How Plasticizer Makes a Ductile Polymer Glass Brittle?

Home , Plasticizer, Polymer solution

HOW PLASTICIZER MAKES A DUCTILE POLYMER GLASS BRITTLE?

A Thesis

Presented to

The Graduate Faculty at The University of Akron

In Partial Fulfillment

of the Requirements for the Degree

Master of Science

Yue Zhao

May, 2016 HOW PLASTICIZER MAKES A DUCTILE POLYMER GLASS BRITTLE?

Yue Zhao

Thesis

Approved: Accepted:

______Advisor Dean of the College Dr. Shi-Qing Wang Dr. Eric J. Amis

______Committee Member Dean of the Graduate School Dr. Mesfin Tsige Dr. Chand Midha

______Department Chair Date Dr. Coleen Pugh

ii ABSTRACT

During uniaxial extension, a polymer glass of high molecular weight is ductile at high temperatures (still below Tg) and turns brittle when the temperature is sufficiently lowered. Incorporation of small-molecular additives to polymer glasses can speed up segmental relaxation considerably. The effect of such plasticization should be to make the polymers more ductile. We examined the effect of blending a few weight percent of either triphenyl phosphate (TPP) or a mineral oil to commercial-grade PS, PMMA and

SAN. Our Instron tests showed that the plasticized PS and SAN were less ductile.

Specifically, at 70 oC, the original PS is ductile at an extensional rate of 0.02 min-1 1 whereas the PS with 4 wt. % mineral oil or TPP is brittle. Similarly, at 50 oC, the original

SAN is ductile at an extensional rate of 0.02 min-1, but the SAN plasticized by 4 wt. %

TPP becomes brittle. No such counterintuitive effect is observed in PMMA. This finding challenges the Eyring type idea2 of activation for flow since the mechanical spectroscopic measurements show that the alpha relaxation time of PS is shortened by more than two orders of magnitude in presence of 4 wt. % TPP. On the other hand, the effect of TPP on PS and SAN can be well anticipated from the recent molecular model 3.

iii ACKNOWLEDGEMENTS

I am sincerely thankful to my master thesis advisor Dr. Shi-Qing Wang for his patience and guidance throughout my graduate study career. All the achievements on my study and the project research rely on his mentorship. I am so impressed and motivated by his passion for truth in the science field. Especially, he helps me realize the importance of critical thinking.

I would also like to thank my thesis reader, Dr. Mesfin Tsige. As a student, I am so lucky to enjoy learning the knowledge about thermodynamics and the algorithm in his classes. Additionally, I am very grateful for his constructive suggestions on my formal seminal.

Finally, I want to express my strong gratitude to my group mates. Dr. Panpan Lin,

Xiaoxiao Li, and Jianning Liu have provided me so much advice on solving academic and experimental problems. And I want to thank Yexin Zheng and Sirui Ge for their help during my study.

iv TABLE OF CONTENTS

Page

LIST OF TABLES ...... vii

LIST OF FIGURES ...... viii

CHAPTER

I. BACKGROUND AND INTRODUCTION ...... 1

Plasticizers and Antiplasticization ...... 1

Traditional interpretations of antiplasticization ...... 2

II. EXPERIMENTAL...... 4

Materials and the Blending Process ...... 4

Tensile Extension ...... 7

Relaxation in Tensile Extension ...... 7

Small Amplitude Oscillatory Shear (SAOS)...... 8

Dynamic Mechanical Analysis (DMA) ...... 9

Differential Scanning Calorimeter (DSC) ...... 10

Gel Permeation Chromatography (GPC) ...... 11

III. EXPERIMENTAL RESULTS13Solubility of Mineral Oil and TPP in Polymers ...... 13

Decrease in Glass Transition Temperature ...... 14

Decrease in Yong’s Modulus and Yield Stress ...... 16

Uniaxial Tensile Extension ...... 17

SAOS Results and Shift of α Relaxation Time ...... 19

Relaxation in Tensile Extension ...... 20

IV. DISCUSSION ...... 23

v What Happened at a Breaking Point for a Brittle Polymer in the Molecular Picture? ...... 23

The Interpretation of Higher Brittleness ...... 25

V. SUMMARY ...... 27

REFERENCES ...... 29

vi LIST OF TABLES

Table Page

2.1 Chemical and Rheological Characteristics of Polymers ...... 4

2.2 Physical and Chemical Properties of Plasticizers ...... 4

vii LIST OF FIGURES

Figure Page

1.1 When the concentration of plasticizers was high enough, they not only filled in the voids but diluted the whole chain ...... 1

2.1 Triphenyl phosphate (TPP) ...... 5

2.2 Styrene acrylonitrile resin (SAN) with 33% acrylonitrile ...... 5

2.3 A diagrammatic sketch from the blending process to specimen making ...... 6

2.4 A diagrammatic sketch of uniaxial tensile extension ...... 7

2.5 Small amplitude oscillatory shear measurements of (a) polystyrene ...... 9

(b) plasticized polystyrene with 4 wt. % TPP, at a reference temperature Tref = 120 °C ...... 9

2.6 Schematic of a basic gel permeation chromatograph, including solvent, solvent delivery system, injector of sample, column(s), detector(s), and data system ...... 12

3.1 Stress strain curves of 4 wt. % TPP/PS blends that underwent different mixing time were compared ...... 14

3.2 Heat flow vs. temperature curves tested by DSC of (a) 0 %, 4% TPP/PS blends ...... 15

(b) 0 %, 2 % TPP/PMMA blends...... 15

3.3 Storage modulus (G’) vs. temperature curves tested by DMA for PS and 4 wt. % TPP/PS blends respectively ...... 16

3.4 Storage modulus (G’) vs. temperature curves tested by DMA for PMMA and 2 wt. % TPP/PMMA blends ...... 17

3.5 Stress strain curves of (a) PS ...... 17

(b) PS plasticized by 4 wt. % mineral oil in uniaxial tensile extension at 70 oC and 80 oC respectively, with a strain rate -1 V/L0 = 0.02 min ...... 17

viii 3.6 Stress strain curves of 4 wt. % TPP TPP/PS blends in uniaxial tensile extension at 70 oC, 75 oC, and 80 oC respectively, with -1 a strain rate V/L0 = 0.02 min ...... 18

3.7 Stress strain curves in uniaxial tensile extension of (a) SAN at 30 oC and 50 oC ...... 19

(b) SAN plasticized by 4 wt. % TPP at 45 oC, 50 oC, and 55 oC, -1 with a strain rate V/L0 = 0.02 min ...... 19

3.8 Stress strain curves in uniaxial tensile extension of PMMA at 55 oC and 2 wt. % TPP/PMMA at 50 oC, 55 oC, and 60 oC, -1 with a strain rate V/L0 = 0.02 min ...... 19

3.9 Comparison of α relaxation time at various temperature between PS and 4 wt. % TPP/PS blends ...... 20

3.10 Relaxation in extension denoting the mobility of polymer chains (b) showed the normalized curve of figure (a) ...... 21

3.11 Stress strain curve of stress relaxation at the same stress,19 MPa, of PS at 70 oC, and 4 wt. % TPP/PS blends at 60 oC, 63 oC, and 68 oC ...... 22

4.1 Description of formation of an LBS and the surrounding primary structure. The lac is a length to characterize the spread of activated primary structure regime. 3 ...... 24

4.2 (a) showed the structure of brittle polymer in our molecular model ...... 24

(b) was zoomed-out view of the former in the XY plane. 1 ...... 24

4.3 (a) Majority of primary structure was activated (drawn as gray dots) due to displacement of LBSs upon sufficient extension, so the material was ready to yield ...... 25

(b) A zoomed-out view of the former in the XY plane where most gray “rings” interpenetrate to show that the activation zones have spread to reach one another. 1 ...... 25

ix CHAPTER I

BACKGROUND AND INTRODUCTION

Before giving details of experiments and results, relative concepts and researches are summarized in this section.

Plasticizers and Antiplasticization

Plasticizers are low-molecular-weight additives in most of daily products to improve processing behavior. A plasticizing component can be gas, vapor, or liquid, and an example of a plasticization process was shown in Figure 1.1. In polymer glasses, effects of plasticizers include reduction in rigidity, increase of the elongation to break at room temperature, and increase of the toughness (impact strength) down to the lowest temperature of serviceability 4.

Figure 1.1 When the concentration of plasticizers was high enough, they not only filled in the voids but diluted the whole chain. 5

However, it had been found since sixty years ago 6, in some plasticized polymers, that the properties changed in the opposite direction compared with normal plasticization behavior when the concentration of the plasticizer was below a certain

value. This phenomenon is called “antiplasticization”, which involves reduction of chain mobility, increase in yield stress, increase in Young’s modulus, and decrease in strain at break 4. Yet, the glass transition temperature always decreases in presence of additives in both plasticization and antiplasticization.

Besides the two phenomena mentioned above, there are some other effects due to low-molecular-weight additives. For example, slightly plasticized polymer glasses can be more brittle, while both the yield stress and the Young’s modulus are decreased 7.

Traditional interpretations of antiplasticization

For antiplasticization, the traditional explanation is mainly based on the free volume concept. Below we briefly review the previous studies.

Anderson et al. 8 blended PS with mineral oil and used the three-point bending technique to test flexural properties in antiplasticization. The results showed the dependence of antiplasticization on polystyrene molecular weight. From positron annihilation spectroscopy (PAS) and nuclear magnetic resonance (NMR) results, their study also proposed a model based on the concept of free volume, where the mineral oil occupied the voids near chain ends to reduce free volume.

Kierkels et al. 9 focused on using the reduction of molecular mobility in antiplasticization to delay aging after rejuvenation of PS and PC with Kenflex (KFX).

However, they did not provide an account for how additives reduced chain mobility.

Morever, they were unable to explain why the transition from antiplasticization into plasticization occured upon increasing temperature.

10 Sundgren et al. mixed poly(vinyl chloride) (PVC) (Mw = 74000) with poly-r- caprolactone (PCL) (Mw = 14000). In the uniaxial tensile creep tests, the α-peak shifted to lower temperatures with increasing amount of PCL in the blend. And it was thought

to be due to a heterogeneous phase structure, because factors like density fluctuations, a certain degree of crystallinity, and uniform mixing existed in the material. However, these cannot explain the similar phenomenon in other amorphous or well-mixed blends systems. Besides, the β-peak was found compressed and shifted to lower temperatures, which was attributed to the pseudocrosslinking effect caused by a strong intermolecular interaction.

Robeson 11 agreed with Jackson and Caldwell on the effect of free volume reduction 7 and crystallinity 12 in antiplasticization. In Jackson and Caldwell’s research, they tested various plasticizers and antiplasticizers. Many literatures cited their conclusion as a definition of antiplasticization. However, the concentration of plasticizers in the polymers in Jackson’s measurements was between 20 % and 30 %, which must have involved dilution effects.

Usually, materials with higher modulus, higher yield stress, and more elongation before breaking are more desirable. Besides plasticization, antiplasticization was also exploited for other purposes. For example, the increasing elastic modulus of PMMA due to antiplasticization could help stabilize the PMMA nanostructures during lithographic processing 13. Jong Suk Lee 14 used incorporation of selected low- molecular-weight diluents into poly(ethylene terephthalate) to significantly improve oxygen and carbon dioxide barrier properties.

The investigation of (anti)plasticization effects can help engineers choose additives to suit a certain polymer material. A fundamental understanding the various phenomena is necessary and desirable. We expect it to come from intimate interactions between theoretically-motived experiments and molecular-level ideas.

CHAPTER II

EXPERIMENTAL

Measurements methods used included blending in a mixer, tensile extension, relaxation in tensile extension, small amplitude oscillatory shear (SAOS), dynamic mechanical analysis (DMA), differential scanning calorimeter (DSC) etc.. More details are shown below about materials and measurements.

Materials and the Blending Process

Table 2.1 Chemical and Rheological Characteristics of Polymers.

Polymer M (kg/mol) M (kg/mol) PDI ℃ W e Tg ( )

PS 319 13 1.44 103

SAN 116 7.2 1.59 107

PMMA 125 10 1.41 113

Table 2.2 Physical and Chemical Properties of Plasticizers.

Plasticizer Physical state Melting point (oC) Boiling point (oC)

Mineral oil Liquid -60 ~ -90 333.00 ~ 609.44

TPP Solid 50 220

The polymer glasses under study were polymethyl methacrylate (PMMA) from

Plaskolite West Inc. (item number CA-86), polystyrene from Dow (Styron 663), and poly (styrene−acrylonitrile) (SAN) from Diamond Polymers (item number SAN 51). 4

Table 2.1 showed their chemical and rheological characteristics including molecular weight (Mw), entanglement weight (Me), polydispersity (PDI), and glass transitiontemperature (Tg). White mineral oil (DRAKEOL® 600 MIN OIL USP) and triphenyl Phosphate (TPP) from TCI America were added into polymers as plasticizers, and their physical and chemical properties were listed in Table 2.2. In manufacture products, TPP performs not only as a plasticizer, but an ingredient in a controversial flame retardant. Anderson and Grulke’s gas chromatogram results showed that the mineral oil was comprised of primarily C-28 to C-46 hudrocarbons 8.

Figure 2.1 Triphenyl phosphate (TPP).

Figure 2.2 Styrene acrylonitrile resin (SAN) with 33% acrylonitrile

PS was blended with small molecular additives, i.e. plasticizers, in a Brabender®

Mixer type 30 EHT with the extension EHT (electric, high temperature) distinguished by their electric temperature conditioning in three control zones with compressed air cooling. The rotor speed ratio of 2:3 (driven to none driven) resulted in a high torque resolution which allowed a well differentiation. Before adding 75g PS and certain amount of plasticizer, the Brabender mixer was set 10 rpm, at 165 °C. While all the raw

materials were softened, the rate was raised to 40 rpm and the temperature stayed at

160 oC until the whole 30-minite process finished. The mixing process for SAN or

PMMA blends was similar. The only difference was that PMMA needed higher processing temperature, 170 oC. In order to guarantee the high precision of comparisons, polymer PS, SAN and PMMA without plasticizer were heated and stirred in Brabender under the same conditions as what the blends underwent.

The bulks of polymer and blends were compressed into transparent films at 180 °C and cut into pieces, before filling a Monsanto Capillary Rheometer to make cylindrical specimens. They were heated up in a barrel and a pressure was applied to compress the resin in the barrel, squeezing out air trapped. For PS, after being rested at 290F for half an hour, the melt was extruded using a capillary die of length L = 15D and diameter D

= 1 mm. The extrusion included three parts. First, the system was heated to 240 F for degas. Then increase the temperature to 290F for around half an hour for relaxation.

Finally, dog-bone-shape samples were extruded between 11 psi and 15 psi. For PMMA, the resin was firstly heated up to 480 F in the rheometer, and after 10 min resting without compressing it was cooled down to 440 F. Then a similar procedure was taken, i.e., compressing at 165 °C and half-an-hour relaxation at 480 F and extrusion. For SAN, the procedure was similar to that of PS. For all of them, the specimens have an effective length of 50 mm.

Figure 2.3 A diagrammatic sketch from the blending process to specimen making.

Tensile Extension

An Instron 5567 with an environmental chamber (model # 3119-406) was used in uniaxial extension measurement. The dog-bone-shape sample was fixed by upper and under clamps. An oven was used to heat the sample to the required temperature, giving thermo control at various temperatures. Before a test, the specimen was left in the oven at a stable temperature for 15 min to minimize the temperature difference between different parts of the specimen. The upper clamp went up at a speed of 1 mm/min.

Considering that the length of the specimens is 50 mm, the extensional rate was 0.02 min-1.

F A0 A

Figure 2.4 A diagrammatic sketch of uniaxial tensile extension. L0 was the original length in the direction of tensile extension. After stretching, the length L0 increased to

L, and the sectional area decreased from A0 to A. Assuming the material was uncompressible, the value of (L/ L0) was equal to the value of (A0/A).

Relaxation in Tensile Extension

Relaxation experiments were also done using the Instron 5567 with an environmental chamber (model # 3119-406). A constant crosshead speed of V = 1 mm/min was used to deform the sample to a prescribed strain or stress. According to

results in extension tests, the elongation in the post-yield relaxation was decided to be

0.1 (∆ l / l0), and the prescribed stress in the pre-yield relaxation was set to be 19 MPa.

The machine switched to the strain-controlled mode to maintain the strain level for one hour, and the decreasing stress was recorded as a function of time. So that the plot obtained was the stress-time curve where the decrease of stress vs. time denoted the mobility of polymer chains.

Small Amplitude Oscillatory Shear (SAOS)

The SAOS was used here to obtain the α relaxation time. The α relaxation time curves of SAOS were recorded on an ARES-G2 rotational rheometer from TA

Instruments. The sample was firstly sit on the 8 mm parallel plates in the rheometer and heated up to the lowest temperature of the series tests. In order to guarantee the good adhesion between the polymer disk and the metal plate, it was necessary to allow the polymer to relax at the required temperature for a time, longer than ten times of τ, before starting tests.

In small amplitude oscillatory shear tests, material functions based on the imposed strain and the stress response are defined quantifying the material behavior. Storage modulus (G’) is defined as ratios of stress and strain amplitudes, based on the amplitude of in-phase stress. Loss modulus (G”) is defined as strain amplitudes, based on the out- of-phase stress. At very low frequencies is G’ ∝ ω2 and G” ∝ ω-1. The crossover between G’ and G” occurs at the frequency ω = 1/π, and π is the characteristic time of

15 material response, i.e. relaxation time. One notation is to call πα the alpha relaxation time, which characterize the segmental mobility. For an entangled uncrosslinked polymer, there are three crossover points in the curve of G’ and G” versus ω, and the point with the largest value of ω is the one wanted, where ω = 1/πα (see Figure 2.5).

120 1%

Temperature of the reference curve = 120 oC. a. Strain maximum = 1%

1/πα

120C 1%

b. Temperature of the reference curve = 120 oC.

Strain maximum = 1%

1/πα

Figure 2.5 Small amplitude oscillatory shear measurements of (a) polystyrene and

(b) plasticized polystyrene with 4 wt. % TPP, at a reference temperature Tref = 120 °C.

Dynamic Mechanical Analysis (DMA)

In the dynamic mechanical analysis, a sinusoidal stress was applied and the strain in the material was measured, which was aimed to characterize many

dynamic properties like the complex modulus and glass transition temperature. The test mode used was temperature sweep mode. This approach could help locate the glass transition temperature of the material, as well as to identify transitions corresponding to other molecular motions. The temperature of the sample was varied in low constant frequency, leading to variations in the complex modulus. At the glass transition, the storage modulus decreased dramatically and the loss modulus reaches a maximum. In this DMA experiment, the strain maximum was 0.01 %, and the frequency is 1 Hz. The sample cut had a scale of 7 mm * 12 mm* 0.5 mm.

Differential Scanning Calorimeter (DSC)

Polymers are not in thermal equilibrium and none of any two polymer materials are exactly the same. A sample was put in DSC (Q10) with a reference and heated from the room-temperature to a high temperature above melting point at a constant heating rate. When thermal effects leaded to difference in temperature, compensated amplifier amplifies the current in hot wires, until the sample and reference were at the same temperature. The compensation of the power during this process was in correspond to the heat flowing, which reflects enthalpy of fusion.

The result of DSC was the curve of heat flowing as a function of temperature. The glass transition temperature of plasticized blends was expected no higher than that of the polymer, around 100 °C , so the test temperature was set from 60 °C to 150 °C. A slice of the sample was cut down, and it was weighed out to be about 5 mg, before being put into the aluminum pan with a lid covered above. After pressing and sealing the pan, it was placed in DSC with an empty reference. The protection atmosphere was N2. The parameters of the equipment and the experiment steps were set before turning on. As settings in order, the experiment process contained data storage off, temperature

jumping to 40 °C, 5 min of isothermal process, data storage on, and ramp 10 °C/min to

150.00°C. Finally, a curve of heat flow vs. temperature was obtained, where the location of a step change anticipated the value of Tg by an anal software automatically.

Gel Permeation Chromatography (GPC)

Gel Permeation Chromatography, abbreviated as GPC, is a widely used method for determination of molecular weight distribution (PDI), number-average molecular weight (Mn), weight-average molecular weight (Mw), Z-average molecular weight

(Mz), and viscosity molecular weight (Mv) 16.

All chromatographic techniques share a feature in common, that is the distribution of a solute between two phases. In GPC, one phase is a liquid stream flowing around the gel packing, called the mobile phase. The other phase is the same liquid stagnant in the pores inside the grains, called stationary phase 17. The interstitial volume around grains in the column, i.e. the mobile phase in column, is represented by 푣푖. The pore volume in the grains, i.e. the stationary phase, is represented by 푣푝. To calculate the elution volume (see equation 2.1), a partition coefficient k, is introduced. It implies the pore volume available to the macromolecules with elution volume v.

Elution volume v ≡ 푣푖 + 푘푣푝 (2.1)

Total valume 푣푡 = 푣푖 + 푣푝 (2.2)

While the total volume 푣푡 is the combination of the volume of mobile phase and the volume of stationary phase (see equation 2.2). For polymer, which can penetrate through the chromatographic columns, the value of k is more than zero and less than one. The bigger the value of k, the smaller the polymer species are.

Separation of polymer using GPC depends on the hydrodynamic volume of the long chains. Larger molecules penetrate through less pores, so they elute earlier than shorter chains. Polystyrene gels crosslink in the specific diluents performing as a stationary phase in columns. Gels are made in fine-mesh bead forming suitable size to pack into chromatographic columns. Polymers were dissolved in tetrahydrofuran (THF).

Thus a series of narrow molecular weight range polymer fractions could flow through columns. Part of the instrument is a continuous differential refractometer, which plays a role of detecting and recording the effluent concentrations. Gels crosslinked with much more diluents had little solvent action of polystyrene. Columns containing these gels can separate high molecular weight polymeric samples. On the contrast, less diluents brought in more solvent action, which led to lower molecular weight permeability limits 18.

Figure 2.6 Schematic of a basic gel permeation chromatograph, including solvent, solvent delivery system, injector of sample, column(s), detector(s), and data system.

CHAPTER III

EXPERIMENTAL RESULTS

In this section, all the results and curves are shown to characterize the mechanical and rheology properties of polymers by adding plasticizers. Information and analysis got directly from the results were also written as below.

Solubility of Mineral Oil and TPP in Polymers

When the concentration of a plasticizer in a polymer glass is beyond the limitation of solubility, phase separation will accrue. When it is the polymer solution of additives, the plasticizer phase without entanglement is brittle, and it cannot undergo as much elongation as ductile polymer materials. So it is necessary to conduct an advanced survey on the solubility between the plasticizer and tested polymer and verify it during practical experiments.

A handbook written by George Wypych 19 says that phosphate is highly recommended for PS and PMMA. To verify the good solubility and well diffusion of the TPP in PS, a blending process was stopped ahead using only 10 min, and the mechanical properties of this bulk were compared with those of standard specimens.

As shown in Figure 3.1, the stress strain curves of different bulks overlapped with each other at the temperatures 70 oC and 80 oC. Both blends mixed for 10 min and 30 min were brittle below 80 oC, and it became ductile at 85 oC, so the brittle-ductile transition happened in the range of 80 – 85 oC. The consistent test results proved the well diffusion of TPP in PS.

Anderson et al. 8 used high-temperature gas chromatography (GC) to analyze the270 000 MW, 128 000 MW, and 40 000 MW PS. And the test bars indicated a maximum concentration of 8, 9, 10 vol % of absorbed mineral oil respectively. They found that blends containing > 8 vol % mineral oil were opaque at room temperature, which was the evidence of phase separation. Similar phenomenon appeared in my experiments, blends containing 4 wt. % mineral oil were transparent, and 8 wt. % TPP blends were opaque. For the blends of PMMA or SAN with plasticizers in the experiment were all transparent.

25  (MPa) 10min 85C engr 70oC  (MPa) 10min 70C engr 20  (MPa) 30min 70C engr  (MPa) 30min 80C engr o 80 C  (MPa) 10min 80C 15 engr

(MPa)

engr  10 85oC

5 4 wt. % TPP/PS blends V/L = 0.02 /min 0 0 1 1.02 1.04 1.06 1.08 1.1 1.12 1.14 1.16 

Figure 3.1 Stress strain curves of 4 wt. % TPP/PS blends that underwent different mixing time were compared.

Decrease in Glass Transition Temperature

From the results in literatures 20, glass transition temperature (Tg) decreases in both plasticization and antiplasticization, no exception. Figure 3.2 (a) and (b) are the

DSC results of PS with 4 wt. % TPP/PS blends, and PMMA with 2 wt. % TPP/PMMA blends. Tg sits at the step change in the DSC curves. In the transition of polymer from 14

glassy state to melt state, its heat capacity became higher, so that the heat flow increased.

The change of heat flow covered a temperature range, because of factors like molecular weight dispersity and complexed molecular structures. Here the upper turning point of the incline was taken to be the Tg. As shown below, for PS, Tg decreased from 105 to

94 oC by adding 4 wt. % TPP, i.e. a reduction of around 10 oC. For PMMA, after adding

2 wt. % TPP, Tg decreased from 119 to 113 oC.

2 (a) DSC resultss PS + 4 wt. % TPP

1.5

Heat Flow (mW) Flow Heat

0.5

o 94 oC 105 C 0 0 50 100 150 200 Temperature (oC)

2 (b) DSC results

1.5 PMMA + 2% TPP

PMMA 1

Heat Flow (mW) Flow Heat

0.5

113 oC 119 oC 0 0 50 100 150 200 Temperature (oC)

Figure 3.2 Heat flow vs. temperature curves tested by DSC of (a) 0 %, 4% TPP/PS blends and (b) 0 %, 2 % TPP/PMMA blends.

Since the storage modulus (G’) decreased sharply in the glass transition temperature range, DMA results also indicated the glass transition temperature shift. As shown in Figure 3.3, the position of upper turning point of G’ vs. temperature curves was taken as Tg. The Tg of PS and 4 wt. % TPP/PS blends were 114 and 102 oC

respectively. So the magnitude of Tg reduction was also about 10 oC, which was consistent with the DSC results.

3 Frequency: 1 Hz

2.5

2 PS

1.5 PS + 4 % TPP

E' (kMPa) E'

0.5

0 40 60 80 100 120 140 160 T (oC)

Figure 3.3 Storage modulus (G’) vs. temperature curves tested by DMA for PS and

4 wt. % TPP/PS blends respectively.

Decrease in Yong’s Modulus and Yield Stress

DMA results gave the storage modulus at various temperature. The strain maximum used in this DMA experiment is 0.01%, rather below the yield strain. So the storage modulus here is equal to Young’s modulus, i.e. the slope of the initial elastic region in a stress strain curve.

In the Figure 3.3, Young’s modulus of plasticized PS was lower than that of PS in the temperature range 50 – 140 oC. In the Figure 3.4, though the concentration of TPP is only 2 %, at low temperature near 50 oC, the Young’s modulus of plasticized PMMA is much lower than that of PMMA. Until now, the effects of plasticizers in PS and

PMMA were similar, i.e. decrease in Tg and G’.

2500 DMA

2000 PMMA

1500

1000 PMMA + 2 wt. % TPP

Storage Moludus (MPa) (MPa) Moludus Storage 500

0 50 100 150 Temperature (oC)

Figure 3.4 Storage modulus (G’) vs. temperature curves tested by DMA for

PMMA and 2 wt. % TPP/PMMA blends.

Uniaxial Tensile Extension

30 30 PS PS + 4 wt. % mineral oil V/L = 0.02 /min V/L = 0.02 /min 25 0 25 0 70 oC

20 20 o o 70 C 80 C

15 15

(MPa) (MPa)

engr engr o

  80 C 10 10

5 (a) 5 (b)

0 0 1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 L/L L/L 0 0 Figure 3.5 Stress strain curves of (a) PS and (b) PS plasticized by 4 wt. % mineral

o o oil in uniaxial tensile extension at 70 C and 80 C respectively, with a strain rate V/L0

= 0.02 min-1.

In the uniaxial tensile extension, not only TPP/PS blends and TPP/PMMA blends, but 4 wt. % TPP/SAN blends were tested. In the Figure 3.5 (a), stress strain curves

o o showed ductile behavior of PS at 70 C and 80 C, which also indicated that TBD of PS was below 70 oC. While in the Figure 3.5 (b), for the blends containing 4 wt. % mineral oil, the stress strain curve showed brittle failure at 70 oC and ductile failure at 80 oC, 17

o o which meant the TBD of the slightly plasticized PS to be between 70 C and 80 C. And in Figure 3.6, the 4 wt. % TPP/PS blends were brittle at 70 oC and 75 oC and ductile at

80 oC, which indicated the brittle-ductile transition located in the temperature range from 75 oC to 80 oC.

For SAN and TPP/SAN blends, phenomenon were similar. Figure 3.7 (a) showed the brittle failure at 30 oC and ductile failure at 50 oC for SAN. For 4 wt. % TPP/SAN blends, specimen broke before necking growth at 45 oC and 50 oC. And at 55 oC, specimen just reached the beginning of necking.

30 PS + 4 wt. % TPP 25 V/L = 0.02 /min 0 70 oC 20 75 oC

(kMPa) o engr 80 C

 10

0 1 1.1 1.2 1.3 1.4 L/L 0 Figure 3.6 Stress strain curves of 4 wt. % TPP TPP/PS blends in uniaxial tensile

o o o -1 extension at 70 C, 75 C, and 80 C respectively, , with a strain rate V/L0 = 0.02 min .

All these results showed that by adding 4 wt. % plasticizer, polymer material became more easy to undergo brittle failure. In PMMA was a different story. As shown in Figure 3.8, 2 wt. % TPP/PMMA blends were ductile at 55 oC and higher temperature.

While PMMA was brittle at 55 oC, and its strain is competitive to the strain of 2 wt. %

TPP/PMMA blends at 50 oC. Here were the plasticization effects.

70 70 SAN SAN + 4 wt. % TPP o 60 30 C V/L = 0.02 /min 60 o V/L = 0.02 /min 0 45 C 0

50 50 50 oC o o 40 50 C 40 55 C

(MPa)

(MPa) (MPa)

engr 30 30 engr



20 20

10 10

0 0 1 1.05 1.1 1.15 1.2 1.25 1 1.05 1.1 1.15 1.2 1.25 L/L L/L 0 0 Figure 3.7 Stress strain curves in uniaxial tensile extension of (a) SAN at 30 oC and 50 oC, and (b) SAN plasticized by 4 wt. % TPP at 45 oC, 50 oC, and 55 oC, with a

-1 strain rate V/L0 = 0.02 min .

50 V/L = 0.02 /min PMMA + 2% TPP, 50 oC 0 40 PMMA, 55 oC PMMA + 2% TPP, 55 oC 30

(MPa) PMMA + 2% TPP, 60 oC

engr

 20

0 1 1.1 1.2 1.3 1.4 1.5 L/L 0

Figure 3.8 Stress strain curves in uniaxial tensile extension of PMMA at 55 oC and

o o o -1 2 wt. % TPP/PMMA at 50 C, 55 C, and 60 C, with a strain rate V/L0 = 0.02 min .

SAOS Results and Shift of α Relaxation Time

Small amplitude oscillatory shear technic was used to obtain α relaxation time of

PS and 4 wt. % TPP/PS blends. As shown in Figure 1.5 (a) and (b), the reciprocal value of the horizontal ordinate of the cross point at high frequency, i.e. 1/ω second, was the

α relaxation time (τα) at the reference temperature. So that various values of α relaxation time at various temperatures can be measured drawing a curve of τα vs. temperature. 19

0.1 PS

0.001

(s)





10-5

10-7 PS + 4% TPP

10-9 60 80 100 120 140 160 180 200 T (oC)

Figure 3.9 Comparison of α relaxation time at various temperature between PS and 4 wt. % TPP/PS blends.

As shown in Figure 3.9, τα decreased with higher temperature. Because the τα is the segmental relaxation time, the smaller the magnitude of τα is, the higher the mobility of corresponding segments. The slope of two curves decreased as well with increasing temperature.

By comparison, the curve of the blends was lower than that of PS. In the temperature range from 105 oC to 120 oC, the α relaxation time decreased by nearly two decades in the presence of 4 wt. % TPP, which indicated higher mobility of segments in slightly plasticized PS.

Relaxation in Tensile Extension

Because the stress level of plasticized blends was almost half of that of the pure one, post-yield relaxation was done at the same strain with different original stress.

Figure 3.10 (a) showed the stress-time curve of relaxation of control sample and 4%

TPP blends, and Figure 3.10 (b) showed the normalized curve. Values on the abscissa 20

axis in Figure 3.10 (b) were equal to the true experiment time minus the time at which point relaxation began in logarithm, i.e. log (t − 푡0). And values on the vertical axis were equal to the engineering stress a divide the engineering stress when the relaxation began, i.e. (휎푒푛푔⁄휎푒푛푔,0), so they are in a range from zero to one.

In Figure 3.10 (b), though the plasticized blends had lower stress, the initial slope of its curve was sharper and the value of (휎푒푛푔⁄휎푒푛푔,0) decreased more than those of the curve belonging to pure polystyrene, which indicated that the molecules in the plasticized blends had higher mobility.

1 25 Relaxation (a) Relaxation (b) PS 0.8 The same strain L/L = 1.1 20 0

15 0.6 PS + 4 wt. % TPP (MPa) PS

engr 0.4

 10

The same strain L/L = 1.1 5 0 0.2

% (stress / stress at strain=1.10) at stress / (stress % PS + 4 wt. % TPP

0 0 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 t (s) relaxation time (s)

Figure 3.10 Relaxation in extension denoting the mobility of polymer chains. Figure (b) showed the normalized curve of figure (a).

Pre-yield relaxation in tensile extension were done at the same stress, 19 MPa, and the stress strain curves were shown in Figure 3.11 The relaxation curve of 4 wt. %

TPP/PS blends at 63oC was below that of PS at 70oC. So it can be expected that the relaxation curve of blends at 70oC would be below that of PS at the same temperature, which anticipated the faster mobility of chains in blends.

Both the pre-yield and post-yield relaxation results prove that the chain mobility in plasticized TPP/PS blends was higher than that in PS.

Relaxation PS + 4 wt. % TPP, 60 oC 15 V/L = 0.02 min-1 0

10 PS, 70 oC

(MPa) (MPa)

engr



o 5 PS + 4 wt. % TPP, 68 C

PS + 4 wt. % TPP, 63 oC

0 0.01 0.1 1 10 100 t (min)

Figure 3.11 Stress strain curve of stress relaxation at the same stress,19 MPa, of

PS at 70 oC, and 4 wt. % TPP/PS blends at 60 oC, 63 oC, and 68 oC.

CHAPTER IV

DISCUSSION

Discussions are based on the results shown above and the molecular model introduced in the papers from the group of Dr Shi-Qing Wang.

What Happened at a Breaking Point for a Brittle Polymer in the Molecular Picture?

The molecular weight of tested PS and SAN was much higher than the entanglement molecular weight. All the long chains intertwined with each other. In the glassy state, their energy state was in nonequiliburium located between liquid state and crystalline state. In tensile extension tests, the affine deformation took place in the initial elastic regime, and it was the intermolecular interaction that mainly contribute to the increasing stress. Near yield point, the stress strain curve deviate from linear regime.

During further deformation, the material could be taken as a hybrid structure, composed of a chain network made of connected load-bearing strands, and a primary structure hold by intermolecular force. A load-bearing strand (LBS) was suggested to be between two junctions along the deformation direction due to uncrossability (see Figure 4.1).

When yielding happens, the segments in the primary structure surrounding LBSs and junctions may be activated by the displacing LBSs. The activation of primary structure means the higher mobility of chain segments and higher energy level.

As shown in the Figure 4.2 (a), for brittle polymer glasses, though the primary structure surround LBSs is activated (gray dots), other parts are still not activated, i.e.,

still in the vitreous state (represented by dark dots). Figure 4.2 (b) was a zoomed-out view of the former in the XY plane for the extension along Z axis, where lac is significantly smaller than the average separation of neighboring LBSs. In contrast, upon yielding, the glass may be depicted by Figure 4.3 (a) - (b).

Figure 4.1 Description of formation of an LBS and the surrounding primary structure. The lac is a length to characterize the spread of activated primary structure regime. 3.

Figure 4.2 The left figure (a) showed the structure of brittle polymer in our molecular model. In the primary structure, gray dots represented the activated part, dark dots represented the vitreous part. The connected red circles were connected segments forming a stretched LBS. Figure (b) was zoomed-out view of the former in the XY plane.2

The Interpretation of Higher Brittleness

In traditional theories, higher mobility was always related to higher ductility. It is universally accepted that upon yielding the α relaxation rate has increased above the applied rate. Thus, addition of more mobile plasticizers should increase the α relaxation and thus make it easier to achieve yielding and subsequent plastic deformation.

Figure 4.3 (a) Majority of primary structure was activated (drawn as gray dots) due to displacement of LBSs upon sufficient extension, so the material was ready to yield. (b) A zoomed-out view of the former in the XY plane where most gray “rings” interpenetrate to show that the activation zones have spread to reach one another.2

Both SAOS and stress relaxation of PS reveal increased chain mobility upon adding 4 wt. % TPP. Indeed, the TPP-containing PS should achieve plasticity more readily than the original PS. Because the concentration of plasticizer was so low that the dilution of the chain network is negligible. However, TPP could also affect the condition for chain pullout. In other words, it is plausible for TPP to lower the critical chain tension required for pullout, making the otherwise ductile PS brittle in a certain temperature range. This is apparently what we observed. We have also observed the opposite behavior. Upon incorporating 2 - 4 % TPP to PMMA, the glass either becomes

more ductile or turns from brittle to ductile. In this case, clearly the effect of TPP is to promote molecular mobility without causing chain pullout that would weaken the chain network as is the case of TPP-containing PS. It is therefore highly intriguing to note that the effects of such additives as TPP on polymer glasses vary, depending on the pair of the glass and the plasticizer. Although no theoretical model is available to explain the different effects of TPP on PS and PMMA respectively, our hybrid picture for polymer glasses under large deformation does encompass these different scenarios.

CHAPTER V

SUMMARY

Upon blending PS with 4 wt. % plasticizer, the PS turns more brittle, although both of its storage modulus and glass transition temperature are lowered. Specifically, at

70 oC, PS is ductile at an extensional rate of 0.02 min-1 whereas the plasticizer- containing PS turns brittle. Mechanical spectroscopic measurements show that the alpha relaxation time is shortened by more than two orders of magnitude due to 4 wt. %

TPP. A similar phenomenon, occurs in the blends of SAN and TPP (4%). On the other hand, the effect of the TPP on PMMA is only to make the PMMA more ductile.

The present findings challenge the conventional wisdom based on the classical

Eyring theory2 of activation for flow. Conversely, the experimental results were anticipated according to a recent molecular picture3.

REFERENCES

1. Li, X.; Wang, S.-Q., Mapping Brittle and Ductile Behaviors of Polymeric Glasses under Large Extension. ACS Macro Letters 2015, 4 (10), 1110-1113.

2. Eyring, H., Viscosity, plasticity, and diffusion as examples of absolute reaction rates. The Journal of chemical physics 1936, 4 (4), 283-291.

3. Wang, S.-Q.; Cheng, S.; Lin, P.; Li, X., A phenomenological molecular model for yielding and brittle-ductile transition of polymer glasses. The Journal of chemical physics 2014, 141 (9), 094905.

4. Immergut, E. H.; Mark, H. F., Principles of plasticization. 1965.

5. Schmeling, N.; Konietzny, R.; Sieffert, D.; Rölling, P.; Staudt, C., Functionalized copolyimide membranes for the separation of gaseous and liquid mixtures. Beilstein journal of organic chemistry 2010, 6 (1), 789-800.

6. Merrill, S. H.; Petrie, S., Block copolymers based on 2, 2‐bis (4‐hydroxyphenyl) propane polycarbonate. II. Effect of block length and composition on physical properties. Journal of Polymer Science Part A: General Papers 1965, 3 (6), 2189-2203.

7. Jackson, W.; Caldwell, J., Antiplasticization. II. Characteristics of antiplasticizers. Journal of Applied Polymer Science 1967, 11 (2), 211-226.

8. Anderson, S.; Grulke, E.; DeLassus, P.; Smith, P.; Kocher, C.; Landes, B., A model for antiplasticization in polystyrene. Macromolecules 1995, 28 (8), 2944-2954.

9. Kierkels, J.; Dona, C. L.; Tervoort, T.; Govaert, L., Kinetics of re‐embrittlement of (anti) plasticized glassy polymers after mechanical rejuvenation. Journal of Polymer Science Part B: Polymer Physics 2008, 46 (2), 134-147.10. Sundgren, N.; Bergman, G.; Shur, Y. J., Antiplasticization and transition to marked nonlinear viscoelasticity in poly (vinyl chloride)(PVC)/poly‐ε‐caprolactone (PCL) blends. Journal of Applied Polymer Science 1978, 22 (5), 1255-1265.

11. Robeson, L.; Faucher, J., Secondary loss transitions in antiplasticized polymers. Journal of Polymer Science Part B: Polymer Letters 1969, 7 (1), 35-40.

12. Jackson Jr, W.; Caldwell, J., Antiplasticizers for bisphenol polycarbonates. 1965.

13. Delcambre, S. P.; Riggleman, R. A.; de Pablo, J. J.; Nealey, P. F., Mechanical properties of antiplasticized polymer nanostructures. Soft Matter 2010, 6 (11), 2475- 2483.

14. Lee, J. S.; Leisen, J.; Choudhury, R. P.; Kriegel, R. M.; Beckham, H. W.; Koros, W. J., Antiplasticization-based enhancement of poly (ethylene terephthalate) barrier properties. Polymer 2012, 53 (1), 213-222.

15. Deshpande, A. P., Techniques in oscillatory shear rheology.

16. Grubisic, Z.; Rempp, P.; Benoit, H., A universal calibration for gel permeation chromatography. Journal of Polymer Science Part B: Polymer Letters 1967, 5 (9), 753- 759.

17. Van Kreveld, M.; Van den Hoed, N., Mechanism of gel permeation chromatography: distribution coefficient. Journal of Chromatography A 1973, 83, 111- 124.

18. Moore, J., Gel permeation chromatography. I. A new method for molecular weight distribution of high polymers. Journal of Polymer Science Part A: General Papers 1964, 2 (2), 835-843.

19. Wypych, G., Handbook of plasticizers. ChemTec Publishing: 2004.

20. Razinskaya, I.; Shtarkman, B.; Izvozchikova, V.; Averbakh, N. Y.; Monich, I.; Bubnova, L.; Pupukina, N., Anti-plasticization of polymethylmethacrylate. Polymer Science USSR 1984, 26 (8), 1806-1813.