBAT melting since the
interaction between CA and PBAT occurs as the PBAT enters the melt stage. As
PBAT in the blend melts, PBAT chains interfere with the hydrogen bonding in
CA, effectively separating CA chains, leading to the elimination of the CA
component melting curve. On cooling, CA crystallizes first and thus its
crystallization temperature is not affected. However the extent of crystallization
is slightly decreased with increased PBAT content. Crystallized CA chains then
inhibit the crystallization rate of PBAT, effectively reducing the extent of
crystallization as well as the crystallization temperature as shown in Figure 15.
Thus, both CA and PBAT have effects on the crystallization properties of the
other. This, again, indicates that there is some interaction occurring between CA

and PBAT.

71

 

1.4

 

O 0% CA, 100% NPBAT
El 30% CA
+ 40% CA
0 50% CA

 

Heat Flow (Wig)
O

L

0.8 -

 

o_3 . . . , . . . , A . . , . . .
4O 60 80 100 120

Temperature (°C)

 

 

 

 

 

Figure 15. DSC crystallization temperatures of preliminary CAPBAT blends

 

 

 

This compositional study has shown that a blend with 30% CA shows
some interaction with CA and is possible a compatible blend. Using the
compositions in this study, we will look at the effects of plasticizer, catalyst, and
the compositional effects of both plasticized and catalyzed samples. Materials,
processing equipment and procedures, and characterization descriptions are as

described earlier.

72

 

fleet of Plasticization and Composition

The compositional make-up of the samples in this segment can be found
in Table 8. The effect of plasticization will be looked at simultaneously with the
effect of composition on plasticized CA samples. All the materials and methods
used are as described earlier. It should be noted that reactive carboxylic end
groups of PBAT may react with accessible hydroxyl groups of CA under reactive
extrusion of two without catalyst.

Molecular weight data for all samples in this segment are shown in Table
8 and Figure 16. CA is cellulose acetate, CT73 is plasticized cellulose acetate
(30% plasticizer), CP37 is unplasticized cellulose acetate and PBAT in a 3:7
ratio, CTP37 is plasticized cellulose acetate and PBAT in a 3:7 ratio, and CTP55
is plasticized cellulose acetate and PBAT in a 5:5 ratio. The physical blend of CA
and PBAT at 30 wt. % CA (CP37) was compared with the plasticized CA and
PBAT blends at 30 (CTP37) and 50 (CTP55) wt. % plasticized CA. The number-
average molecular weight for CP37 decreased dramatically as compared with
CA, and had a slight increase in PDI due to the addition of a lower molecular
weight component, as expected. Compared to NPBAT there was an increase in
Mn and decrease in PDI. Plasticized blends of CA show higher molecular weight
values due to the breaking up of the hydrogen bonding and increasing access to

available CA hydroxyl groups to react with PBAT carboxylic end groups.

73

 

 

Sample Mn (Dalt.) Mw (Dalt.) PDI Mum], Mm % diff. Mn % diff. Mw
CA 76640 1 45521 1 .90 --- --- --- ---
CT73 66406 126976 1.91 --- --- --- ---
CP37 48800 95531 1 .96 50755 105579 4% 1 1 °/o
CTP37 51 851 1 01 667 1 .96 49248 1 0001 6 -5% -2%
CTP55 54868 1 101 60 2.01 531 73 10771 9 -3% -2%
NPBAT 44338 88462 2.00

 

 

Table 8. Molecular weight data for “unplasticized” and plasticized CA-PBAT
samples showing actual and theoretical Mm“, and MW“, values and their %
difference values

 

 

 

 

 

 

 

 

0 fl 1 r

14 16 18 20 22
Minutes

 

 

 

 

 

Figure 16. Chromatogram of samples affected by varying
contents of plasticizer and cellulose acetate: CP37 (0%
plasticizer, CTP37 (9% plasticizer), CTP55 (15% plasticizer)

 

 

 

It can be seen from Table 8 and Figure 16 that amongst the blends, the
unplasticized blend had the lowest average molecular weight, while the
plasticized blend had higher average molecular weight properties for reasons
explained earlier. The theoretical values of Mn and Mw for the samples are

shown in Table 8 with the percent differences as compared to the actual values

74

(first set of average molecular weights). Theoretically, since plasticizing CA
allows for higher reactions between PBAT and CA, blends of CA and PBAT will
naturally exhibit better than additive properties, and will not follow the
incompatible curve as predicted by the Kienle equation.

Effect of plasticizer and plasticized compositions on the thermal stability of
the blend can be seen in Figure 17 and Table 9. Addition of plasticizer
decreases the thermal stability of the blend by being easily volatilized or forced
out of the blend. The lower degradation temperatures could also be due to small
amounts of triacetin reacting as well. This decrease in thermal stability is

increased for higher CA loadings because increased CA in the blend increases

the overall percentage of plasticizer in the blend as well.

 

Sample TDTA (°C) TDCA (°C) TDPBAT (°C)TDCA-TA (°C)
CA 344.59
CT73 267.24 351 .26 --- ---
CP37 353.03 376.70
CTP37 140.98 352.68 375.73 ---
CTP55 1 60.65 350.88 383.04 ---
PBAT 373.88
NPBAT 375.24

 

 

 

Table 9. Degradation temperatures for
"unplasticized" and plasticized blends

 

75

 

 

 

 

 

 

 

120
o CP37
‘ I:I CTP37
100 .. g - + CTP55
: o NPBAT
80-
g .
40-
20-
: Eli—~32
0 T—‘ I I I I 1 r v u I v u v
100 150 200 250 300 350 400 450 500
Termeratue('C)

 

 

 

 

Figure 17. TGA weight loss curves for "unplasticiz " and plasticized
samples

 

 

 

Similar to compositional trends earlier, as the CA component increases,
degradation temperatures shift towards the degradation temperature of 100%
CA, while PBAT component temperatures remain about the same for all blends
except CTP55. The derivative weight loss curves, whose peak gives the
degradation temperature, seem to be shifting apart as shown in Figure 18.

Again, at a composition of 30 wt. % plasticized CA, the blend displays a slightly

76

higher degradation temperature for the CA component suggesting some

compatibility between the two components.

 

 

 

 

 

 

 

 

 

0.05
l
o CTP37
004‘ El CTP55
E"
50.03-
E
E
8’
.1
30.021
0.01-
0,00——.‘_L-,...,...,-..,fi..,..-..,..—
320 340 360 380 400 420 440 460 4130
Temperature ('C)
Figure 18. TGA derivative weight loss curves comparing plasticized
compositions

 

 

 

Unlike thermal transitions in DSC for physical blends of CA and PBAT
without plasticizer, the DSC results for plasticized blends of CA and PBAT show
no transitions for the CA component of the blend, as shown in Table 10. The
glass transition temperatures in Table 10 that are below zero were taken from

DMA results for a better approximations of T9. Values of T9 above zero

77

 

93mph -__°P°(°I ,T- 46°) Tm l° __ _ “1c" ( 7,0,, (° C) “and,“

      

 

CA -- 207.00 234.25 9.39 --- --- 202.35 5.48
CP37 -24.37 193.81 1 15.44 10.17 59.67 11.72 203.21 2.48
CTP37 -39.63 --- 105.95 5.95 61.91 8.29 ---
CTP55 -36.46 --- 101.79 4.46 54.36 4.48

PBAT -26.26 --- 1 15.27 18.49 25.88 13.99 ---
NPBAT -27.90 --- 123.31 13.33 76.55 17.09 --- ---

 

 

Table 10. DSC results for "unplasticized" and plasticized blends with varying
plasticized blend compositions; * indicates values from DMA

 

 

 

were similar for both DSC and DMA. Blends that were plasticized showed a
decrease in the PBAT component Tg values with no identified glass transition
temperatures present for CA, while unplasticized blends had the opposite effect
on PBAT component Tg values and did have glass transition temperature for CA.
Plasticized CA and PBAT blends also showed an appreciable decrease in
melting temperatures and extent of crystallization of PBAT as compared to
unplasticized blends. The higher crystallization rates temperatures of PBAT are
attributed to the more flexible amorphous segments of PBAT and smaller
plasticizer molecules allowing for chains to order themselves quickly, but not
necessarily to the extent that they would have without the increased flexibility.
Compositionally, melting and crystallization properties (temperatures and
enthalpies) of the PBAT component decrease with increasing CA content in the
blend.

Dynamic mechanical results of the blends in this segment are shown in
Figures 19, 20 and 21. The storage modulus (Figure 19) is higher for plasticized
blends in agreement with molecular weight data. Increasing the amount of CA in

the blend, increases the storage modulus. Plasticized samples show higher

78

storage moduli’s due to the enhanced reactions occurring as compared to
unplasticized blends of CA and PBAT. This is the cause for the great difference
in the storage modulus at low temperatures. It is interesting to note that the
storage modulus for the unplasticized blend of CA is about the same as NPBAT
at cooler temperatures. The difference in the slopes of the storage modulus
curves can be seen in the loss modulus curves (Figure 20) which show lower Tg
values for the plasticized samples, as discussed earlier. This difference in the
slopes of the storage modulus curves, as well as the difference in the width of the
loss modulus curves results in broad tan delta curves for the samples with higher
CA loading, as shown in Figure 21. Broader curves indicate less interaction
between the two components. Broadening of the curve usually lowers the curve

peaks as well [77].

79

 

 

25

 

  
 
    

; O CP37
20‘ D CTP37
. + CTP55
O NPBAT
a .
g 15~
fl .
2
g I
m 1
g 10«

 

 

Temperature (°C)

 

 

 

Figure 19. DMA storage modulus of "unplasticiz " and plasticized
blends along with varying compositions of plasticized blends

 

80

 

 

 

 

 

 

 

2.5
c CP37
20‘ u CTP37
+ CTP55
O NPBAT
$1.54
‘3 .
3
3
’3 .
2 I
1.04
3 .
0.58
0 0 r I , I , I ‘2 mgr—r _ _
80 .60 -40 -20 0 20 40 60
Temperature (’C)

 

 

 

Figure 20. DMA loss modulus of "unplastlciz " and plasticized blends
along with varying compositions of plasticized blends

 

81

 

 

0.5

 

. e CP37
0'4, [ll :1 CTP37
‘ I + CTP55
/ , o NPBAT

0.3 -

Tan Delta

 

  
  

 

W8060-4E7-20'I'01'IZOYI'40
TemperatueCC)

 

 

 

 

 

Figure 21. DMA tan delta of "unplasticized" and plasticized blends along
with varying compositions of plasticized blends

 

 

 

The use of ESEM images of the surface of blend pellets from reactive
extrusion allowed for an enhanced view of the morphology. The results from
ESEM are in Figure 22 at magnifications of 1000x, 2000x and 5000x. Each
column represents a different sample while each row represents the
magnification. Cellulose acetate and PBAT form a heterogenous blend from two
homogenous polymers. In these images spherical particles of CA are dispersed

in a PBAT matrix. Only shape and size of particles are discussed.

82

In Figure 22, at 1000x magnification, samples CP37 and CTP37 both
show some distinct spherical particles, whereas sample CTP55 shows
expanded, distorted spherical shapes suggesting CA aggregation. At 2000x
magnification, there seems to be some adhesion and deformation of CA particles
in CP37 and CTP37. The plasticized blend (CTP37) shows increased interaction
and deformation over the unplasticized (CP37) blend. Thus the addition of
plasticizer helps increase the deformation of CA particles, and enhance the
interaction between CA and PBAT phases. Compositionally, additional CA
seems to decrease interactions between the two components and increase the
interactions of CA with itself as seen for CTP55 at a 2000x magnification. The
spherical particles are less distinct and CA material is spreading across the
matrix. Magnifications at 5000x estimates particle size to be the lowest for the
plasticized blend of CA and PBAT at 30 wt. % plasticized CA at 1 — 2.5 pm. The
unplasticized blend had comparable particle size results (1 — 3 pm). Increased
CA content showed higher range of particle sizes (2.5 — 4 pm). Smaller particle

sizes are associated with higher dispersity of one phase in another [42].

83

 

 

Figure 22. ESEM images of CP37, CTP37 and CTP55 by column from left to right,
at 1000x, 2000x, SOOOx by row from top to bottom. Values displayed at the bottom
represent particle sizes for each sample

 

 

 

Effect of Catalyst and Composition of Plasticized Blends

Catalyst addition was expected to cause transesterification reactions
between the CA and PBAT. Table 11 shows the average molecular weight

values and Figure 23 shows the chromatogram for the uncatalyzed and

84

 

catalyzed blends. In Table 11, CTP37 is plasticized cellulose acetate and PBAT
in a 3:7 ratio, CTP55 is plasticized cellulose acetate and PBAT in a 5:5 ratio,
CTPD37 is plasticized cellulose acetate and PBAT in a 3:7 ratio with catalyst,
and CTPD55 is plasticized cellulose acetate and PBAT in a 5:5 ration with
catalyst. The “D” in the sample names represent the catalyst dibutyltin oxide.

Transesterification is expected to be the main reason for the significant
decrease in both Mn and MW values, and the increase in PDI for catalyzed blends,
as compared to uncatalyzed blends of the same composition. Differences in
composition of catalyzed blends are similar to before in that increased CA
content increases average molecular weight properties of the sample. The
significant decrease of the molecular weight properties of the catalyzed blends is
also expected to be affected by the triacetin in the blend. This is because
triacetin also has ester components and is most likely involved in the

transesterification reactions.

 

Sample M..(Dalt.) Mw(Dalt.) PDI

CTP37 51851 101667 1 .96
CTP55 54868 1 10160 2.01
CTPD37 36930 78492 2.1 3
CTPD55 42135 91302 2.17
NPBAT 44338 88462 2.00

 

Table 11 . GPC average molecular weight data of
uncatalyzed and catalyzed blends

 

 

 

85

 

 

-—-—-— CT P37
---—-— CT P55
-- - ——— CTPD37
— CTPDSS
— NPBAT

  
  
   

 

 

 

14 16 18 20 22
Minutes

 

 

 

 

Figure 23. GPC chromatogram of uncatalyzed and catalyzed
blends

 

 

 

Thermal properties give an even greater indicator that plasticizer is
participating in transesterification reactions. The degradation temperatures are
tabulated (Table 12) for the uncatalyzed and catalyzed samples. Catalyzed
samples show the least stability of all when compared to the unplasticized and
plasticized blends (Figure 24). Cleavage of the triacetin chains within the

catalyzed blends causes the volatile triacetin molecules to escape and/or

 

Sample TDTA (°C) TDCA (°C) TDPBAT (°C) TDCA-TA (°C)
CTP37 1 40.98 352.68 375.73 ---
CTP55 1 60.65 350.88 383.04 ---

CTPD37 127.51 346.81 375.44 307.02
CTPD55 137.65 346.18 381 .82 293.14
NPBAT 375.24 ---

 

Table 12. TGA degradation temperatures of uncatalyzed and
catalyzed blends

 

 

 

86

 

 

 

 

 

 

120
100~ .
, M
' ERR
80~
.3 i
g, 60: “00% CA
g ‘ EICTP37
+CTP55
40: OCTPD37
J ICTPDSS
20~
o..%,efi.....,..-,...
0 100 200 300 400 500
Tempereture(°C)

 

 

 

 

Figure 24. TGA weight loss curve of uncatalyzed and catalyzed blends

 

 

 

degrade at lower temperatures as compared to uncatalyzed blends. The
degradation temperatures of the CA component are lower for catalyzed blends
than they are for the uncatalyzed blends. Degradation temperature effects on
the composition of catalyzed blends are the same as before in that the TGA
derivative weight loss curves of each component shift apart from each other

suggesting increased incompatibility (Figure 25).

87

 

 

 

 

 

 

0.04
eCTPD37
DCTPD55

0.034

C

3:

i?

D

20.02

3‘?

E .

0.01-

0.00

250 300 350 400 450 500
Temperature (°C)

 

 

 

 

Figure 25. TGA derivative weight loss curve comparing compositions of
catalyzed blends

 

 

 

Thermal transitions from DSC show that catalyzed blends decrease Tg

further as compared to the uncatalyzed blend (Table 13). The glass transition

88

 

Sample Tgpa *(OC) TmPB (°C) A"Imps (Jlg) TcPB (°C) AHcPB (Jlg)

CTP37 -39.63 105.95 5.95 61.91 8.29
CTP55 -36.46 101 .79 4.46 54.36 4.48
CTPD37 40.87 1 10.58 7.62 68.89 9.95
CTPD55 -41.56 105.01 4.89 61.59 5.31
NPBAT -27.90 123.31 13.33 76.55 17.09

 

 

Table 13. DSC transitions for uncatalyzed and catalyzed blends; *
indicates values from DMA

 

 

 

temperature below zero is obtained from DMA. The decrease in T9 between the
uncatalyzed and catalyzed blend with higher CA wt. % is greater than the T9 of
the catalyzed and uncatalyzed blends at lower CA wt. °/o. These lower Tg values
are a result of shorter chains from transesterification. These shorter chains are
also responsible for increased crystallization and melting properties of catalyzed
blends as compared to their uncatalyzed counterparts. The differences in the
melting and crystallization properties of the uncatalyzed and catalyzed blend with
higher CA wt. % are, again, greater than the differences for the uncatalyzed and
catalyzed blend with lower CA wt °/o. Compositionally, melting and crystallization
properties (temperatures and enthalpies) of the PBAT component decrease with
increasing CA content in the blend.

Dynamic mechanical results of the blends in this segment are shown in
Figures 26, 27 and 28. The storage modulus (Figure 26) shows opposite effects
for uncatalyzed versus catalyzed blends at the varying compositions. The blend
that is somewhat compatible (30% plasticized CA content) shows a higher

storage modulus at low temperatures compared to its uncatalyzed counterpart.

89

On the other hand, the blend that is less compatible or even incompatible (50%
plasticized CA content) shows a lower storage modulus at low temperatures
compared to its uncatalyzed counterpart. The reason for this is that the
catalyzed sample with higher CA loading may have greater aggregation that
forced the triacetin to separate out and reduce the overall plasticizer in the blend,
reducing the storage modulus at low temperatures. The difference in the slopes
of the storage modulus curves can be seen in the loss modulus curves (Figure
27) which show lower Tg values for the catalyzed samples as discussed earlier.
This difference in the slopes of the storage modulus curves, as well as the
difference in the width of the loss modulus curves result, again, in broad tan delta
curves for the samples with higher CA loading as shown in Figure 28. Again,
broader curves indicate less interaction between the two components and more

interaction among the chains within their respective components.

90

 

 

 

. o CTP37
20. D CTP55

‘ + CTPD37
‘ O CTPD55
‘ I NPBAT

Storms Modulus (GPa)

L

 

 

 

 

-£l0'r'-60H'-4'0H1-2'0H'0'It2‘0‘ 4,0...60. '80
Temperature(‘C)

 

 

 

Figure 26. DMA storage modulus curves for uncatalyzed and catalyzed
blends

 

91

 

 

 

 

 

O CTP37
D CTP55
+ CTPD37
O CTPD55
l NPBAT

Loss Modulus (GPa)

 

 

\

 

 

 

 

Figure 27. DMA loss modulus curves for uncatalyzed and catalyzed
blends

 

92

 

 

 

 

0.5

o CTP37

0 4‘ n CTP55
- + CTPD37
‘ \ o crposs

- NPBAT

Tan Della

 

 

 

 

\~ .._..-.--
\-...

‘W’
0,0....,...,......e,...,...
-80 ~60 -40 -20 0 20 40

 

 

Temperature (’C)

 

 

Figure 28. DMA tan delta curves for uncatalyzed and catalyzed blends

 

 

 

The results from ESEM are in Figure 29 at magnifications of 1000x, 2000x
and 5000x. Each column represents a different sample while each row
represents the magnification. Each column representing the catalyzed sample
images is labeled as such. Again, in these images spherical particles of CA are
dispersed in a PBAT matrix. Only shape and size of particles are discussed
with a focus on the catalyzed samples since the uncatalyzed samples were

discussed earlier.

93

In Figure 29, all the images of the catalyzed samples show less interaction
with PBAT. Spherical particles are clearly defined and aggregated in the
CTP037 sample unlike in the uncatalyzed counterpart. Further distorted, (as
compared to uncatalyzed counterpart) spherical shapes suggests increased
aggregation of CA in CTPD55. The same results are seen at 2000x
magnification and with evidence of less adhesion and interactions of the two
components with each other and more adhesion and interaction of the two
components with themselves. Compositionally, the CTPDSS shows even further
spreading of CA material across the matrix. Magnifications at 5000x shows
enlarged particle size for catalyzed blends with largest particle size of ~ 4pm
related to CTPDSS. The particle shapes in this sample almost seems cylindrical

with spheres being hard to identify.

94

Catalyzed _ ‘ . Catalyzed

 

 

Figure 29. ESEM images of CTP37, CTPDa7, CTP55 and CTPD55 by column from
eft to right, at 1000x, 2000x, 5000x by row from top to bottom. Values displayed at
the bottom represent particle sizes for each sample.

 

 

 

4.4 Summary

Blends of CA with PBAT were studied as unplasticized and plasticized
blends at different compositons, and as uncatalyzed and catalyzed blends at
varying composition. Experimental results show that CA and PBAT can be
reactively extruded to yield a compatible blend. Interactions between the
polymers are confirmed by various characterization techniques. One example is

the increased molecular weight and decreased PDI values of the blends as

95

compared to PBAT alone along with other characterization identifiers.
Regardless, there is a limit to the compatibility of CA in PBAT.

The overall results show cellulose acetate and PBAT do in fact disrupt the
crystallinity of each other. Thus, increasing CA content increases the disruption
of PBAT crystallinity. In this study, PBAT proved to be a polymeric plasticizer for
CA in that the melting endotherm was eliminated and the T9 of CA was reduced
with increasing PBAT content. While external plasticizers did show increases in
molecular weight as compared to unplasticized blends and to PBAT, they also
resulted in decreased thermal stability most likely due to their high volatility. An
increase in plasticizer content seemed to increase access to CA hydroxyl groups
allowing for the reaction with PBAT acidic end groups. While higher average
molecular weight properties were attained with higher CA loading, TGA results
suggested decreased compatibility with increased CA loading.

The addition of catalyst in the blends showed a decrease in blend Tg
values and an increase in crystallization properties as compared to uncatalyzed
blends. The morphology of the blends was significantly affected by catalysts.
Catalyst addition promoted decreased interaction between the two components.
Triacetin was assumed active in transesterification reactions and, also believed
to play a role in how the properties of the blends were affected.

While plasticization seemed to increase compatibility, it severely
decreased thermal stability. Comparing the unplasticized and plasticized
samples at a ratio of 3:7 CA (or plasticized CA) to PBAT, both showed promising

behavior in terms of interaction with PBAT. In line with the goals of this study,

96

however, the unplasticized blend of CA and PBAT would be further developed in
order to yield a higher amount of biobased material being incorporated into
PBAT. Therefore, the physical blend of CA and PBAT, with a 30 wt. % loading of

CA, proved to be the best blend suiting the purposes of this work.

97

Chapter 5

 

Synthesis of Glycerol-based Materials

 

5.1 Overview

The development of biobased, biodegradable blends plays an important
role in the overall success of sustainability in the 21St century. While polymer
blending is a common route to achieving polymers with desired properties,
synthesis of new polymers is another essential route to producing sustainable
polymers. Development of these new polymers can increase the possibilities for
many more blends to achieve desired properties.

To develop such a polymer, Carothers’ theory, as described earlier, is
used to determine the relationship between percent conversion and degree of
polymerization. One focus of synthesis reactions is to build molecular weight,
and molecular weight is built by pushing the reaction to as high a conversion as
possible without gelation, which would correspond to a high degree of
polymerization. This work is focused on the transesterification of dimethyl
terephthalate with glycerol. Glycerol has a functionality of 3 and dimethyl
terephthalate has a functionality of 2. Calculating the average functionality as
givien in equation 21, Carothers’ equation (equation 24) can be used to estimate
degree of polymerization as a function of percent conversion.

Figure 30 shows the results of Carothers’ equation for a 1:1 and 2:3 ratio
of glycerol to DMT. It would seem natural to take two moles of glycerol and

react it with three moles of DMT since glycerol has three hydroxyl groups and

98

DMT has two methyl ester groups for reaction. However, for a two to three molar
ratio of glycerol to DMT, gelation occurs at a conversion of about 85% and the
maximum degree of polymerization is about 60. On the other hand, 1:1 ratio of

glycerol to DMT would have the potential to get to higher degrees of

 

 

 

 

 

 

 

 

 

 

 

 

Number Average DP (Xn)

 

 

 

 

 

O 0.2 0.4 0.6 0.8 1 1.2
% Conversion

 

 

 

Figure 30. Degree of polymerization and gel point for a 1:1 and 2:3
mixture of glycerol:dimethyl Terephthalate

 

 

 

polymerization (~upwards of 95) and gelation would only occur at extremely high
conversions. It is important to note that Carothers’ theory is simply an estimate
and in practice gelation is known to occur before the theoretical value. In this
study, a 1:1 ratio of glycerol to DMT is used as the basis for forming new
biobased polyesters. Overall, this study focuses on the development of a new
biobased, biodegradable polyester starting with a multifunctional alcohol and
reports on the molecular weight, hydroxyl values, and thermal properties

associated with the new pre-polymer.

99

5.2 Materials and Methods

The monomers used in this synthesis were anhydrous glycerol and
dimethyl terephthalate (DMT). Two catalysts, dibutyltin(lV) oxide and sodium
carbonate, were used. Other chemicals used for characterization techniques
include pyridine, acetic anhydride, n-butanol, tetrahydrofuran, and inhibitor-free
tetrahydrofuran.

The synthesis procedure used in this study is described in the materials
and methodology chapter. The characterization experiments include methanol
collection, hydroxyl value, and gel permeation chromatography, along with

thermal analysis of the reaction products.

5.3 Results and Discussion

In the transesterification reaction to produce glycerol-based biopolyesters,
various catalyst concentrations were run to determine catalyst performance and
identify an optimal catalyst and catalyst concentration. Specific reaction
conditions were described earlier in this chapter, but are summarized and shown
with the reaction matrix in Figure 31. All reactions were run under the same
conditions using a 1:1 molar ratio, varying only the catalyst type and the catalyst
amount. Sodium carbonate, denoted as NC, and dibutyltin (IV) oxide (DBTO)
were used in concentrations of 0.1, 0.2 and 0.3 wt. % of dimethyl terephthalate
(DMT). The reaction start time was defined as when the reaction vessel reached
a temperature of 200 °C. Methanol collection was recorded over the four-hour

reaction period and was expected to be about 202 mL based on stoichiometry.

100

The methanol data was used to estimate percent conversion (P) and the number-

average degree of polymerization (Xn).

 

 

 

 

 

 

Figure 31. a. Reaction conditions b. Reaction Matrix

 

Effect of Catalyst Type and Concentration on Conversion

The synthesis of glycerol-based pre-polymers (GDMT) led to an
understanding of the reaction of glycerol with the methyl ester, dimethyl
terephthalate, and its behavior with two types of catalyst. Sodium carbonate and
dibutyltin oxide were the catalysts of choice. This study focused on the
development of pre-polymers for further development of glycerol-based
materials. Using the reaction conditions shown in Figure 31, the conversion over
time and the degree of polymerization as a function of conversion are shown in

Figures 32 and 33.

101

 

101%
”lo
31%
70%
HP/o
51%
(KP/o
31%
21%
10%

0%

Corwersion (%)

 

0 50 100 150 200 250 300 $0
Th'e(rdn)

 

 

 

 

Figure 32. Conversion over time based on recorded methanol
collection

 

 

 

The percent conversion was calculated based on recorded methanol
collection. A theoretical value of 202.54 ml of methanol based on stoichiometry
was expected to form upon full conversion of the reactants to products. Thus,
methanol collection provided a crude estimate of the conversion of reactants that
could be gathered quickly and easily. Literature shows that 50% conversion is
seen within the first hour of the reaction [51, 52]. In this study, 50% conversion
was seen within the first hour and a half of the reaction. Figure 33 shows the
theoretical behavior of Xn as percent conversion increases with respect to all the
reactions in the study as a comparison to see how far along the reactions have

proceeded from a degree of conversion perspective.

.102

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10 l

9 + carom i'
g + DBT002 I
- 0
E DBTOO3 I,
“5 ' 7 ,. NC01 1
E, 6 -— NC02
3 + N003
9-6 5 — -Theoretical Data
3 4
g 3
X 1 “HIM

o T I i I

0.0 0.2 0.4 0.6 0.8 1.0
Percent Coverslon

 

 

 

 

Figure 33. Degree of polymerization as calculated by Carothers’
Theory for the theoretical curve and by experimental data for all
other curves

 

 

 

It is important to note that methanol formation began at lower
temperatures (150 °C) with sodium carbonate and at higher temperatures (170
°C) when DBTO was used. After four hours, DBT002 and DBTOO3 reactions
seemed to lag behind reactions DBTOO1 and all the NC catalyst reactions.
Amongst the reactions, conversions ranged from about 70 to 80% and the
degree of polymerization, from about 3 to 6. In order to get a high degree of
polymerization for high molecular weight in step-growth reactions, conversions
approaching 100% should be attained as shown in Figure 33 as Xn approaches
infinity. The final conversion and degree of polymerization values calculated

from methanol formation are tabulated in Table 14.

103

As a more reliable estimation of the extent of reaction, hydroxyl tests were
performed for all the reactions completed and the results are also shown in Table
14. The acid and hydroxyl value yielded conversions ranging from roughly 75-
77% and degrees of polymerization averaging about 4. Reactions NC02 and
DBTOOl showed the highest conversion and degree of polymerization values

within their respective catalyst group, relative to the results based on other

 

 

reactions.
Acid Hydroxyl % %
Value Value Conversion Conversion Xn" Xn*
Prepolymer (AV) (HV) (OH ff) (MEOH) (OH #) (MEOH)

NC01 0.8 310 76.78 84.04 4.3 6.3
NC02 0.4 300 77.54 83.10 4.5 6.0
N003 0.4 336 74.88 84.14 4.0 6.4
DBTO-01 6.1 297 77.77 83.74 4.5 6.5
DBTO-02 2.1 299 77.59 78.62 4.5 4.7
DBTO-03 5.0 305 77.19 70.54 4.4 3.4

 

*Xn is the degree of polymerization

 

Table 14. Comparison of percent conversion and Xn based on hydroxyl
values versus methanol formation

 

 

 

In the later stages of the reaction the viscosity increased significantly,
possibly trapping some volatile materials (mostly methanol). The amount of
methanol mixed in with the reaction product could potentially influence the
hydroxyl values significantly. To account for the possibly trapped methanol in the
samples, TGA was used to estimate the amount of volatiles from room
temperature up to 200°C (the run temperature of the reaction).

Figures 34 and 35 show the degradation curves for the NC and DBTO

catalysts sets. It was assumed that the majority of the volatile material

104

represented trapped methanol. For the sodium carbonate catalyst results shown
in Figure 34, reaction NC02 showed the least amount of trapped methanol,
suggesting that this catalyst composition was the either the furthest along in
terms of conversion or highly branched. For the dibutyltin oxide catalyst, the
reaction DBTOO1 showed the least amount of trapped methanol while the other
reactions in this catalyst group showed around 2.5% weight loss up to 200°C
(Figure 34). Comparing the results of the two catalysts suggests that the sodium
carbonate catalyst, when compared by weight %, did show lower amounts of
trapped methanol possibly indicating being slight further along, in terms of

conversion, than the reactions using dibutyltin oxide.

105

 

 

 

 

 

 

 

120
' 1.856% oNCOi
. 1.575% nNC02
1001 g _ 1.929% +N003
: \
80-
g .
z 60-
g .
40-
20-
o...,...,...,.-.,...
0 100 200 300 400 500

Temperature (°C)

 

 

 

Figure 34. TGA derivative weight loss curve for NC samples showing %
trapped methanol in the samples

 

 

 

106

 

 

1.745%

1 2.652% e DBTOO1

100. __ V_ f “55% D DBT002
‘ -‘ ~ - ' + DBT003

 

80-

Weight (%)

 

I I I I '

UIUI100"'200"'300 400II'5lJO
Temperatue('C)

 

 

 

 

 

 

Figure 35. TGA derivative weight loss curve for DBTO samples showing %
trapped methanol in the samples

 

 

 

The results of the amount of methanol trapped are then used to
recalculate the hydroxyl values as shown in Table 15. By subtracting the weight
of methanol in the sample from the weight of the sample for the hydroxyl value
test the hydroxyl values could be adjusted. While reaction conversions are close,
and may not be statistically different, the samples N002 and DBTOOl exhibit the

highest conversions based on hydroxyl value relative to the other samples.

107

 

Acid Hydroxyl % %
Value Value Conversion Conversion Xn* Xn*

 

Prepolymer (AV) (HV) (OH 11) (MEOH) (OH if) (MEOH)
mm 0.9 316 76.36 94.04 4.2 6.3
NC02 0.4 305 77.20 93.10 4.4 6.0
N003 0.4 342 74.41 94.14 3.9 6.4
DBTO-O1 6.1 310 76.80 93.74 4.3 6.5
DBTO-02 2.1 330 75.29 79.92 4.0 4.7
DBTO-03 5.0 312 76.62 70.54 4.3 3.4

 

‘Xn is the degree of polymerization

 

Table 15. Comparison of percent conversion and Xn based on
hydroxyl values versus methanol formation after adjustments for
traooed methanol

 

 

 

Conversion based on methanol displays maximums both within the NC
and DBTO catalyst group (Figures 36 and 37). Conversion based on OH passes
through a maximum at 0.2 wt. % NC, while in the DBTO catalyst group
conversion shows a decreasing trend with 0.1 wt. % DBTO as the optimal

catalyst amount for that group. Hydroxyl value tests were only performed on

 

0.9)
0.90 1
0.70 1
0.60
0.50
0.401“ ~------ -—- - — __.__,., -—- —' »—~* -
03+ . fl . 7W, .2, ._

0.201 W , *A 5% *5 -, 55-7 ~

0.10 -. _, ._ -._ 2. *9 ,, 44. ”-2, .-__ . . -_-. _. __.._-.

00m T l T T I 1
MI) 015 0.10 0.15 0.20 0.25 0.1!) 0.35

Catalyst Amount (we/own

 

 

 

 

 

 

 

 

 

 

 

 

 

Conversion based on MEOH

 

 

 

 

 

Figure 36. Catalyst perfomance based on MEOH for different
types of catalysts at different concentrations

 

 

 

108

 

0.78
0.78 "" [BTO

 

 

 

 

 

 

 

077 ~——— -

0.77 .__- ~f~

0.79 -— *

076~~—

0.75 ~——*

 

Conversion based on OH

 

0.75 2,- ~-

 

 

 

0-74 I I I T l I
0.1!) 0.05 0.10 0.15 0.20 0.25 0.30 0.35

Catalyst Amount (wl% DMT)

 

 

 

 

 

Figure 37. Catalyst perfomance based on hydroxyl # for different
types of catalysts at different concentrations

 

 

 

samples with a high enough conversion to yield a transparent solid mass. This
resulted in the methanol-based catalyst performance curves having more points

than the hydroxyl-based catalyst performance curves.

_Effect of C_atalvst Type and Concentrat_i9n on Molecular Weight

Although the conversion results only ranged from 74 — 77% conversion,
the trends seen from hydroxyl values agreed closely with the weight-average
molecular weights derived from GPC. The curves associated with the products
of these reactions were very broad (Figure 38) indicating high polydispersity
indices as shown in Table 16. It is worthwhile to note the high MW associated
with N002 and DBTOOl reactions. These results suggest the high conversion of

the glycerol hydroxyls and, possibly, the formation of a highly branched structure.

109

From a catalyst standpoint, sodium carbonate yielded higher Mn and MW values
overall when compared to DBTO. However the reaction DBTOO1 had

comparable results to the sodium catalyst group.

 

 

 

 

 

 

 

 

 

 

 

 

 

30.00-
20.00;
4
.1
10.00:
> -
I so
a . . ,
0 N —0 “PW;
-1000:
1
cm I I I?“ I I I Id I I I15l00] I—Ifzolmf I I 125Im] I VTaOIJ ffT F I I I IwImI 1 fifitm
Minutes
Figure 38. Chromatogram for all reactions with different catalysts and
different catalyst amounts

 

 

 

 

 

Mn Mw

Prepolymer (Daltons) (Daltons) PDI
N001 2455 7880 3.20
N002 2750 10767 3.92
N003 2487 7778 3.13
DBTO-01 2666 12542 4.70
DBTO-02 2233 4846 2.21
DBTO-03 2209 6257 2.83

 

 

Table 16. GPC Molecular weights and
polydispersity index for all reactions

 

 

 

110

_Effect of (fitalyst Type and Concentration on Thermal Properties

The thermal results follow the trends of the results discussed previously.
Table 17 summarizes the degradation temperatures for the polymer systems with
different catalysts and varying amounts of catalyst. The Figures 39 and 40 show
the degradation curves for the NC and DBTO catalysts groups, respectively.

Each sample exhibits a small weight loss at around 290 °C and 385 °C. The

 

 

Sample To (°C) TDGDM‘I’ (°C)
NCO1 293.58 386.23
NC02 285.59 386.41
NC03 282.65 385.79
DBTOO1 287.1 7 386.04
DBT002 267.47 384.41
DBTOO3 265.87 382.80

 

 

Table 17. TGA degradation
temperature values for both catalyst
sets

 

 

 

lower degradation temperatures are attributed to glycerol—based oligomers while
the higher degradation temperatures are associated with the glycerol-based pre-
polymers. For each catalyst set, the optimal concentration based on degradation
is highlighted. All samples show an overall increase in thermal stability as

compared to glycerol or DMT alone.

The degradation temperatures for the systems containing sodium
carbonate are at about 385°C, agreeing strongly with chemical results showing
reaction N002 with the highest degradation temperature relative to the varying

amounts of the sodium carbonate catalyst. On the other hand, degradation

lll

temperatures range from 381 to 385 °C for the DBTO catalyst group. There is a
slight difference between the DBTOO1 reaction and the other DBTO reactions as
illustrated in Figure 6. For DBTOOl a weight loss of about 9% can be seen at
294°C while this same weight loss is at 263°C for the other DBTO catalyst

amounts. This supports the idea of a highly branched system possibly forming a

 

 

 

 

 

network.
120i 009
1 eN001
100. nNcoz
‘ \E“ . +N003 ~004
80‘ \ .
6‘
1 \ ' t:
it? ~ '5.
z 90- t ~002 i
i .. g
. _ \ z
40 - . 3
+000
20-
o.fi.....,...4...,.-.,.......0.02
150 200 250 300 350 400 450 500
Temperature('0)

 

 

 

Figure 39. Sodium Carbonate percent and derivative weight loss curves
based on different catalyst amounts

 

 

 

112

 

 

 

 

 

 

 

 

120 0.05
‘ 1 e DBTOO1 .
. D DBT002
,0, .1 -0...
‘ ... + 091003 _
80~ ~0.03 A
. _ o
i. a
. ll '
o-I _ j '; _ -—
g 60‘ I 0.02 g
5 II 2 5‘5
‘ 1 I I), . E
40— 11\ ~0-01 8
2, _, \.
o . . . , . . . , . . . , . . . , . . . . - . . 001
150 200 250 300 350 400 450 500
Temperature ('0)

 

 

 

Figure 40. Dibutyltin oxide percent and derivative weight loss curves based

on different catalyst amounts

 

The only transitions identified from DSC were glass transitions; this, and

the absence of melting properies (temperature and endotherm) indicates the

formation of an amorphous polymer typical of alkyd resins. Figure 41 shows the

glass transition behavior for all reaction products of the study. Glass transition

temperatures ranged from 26 to 33 °C for NC reactions, and 23 to 31 °C for

DBTO reactions, as shown in Table 18. The highest glass transition

temperatures corresponded to DBTOO1, as expected from earlier results. Within

113

 

 

 

 

0.2

 

0 NC01
El NC02
+ NC03
O DBTOO1
I DBT002
<> DBTOO3

0.1~

Heat Flow (Wig)
O
G

-0.1-

 

 

 

 

-0.2 . , . . . , . . . . . .
- 20 40 60 80
Temperature (°C)

M
C)
o -1

 

 

 

Figure 41. DSC glass transition temperature results for reaction products
with varying amounts of a. Sodium Carbonate and b. Dibutyltin oxide

 

 

Sample TgGDMT *(oc)

 

N001 26.09
N002 26.21
N003 32.94
DBTOO1 31 .01
DBT002 23.26
DBTOO3 1 9,14

 

 

Table 18. DSC glass transition
temperatures of all reaction
samples

 

 

 

114

 

 

the NC reactions, however, NCO3 yielded the highest glass transition
temperature followed by N002. The reaction N003 also displayed the most

volatiles, suggesting it was the least along in terms of of conversion. This

result may be attributed to branching of the N002 reaction. Branched structures
help to lower Tg values. Thus sample N002, although believed to have reacted

further than sample N003, exhibits a lower Tg value than N003.

5.4 Summary

Homogenous, pre-polymeric glyceryl materials were synthesized from
glycerol and dimethyl terephthalate yielding an amorphous highly branched or
network polymer. Conversions varied slightly depending on the catalyst system
and amount of catalyst used, but all conversions ranged from about 74 - 77%
based on hydroxyl values and 78 — 84% based on methanol formation. Values
for Xn ranged from 4 to 6.5 depending on the basis of the calculations.

Molecular weight data revealed broad curves indicative of highly branched
structures. The highest number average molecular weight was 2750 Daltons
with PDI’s as high as 4.7. Thermal results displayed behavior suggesting
trapped volatile materials in the reaction product but showed increase in thermal
stability as compared to glycerol or dimethyl terephthalate alone.

Based on the overall results, dibutyltin oxide proved to be the optimal
catalyst at 0.1 wt. % DMT. Although sodium carbonate at 0.2 wt. % showed
comparable results, less dibutyltin oxide on a weight basis is necessary for the

similar results. On a molar basis, even less dibutyltin oxide is necessary to

115

produce the results as compared to sodium carbonate. Successful preparation
of glyceryl pro-polymers shows future potential for the development of network

polymers based on glycerol.

116

Chapter 6

 

PBAT-Biopolyester Preliminary Blends

 

6.1 Overview

The exciting part of developing a new polyester is the really the prospect
of all the potential blends that could be formed with new properties. This is
exciting because it helps to increase the understanding of the behavior of the
polymers interacting. What is even more intriguing is the fact that this new
biopolyester incorporates at least 50% glycerol and the blending of this
biopolyester to any polymer would add biobased content and, likely,
biodegradation properties to the blend.

In this study PBAT is being used as the preliminary polymer for blending
with the pre-polymer developed in this work (G DMT). PBAT was chosen over
cellulose acetate because of the greater potential of PBAT to be affected by the
addition of this pre-polymer and the ease with which it could be processed.
Benefits and properties of both PBAT and GDMT were discussed in previous
chapters. The objective of this work was to begin to explore the effects of GDMT
blending with other polymers by studying the thermal and dynamic mechanical

properties of the blend

6.2 Materials & Methods

The materials used in this study include PBAT and GDMT. The methods

for processing included a reactive extrusion system and the DSM for preparation

117

of test specimens. The characterization techniques employed included
differential scanning calorimetry, thermal gravimetric analysis and dynamic

mechanical analysis.

6.2.1 Specific Processing Methodology

The PBAT was used as received from BASF. The GDMT was crushed as
uniformly as possible by using a flathead hammer against small blocks of the
material wrapped in a thick cloth. Reactive extrusion of PBAT and GDMT
occurred using the following temperature profile from feed to die:
25/95/155/180/180/180/180/180/180/170 (°C). The melt temperature was around
175 °C and the motor speed was set at 100 RPM. The PBAT, as received from
BASF, as well as the blends of 60 and 40 wt. % of PBAT and GDMT,
respectively, were prepared using the twin-screw extruder system under the

given conditions.

6. 2. 1 Specific Characterization Methods

For DSC, the sample temperature equilibrated at -50 °C and was then
held isothermally at that temperature for two minutes. The temperature was then
increased to 200 °C at a rate of 10 °C/min and held isothermally for two minutes.
Finally the temperature was returned to 0 °C at a rate of 10 °C/min. Thermal
data was gathered from this scan.

The TGA and DMA methods used are as described earlier in the general

characterization section.

118

6.3 Results and Discussion

6.3.1 Summary of Runs

Three different GDMT samples were used in this experiment to explore
catalytic effects at different concentrations. All sample compositions were 40 wt.
% GDMT and 60 wt. % PBAT. Table 19 shows the compositional details for the

three samples in this study.

 

 

 

 

 

 

 

 

. PBAT! GD .33
m& W96 '
PGDNC01 60 4O
P_GDNC02 00 40
PGDDBO1 60 40

 

 

 

Table 19. Summary of Runs for GDMT-PBAT
blends

 

 

 

Effect of Residual Catalyst Type and Concentration on Thermal Properties
Incorporating the low molecular weight prepolymer developed in this work,
under reactive extrusion conditions, resulted in two degradation temperatures
with the lower degradation temperature being associated with a small weight %
loss of possibly transesterified material due to the residual transesterification
catalyst in the blend. The lower degradation temperature (TDESTER) is attributed
to either small segments of the PBAT polymer chains or glycerol residues
degrading as a result of transesterification reactions with GDMT. The GDMT-

PBAT degradation temperatures, regardless of the sample, remained around the

119

degradation tempearature of PBAT. Table 20 summarizes the degradation
temperatures associated with the samples in this study. The bulk of the weight

loss for all samples occurred closer to the PBAT degradation temperatures.

 

Sample Toesren (°C) TDPBAT (°C)
PBAT --- 373.68
NPBAT 375.24
PG DNCO1 326.66 374.54
PG DNC02 329.39 373.88
PGDDBO1 341.85 376.42

 

 

Table 20. TGA degradation
temperature for GDMT-PBAT samples

 

 

 

Figure 42 shows the TGA weight loss curve for the samples. All blends
showed a slight decrease in thermal stability as compared to NPBAT (again,
referring to PBAT extruded once) as the oligoesters are degrading. The sample
with 0.1 wt. % dibutyltin oxide (PGDDBO1) dislayed the greatest decrease in
weight loss with increasing temperature towards GDMT-PBAT, GDMT and PBAT
degradation temperatures, compared to the other samples. This could be due to
highly branched GDMT reacting with PBAT and forming slightly greater network

structures, than lower branched GDMT reacting with PBAT (PGDNCO1).

120

 

 

120

0 PBAT

D NPBAT
+ PGDNCO1
O PGDNC02
l PGDNCOB

100

 
 
   

801

60-

Weight (%)

401

20-

 

 

 

do

T I I U I I I I V T T f I Y I V I

00' ' '150‘ ' Y200 250 300 350 400 450 ' '500
Temperature (°C)

 

 

 

Figure 42. TGA weight loss curve for GDMT-PBAT samples

 

 

 

The DSC results support the idea of further reaction of GDMT with PBAT
for all samples. Table 21 summarizes the thermal data for the samples. For
glass transitions below zero the table shows values obtained from DMA. For
glass transition temperatures above zero, DMA and DSC results are shown for

comparison.

121

 

53me TgPBAT i"(”c) TgGDMT (°C) TgGDMT *(oc) Tm (°C) "tn (”9) ToPeAr (°C) HcPBAT1J’9)

 

PBAT -26.26 --- 1 15.27 18.49 25.88 13.99
NPBAT -27.90 --- 123.31 13.33 76.55 17.09
PGDNC01 -22.01 38.01 38.56 105.96 10.25 63.59 7.88

PG DNCOZ --- 33.42 m 101.20 13.20 58.42 12.28
PG D0801 -23.77 37.22 38.58 108.73 12.22 65.34 10.48

 

 

Table 21 . DSC thermal transitions for blends of GDMT-PBAT; * indicates values
obtained from DMA

 

 

 

Addition of GDMT to PBAT resulted in an increase in the glass transition
temperatures of both GDMT and PBAT components of the blend as revealed by
DSC and DMA results. This increase in the Tg suggests a decrease in chain
flexibility possibly caused by the reaction of several chains of PBAT to various
OH groups of GDMT. Glass transition temperatures from DSC above 0 °C show
about the same values as those from DMA, as expected.

The decrease in chain flexibility of all the blends causes a slight decrease
in the melting temperature and the melting endotherm as compared to NPBAT
because stiffer, amorphous chains are not allowing the chains of PBAT to order
as quickly to form more crystalline material. Thus the lower amount of chains
that have ordered and crystallized melt at about 7 to 14 °C lower than PBAT.
Less flexibility of the amorphous segments of PBAT also hinder crystallization to
a small extent. In addition, branching could be playing a role in the extent of
crystallization as seen in the relative comparison of samples that are thought to
have higher branching (PGDNCO2 and PGDDBO1) than others (PGDNCO1 ).

Lower branched GDMT material would affect longer chain segments of PBAT,

122

whereas higher branched GDMT material would be limited to shorter PBAT chain
segments.

The decrease in crystallization endotherms show a similar effect that could
be explained by the branching explanation given earlier. Figure 43 depicts the
slight shifts in crystallization temperatures for the samples. Comparing residual
catalysts, the sample PGDDBO1 had higher melting and crystallization properties
as compared to PGDNCOl. Increase in residual catalyst resulted in the slight
lowering of melting and crystallization temperatures, while increasing the extent
of crystallization. The higher branching at 0.2 wt. % sodium carbonate
(PGDNCO2) is, again, potentially responsible for this behavior by interfering with

less of the PBAT chain segments.

123

 

 

0.9

0.8- . e PGDNCOl
1 it 1:: PGDN002
' '\ + PGDDBO1
‘ 1 . o NPBAT
07“ I PBAT

 

Heat Flow (Wig)
o
O)

    
   

(I W.

1 ‘ v 7 ' 53:6,;
.Jf _

0.4 -

.‘3
01

 

 

 

0.3
0 20 40 60 90 100 120 140

Temperatue ('0)

 

 

Figure 43. DSC crystallization transition for blends of GDMT-PBAT

 

 

 

Effect of Residual Catalyst Type and Concentration on _Dvnamic Mecha_nipal
Properties

The possible formation of network-like structures with the reaction of
GDMT and PBAT is the reason for the increased storage modulus seen for the
sample with 0.1 wt. % dibutyltin oxide (PGDDBO1) shown in Figure 44. The
storage moduli comparing blends with residual catalysts are shown in Figure 44
with PBAT and NPBAT. While the sodium catalyst carbonate blend shows only a

slight increase in storage modulus at low temperatures as compared to NPBAT,

124

the blend containing residual dibutyltin oxide showed a greater increase at lower

temperatures as compared to NPBAT.

 

 

 

 

 

 

 

 

 

20
j e PBAT
* \ D NPBAT
15- \ + PGDN001
\ o PGDD801
' \
Ti \
i=9, j \
S
E 101 ‘ \
.9
(I)
5 1 \
. “k...
I RR
\
0 T I r I f r F fi' 1 bk-
90 00 40 -20 o 20 40 60 80
Temperature ('0)
Figure 44. Storage modulus of GDMT-PBAT blends

 

 

 

The damping curves shown in Figure 45 support the idea that
transesterification is taking place. The ratio of the height of the two peaks per
sample shown in Figure 45 is reported to approximate the amount of each
material in the mixture [77]. In this case, the tan delta representing the PBAT-
based component of the mixture is lower for the PGDDBOf because more

reactions have taken place in this mixture decreasing the amount of the pure

125

PBAT in the mixture. As the temperature increases, there is a shift in the GDMT-
based tan delta curve and the sample PGDDBOl is now higher than the

PGDN001 curve indicating a higher concentration of PBAT-GDMT materials in

the PGDDBOf mixture.

 

 

0-6

0 PBAT r1
1: NPBAT / .

 
   

+ PGDNC01
O PGDDBO1

 

Tan Delta

 

 

 

 

'50 "40' ' '-20' "0' ' '2'0' ' '40' ' '60" 90
Temperature(°C)

 

 

 

Figure 45. DMA tan delta curve for GDMT-PBAT blends

 

 

 

126

6.4 Summary

Based on the results of the thermal and dynamic mechanical data,
reaction of GDMT with PBAT is believed to have caused the formation of some
network-like structures. These structures only cause slight decreases in the
melting and crystallization properties (temperatures and endotherms) of pure
PBAT. Inaddition, the amorphous chains of both PBAT and GDMT components
increase slightly resulting in a stiffer amorphous segment. In this study, while
transesterification did result in the degradation of some oligomers at lower
temperatures as compared to PBAT, transesterification also resulted in stronger
networks between PBAT and GDMT components. The stronger network allows
for an the addition of biobased materials to PBAT without dramatically affecting
the thermal properties of PBAT. More transeterification occurred with the
dibutyltin oxide catalyst at 0.1 wt%, and increasing the concentration of the '
catalyst (at least in the case of sodium carbonate) seemed to promote more
crystallization as compared to sodium carbonate at 0.1 wt% (most likely due to
the branching effects of the sample containing 0.2 wt. % residual sodium

carbonate).

127

Chapter 7

 

Conclusions and Recommendations for Future Work

 

7.1 Conclusions
The goal of this research was to increase the biobased content of
synthetic materials by the addition of the biobased materials glycerol and
cellulose acetate. To accomplish this goal, three main objectives were met:
0 The effects of composition, plasticizer and catalyst on blends of CA and
PBAT were studied
0 Catalyst performance on the glycerol-based biopolyester were determined
a The effects of catalyst type and catalyst amount on blends of GDMT with
PBAT were studied
The reactive extrusion route to incorporating biocontent into PBAT showed
promising results. Results showed that when PBAT and CA were physically
blended with no additives, PBAT acted as a plasticizer, reducing the glass
transition temperature of CA with increasing PBAT content. While the addition of
CA reduced the extent of crystallinity and slightly increased the T9 of the PBAT
component in the blend, the overall blends exhibited a higher Mn as compared to
PBAT alone due to acidic end groups of PBAT reacting with hydroxyl groups of
CA. Results of plasticized CA blends versus unplasticized CA blends showed
that plasticized blends allowed for more interaction of CA with PBAT, thereby
increasing reactions between the two resulting in increased molecular weights as

compared to the unplasticized blends. All of the plasticized blends decreased

128

the extent of PBAT crystallinity further than the unplasticized blend. Catalysis
results showed that the properties of the uncatalyzed blends were superior to
those of the catalyzed blends, for reasons mostly associated with the
interference of triacetin in the esterification reactions as triacetin is of low

molecular weight and does have ester groups.

Overall, blends of CA and PBAT did show interactions between the CA
and PBAT, which suggests that although the blend was not miscible, blends were
compatibilized via the reactive extrusion of the two polymers. In addition, the
plasticized CA sample at 70% loading of PBAT showed some of the best
properties as compared to the other blends. However, the unmodified blend of
CA with 70% PBAT loading was more promising given the goals of this study.
This is because of the larger amount of CA actually within the blend. (30 versus

21 wt. % for unplasticized (CP37) and plasticized (CTP37) blends, respectively).

Biocontent was incorporated in the synthesis design phase of the
transesterification reaction, which resulted in a transparent solid. Catalyst
performance was a crucial step in obtaining a maximum degree of
polymerization. For the NC catalyst, a concentration of 0.2 wt. % catalyst was
determined to have the highest conversion and molecular weight values. At the
same time the polydispersity increased significantly, suggesting branched pre-
polymers. The sample containing the DBTO catalyst at 0.1 wt. % was found to
be the optimal catalyst given the range of concentrations studied. While the M.
of the mixture was slightly lower, as compared to sodium carbonate at 0.2 wt. %,

the MW and PDI at this concentration were the highest of all the samples studied.

129

This is most likely a result of how much branching occurred and which end
groups were added to the pre-polymer (acidic DMT or basic glycerol). Themlal
studies showed that degradation temperatures associated with both catalyst
groups were about the same at around 385 °C. However at 0.1 wt. % DBTO,
degradation temperatures were comparable to the NC temperatures. Results
from DSC showed higher glass transition temperatures overall for the samples in
the NC catalyst group and lower glass transition temperatures for the samples in

the DBTO catalyst group, except at 0.1 wt. % of DBTO.

Overall, while some major differences could be seen amongst the
concentrations within the DBTO catalyst group, the differences within the NC
catalyst group were small. However these small differences did show a specific
trend. This suggests that a wider range of catalyst concentrations may have
shown more of a difference in results. While on a weight-basis the samples
containing catalyst DBTOOl and N002 exhibited similar results, on a molar basis
much less DBTO was necessary to produce about the same or, in some cases,
better results. Successful incorporation of glycerol at over 50% in the synthesis
of GDMT resulted in a significant amount of biocontent in the developed pre-

polymer.

Biocontent was, again, incorporated into PBAT via GDMT. The use of
these pre-polymeric glycerol-based materials in blends seems promising.
Reactive extrusion of GDMT and PBAT showed that increasing catalyst
concentration decreased melting and crystallization temperatures due to the

branching effects of the GDMT with 0.2 wt. % residual sodium carbonate. The

130

addition of GDMT also slightly decreased the extent of PBAT crystallinity, but not
to the level that it was decreased in the plasticized CA blends. This could be
attributed to these prepolymers only affecting small segments of the PBAT
chains. Comparing the catalysts, DBTO promoted more reactions as compared
to NC. This resulted in slightly higher DSC transition results for DBTO versus NC
at equal wt. %. Although incorporation of biocontent to PBAT was the
overarching goal of this work, the objective of this study was to study residual
catalyst performance in promoting reactions with PBAT. Based on this study the
residual dibutyltin oxide at 0.1 wt. % did the best job (on wt. and molar basis) in

maintaining overall PBAT properties while adding biocontent.

Addition of biocontent through PBAT and DMT with the use of CA and
glycerol proved to be a viable route to the advancement of biobased polymers.
Out of this work a novel bio pre-polymer that is solid and transparent was
produced, and has a high potential for being a biopolyester in the near future.

This work also resulted in biobased, biodegradable blends of CA and PBAT.

7.2 Recommendations for Future Work

Completion of this work has led to ideas for more work in this area of

increasing biocontent. Some areas for future work are listed below.

0 Transesterification of physical blends of CA and PBAT
o Maleation of PBAT for use in blends with CA

0 Compatibilization studies on CA and PBAT with maleated PBAT and the

determination of optimal compositions

131

Further development and characterization of glycerol-based polyesters to
higher degrees of polymerization

Blending of developed glycerol-based polyesters for applications coatings
as well as injection molded plastics

Re-blending of developed glycerol-based polyesters with PBAT

132

[1]

[2]

[31

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[111

[12]

[13]

References

Narayan, Ramani. Drivers for biodegradable/compostable plastics & role
of composting in waste management & sustainable agriculture.
Bioprocessing of Solid Waste & Sludge 1(1) 2001.

Johnson, Jeff. Ethanol-Is it Worth It?. Chemical & Engineering 85(1), 19-
21 , January 2007.

National Research Council. Biobased lnsdustrial Products: Priorities for
Research and Commercialization. The National Academy Press,
Washington, DC 2000.

Nabar, Y, Narayan, R, Schindler M. Twin-screw extrusion production and
characterization of starch foam products for use in cushioning and
insulation applications. Polymer Engineering and Science, 46 (4), 438-
451 April 2006.

Biobased Plastic Flexes Its Muscle. ScienceDaily.
http://www.sciencedaily.com/releases/2005/12/051 21 61 85231 .htm. 2005

Summary of the Kyoto Protocol. Energy Information Administration. '
http://www.eia.doe.gov/oiaf/kyoto/kyotobrf.html. 2002.

Duncan, Marvin. US. Federal Initiatives to Support Biomass Research
and Development. The Journal of Industrial Ecology, 7 (3-4) 2004.

Narayan, Ramani. Rationale and Design of Environmentally Degradable
Plastics. Standardization News, 17 (11), 40 (1989).

Brookes, Chisa K. A comparative life cycle assessment: polyethylene and
starch foams. MS. Thesis. Michigan State University, East Lansing, MI
2004.

Cellulose. Wikipedia: The Free Encyclopedia.
http://en.wikipedia.org/wiki/Cel|ulose 2007.

Encyclopedia of Polymer Science and Technology: Casting to Cohesive
Energy. John Wiley & Sons, Inc. Volume 3 New York 1965.

Fibrous Cellulose Acetate in Engineered Structures. Eastman Chemical
Company. May 1999.

Brydson, J. A. Plastic Materials. Newnes-Buttervvorths. London, (1975).

133

[14]

[15]

[15]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

Technical Glossary of Common Audiovisual Terms: Diacetate. National
Film and Sound Archive, ©Commonwealth of Australia.
http://screensound.gov.au/glossary.nsf/Pages/Diacetate?OpenDocument
2002.

Edgar, K.J., Buchanan, C.M. et al. Advances in cellulose ester
performance and application. Progress in Polymer Science 26 (2001 ).

Schauber, T., De Vos, S., et al. Phase behavior and mechanical
properties of blends of cellulose propionate and an alternating propene-
carbon monoxide copolymer. Macromolecular Chemistry and Physics
200(3) 1999.

Light R. R., Dombroski J.R., Hawkins J.M. Cellulose ester/copolyacrylate
composition. US Patent 4,263,183 (1981).

Nabar, Y, Gupta, A, Narayan, Ramani. Isothermal Crystallization Kinetics
of Poly(Ethylene Terephthalate)—Cellulose Acetate Blends. Polymer
Bulletin, 53, 117-125 (2005).

Ghiya, V. P., Dave, V. et al. Plasticization and Biodegradation of
Cellulose Acetate. SPE ANTEC 3726-3727 (1995).

Mohanty, A. K.; Wibowo, A.; Misra, M.; Drzal, L. T. Development of
renewable resource-based cellulose acetate bioplastic: Effect of process
engineering on the performance of cellulosic plastics. Polymer
Engineering and Science 43(5), 1151-1161 (2003).

Wibowo, A., Misra, M., et. al. Biodegradable nanocomposites from
cellulose acetate: Mechanical, morphological, and thermal properties.
Composites: Part A 37, 1428-1433 (2006).

Park, H., Mohanty, A. K., et al. Effect of Sequential Mixing and
Compounding Conditions on Cellulose Acetate/Layered Silicate
Nanocomposites. Journal of Polymers and the Environment, 14(1), 27-35
January 2006.

Rosa, D. 8., Guedes, C. G. F., Casarin, F., et al. The effect of the M-w of
PEG in PCUCA blends. Polymer Testing 24 (5), 542-548 August 2005.

Nie, L., Narayan, R. Grafting Cellulose Acetate with Styrene Maleic
Anhydride Random Copolymers for Improved Dimensional Stability of
Cellulose Acetate. Journal of Applied Polymer Science 54 (5), 601 -61 7
October 1994.

134

[251

[26]

[27]

[28]

[29]

[301

[31]

[32]

[33]

[34]

[35]

Lee SH, Yoshioka M, Shiraishi N. Polymer blend of cellulose acetate
butyrate and aliphatic polyestercarbonate. Journal of Applied Polymer
Science 77 (13) September 2000.

Maheras, J.C., Hopkins Jr., J.B., Tatzlaff, H. Polymer blend composed of
cellulose acetate and starch acetate and starch used to form fibers, films
and plastic materials and a process to prepare said blend. US Patent
5,446,140 (1995).

Buchanan C.M., Pearcy B.G., White A.W., Wood MD, at. Al. The
relationship between blend miscibility and biodegradation of cellulose
acetate and poly(ethylene succinate) blends. Journal of Environmental
Polymer Degradation 5(4), 209-223 October 1997.

Guruprasad, K. H., Shashidhara, G. M. Grafting, blending, and
biodegradability of cellulose acetate. Journal of Applied Polymer Science
91 (3), 1716-1723 February 2004.

Teramoto, Y., Nishio, Y. Biodegradable cellulose diacetate-graft-poly(L-
Iactide)s: Thermal treatment effect on the development of supramolecular
structures. Biomacromolecules 5 (2), 397-406 March-April 2004.

Teramoto, Y., Nishio, Y. Biodegradable cellulose diacetate-graft-poly(L-
Iactide)s: Enzymatic hydrolysis behavior and surface morphological
characterization. Biomacromolecules 5 (2), 407-414 March-April 2004.

Teramoto, Y., Nishio, Y. Cellulose diacetate-graft-poly(lactic acid)s:
synthesis of wide-ranging compositions and their thermal and mechanical
properties. Polymer 44 (9), 2701 -2709 April 2003

Braganca, F. 0., Rosa, D. S. Thermal, mechanical and morphological
analysis of poly(epsiIon-caprolactone), cellulose acetate and their blends.
Polymers for Advanced Technologies 14 (10), 669-675 October 2003.

Yoshioka, M., Hagiwara, N., Shiraishi, N. Thermoplasticization of
cellulose acetates by grafting of cyclic esters. Cellulose 6 (3), 193-212
September 1999.

Yoshioka, M., Okajima, K., Miyazaki, T., et al. Plasticization of cellulose
derivatives by reactive plasticizers II: characterization of plasticized
cellulose acetates and their biodegradability. Journal of Wood Science 46
(1), 22-31 (2000).

Teramoto, Y., Yoshioka, M., Shiraishi, N., et al. Plasticization of cellulose

diacetate by graft copolymerization of epsilon-caprolactone and lactic acid.
Journal of Applied Polymer Science 84 (14), 2621 -2628 June 2002.

135

[36]

[37]

[38]

[391

[40]

[41]

[421

[43]

[44]

[45]

[46]

[47]

Lee, S. H., Shiraishi, N. Plasticization of cellulose diacetate by reaction
with maleic anhydride, glycerol, and citrate esters during melt processing.
Journal of Applied Polymer Science 81 (1 ), 243-250 July 2001.

Suvorova, A. I. and Demchik, L. Y. Chemical structure of plasticizers,
compatibility of components and phase equilibrium in plasticized cellulose
diacetate. Die Makromolekulare Chemie, Volume 195 (5) 1315-1321
(2003)

Someya, y., Sugahara, Y. and Shibata, M. Nanocomposites Based on
Poly(butylene adipate-co-terephthalate) and Montmorillonite. Journal of
Applied Polymer Science, Volume 95, 386-392 9 (2005).

Nabar, Yogaraj U., Draybuck D. and Narayan, R. Physicomechanical and
Hydrophobic Properties of Starch Foams Extruded with Different
Biodegradable Polymers. Journal of Applied Polymer Science, Volume
102, 58-68 (2006).

Nabar, Y., Raquez, J. M., Dubois, P., et al. Production of starch foams by
twin-screw extrusion: Effect of maleated poly(butylene adipate-co-
terephthalate) as a compatibilizer. Biomacromolecules 6 (2), 807-817
March-April 2005.

Jiang, L., Wolcott, M. P., Zhang, J. Study of Biodegradable
Polylactide/Poly(butylenes adipate-co—terephthalate) Blends.
Biomacromolecules 7, 199-207 (2006).

Paul, D. R., Newman, S. Polymer Blends: Volume 1. Academic Press
Inc. New York 1978.

Rudin, Alfred. The Elements of Polymer Science and Engineering, 2nd
Edition. Academic Press, San Diego, CA 1999.

Young R.J., Lovell, P.A. Introduction to Polymers, 2nd Edition. Chapman
and Hall, Cambridge 1991.

Baber, T, M. A Novel Application of Ozone Chemistry for Biodiesel
Improvement: Product Development and Characterization. Doctoral
Dissertation. Michigan State University, East Lansing, MI 2005.

Rodriguez, Ferdinand. Principles of Polymer Systems, 4th Edition. Taylor
and Francis, Washington, DC 1996.

Kienzle, S. Y. Polymer Blends and Alloys: Guidebook to Commercial
Products. Technomic Publishing pp. 1-6 (1988).

136

[43]

[49]

[50]

[51]

[52]

[53]

[54]

[55]

[56]

[57]

[58]

Schlumberger oilfield glossary. Schlumberger, Ltd.
http://www.glossary.oilfield.slb.com/search.cfm 2007.

Scheirs, J. and Long, T. Modern Polyesters. John, Wiley and Sons, Ltd.,
West Sussex, England 2003.

Kienle, R. H.. Alkyd resins, development of and contributions to polymer
theory. Journal of Industrial and Engineering Chemistry (Washington,
DC), 41, 726-729 (1949).

Kienle, R. H.; Hovey, A. G. Polyhydric; alcohol-polybasic acid reaction, I.
Glycerol-phthalic anhydride. Journal of the American Chemical Society,
51, 509-519 (1929).

Kienle, R. H.; Hovey, A. G. Polyhydric alcohol-polybasic acid reaction. ll.
Ethylene glycol-phthalic anhydride. Journal of the American Chemical
Society, 52, 3636-3645 (1930).

Kienle, R. H.; Van der Meulen, P. A.; Petke, F. E. Polyhydric alcohol-
polybasic acid reaction. Ill. Further studies of the glycerol-phthalic

anhydride reaction. Journal of the American Chemical Society, 61,
2258-2268 (1939).

Kienle, R. H.; Van der Meulen, P. A.; Petke, F. E. Polyhydric alcohol-
polybasic acid reaction. IV. Glyceryl phthalate from phthalic acid. Journal
of the American Chemical Society, 61, 2268-2271 (1939).

Kienle, R. H.; Petke, F. E. Polyhydric alcohol-polybasic acid reaction. VI.
The glyceryl adipate and glyceryl sebacate polyesters. Journal of the
American Chemical Society, 63, 481 -484 (1941 ).

Kienle, R. H.; Petke, F. E. Polyhydric alcohol-polybasic acid reaction. V.
The glyceryl succinate and glyceryl maleate polyesters. Journal of the
American Chemical Society, 62, 1053-1056 (1940).

Odian, George. Principles of Polymerization, 4th Edition. John, Wiley and
Sons, Inc., Hoboken, NJ 2004.

Otton, Jean; Ratton, Serge. Investigation of the formation of poly(ethylene
terephthalate) with model molecules: kinetics and mechanism of the
catalytic esterification and alcoholysis reactions. l. Carboxylic acid
catalysis (monofunctional reactants). Journal of Polymer Science, Part
A: Polymer Chemistry, 26(8), 2183-2197 (1988).

137

[59]

[60]

[51]

[62]

[53]

[64]

[55]

[66]

Otton, Jean; Ratton, Serge; Vasnev, V. A.; Markova, G. D.; Nametov, K.
M.; Bakhmutov, V. l.; Komarova, L. I.; Vinogradova, S. V.; Korshak, V. V.
Investigation of the formation of poly(ethylene terephthalate) with model
molecules: kinetics and mechanisms of the catalytic esterification and
alcoholysis reactions. ll. Catalysis by metallic derivatives (monofunctional
reactants). Journal of Polymer Science, Part A: Polymer Chemistry,
26(8), 2199-2224 (1988).

Otton, Jean; Ratton, Serge; Markova, G. D.; Nametov, K. M.; Bakhmutov,
V. l.; Vinogradova, S. V.; Korshak, V. V. Investigation of the formation of
poly(ethylene terephthalate) with model molecules. Ill. Metal-catalyzed
esterification and alcoholysis reactions: influence of the structure of the
reactants and of the nature of the reaction medium. Journal of Polymer
Science, Part A: Polymer Chemistry, 27(11), 3535-3550 (1989).

Otton, Jean; Ratton, Serge. Investigation of the formation of poly(ethylene
terephthalate) with model molecules. IV. Catalysis of the esterification of
ethylene glycol with benzoic acid and of the condensation of ethylene
glycol monobenzoate. Journal of Polymer Science, Part A: Polymer
Chemistry, 29(3), 377-391 (1991).

Tsutsumi, Naoto; Chen, Ying Hsien; Kiyotsukuri, Tsuyoshi. Novel regular
network polyester films from aromatic tetracarboxylic acids and ethylene
glycol. Journal of Polymer Science, Part A: Polymer Chemistry, 29(13),
1963-1970 (1991).

Kiyotsukuri, Tsuyoshi; Tsutsumi, Naoto; Chen, Ying Hsien. Novel regular
network polyester films from benzenetricarboxylic acids and glycols.
Journal of Polymer Science, Part A: Polymer Chemistry, 28(5), 1197-
1208(1990)

Kiyotsukuri, Tsuyoshi; Tsutsumi, Naoto; Sugimoto, Mariko; Saito, Naomi.
Novel regular network polyamide films from trimesic acid and aliphatic
diamines. Polymer Communications, 31(2), 56-67 (1990).

Kiyotsukuri, Tsuyoshi; Kanaboshi, Masato; Tsutsumi, Naoto. Network
polyester films from glycerol and dicarboxylic acids. Polymer
International, 33(1), 1-8 (1994).

Nagata, M., Machida, T., Wataru, S. and Tsutsumi, N. Synthesis,
Characterization and Enzymatic Degradation of Network Aliphatic
Copolyesters. Journal of Polymer Science: Part A: Polymer Chemistry,
Volume 37, 2005-2011 (1999).

138

[67]

[68]

[69]

[70]

[71]

[72]

[73]

[74]

[75]

[76]

[77]

Pramanick, D., Pramanick, R., and Betal, G. Copolyester of Trimellitic
Acid, Glycerol, and Poly(ethylene glycol): Synthesis and Characterization.
Journal of Applied Polymer Science, volume 91, 343-346 (2004).

Shenoy, M., Parikh, P. Aromatic Polyesters Based on BisphenoI-A for
Liquid Insulating Systems: A Synthetic Approach. Journal of Applied
Polymer Science, Volume 95, 606-614 (2005).

Jikei, Mitsutoshi; Kakimoto, Masa-Aki. Hyperbranched aromatic
polyamides prepared by direct polycondensation. High Performance
Polymers, 13(2), 833-843 (2001).

Hao, Jianjun; Jikei, Mitsutoshi; Kakimoto, Masaaki. Preparation of
Hyperbranched Aromatic Polyimides via A2 + B3 Approach.
Macromolecules, 35(14), 5372-5381 (2002).

Lin, Qin; Long, Timothy E.. Polymerization of A2 with B3 monomers: A
facile approach to hyperbranched poly(aryl ester)s. Macromolecules,
36(26), 9809-9816 (2003).

Unal, Serkan; Long, Timothy E. Highly Branched Poly(ether ester)s via
Cyclization-Free Melt Condensation of A2 Oligomers and B3 Monomers.
Macromolecules, 39(8), 2788-2793 (2006).

Stumbe, Jean-Francois; Bruchmann, Bernd. Hyperbranched polyesters
based on adipic acid and glycerol. Macromolecular Rapid
Communications, 25(9), 921-924 (2004).

Wyatt, Victor T.; Nunez, Alberto; Foglia, Thomas A.; Marmer, William N.
Synthesis of hyperbranched poly(glycerol-diacid) oligomers. Journal of
the American Oil Chemists' Society, 83(12), 1033-1039 (2006).

ASTM D1980. Standard Test Method for Acid Value of Fatty Acids and
Polymerized Fatty Acids. Annual Book of ASTM Standards, Volume
11.01, 397-398 (1998).

ASTM 01980. Standard Test Method for Hydroxyl Value of Fatty Oils and
Acids. Annual Book of ASTM Standards, Volume 11.01, 375-376 (2001).

Nielson, Lawrence E. Mechanical Properties of Polymers. Reinhold
Publishing Corporation, New York 1962.

139

 

 

IBRARIES

I

1293 02956 4

L

E UNIVERSITY

A
T
S
N
m
H
m
M.

III
III‘

I
_ I

34

111”,

I

5

 

 

 

 

Z.
.11....

.n