DISS. ETH NO. 19405
1st December 2010
O LIGOMERE T ECHNOLOGIES FOR
C OST-E FFECTIVE P ROCESSING
H IGH -P ERFORMANCE
P OLYPHTHALAMIDE C OMPOSITES
A dissertation submitted to
ETH Z URICH
for the degree of
Doctor of Sciences ETH Zurich
presented by
Chiara Zaniboni
Laurea in Ingegneria dei Materiali, Università degli studi di Trento
accepted on recommendation of
Prof. Dr. P. Ermanni, examiner
Prof. Dr. Ignaas Verpoest, co-examiner
Zurich 2010
ii
to my husband Andrea and my son Pietro
iv
Contents
Title
i
Abstract
xix
Zusammenfassung
xxi
1
Introduction
1.1 Thermoset and Thermoplastic Carbon Fiber Composites . . . .
1.2 State-of-the-art of Thermoplastic Continuous Fiber Composites .
1.2.1 Impregnation Processes . . . . . . . . . . . . . . . . .
1.2.2 Reactive Processing . . . . . . . . . . . . . . . . . . .
1.3 Research needs . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4 Objectives of this work . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
1
. 2
. 4
. 6
. 11
. 12
. 13
2
Polyphthalamide
21
2.1 Polyphthalamide PA 6T/6I . . . . . . . . . . . . . . . . . . . . . . . 23
2.1.1 PA 6T/6I Prepolymers . . . . . . . . . . . . . . . . . . . . . 24
3
Reactive Processing of Polyphthalamide in Carbon Fiber Composites
3.1 PA 6T/6I Reactive Processing Route . . . . . . . . . . . . . . . . .
3.1.1 Experimental . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.2 Results and discussion . . . . . . . . . . . . . . . . . . . .
3.1.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
29
31
31
32
36
vi
4
5
6
CONTENTS
Material Characterization of the Oligomer Precursor
4.1 Experimental . . . . . . . . . . . . . . . . . . . . . . . .
4.1.1 Thermal Characterization . . . . . . . . . . . . . .
4.1.2 Rheological characterization . . . . . . . . . . . .
4.2 Results and discussion . . . . . . . . . . . . . . . . . . .
4.2.1 Thermal characterization of PA6T/6I oligomer . .
4.2.2 Glass transition of PA6T/6I oligomer . . . . . . .
4.2.3 Cold crystallization of PA6T/6I oligomer . . . . .
4.2.4 Rheological characterization of PA6T/6I oligomer
4.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
43
45
45
46
49
49
49
53
54
59
Electrostatic Dry Powder Impregnation Process
5.1 Dry powder Impregnation . . . . . . . . . . . . . . . . . . . . . .
5.1.1 Tow spreading stage . . . . . . . . . . . . . . . . . . . .
5.1.2 Impregnation Stage . . . . . . . . . . . . . . . . . . . . .
5.1.3 Stabilization Unit . . . . . . . . . . . . . . . . . . . . . .
5.2 Electrostatic Powder Spray Process . . . . . . . . . . . . . . . . .
5.2.1 Transport and Deposition on Carbon Fibers . . . . . . . .
5.3 Set up of the electrostatic powder spray impregnation process . . .
5.4 Optimization of the electrostatic powder spray impregnation plant
5.4.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.2 Fiber Spreading . . . . . . . . . . . . . . . . . . . . . . .
5.4.3 Powder Fluidization . . . . . . . . . . . . . . . . . . . .
5.4.4 Powder Coating . . . . . . . . . . . . . . . . . . . . . . .
5.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
65
66
66
70
72
72
74
78
79
80
81
90
97
106
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Characterization of PA6T/6I Solid State Polymerization
115
6.1 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
6.1.1 Tooling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
6.1.2 Rheological characterization . . . . . . . . . . . . . . . . . . 117
6.1.3 Thermal Characterization . . . . . . . . . . . . . . . . . . . . 117
6.1.4 Mechanical Characterization . . . . . . . . . . . . . . . . . . 118
6.2 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . 118
6.2.1 Thermal characterization of pre-extrusion PA6T/6I prepolymer 118
CONTENTS
vii
6.2.2
6.3
Thermal characterization of PA6T/6I post-extrusion polymerized slabs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
6.2.3 Thermomechanical characterization of PA6T/6I post-extrusion polymerized slabs . . . . . . . . . . . . . . . . . . . . . 129
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7
Consolidation and Polymerization of PA6T/6I Oligomer Pre-forms
135
7.1 Processing of aligned thermoplastic powder impregnated composites . 136
7.2 Processing of aligned PA6T/6I prepolymer powder impregnated composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
7.3 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
7.3.1 Powder impregnation and characterization of carbon fiber tows 138
7.3.2 Laminate preparation and characterization . . . . . . . . . . . 139
7.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 142
7.4.1 Microstructure . . . . . . . . . . . . . . . . . . . . . . . . . 142
7.4.2 PA6T/6I laminate flexural properties . . . . . . . . . . . . . . 146
7.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
8
Concluding remarks and outlook
157
8.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
8.2 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Acknowledgements
161
Publication List
163
Curriculum Vitae
164
viii
CONTENTS
List of Figures
1.1
1.2
1.3
1.4
New Toyota’s hybrid car with carbon body structure. . . . . . . . . . 2
Illustration of the impregnation of a fiber tow at macro–and microlevel. 5
Manufacturing routes for composites materials [23]. . . . . . . . . . . 7
Melt viscosity and processing temperature for reactive and non reactive processing [16]. . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1
2.2
Polyamide 6-3-T [Degussa]. . . . . . . . . . . . . . . . . . . . . . . 22
Polyphthalamide PA6T/6I. . . . . . . . . . . . . . . . . . . . . . . . 23
3.1
Micrograph of PA 6T/6I precondensates as emerges from the production process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Particle size distribution of PA6T/6I prepolymers after milling. . . .
DSC temperatures scan of PA 6T/6I precondensates from 80◦ C to
350◦ C and from 80◦ C to 350◦ C at 10◦ C/min: a) heating curve b)
cooling curve. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Superposition of a DSC and a TGA scan at 10◦ C/min under nitrogen
of the PA6T/6I precondensates. . . . . . . . . . . . . . . . . . . . .
Schematic of the reactive processing route of PA6T/6I. . . . . . . .
3.2
3.3
3.4
3.5
4.1
4.2
4.3
. 33
. 33
. 34
. 34
. 37
Powder impregnation technologies. . . . . . . . . . . . . . . . . . . . 43
Experimental set up for the study of the cold crystallization kinetics
of PPA oligomers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Schematic of the torsional dynamic resonance rheometer [13]. . . . . 48
x
LIST OF FIGURES
4.4
4.5
4.6
4.7
4.8
4.9
4.10
4.11
4.12
4.13
5.1
5.2
5.3
5.4
5.5
5.6
5.7
Schematic of the apparatus for the production of pills. . . . . . . . . .
DSC scan of PA 6T/6I oligomer between 80◦ C and 350◦ C at 10◦ C/min:
a) heating curve b) cooling curve. . . . . . . . . . . . . . . . . . . .
DSC scan from 80◦ C to 140◦ C at 10◦ C/min showing glass transition temperature and enthalpy relaxation of PA6T/6I oligomer dried
at 80◦ C and 4 mbar for 48 hours. . . . . . . . . . . . . . . . . . . . .
DSC scan from 80◦ C to 140◦ C at 10◦ C/min of PA6T/6I oligomer
dried at 80◦ C and 4 mbar for several months: a) first heating b) second
heating of the same sample. . . . . . . . . . . . . . . . . . . . . . . .
DSC scan at 10◦ C/min of PA 6T/6I oligomer: a) first heating from
80◦ C to 180◦ C; b) second heating from 80◦ C to 350◦ C. . . . . . . .
Crystallization half time of PA 6T/6I oligomer under isothermic conditions. The solid line is a polynomial fit curve. . . . . . . . . . . . .
Complex viscosity versus time of PA6T/6I oligomer powder measured on pills prepared at 140◦ C with 160 MPa pressure. . . . . . . .
Complex viscosity versus time of PA6T/6I oligomer powder measured on pills prepared at 140◦ C with 1 MPa pressure. . . . . . . . .
Viscosity values versus time calculated with torsional resonance rheometer of PA6T/6I prepolymer powder sample pressed at 140◦ C with different pressures for the same time. . . . . . . . . . . . . . . . . . . .
Viscosity versus time determined with a torsional resonance rheometer for PA6T/6I oligomer powder pressed at 140◦ C with 4.7 MPa pressure for different times (c) and d)). . . . . . . . . . . . . . . . . . . .
12K Carbon fiber tow. . . . . . . . . . . . . . . . . . . . . . . . . .
Different methods for spreading fibers. . . . . . . . . . . . . . . . .
Mechanical tow spreading with pins: a) typical set up with pins b)
force balance and c) cross section view [63]. . . . . . . . . . . . . .
Enhance mechanical spreading with curved surface. . . . . . . . . .
Schematic of the pneumatic spreading unit [63]. . . . . . . . . . . .
Schematic drawing of a) electrostatic fluidized bed technology, b)
acoustically fluidized bed technology and c) recirculating bad technology [16]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Schematic of the powder seeding processes with a) extruder powder
feed b) helix driven powder feed. . . . . . . . . . . . . . . . . . . .
48
50
51
52
52
54
56
56
57
58
. 66
. 67
. 68
. 69
. 69
. 71
. 72
LIST OF FIGURES
5.8
5.9
5.10
5.11
5.12
5.13
5.14
5.15
5.16
5.17
5.18
5.19
5.20
5.21
5.22
5.23
5.24
Schematic representation of the electrostatic powder coating system
and process [9]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Schematic representation of the electrical field and electrical field
lines [2]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Schematic of electrostatic spray impregnation plan used in this work. .
Photograph of the electrically-grounded substrate. . . . . . . . . . . .
Dependency of the powder coating from the fiber spreading and tow
velocity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Geodesic fiber spreader. . . . . . . . . . . . . . . . . . . . . . . . . .
Edge fiber spreader. . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vacuum Spreader. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Schematic representation of the compressed air spreader and detailed
drawing of the cylindrical cage. . . . . . . . . . . . . . . . . . . . . .
Amount of sprayed powder in function of time and for different fluidization pressure. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Amount of sprayed powder in function of time with the same fluidization pressure and for different vibration energy. . . . . . . . . . . . .
Amount of sprayed powder in function of time with the same fluidization pressure and for highest investigated fluidization pressure, for the
highest investigated horizontal vibration energy and for tow different
vertical amplitudes. . . . . . . . . . . . . . . . . . . . . . . . . . . .
Powder flow index for highest investigated fluidization pressure, for
the highest investigated horizontal vibration energy and for two different vertical amplitudes. . . . . . . . . . . . . . . . . . . . . . . .
Amount of sprayed powder in function of time different powder sizes.
Powder flow index in fuction of different powder sizes. . . . . . . . .
Schematic representation of the coating chamber and of the several
operating variables affecting the electrostatic deposition of oligomers
powder. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Total amount of powder electrostatically deposited onto the grounded
substrate after a spraying time of 10 s as a function of the applied
corona voltage; the amount of deposited powder respectively on the
upper and lower surface is also represented. . . . . . . . . . . . . . .
xi
73
77
79
81
82
84
85
86
89
91
92
94
95
96
96
97
99
xii
LIST OF FIGURES
5.25 Total amount of powder electrostatically deposited onto the grounded
substrate after a spraying time of 10 s as a function of the applied
corona current; the amount of deposited powder respectively on the
upper and lower surface is also represented. . . . . . . . . . . . . .
5.26 Distribution of the powder along the conductive substrate as a function of the air flow velocity. . . . . . . . . . . . . . . . . . . . . . .
5.27 Total amount of powder electrostatically deposited onto the grounded
substrate after a spraying time of 10 s as a function of the power of
the recovery system and of the impregnation chamber geometry. . .
5.28 Total amount of powder electrostatically deposited onto the grounded
substrate as a function of the distance of the gun nuzzle from the
substrate; the deposition on the upper and on the lower surface of the
substrate is also plotted in the graph. . . . . . . . . . . . . . . . . .
5.29 Total amount of powder electrostatically deposited onto the grounded
substrate as a function of the powder size; the deposition on the upper
and on the lower surface of the substrate is also plotted in the graph.
6.1
6.2
6.3
6.4
6.5
6.6
6.7
Match plate mould for the production of PA6T/6I polymer slabs. . .
Solid state polymerization of PA 6T/6I after cold crystallization. . .
DSC thermograms of a prepolymer powder sample heated to 290◦ C
(curve a)), polymerized for 10 min and cooled to 80◦ C (curve b)).
Second heating to 350◦ C (curve c)) and subsequent second cooling to
80◦ C (curve d)) are also shown. Heating/cooling rate 10◦ C/min. . .
Weight loss of PA6T/6I prepolymer samples during polymerization
reaction at different temperatures and for different times, as determined by TGA. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Prepolymer SSP reaction at 290◦ C for ten minutes with constant powder particle size (100 µm) but different powder sample weights. . .
Prepolymer SSP reaction at 290◦ C for ten minutes with constant powder sample weight but with different powder particle size. . . . . . .
Effect of the reaction temperature on the inherent viscosity of PA
6T/6I after SSP for different reaction times. Solution concentration
of 0.3 g/dL in 98% sulphuric acid at 25◦ C. . . . . . . . . . . . . . .
. 100
. 101
. 102
. 104
. 105
. 117
. 119
. 120
. 121
. 122
. 123
. 124
LIST OF FIGURES
xiii
6.8
DSC heating curves from 80 to 350◦ C for polymer samples previously solid state polymerized at different temperatures: a) 270◦ C for
10 min; b) 280◦ C for 10 min; c) 290◦ C for 10 min; d) 300◦ C for 10
min. Heating rate 10◦ C/min. . . . . . . . . . . . . . . . . . . . . . . 126
6.9
Superposition of a DSC and a TGA scan obtained from two polymer
samples cut form the same prepolymer slab and polymerized in the
DSC at 290◦ for 10 min under nitrogen. The initial weight of the two
samples was identical. Heating rate 10◦ C/min. . . . . . . . . . . . . . 127
6.10 DSC heating curves from 80 to 350◦ C for polymer samples previously melt polymerized at different temperatures: a) 305◦ C for 10 min;
b) 310◦ C for 10 min; c) 320◦ C for 10 min; d) 340◦ C for 10 min.
Heating rate 10◦ C/min. . . . . . . . . . . . . . . . . . . . . . . . . . 128
6.11 Storage modulus from -10◦ C to 250◦ C for polymer samples previously polymerized at different temperatures in a hot press: a) 290◦ C
for 15 min; b) 300◦ C for 15 min; c) 320◦ C for 15 min; d) 340◦ C for
15 min. All samples had been heated to the polymerization temperatures at 6◦ C and cooled to 20◦ C at 33◦ C/min. DMA heating rate
10◦ C/min . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
6.12 Storage modulus from -10◦ C to 250◦ C for polymer samples with different thicknesses previously polymerized at 290◦ C in a hot press: a)
0.5 mm; b) 0.9 mm; c) 2 mm. All samples had been heated to the
polymerization temperature at 6◦ C/min, polymerized for 15 minutes
and cooled to 20◦ C at 33◦ C/min. . . . . . . . . . . . . . . . . . . . . 131
7.1
Processing cycle of aligned PA6T/6I prepolymer powder impregnated
composites. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
7.2
Photograph of a UD laminate. . . . . . . . . . . . . . . . . . . . . . 140
7.3
Photograph and micrograph of an electrostatically spray impregnated
carbon fiber tow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
7.4
Picture of the fiber spreader and micrographs of an electrostatically
spray impregnated carbon fiber. . . . . . . . . . . . . . . . . . . . . . 145
7.5
Micrograph of a) a servo-based tensioner and b) a brake-based tensioner.146
xiv
LIST OF FIGURES
Micrographs of consolidated prepolymer laminates prepared at 140◦ C
with holding time of 20 min for two different pressures: a) 0.5 MPa
and b) 8 MPa, respectively. The laminates were prepared via fabric
stacking with prepolymer matrix powder. . . . . . . . . . . . . . . . 147
7.7 Micrographs of PA6T/6I polymer UD laminates obtained from towpregs impregnated with a) d<50µm, b) 50<d<100µm and c) 100<d<250µm
PA6I/6I prepolymer powder matrix and polymerized at 300◦ C for 15
min in a hot press. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
7.8 Flexural modulus of PA6T/6I polymer laminates obtained from towpregs impregnated with d<50 µm, 50<d<100 µm and 100<d<250 µm
PA6I/6I prepolymer powder matrix and polymerized at 300◦ C for
15 min in a hot press. . . . . . . . . . . . . . . . . . . . . . . . . . . 149
7.9 Flexural strength of PA6T/6I polymer laminates obtained from towpregs impregnated with d<50 µm, 50<d<100 µm and 100<d<250 µm
PA6I/6I prepolymer powder matrix and polymerized at 300◦ C for
15 min in a hot press. . . . . . . . . . . . . . . . . . . . . . . . . . . 149
7.10 Flexural modulus of PA6T/6I polymer laminates with 1.5 mm, 1.8
mm and 2.6 mm thickness, polymerized at 300◦ C for 15 min in a
hot press and obtained from towpregs impregnated with d<50µm prepolymer powder particles. . . . . . . . . . . . . . . . . . . . . . . . . 151
7.11 Flexural Strength of PA6T/6I polymer laminates with 1.5 mm, 1.8
mm and 2.6 mm thickness, polymerized at 300◦ C for 15 min in a
hot press and obtained from towpregs impregnated with d<50µm prepolymer powder particles. . . . . . . . . . . . . . . . . . . . . . . . . 151
7.6
List of Tables
1.1
1.2
Advantages of the different impregnation technologies for the production of composites with high performance, high viscosity thermoplastic matrix. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Advantages of the processing route developed in this work. . . . . . . 14
2.1
Overview of the properties of PPA and its competitors [3]. . . . . . . 25
4.1
Characteristic data of non-isothermal cold crystallization exotherms
for PA6T/6I oligomer as measured by DSC at different heating rates. . 53
Characteristic data of non-isothermal cold crystallization exotherms
for PA6T/6I oligomer as loose powder and pressed at different pressure as determined by DSC at 10◦ C/min. . . . . . . . . . . . . . . . . 55
4.2
5.1
5.2
5.3
5.4
5.5
5.6
5.7
Advantages of the different powder impregnation technologies. . . .
Advantages of different dry powder processing technologies. . . . .
Air velocities in powder spray gun. . . . . . . . . . . . . . . . . . .
PA6T/6I prepolymers particle size distribution. . . . . . . . . . . .
Carbon fibers properties. . . . . . . . . . . . . . . . . . . . . . . .
Standard operating conditions during spraying optimization. . . . .
Mean values and standard deviations of the amount of sprayed powder after 5 seconds and for different fluidization pressure. Each mean
value was calculated over 12 consecutive measurements. . . . . . .
.
.
.
.
.
.
65
71
76
80
81
90
. 91
xvi
LIST OF TABLES
5.8
5.9
5.10
5.11
5.12
5.13
5.14
6.1
6.2
6.3
6.4
6.5
Mean values and standard deviations of the amount of sprayed powder after 5 seconds and for different vibration energy. Each mean
value was calculated over 10 consecutive measurements. . . . . . .
Mean values and standard deviations of the amount of sprayed powder after 5 seconds for highest investigated fluidization pressure, for
the highest investigated horizontal vibration energy and for tow different vertical amplitude. Each mean value was calculated over 10
consecutive measurements. . . . . . . . . . . . . . . . . . . . . .
Mean values and standard deviations of the amount of sprayed powder after 5 seconds for different powder sizes. . . . . . . . . . . . .
Standard operating conditions during powder coating optimization. .
Recovery of oversprayed powder by the cyclone separator. . . . . .
Standard operating parameters during powder coating. . . . . . . .
Amount of powder electrostatically deposited onto the grounded substrate as a function of the powder size. . . . . . . . . . . . . . . . .
. 92
. 93
.
.
.
.
95
98
103
103
. 105
Characteristic data of non-isothermal exothermal crystallization peaks
determined with DSC by cooling the sample at 10◦ C/min from different polymerization temperatures. All the PA6T/6I prepolymer powder samples were polymerized for 10 minutes under nitrogen. . . . . .
Characteristic DSC data of a second temperature scan between 80 and
350◦ C for PA6T/6I polymers at 10◦ C/min heating/cooling rate. The
samples were in a first temperature run SS polymerized at 290◦ for
10 minutes. The powder sample weight was kept constant but the
particle powder sizes were varied. . . . . . . . . . . . . . . . . . . .
Inherent viscosity of samples polymerized at different temperatures
for different times. Solution concentration of 0.3 g/dL in 98% sulphuric acid at 25◦ C. . . . . . . . . . . . . . . . . . . . . . . . . . . .
Characteristic DSC data for polymer samples previously solid state
polymerized for 10 min at different temperatures. The sample weight
was kept identical. Heating rate 10◦ C/min. . . . . . . . . . . . . . . .
Characteristic DSC data for polymer samples previously melt polymerized for 10 min. at different temperatures. The sample weight
was kept identical. Heating rate 10◦ C/min. . . . . . . . . . . . . . . .
121
123
125
126
128
LIST OF TABLES
6.6
6.7
7.1
7.2
7.3
xvii
Glass transition temperature as function of polymerization temperature. The samples had been polymerized in a hot press at different
polymerization temperatures for 15 minutes from prepolymer slabs
of identical thicknesses. . . . . . . . . . . . . . . . . . . . . . . . . . 130
Glass transition temperature as function of sample thickness. The
samples had been polymerized in a hot press at 290◦ C for 15 minutes
from prepolymer slabs of different thicknesses. . . . . . . . . . . . . 131
Typical coating line operational condition. . . . . . . . . . . . . . . . 139
Consolidation and reactive processing parameters for the manufacturing of PA6T/6I laminates. . . . . . . . . . . . . . . . . . . . . . . . . 140
Flexural properties of PA6T/6I polymer laminates obtained from towpregs impregnated with d<50 µm, 50<d<100 µm and 100<d<250 µm
PA6I/6I prepolymer powder matrix and polymerized at 300◦ C for
15 min in a hot press (Vf = 39%). . . . . . . . . . . . . . . . . . . . 150
xviii
LIST OF TABLES
Abstract
The main drawback to the development and applications of continuous carbon fiber
thermoplastic composites have been caused essentially from the high melt viscosities
of the thermoplastic material. Achieving good wetting of the thin continuous reinforcing carbon fibers by highly viscous thermoplastic melts is not a simple matter
and furthermore high processing temperatures and pressures are required limiting the
product maximal surface. Due to the very low melt viscosity of in-situ polymerized
systems, new low pressure forming techniques can be developed and applied to costeffective produce large continuous fiber thermoplastic composites.
The main goal of this work is to develop a new cost-effective press forming route
for an in situ polymerizing thermoplastic matrix system with outstanding mechanical,
thermal and chemical properties not yet applied in the market of high-performance
composites. The oligomers selected for these goals are an intermediate step of the
production of polyphthalamides PA6T/6I, a thermoplastic material with outstanding
properties and low raw material costs. As intermediate of a high volume production,
the costs of the oligomers are very low and therefore are suitable for mass production.
The reactive press forming process of polyphthalamides PA6T/6I oligomers under
development in this work is based on prepreg technologies, in situ solid state polymerization and press forming and was divided in two separate process steps. In the first
process step the carbon fiber rovings were mingled with the PA 6T/6I precondensate
powder in order to produce prepolymer powder coated pre-form. Taking advantage of
many years of technological developments in powder coating, an electrostatic powder spray coating impregnation plan was developed and optimized in our laboratory.
xx
Abstract
With this electrostatic powder spray coating high mingling of the fiber with the powder matrix could be achieved and a wide range of powder size could be used with no
need of expensive cryogenic milling.
In the second process step, the prepolymer powder coated prepregs were heated
to the softening temperature of the oligomers in a hot press and pressed in order to
achieve full impregnation of the carbon fiber by low pressure while the prepreg is
shaped to the final form. Between softening and polymerization temperature, the
oligomers were found to undergo a non-reversible cold crystallization, which is promoted not only by increasing temperature but also by increasing applied pressure
forces. Due to the compression forces during the molding step in the hot press, further impregnation of the carbon fibers through the low melt viscosity oligomers could
not be achieved.
Finally, the shaped PA6T/6I prepolymer laminate were polymerized in the hot
press at a temperature between the glass transition and the onset of melting in order
to induce the oligomers to crystallize and subsequent to polymerize in solid state.
Crystallization, consolidation and shaping of PA6T/6I matrix was found to occur simultaneously in few seconds under isothermal condition before SSP with a consistent
reduction of proceeding time and pressure. High flexural modulus and high Tg matrix
could be produce without catalyst within ten minutes at a polymerization temperature
below the melting temperature of the resulting polymer. Additionally, the low reaction
temperature of PA6T/6I SSP restrains side reactions and thermal degradation leading
to higher degree of chain reactions and recyclability of the final polymer. However,
an increase in thickness of the final laminate polymerized by post-extrusion SSP reaction was found to adversely affect the mechanical properties due to the increase of the
diffusion path around the fibers in the reacting mass of the polymerization by-product.
Zusammenfassung
Der Hauptnachteil für die Entwicklung und Anwendung von thermoplastischen Verbundwerkstoffen aus kontinuierlichen Kohlefasern wird im Wesentlichen durch die
hohe Viskosität der Schmelze des thermoplastischen Materials hervorgerufen. Das
Erreichen einer guten Benetzung der dünnen kontinuierlichen kohlenstoffverstärkten
Fasern mit einer hochviskosen thermoplastischen Schmelze ist keine einfache Angelegenheit und zudem begrenzen die hohen Prozesstemperaturen und der benötigte
Druck die maximale Oberfläche des Produkts. Aufgrund der sehr geringen Schmelzviskosität der in-situ polymerisierten Systeme, können neue formgebende Techniken
mit niedrigem Druck entwickelt werden und angewendet werden, um kostengünstig
grosse Bauteile aus thermoplastischen Endlosfasern zu produzieren.
Das Hauptziel dieser Arbeit ist die Entwicklung eines kostengünstigen Umformprozesses für ein in-situ polymerisiertes thermoplastisches Matrixsystem mit herausragenden mechanischen, thermischen und chemischen Eigenschaften, das bisher noch
nicht auf dem Markt der Hochleistungsverbundwerkstoffen vorhanden ist. Die für
diese Ziele ausgewählte Oligomere sind ein Zwischenprodukt bei der Herstellung
von Polyphthalamiden PA6T/6I, ein thermoplastischer Kunststoff mit hervorragenden
Eigenschaften und niedrigen Rohstoffkosten. Als Zwischenprodukt der Grossserienproduktion sind die Kosten für die Oligomere sehr gering und daher gut geeignet für
die Massenproduktion.
Der reaktive Umformprozess der polyphthalamiden PA6T/6I Oligomere, der in
dieser Arbeit entwickelt wird, basiert auf Prepreg Technologien, in-situ Polymerisation im festen Zustand und Umformung und ist unterteil in zwei separate Prozess-
xxii
Zusammenfassung
schritte. Im ersten Prozessschritt werden die Kohlefaserrovinge mit dem vorkondensierten Pulver PA6T/6I vermischt, um vorpolymerisierte pulverbeschichtete Vorformen herzustellen. Unter Ausnutzung der langjährigen technologischen Entwicklungen im Bereich Pulverbeschichtung wurde eine elektrostatische Pulverbeschichtungsanlage zur Imprägnierung in unserem Labor entwickelt und optimiert. Mit dieser
elektrostatischen Pulverbeschichtungsanlage wurde eine hohe Durchmischung der
Fasern mit der Pulvermatrix erreicht. Verschiedene Pulvergrössen konnten verwendet
werden ohne auf eine teure kryogene Zerkleinerung zurückgreifen zu müssen.
Im zweiten Prozessschritt werden die vorpolymerisierten pulverbeschichteten Prepregs bis zur Erweichungstemperatur der Oligomere in einer heissen Presse erhitzt
und mit geringem Druck gepresst, um eine vollständige Imprägnierung der Kohlefasern zu erreichen, während das Prepreg in die endgültige Form gebracht wird. Zwischen der Erweichungs- und Polymerisationstemperatur durchlaufen die Oligomere
eine irreversible kalte Kristallisation, welche nicht nur durch eine steigende Temperatur, sondern auch durch steigende Druckkräfte gefördert wird. Durch die Druckkräfte
während des Formens in der heissen Presse kann keine weitere Imprägnierung der
Kohlefasern durch die niedrigschmelzviskosen Oligomere erreicht werden.
Schliesslich wurden die geformten PA6T/6I polymer Laminate in der heissen
Presse bei einer Temperatur, die zwischen der Glasübergangstemperatur und dem
Beginn der Schmelztemperatur von dem endgültigen Polymer liegt, polymerisiert,
um die Oligomere zu kristallisieren und dadurch in einem festen Zustand zu polymerisieren. Kristallisation, Konsolidierung und Formung der PA6T/6I Matrix geschehen im isothermen Zustand innerhalb weniger Sekunden zeitgleich vor der SSP
mit einer Reduktion der Taktzeit. Eine Matrix mit hohem Biegemodul und hoher
Glasübergangstemperatur können ohne Katalysator innerhalb von zehn Minuten bei
einer Polymerisationstemperatur unterhalb der Schmelztemperatur des endgültigen
Polymers produziert werden. Darüber hinaus schränkt die niedrige Reaktionstemperatur von PA6T/6I SSP die Nebenreaktionen und die thermische Degradation ein
und führt zu einem höheren Grad von Kettenreaktionen und Recyclingfähigkeit des
endgültigen Polymers. Jedoch führt eine zunehmende Dicke des fertigen Laminates,
das durch eine SSP Reaktion polymerisiert wurde, zu einer Beeinträchtigung der mechanischen Eigenschaften durch die Erhöhung der Diffusionsstrecke um die Fasern
in der reagierenden Masse der Polymerisationsnebenprodukte.
Chapter 1
Introduction
Reduction of CO2 and N Ox emissions is today a driving force for further weight reduction in automobiles in combination with improved internal combustion technology
and advanced powertrain technologies such as hybrid gasoline/electric and fuel cells
[23]. Given that the major improvements in powertrain technologies will be realized
in the medium-to-long term, weight reduction through new and advanced material
applications will be the key element in the short-term [21]. The potential of carbon
fiber composites in reducing weight and therefore fuel consumption in vehicles had
been already attested in many applications [19], [12]. Toyota, the second biggest automaker in the world, has recently presented his new concept of hybrid car, which
weights 67% less than the Prius because of a carbon body structure (Fig. 1.1). The
new carbon fiber hybrid car uses half as much fuel as the Prius and could travel with
2.55 liters per 100 km. [6].
Continuous carbon fiber composites exhibit the highest specific stiffness of any
widely available engineering material and have an extremely high strength-to-weight
ratio coupled with exceptional corrosion resistance and fatigue. In addition, due to
their high energy absorbing capability, carbon fiber composite structures can also increase safety in vehicles, as already shown in Formula 1 [17]. However, despite all
these advantages carbon fiber composites are not widely used in vehicles. The fundamental limitations are:
2
Chapter 1. Introduction
Figure 1.1: New Toyota’s hybrid car with carbon body structure.
1. Lack of experience and knowledge in how to design advanced composite structures
2. High cost of the raw materials
3. No affordable process for producing structural composites parts in high volume
Anticipating automakers’ possible use of carbon fibers, big efforts are carried out
today by the carbon fibers industry in research and development to reduce the raw
material price and make it suitable for mass production in cars [6]. On the other hand,
affordable processes are needed in order to produce at low cost structural composite
parts. In response to this demand, the main goal of this research work is to develop
a new cost effective manufacturing route for continuous carbon fiber composites in
high volume applications.
1.1
Thermoset and Thermoplastic Carbon Fiber Composites
Matrices for polymeric composites can be either thermosets or thermoplastics but today the continuous carbon fiber composite market is dominated by composites with
1.1 Thermoset and Thermoplastic Carbon Fiber Composites
3
thermoset matrix resins [20]. Thermoset resins are low viscosity short chains that
during the curing reaction form a series of crosslinks so that one large molecular network is formed. The resulting matrix is an intractable solid that cannot be reprocessed
by reheating [3] and therefore welded or thermally recycled [1]. On the other hand,
thermoplastics start as fully reacted, high viscosity materials that do not chemically
react on heating. On heating to a high enough temperature, they either soften or melt,
so they can be reprocessed a number of times and therefore easily recycled [3].
Recyclable materials and therefore recyclable composites are becoming more and
more important in the automotive industry [11], [1]. The new EU regulation gives the
responsibility to the manufacturers of the end-of-life vehicle disposal and permits by
2015 a maximum 5% disposable waste [11]. The new EU environmental directives
regard also healthy and environmentally friendly production method. Since thermoplastic materials are fully reacted, there is no danger to the worker from emission,
fogging and smelling during the fabrication process.
The melt fusible nature of thermoplastics also offers a number of attractive joining
options such as melt fusion, resistance welding, ultrasonic welding, and induction
welding, in addition to conventional adhesive bonding and mechanical fastening.
Another important advantage of thermoplastics is the low moisture absorption combined with high strain to failure, high impact and temperature resistance. Matrixdominated properties in composites are reduced when the glass transition temperature
is exceeded. The moisture absorption reduces the glass transition temperature and
hence limits the application of many thermoset composites to less than 120◦ C. Moreover, thermoset systems intended for high-temperature applications may undergo curing at temperatures up to 350◦ C and a compromise with the other desirable mechanical characteristics has to be made [19].
Finally, the use of composites to reduce weight in automotive applications requires
special attention not only on material performance but also on material and manufacturing costs: customers expect a contribution to environmental improvements without
additional cost [21]. High degree of automated processing and high processing velocity has been recently implemented in SMC [18], but the cure time required for
structural thermoset composites is usually of several minutes and therefore limits the
production speed and increase manufacturing costs. Thermoplastic matrix does not
need to be crosslinked and their shaping process is a purely physical one, making
possible to produced structures in minutes with a consequently potential for high volume production. In addition, there are new added strategic benefits associated with
4
Chapter 1. Introduction
carbon fiber composites such as component integration, modularity, lower assembly
costs, and potential to eliminate conventional painting and its corresponding capital
and environmental cost which could make them interesting for mass production in
cars [21].
1.2
State-of-the-art of Thermoplastic Continuous Fiber
Composites
The main drawbacks to the development and applications of continuous fiber thermoplastic composites have been caused essentially from the high melt viscosities of
thermoplastic matrices [7], [5].
A strong interface between matrix and reinforcement is a prerequisite for high composites mechanical properties, in particular transverse flexural strength and interlaminar fracture toughness [4]. It is important to place the resin and the fibers in intimate
contact to achieve good adhesion between the fiber and the matrix. However, achieving good wetting of the thin continuous reinforcing fibers by highly viscous thermoplastic melts is not a simple matter and furthermore high processing temperatures and
pressures are required [4].
Impregnation of the continuous fibers by the thermoplastic matrix is generally the rate
determining step in the production of structural composites [10]. The impregnation
rate depends on the resin viscosity, as described by Darcy’s low [5]. A flow length,
L, can be used to describe the distance that the resin must flow to achieve a full impregnation of the fiber bed or tow, as shown in fig. 1.2. If only one dimensional flow
is considered, Darcy’s law appears as:
dL
Kp ∂P
=−
(1.1)
dt
η ∂x
where uliq is the superficial velocity of the fluid, vv is the porosity of the porous
medium, L is the penetration distance in the x direction, t is the time, Kp is the permeability of the porous medium, η the viscosity of the fluid, and P is the pressure acting
along x to enable the fluid to advance through the porous medium. By integrating this
equation and assuming a constant permeability during the impregnation process, the
time required to fully impregnate the bed can be estimated as:
uliq = vv
1.2 State-of-the-art of Thermoplastic Continuous Fiber Composites
5
Pa
Macrolevel
Matrix
x
Faser
Po
L
Microlevel
L
x
Pa
Pa
Po
Figure 1.2: Illustration of the impregnation of a fiber tow at macro–and microlevel.
t = vv
ηL2
2Kp (Pa − P0 )
(1.2)
where Pa designates the applied pressure to enhance flow and P0 is the atmospheric pressure. The impregnation time is proportional to the resin viscosity. As the
viscosity of thermoplastic resin is three order of magnitude higher than for thermosets,
it follows that thermoplastic systems cannot be processed using the same techniques
as thermosets. The impregnation process can be accelerated by increasing the applied
pressure, but fiber damage may occur with consequent considerable reduction of the
mechanical properties of the composite. In addition, the permeability of the bed, Kp ,
decreases with increasing applied pressure, increasing the difficulty of impregnation
[10].
Equation 1.2 demonstrates also that the most influential factor on the impregnation
time is the impregnation length in the squared term L, therefore an effective solution
6
Chapter 1. Introduction
to reduce impregnation time with high melts viscosity thermoplastic is to minimize
the flow distance for impregnation. Based on this consideration, different techniques
have been developed in order to cost-effectively produce high quality thermoplastic
continuous fiber composites, as shown in Fig. 1.3. The majority of these techniques
consist in two separate process step [10]:
1. The first process step involves an impregnation process for the production of
thermoplastic continuous fiber laminate sheets or tapes and tows, which can be
woven to fabric preforms. These preforms can also be stacked, impregnated
and consolidated to laminate sheets in a separate process step by applying heat
and pressure with an autoclave or press.
2. The second process step involves a consolidation and shaping process to the
final composite parts using a combination of heat and pressure in autoclave or
press. In case of the thermoplastic continuous fiber preforms can also involve
complete impregnation of the fibers. On the other hand, the pre-impregnated,
pre-consolidated sheets are shaped with short cycle times to composite parts
with a good, reproducible fiber matrix interface but due to the stiff nature of the
laminate with a reduced design freedom [10].
1.2.1
Impregnation Processes
In general the impregnation processing techniques for the production of continuous fiber thermoplastic composites can be divided in two process strategies: preimpregnation processes and post shaping impregnation processes 1 , depending in
which process step the fully impregnation of the fibers takes place. Unlike in the case
of impregnation after shaping processes where fully impregnation of the fibers and
consolidation of the laminates take place during shaping, in pre-impregnation processes the impregnation and shaping steps are separated [10]. The resulting prepreg
consists of reinforcing fibers completely wetted by the matrix with a good, reproducible fiber-matrix interface. In order to achieve full impregnation of the reinforcement fibers and short impregnation time, pre-impregnation processes aim for reducing the viscosity of the thermoplastic matrix. Hereafter, an enumeration of the preimpregnation techniques is presented:
1 other names found in literature for post shaping impregnation are: impregnation after shaping [4],
intimate mixing techniques [7] or intimate mingling processes [10]
Impregnation and Preform Types
Row Material
1.2 State-of-the-art of Thermoplastic Continuous Fiber Composites
Granulate
Reinforcement Fiber
Matrix
Extrusion
(Milling)
Film
Powder
Fiber
Pre Impregnation Process
Post Impregnation Process
Impregnation
Mingling
Reinforcement Fiber
Tape, Tow, Yarn
Discontinuous, Semi-continuous
and Continuous Process
Impregnation
Consolidation
(Weawing and Brading)
Fabric
2D-3D Preform
Pre-preg
Sheet Stock
Manufacturing
Autoclave, Diaphragma, Press Forming
Impregnation
Tape Laying, Stamp Forming
Pultrusion, Roll Forming, Stamp Forming
Consolidation
Shaping
Final Part
Figure 1.3: Manufacturing routes for composites materials [23].
7
8
Chapter 1. Introduction
1. Melt impregnation is based upon forcing molten polymer by high shear rate
and/or pressure into a fiber tow to form a prepreg tape. This technique has
been found to operate well only with low viscosity thermoplastic melts, such
as Nylon 66, at line speeds up to 30 m/min but the process presents limitations
such as low processing velocity and low impregnation quality when high resin
viscosities and high fiber volume fractions are employed [10].
2. Solution impregnation is primarily used for amorphous thermoplastics (as most
crystalline thermoplastics are not readily soluble in solvents) by dissolving the
polymer in a suitable solvent or plasticizer with the advantages to considerably
reduce the resin viscosity [10]. After complete wet-out of both unidirectional
and woven reinforcement with the dilute solution, the solvent has to be removed. Many are the technical limitations of this technique. Many high performance thermoplastic polymers have no suitable solvents. Moreover, solvents
are very difficult to remove from the matrix [22] and the presence of residual
trace of solvent or void formation after the removal of the solvent may compromise processing and reduce service performance of structural composites, particularly at high temperature [5]. Solvents can be toxic, therefore environmental
and health concerns as well as the costs associated with using and reclaiming
high boiling solvents makes their use undesirable [14].
In intimate mixing techniques the reinforcing fibers and the thermoplastic matrix
in solid form are mingled together in order to produce a preform where the reinforcing
fibers are not fully wetted by the matrix [2]. Shifting the full impregnation of the
fibers in the final shaping step gives the advantage of a more flexible and drapeable
preform and therefore more freedom in forming complex parts. On the other hand
the mixing process reduces the distance to flow L of the thermoplastic matrix by
bringing resin and fibers mechanically in contact [10]. Depending on the degree of
intimacy reached in the mingling process, the impregnation time and/or pressure can
be therefore drastically reduced. Hereafter, an enumeration of the intimate mixing
techniques is presented:
1. Film stacking can be used with thermoplastic material in film form and consists in stacking layers of fiber reinforcement, in form of unidirectional tows or
woven fabrics, alternated with layer of thermoplastic thin films before applying
1.2 State-of-the-art of Thermoplastic Continuous Fiber Composites
9
heat and pressure. Due to the long impregnation distance and to the flow transverse to the fiber, which requires more energy compared with the flow parallel
to the fiber [5], high pressure and long molding time are required to wet the
fibers. This impregnation technique is adequate for the low volume production
of composites not commercially available in prepreg form. Complex shape can
hardly be achieved due to the wrinkling and buckling of the polymer films [10].
2. Commingling also known as hybridization can be used with thermoplastic material in fiber form and involves an intimate mixing of polymer and reinforcing
fibers into a single tow. The resultant hybrid tow (or yarn) is normally woven
into a fabric or in a unidirectional preform, which have the advantage of being
very flexible and drapeable, or braided directly in a complicated 3D form [8].
Unfortunately not all thermoplastic matrices can be spun into fibers and additional costs are involved in producing thermoplastic yarn and weaving it with
the reinforcing fibers. In order to achieve a good distribution of the polymer
and reinforcing fibers in the preform, the matrix fiber diameter has to match
that of the reinforcement, additionally increasing the spinning costs [10]. An
alternative to strongly reduce costs are co-woven fabrics which consist of reinforcement fiber tows and polymer fiber tows woven together without being
combined into a hybrid yarn. However, the thermoplastic fibers and the reinforcement fibers are in this case not intimately mingled, increasing the impregnation flow distance L and therefore the overall impregnation time during
shaping [10]. Moreover, both processing strategies lead to increased impregnation time due to the flow transverse to the fiber direction, which is the direction
of lower permeability of the fiber bed [10].
3. Powder impregnation consists in mingling the reinforcement fiber tows with
the thermoplastic matrix in powder form. A considerable advantage of this impregnation technique is the availability through mechanical grinding or cryogenic milling of any kind of polymer in powder form [13]. Moreover, many
high performance polymers emerge from the polymerization reaction in controlled powder form [5], and thus no additional milling process is required with
a resulting cost benefit. In opposition to the other mixing techniques, powder impregnation produce impregnated tows, where the resin flow takes place
parallel, rather than perpendicular to the fiber direction [5]. As the permeability along the fiber direction is at least one order of magnitude higher than in
10
Chapter 1. Introduction
the transverse direction [10], lower impregnation time and pressure are needed.
The degree of intimate mingling of the powder matrix with the reinforcement
fibers increase with the reduction of the powder matrix diameter to match that
of the fibers. However, the reduction of the particle size increases the costs of
the milling process.
In case of thermoplastic structural composites, continuous carbon fibers and high
performance thermoplastics are used. High performance thermoplastic materials are
often crystalline, are not soluble in solvents and show very high melt viscosity. Moreover, the diameter of the fragile carbon fibers (7 µm) is usually smaller than that of
other reinforcement fibers, like glass fibers (17 µm). This difference combined with
the high fiber volume content up to 65% in structural composites leads to higher flow
resistance during melt impregnation and therefore to low processing velocity, high
fiber damage and low impregnation quality [24]. Considering that not all thermoplastic polymers can be spun into fibers, a promising intimate mixing impregnation
technology to process carbon fiber thermoplastic composites is powder impregnation,
as shown in Tab. 1.1. The advantages can be summarized as follows [10]:
1. The availability of all thermoplastic material in powder form through grinding
and the polymerization reaction of many high performance polymers into powders with finally controlled size which reduce row material costs especially if
compared with fiber spinning
2. The no need of solvent
3. The very low emissions during impregnation
4. The reduction in flow length for impregnation which takes place parallel to the
fibers
5. The high flexibility of the towpregs, which can be woven in drapeable fabric or
braided to complex 3D forms
In the last years a new manufacturing route for continuous fiber thermoplastic
composite named reactive processing was developed.
1.2 State-of-the-art of Thermoplastic Continuous Fiber Composites
Advantages
Availability matrix material form
Low costs of matrix material form
Low costs impregnation plan
Low impregnation length
Reduced emissions
Complex composite shape
Solution
-,+
-,-,-
Melt
+
++
+
-
Film
+
+
+
+
Faser
+
++
++
++
11
Powder
+
+
++
+++
++
++
Table 1.1: Advantages of the different impregnation technologies for the production
of composites with high performance, high viscosity thermoplastic matrix.
1.2.2
Reactive Processing
Reactive processing or in-situ polymerization is based on impregnating the fiber tow
or fiber bed with a monomer or oligomer and on subsequent in-situ polymerization to
linear chain thermoplastic polymers. These systems have a low viscosity (Fig. 1.4)
and are able to quickly and effectively impregnate the continuous reinforcing fibers
before being in situ polymerized, i.e. before being converted into high-molecularweight polymer directly on the fiber substrate [16].
Reactive processing of thermoplastic essentially combines the basic processing routes
and advantages associated with thermoset processing with the advantages of the final
material properties of thermoplastic. Due to their low melt viscosity, the in-situ polymerized systems can be processed without the need of high pressures like all classic manufacturing processes of thermoplastics composites and with the possibility of
creating a chemical bond between fibers and matrices through a for this purpose optimized sizing [16], [4]. Moreover, the impregnation of the reinforcement fiber and
shaping to the final parts can take place in only one process step, like in the case of
RTM technologies. Another advantage associated with this process is the possibility
of adding to the unreacted monomer nano-particles in order to obtain a fiber reinforced polymer nano-composite.
Disadvantages associated with this process include also the need of catalyst systems
and the need to remove undesired materials, such as solvent, inhibitor or by-products
of the polymerization reactions [9]. Such polymerization processes are usually accompanied by significant shrinkage of the resin phase, which may leave voids in the
12
Chapter 1. Introduction
100000
10000
Melt viscosity [Pa s]
1000
.
PES
PMMA
PA-12
PA-6
Melt processing of
engineering plastics
PPS
Melting processing of highperformance plastics
Reactive processing of
thermoset resins
PC
1
vinylester epoxy
ETPU
Reactive processing
of thermoplastic
oligomers
polyester
0.1
PMMA
PA-6
0.01
0.001
PEKK
PEI
PBT
100
10
PEEK
PBT
PEK
PA-12
Reactive processing of
thermoplastic monomers
0
50
100 150 200 250 300 350 400 450
Processing temperature [°C]
Figure 1.4: Melt viscosity and processing temperature for reactive and non reactive
processing [16].
structure [4].
1.3
Research needs
Continuous fiber reinforced thermoplastics are traditionally compression molded and
high consolidation pressures are used, typically 100-200 bars. Due to the high temperatures and pressures, heavy steel moulds are required, increasing tooling costs. In
addition, especially for larger products, very heavy presses are needed to provide the
required consolidation pressure, dramatically increasing machine costs and limiting
the product maximal surface. Due to the very low melt viscosity of the in-situ polymerized systems, new low pressure forming techniques can be developed and applied
to cost-effectively produce large continuous fiber thermoplastic composites. Moreover, all the classical manufacturing processes of thermoset composite can be used
for reactive thermoplastic [4].
1.4 Objectives of this work
13
While reactive thermoset composites dominate the high performance composites market, the in situ polymerization process for the production of structural thermoplastic composites is rather an exception than an established process. Possible reasons
for this low number of applications are the high materials and tooling costs combined with the inferior intrinsic properties of the engineering thermoplasts investigated so far. The costs of the oligomers CBT TM , for example, currently being marketed specifically for the production of composites by the Cyclics Corporation are
higher compared to the cost of row PBT. Moreover, the thermal and chemical performance of PBT matrix, even if very good is not outstanding and can not really fill
a cost-performance gap that can justify the increase in tooling and matrix costs for
the reactive processing (Tab. 2.1). Up to now these thermoplastic matrix composites
are still preferably produced by conventional thermoplastic processing routes. On
the other hand, an in situ polymerized thermoplastic matrix system with outstanding
mechanical, thermal and chemical properties not yet applied in the market of highperformance composites would justify the costs and the development efforts of a new
processing route.
Long cycle times and high temperatures during processing represent also drawbacks
of this technology, especially if we consider that rapid fabrication techniques are one
of the main advantages of thermoplastic composites. Therefore, in order to costeffectively produce thermoplastic structural composites by reactive processing routes
the in-situ polymerization reaction should take place in a few minutes.
1.4
Objectives of this work
In response to this driving force, the main goal of this work is to develop a new costeffective press forming route for an in situ polymerizing thermoplastic matrix system
with outstanding mechanical, thermal and chemical properties not yet applied in the
market of high-performance composites. Due to its very low melt viscosity, this insitu polymerized system could be processed without the need of high pressures and
with the possibility of obtaining a strong chemical fiber-to-matrix bond. This would
also open the possibility for new low pressure forming techniques, which could be
applied to cost-effectively produce large continuous fiber thermoplastic composites.
The oligomers selected for these goals are an intermediate step of the production
of polyphthalamides PA6T/6I, a thermoplastic material with outstanding properties
14
Chapter 1. Introduction
Advantages
Reduced processing pressure
Short processing time
Reduced processing temperature
No emissions
Improved material properties
Complex Composite shape
Powder impregnation
+
++
\
+
+
+
In-situ polymerization
+
–
+
+
++
+
Table 1.2: Advantages of the processing route developed in this work.
and low raw material costs never used before in the field of high performance composites (chapter 2). As intermediate of a high volume production, the costs of the
oligomers are very low and therefore are suitable for mass production.
The reactive press forming process of polyphthalamides PA6T/6I oligomers under
development in this work (chapter 3) will be based on prepreg technologies, in situ
solid state polymerization and press forming and will be basically divided in two
process steps [15]. In the first process step the carbon fiber rovings have to be mingled with the PA6T/6I precondensate in powder form. The oligomers emerge from
the polymerization reaction in powder form, and thus no additional processing is required for their use in powder impregnation technologies, resulting in a cost benefit
for powder coated prepregs. In order to develop and optimize the impregnation process, the oligomer precursors have to be characterized and the most pertinent material
parameters and their influence on the desired impregnated prepreg have to be identified (chapter 4). Taking advantage of many years of technological developments in
powder coating, an electrostatic powder spray coating impregnation plan has to be developed and optimized in our laboratory (chapter 5). With electrostatic powder spray
coating a wide range of powder size can be used and therefore there is no need of
expensive cryogenic milling, accomplished by recycling oversprayed powder, which
additionally reduces the costs for the matrix material. In the second separate process
step, the prepolymer powder coated prepregs will be heated to the softening temperature of the oligomers in a hot press and pressed so that the oligomers - thanks to their
low melt viscosity - can easily wet the fibers by low pressures while the material is
shaped to the final form. Finally, the temperature have to be raised to a value between
the glass transition and the onset of melting in order to induce the oligomers to crys-
1.4 Objectives of this work
15
tallize and subsequent to polymerize in solid state (chapter 7). In order to establish
a useful process window for the reactive molding, the reaction kinetic and the most
important variables affecting the solid state polymerization of PA6T/6I prepolymers
have to be investigated (chapter 6).
16
Chapter 1. Introduction
Bibliography
[1] S. Béland, High Performance Thermoplastic Resins and their Composites,
Noyes Publications, 1990, New Jersey.
[2] U. Breuer, M. Ostgathe, Halbzeug und Bauteilherstellung - Umformverfahren,
Faserverbundwerkstoffe mit thermoplastischer Matrix, Expert Verlag, 1996.
[3] F. Campbell Jr, Manufacturing Technology for Aerospace Structural Materials,
Elsevier 2006.
[4] F.N. Cogswell, Thermoplastic Aromatic Polymer Composites, ButterworthHeinemann, 1992.
[5] M. Connor, Consolidation Mechanisms and Interfacial Phenomena in Thermoplastic Powder Impregnated Composites, These N◦ 1423, 1995.
[6] N. Fujimura, M. Kitamura, Automakers Turn to Carbon Fiber and Aluminum in
Bid to Increase Gas Mileage, International Herald Tribune, October 23, 2007.
[7] A.G. Gibson ,J.-A. Manson, Impregnation Technology For Thermoplastic Matrix Composites, Composites Manufacturing, 3, 4 1992, 223.
[8] S.A. Hasselbrack, C. Perderson, J.C. Seferis, Evaluation of Carbon-FiberReinforced Thermoplastic Matrices in a Flat Braid Process, Journal of Polymer
Composites, 13, 1, 1992, 38-46.
18
BIBLIOGRAPHY
[9] A. Luisier, P.-E. Bourban, J.-A.E. Månson, Time-Temperature-Transformation
Diagram for Reactive Processing of Polyamide 12, J. Appl. Polym. Sci., 81,
2001, 963-972.
[10] A. Miller, A.G. Gibson, Impregnation Techniques for Thermoplastic Matrix
Composites, Polymers and Polymer Composites, Vol. 4, No. 7, 1996.
[11] J.-B. Monteil, "The European Automotive Industry Will Use more and more
Composites", JEC Composites Magazine, nr. 34 July-August 2007.
[12] A.P. Mouritz , E. Gellert , P. Burchill , K. Challis, Review of Advanced Composite Structures for Naval Ships and Submarines, Composite Structures 53, 2001,
21-41.
[13] M. Neitzel, P. Mitschang, Handbuch Verbundwekstoffe, Carl Hanser Verlag
München Wien, 2004.
[14] S. Padaki, L.T. Drzal, A Simulation Study on the Effects of Particle Size on the
Consolidation of Polymer Powder Impregnated Tapes, Composites: Part A 30
1999, 325-337.
[15] N. Pini, C. Zaniboni, S. Busato, P. Ermanni, Perspectives for Reactive Molding
of PPA Matrix for High-Performance Composite Materials, FPCM7 2004.
[16] K. van Rijswijk, H.E.N. Bersee, Reactive Processing of Textile Fiber-Reinforced
Thermoplastic Composites - An Overview, Composites: Part A, 38, 2007, 666681.
[17] S. Savage, I. Bomphray, M. Oxley, Exploiting the Fracture Properties of Carbon Fiber Composites to Design Lightweight Energy Absorbing Structures, Engineering failure analysis 11, 2004, 677-694.
[18] M. Sommer, Changing Demands for Composites in the Automotive Industry,
JEC Composites Magazine, nr. 34, July-August 2007.
[19] C. Soutis, Fiber Reinforced Composites in Aircraft Construction, Progress in
Aero-space Sciences 41, 2005, 143-151.
BIBLIOGRAPHY
19
[20] K.-H. Sprenger, Aufbau und Eigenschaften von thermoplastischen Hochleistungs-FVW, Faserverbundwerkstoffe mit thermoplastischer Matrix, Band
529, 1996, 86-94.
[21] J. Sullivan, Use of Plastics in Body Construction in Future Automobile Design
and Production, Kunstoffe im Automobilbau, 2003, pp 3-25.
[22] Patent US 5,364,657, J. L. Throne, Method of Depositing and Fusing Polymer
Particles onto Moistened Continuous Filaments, 1994.
[23] J. Verrey, M.D. Wakeman, V. Michaud, J.-A.E. Månson, Manufacturing Cost
Comparison of Thermoplastic and Thermoset RTM for an Automotive Floor
Pan, Composites, Part A 37, 2006, 9-22.
[24] A.M. Vodermayer, J.C. Kaerger, G. Hinrichsen, Manufacture of High Performance Fiber-Reinforced Thermoplastics by Aqueous Powder Impregnation,
Composites Manufacturing, Vol 4 No 3. 1993.
20
BIBLIOGRAPHY
Chapter 2
Polyphthalamide
Polyphtalamid (PPA is the ISO abbreviation, see ISO 1874) is a material classification
(ASTM) for copolyamides of isophthalic acid and terephthalic acid. These partially
aromatic polyamides are high performance engineering thermoplastics that bridge the
cost-performance gap (Tab. 2.1) between traditional engineering thermoplastics such
as polycarbonate (PC), polyamides (PA), polyesters (PET, PBT), acetals (POM) and
higher-cost specialty polymers such as liquid crystal polymers (LCP), polyphenylene
sulfide (PPS) and polyetherimide (PEI) [4].
In general polyamides containing aromatic residues are practically insolubly crystalline polymers with very high melting temperature and remarkable heat resistance
and therefore are very difficult to process. The high melting points which characterize such materials are the result of the ability of the linear chains to pack in a regular
crystal lattice and of the interchain hydrogen bonding between amide links [3]. The
attachment of methyl, ethyl or higher substituents to the same in-chain carbon along
the linear aliphatic chain or the introduction of monomers which possess some measure of stereochemical asymmetry (asymmetric monomers) are all good approaches
to reduce crystallization and to improve the processability of aromatic polyamides
[1].
In the year 1967, Gabler and al. reported, for example, the effect of using different
alkyl substituted hexamethylenediamines on the melting point and physical state of
corresponding polyterephthalamides [2]. Therefore, in the ’70s Dynamit Nobel de-
22
Chapter 2. Polyphthalamide
H
H
C
C
O
O
N
(CH2)6
N
(CH3)3
n
Figure 2.1: Polyamide 6-3-T [Degussa].
veloped a more easily processed amorphous polyterephthalamide with good mechanical properties with the commercial name TROGAMID TTM . TROGAMID TTM is
a polyamide 6-3-T (DIN 16773), a copolymer of terephthalic acid and trimethylhexamethylenediamine (Fig. 2.1). The inclusion of alkyl substituted hexamethylenediamines is able to prevent regular chain packing and therefore crystalline regions
which lead to formation of an amorphous material of generally lower melting point.
Semi-crystalline polyphthalamides, copolyamides of isophthalic acid and terephthalic acid with aliphatic or substituted diamines, were first developed during the ’80s
by Amoco Chemical, USA. The introduction of asymmetric monomers, like isophthalic acid, to inhibit crystallization of aromatic polyamides is less obvious. The
isophthalic acid / aliphatic diamine copolymer unit introduce a point of inflection
in the polymer chain do to the stereochemical configuration of the carbonyl groups
adjacent to the benzene ring in the isophthalic acid. The benzene ring results asymmetrically positioned with respect to the main chain axis in contrast with the linear
configuration of the terephthalic acid / aliphatic diamine copolymer units. The point
of inflection created in the linear polyamide chain prevents the growth of ordered layers of crystallite and increase the amorphous character of the polyamide. The more
crystalline or amorphous character of PPA is obviously dependent to the molar concentration of asymmetric monomer.
23
%
2.1 Polyphthalamide PA 6T/6I
30
C
O
H
H
70%
C
N
(CH2)6
N
O
n
Figure 2.2: Polyphthalamide PA6T/6I.
2.1
Polyphthalamide PA 6T/6I
Polyphthalamide PA 6T/6I (ISO 1043-1) used in this work is a semi-crystalline copolyamide based on hexamethylenediamine, terephthalic acid and isophthalic acid,
where the major monomer (70%) is the unit with terephthalic acid (6T) and the minor
monomer (30%) is the unit with isophthalic acid (6I) (Fig. 2.2). The proportion of the
constituent monomers determine the bulk physical properties, the processability, the
whether crystalline or amorphous structure and therefore the mechanical properties
of the final PA 6T/6I resin.
Polyphthalamides are low costs matrices which present very high chemical resistance and good mechanical properties (e.g. strength, stiffness, fatigue, creep resistance) over a broad temperature range.
In particular, PA 6T/6I resin features excellent strength, very good rigidity, and
high dimensional stability at working temperature up to 150◦ C. Its thermal performance (heat deflection temperature at 1.8 MPa up to 120◦ C and melting temperature
around 325◦ C) is superior relative to polyamide 12 (PA12), polybutylene terephthalate (PBT), and polyphenylsulfone (PPS) but it is exceeded by polyetheretherketone
(PEEK).
Polyamides are limited by the fact that moisture uptake affects material properties. PA 6T/6I resin also absorbs moisture but at a lower degree (< 30%) and more
slowly than PA 6,6. However, at high value of water absorption stiffness and strength
are significantly reduced.
24
Chapter 2. Polyphthalamide
Due to the low costs, high temperature performance and chemical resistance, PPA
resins can be found in automotive applications (such as fuel components systems), in
hot water applications in the sanitary field as well as in industrial applications (coffee
machines, golf club heads, pumps in the petrochemical industry, gas cylinder hooks,
etc). Furthermore, due to the high dimensional stability of the injection molded parts
made from PPA matrices, PPA are used in the electrical and electronic fields, where
requirements regarding performance under extreme temperatures and reproducibility
are becoming more demanding.
2.1.1
PA 6T/6I Prepolymers
The PPA prepolymers (PA 6T/6I, XE 3733 VK) utilized in this work are the intermediates in a two-stage method developed by EMS-CHEMIE AG for the production of
partially aromatic polyamides. Because of the typically very high structure-dictated
melting viscosities of such polyphthalamides, the polycondensation with a singlestage batch process must be stopped at a very early stage to enable discharging the
melt from the autoclaves and processing it into granular material. Limiting the average molecular weight of these polymers to comparatively low values has a negative
effect on their mechanical properties.
In a first attempt to overcome this problem, the production process of partially aromatic polyamides was developed in a two-stage method: a batchwise process for the
production of precondensates (prepolymers), and a final reaction stage where the precondensates were melt polymerized in a double- or single-screw extruder into highmolecular weight polyphthalamide [5], [6].
However a conventional batch process is not suitable for the production of high performance PA 6T/6I, or in general for high-viscosity partially aromatic polyamides.
The major difficulties arise from the fact that the extruder must be supplied with two
prepolymers in stoichiometric rations and homogeneous mixing of the two prepolymers have to occur during polycondensation which results in technically impossible
precision.
In order to overcome these difficulties, EMS-CHEMIE AG developed a multi-step
batch process for the production of precondensates of partially aromatic polyamides
[7], [8], [9]. This process is also applicable to the production of other polycondensation polymers. The process involved a salt forming phase for the production of aque-
25
2.1 Polyphthalamide PA 6T/6I
Properties
Density (gcm3 )
Thermal exp. coeff.
Melting temp. (◦ C)
Glas trans. temp. (◦ C)
HDT at 1.8 MPa (◦ C)
Young’s modulus (GPa)
Tensile strength (MPa)
Elongation at break (%)
Price reference ($/Kg)
PA12
1.04
9.7·10−5
180
40-50
40-50
1.4
50-60
300
PBT
1.31
10·10−5
220-230
50-60
48
2.5-3.5
64
200
4-5
PPA
1.18
8·10−5
310-330
121-138
120
2.5-3.5
90
6
6-7
PPS
1.35
4.9·10−5
280-290
90
110
2.5
75
15
7-10
PEEK
1.32
3.2·10−5
365-375
150-160
152
4.5
90-100
20
80-90
Table 2.1: Overview of the properties of PPA and its competitors [3].
ous salt solution of diamines and dicarboxylic acids. In a subsequent reaction stage,
the precondensates are formed and during a stationary stage the precondensates reach
a virtually stable state, in which no further substantial change in its average molecular
weight and in its related properties occurs. Finally, the partially crystalline PA 6T/6I
precondensates are discharged in the form of solid particles by being spayed into an
inert cyclone or spray tower. The constancy of the precondensate quality obtained by
EMS-CHEMIE AG assures that the polycondensation in the final reaction also produces polymer of constant quality, without need of catalyst or of mixing prepolymers.
Moreover, the PA 6T/6I prepolymers are readily produced especially if compared to
cyclic prepolymers and it is possible to develop an in-situ polymerization process for
large manufacturing volume and with low raw materials costs.
26
Chapter 2. Polyphthalamide
Bibliography
[1] J.G. Dolden, Structure-Property Relationships in Amorphous Polyamides,
POLYMER, Vol 17, 1976.
[2] R. Gabler, M. Muller, G.E. Ashby, E.R. Agouri, H.L. Meyer and G. Kabos,
Chimia, Vol. 21, No. 65, 1967.
[3] N. Pini, C. Zaniboni, S. Busato, P. Ermanni, Perspectives for Reactive Molding
of PPA Matrix for High-Performance Composite Materials, FPCM7 2004.
[4] H.P. Schmeer, Polyphthalamid, KGK Kautschuk Gummi Kunststoffe 46,
Jahrgang, Nr. 10/93, 1993.
[5] H.-J. Liedloff, M. Schmid, Process for Producing Precondensates of Partially
Crystalline or Amorphous, Thermoplastically Processable, Partially Aromatic
Polyamides or Copolyamides, European Patent Disclosure EP A 410 649 1995.
[6] G. Galland, J. Coquard, Preparation of Amorphous Polyamides based on
Aromatic Dicarboxylic Acids and on Aliphatic Diamines from Unbalanced
Polyamide Prepolymer, U.S. Pat. No. 4,963,646, Oct. 16, 1990.
[7] H.-J. Liedloff, M. Schmid, Process for Producing Precondensates of Partially
Crystalline or Amorphous, Thermoplastically Processable,Partially Aromatic
Polyamides or Copolyamides, U.S. Pat. No. 4,831,108, Jan. 13, 1998.
[8] G.T. Brooks, Naperville, Crystalline Polyphthalamide Composition Having Improved Properties, U.S. Pat. No. 5,098,940, Mar. 24, 1992.
28
BIBLIOGRAPHY
[9] H.-J. Liedloff, M. Schmid, Process for Producing Precondensates of Partially
Crystalline or Amorphous, Thermoplastically Processable, Partially Aromatic
Polyamides or Copolyamides, U.S. Pat. No. 5,708,125, Jan. 13, 1998.
Chapter 3
Reactive Processing of
Polyphthalamide in Carbon
Fiber Composites
The reactive processes developed so far for the manufacturing of fiber-reinforced thermoplastic composites can be divided into three groups: reactive liquid composite
molding (LCM), reactive injection pultrusion and reactive prepregs.
Reactive Liquid Composite Molding
Reactive liquid composite molding includes methods like infusion processes or lowpressure resin injection such as vacuum assisted resin infusion (VARI), resin film
infusion (RFI), resin transfer molding (RTM) or compression resin transfer molding
(CRTM). Prerequisites for these closed mold production methods, originally developed for thermosets, are that the thermoplastic precursor viscosity during impregnation has to be below 1 Pa·s [14] and that the polymerization reaction should proceed
without the formation of any by-products [1]. Resin injection and infusion into fibrous preforms has in general three advantages. The fibrous preforms made with reinforcement in several directions can eliminate interlaminar weakness. This is a real
30 Chapter 3. Reactive Processing of Polyphthalamide in Carbon Fiber Composites
advantage with brittle thermoset composites, but with tough thermoplastics the additional toughness gain not really compensates the loss in stiffness due to fiber crimping and out of plane fibers [2]. Real advantages are to be gained in making complex
shape by exploiting 10,000 years of textile technology [2]. Disadvantages of reactive
resin injection are the high tooling costs due to the involved pressure (5-10 bar) and
polymerization temperatures. Moreover, part size is limited, as a consequence of the
increase in clamping force required during injection to keep the mold closed as the
size of the product increases [3]. On the other hand, in infusion processes like reactive
VARI the atmospheric pressure is sufficient for mold clamping, making possible to
achieve virtually unlimited size of the parts and low tooling costs due to the low pressures involved [3]. The disadvantages are related to the flexible mold half, which can
be often used only once and leads to poor surface quality on one side of the product
[3]. A vacuum infusion process for manufacturing of polyamide-6 (PA6) composites
wind turbine blades is currently being developed at TU Delft [4], [5], [6], [7], while
extensive work has been conducted with the anionic polymerization of laurolactam
into polyamide-12 (PA12) at EPFL and IVW (Keiserslautern) [3], [8], [9], [1], [11],
[12]. Reactive oligomer precursors of polybutylene terephthalate (PBT) are currently
being investigated at K.U. Leuven and commercialized [3], [13], [14], [15], [16].
Reaction Injection Pultrusion
One of the major advantages of the pultrusion process is its continuous nature, which,
in principle, may enable a high degree of automation and process speed. However,
a limitation of the pultrusion process is that it is adapted only to produce straight,
elongated parts of uniform cross-section [10]. In case of reactive systems, the very
low viscosity of the prepolymers leads to very short impregnation time, improved
wetting, less void, higher fiber content and has the potential of high pultrusion line
speeds. The limiting step for an in-situ pultrusion process is the polymerization time
and therefore is the key factor in evaluating the cost-effectiveness of a reactive pultrusion process. Reaction injection pultrusion (RIP) has been investigated with different reactive engineering thermoplastics and in contrast to reactive LCM has already found industrial application, [3], [17], [18], [19], [20], [21], [10]. The reactive pultrusion processing of rigid thermoplastic polyurethanes (RTPU) based on a
depolymerization-repolymerization (DPRP) mechanism was recently developed by
Dow Chemicals (USA) and currently applied by the Fulcrum Composites Company
3.1 PA 6T/6I Reactive Processing Route
31
[3], [22], [23]. The DRTP allow for fast process speed (10m/min) without sacrificing
degree of impregnation.
Reactive Prepregs
Reactive prepregs are prepared by preimpregnating reinforcement fibers with a prepolymer. These prepolymer prepregs are subsequently shaped and polymerized in a
separate process step. This approach allows taking advantage of the existing technologies in the field of pre-impregnation processes [3], which manage remarkably the
fiber volume content and allow to achieve very high quality structural parts. A further advantage is found for partially impregnated prepregs in which the fibers are not
fully wetted out at the "prepreg" stage [3]. The voids present in that case represent an
important pathway for the extraction of volatile materials with condensation reaction
polymers, as shown by Gibbs [24]. Recently, Weixia et al. [3], [25], [26], developed
an in-situ polymerization process for long glass fiber-reinforced composites with the
oligomers of polycondensation polymers, such as oligomers of polyethylene terephthalate (PET), PBT, polycarbonate, PA6, polyamide-66, and polyamide-1010. This
process, named in situ solid-state polymerization (INSITU SSP) process, has the advantage of using intermediates of polycondensation polymers with a reduction in raw
material costs, especially if compared to cyclic oligomers [26].
Since PA6T/6I prepolymers undergo a polycondensation reaction with water as by
product, reactive prepregs is the most adequate reactive processing strategy for this
material system.
3.1
PA 6T/6I Reactive Processing Route
In order to develop a suitable processing strategy for PA 6T/6I prepolymers, the intermediate matrix material was first characterized from a thermal (DSC, TGA) and
rheological point of view.
3.1.1
Experimental
Because of their sensitivity to moisture, the PA 6T/6I oligomers (XE 3733 VK) were
dried in a vacuum oven at 80◦ C and 4 mbar for at least 24 hours before further use.
32 Chapter 3. Reactive Processing of Polyphthalamide in Carbon Fiber Composites
Optical Analysis
The PA6T/6I prepolymer powder distribution was observed by optical analysis with
LEICA RX DMA microscope equipped with a LEICA DC 480 digital camera.
Diffraction Spectrometer
A Helios Sympatec Laser Diffraction Spectrometer with different lens was used for
the determination of the size distribution of the powder particles.
Thermal Characterization
A Perkin Elmer Pyris 1 DSC instrument calibrated using indium was used to study
the melting, cold-crystallization and polymerization behavior of PPA oligomers. All
DSC non-isothermal tests were performed from 80◦ to 350◦ C at a constant heating
rate of 10◦ C/min under nitrogen atmosphere to prevent high temperature oxidation.
Rheological characterization
The complex viscosity of the neat PPA oligomers was investigated with a Paar Physica
UDS200 rheometer in a parallel plate configuration at different temperatures. All
measurements were carried out under nitrogen atmosphere to prevent degradation
or absorption of moisture. The prepolymer powder was pressed to pills of 20 mm
diameter at 25◦ C.
3.1.2
Results and discussion
The PA 6T/6I precondensates emerge from the production process in powder form,
as shown in Fig. 3.1, with a particle size distribution ranging from few µm to several
mm.
However, the prepolymers after cryogenic milling show a uniform grain size distribution below 20 µm, as shown in Fig. 3.2.
A typical DSC temperature scan of PA 6T/6I precondensates from 80◦ C to 350◦ C
at 10◦ C/min is shown in Fig. 3.3.
The as-received oligomers show seven distinct transitions in the heating (upper
curve) and cooling curve (lower curve).
33
3.1 PA 6T/6I Reactive Processing Route
ȝP
Figure 3.1: Micrograph of PA 6T/6I precondensates as emerges from the production
process.
16
PA6T/6I milled
14
12
Fraction
10
8
6
4
2
0
0
5
10
15
20
25
30
35
40
45
50
Particle size [µm]
Figure 3.2: Particle size distribution of PA6T/6I prepolymers after milling.
34 Chapter 3. Reactive Processing of Polyphthalamide in Carbon Fiber Composites
heat flow endo -> [MW]
15
DSC scan at 10°C/min
4
5
10
1
a)
5
0
3
2
7
b)
6
-5
-10
80
110
140
170
200
230
260
290
320
350
Temperature [°C]
Figure 3.3: DSC temperatures scan of PA 6T/6I precondensates from 80◦ C to 350◦ C
and from 80◦ C to 350◦ C at 10◦ C/min: a) heating curve b) cooling curve.
heat flow endo -> [MW]
15
4
5
TGA
DSC
10
1
99
3
2
5
100
98
97
0
96
-5
-10
95
80
110
140
170
200
230
260
290
320
94
350
Temperature [°C]
Figure 3.4: Superposition of a DSC and a TGA scan at 10◦ C/min under nitrogen of
the PA6T/6I precondensates.
3.1 PA 6T/6I Reactive Processing Route
35
TGA experiment under nitrogen at 10◦ C/min showed that up to 200◦ C no volatile
by-products are emitted, as shown in Fig. 3.4. Therefore, DSC peaks 1 and 2 are no
polycondensation peaks. Peak 1 represents a glass transition (Tg) step with a strong
pronounced enthalpy relaxation where the oligomers undergo softening. On further
heating from the glass transition the oligomers are brought to crystallize (peak 2).
PA6T/6I prepolymers are short chain molecules: in their solid state their mobility
is strongly reduced but with increasing temperature they can easily rearrange and
crystallize in a paraffin-like structure (fully extended chain crystals) without any entanglement between the chains. The stability of this configuration is further increased
in the case of polyamides, which can form hydrogen bonds between the chains. Peak
2 could also be interpreted as a first polymerization step, but the narrow peak is more
typical for crystallization processes and TGA analysis does not show evidence of
volatile reaction by-products at this temperature, which would be expected in case of
a polycondensation.
Between 220◦ C and 290◦ C about 3.5 wt% water is released, due to the polycondensation reaction taking place in correspondence to DSC peak 3. Therefore, between
220◦ C and 305◦ C peak 3 can be identified as a broad endothermal peak where solid
state polycondensation and evaporation of the reaction by-product i.e. water take
place. Peak 4 between 305◦ C and 320◦ C is probably a melting peak of the formed
polymer while Peak 5 between 320◦ and 345◦ C could be interpreted as a melt polymerization peak.
In the cooling curve (lower curve in Fig. 3.3) one crystallization peak from the melt
can be identified between 280◦ C and about 210◦ C and one glass transition temperature at 130◦ C of the final polymer.
During processing the PA6T/6I prepolymer viscosity changes by several orders of
magnitude and shows a pronounced shear thinning behavior, which causes the low
viscosity to be evident only at higher shear rates. The lowest observed viscosity is
more than 50 Pa·s at 180◦ C, which is still too high to allow injection of the oligomers
in reactive liquid composite molding, but it is one order of magnitude lower than that
of thermoplastic melts, so ensuring a more efficient fiber impregnation.
After heating the PA 6T/6I precondensates to 290◦ C, holding the temperatures for
10 minutes at 290◦ C and cooling to 80◦ C (Fig. 6.3), the DSC curve shows complete
absence of any peaks in the cooling curve, thus implying that the oligomers reacted
in solid state during the heating and holding phase. On second heating of the in-situ
polymerized polymer to 350◦ C only the melting peak between 320◦ C and 350◦ C can
36 Chapter 3. Reactive Processing of Polyphthalamide in Carbon Fiber Composites
be observed.
Therefore, PA6T/6I polymer matrix can be produced by solid state polymerization
of PA6T/6I precondensate without catalyst, in few minutes but with the release of
3.5 wt% water. These results open up new opportunities to process PA 6T/6I precondensates directly in solid phase at a polymerization temperature lower than the
melting temperature of the final polymer.
3.1.3
Conclusions
The viscosity of the molten PA 6T/6I precondensates and their polycondensation reaction do not allow them to be injected in a RTM process, since prerequisites for
closed mold production methods are that the thermoplastic precursor viscosity during impregnation has to be below 1 Pa·s and that the polymerization reaction should
take place without any by-products. On the contrary, reactive prepregs and in particularly partially impregnated prepregs in which the fibers are not fully wetted out at
the prepreg stage represent a viable processing solution in case of condensation reaction polymers. Moreover, the PA 6T/6I precondensates emerge from the production
process in powder form and the precondensate powder after milling shows a uniform
grain distribution around 20 µm, which allows a homogeneous powder impregnation
of carbon fiber rovings, as well by dry as by wet powder impregnation methods.
The reactive forming process can be divided therefore in three separate processes
steps (Fig. 3.5):
1. Impregnation of carbon fiber tows with the prepolymer by powder impregnation
technologies.
2. Textile process for the production of fabrics or preforms with the prepolymer
powder impregnated tows.
3. Fully impregnation of the fibers, shape and polymerization of the prepolymer
prepreg in a press under nitrogen atmosphere, to prevent high temperature oxidation.
The third process step in the press can be also divided in two separate stages.
In the first stage the prepolymer powder coated prepregs are heated to the softening
temperature of the oligomers in a hot press and pressed so that the oligomers - thanks
37
3.1 PA 6T/6I Reactive Processing Route
1. Prepreg Technologies
2. Textile Technologies
Powder Impregnation
Prepolymer Prepregs
Textile Process
Prepolymer Preforms
3. Reactive Forming
Shaping, Impregnation
Polymer Composite Parts
Crystallization, In-situ Polymerization
Figure 3.5: Schematic of the reactive processing route of PA6T/6I.
38 Chapter 3. Reactive Processing of Polyphthalamide in Carbon Fiber Composites
to their low melt viscosity - can easily flow and achieve high wetting and impregnation
quality of the fibers. In the second step the temperature is raised to a value between
the prepolymer glass transition and the onset of polymer melting in order to induce
the oligomers to first crystallize and then to polymerize in solid state while the entire
mass of material is shaped. Polymerization, crystallization, consolidation and shaping
can occur below the melt temperature of the polymer, with a consistent reduction of
the processing time and energy consumption.
Bibliography
[1] A. Luisier, P.-E. Bourban, J.-A.E. Månson, Time-Temperature-Transformation
Diagram for Reactive Processing of Polyamide 12, J. Appl. Polym. Sci., 81,
2001, 963-972.
[2] F.N. Cogswell, Thermoplastic Aromatic Polymer Composites, ButterworthHeinemann, 1992.
[3] K. van Rijswijk, H.E.N. Bersee, Reactive Processing of Textile Fiber-Reinforced
Thermoplastic Composites - An Overview, Composites: Part A, 38, 2007, 666681.
[4] K. van Rijswijk, S. Joncas, H.E.N. Bersee, O.K. Bergsma, A. Beukers, Sustainable Vacuum-Infused Thermoplastic Composites for MW-Size Wind Turbine
Blades-Preliminary Design and Manufacturing Issues, Journal of Solar Energy
Engineering, Vol. 127, November 2005, 570-580.
[5] K. van Rijswijk, H.E.N. Bersee, W.F. Jager, S.J. Picken, Optimisation of Anionic Polyamide-6 for Vacuum Infusion of Thermoplastic Composites: Choice
of Activator and Initiator, Composites: Part A37, 2006, 949-956.
[6] K. van Rijswijk, S. Lindstedt, D.P.N. Vlasveld, H.E.N. Bersee, A. Beukers, Reactive Processing of Anionic Polyamide-6 for Application in Fiber Composites:
A Comparative Study with Melt Processed Polyamides and Nanocomposites,
Polymer Testing 25, 2006, 873-887.
40
BIBLIOGRAPHY
[7] K. van Rijswijk, H.E.N. Bersee, A. Beukers, S.J. Picken, A.A. van Geenen, Optimisation of Anionic Polyamide-6 for Vacuum Infusion of Thermoplastic Composites: Influence of Polymerization Temperature on Matrix Properties, Polymer
Testing 25, 2006, 392-404.
[8] J. Verrey, M.D. Wakeman , V. Michaud, and J.-A.E. Månson, Manufacturing Cost Comparison of Thermoplastic and Thermoset RTM for an Automotive
Floor Pan, Composites, Part A 37, 2006, 9-22.
[9] M.D. Wakeman, L. Zingraff, P.-E. Bourban, J.-A.E. Månson, P. Blanchard,
Stamp Forming of Carbon Fiber/PA12 Composites - A Comparison of a Reactive Impregnation Process and a Commingled Yarn System, Composites Science
and Techology 66, 2006, 19-35.
[10] A. Luisier, P.-E. Bourban, J.-A.E. Månson, Reaction Infection Pultrusion of
PA12 Composites: Process and Modelling, Composites: Part A 34, 2003, 583595.
[11] P. Rosso, K. Friedrich, A. Wollny, Evaluation of the Adhesion Quality between
differently Treated Carbon Fibers and an in-Situ Polymerized Polyamide 12 System, Journal of Macromolecular Science, Part B - Physics, Vol. B41, Nos. 4-6,
2002, 745-759.
[12] P. Rosso, K. Friedrich, A. Wollny, R. Mülhaupt, A Novel Polyamide 12 Polymerization System and its Use for a LCM-process to Produce CRFP, Journal of
thermoplastic Composite Materials, Vol. 18, 2005, 745-759.
[13] Presentation of Cyclics Coorporations thanks to Mr. Rösch of Cyclics Coorporation Europe for this document
[14] H. Parton, I. Verpoest, In situ Polymerization of Thermoplastic Composites
Based on Cyclic Oligomers, Polymer Composites, Vol. 26, Issue 1, February
2005, 60-65.
[15] H. Parton, J. Baets, P. Lipnik, B. Goderis, J. Devaux, I. Verpoest, Properties
of poly(butylene terephthatlate) polymerized from cyclic oligomers and its composites, Polymer 46, 2005, 9871-9880.
BIBLIOGRAPHY
41
[16] H. Parton, I. Verpoest, Thermoplastic Liquid Composite Molding - Production
and Characterization of Composites based on Cyclic Oligomers, Proceedings of
7th international conference on composite materials, Newark (DE, USA), July
14-18, 2003.
[17] C-C.M. Ma, C-H. Chen, Pultruded Fiber-Reinforced Poly(methyl Methacrylate)
Composites. I. Effect of Processing Parameters on Mechanical Properties, Journal of Applied Polymer Science, Vol. 44, 1992, 807-817.
[18] C-C.M. Ma, C-H. Chen, Pultruded Fiber-Reinforced Poly(methyl Methacrylate)
Composites. Part II. Mechanical and Thermal Properties, Polymer Engineering
and Science, Vol. 31, No. 15, 1991, 1094-1100.
[19] C-C. M. Ma, C-H. Chen, Pultruded Fiber-Reinforced Poly(methyl Methacrylate)
Composites. Part I. Correlation of Processing Parameters for Optimizing the
Process, Polymer Engineering and Science, Vol. 31, No. 15, 1991, 1086-1092.
[20] B-G. Cho, S.P. Mccarthy, Fiber Reinforced Nylon-6 Composites Produced by
the Reatcion Injection Pultrusion Process, Polymer Composites, Vol.17, No. 5,
1996, 673-681.
[21] M.G. Dubé, G.L. Batch, J.H. Vogel, C.W. Macosko, Reaction Injection Putrusion of Thermoplastic and Thermoset Composites, Polymer Composites, Vol.
16, No. 5, 1995, 378-385.
[22] Edwards et al. Fiber-Reinforced Composite and Method of Making Same, US
Patent 5,891,560, 1999.
[23] Edwards et al. Process for in-line Forming of Pultruded Composites, US Patent
6,872,343, 2005.
[24] H.H. Gibbs, K-Polymer: a New Experimental Thermoplastic Matrix Resin for
Advanced Structural Aerospace Composites, 29Th National SAMPE Symposium, 1585-1601.
[25] Weixia Yan, KeQing Han, LiuLiang Qin, MuHuo Yu, Study on Long FiberReinforced Thermoplastic Composites Prepared by In Situ Solid-State Polycondensation, Journal of Applied Polymer Science, Vol. 91, 2004, 3959-3965.
42
BIBLIOGRAPHY
[26] Weixia Yan, KeQing Han, Hongmei Zhou, MuHuo Yu, Study on the Grafting of
PET onto the Glass Fiber Surface during In Situ Solid-State Polycondensation,
Journal of Applied Polymer Science, Vol. 99, 2006, 775-781.
Chapter 4
Material Characterization of
the Oligomer Precursor
Impregnation of continuous fiber reinforcement with the thermoplastic matrix in powder form, basically consists of three distinct process steps (Fig. 4.1): spreading of the
tow fiber filaments to a tape, impregnation of the spread fibers with polymer powder and subsequent thermal process to attach the powder particles on the reinforcing
filaments.
The result is a partially impregnated tow or tape preform or prepreg. Various
techniques have been developed by several research groups for spreading the fiber
filaments, for dispersing the polymer powder and for impregnating the spread fiber
Tow spreading Powder Impregnation Stabilization
Tow take off spool
Tow take up spool
Figure 4.1: Powder impregnation technologies.
44
Chapter 4. Material Characterization of the Oligomer Precursor
tow. Based on the method used for dispersing the polymer powder, we will classify
these processes in two distinct groups:
1. Wet powder impregnation techniques also comprehending suspension, emulsion or slurry coating techniques, which consist in dispersing the polymer powder in a non-solvent liquid carrier to form a suspension or emulsion through
which the spread fiber tow is drawn [1], [2], [3], [4], [5] , [6] , [7], [8]. The
need of an additional drying process for evaporating the liquid carrier and the
need of high content of emulsifiers or suspending agents, which often remain
in the finished composite, adversely affect the processing and the mechanical
performance of the final composite [9].
2. Dry powder impregnation techniques also called powder coating techniques do
not involve the use of a liquid carrier. The powder can be directly seeded on the
spread fibers or suspended by a gas forming a fluidized bed where the spread
fibers are drawn. The powder loaded on the fiber tow or tape has to be subsequently fixed in order to prevent loss of powder during further manipulation.
Dry powder coating has the significant advantages that no liquid carrier is required
and therefore high process speed i.e. cost-effective processing can be achieved. For
the manufacturing of PA 6T/6I prepolymer powder impregnated preforms, we have
also to take in consideration that residual moisture and emulsifiers or suspending
agents negatively affect the subsequent in-situ polymerization reaction. Moreover,
the fiber tow in wet powder impregnation is spread by the powder particles which
are forced by the pins and act as spacers between the fibers [1]. As a consequence
fiber spreading, the degree of fiber matrix mingling and the fiber volume content are
strongly related with the size of the polymer powder: for a fiber-matrix volume fraction of 60%, a corresponding particle diameter of 5 µm should be used [1]. However,
such small particle size can be obtained only through expensive cryogenic milling
[10]. In addition with small particles diameter the attractive van der Waals’s forces
become important giving rise to the problem of agglomeration, which can only be
prevented using a high content of additives in the impregnation bad [1].
In this chapter the prepolymer material system is characterized in order to determine a suitable process window for the impregnation of carbon fiber rovings with
45
4.1 Experimental
the prepolymers and in order to develop a dry impregnation process that suit to the
properties of the material system.
4.1
Experimental
Because of their sensitivity to moisture, the PA 6T/6I oligomers (XE 3733 VK) were
dried in a vacuum oven at 80◦ C and 4 mbar for at least 24 hours before further use.
4.1.1
Thermal Characterization
A Perkin Elmer Pyris 1 DSC instrument calibrated using indium was used to study
the melting, cold crystallization and polymerization behavior of the PPA oligomer.
All DSC non-isothermal tests were performed from 80◦ to 350◦ C at a heating rate
of 10◦ C/min under nitrogen atmosphere to prevent high temperature oxidation. Cold
crystallization of the low molecular weight PPA oligomer was found to be too rapid
to be studied by DSC under isothermal conditions [11]: The cold-crystallization half
times (i.e. the time at which one-half of the crystallization has taken place) of the
isothermal crystallization were shorter then heating times or were comparable to the
time constant of the DSC instrument. In order to study the rapid kinetics of the cold
crystallization a different experimental approach had to be chosen.
An aluminum pan with the oligomer powder and joined to an aluminum tube was
immersed in a thermostated oil bath at various temperatures and for different times
and then rapidly quenched in water at 0◦ C (Fig. 4.2). The aluminum tube allowed
maintaining constant atmospheric pressure in the pan. By subsequent DSC analysis
of the partially crystallized samples the extent of induced crystallization and the cold
crystallization half times were determined. The area above the cold crystallization
curve of the as received PA 6T/6I oligomer measured by DSC scan at 10◦ C/min from
80 to 350◦ C was taken as a repeatable reference value of 100% crystallization. The
percentage of cold crystallinity was determined using Equation 4.1 [12]:
%crystallinity = 100% − (
∆Hc0 − ∆Hc
· 100%)
∆Hc0
(4.1)
Where ∆Hc0 is the cold crystallization enthalpy of the as-received oligomer and
∆Hc is the cold crystallization enthalpy of the oligomer previously heated to various
46
Chapter 4. Material Characterization of the Oligomer Precursor
Patm
Oligomer Powder
7 mm
Oil bath
9 mm
Figure 4.2: Experimental set up for the study of the cold crystallization kinetics of
PPA oligomers.
temperatures and for various time.
4.1.2
Rheological characterization
Solution Viscosity
Viscosity measurements of oligomer solution in 98% sulphuric acid were carried out
at concentrations of 0.1 g/dL, 0.3 g/dL, 0.5 g/dL und 0.7 g/dL in a Paar Physica MC
300 rheometer using Couette geometry. The intrinsic viscosity [µ] of the oligomers
was calculated from the reduced viscosity by extrapolation of ηsp /c to zero concentration.
Complex Viscosity
The complex viscosity of the neat PPA oligomer was investigated with a Paar Physica
UDS200 rheometer in a parallel plate configuration. All measurements were carried
out under nitrogen atmosphere to prevent degradation or absorption of moisture. The
4.1 Experimental
47
oligomer powder was pressed to pills of 20 mm diameter with different pressure between 160 MPa and 1 MPa at 140◦ C.
Torsional Dynamic Resonance Rheometer
When a rod is performing torsional vibrations at one of its resonance frequencies, any
interaction with a viscoelastic medium changes both its resonance frequency and its
damping characteristics. By measuring this change, one can obtain the viscoelastic
properties of the medium.
In this study a new high frequency torsional dynamic resonance rheometer developed by the Institute of Mechanical Systems of ETH Zurich was used to characterise
the oligomeric material. The dynamic rheometer consists of an outer tube rigidly
joined at one end to a cylindrical inner rod through an end plate (Fig. 4.3, A). The
tube is free of loading along its lateral surface and attached at the other end (Fig. 4.3,
B) to a thick plate of large diameter in comparison to the diameter of the tube. Since
the torsional rigidity of the plate is much larger than the tube, it acts as a decoupling
mass enforcing a node of the torsional vibration mode in its immediate vicinity. With
an electromagnetic transducer, fixed at the free end of the internal rod, the system is
forced to perform high frequency vibrations of very low amplitude at one of its torsional eigenmodes. The frequency is stabilized with the help of a phase-locked loop
fixing the phase between the applied torque and the measured angle of rotation.
With the help of a suitable theoretical model of the vibrating system, one can calculate
from the measured frequencies for given phase angles the real part (G’) and imaginary
part (G”) of the complex shear modulus (G*) of the material system. Further details
can be found in literature [13]. The elastic part, mainly related to the increase in
resonance frequency, and the viscous constant, mainly connected with the frequency
difference, measured by the torsional dynamic resonance rheometer were not correlated with absolute value of viscosity. The measured values are simply compared in
relation to the variation of the processing parameters of the produced samples.
The oligomer powder was pressed to pills with 8 mm diameter at 140◦ C with different pressure (Fig. 4.4) and viscosity measurements were performed with the high
frequency torsional dynamic resonance rheometer at different constant temperatures.
48
Chapter 4. Material Characterization of the Oligomer Precursor
Electromagnetic Transducer
B
Decoupling Mass
Inner Rod
Outer Tube
End Plate
A
Polymer
Figure 4.3: Schematic of the torsional dynamic resonance rheometer [13].
Compressed air
Control valve
Moving rod
Precondensate powder
Die
8 mm
Figure 4.4: Schematic of the apparatus for the production of pills.
4.2 Results and discussion
4.2
4.2.1
49
Results and discussion
Thermal characterization of PA6T/6I oligomer
A typical DSC temperature scan of PA 6T/6I oligomer from 80◦ C to 350◦ C at 10◦ C/min
is shown in Fig. 4.5. The as-received material shows seven distinct transitions in the
heating (upper) and cooling curve (lower).
TGA measurements had shown that up to 200◦ C no volatile by-products are emitted
(Fig. 3.4). Therefore, DSC peaks 1 and 2 are no polycondensation peaks. Rather,
Peak 1 represents a glass transition (Tg ) with a pronounced enthalpy relaxation where
the oligomers undergoes softening. On further heating from the glass transition the
oligomer is induce to crystallize (peak 2). Peak 3 between 220◦ C and 305◦ C can be
identified as a broad endothermal peak where polycondensation and evaporation of
the reaction by-product water take place. Peaks 4 and 5 between 305◦ C and 345◦ C
are probably a melting polymerization and melting peak of the formed polymer. In
the cooling curve (lower curve in Fig. 4.5) one crystallization peak (6) from the melt
between 280◦ C and about 210◦ C as well as the glass transition (7) at 130◦ C of the
final polymer can be identified.
In this chapter, we focus on the first two peaks, in the temperature range where
a powder impregnation process for the production of oligomer prepregs is viable.
Polymerization (peak 3) is to be induced in a second separate process step after consolidation of the powder impregnated intermediate materials and while the workpiece
is being shaped. The polymerization has to be conducted in vacuum or protective
atmosphere in order to avoid oxidative degradation of the material. Additionally,
the oligomer powder does require drying before reactive processing. Any humidity
present would lead to agglomeration of the powder particles during powder impregnation and negatively affect the polymerization reaction in the reactive stamp forming
process.
4.2.2
Glass transition of PA6T/6I oligomer
Peak 1 represents a glass transition (Tg ) with a pronounced enthalpy relaxation [14],
[15], [16] where the oligomer undergoes softening. Fig. 4.6 shows a DSC scan of a
sample of PA6T/6I oligomer as-received from the polymerization reactor and dried
50
Chapter 4. Material Characterization of the Oligomer Precursor
heat flow endo -> [mW]
15
10°C/min
4
5
10
1
5
a)
0
b)
3
2
7
6
-5
-10
80
110
140
170
200
230
260
290
320
350
Temperature [°C]
Figure 4.5: DSC scan of PA 6T/6I oligomer between 80◦ C and 350◦ C at 10◦ C/min:
a) heating curve b) cooling curve.
at 80◦ C and 4 mbar for 48 hours, representing typical drying conditions. The glass
transition temperature (Tp ) and the relaxation enthalpy (∆H) of the PA6T/6I oligomer
are strongly dependent on the thermal history of the sample, as shown in Fig.4.7. In
a first DSC scan a sample dried at 80◦ C and 4 mbar for several months was heated
to 140◦ C at 10◦ C/min and held at this temperature for several minutes in order to
erase its thermal history. Afterwards the sample was cooled to 80◦ C at 10◦ C/min in
order to induce a defined and reproducible thermal history, i.e. a "defined amorphous"
condition [15]. Finally the glass transition temperature was redetermined in a second
DSC scan.
Thus, a large difference of 26.8◦ C between the glass transition temperatures of
the same oligomer sample was observed, due only to the different thermal history. On
the other hand, the glass transition temperature of the PA6T/6I oligomer after holding
the samples at 140◦ for one minute was found to be very reproducible (Tg = 127.9◦ C
± 0.54◦ C on three different samples).
This phenomenon well known in amorphous polymers is called physical aging
and in contrast to chemical or biological aging does not involve an irreversible modification of the material structure [16]. On slow cooling, the molecules can better
51
4.2 Results and discussion
1st
Tp=136.96°C
heat flow endo -> [mW]
6
Delta H= 5.3 J/g
1
4
Tg= 129.89°C
2
0
90
100
110
120
130
140
150
Temperature [°C]
Figure 4.6: DSC scan from 80◦ C to 140◦ C at 10◦ C/min showing glass transition
temperature and enthalpy relaxation of PA6T/6I oligomer dried at 80◦ C and 4 mbar
for 48 hours.
rearrange (internal relaxation), shifting the glass transition temperature to lower temperatures. Moreover, amorphous polymers not in thermodynamic equilibrium tend to
gradually approach equilibrium at temperatures well below the glass transition [17].
The enthalpy lost during this relaxation process, i.e. the enthalpy difference ∆H (enthalpy relaxation) between an annealed (relaxed) and non-annealed glass, is a simple
indication of the structural changes on annealing. For the present material, this phenomenon is enhanced by the short chain length of the PA6T/6I oligomer and explains
the large differences in glass transition temperatures and relaxation enthalpies observed for samples dried at 80◦ for 48 hours (Fig. 4.6) and several months (Fig. 4.7),
respectively.
From the analysis of the glass transition temperature we can conclude that the impregnation of carbon fiber with the PA6T/6I oligomer in the fixing unit of the powder
impregnation plant should be performed at temperatures above 140◦ C, a temperature
above Tg as measured after the typical drying procedure (Fig. 4.6).
52
Chapter 4. Material Characterization of the Oligomer Precursor
heat flow endo -> [mW]
5
1st
2nd
Delta H = 0.3J/g Tp= 105.7°C
4
1
Tg= 100.6°C
Tp= 132.5°C
Delta H = 0.6J/g
Tg= 128.5°C
a)
3
2
1
b)
80
90
100
110
120
130
140
Temperature [°C]
Figure 4.7: DSC scan from 80◦ C to 140◦ C at 10◦ C/min of PA6T/6I oligomer dried
at 80◦ C and 4 mbar for several months: a) first heating b) second heating of the same
sample.
1st
2nd
heat flow endo -> [mW]
1.0
0.5
1
2
a)
1
b)
0.0
-0.5
-1.0
-1.5
2
80
100
120
140
160
180
Temperature [°C]
Figure 4.8: DSC scan at 10◦ C/min of PA 6T/6I oligomer: a) first heating from 80◦ C
to 180◦ C; b) second heating from 80◦ C to 350◦ C.
53
4.2 Results and discussion
Heating rate (◦ C/min)
10
50
100
T0.01 (◦ C)
157.9
155.4
151.5
Tc (◦ C)
171.2
181.2
183.6
T0.01 (◦ C)
180.7
196.7
205.8
∆Hc (J/g)
36.7
41.4
41.4
Table 4.1: Characteristic data of non-isothermal cold crystallization exotherms for
PA6T/6I oligomer as measured by DSC at different heating rates.
4.2.3
Cold crystallization of PA6T/6I oligomer
On further heating beyond the glass transition temperature the PPA oligomer is induced to crystallize at about 170◦ C (peak 2). This cold crystallization behavior arises
from the short chain lengths and low level of entanglements of the molecules: in the
glassy state the mobility of the short chain molecules is strongly reduced but above
the glass transition temperature they can easily rearrange and crystallize in a paraffinlike structure (fully extended chain crystals). The stability of this configuration is
further increased in the case of polyamides, which can form hydrogen bonds between
the chains [11]. The crystallized oligomer cannot be softened anymore, as shown in
Fig. 4.8 from the absence of a melting peak in the second heating curve. On further heating the oligomer polymerizes in the crystalline state and the material can be
melted only at the melting temperature of the final polymer (between 300 and 350◦ C).
The crystallization process is controlled either by the formation of the nuclei or
by the diffusion of the molecular segments. The non-isothermal cold crystallization
exotherms of PA 6T/6I oligomer were determined by DSC for three different heating rates. The temperature at 1% relative crystallinity (T0.01 ), the temperature of
the maximum crystallization rate, i.e., the peak temperature (Tc ), the temperature at
99% relative crystallinity (T0.01 ) and the crystallization enthalpy are summarized in
Tab. 4.1. The crystallization exotherm becomes wider and the maximum crystallization rate shifts toward higher temperature with increasing heating rate, as expected
for a crystallization in a diffusion-controlled region.
It is very important during the melting stage of the powder impregnation process
to avoid cold crystallization of the PA 6T/6I oligomer, in order to permit subsequent
final consolidation and shape of the preforms in the molding process. The cold crystallization rate of the oligomer was investigated by determining the cold crystalliza-
54
Chapter 4. Material Characterization of the Oligomer Precursor
Crystallization half time [s]
1000
100
10
50% Crystallization
150
160
170
180
190
Temperature [°C]
Figure 4.9: Crystallization half time of PA 6T/6I oligomer under isothermic conditions. The solid line is a polynomial fit curve.
tion half time at different constant temperatures, i.e. the time required to achieve
50% crystallization in the material system. At a temperature lower than 155◦ C,
50% of crystallization occurs in a matter of minutes, while above 155◦ C the halfcrystallization time decreases to seconds. At 180◦ the oligomer is fully crystalline
after a few seconds.
4.2.4
Rheological characterization of PA6T/6I oligomer
The intrinsic viscosity of short chain PPA oligomer was found to be very low [µ]
= 0.07 dl/g especially if compared, for example, to low molecular weight PET and
PEN oligomers, [µ] = 0.2 dl/g [17]. The low molecular weight oligomer softens
between 105◦ and 135◦ C, but was found to melt only under shear force. In order to
measure its complex viscosity [µ∗ ] the oligomer powder was pressed to pills of 20 mm
diameter at 140◦ C under various pressures. Measured viscosity values were found to
be a function of the pressure applied during the manufacturing of the pills. Complex
viscosity of pills pressed at 140◦ C and 160 MPa pressure, measured versus time at
140◦ C is very high ([µ∗ ] > 104 Pa·s, Fig. 4.11). On the other hand, complex viscosity
55
4.2 Results and discussion
Applied pressure(MPa)
Powder as-received
7
8
9
T0.01 (◦ C)
157.9
156.9
155.4
153.4
Tc (◦ C)
171.2
170.8
168.2
165.5
T0.01 (◦ C)
180.7
177.9
178.5
174.9
∆Hc (J/g)
36.7
33.3
33.3
27.5
Table 4.2: Characteristic data of non-isothermal cold crystallization exotherms for
PA6T/6I oligomer as loose powder and pressed at different pressure as determined by
DSC at 10◦ C/min.
of pills pressed at 140◦ C and 1 MPa pressure show much lower values ([µ∗ ] > 102
(Fig. 4.10)).
The temperature at 1% relative crystallinity (T0.01 ), the temperature of the maximum crystallization rate, i.e., the peak temperature (Tc ), the temperature at 99% relative crystallinity (T0.99 ) and the crystallization enthalpy of the PPA oligomer samples,
both as loose powder and pressed under different pressure, are shown in Tab. 4.2.
Compared to the unpressed oligomer powder, the cold crystallization peaks of the
pressed powder are found at lower temperatures. Crystallization enthalpy tends to
decrease with increasing pressure, as well as the onset temperature at which the crystallization begins and the end temperature at which the crystallization ceases. This
behavior could be rationalized by an increase of orientation of the short oligomer
molecules under applied pressure; in other words, the pressure gives rise to a more
ordered structure as in the case of drawing and cold crystallization in poly(ethylene
terephthalate) fibers [18]. The ordered oligomer segments can easily arrange into the
crystal lattice, shifting the onset and the end temperature of crystallization to lower
temperatures; also, the orientation of amorphous phase promotes a substantial increase in crystallinity, as evidenced by the lower crystallization enthalpies. No melting peak is observed between the crystallization of the oligomer and polymerization.
The existing crystalline component is not destroyed.
The rotational movement of the rheometer in the parallel plate configuration was
found to also induce crystallization of the oligomer. Thus, instead of preparing pills
by pressing the powder, pills could also be manufactured at 140◦ C with a rotating rod
of equal diameter to that of the pills. The induced crystallization of the oligomer as a
function of time was clearly evidenced by DSC and visual inspection.
56
Chapter 4. Material Characterization of the Oligomer Precursor
107
Complex viscosity [Pa s]
106
105
104
103
measuremet at 140°,
ω = 10 (rad/s), Fn = 2N
102
101
0
200
400
600
800
1000
1200
Time [s]
Figure 4.10: Complex viscosity versus time of PA6T/6I oligomer powder measured
on pills prepared at 140◦ C with 160 MPa pressure.
107
mesurement at 140°,
ω = 10 rad/s, Fn = 2N
Complex viscosity [Pa s]
106
105
104
103
102
101
0
50
100
150
200
250
300
Time [s]
Figure 4.11: Complex viscosity versus time of PA6T/6I oligomer powder measured
on pills prepared at 140◦ C with 1 MPa pressure.
57
4.2 Results and discussion
20
9.4 MPa
7.8 MPa
4.7 MPa
18
Viscous Constant (Hz)
16
14
12
10
8
6
4
2
0
0
20
40
60
80
100
120
Time (s)
Resonance Frequency (Hz)
8000
9.4 MPa
7.8 MPa
4.7 MPa
7800
7600
7400
7200
7000
0
20
40
60
80
100
120
Time (s)
Figure 4.12: Viscosity values versus time calculated with torsional resonance
rheometer of PA6T/6I prepolymer powder sample pressed at 140◦ C with different
pressures for the same time.
58
Chapter 4. Material Characterization of the Oligomer Precursor
20
1s
1min
8min
Viscous Constant (Hz)
18
16
14
12
10
8
6
4
2
0
0
20
40
60
80
Time (s)
100
Resonance Frequency (Hz)
8000
120
1s
1min
8min
7800
7600
7400
7200
7000
0
20
40
60
Time (s)
80
100
120
Figure 4.13: Viscosity versus time determined with a torsional resonance rheometer
for PA6T/6I oligomer powder pressed at 140◦ C with 4.7 MPa pressure for different
times (c) and d)).
4.3 Conclusions
59
In order to study the viscosity behavior of the oligomer without the influence of
the parallel plate rheometer, the effect of pressure on the cold crystallization was investigated with a torsional resonance rheometer. In the torsional resonance rheometer
the pills were placed on a hot plate (140◦ C) and the measurement started almost instantly. Both the viscous constant and the elastic part (resonance frequency) increased
with increasing pressure applied in the preparation of the pills (Fig. 4.12), indicating
a stiffening of the material that can be attributed to partial crystallization. The effect
is more pronounced in 4.13, where the oligomers was pressed at 140◦ C with the
same pressure (4.7 MPa), but for different times. In conclusion, it is very difficult
to give a precise value of the viscosity of the PA 6T/6I oligomer. For the viscosity
measurements the oligomer has to be pressed to pills at 140◦ C, which in turn induces
partial crystallization. Crystallization is also promoted by the rotational movement of
the rheometer plates. Additionally, for viscosity measurements the pills are brought
to the desiderated temperature by heating the rotating plates. This again is likely to
promote further crystallization in the material before the start of the actual measurement.
4.3
Conclusions
A reactive oligomeric PPA precursor was characterized in order to optimize the impregnation of reinforcing fibers. Low molecular weight PA6T/6I oligomer softens
at 135◦ C and polymerizes above 200◦ C. The polymerization process has to be conducted in vacuum or protective atmosphere; polymerization has to be avoided during powder impregnation. Between softening and polymerization temperature, the
oligomer undergoes an irreversible cold crystallization. It is important during the
melting stage of the powder impregnation process to avoid cold crystallization, in
order to consolidate and shape the preform in the subsequent molding stage. Above
155◦ half crystallization of the oligomer occurs in only few seconds. However, the
cold crystallization of the oligomer is promoted not only by elevated temperature but
also by increasing applied pressure: high pressure induces the oligomer to crystallize even at temperatures below its softening temperature. Unfortunately, the low
viscosity of the oligomer could not be measured due to the necessity of pressing the
oligomer powder to pills for the rheological measurements, leading to irreversible
cold crystallization.
60
Chapter 4. Material Characterization of the Oligomer Precursor
Bibliography
[1] A.M. Vodermayer, J.C. Kaerger, G. Hinrichsen, Manufacture of High Performance Fiber-Reinforced Thermoplastics by Aqueous Powder Impregnation,
Composites Manufacturing, Vol 4 No 3. 1993.
[2] F.V. Lacrox, H.Q. Lu, K. Schulte, Wet Powder Impregnation for Polyethylene
Composites: Preparation and Mechanical Propeties, Composites: Part A 30,
1999, 369-373.
[3] D.A. Evans, Hybrid Composite Articles and Methods for their Production,
Patent US 6,861,131, 2005.
[4] D. A. Soules, Apparatus and Process for Improved Thermoplastic Prepreg Materials, Patent US 5,019,427, 1991.
[5] B. Pfeiffer, D. Skaletz, H. Heckel, A. Texier, J. Heydweiller, Production of
Fiber-Reinforced Composites by Pultrusion with Thermoplastic Powder Pretreatment, Patent US 5,725,710, 1998.
[6] A. M. Vodermayer, J. Krger, S. Kaufmann, H. Erlach, Manufacture of Unidirectional Fiber Reinforced Thermoplastic, Patent US 6,372,294, 2002.
[7] G. Hinrichsen, A. Vodermayer, K.-H. Reichert, L. Kuhnert, W.G. Lindner, G.
Goldmann, Production of Composites from Polymer Powder Dispersions, Patent
US 5,888,580, 1999.
62
BIBLIOGRAPHY
[8] E.M.J. Morel, G.M. Richert, Process for Impregnating with Thermoplastic
or Thermosetting Polymers in Solid State Fiber of Great Length, Patent US
4,828,776, 1989.
[9] J. L. Throne, Method of Depositing and Fusing Polymer Particles onto Moistened Continuous Filaments, Patent US 5,364,657, 1994.
[10] M. Connor, Consolidation Mechanisms and Interfacial Phenomena in Thermoplastic Powder Impregnated Composites, These N◦ 1423, 1995.
[11] H. G. Kim, R.E. Robertson, A New Approach for Estimating the Recrystallization Rate and Equilibrium Melting Temperature, Journal of Polymer Science:
Part B: Polymer Physics„ Vol. 36, 1998, 133-141.
[12] W.J. Sichina, DSC as Problem Solving Tool: Measurement of Percent Crystallinity of Thermoplastics, PerkinElmer Insturments.
[13] J. Goodbread, M. Sayir, K. Häusler, J. Dual, Method and Device for Measuring
the Characteristics of an Oscillating System, US Patent No. 5,837,885, European Patent No. 0749570. 1998
[14] S. Affolter, A. Ritter, M. Schmid, Interlaboratory Tests on Polymers by Differential Scanning Calorimetry (DSC): Determination of Glass Transition Temperature (Tg ), Macromol. Mater. Eng., 286, 2001, 605-610.
[15] J. M. Hutchinson, Physical Aging of Polymers, Progress in Polymer Science,
Vol. 20, 1995, 703-760.
[16] Y. Di, A. D’Amore, G. Marino, L. Nicolais and B. Li, Physical Aging of HighPerformance Thermoplastics: Entalpy Relaxation in PEK-C and PES-C, Journal
of Applied Polymer Science, Vol. 57, 1995, 989-995.
[17] N.R. James, C. Ramesh, S. Sivaram, Development of Structure and Morphology
during Crystallization and Solid State Polymerization of Polyester Oligomers,
Macromol. Chem. Phys., 202, 2001, 1200-1206.
[18] Z. Zhang, S. Wu, M. Ren, C. Xiao, Model of Cold Crystallization of Uniaxially
Oriented Poly(ethylene Terephthalate) Fibers, Polymer 45, 2004, 4361-4365.
BIBLIOGRAPHY
63
[19] P. Supaphol, N. Apiwanthanakorn, Nonisothermal Cold-Crystallization Kinetics of Poly(Trimethylene Terephthalate), Journal of Polymer Science: Part B:
Polymer Physics, Vol. 42, 2004, 4151-4163.
64
BIBLIOGRAPHY
Chapter 5
Electrostatic Dry Powder
Impregnation Process
A dry powder impregnation process was developed for the production of PA6T/6I prepolymer impregnated tows, considering, as summarized in Tab. 5.1, the many positive
aspects in correlation with in-situ polymerizing systems and its cost effectiveness. In
this chapter the most important aspects of dry powder impregnation processes are
reviewed and finally the powder coating line developed at the Center of Structure
Technologies will be presented.
Advantages
High process speed
No need of additives
Low cost of powder
Wet powder impregnation
-
Dry powder impregnation
+
+
+
Table 5.1: Advantages of the different powder impregnation technologies.
66
Chapter 5. Electrostatic Dry Powder Impregnation Process
win = 3 mm
wmax = 85.2 mm
Figure 5.1: 12K Carbon fiber tow.
5.1
Dry powder Impregnation
The control of the tow spreading stage is the first step in order to achieve an even
distribution of the matrix around the fibers and thus an homogeneous fiber volume
content throughout the composite.
5.1.1
Tow spreading stage
A tow well spread to a fiber tape increases the fiber surface area exposed to the polymer powder and reduces the distance powder particles need to travel to reach the
center of the tape [11], [4]. Therefore, a well spread fiber tape with a high constant
width, w, is a condition for a good distribution of the matrix in the fibers and high
coating velocity [4].
In a fully spread tow, where each single filament is in contact and lies on the same
plane of the neighbor fibers to form a tape, the maximal theoretical tape width, wmax ,
for a carbon fiber tow with 12K filaments (K = 1000 filaments) can be calculated as
follow:
wmax = K · d = 120 000 · 7.1µm = 85.2mm
(5.1)
hmin = 7.1µm
(5.2)
67
5.1 Dry powder Impregnation
Form
Mechanical
Vibration
Fiber Spreading
Vacuum
Pneumatic
Compressed Air
Figure 5.2: Different methods for spreading fibers.
with d the diameter of the carbon fibers (7.1 µm) and hmin the minimum height
of the tape. An other important aspect correlated with the amount of fiber spreading is the flexibility of the tow after the impregnation and stabilization process: the
flexibility of the final impregnated tape increases with decreasing tape thickness [63].
Basically two different methods have been developed for spreading a multi-filament
bundle into tape: mechanical and pneumatic spreading, as shown in Fig. 5.2. There
are also several other more creative methods [56], like comb spreader [43] , [58], or
corona discharge spreader [47], [27], [26], [20], [37], [25], which will not be taken in
consideration in this review.
Mechanical spreading
Mechanical fiber spreading can be achieved drawing the fiber tow at sharp angles and
under tension over and under a series of pins [36] , [63], [64], [58], convex roller [44],
or curved smooth surfaces like sphere [17] or discs [4].
The tension (T) acting on the tow pulls each filament against the pins with a force
(F) normal to the pin surface and equal to:
F = T sin ψ
(5.3)
68
Chapter 5. Electrostatic Dry Powder Impregnation Process
Tow take off spool
Tow spreading
a)
To
T
F
t=0
θ
F ψ
r
T
r
F
b)
c)
Figure 5.3: Mechanical tow spreading with pins: a) typical set up with pins b) force
balance and c) cross section view [63].
where ψ is the tow wrapping angle and ϑ the contact angle, as shown in Fig. 5.3.
Since the filaments on the top of the tow seek the shortest route, the fibers in contact
with the pin are force to move laterally and consequently the tow is spread over the
pins or over the curved surface [4], as shown in Fig. 5.4.
To enhance the spreading the pins can be immersed in a gaseous medium and the
fibers can be additional spread from the acoustic vibration generated from a speaker
[21], [39], [46]. The tow can also be drawn and pressed through revolving rollers to a
flat squeezed spread shape.
Pneumatic spreading
The most frequently used method of pneumatic spreading consists of passing the
fiber tow through a Venturi slot tunnel [63], [22], [24], [54], [43], [53] , [52], [14],
[40]. The bundle enters the spreader at the inlet of a flat expansion section in which
the fibers are spread by the air entering from the outlet and traveling in the opposite
69
5.1 Dry powder Impregnation
t=0
t
F
Figure 5.4: Enhance mechanical spreading with curved surface.
Vacuum
Fiber Tow
Countercurrent
air flow
Figure 5.5: Schematic of the pneumatic spreading unit [63].
direction, as shown in Fig. 5.5. The air is forced by the holes of the side walls into a
vacuum manifold and the air drag is proportional to the pressure difference drawn in
the chamber.
Another method mentioned in the literature is to use compressed air in form of a
gas or air jet direct toward the fiber tow [60], [41], [61] or fed into a box in which
one or more exit ports in the box force the gas to move perpendicular to the fiber tow
[31], [34], [51], [57], [30]]. The fiber tow can also be spread by being drawn over a
suction cavity so that is bend and spread from the suction air; this spreading methods
is also known as the Fukui principle [55].
70
Chapter 5. Electrostatic Dry Powder Impregnation Process
5.1.2
Impregnation Stage
After being spread the fibers have to be coated with the polymer powder. A great variety of techniques have been developed for mingling carbon fibers with a dry polymer
powder. The different methods can be subdivided as follow:
1. Conventional Fluidized Bed Technology. The fiber tows are passed continuously over and under a series of pins in a bed of powder resin or powder resin
fluidized by the introduction of a gas stream [33], [3], [16], [12], [28], [32],
[49], [51]. However, little control over the fiber-matrix volume fraction could
be achieved in the impregnation step [16].
2. Electrostatic Fluidized bed Technology. The spread tows are pulled through
a fluidized suspension of charged polymer particles which adhere electrostatically to the fibers, as shown in Fig. 5.6 [23],[25], [29], [43], [50], [18]. The
resin powder is charged by the fluidizing gas, which in turn is charged passing
through a distributor plate connected to an electrostatic generator.
3. Acoustically Fluidized Bed Technology (Fig. 5.6). The spread tow is passed
through a fluidized bed of tribocharged polymer powder which is created using
both gas flow and a low frequency vibration produced by an acoustic source
[46], [45]. Tribocharged polymer particles adhered electrostatically to the fibers
and the rate of deposition is directly controlled by the amplitude of the acoustic
energy.
4. Recirculating Powder Deposition Technologies (Fig. 5.6). The spread tow is
drown in a recirculating powder deposition chamber, which consists of a powder feeder that showers particles from the top with and an exhaust fan which
recirculates the powder that accumulate at the bottom [42], [16], [14], [15].
5. Powder seeding (Fig. 5.7). The process simply consists of sprinkling the thermoplastic powder using an oscillating mesh or a powder extruder feeder or a
helix driven powder feeder [63], [11], [1]. A major drawback of this technique
is that only the upper surface of the spread tow comes into contact with the
powder. An alternative method consists to spread the fiber tow through an flux
of air so that the powder, which may enter through the same inlet as the air,
penetrates to the center of the bundle [48].
71
5.1 Dry powder Impregnation
Vacuum
Gas Vent
Fiber Tow
Fiber Tow
Fiber Tow
Gas Inlet
Porous Plate
Charging Plate
Diaphram
Recirculating Exhaust Fan
Dry Air Input
a)
b)
c)
Figure 5.6: Schematic drawing of a) electrostatic fluidized bed technology, b) acoustically fluidized bed technology and c) recirculating bad technology [16].
Technology
Conventional fluidized bed
Electrostatic fluidized bed
Acoustically fluidized bed
Recirculating fluidized bed
Electrostatic powder spray
Low costs
++
+
+
+
Powder size
+(50-200 µm)
+(50-200 µm)
-(≤50 µm)
+(50-200 µm)
++(1-250 µm)
Vf control
++
++
+
Mingling
++
++
++
Table 5.2: Advantages of different dry powder processing technologies.
6. Electrostatic Powder Spray Process. The spread tow is grounded and coated
with an air stream of electrostatically charged particles. The charged particles
air stream is formed by aspirating the polymer particles from a fluidization
chamber. The electrostatic field is supplied by means of an applied voltage [53]
or of a negative corona electrostatic spray gun [4], [17]. The powder particles
are able to coat both sides of the spread fiber tow.
Tab. 5.2 lists the advantages and disadvantages associated with each process.
72
Chapter 5. Electrostatic Dry Powder Impregnation Process
Powder Feed
Powder Feed
air
Fiber Tow
a)
Fiber Tow
b)
Figure 5.7: Schematic of the powder seeding processes with a) extruder powder feed
b) helix driven powder feed.
5.1.3
Stabilization Unit
They are basically two methods for fixing the powder on the fiber tow in order to
prevent loss of powder during the subsequent processing: to melt-fuse the powder
on the tow or to cover the towpreg with an extruded sheet of resin material. The
major advantage of filament bundle coating also known as FIT (Fibers Imprégnées
de Thermoplastique) is the control of the fiber-matrix volume fraction through the
extruded resin sheet [63], [11]. However, as the extruded sheath contain up to 70% of
the matrix fraction, consolidating this towpreg requires higher pressures under similar
processing conditions in comparison to a powder coated towpreg.
5.2
Electrostatic Powder Spray Process
A cost-effective, flexible and easy manageable electrostatic powder spray process was
selected within the scope of prepreg development for the production of continuous
carbon fiber PA6T/6I composites. Electrostatic powder coating is a process widely
used in the metal finishing industry for painting parts without using solvents. Advantages of powder coating are the use of a wide range of powder size and the recycling
of oversprayed powder [6], which reduces the costs for the resin powder material.
Moreover, recent technological developments have allowed leading equipment man-
73
5.2 Electrostatic Powder Spray Process
Dosing air
Conveying air
0-100 KV tension cable
Powder Feed
Silica gel drying column
High voltage converter
A.C. + D.C.
Powder-air mixture
Compressed fluidzation air
Vibrating table
Fiber Tow
A.C.
Point corona electrode
Powder feed
Powder cloud deflector
Electrode protecting air
Figure 5.8: Schematic representation of the electrostatic powder coating system and
process [9].
ufacturers to offer new equipment features able to maximize the transfer efficiency,
the coating effectiveness and uniformity [8]. Taking advantage of these technology
improvements, an ITWGema EASY 1-L manual powder coating system was used
and implemented in an electrostatic powder spray impregnation process. A schematic
representation of the electrostatic powder coating system and process is shown in
Fig. 5.8.
The polymer particles in the hopper are fluidized by air flowing through a porous
plate into the powder bed. The uncharged fluidized polymer particles are aspirated
into an air stream through the partial vacuum created from the conveying air and are
74
Chapter 5. Electrostatic Dry Powder Impregnation Process
fed to the spray gun through the air stream. The velocity of the air stream and the
powder flow rate can be controlled by controlling the pressure and therefore the velocity of the conveying air. The powder dosage rate can be reduced by the same air
stream velocity increasing the flow of dosing air. Increase in dosing air rate decrease
the partial vacuum created from the conveying air, reducing the rate of polymer particles aspiration. The polymer powder becomes electrostatic charged in the negative
corona charged gun. As shown in Fig. 5.8, the powder spray gun basically consists
of a pointed wire electrode at the nozzle, a tube made of a plastic insulator for transporting the powder and a stationary powder deflector attached to the nozzle to shape
the spray pattern [9]. High electric field can be produced in a small region near a
pointed electrode. When the field reaches a critical value characteristic of the gaseous
medium, a self-sustaining local electric breakdown will occur. This is known as the
corona effect. Corona charging arises from attachment of the ions produced in the
corona to the sprayed powder particles. The corona onset voltage is usually about
to 20 to 30 kV and negative corona charging guns are usually used, since negatively
charged powder particles tend to deposit more uniformly and efficiently than positively charged ones [19]. Sticky powders, powders which have been exposed to high
humidity and powder particles which acquire tribocharges of opposite polarity to the
corona will deposit onto the corona electrode in the course of spraying. Therefore
pressured air is blown around the electrode to prevent the sintering of powder on its
surface [19], [9].
5.2.1
Transport and Deposition on Carbon Fibers
The gun nozzle charges the powder particles, which are then directed towards the
grounded fibers by a force (Fnet ) under the influence of aerodynamic (Fa ), electrostatic (Fe ), and gravitational forces (Fg ).
Fnet = Fa + Fg + Fe
(5.4)
Optimizing air flow rate and electrostatic voltage, as well as achieving correct target distance and proper grounding, will produce the best powder coating uniformity.
Aerodynamics forces dominate other forces in the region close to the gun nuzzle but
the electrical forces dominated near the conductive substrate [2]. In other words, the
air flow from the powder coating gun is responsible to capture the powder particle
75
5.2 Electrostatic Powder Spray Process
and to convey them to the object, while the electrical forces control the deposition
process.
Aerodynamic Forces
As air flows are used to blow the powder from the gun toward the substrate, aerodynamic effects are extremely important and the trajectories followed by particles
depend on a balance between electrostatic and aerodynamic forces and are, in addition, particle-size dependent. The gun nozzle charges the powder particles, which are
then directed toward the grounded fibers. The aerodynamic force is responsible for
particle transport from the gun. This force acting on a single particle can be modeled
as flow past a sphere using the following equation [17], [4]:
Fa = CD · π ·
ρf d p 2
·
Vrel 2
2
4
(5.5)
where dp is the diameter of a particle, ρf is the density of the air, Vrel is the velocity
of a particle relative to the air flow, and CD is the drag coefficient which is related
to the Reynolds number (Rep ). The electrostatic gun can spray air from 1.8 Nm3 /h
to 8 Nm3 /h and has a nozzle diameter of 12 mm. The air velocity (Vair ) is then
expressed by:
Vair =
V̇
A
(5.6)
where V̇ is the air flow flowing through the nozzle area A. In Tab. 5.3 are shown
the air velocities which corresponds to a given air flow.
The drag coefficient is empirically found for Rep < 1000 to be [17]:
CD = (
24
) · (1.0 + 0.15 · Re0.687
)
p
Rep
(5.7)
ρf · Vrel · dp
µf
(5.8)
with
Rep =
where µf is the air viscosity.
76
Chapter 5. Electrostatic Dry Powder Impregnation Process
Air flow [Nm3 /h]
1.8
2
3
4
6
8
Air velocity [m/s]
4.4
4.9
7.4
9.8
14.7
19.7
Table 5.3: Air velocities in powder spray gun.
Gravitational Forces
The force of gravity on a solid particle in a fluid medium is given by:
Fg = π ·
d3p
· (ρp − ρf ) · g
6
(5.9)
where ρp and ρf are respectively the particle and air density, g is the earth gravitational acceleration. This force depends only from the diameter of the powder and
from the density difference between the particle and the air.
Electric forces
Electric field plays an important role in generating the corona for particle charging
and in driving the charged particles to the substrate. The electrical force acting on a
particle is the product of the electric field (Etot ), which the particle is subjected, and
the charge of the particle qp [17]:
Fe = qp · Etot
(5.10)
The electric field in the space between the electrode at the gun nozzle and the grounded
conductive substrate, i.e. in the transport region, arises from different sources: the applied voltage at the corona electrode (Eelectrode ), the space charge due to the cloud
of the charged powder particles and the ions generated form the corona (Espace ) and
the field resulting from the deposited layer of charge particles on the grounded fibers
77
5.2 Electrostatic Powder Spray Process
Electric field lines
E electrode
Flow field
E deposited
E space
Fiber tow
E image
Wrap-around effect
Figure 5.9: Schematic representation of the electrical field and electrical field lines
[2].
(Edeposited ):
Etot = Espace + Eelectrode + Eimage + Edeposited
(5.11)
The charged powder particles contribute 75% to 95% of the total space charge [19],
and the copious ions present in the transport region considerably enhance the space
charge field. The electric field in the transport region terminates on the grounded
substrate and ensures that the trajectories of the charged particles are directed toward
and onto the substrate (Fig. 5.9). The adhesion of the charged powder to the grounded
fiber tow is predominantly given by the image force [17]:
Fimage = qp · Eimage =
d2p
4π · 0 · L2
(5.12)
A negative charge in the vicinity of a conductor will induce a pair of positive and
negative charges in the conductor. If the conductor is grounded, the induced negative
charge will leak to the ground, leaving only the induced positive charge in the conductor. The coulombic attraction between the negative charge and the induced positive
78
Chapter 5. Electrostatic Dry Powder Impregnation Process
charge is the image force. Charged powder particles adhere to and form a coherent,
although friable, layer on the grounded conductive substrate but as the charges on the
powder gradually leak off, the adhesion is weakened.
5.3
Set up of the electrostatic powder spray impregnation process
A schematic of the electrostatic powder spray impregnation plan developed at the
Center of Structure Technologies for the impregnation of carbon fiber rovings with
the PA6T/6I oligomer powder is shown in Fig. 5.10. The process uses a brake fiber
tensioner to maintain low and constant tension in the fiber tow. The carbon fibers roving used in this work are not free of false twists, which can interfere with the spreading
process. In order to avoid these twists, the take-off mandrel was placed at a large distance from the fiber spreaders, as shown in Fig. 5.10. This provides an accumulation
of twist, which can be compensated due to their opposite direction. Two pneumatic
fiber spreaders spread the fibers before being coated by the negative corona electrostatic powder coating equipment in the coating chamber. The oversprayed powder in
the coating chamber is collected and recycled through a Dyson DC08 cyclone separator. The coated fibers are then passed through two infrared heaters to melt and
therefore to fix the powder on the fibers. The temperature set-up for the heaters is
a function of process speed, melting temperature and area weight of the prepolymer.
The final towpreg is then wound onto a take-up mandrel. The electrostatic spraying
equipment used throughout this study was an ITWGema EASY 1-L manual powder
coating system. It consists in an electrostatic hand spray gun and powder fluidization
reservoir. The system is also equipped with a control panel for adjusting powder delivery (0-290 g/min), total air flow (deliver air 5.9-26.1 m/s), electrode scavenging air
flow (0-6.2 Nm3 /h), corona charging voltage (0-100 kV), and corona charging current
(0-100 µA). The spray gun was held perpendicularly or at 45◦ angle and stationary
relative to the grounded substrate at any desired spraying distance between 0 mm and
70 mm by means of a special fixture.
5.4 Optimization of the electrostatic powder spray impregnation plant
79
Figure 5.10: Schematic of electrostatic spray impregnation plan used in this work.
5.4
Optimization of the electrostatic powder spray impregnation plant
The most important goals in the optimization of the electrostatic powder spray impregnation plant and processing parameters are summarized as follow:
• High process speed, i.e. tow velocity
• High control of the fiber-matrix volume fraction
80
Chapter 5. Electrostatic Dry Powder Impregnation Process
Powder diameter (µm)
≥560
≥250
≥100
≥71
≥50
≤50
Percent by weight
31%
13%
21%
11%
8%
16%
Table 5.4: PA6T/6I prepolymers particle size distribution.
• High mingling of the fiber with the powder matrix
In order to achieve these targets, the distinct process steps - spreading of the fiber
tow, fluidization of the polymer powder, and powder coating process - were first optimized separately and statically, as shown in the following paragraphs.
5.4.1
Materials
PA 6T/6I Oligomers
Since the PA 6T/6I precondensates (PA 6T/6I, XE 3733 VK) are sensitivity to moisture and cohesive force, therefore were dried in a vacuum oven at 80◦ C and 4 mbar
for at least 24 hours before further use. The PA6T/6I precondensates as received from
the polymerization reaction have a particle size distribution from few µm to several
mm, with more than 50% percent by weight of oligomers with a diameter size below
250 µm, as shown in Tab. 5.4.
Oligomers of sieved size range between 5-250 µm were used in the experiments.
Substrate
In order to obtain an initial knowledge of the effects of operating variables on the
distribution of the powder and coating uniformity, a metallic tape with electrical resistivity of 1 x 10−7 Ω · m, length of 300 mm and width of 12.7 mm was used.
According to the literature [13] it is well known that high temperature thermoplastic polymers exhibit a stronger interface bonding toward desized carbon fibers.
5.4 Optimization of the electrostatic powder spray impregnation plant
300 mm
z2
z3
81
z1
50 mm
z2
z1
12.7 mm
Figure 5.11: Photograph of the electrically-grounded substrate.
Properties
Filament diameter (µm)
Twist
Filament shape
Tow cross-sectional area (mm2 )
Resistivity ρ (Ω·m)
Electrical resistance, 12K (Ω/m)
Density (g/m3 )
Weight/length (g/m)
Hexcel AS4 PAN carbon fiber,12K
7.1
None
Round
0.48
1.53 × 10−5
35 × 102
1.79 × 106
0.858
Table 5.5: Carbon fibers properties.
Therefore, desized carbon fibers were used in continuous carbon fiber PA6T/6I composites. The properties of the desized carbon fibers and the carbon fiber tow, which
were used for composite production, are listed in Tab. 5.5.
5.4.2
Fiber Spreading
Unlike wet powder impregnation, where the fiber tow is spread by the powder particles which act as spacers during impregnation [64], in dry powder impregnation technologies spreading and impregnation are two separate process steps. As mentioned
82
Chapter 5. Electrostatic Dry Powder Impregnation Process
Electrostatic Spray Gun
Costant Spreading
Spread Fiber Tow
Fiber Tow Velocity
Spreading Width
Spraying distance
d
w
Totaly Spread Fiber tow
Figure 5.12: Dependency of the powder coating from the fiber spreading and tow
velocity.
before, a well spread tow to a fiber tape increases the fiber surface area exposed the
polymer powder (Fig. 5.12), i.e. the coating velocity and therefore the maximal permitted fiber tow velocity. High fiber tow velocity in preform impregnation processes
enables low manufacturing costs. Moreover, increasing the ratio between the spreading width, w, and the diameter of the gun nozzle, d, rises also the efficiency in the
deposition on the fibers [6]. More precisely, we define as deposition efficiency the
ratio between the weight of the deposited powder and the weight of the total sprayed
powder. When w/d is increased from 0.5 to 3, the deposition efficiency rises from
20% to 65%, which also represents the upper limit of the deposition efficiency. In our
set-up, where d=15 mm, to reach the limit efficiency value a fiber spreading width
higher than 45 mm is necessary.
The most important goals in developing a fiber spreader are summarized as follow:
• achieving very low fiber damage
• operating over a wide range of tow speeds
5.4 Optimization of the electrostatic powder spray impregnation plant
83
• obtaining constant amount of spreading
• obtaining high amount of spreading (w ≥ 45 mm)
In order to achieve these goals, different mechanical and pneumatic fiber spreaders
were developed, the different factors influencing the spreading process were analyzed
and the results compared.
Mechanical Spreaders
Three different types of mechanical fiber spreaders were developed and investigated
as explained in the following section: a geodesic fiber spreader, an edge fiber spreader
and a vibration fiber spreader. The spreading width by the different set-up and velocity was measured optically and averaged over at least 10 measurements by the same
set-up and velocity of a spread tow in which all the filaments lie adjacent.
Geodesic fiber spreader. Several factors affect the final spreading width in geodesic
fiber spreaders, including fiber tension, angle of the curved spreading surface, contact
angle and tow running speed. In order to investigate all these parameters, a special
set-up provided with several, different interchangeable geodesic forms or roller was
developed, as shown in Fig. 5.13. The set-up forces the fiber bundle to run along the
roller elements or surfaces with a determinate contact angle ϑ, which could be varied before each experiment. The geodesic pins were fastened to the support in two
configurations: free to rotate or fixed.
Contact angles at about 45◦ or less with a fixed radius of curvature were found
to increase the fiber spreading width (w). The smaller the curvature radius of the
geodesic surface, the larger is the width of fiber spreading and a spreading width up
to 25 mm could be achieved, as shown in Fig. 5.13. However, decreasing the curvature radius increases the tendency of the fiber tow to slide to one side of the spherical
surfaces or to split into two fiber bundles. The stability of the spreading process
is therefore compromised and this tendency rises with increasing fiber tow velocity.
Moreover, the false twist in the fiber tow just before the compensation phase causes
a decrease in the spreading width, which also increases the problem of the instability.
A stable spreading could be achieved with a set-up which provides spreading width
not higher then 10 mm with 18% variation by a fiber tow velocity between 0.8 and
2 m/min and up to 25% variation by higher velocity. The geodetic fiber spreader uses
84
Chapter 5. Electrostatic Dry Powder Impregnation Process
MM
Figure 5.13: Geodesic fiber spreader.
physical contact to spread the fiber bundle, and consequently causes abrasion damage
and breakage of the small brittle un-sized carbon fibers. The friction can be reduced
if the pins are free to rotate with a resulting decrease in spreading width. The considerable advantages of geodesic fiber spreaders lie in the low costs for the development
and realization of the apparatus, in the absence of maintenance in service and of external energy sources.
Edge Fiber Spreader. In this type of mechanical fiber spreader the fiber tow was
5.4 Optimization of the electrostatic powder spray impregnation plant
85
m
2m
m
4m
m
2m
m
4m
Figure 5.14: Edge fiber spreader.
drown above more sharp angles with different radius of curvature and different contact
angles. In case of edges with small radius of curvature, the wrapping angle is small,
which reduces the negative effect of friction. Moreover, the contact angle can also
be reduced to a smaller value, which together with high tension increases the pulling
force F and therefore the spreading width. On the other hand, the increased shearing
force acting on the fiber causes fiber damage, due to the low mechanical properties
of the carbon fiber in the direction perpendicular to the fiber length. A compromise
between spreading width and fiber damage was found with the experimental set-up
shown in Fig. 5.14. Maximal spreading width up to 13 ± 15% mm was achieved
at 6 m/min. At higher velocity the mean value decreases to 7 mm with a variation
of 25% while the fiber damage increases. The advantage of the edge spreader compared to the geodetic spreader is the stability of the process, the higher spreading at
higher velocity and the higher constancy of the spreading. Edge fiber spreader setup is very compact and easy to build and do not need maintenance or external sources.
Vibration Spreader. In this prototype of fiber spreader the fiber tow is drown over
a pin fixed to a 18 N shaker. The pin was moved perpendicularly to the fibers by the
vibrating diaphragm of the shaker, which oscillates with a predetermined amplitude
and frequency between 100 Hz and 50 KHz. Different surface finishes of the pin
were tested in order to increase or reduce the surface roughness and therefore to find
a compromise between high spreading widths and low fiber damage. The advantage
86
Chapter 5. Electrostatic Dry Powder Impregnation Process
Vacuum
4°
L1
L2
500
Figure 5.15: Vacuum Spreader.
of spreading the fibers with vibration sources lies in the high control of the process
and therefore in the high control of the spreading width and for this reason this kind
of fiber spreader has found application in industrial plan. However, throughout the
frequency range fiber spreading widths not higher then 5 mm were achieved.
Pneumatic Spreaders
Three different types of pneumatic fiber spreaders were developed and investigated as
explained in the following section: two vacuum fiber spreaders and a compressed air
spreader. Contrary to mechanical fiber spreading process, it is important to maintain
very low tension in the tow in order to achieve high spreading of the fiber bundle.
For this reason the servo-based tension control system, which provides a tension not
lower than 0.5 kg, was replaced with a feed spool brake system, where the tension
could be controlled between 0.1 and 0.5 kg.
Vacuum Spreader. Two simple experimental devices were developed in order to
spread the fiber using vacuum as energy sources:
1. a prototype venturi slot tunnel with constant height (3 mm), variable inlet (L1 )
and outlet length (L2 ) and with a transparent top cover for a qualitative optical
5.4 Optimization of the electrostatic powder spray impregnation plant
87
observation, as shown in Fig. 5.15.
2. a pin with smooth surface and with a slot through which the air is drawn by a
vacuum pump. The length (Ls ) and the width (ws ) of the slot were also varied
in order to find the optimal set-up.
The design of a venturi slot tunnel was found not to be a simple task: high countercurrent air velocity increase the drag force which is responsible of the fiber spreading
but on the same time increases the turbulent behavior which causes fiber damage and
entanglement of the filaments. The velocity of the countercurrent is proportional to
the pressure drop in the tunnel and is related to the cross-sectional area of the inlet and
outlet; therefore the height and the length of the inlet and outlet are design parameters
to be optimized. However, an optimized geometry is related to a fiber tow with a determinate number of fibers in the bundle. Changing the number of filaments in the tow
change also the optimal design parameters of the spreading chamber; this impairs the
flexibility of one venture slot tunnel system in using different type of rovings. Moreover, the bundle has to be spread without gaps between the fibers or without being
split into two bundles, as occurred in our experiments. Further, a Venturi slot tunnel
apparatus is not compact since it needs a tunnel length of about 450 mm.
The vacuum spreading system with the variable length and width of the slot was simpler to developed and optimize, however the achieved maximal fiber spreading width
was of 8 ± 4% mm by 0.8 m/min with ws =60 mm and Ls =6 mm. Moreover, the
length of the slot was found to have a higher influence on the maximal spreading
width than the slot width. This is also a limiting factor for carbon fiber roving with
higher number of filaments, i.e. with higher achievable maximal spreading width.
However, this system was found to be an additional adjuvant step to a spreading stage
for its ability to fix the spreading while centering the fiber tow. In general, the advantage of using vacuum for spreading the fiber filaments was found to lie in the very
low fiber damage and in the stability of the process. For a possible industrial application should be taken into consideration that damaged fiber filaments are incline to be
drawn in the vacuum system causing plugging of the tubes or damage to the vacuum
pump.
Compressed Air Spreader. The compressed air spreader developed in this work is
an apparatus based on the principle that a collimated fiber tow can be spread into its
88
Chapter 5. Electrostatic Dry Powder Impregnation Process
individual filaments through one or more high energy air flow directed toward the fiber
bundle. In this fiber spreader the fiber tow is drawn under two free to rotate rolls and
over a cylindrical cage between the two rolls, where the spreading of the fibers take
place. A schematic representation of the compressed air spreader is shown in Fig. 5.16
as well as a detailed drawing of the cylindrical cage. Two compressed air injectors
are directed toward the fiber tow on the free to rotate cylindrical cage. The spreading
is fixed by the eight pins of the cage while the air flow is free to escape through the
gaps. By varying the position of the two rollers it is possible to set different contact
angles of the fiber tow with the cylindrical cage. The contact angle of the bundle
with the gage, the number, distance and angle of the injectors and the pressure of
the blowing air were adjusted to find an ideal condition for the spreading of the fiber
tow. Moreover, one injector was clamped on an arm connected to an electrical motor,
whose amplitude and oscillation frequency was given by a microcontroller. Due to
the movement of the arm the injector could oscillate perpendicularly to the running
direction of the tow.
In general, increasing the pressure of the blowing air increases the spreading width
and concurrently the fiber damage, but the swinging of the blowing air on the fiber
filaments was found to greatly improve the spreading width by reduced fiber damage.
The spreading width was found also to increase by constant pressure adding blowing air injectors; however, more than two air injectors blowing toward the spreading
cylindrical cage were found to create turbulent air behavior, which leads to entanglement of the fiber filaments and inconstancy in the spreading width. The optimized
solution with two separate spreading stage and only two spreading injectors directed
toward the cylindrical cage is shown in Fig. 5.16 . The optimized set-up was found to
provide spreading width up to 30 ± 5 mm by fiber tow velocity of 6 m/min. For constant set-up and process parameter, a dependency exists between the constancy of the
fiber spreading width and the constancy of the fiber-matrix fraction in the final towpreg. The advantage of this experimental set-up consists in the possibility to measure
the fiber spreading width through an on line optical device and to consequently adjust
the pressure value in order to control the fiber content of the final tape. In case of a
negative deviation from the nominal value of the fiber spreading width, an increase
in the applied air pressure could be quickly provided to increase the spreading degree.
5.4 Optimization of the electrostatic powder spray impregnation plant
Compressed air
89
Oscillating compressed air
Compressed air
Oscillating compressed air
Figure 5.16: Schematic representation of the compressed air spreader and detailed
drawing of the cylindrical cage.
90
Chapter 5. Electrostatic Dry Powder Impregnation Process
Corona Voltage [KV]
Corona Current [mA]
Fluidization Pressure
Air flow velocity [m/s]
Powder size [µm]
Powder amount pro test [g]
0
0
6 bar if not specified
4.4
100 < d < 250 if not specified
250
Table 5.6: Standard operating conditions during spraying optimization.
5.4.3
Powder Fluidization
The fluidization behavior of powder fluidized by gases can be classified in bubbling
(aggregative, heterogeneous) and non-bubbling (particulate, homogeneous) fluidization and depends by the same experimental parameters from the density, but also from
the particle size of the powder [5]. A given powder system with a given particle size
can exhibits both behavior depending on the fluid velocity. Therefore, the pressure
applied to the fluidization hopper influence the fluidization and, as a consequence,
the powder mass flow, as shown in Fig. 5.17. The graph shows the total amount of
sprayed powder in function of time for three different fluidization pressures. The
standard operating conditions during spraying are listed in Tab. 5.6.
The fluidizing pressure in the powder fluidization vessel has an effect on the consistency of the powder flow and therefore on the total amount of sprayed powder, as
shown in Tab. 5.7. However, as the amount of powder in the fluidization vessel reduced, the size distribution of the powder in the hopper shifts to lower values, which
are more difficult to fluidize, resulting in lower powder flow with time, as shown from
the inclination of the linear fit curves and from the high standard deviation value.
The fluidization difficulty arises from the cohesive behavior of small powder:
the interparticle forces increase reducing the size of the particles and when they are
greater than the forces exerted from the fluid, the powder particles tend to agglomerate. In this case fluidization can generally be possible or improved by the use of
mechanical vibration, which brakes up the agglomerates. The fluidization hopper
was fixed over an oscillating sander, whose power was controlled over a dimmer, as
shown in Fig. 5.18. The maximal power used was 20% of the vibration energy of the
system. In all experiments the fluidization air was dried in a silica gel column in order
5.4 Optimization of the electrostatic powder spray impregnation plant
Amount of sprayed powder, g
14
91
3 bar
6 bar
7 bar
Linear Fit of 3 bar
Linear Fit of 6 bar
Linear Fit of 7 bar
12
10
8
6
4
2
0
0
10
20
30
40
50
60
Time, s
Figure 5.17: Amount of sprayed powder in function of time and for different fluidization pressure.
Fluidization Pressure
3 bar
6 bar
7 bar
¨
Mean Value
0.195
0.48667
0.665
Standard Deviation
± 0.29596 (152%)
± 0.27632 (57%)
± 0.58759 (88%)
Table 5.7: Mean values and standard deviations of the amount of sprayed powder after 5 seconds and for different fluidization pressure. Each mean value was calculated
over 12 consecutive measurements.
to reduce air humidity, which also causes powder agglomeration. As clearly shown
from the figure, the vibration has a much higher positive effect on the fluidization
and on the powder mass flow than the fluidization pressure. Increasing the vibration
energy, increase the total amount of sprayed powder and decrease the tendency of
shifting toward lower powder mass flow with time, as shown from the higher inclination of the linear fit curves with increasing vibration energy and from the decrease in
the standard deviation values shown in Tab. 5.8.
92
Chapter 5. Electrostatic Dry Powder Impregnation Process
60
0%
15%
18%
20%
Linear Fit 0%
Linear Fit 15%
Linear Fit 18%
Linear Fit 20%
Amount of sprayed powder, g
50
40
30
20
10
0
0
10
20
30
40
50
60
Time, s
Figure 5.18: Amount of sprayed powder in function of time with the same fluidization pressure and for different vibration energy.
Vibration Energy
7 bar and 0%
15%
18%
20%
¨
Mean Value
0.665
3.615
3.88667
4.65667
Standard Deviation
± 0.58759 (88%)
± 2.48259 (69%)
± 2.00969 (52%)
± 1.15766 (25%)
Table 5.8: Mean values and standard deviations of the amount of sprayed powder
after 5 seconds and for different vibration energy. Each mean value was calculated
over 10 consecutive measurements.
5.4 Optimization of the electrostatic powder spray impregnation plant
Variable
7 bar
20% horizontal vibration
0.41 mm vertical vibration
0.94 mm vertical vibration
¨
Mean Value
0.665
4.65667
12.979
14.139
93
Standard Deviation
± 0.58759 (88%)
± 1.15766 (25%)
± 0.92725 (7%)
± 0.37946 (3%)
Table 5.9: Mean values and standard deviations of the amount of sprayed powder
after 5 seconds for highest investigated fluidization pressure, for the highest investigated horizontal vibration energy and for tow different vertical amplitude. Each mean
value was calculated over 10 consecutive measurements.
An oscillating sander functions like a horizontal shaker with variable frequency,
which sets the powder in the hopper to a horizontal movement. In contrast, throwaction shakers, also known as vibratory sieve shakers, force the powder particles to
a uniformly distributed 3-dimensional movement in the hopper. An electromagnetic
drive sets a spring-mass system in motion and transfers the oscillations to the sieve
stack or in our case to the fluidization hopper [62]. Moreover, throw-action shakers are characterized by the fact that their mechanical parameters, such as amplitude and frequency, can be set with high precision. In order to test the effect of the
3-dimensional movement of the powder particles in the hopper on the fluidization
behavior and therefore on the powder mass flow the hopper was fixed on a Retsch
analytical sieve shakers AS 200. The powder mass flow was investigated for different
amplitudes in the range from 0.2 mm and 1.6 mm. The effects of the 3-dimensional
movement for two different amplitudes (0.41 and 0.94) on the amount of sprayed
powder in dependence of time are shown in Fig. 5.19. In the same graph are plotted
as comparison the powder mass flows in function of time for the highest investigated
fluidization pressure and for the highest investigated horizontal vibration energy. The
added vertical movement to the powder particles has the most effective positive influence on the powder fluidization and therefore on the amount and constancy with
time of the sprayed powder, as shown in Tab. 5.9. It is also evident form the higher
inclinations of the linear fit curves for the two amplitude values in the Graph 5.19.
However, in order to better represent the constancy of the amount of the sprayed
powder with time and for the different fluidization variables, the concept of powder
94
Amount of spryed powder, g
Chapter 5. Electrostatic Dry Powder Impregnation Process
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0
7bar
20%
0.41mm
0.94mm
Linear Fit 7bar
Linear Fit 20%
Linear Fit 0.41mm
Linear Fit 0.94mm
0
10
20
30
40
50
Time, s
Figure 5.19: Amount of sprayed powder in function of time with the same fluidization pressure and for highest investigated fluidization pressure, for the highest investigated horizontal vibration energy and for tow different vertical amplitudes.
flow index was introduced and plotted in function of time and for the different fluidization variables, as shown in Graph 5.20. The powder mass flow index is defined
as the ratio between the amount of sprayed powder during five seconds to the mean
value calculated over 10 consecutive measurements. From Graph 5.20 is evident that
the introduction of a vibration shaker, which set the powder in the hopper in a 3dimensional movement, guarantees the constancy of the powder mass flow with time
and therefore throughout the impregnation process. This is the first prerequisite to
guarantee the constancy and uniformity of the powder coating process on the carbon
fibers and consequently the constancy of the resin fiber fraction in the final towpregs.
As mentioned before, decreasing the particle size increases the cohesive forces
and makes the powder fluidization more difficult, as shown in Fig. 5.21. A reduction
of the constancy of the powder mass flow with time is evident only with particles size
below 50µm, as shown in Fig. 5.22. However, the standard deviation throughout the
measurements is about 9% (Tab. 5.10), which indicates good fluidization behavior of
the powder. Therefore, mounting the fluidization hopper on a throw-action shaker is
possible to fluidized powder size ranging form few µm to 250 µm. The possibility
5.4 Optimization of the electrostatic powder spray impregnation plant
Powder mass flow index
3.0
95
7bar
20%
0.41mm
0.94mm
Polinomial Fit 7bar
Linear Fit 20%
Linear Fit 0.41mm
Linear Fit 0.94mm
2.5
2.0
1.5
1.0
0.5
0.0
0
10
20
30
40
50
Time, s
Figure 5.20: Powder flow index for highest investigated fluidization pressure, for the
highest investigated horizontal vibration energy and for two different vertical amplitudes.
Powder size [µm]
100 < d < 250
50 < d < 100
d < 50
Mean Value
14.139
13.948
9.41
Standard Deviation
± 0.37946 (3%)
± 1.22085 (9%)
± 0.89315 (9%)
Table 5.10: Mean values and standard deviations of the amount of sprayed powder
after 5 seconds for different powder sizes.
of using a wide range of powder size guarantee low costs associated with the impregnation of carbon fiber tow with PA6T/6I prepolymers, since the prepolymers emerge
from the polymerization reaction in powder form and no expensive cryogenic milling
is necessary.
96
Chapter 5. Electrostatic Dry Powder Impregnation Process
160
100<d<250µm
50<d<100µm
d<50µm
Linear Fit 100<d<250µm
Linear Fit 50<d<100µm
Linear Fit d<50µm
Amount of spryed powdere, g
140
120
100
80
60
40
20
0
0
10
20
30
40
50
Time, s
Figure 5.21: Amount of sprayed powder in function of time different powder sizes.
100<d<250µm
d<50µm
Linear Fit 100<d<250µm
Linear Fitd<50µm
Powder mass flow index
1.4
1.2
1.0
0.8
0.6
0
10
20
30
40
50
Time, s
Figure 5.22: Powder flow index in fuction of different powder sizes.
5.4 Optimization of the electrostatic powder spray impregnation plant
Gun Air Flow
Corona Current
97
45°
Secondary Air Flow
Charging Voltage
d
Conductive substrate
Powder Fluidization
Particle Size
Figure 5.23: Schematic representation of the coating chamber and of the several
operating variables affecting the electrostatic deposition of oligomers powder.
5.4.4
Powder Coating
The effects of various spraying variables on the electrostatic deposition of PA6T/6I
prepolymer powder were investigated in order to optimize the coating uniformity, the
upper and lower area coverage, and the deposition efficiency. The effect of the various operating parameters was investigated by spraying onto a stationary preweighted
metal tape for 10 s by changing one variable at a time. Powder was sprayed after
the gun was placed on its holding fixture at different distances from the grounded
substrate. The total amount of powder deposited on the upper surface, on the lower
surface and on the three different areas of the grounded substrate was weighted with
an accuracy of 0.01g. Each test consisted of six to ten runs, but just the resulting average and range values are reported. The use of metal tapes permits to investigate the
various spraying variables without the influence of the spreading process. Moreover,
it is also possible to characterize the coating process and deposition uniformity on the
upper and lower surface with higher precision.
98
Chapter 5. Electrostatic Dry Powder Impregnation Process
Corona Voltage [KV]
Corona Current [mA]
Fluidization Pressure
Fluidization Vibration
Air flow velocity [m/s]
Powder size [µm]
Powder amount pro test [g]
100
50
4.5 bar if not specified
20% horizontal vibration
4.4
100 < d < 250 if not specified
250
Table 5.11: Standard operating conditions during powder coating optimization.
Charging Voltage
The amount of powder deposited onto the grounded substrate and respectively on
the upper and lower surface is shown as a function of the applied corona voltage in
Fig. 5.24. The powder was fluidized by applying to the hopper a horizontal vibration with 20% of the maximal system vibration energy. The distance of the nozzle
from the substrate was set at 10 mm and the corona voltage was varied from -40 kV
to -100 kV. The Fig. 5.24 shows clearly that the powder deposition increases with
increasing corona charging voltage. Charging voltage influences the powder deposition efficiency since it affects the balance between the electrostatic and hydrodynamic
forces at a determinate spraying distance [7]. With an increase in charging voltage the
electric field between spraying gun and substrate surface increases. Since this field is
responsible for delivering the charged particles to the surface, an increase in electric
field usually increases deposition efficiency [6].
Corona Current
The amount of powder deposited onto the grounded substrate and respectively on
the upper and lower surface is shown as a function of the applied corona current in
Fig. 5.25. The powder was fluidized by applying to the hopper a horizontal vibration
with 20% of the maximal system vibration energy. The distance of the nozzle form the
substrate was set to 10 mm and the corona current was varied form 20 µA to 100 µA.
The Fig. 5.25 shows clearly that the powder deposition increases with decreasing
corona charging current.
The corona current generated from the spraying gun flows through the powder
5.4 Optimization of the electrostatic powder spray impregnation plant
99
0.55
Total deposited powder
Upper surface
Lower surface
Weight of deposited powder, g
0.50
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
40
50
60
70
80
90
100
Corona charging voltage, kV
Figure 5.24: Total amount of powder electrostatically deposited onto the grounded
substrate after a spraying time of 10 s as a function of the applied corona voltage;
the amount of deposited powder respectively on the upper and lower surface is also
represented.
layer and generates, if the prevalent conduction mechanism is ohmic, an ohmic field.
This electric field is directly proportional to the corona current and can interfere with
the space-charge field [19].
Powder Mass flow and Secondary Mass Flow
The air flow field in the electrostatic powder coating system is the resultant of two
separate air flow: one due to the powder feed system via the powder spray gun, and
the other a secondary flow due to the powder recovery system. The Fig. 5.26 shows
the fraction in percentage of the total deposited powder on three different areas of the
upper substrate surface as a function of the applied air flow velocity. Z3 represents
the substrate surface just below the spraying gun and Z1 the surface with the greatest distance from the gun nozzle while the surface Z2 lies in between, as shown in
chap. 5.4.1. The air flow velocity was varied from 1.4 to 9.8 m/s.
As shown in Fig. 5.26, increasing the air flow velocity the amount of powder
100
Chapter 5. Electrostatic Dry Powder Impregnation Process
1.0
Total deposited powder
Upper surface
Lower lower
Weight of deposited powder, g
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
20
40
60
80
100
Corona current, µA
Figure 5.25: Total amount of powder electrostatically deposited onto the grounded
substrate after a spraying time of 10 s as a function of the applied corona current;
the amount of deposited powder respectively on the upper and lower surface is also
represented.
deposited on the surface below the gun nozzle decreases, while on the boundary region increases. The air stream delivers particles to the part while the electric force
draws the particles towards the grounded substrate. As the powder mass flow rate is
increased, the air and powder exit velocities increase. High powder velocities overcome the electrostatic image force near the conductive substrate allowing the attraction force between the charged particle and grounded metal surface to be established
[17]. However, if the field strength is not sufficient or the velocities are too high,
the particle will be carried away by the air stream. In order to improve the balance
between the aerodynamic and electrostatic forces the powder spray gun was inclined
of 45◦ respect to the direction of the fiber. The secondary flow due to the powder
recovery system and the geometry of the impregnation chamber were also found to
have an influence on the deposition efficiency of the charged powder, as shown in
Fig. 5.27. The secondary air flow was kept parallel to the spraying direction and the
power of the cyclone separator was varied between 0% and 100% of the 1400 kW.
As expected, increasing the power of the cyclone separator i.e. the secondary flow air
5.4 Optimization of the electrostatic powder spray impregnation plant
101
Distribution of deposited powder, %
120
Z3
Z2
Z1
100
80
60
40
20
0
1
2
3
4
5
6
7
8
9
10
11
Air flow velocity, m/s
Figure 5.26: Distribution of the powder along the conductive substrate as a function
of the air flow velocity.
velocity decreases the powder deposited on the substrate, since the charged particles
are entrained from the air flow to the cyclone separator. In order to maximize the
deposition efficiency the powder of the recovery system should be kept at a very low
value in order not to modify significantly the air flow of the spry gun and thus not to
affect the particle trajectory.
One of the advantages of industrial powder coatings is the efficient utilization of
powder materials, accomplished by recycling oversprayed powder. In order to achieve
maximum economy in powder coating, the powder spray booth should ensure that the
oversprayed powder is transferred and collected efficiently and recycled back into the
feed system. In Tab. 5.12 the value of recovered powder in function of different
particle size are shown. The powder was weighted before being fluidized and sprayed
in the impregnation chamber without conductive substrate. The recovered powder
from the cyclone separator was finally weighted and compared with the value in the
fluidization hoper, as shown in Tab. 5.12. With the coating set-up developed in this
work it is possible to recover between 86 and 97% of the oversprayed powder, which
reduces the costs in tow-preg manufacturing.
102
Chapter 5. Electrostatic Dry Powder Impregnation Process
Figure 5.27: Total amount of powder electrostatically deposited onto the grounded
substrate after a spraying time of 10 s as a function of the power of the recovery
system and of the impregnation chamber geometry.
5.4 Optimization of the electrostatic powder spray impregnation plant
Powder size [µm]
d > 560
250 < d < 560
100 < d < 250
70 < d < 100
50 < d < 70
d < 50
Powder hoper [g]
140.9
75.1
67.3
29.8
33.2
93.6
103
Powder recovered [g]
137.2 (97.4%)
72.4 (96.4%)
62.2 (92.4)%)
28.7 (96.3%)
28.8 (86.%7)
80.7 (86.2%)
Table 5.12: Recovery of oversprayed powder by the cyclone separator.
Corona Voltage [KV]
Corona Current [mA]
Fluidization Pressure [bar]
Fluidization Vibration
Air flow velocity [m/s]
Powder size [µm]
Powder amount pro test [g]
100
50
4.5
Throw-action shakers
4.4
100 < d < 250
250
Table 5.13: Standard operating parameters during powder coating.
Spray Distance
The effect of spraying distance on deposition efficiency at corona charging voltage
of 100 kV is shown in Fig. 5.28. The total amount of powder deposited onto the
grounded substrate and the fraction of powder deposited on the upper and lower surface of the substrate is shown in Fig. 5.28 as a function of the distance. The distance
of the spraying nozzle form the substrate was varied between 20 and 50 mm and the
spray gun was held with an angle of 45◦ respect to the fiber direction. The substrate
width was 17 mm in order to reproduce a realistic spread fiber tow. The operating
parameters are listed in Tab. 5.13.
As the spray distance increases, both the particle charge and the electric field
will decrease [6]. Moreover, as the spray distance increases, increases also the time
span in which the particles are exposed to the influence of air flow and self-repulsion
due to their similar charges. Thus the deposition efficiency is expected to decrease
104
Chapter 5. Electrostatic Dry Powder Impregnation Process
1.6
Total deposited powder
Upper surface
Lower surface
Weight of deposited powder, g
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
20
25
30
35
40
45
50
Distance, mm
Figure 5.28: Total amount of powder electrostatically deposited onto the grounded
substrate as a function of the distance of the gun nuzzle from the substrate; the deposition on the upper and on the lower surface of the substrate is also plotted in the
graph.
with increasing of the spray distance. However, as shown in Fig. 5.28 the deposition
efficiency decreases reducing the spraying distance until it reach a plateau between 40
and 20 mm. This is probably due to the drop of voltage on the corona (safety device)
as the spray gun nears the conductive substrate in order to avoid sparking [6]. By
40 mm the mean value of deposited powder is 0.77 g and presents the lowest standard
deviation (± 8.8% between six measurements), which is an important parameter for
the control of the fiber matrix fraction.
Powder Size
The total amount of powder deposited onto the grounded substrate and the fraction of
powder deposited on the upper and lower surface of the substrate is shown in Fig. 5.29
as a function of the powder size. The distance of the spraying nozzle from the substrate was fixed at 40 mm and the spray gun was held with an angle of 45◦ respect
to the fiber direction. The substrate width was 17 mm and the coating parameters
5.4 Optimization of the electrostatic powder spray impregnation plant
105
2.0
Total deposited powder
Upper surface
Lower surface
Weight of deposited powder, g
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
25
50
75
100
125
150
175
200
225
250
275
Powder size, µm
Figure 5.29: Total amount of powder electrostatically deposited onto the grounded
substrate as a function of the powder size; the deposition on the upper and on the
lower surface of the substrate is also plotted in the graph.
Powder size [µm]
100 < d < 250
70 < d < 100
50 < d < 70
d < 50
Mean Value [g]
1.435
1.25667
1.00833
1.07
Standard Deviation
0.25058 (17%)
0.15253 (12%)
0.14932 (14%7)
0.08075 (7%)
Table 5.14: Amount of powder electrostatically deposited onto the grounded substrate as a function of the powder size.
optimized in the previous experiments were set.
The weight of deposited powder tends to increase with increasing powder size
but also increases the standard deviation between the six different measurements, as
shown in Tab. 5.14.
Moreover, the experiments showed that with increasing powder size the deposition of the powder on the metal substrate was not uniformly distributed. This is
probably related with the increase in the aerodynamic force acting on the powder,
106
Chapter 5. Electrostatic Dry Powder Impregnation Process
which is direct proportional to the diameter of the particles.
5.5
Conclusions
A cost-effective, flexible and easy manageable electrostatic powder spray impregnation process was developed and optimized within the scope of prepreg development
for the production of continuous carbon fiber PA6T/6I composites. In order to achieve
high mingling of the fiber with the powder matrix a fiber spreader able to obtain high
amount of spreading (w > 52 mm) with very low fiber damage and operating over
a wide range of tow speeds was developed. A well spread tow to a fiber tape increases the fiber surface area exposed the polymer powder, i.e. the coating velocity
and therefore the fiber tow velocity. High fiber tow velocity in preform impregnation
processes enables low manufacturing costs. In addition, another potential advantage
of the pneumatic fiber spreader developed in this work consists in the possibility to
measure the fiber spreading width by an on-line optical device and to consequently
adjust the pressure value in order to control the spreading width of the fiber tape. The
constant width of the fiber spreading is very important for the control of the fiber matrix ratio in the final tow-preg. In case of upgrading the powder coating impregnation
process to an industrial production obtaining high quality in the final prepreg is crucial. The first step to guarantee the constancy and uniformity of the powder coating
process on the carbon fibers and consequently the constancy of the resin fiber fraction
in the final tow-pregs is the control of the powder mass flow through the fluidization
process. The introduction of a vibration shaker, which sets the powder in the hopper
in a 3-dimensional movement guarantees the constancy of the powder mass flow with
time and therefore throughout the impregnation process. Moreover, the throw-action
shaker was found to fluidize powder sizes ranging form few µm to 250 µm. The possibility of using a wide range of powder sizes guarantees low costs associated with the
impregnation of carbon fiber tow with PA6T/6I prepolymers, since the prepolymers
emerge from the polymerization reaction in powder form and no expensive cryogenic
milling is necessary. The spraying variables were optimized in order to achieve high
and constant deposition efficiency. Since the amount of deposited powder depends
also on the width and constancy of the fiber spreading, the influence of the spraying variables were tested statically on a grounded metal tape. High charging voltage,
low corona current, low gun and low recovery system air velocity result in higher,
5.5 Conclusions
107
constant and uniform deposition efficiency. The form of the coating chamber was
also optimized in order to obtain uniform deposition on the lower substrate surface.
The optimal gun-to-substrate distance with a standard deviation of the deposition of
only 8.8% was found around 40 mm. However, with larger particle size the standard
deviation was found to increase. Finally of the advantages of the powder coating
impregnation process is the use of a wide range of powder size accomplished by recycling oversprayed powder, which reduces the costs for the resin powder material.
In fact with the coating set-up developed in this work it is possible to recover between
86 and 97% of the oversprayed powder.
108
Chapter 5. Electrostatic Dry Powder Impregnation Process
Bibliography
[1] R.E. Allred, S.P. Wesson, D.A. Babow, Dry Powder Impregnaiton Studies
for High Temperature Towpregs, Sampe Journal, Vol. 40, No. 6, November/December 2004, 40-48.
[2] M.L. Ang, P.J. Lloyd, Investigation of Charged Particle Trajectories in Electrostatic Powder Coating Systems, Int. J. Multiphase Flow, Vol.13. No. 6, 1987,
823-836.
[3] N. Chand, S. Kreuzberger, G. Hinrichsen, Influence of Processing Conditions
on the Tensile Properties of Unidirectional UHMWPE Fiber/LDPE Composites
Composites, Vol. 25, No. 9, 1994, 878-880.
[4] M. Duval, K. Ramani, M. Bays, F. Caillat, In-Situ Composite Manufacture Using an Electrostatic Powder Spray Process and Filament Winding, Polymers and
Polymer Composites, Vol.4, No. 5, 1996, 325-334.
[5] D. Geldart, Types of Gas Fluidization, Powder Technology, 7, 1973, 285-292.
[6] A. Golovoy, Deposition Efficiency in Electrostatic Spraying of Powder Coating,
Journal of Paint Technology, Vol. 45, No. 580 1973, 42-48.
[7] A. Golovoy, Particle Deposition in Electrostatic Spraying Of Powder Coating,
Journal of Paint Technology, Vol. 45, No. 585 1973, 68-73.
[8] S. Guskov, Application of Powder Coatings: Old Problems, New Findings,
New Developments, Powder Coating ’96 Conference in Indianapolis, Indiana,
September 18, 1996.
110
BIBLIOGRAPHY
[9] http://www.itwgema.com/index.html
[10] F.V. Lacrox, H.Q. Lu, K. Schulte, Wet Powder Impregnation for Polyethylene
Composites: Preparation and Mechanical Propeties, Composites: Part A 30,
1999, 369-373.
[11] A. Miller, A.G. Gibson, Impregnation Techniques for Thermoplastic Matrix
Composites, Polymers and Polymer Composites, Vol. 4, No. 7, 1996.
[12] A.H. Miller, H. Dodds, J.M. Hale, A.G. Gibson, High Speed Pultrusion of Thermoplastic Matrix Composites, Composites part A 29A, 1998, 773-782.
[13] P. Mitschang, M. Blinzler, A. Wöginger, Processing Technologies for Continuous Fiber Reinforced Thermoplastics with Novel Polymer Blends, Composites
Science and Technology 63 2003, 2099-2110.
[14] J.P. Nunes, J.F. Silva, A.T. Marques, N. Crainic, S. Cabral-Fonseca, Production
of Powder-Coated Towpregs and Composites, Journal of Thermoplastic Composite Materials, Vol. 16, May 2003.
[15] J.P. Nunes, F.W.J. Van Hattum, C.A. Bernardo, Advances in Thermoplastic Matrix Towpregs Processing, Journal of Thermoplastic Composite Materials, Vol.
17, November 2004.
[16] S. Padaki, L.T. Drzal, A simulation Study on the Effects of Particle Size on the
Consolidation of Polymer Powder Impregnated Tapes, Composites, Part A, 325337, 1999.
[17] K. Ramani, D.E. Woolard, M.S. Duvall, An Electrostatic Powder Spray Process
for Manufacturing Thermoplastic Composites, Polymer Composites, Vol.16,
No.6, December 1995, pp 459-469.
[18] M. Rath, S. Kreuzberger, G. Hinrichsen, Manufacture of Aramid Fiber Reinforced Nylon-12 by Dry Powder Impregnation Process, Composites Part A, 29A,
1998, 933-938.
[19] S. Wu, Electrostatic Charging and Deposition of Powder Coating, Polym. Plast.
Technol. Eng., 7(2), 1976, 119-220.
BIBLIOGRAPHY
111
[20] Patent US 3,456,156, An Apparatus for Applying an Electrostatic Charge to
Fibrous Material, July 15, 1969.
[21] Patent US. 3,704,485, J. N. Hall, Apparatus for Spreading a Graphite Fiber Tow
into a Ribbon of Graphite Filaments, 1972.
[22] Patent US 3,795,944, C.G. Daniels, Pneumatic Spreading of Filaments, 1974.
[23] Patent US 3,817,211, A.W. Brown, R.J. Moran, Apparatus for Impregnationg
Strands, Webs, Fabrics and the Like, 1974.
[24] Patent US 3,873,389, C.G. Daniels, Pneumatic Spreading of Filaments, 1975.
[25] Patent US 3,919,437, A.W. Brown, R.J. Moran, Method for Electrostatically
Impregnating the Strand, 1975.
[26] Patent US 3,967,118, E.M. Sternberg, C. Hill, Method and Apparatus For
Charging a Bundle od Filaments, June 29, 1976.
[27] Patent US 4,081,856, E.M. Sternberg, C. Hill, Apparatus for Forwarding and
Charging a Bundle of Filaments, Mar. 28, 1978.
[28] Patent US 3,940,464, J.H. Davis, R.C. Harwood, B. Davis, Production of Laces
and Granules of Fiber-Reinforced Thermoplastics, 1976.
[29] Patent US 4,427,712, P.N.Y. Pan, Electrodynamic Coating Process, 1984.
[30] Patent US 4,575,999, M. Anahara, H. Omori, Pneumatic Nozzle Utilized in the
Process of Producing a Fasciated Yarn, 1986.
[31] Patent US 4,588,538, T.-S. Chung, H. Furst, Z. Gurion, P.E. McMahon, R.D.
Orwoll, D. Palangio, Process for Preparing Tapes from Thermoplastic Polymers
and Carbon Fibers, 1986.
[32] Patent US 4,626,306, G. Chabrier, G. Moine, R. Maurion, R. Szabo, Process for
Manufacturing Profiled Strips in Fiber-Loaded Thermoplasticm Resin, 1986.
[33] Patent US 4,713,139, R.A. Ganga, Apparatus for Producing Flexible Composite
Material, 1987.
112
BIBLIOGRAPHY
[34] Patent US 4,799,985, P.E. McMahon, T.-S. Chung, Method of Forming Composite Fiber Blends and Molding Same, 1989.
[35] Patent US 4,828,776, E.M. J. Morel, G.M. Richert, Process for Impregnating
with Thermoplastic or Thermosetting Polymers in Solid State Fiber of Great
Length, 1989.
[36] Patent US 4,994,303, R.W. Calkins, Fiber Impregnation Process, 1991.
[37] Patent US 4,999,733, T. Kakuda, Web Electricity Removing Roller, Mar. 12,
1991.
[38] Patent US 5,019,427, D.A. Soules, Apparatus and Process for Improved Thermoplastic Prepreg Materials, 1991.
[39] Patent U.S. 5,042,111, Iyer at al., Method and System for Spreading a Tow of
Fibers, 1991
[40] Patent US 5,057,338, R.M. Baucom, J.J. Snoha, J.M. Marchello, Process for
Application of Powder Particles to Filamentary Materials, 1991.
[41] Patent US 5,060,351, R.L. Street Process and Apparatus for Blowing Continuous Filament Tow, 1991.
[42] Patent US. 5,057,338, R.M. Baucom, N. News, J.J. Snoha, J.M. Marchello, Process for Application of Powder Particles to Filamentary Materials Oct. 15, 1991.
[43] Patent US 5,094,883, J.D. Muzzy, B. Varughese, Flexible Multiply Towpreg and
Method of Production therefore, 1992.
[44] Patent US 5,101,542, N. Nakagawa, Y. Ohsora, Fiber Separator for Producing
Fiber Reinforced Metallic or Resin Body, 1992.
[45] Patent US 5,102,690, S. Iyer, L.T. Drzal, K. Jayaraman, Method Coating Fibers
with Particles by Fluidization in a Gas, 1992
[46] Patent US 5,128,199, S. Iyer, L.T. Drzal, Method for Fiber Coating with Particles, 1992.
BIBLIOGRAPHY
113
[47] Patent US 5,200,620, J.M. Peritt, R. Everett, A. Edelstein, Electrostatic Fiber
Spreader Including a Corona Discharge Device, 1993.
[48] Patent U.S. 5,213,843, D.C. Working, Vacuum Powder Injector and Method of
Impregnating Fiber with Powder, 1993.
[49] Patent US 5,275,883, M. Leone, R.F. Carbone, R.A. Ganga, Composite Material the Characteristics of which can be Modulated by Preimpregnation of a
Continuous Fiber, 1994.
[50] Patent US 5,296,064, D. Muzzy, J.S. Colton, Flexible Multiply Towpreg Tape
from Powder Fusion Coated Towpreg and Method for Production thereof, 1994.
[51] Patent US 5,334,414, D.D. Edie, J.W. Klett, Process for Coating Carbon Fibers
with Pitch and Composites Made therefrom, 1994.
[52] Patent US 5,364,657, J.L. Throne, Method of Depositing and Fusing Polymer
Particles onto Moistened Continuous Filaments, 1994.
[53] Patent US 5,370,911, J.L. Throne, A. L. Ogden, Method of Depositing and Fusing Charged Polymer Particles on Continuous Filaments, 1994.
[54] Patent US 5,446,952, C. Kim, R.A. Gray Jr., Pneumatic Induction Fiber
Spreader with Lateral Venturi Restrictors, 1995.
[55] Patent US 6,032,342, K. Kawabe, S. Tomoda, Multi-Filament Split-Yarn Sheet
and Method and Device for the Manufacture thereof, 2000.
[56] Patent US 6,049,956, J.L. Lifke, L.D. Busselle, D.J. Finley, B.W. Gordon,
Method and Apparatus for Spreading Fiber Bundles, 2000.
[57] Patent US 6,253,431, T. Ames, W. West, R. Kenley, W.T. Wygand, E.J. Powers,
Air Opening Jet Apparatus, 2001.
[58] Patent US 6,311,377, E.G. Krueger, Apparatus and Method for Spreading Fibrous Tows into Linear Arrays of Generally Uniform Density and Products
Made thereby, 2001.
114
BIBLIOGRAPHY
[59] Patent US 6,372,294, A.M. Vodermayer, J. Krger, S. Kaufmann, H. Erlach,
Manufacture of Unidirectional Fiber Reinforced Thermoplastic, 2002.
[60] Patent US 6,836,939, J.-M. Guirman, B. Lecerf, A. Memphis, Method and Device for Producing a Textile Web by Spreading Tows, 2005.
[61] Patent US 6,983,520, T. Ames, R.L. Kenley, W.T. Wygand, B.R. Lomax, Method
and Apparatus for Making an Absorbent Composite, 2006.
[62] Patent EP 0 642 844 B1, H. Fischer, Laboratory Sieving Device, 1994.
[63] M. Connor, Consolidation Mechanisms and Interfacial Phenomena in Thermoplastic Powder Impregnated Composites, These N◦ 1423, 1995.
[64] A.M. Vodermayer, J.C. Kaerger, G. Hinrichsen Manufacture of High Performance Fiber-Reinforced Thermoplastics by Aqueous Powder Impregnation,
Composites Manufacturing, Vol 4 No 3. 1993.
Chapter 6
Characterization of PA6T/6I
Solid State Polymerization
The advantage of the in-situ solid state polymerization (in-situ SSP) is that shaping,
consolidation, crystallization and polymerization can occur below the melt temperature of the polymer, with a reduction of the processing time and energy consumption.
Moreover, it is well known that solid state polymerization increases molecular weight
and concomitant material properties in polymers such as polyesters and nylons [11],
[12]. Solid state polymerization of polymer chips or powder before shaping, also
known as pre-extrusion SSP, is already an important industrial process with a wide
application in the manufacture of soft-drink and beverage bottles, tire cord filaments,
and industrial fibers. However, there is a practical limit to the achievable molecular weight through pre-extrusion SSP, associated with the subsequent processing:
high molecular weight polymers require high processing pressure and can undergo
thermo-mechanical degradation. On the other hand, in case of post-extrusion SSP
(i.e. solid state polymerization after shaping of the final part) there is no limit on
the final molecular weight since polymerization occurs after full impregnation of the
fibers and preform shaping. In this chapter the reaction kinetics of PA 6T/6I postextrusion solid state polymerization (SSP), the most important variables affecting the
SSP reaction and the mechanical properties of the final polymer were investigated in
116
Chapter 6. Characterization of PA6T/6I Solid State Polymerization
order to determine the optimum operating conditions and establish a useful process
window for the reactive molding process on carbon fibers.
6.1
Experimental
The PA 6T/6I prepolymer powder (XE 3733 VK) was dried in a vacuum oven at 80◦ C
and 4 mbar for at least 24 hours before further use. Polymerized samples were prepared at different temperatures and for different times in a hot press under protective
atmosphere.
6.1.1
Tooling
The influence of the impregnation temperature, heating rate and polymerization temperature on the PA6T/6I material properties was investigated by utilizing the matched
plate mould illustrated in Fig. 6.1. The molding tool was mounted inside the press
chamber of a 400 kN hydraulic hot press. PA6T/6I slabs were press formed by distributing the prepolymer powder between two thin aluminum sheets embedded in two
aluminum frames. The press plates (upper and lower mould in Fig. 6.1) were heated
to the desired temperature. As the press chamber was closed and evacuated, the frame
was brought into contact with the upper mould by the movement of the lower mould
through the hydraulic cylinder. Finally, the frame came into contact with the lower
mould through the springs of the holding pins. In order to determine the optimal
process window for PA6T/6I prepolymer powder coated preforms, two different experimental investigations were conducted. First slabs of PA6T/6I prepolymer were
formed in a hot press and the polymerization reaction characterized by DSC measurements. The prepolymer powder was pressed with the lowest pressure allowed by
the hydraulic system (0.4 MPa) for 30 seconds with the press plate heated to 172◦ C.
Under these conditions, fully crystallized prepolymer slabs could be prepared within
few seconds and with only one process step. Finally slabs of PA6T/6I were formed
and polymerized in the hot press under vacuum, in order to characterize the polymerization reaction in the hot press.
6.1 Experimental
117
Figure 6.1: Match plate mould for the production of PA6T/6I polymer slabs.
6.1.2
Rheological characterization
Solution viscosity measurements of the prepolymer were carried out at concentrations
of 0.1 g/dL, 0.3 g/dL, 0.5 g/dL und 0.7 g/dL in 98% sulphuric acid in a Paar Physica
MC 300 rheometer using Couette geometry. The intrinsic viscosity [η] was calculated
from the reduced viscosity ηsp /c by extrapolation to zero concentration.
6.1.3
Thermal Characterization
A Perkin Elmer Pyris 1 DSC instrument was used to study the melting, cold-crystallization
and polymerization behavior of PA 6T/6I prepolymer. All DSC non-isothermal tests
were performed from 80◦ to 350◦ C at different heating rates under nitrogen atmosphere to prevent high temperature oxidation. Since the PA6T/6I prepolymer undergo
loss of weight during polymerization with water as by-product, TGA analysis was
performed on a Perkin Elmer Pyris 1 TGA under nitrogen atmosphere to study the
polymerization behavior. Samples were heated at 30◦ /min to different polymerization temperatures and held at the polymerization temperatures for different times.
118
6.1.4
Chapter 6. Characterization of PA6T/6I Solid State Polymerization
Mechanical Characterization
A Perkin Elmer 7 dynamic mechanical analyzer (DMA) in three-point bending deformation mode was used to conduct thermomechanical tests. Samples 5 mm wide
and 20 mm long were cut from PA6T/6I prepolymer slabs with different thickness
previously polymerized in a hot press for 15 min at different temperatures. The sample temperature was varied between -10◦ C and 250◦ C while the static force, dynamic
force and frequency were held constant throughout the experiments. The glass transition temperature was defined in terms of the maximum loss modulus (G"max ) as
recommended by ASTM D 4065-2001.
6.2
6.2.1
Results and discussion
Thermal characterization of pre-extrusion PA6T/6I prepolymer
A typical DSC temperature scan of PA 6T/6I prepolymer powder between 80◦ C and
350◦ C is shown in chapter 4.2.1, Fig. 4.5. The as-received amorphous prepolymer
undergoes a cold-crystallization upon heating (peak 2), at a temperature between the
glass transition temperature (peak 1) and the onset of the solid state polymerization
(SSP) (peak 3). This cold-crystallization observed for PA 6T/6I arises from the short
chain lengths and low level of entanglements of the molecules: below the glass transition temperature the mobility of the short chain molecules is strongly reduced, but
above the glass transition temperature they can easily rearrange and crystallize presumably in a paraffin-like structure (fully extended chain crystals). The stability of
such a structure is further increased in the case of polyamides, which can form hydrogen bonds between the chains. The crystallization of the prepolymer is an important
step to produce high molecular weight PA 6T/6I polymer by SSP. The reactive end
groups are excluded from the crystalline region of the prepolymer, i.e. they are not
present as isolated defects in crystallites [5], [6] . Thus, they are confined in the
amorphous regions which have a relatively high mobility. As a consequence, the concentration of end groups in the amorphous regions of the semicrystalline prepolymer
is many times higher than the concentration that would be present in a completely
non-crystalline prepolymer at the same temperature [7].
6.2 Results and discussion
119
Figure 6.2: Solid state polymerization of PA 6T/6I after cold crystallization.
Between 220◦ C and 300◦ C peak 3 can be identified as a broad endothermal peak
where solid state polycondensation and evaporation of the reaction by-product water
take place. Peak 4 between 300◦ and 320◦ C and peak 5 between 320◦ and 345◦ C are
presumably the melting peak of the formed polymer and a melt polymerization (MP)
peak, respectively. In the cooling curve (lower curve in Fig. 4.5) one crystallization
peak can be identified between 280◦ C and about 210◦ C as well as the glass transition
temperature of the final polymer at 130◦ C.
The demarcation between SSP and MP reactions was further investigated by determining the crystallization temperature (TCC ) of samples upon cooling from the
polymerization temperature. Fig. 6.3 presents the DSC thermograms of a prepolymer
powder sample heated to 290◦ C (curve a)), polymerized at 290◦ C for 10 minutes and
cooled to 80◦ C (curve b)) at 10◦ C/min.
The absence of crystallization on cooling from the polymerization temperature of
the above mentioned polymer indicates that a SSP reaction proceeds in the amorphous
region of the cold-crystallized prepolymer, i.e. the chains grow via the reaction between the chain ends of the prepolymer that are already ordered in different lamella.
Due to the above mentioned short chain lengths and the low level of entanglements
of the PA 6T/6I prepolymer chains, larger crystals with fewer defects are expected
especially if compared with high molecular weight polymers [8]. This can be confirmed from the presence of only one sharp melting peak at 324◦ C in a subsequent
second heating run from 80 to 350◦ C (curve c)) at 10◦ C/min. The demarcation between SSP and MP was found for powder samples polymerized at 300◦ C and 305◦ C,
respectively. At 305◦ C solid state and melt reactions likely take place simultaneously,
as indicated by the presence of a crystallization peak on cooling from the polymer-
120
Chapter 6. Characterization of PA6T/6I Solid State Polymerization
st
1 Heating to 290°
st
1 Cooling
nd
2 Heating to 350°
nd
2 Cooling
14
Heat flow endo >, mW
12
10
131.954
8
1
4
c)
5
2
b)
4
0
d)
8
-2
6
3
a)
2
6
324.507
253.274
7
-4
177.986
100
125
150
175
200
225
250
275
300
325
350
Temperature, °C
Figure 6.3: DSC thermograms of a prepolymer powder sample heated to 290◦ C
(curve a)), polymerized for 10 min and cooled to 80◦ C (curve b)). Second heating to
350◦ C (curve c)) and subsequent second cooling to 80◦ C (curve d)) are also shown.
Heating/cooling rate 10◦ C/min.
ization temperature (Tab. 6.1). On the other hand, after polymerization at 300◦ C no
crystallization peak was observed on cooling, indicative of pure SSP. At the polymerization temperature of 310◦ C the enthalpy of crystallization on cooling increases.
Moreover, the crystallization peak shifts to higher temperature, due probably to the
formation of lower molecular weight molecules by MP which readily crystallize.
Chain building reactions are usually slow at SSP temperatures as compared to
polymerization in the melt, because of the reduced mobility of the reacting species
and the slow diffusion of the by-products [7]. However, for PA6T/6I prepolymer
the polymerization was found to be very rapid and to take place within the first ten
minutes, as shown in Fig. 6.4. The progress of the polymerization reaction in powder
was assessed by TGA analysis by determining the amount of released reaction byproduct (water) during the reaction. It may be noted that the final weight loss at
290◦ C is higher than at 300◦ C. Probably at 300◦ C the prepolymers are very close
121
6.2 Results and discussion
Polymerization Temperature [◦ C]
290
300
305
310
TCC [◦ C]
No peak
No peak
265
278
∆Hcc [J/g]
No peak
No peak
11
17
Table 6.1: Characteristic data of non-isothermal exothermal crystallization peaks determined with DSC by cooling the sample at 10◦ C/min from different polymerization
temperatures. All the PA6T/6I prepolymer powder samples were polymerized for 10
minutes under nitrogen.
Weight [% ]
100
320°
300°
290°
280°
98
96
94
5
9.5
12
128
Reac tion Time [min]
Figure 6.4: Weight loss of PA6T/6I prepolymer samples during polymerization reaction at different temperatures and for different times, as determined by TGA.
to their melting temperatures and partial melting may occur with reduced by-product
diffusion and particle agglomeration.
Diffusion of the condensate within the solid reacting mass (interior diffusion) and
from the reacting mass to the inert gas (surface diffusion) was found to affect the
progress of the SSP reaction as much as temperature. Fig. 6.5 and Fig. 6.6 show the
progress of the SSP reaction monitored by TGA upon heating to 290◦ C and holding
122
Chapter 6. Characterization of PA6T/6I Solid State Polymerization
Thermal Analysis different sample weight
350
1
15.3 mg
8.2 mg
3.6 mg
Temperature
Weight [%]
0.98
300
250
0.97
200
0.96
150
Temperature [°C]
7.9 mg
0.99
0.95
100
0.94
0.93
50
0
10
20
30
40
Time [min]
Figure 6.5: Prepolymer SSP reaction at 290◦ C for ten minutes with constant powder
particle size (100 µm) but different powder sample weights.
the different samples for ten minutes at the SSP temperature. In Fig. 6.5 the prepolymer was sieved to obtain a uniform particle size of 100 µm and different total sample
weights were polymerized, while in Fig. 6.6 the prepolymer was sieved to obtain several different particle sizes (d<50µm, 50<d<100 µm and 100<d<250 µm) but the total
sample weights were held constant.
The figures show a reduction of the progress of the SSP reaction with increasing
sample weights or diameter of prepolymer particles, indicating that diffusion of the
by-product through the solid polymer is rate-determining in the PA6T/6I SS polymerization process. Increasing the diameter of prepolymer particles was found to slightly
increase the melting temperature of the final PA6T/6I polymer, as shown in Tab. 6.2.
Higher melting temperatures of the samples with larger powder sizes can be explained considering that the extent and depth of crystalline regions is presumably
higher the larger the powder size. Since the polymerization reaction can occur between molecules of the same lamella, the force of attraction between the chains in the
well-ordered depths of the crystal is greater than at the surface, so that thicker crystals
have higher melting point [8], [9]. The lower crystallization temperature (Tab. 6.2)
could also indicate the formation of higher molecular weight chains, which could be
123
6.2 Results and discussion
T hermal Analysis different powder sizes
400
50 μm 1
50 μm 2
1
350
100 μm 1
0.99
100 μm 2
250 μm 1
250 μm 2
300
Weight [%]
Temperature
250
0.97
0.96
200
0.95
150
0.94
100
0.93
Temperature [°C]
0.98
50
0
10
20
30
40
Time [min]
Figure 6.6: Prepolymer SSP reaction at 290◦ C for ten minutes with constant powder
sample weight but with different powder particle size.
Powder size [µm]
100<d<250
50<d<100
d<50
Tg [◦ C]
134.2
133.7
132.9
Tm [◦ C]
324.5 sharp
322.7
321.2
∆Hm [J/g]
40.9
43.1
42.2
Tc [◦ C]
254.3
262.1
262.1
∆Hc [J/g]
33
28.9
29.8
Table 6.2: Characteristic DSC data of a second temperature scan between 80 and
350◦ C for PA6T/6I polymers at 10◦ C/min heating/cooling rate. The samples were in
a first temperature run SS polymerized at 290◦ for 10 minutes. The powder sample
weight was kept constant but the particle powder sizes were varied.
explained with the increased depth of the prepolymer crystals and with the increased
number of adjacent prepolymer crystals in larger powder. Therefore an increased
number of polymerization reactions in the same crystal and an increased possibility
of reaction between chains of different crystals should be expected.
In order to assess the molecular weight of the final polymer, the inherent viscosity of the PA 6T/6I SSP polymerized prepolymer was measured for different poly-
124
Chapter 6. Characterization of PA6T/6I Solid State Polymerization
Inherent Viscosity [g/dL]
1.2
225°C
1.0
250°C
275°C
0.8
Oligomer
0.6
0.4
0.2
0.0
0
5
10
15
20
Reaction Time [min]
Figure 6.7: Effect of the reaction temperature on the inherent viscosity of PA 6T/6I
after SSP for different reaction times. Solution concentration of 0.3 g/dL in 98%
sulphuric acid at 25◦ C.
merization temperatures (Fig. 6.7). Elevated polymerization temperatures combined
with high residence times were found to encourage thermal degradation and undesirable side reaction. The side reactions include three-dimensional network formation
(Tab. 6.3), which can drastically impair the quality and the recyclability of the end PA
6T/6I product.
6.2.2
Thermal characterization of PA6T/6I post-extrusion polymerized slabs
Samples of prepolymer were cut from a slab of PA6T/6I prepolymer pressed at 172◦ C
for 30 seconds with 20 kN. All the samples had the same weight (10 mg ± 0.1mg)
and were polymerized in a first DSC heating run in SSP reaction for 10 minutes at
different temperature between 270◦ C and 300◦ C, i.e. below the melting point of the
125
6.2 Results and discussion
Time [min]
225◦ C (SSP Temperature)
250◦ C (SSP Temperature)
275◦ C (SSP Temperature)
300◦ C (SSP Temperature)
325◦ C (MP Temperature)
350◦ C (MP Temperature)
* Insoluble; only swelling
2
0.118
0.462
0.734
0.743
0.811
0.7
5
0.245
0.603
0.846
*
*
10
0.285
0.774
1.141
*
*
*
20
0.480
*
*
Table 6.3: Inherent viscosity of samples polymerized at different temperatures for
different times. Solution concentration of 0.3 g/dL in 98% sulphuric acid at 25◦ C.
PA6T/6I polymers. The DSC thermograms of the second heating run after the polymerization reaction are shown in Fig. 6.8. With increasing polymerization temperature from 270◦ C to 300◦ C, the temperature of peak 4 rose from 305.8◦ C to 318.8◦ C,
while the area under the peak 5 diminished (Tab. 6.4). Peak 4 could represent the
melting peak of the polymer while peak 5 can be related to the evaporation of water
which had not diffused out of the material during the initial polymerization process
or to the evaporation of water as by-product from a post polymerization.
A superposition of a DSC and a TGA scan obtained from two polymer samples is
shown in Fig. 6.9. Both polymer samples had the same weight and were polymerized
from the same prepolymer slab at 290◦ C for ten minutes under nitrogen. The TGA
curve indicates that in correspondence to DSC peak 5 there is an evident but very low
weight loss. It is interesting to compare this DSC thermogram with the DSC thermogram of a PA6T/6I prepolymer powder sample polymerized at the same temperature
and for the same time. In the case of the powder sample (curve c) in Fig. 6.3) only
one sharp peak at 324.5◦ C is present, indicating the absence of evaporation of water
or of a post polymerization reaction. This difference between the two samples can be
explained with the diffusion distance of the polymerization by-product in the range of
millimeters for the prepolymer slab and in the range of micrometers for the powder
sample.
DSC analysis of polymer samples previously melt polymerized in DSC at different temperatures for 10 min is shown in Fig. 6.10 and Tab. 6.5. All samples had the
126
Chapter 6. Characterization of PA6T/6I Solid State Polymerization
36
Samples polymerized at
270°for 10 min
280°for 10 min
290°for 10 min
300°for 10 min
Heat flow endo up >, mW
33
30
27
24
d)
21
18
c)
15
12
a)
6
75
100
125
150
175
200
225
T5
∆HTOT
Tg,h
3
0
T4
b)
9
250
275
300
325
350
Temperature, °C
Figure 6.8: DSC heating curves from 80 to 350◦ C for polymer samples previously
solid state polymerized at different temperatures: a) 270◦ C for 10 min; b) 280◦ C for
10 min; c) 290◦ C for 10 min; d) 300◦ C for 10 min. Heating rate 10◦ C/min.
Polymerization Temperature [◦ C]
Tg,h [◦ C]
T4 [◦ C]
∆Htot [J/g]
T5 [◦ C]
Tc [◦ C]
∆Hc [J/g]
Tg,c [◦ C]
300
136.5
318.8
57.7
330.9
256.1
31
135.1
290
137.4
312.1
56.6
328.6
253.1
30.7
135.1
280
140.5
309.5
55.7
328.2
249.8
29.5
135.2
270
136.9
305.8
58.3
326.9
250.4
29.8
135.9
Table 6.4: Characteristic DSC data for polymer samples previously solid state polymerized for 10 min at different temperatures. The sample weight was kept identical.
Heating rate 10◦ C/min.
127
6.2 Results and discussion
Heat flow endo up > mW
225
250
275
300
325
350
312.281
TGA
DSC
100.4
100.2
327
20
100.0
318.5
99.8
15
99.6
Weight loss, %
200
25
99.4
10
200
99.2
225
250
275
300
325
350
99.0
Temperature, °C
Figure 6.9: Superposition of a DSC and a TGA scan obtained from two polymer
samples cut form the same prepolymer slab and polymerized in the DSC at 290◦ for
10 min under nitrogen. The initial weight of the two samples was identical. Heating
rate 10◦ C/min.
same weight and were taken from the same slab and heated to the polymerization
temperatures at 6◦ C/min and cooled to 80◦ C at 33◦ C/min. The results of the DSC
analysis clearly show higher and sharper melting peaks for the samples polymerized
at lower temperature especially if compared with the broad melting peak of the sample polymerized at 340◦ C. In the latter case, the reduced mobility of high molecular
weight chains compared to short chain prepolymer presumably leads to crystalline domains with a distribution of fold periods and hence melting temperatures. Moreover,
the morphology, the size and perfection of the crystalline domains are also significantly affected by the cooling rate (33◦ C/min). On the contrary, by SSP the lamellas
are formed during cold-crystallization and finally in SSP and are not affected by the
cooling rate.
128
Chapter 6. Characterization of PA6T/6I Solid State Polymerization
Polymerization Temperature [◦ C]
Tg,h [◦ C]
T4 [◦ C]
∆H4 [J/g]
Tc [◦ C]
∆Hc [J/g]
Tg,c [◦ C]
310
136.5
326.2
45.3
253.4
29.5
136.5
320
137.3
334.2
30.7
258.4
31.5
137.3
340
130.6
321.9
25.8
285.6
31.5
130.6
Table 6.5: Characteristic DSC data for polymer samples previously melt polymerized
for 10 min. at different temperatures. The sample weight was kept identical. Heating
rate 10◦ C/min.
28
310°C for 10 min
320°C for 10 min
340°C for 10 min
26
Heat flow endo up > mW
24
22
T4
20
18
c)
16
∆H4
b)
14
12
a)
10
Tg,h
8
6
4
100
125
150
175
200
225
250
275
300
325
350
Temperature, °C
Figure 6.10: DSC heating curves from 80 to 350◦ C for polymer samples previously melt polymerized at different temperatures: a) 305◦ C for 10 min; b) 310◦ C
for 10 min; c) 320◦ C for 10 min; d) 340◦ C for 10 min. Heating rate 10◦ C/min.
6.3 Conclusions
6.2.3
129
Thermomechanical characterization of PA6T/6I post-extrusion polymerized slabs
The mechanical properties of PA6T/6I polymers polymerized for 10 minutes at different SSP and MP temperatures in a hot press were determined by dynamic mechanical
analysis (DMA) between -10◦ C and 250◦ C. The slabs of PA6T/6I prepolymer had
been pressed at 172◦ C for 30 seconds with 20 kN. Storage modulus as a function of
SSP and MP temperature is illustrated in Fig. 6.11, and the glass transition temperatures are summarized in Tab. 6.6. The graphs in Fig. 6.11 show that the mechanical
properties of the PA6T/6I polymers are directly related to the polymerization process
and polymerization temperature. Higher values of the storage modulus and glass transition temperature were found for samples polymerized in SSP at 290◦ C and 300◦ C
as compared to samples polymerized in MP at 320◦ C and 340◦ C. Generally, the mechanical properties of the PA6T/6I polymer matrix decrease with increasing polymerization temperatures. However, the values of the storage modulus and glass transition
temperature (defined in terms of the maximum loss modulus as measured by DMA)
for samples polymerized in SSP are negatively affected by the thickness of the prepolymer slabs. Fig. 6.12 and Tab. 6.7, respectively, show the storage modulus and
the glass transition temperatures of PA6T/6I polymer samples polymerized in SSP at
290◦ C for 10 min from prepolymer slabs with different thicknesses. The mechanical properties of the final polymer are adversely affected with increasing thickness of
the reacting mass due presumably to the longer diffusion path for the polymerization
by-product (water) to reach the reacting mass surface (interior diffusion). The presence of the reaction by-product can decrease the polymerization reaction rate, can
lead to de-polymerization or increased porosity, negatively affecting the mechanical
properties of the resulting polymer.
6.3
Conclusions
The post-extrusion solid state polymerization (SSP) process of PA 6T/6I prepolymer
slabs presented in this work offers several advantages in terms of processing time, reduced energy consumption, tooling costs, and improved mechanical properties of the
final polymer. High flexural modulus (4-5 GPa) and high Tg ( 150◦ C) PA 6T/6I polymer can be produced without catalyst by SSP of PA 6T/6I prepolymer in a hot press at
130
Chapter 6. Characterization of PA6T/6I Solid State Polymerization
5x109
290°C
300°C
320°C
340°C
a)
9
Storage modulus, Pa
4x10
b)
3x109
c)
2x109
d)
1x109
0
0
25
50
75
100
125
150
175
200
225
250
Temperature [°C]
Figure 6.11: Storage modulus from -10◦ C to 250◦ C for polymer samples previously
polymerized at different temperatures in a hot press: a) 290◦ C for 15 min; b) 300◦ C
for 15 min; c) 320◦ C for 15 min; d) 340◦ C for 15 min. All samples had been heated
to the polymerization temperatures at 6◦ C and cooled to 20◦ C at 33◦ C/min. DMA
heating rate 10◦ C/min
Polymerization Temperature [◦ C]
290 (SSP)
300 (SSP)
320 (MP)
340 (MP)
Tg [◦ C]
151.8
151.2
149.8
147.3
Table 6.6: Glass transition temperature as function of polymerization temperature.
The samples had been polymerized in a hot press at different polymerization temperatures for 15 minutes from prepolymer slabs of identical thicknesses.
131
6.3 Conclusions
5.0x109
a)
Storage modulus, Pa
4.5x109
0.5 mm
0.90 mm
2 mm
4.0x109
3.5x109
3.0x109
b)
2.5x109
c)
2.0x109
9
1.5x10
1.0x109
5.0x108
0.0
0
25
50
75
100
125
150
175
200
225
250
Temperature, °C
Figure 6.12: Storage modulus from -10◦ C to 250◦ C for polymer samples with different thicknesses previously polymerized at 290◦ C in a hot press: a) 0.5 mm; b)
0.9 mm; c) 2 mm. All samples had been heated to the polymerization temperature at
6◦ C/min, polymerized for 15 minutes and cooled to 20◦ C at 33◦ C/min.
Sample thickness [mm]
0.5
0.9
2
Tg [◦ C]
151.8
140.7
138
Table 6.7: Glass transition temperature as function of sample thickness. The samples
had been polymerized in a hot press at 290◦ C for 15 minutes from prepolymer slabs
of different thicknesses.
132
Chapter 6. Characterization of PA6T/6I Solid State Polymerization
290◦ C, a polymerization temperature below the melting temperature of the resulting
polymer. Additionally, the low reaction temperature of PA 6T/6I SSP restrains side
reactions and thermal degradation, leading to a higher degree of reactions in the main
polymer chain and recyclability of the final polymer. The polymerization reaction,
typically slow for SSP compared to melt phase polymerization, was found to proceed
within ten minutes for the PA 6T/6I prepolymer. Crystallization, consolidation and
shaping of PA6T/6I prepolymer occur simultaneously in a few seconds at 172◦ C under isothermal condition and before SSP, with a consistent reduction of processing
time and pressure. Moreover, due to the short chain lengths and low level of entanglements of the PA6T/6I prepolymer molecules, higher extent and depth of crystalline
regions with higher melting point can be achieved through SSP as compared to melt
polymerized material. However, an increase in the thickness of the prepolymer slabs
polymerized by SSP reaction was found to adversely affect the mechanical properties
of the resulting polymer, due to an increase of the diffusion path length in the reacting
mass of the polymerization by-product. Therefore, a prerequisite to obtaining a PA
6T/6I polymer product of high molecular weight from prepolymer slabs of elevated
thickness consists in increasing the coefficient of diffusion of the polymerization byproduct to prevent depolymerization or increased porosity.
Bibliography
[1] K. van Rijswijk, H.E.N. Bersee, Reactive Processing of Textile Fiber-Reinforced
Thermoplastic Composites - An overview, Composites: Part A, 38, 666-681,
2007.
[2] H.P. Schmeer, Polyphthalamid, Kautschuk.Gummi Kunstoffe, 46, Jahrgang, Nr.
10/93,1993.
[3] C. Zaniboni, P. Ermanni, An Electrostatic Powder Spray Process for Manufacturing Polyphthalamide High Performance Composite, FPCM8, 445-451, 2006.
[4] C. Zaniboni, P. Ermanni, Reactive Stamp Forming of Polyphthalamide as Matrix
Material in High Performance Composites, ESAFORM 2006, 763-766, 2006.
[5] Zimmerman J, Equilibria in Solid Phase Polyamidation, Journal of Polymer
Letter, Vol.2, 955-958, 1964.
[6] J. Zimmerman, M. Kohan, Nylon-Selected Topics, Journal of Polymer Science,
Part A: Polymer Chemistry, 39: 2565-2570, 2001.
[7] S.N. Vouyiouka, E.K. Karakatsani, C.D. Papaspyrides, Solid State Polymerization, Progress in Polymer Science, 30, 10-37, 2005.
[8] N.R. James, C.Ramesh and S. Sivaram, Development of Structure and Morphology during Crystallization and Solid State Polymerization of Polyester
Oligomers, Macromol. Chem. Phys 202, 1200-1206, 2001.
134
BIBLIOGRAPHY
[9] P.C. Painter, M.M. Coleman, Fundamentals of Polymer Science, Technomic
Publishing Company, 1997.
[10] K. Yao, W. Ray, Modeling and Analysis of New Processes for Polyester and
Nylon Production, AIChE J 47 (2), 401-412, 2001.
[11] A.K. Agrawal, V.T. Mhaisgawali, Post-Extrusion Solid-State Polymeritation of
Fully Drawn Polyester Yarns, Journal of Applied Polymer Science, Vol. 102,
5113-5122, 2006.
[12] R.J. Schiavone, Solid State Polymerization (SSP) of Low Molecular Weight
Poly(ethylene terephthalate) (PET) Copolyesters Compared to Conventional
SSP of PET, Journal of Applied Polymer Science, Vol. 86, 230-238, 2002.
[13] W Yan, K. Han, L. Qin, M. Yu, Study on Long-Fiber-Reinforced Thermoplastic
Composites Prepared by In Situ Solid-State Polycondensation, Journal of Applied Polymer Science, Vol. 91, 3059-3965, 2004.
[14] N. Pini, C. Zaniboni, S. Busato, P. Ermanni, Perspectives for Reactive Molding of PPA as Matrix for High-performance Composite Materials, Journal of
Thermoplastic Composite Materials, Vol. 19, No. 2, 207-216 (2006)
[15] C. Zaniboni, N. Pini, S. Busato, P. Ermanni, Reactive Processing of Polyphthalamide as Temperature Resistant Matrix for High Performance Composite
Materials, THEPLAC 2005, 15-16 September 2005.
Chapter 7
Consolidation and
Polymerization of PA6T/6I
Oligomer Pre-forms
The processing of thermoplastic composites generally requires the application of heat
and pressure in order to melt the polymer, to fully impregnate the reinforcement fibers
and to establish the geometry of the final part [1]. The induced pressure is intended
to squeeze air and resin out, to suppress void, to increase and to make uniform the
fiber volume fraction. This process is called consolidation [2]. In case of PA6T/6I
thermoplastic composites an additional polymerization step at a temperature above
the consolidation temperature and below the melting temperature of the final polymer
is required. This chapter describes the principal factors involved in the consolidation
and polymerization stage of processing PA6T/6I powder-impregnated composites, focusing on the investigation of the relationship between oligomers powder size, processing parameters and laminate quality in terms of mechanical properties and void
content.
136
7.1
Chapter 7. Consolidation and Polymerization of PA6T/6I Oligomer Pre-forms
Processing of aligned thermoplastic powder impregnated composites
The processing of aligned thermoplastic composites can be divided into three basic
phases: heating, consolidation and cooling [3]. During consolidation heat and pressure are applied to enable healing between plies, to bring in intimate contact the tows,
and to induce melt flowing of the polymer matrix in order to fully impregnate the
reinforcement fibers [2]. The aim of powder impregnation processing techniques is
to minimize the distance which the resin must flow to wet the reinforcement fibers
in order to achieve good impregnation during consolidation at economic cycle times
[4]. These techniques have the advantage that the matrix in powder form is intimately
mixed with the reinforcement fibers independently of thermoplastic matrix viscosity
[5]. The consolidation process of towpregs with uniform distributed powder involves
two steps: 1) drops of molten resin powder form a bridge between adjacent fibers and
spread along the fibers until all adjacent bridges come in intimate contact and physical deformation occurs at the polymer-polymer interface [4], [5]. 2) Autoadhesion or
interdiffusion of polymer chains across the interface [5]. It is generally agreed that the
dominant mechanisms of consolidation involves flow of resin along the fibers, since
the axial permeability is about an order of magnitude greater than that in the transverse direction [4], [6], [7]. However, with increasing particles size the likelihood of
transverse flow can not be further neglected [8]. The flow pressure, distance and time
required from the resin to flow in the direction parallel to the fiber are usually small
but increase with interparticle distance and particle size [3], [9]. Ideally, particles
should be the same diameter as the fibers and uniformly distributed among the fiber
reinforcement [5].
7.2
Processing of aligned PA6T/6I prepolymer powder
impregnated composites
The innovative reactive forming process of aligned PA 6T/6I prepolymer powder
coated composites can be divided into five basic phases: heating, consolidation, heating, polymerization and cooling (Fig. 7.1).
In the first two stages, the prepolymer powder coated preforms are heated in a hot
7.2 Processing of aligned PA6T/6I prepolymer powder impregnated composites 137
Figure 7.1: Processing cycle of aligned PA6T/6I prepolymer powder impregnated
composites.
138
Chapter 7. Consolidation and Polymerization of PA6T/6I Oligomer Pre-forms
press to the consolidation temperature i.e. to the softening temperature of the prepolymer, so that the prepolymers can flow under pressure and wet the fibers in a short
time while the preforms are shaped to the final forms [10]. When semicrystalline
thermoplastics are involved, solidification and crystallization are also considered as
part of the consolidation process [1]. Usually the crystallization process of semicrystalline polymers occurs during cooling from the consolidation temperature but in case
of PA6T/6I prepolymer powder coated preforms, crystallization takes place after consolidation on further heating (third stage of the reactive forming process) from the
melting temperature to the polymerization temperature of the prepolymers. This
cold-crystallization process proceeds at very high heating rates since the prepolymer molecules can easily rearrange to a crystalline structure due to the short chain
lengths and low level of entanglements. In the fourth process step the fully impregnated, consolidated and shaped prepolymer laminates are further heated under vacuum to the polymerization temperature of the prepolymers in order to obtain the final
PA6T/6I polymer. By the polymerization temperature the end groups of the PA6T/6I
prepolymers are sufficiently mobilized to enable polymerization in the solid state and
in-situ, i.e. directly on the carbon fibers [11]. The polymerization temperature of the
prepolymer occurs below the melt temperature of the final PA6T/6I polymer, with a
consequent reduction of the processing time and energy consumption. After the polymerization reaction is completed, the PA6T/6I laminates which were already crystallized and solidified to the final form are cooled to a temperature below 200◦ C, i.e. a
temperature at which the laminate do not need to be kept under vacuum in order to
avoid degradation reaction with oxygen.
7.3
7.3.1
Experimental
Powder impregnation and characterization of carbon fiber
tows
Powder impregnated tows with three different powder sizes were employed to produce laminates with unidirectional reinforcement fiber orientation. The three different powder impregnated tows were produced with the same experimental procedure
and powder impregnation parameters. The typical coating line operating conditions
used to produce towpregs in this work are shown in Tab. 7.1. To investigate the
139
7.3 Experimental
Variable
Linear Fiber Pull Speed
Corona Charge
Corona Voltage
Fluidization Pressure
Vertical Vibration
Air Flow Velocity
Unit
m/s
mA
KV
bar
mm
m/s
Value
4
50
100
4.5
0.94
4.4
Table 7.1: Typical coating line operational condition.
micro-impregnation quality of the different powder impregnated tows, micrographs
were captured with a LEICA DM RXA Microscope equipped with a LEICA DC 480
digital camera.
7.3.2
Laminate preparation and characterization
Filament winding laid preforms with powder impregnated tows
To obtain laminates with unidirectional (UD) fiber arrangement, the different powder impregnated fiber tows were individually laid on a frame by filament winding
(Fig. 7.2). According to the powder fiber tow, the feed rate per revolution was set to
0.8 mm, resulting in a filament angle of 89.9◦ relative to the rotation axis. In order
to investigate the relationship between powder coating processing parameters and final laminate quality, laminates with as-molded thickness of 1.5 mm were obtained
from tows previously impregnated with three different powder size (100<d<250 µm,
50<d<100 µm, and d<50 µm). To investigate the effect of increasing laminate thickness on the final PA6T/6I polymer matrix, several layers of prepreg tows were wound
to the frame, resulting after consolidation in laminates with different thickness (1.5 mm,
1.8 mm and 2.6 mm). The laminates were obtained from prepreg tows previously impregnated with d<50 µm PA6I/6I prepolymer powder matrix.
The frame containing the wound powder prepreg tows was then transferred to a
hot press and consolidated to the final laminates. To reduce air entrapment in the final laminate and to ensure easy air transportation during the impregnation stage, the
140
Chapter 7. Consolidation and Polymerization of PA6T/6I Oligomer Pre-forms
Powder Size [µm]
Impregnation Pressure [MPa]
Impregnation Temperature [◦ C]
Impregnation Time [min]
Polymerization Temperature [◦ C]
Polymerization Time [min]
<50
0.5
140
10
300
15
50÷100
0.5
140
10
300
15
100÷250
0.5
140
10
300
15
Table 7.2: Consolidation and reactive processing parameters for the manufacturing
of PA6T/6I laminates.
Figure 7.2: Photograph of a UD laminate.
mould was open on both sides. The overall optimized pressing cycle in the hot press
can be divided in two process steps. (1) The frame containing the wound powder
prepreg tow is placed in a hot press, previously heated up to 140◦ C, and pressed at
the lowest pressure (0.5 MPa) allowed by the hydraulic system. In this first stage fully
micro impregnation of the carbon fibers by means of the PA6T/6I prepolymers and
subsequent crystallization of the matrix material take place. (2) The press plates are
heated to the polymerization temperature of the PA6T/6I prepolymer matrix without
applied pressure, in order to not hinder the diffusion of the polymerization by-product
through the laminate thickness. Then the laminate is cooled down to room temperature. The cooling rates are limited to approximately 30◦ C/min due to the machinery
system. An overview of the optimized processing parameters for the manufacturing
of the PA6T/6I laminates is given in Tab. 7.2.
7.3 Experimental
141
Fabric stacking with matrix powder
Percolation of matrix through the fiber reinforcement plays an essential role in the
production of composites, since reduces the void content and, in particular, allows the
bonding of different layers of towpregs [12]. In order to investigate the percolation
through the carbon fiber reinforcement by the PA6T/6I prepolymer matrix in the hot
press a technique similar to fabric stacking was used to prepare unidirectional carbon
fiber composites with powder and without prepregging. Instead of the polymer films,
the prepolymer powder was electrostatically sprayed to each layer of unidirectional
fabric during the stacking sequence [13]. The final stacked layers were transferred for
the final consolidation to laminates to a hot press heated at 140◦ C. Different impregnation pressure (0.5-8 MPa) and holding time (5 - 20 min) were selected to identify
the impregnation mechanisms of PA6T/6I prepolymer as a function of the prepolymer
processing conditions in a hot press.
Tensile and flexural tests
In order to preliminarily characterize the new PA6T/6I composites, to study the effects
of the processing parameters on the quality of the final laminates and in order to permit preliminary comparisons with other existing matrix systems, flexural strength and
modulus were investigated. Flexural properties of unidirectional fiber composites,
when tested parallel to a principal fiber direction, are much more affected by matrix
properties then tensile properties, which are far more dependent upon the continuous
fiber reinforcement properties. Moreover, the apparent interlaminar shear strength
determined by short-beam method is usually more directly related to the void content
then flexural properties and for processing parameters leading to low void contents
could fail to consider other important microstructural parameters related to the matrix
system and fiber-matrix interface [14].
The UD laminates were cut parallel to the fiber direction to obtain samples for
flexural tests parallel to the fiber direction according to the European Standard EN
2562. Five test specimens were taken out of each laminate. The samples were subjected to a cross head speed of 0.77 mm/min and the flexural modulus was measured
between a strain of 0.05 and 0.25% and strength was determined at flexural failure.
All experiments were carried out on a Zwick universal testing machine.
142
Chapter 7. Consolidation and Polymerization of PA6T/6I Oligomer Pre-forms
Porosity and fiber volume content tests
Optical microscopy is a proper and inexpensive method for evaluate the fiber distribution and impregnation quality of cross-sections of the final laminates. Due to the
very toxic nature of the suggested solvent and to the high decomposition temperatures and time for PA6T/6I matrix, the void content and the fiber volume fraction
of the final laminates were determined by means of photomicrographic method and
imaging analysis using LEICA QWin software package. The fiber volume fraction
was determined as:
Af
(7.1)
A
where Af and A are the total fiber area and the area of the selected region of the
micrograph respectively.
Vf =
7.4
Results and Discussion
In order to produce PA6T/6I laminates of good quality and mechanical performance
it is necessary to initially gain a fundamental understanding of the effects of prepreg
and molding processing conditions on the structure and morphology of the resultant
laminates.
7.4.1
Microstructure
It is important to evaluate in which process step of the whole processing route the
micro-impregnation of the carbon fiber tows occurs. Fig. 7.3 shows a micrograph
of the cross-sections of a carbon fiber tow coated using the spray impregnation process, exhibiting full micro-impregnation of the fiber bundles. Because of the low
melt viscosity of PA6T/6I prepolymers and the high fiber spreading (Fig. 7.3) microimpregnation or wetting of single filaments inside the rovings occurs during the powder impregnation process. Moreover, the powder impregnated preform shows a high
shape ratio, i.e. a high ratio between the width and the height of the powder impregnated tow, which reduces the impregnation time and may improve the impregnation
and consolidation of laminates made from these prepreg in a subsequent compression
molding step [17], [18]. However, due to the false twist and small variation of the
7.4 Results and Discussion
143
tension applied on the fiber it is difficult during the process to hold the fiber spreading
constant in terms of spreading width and fiber distribution, as shown in Fig. 7.4. As a
consequence the fibers tend to shift from the centre of the tow and to accumulate at the
boarders leading to resin rich areas in the centre and fibers rich areas at the borders of
the final towpreg. In order to reduce the influence of the false twist, a first spreading
step was introduced, as shown in Fig. 7.3. Moreover, a brake tensioner (Fig. 7.5) was
introduced to replace the servo based system achieving a better control of the tension
applied on the fiber tow at very low value. However, without a strain gauge tension
monitor and a phase close loop it is not possible to hold the applied tension on the
carbon fibers perfectly constant.
Fig. 7.6 shows micrographs of consolidated prepolymer UD laminates prepared
at two different values of applied pressure and for the longer holding time (20 min);
the processing temperature was held constant at 140◦ C. The laminates were prepared
via fabric stacking with prepolymer matrix powder. According to the literature [13],
this technique permits the production of good quality laminates with moderate melt
viscosity polymer matrix, like J-polymer. Although the very low prepolymer matrix
viscosity and the very high consolidation time it is clearly evident from the micrographs, that no percolation of the prepolymer matrix and no wetting of the fiber takes
place. Applied pressure, temperature and holding time do not affect the quality of impregnation and consolidation of PA6T/6I prepolymer preforms during compression
molding in a hot press, probably due to the shear thinning and cold crystallization
behavior of the matrix system. As already demonstrated in chapter 4.2.4, increasing applied pressures, temperatures and times induce the prepolymer to crystallize.
Moreover, the viscosity of the prepolymer matrix decreases with increasing shearing
forces which are in a pressing process usually very low. In conclusion it is reasonable
to assume that the impregnation step which largely affects the properties of the final
PA/6T/6I composites occurs during the powder coating process of the fiber reinforcement for the production of PA6T/6I prepolymer towpregs.
Fig. 7.7 shows the micrographs of the PA6T/6I polymer laminates prepared laying
on a frame several layer of towpregs previously impregnated with different powder
sizes; the laminates were processed with the same processing cycle. The laminates
show fairly large resin rich areas as well as densely packed flattened fiber agglomerations. This separation in fiber and resin rich areas increases with increasing prepolymer powder size. The resin rich areas of the composite specimens show large
inter-tow voids and the densely packed agglomerations small intra-tow porosities, re-
144
Chapter 7. Consolidation and Polymerization of PA6T/6I Oligomer Pre-forms
First Spreading Step
Second Spreading Step
120µm
Figure 7.3: Photograph and micrograph of an electrostatically spray impregnated
carbon fiber tow.
sulting in an overall void content of 22.2 ± 2.2 %, 14.2 ± 1.7% 15.8 ± 1.2% for
laminates prepared from towpregs impregnated with 100<d<250 µm, 50<d<100 µm
and d<50 µm prepolymer powder, respectively.
When modeling the consolidation time of laminates produced from powder impregnated tows, many authors [8], [4] assume the particle flow as the major mechanism of consolidation and consider the powder as reduced void content between the
fibers. As the particle size increases, decreases the degree of mingling of the fiber
7.4 Results and Discussion
145
Figure 7.4: Picture of the fiber spreader and micrographs of an electrostatically spray
impregnated carbon fiber.
with the matrix and increases the distance which a single particle has to flow parallel
or transverse to the fibers to obtain a void free laminates [9]. Moreover, the flow pressure, i.e. the pressure necessary for the full consolidation of the laminate, was found
for different matrix system to increase by three order of magnitude as the particle
size change by one order of magnitude (10-100 µm) [19]. The static experiments for
the optimization of the electrostatic powder deposition (chapter 5.4) also showed that
with increasing powder size decreases the uniformity of prepolymer powder deposition, due probably to the increase in the aerodynamic forces acting on the powder.
Therefore, laminates impregnation quality and void content can only be improved
during powder coating decreasing the powder size [15]. However, due to the increase
146
Chapter 7. Consolidation and Polymerization of PA6T/6I Oligomer Pre-forms
a) servo-based tensioner
b) brake-based tensioner
Figure 7.5: Micrograph of a) a servo-based tensioner and b) a brake-based tensioner.
of interparticle forces with decreasing powder size, PA6I/6T prepolymer particles in
the range of 50 µm tend to agglomerate and consequently to act as bigger particles.
7.4.2
PA6T/6I laminate flexural properties
The flexural modulus and flexural strength results of the final PA6T/6I polymer laminates obtained from towpregs impregnated with different powder sizes are reported
in Fig. 7.8 and Fig. 7.9, respectively. The mechanical properties were assessed by
three-point bending tests in the fiber direction and compared to the theoretical values
(Tab. 7.3) using the rules of mixture (Vf = 39%) [16].
The values of flexural strength and modulus calculated according to the theory using unreinforced matrix and fiber data have higher values than the one experimentally
calculated. The difference could be explained with the very high void contents ascertained in all PA6T/6I laminates. Decreasing the PA6T/6I prepolymer particle size
during the powder impregnation process positively affects the flexural properties due
probably to the decreased void content in the final laminated. Another factor which
could negatively affects the flexural properties is the fiber-matrix adhesion [14]: low
interfacial strength between carbon fibers and PA6T/6I polymer matrix can also negatively influence the final flexural strength.
147
7.4 Results and Discussion
a)
b)
Figure 7.6: Micrographs of consolidated prepolymer laminates prepared at 140◦ C
with holding time of 20 min for two different pressures: a) 0.5 MPa and b) 8 MPa,
respectively. The laminates were prepared via fabric stacking with prepolymer matrix
powder.
148
Chapter 7. Consolidation and Polymerization of PA6T/6I Oligomer Pre-forms
a)
b)
c)
Figure 7.7: Micrographs of PA6T/6I polymer UD laminates obtained from towpregs
impregnated with a) d<50µm, b) 50<d<100µm and c) 100<d<250µm PA6I/6I prepolymer powder matrix and polymerized at 300◦ C for 15 min in a hot press.
149
7.4 Results and Discussion
100
250 - 100 µm
100 - 50 µm
< 50 µm
90
Flexural Modulus, GPa
80
70
60
50
40
30
20
10
0
250
100
50
Powder size, µm
Figure 7.8: Flexural modulus of PA6T/6I polymer laminates obtained from towpregs
impregnated with d<50 µm, 50<d<100 µm and 100<d<250 µm PA6I/6I prepolymer
powder matrix and polymerized at 300◦ C for 15 min in a hot press.
1400
Flexural Strenght, MPa
1200
250 - 100 µm
100 - 50 µm
< 50 µm
1000
800
600
400
200
0
250
100
50
Powder size, µm
Figure 7.9: Flexural strength of PA6T/6I polymer laminates obtained from towpregs
impregnated with d<50 µm, 50<d<100 µm and 100<d<250 µm PA6I/6I prepolymer
powder matrix and polymerized at 300◦ C for 15 min in a hot press.
150
Chapter 7. Consolidation and Polymerization of PA6T/6I Oligomer Pre-forms
Powder Size [µm]
100<d<250
50<d<100
d<50
E (GP a)
Experimental Theoretical
40.5 ± 9
91.3
62.5 ± 8
91.3
58.2 ± 6
91.3
σR (M P a)
Experimental Theoretical
874 ± 91
1821.2
1052 ± 68
1821.2
1069 ± 89
1821.2
Table 7.3: Flexural properties of PA6T/6I polymer laminates obtained from towpregs
impregnated with d<50 µm, 50<d<100 µm and 100<d<250 µm PA6I/6I prepolymer
powder matrix and polymerized at 300◦ C for 15 min in a hot press (Vf = 39%).
Fig. 7.10 and 7.11 show the values of the flexural modulus and flexural strength
of PA6T/6I polymer laminates with different thickness (1.5 mm, 1.8 mm and 2.6 mm,
respectively) prepared from towpregs impregnated with d<50µm powder particles.
The flexural properties of the final laminates seem to be adversely affected from the
increasing thickness of the laminate. This behavior could be explained with the higher
diffusion path which the polymerization reaction by-product has to travel in the composites in order to reach the reacting surface. The presence of the condensate could
lead to de-polymerization in the matrix or increases the voids volume fraction, adversely affecting the laminate flexural properties [14].
7.5
Conclusions
In this chapter processing of PA6T/6I prepolymer powder impregnated intermediate
materials with different powder sizes and thicknesses has been examined at different
processing conditions.
It is shown that the laminate quality in terms of void content and mechanical
properties are strongly influenced from the prepolymer powder size used during the
deposition process and on the thickness of the final laminate. Because of the low melt
viscosity of PA6T/6I prepolymers, the high fiber spreading, the melting stage in the
infrared heater and the capillary forces high degree of fiber wet-out can be achieved
during the powder deposition process. The consolidation step in the hot press should
also contribute to the reduction of the final void volume fraction and to the increase
of the degree of mingling. However, due to the low shear forces and to the cold crys-
151
7.5 Conclusions
100
1.5 mm
1.8 mm
2.6 mm
90
Flexural Modulus, GPa
80
70
60
50
40
30
20
10
0
1.5
1.8
2.6
Thickness, mm
Figure 7.10: Flexural modulus of PA6T/6I polymer laminates with 1.5 mm, 1.8 mm
and 2.6 mm thickness, polymerized at 300◦ C for 15 min in a hot press and obtained
from towpregs impregnated with d<50µm prepolymer powder particles.
1400
1.6
1.8mm
2.6mm
Flexural Strenght, MPa
1200
1000
800
600
400
200
0
1.5
1.8
2.6
Thickness, mm
Figure 7.11: Flexural Strength of PA6T/6I polymer laminates with 1.5 mm, 1.8 mm
and 2.6 mm thickness, polymerized at 300◦ C for 15 min in a hot press and obtained
from towpregs impregnated with d<50µm prepolymer powder particles.
152
Chapter 7. Consolidation and Polymerization of PA6T/6I Oligomer Pre-forms
tallization induced on the prepolymers from compression forces, very low degree of
fiber wet-out occurs during the molding process. Therefore, the laminate impregnation quality and consequently the final laminates mechanical properties can be mostly
improved during powder impregnation decreasing the powder size. However, there is
a lower limit (50µm for PA6I/6T prepolymer particles) due to the increase of interparticle forces with decreasing powder size which crate particle agglomerates.
Another factor influencing the final PA6T/6I laminate mechanical properties is
probably the presence of moisture as polymerization reaction by-product of the final
polymerization step. In fact, the mechanical properties of the final polymer matrix
seems to be adversely affected by increasing laminate thickness due presumably to
the higher diffusion path around the fibers which the condensate has to travel in order
to reach the reacting surface. The presence of the condensate in the final composite
could lead to de-polymerization in the matrix or increase the voids content.
Bibliography
[1] P.-E. Bourban, N. Bernet, J.-E. Zanetto, J.A.E. Månson, Material Phenomena
Controlling Rapid Processing of Thermoplastic Composites, Composites: Part
A, 32, 1045-1057, 2001.
[2] T.G. Gutowski, Z. Cai, S. Bauer, D. Boucher, J. Kingery and S. Wineman, Consolidation Experiments for Laminate Composites, Jounal of Composite Materials, Vol. 21, 650-669, 1987.
[3] S. Padaki, L.T. Drzal, A Consolidation Model for Polymer Powder Impregnated
Tapes, Journal of Composite Material, Vol. 31, No. 21, 2202-2227, 1997.
[4] M. Connor, S. Toll, J.-A. E. Månson, A Model for the Consolidation of Aligned
Thermoplastic Powder Impregnated Composites, Jounal of Thermoplastic Composite Materials, Vol. 8, 138-162, 1995.
[5] S.R. Iyer, L.T. Drzal, Manufacture of Powder-Impregnated Thermoplastic Composites, Journal of Thermoplastic Composite Materials, Vol. 3, 325-355, 1990.
[6] J.P. Nunes, A.M. Brito, A.S. Pouzada, C.A. Bernardo, Isothermal and NonIsothermal Consolidation of Carbon Fiber Towpregs, Polymer Composites, Vol.
22, No. 1, 71-79, 2001.
[7] J.F. Silva, J.P. Nunes, A.T. Marques, Consolidation of Glass FiberPolypropylene Towpregs by Compression Moulding, Materials Science Forum,
Vol. 514-516, 677-681, 2006.
154
BIBLIOGRAPHY
[8] A. Miller, G. Wei, A.G. Gibson, Manufacture of Polyphenylene Sulfide (PPS)
Matrix Composites via the Powder Impregnation Route, Composite PART A,
Vol. 27A, No. 1, 49-56, 1996.
[9] A. M. Vodermayer, J.C. Kaerger, G. Hinrichsen, Manufacture of High Performance Fiber-Reinforced Thermoplastics by Aqueous Powder Impregnation,
Composites Manufacturing, Vol. 4, No. 3, 123-132, 1993.
[10] C. Zaniboni, P. Ermanni, An Electrostatic Powder Spray Process for Manufacturing Polyphthalamide High Performance Composite, FPCM8, 445-451, 2006.
[11] C. Zaniboni, P. Ermanni, Reactive Stamp Forming of Polyphthalamide as Matrix
Material in High Performance Composites, ESAFORM 2006, 763-766, 2006.
[12] J.A. Barnes, F.N. Cogswell, Transverse Flow Process in Continuous FiberReinforced Thermoplastic Composites, Composites, Vol. 20, No. 1, 38-42, 1969.
R
[13] I.Y. Chang, Thermoplastic Continuous Filament Composites of KevlarAramid
or Graphite Fiber, Composites Science and Technology, Vol. 24, 61-79, 1985.
[14] M.N. Bureau, J. Denault, Fatigue Resistance of Continuous Glass
Fiber/Polypropylene Composites: Consolidation Dependence, Composites Science and Technology, 64, 1785-1794, 2004.
[15] A.G. Bailey, The Science and Technology of Electrostatic Powder Spraying
Transport and Coating, Polym. Plast. Technol. Eng., Journal of Electrostatics,
45, 85-120, 1998.
[16] K. Ramani, D. E. Woolard, M. S. Duvall, An Electrostatic Powder Spray Process
for Manufacturing Thermoplastic Composites, Polymer Composites, Vol.16,
No.6, 459-469, 1995.
[17] L. Ye, K. Friedrich, D. Cutolo, A. Savadori, Manufacture of CF/PEEK Composites from Powder/Sheath Fiber Preforms, Composites Manufacturing, Vol.5,
No.1, 41-50, 1994.
[18] A.G. Gibson, J. -A. Månson, Impregnation Technology for Thermoplastic Matrix Composites, Composites Manufacturing, Vol.3, No.4, 223-133, 1992.
BIBLIOGRAPHY
155
[19] S. Padaki, L.T. Drzal, A Simulation Study on the Effects of Particle Size on the
Consolidation of Polymer Powder Impregnated Tapes, Composites: Part A, 30,
325-337, 1999.
156
BIBLIOGRAPHY
Chapter 8
Concluding remarks and
outlook
A new cost-effective press forming route for an in situ polymerizing thermoplastic
matrix system with outstanding mechanical, thermal and chemical properties not yet
applied in the market of high-performance composites was developed in this research
work. In this last short chapter we will try to resume the results and sketch an outlook
for the project.
8.1
Conclusions
The major goal of the project was to develop a reactive processing route based on
prepreg technologies, in situ solid state polymerization (SSP) and press forming for
the use of PA6T/6I polyphthalamide as matrix material in high performance, high
temperature composites. The selected PA6T/6I prepolymers are an intermediate step
of a high volume production and therefore a low cost material system suitable for
mass production.
Due to their very low melt viscosity, this in-situ polymerizing system could be
processed through low pressure forming techniques unlike all classic manufacturing
processes for thermoplastics composites. However, between softening and polymer-
158
Chapter 8. Concluding remarks and outlook
ization temperature, the prepolymers undergo a non-reversible cold crystallization
process. Unfortunately the cold crystallization of PA6T/6I prepolymers is promoted
not only by increasing temperature but also by increasing applied pressures. Due to
the low shear forces and to the compression forces during the molding step, impregnation of the carbon fibers through the low melt viscosity oligomers can not be achieved
in a hot press, i.e. while the preforms are heated to the softening temperature of the
prepolymers and shaped to the final form. The fiber wet-out takes place principally
during powder deposition in the prepreg production process.
The cost-effective and flexible electrostatic powder spray impregnation process
developed in this work achieve high mingling of the fiber with the prepolymer powder
matrix, due to the to fiber spreader able to obtain high amount of spreading with very
low fiber damage for a wide range of tow velocity. However, powder spray processes
aim to mingle the reinforcing fibers and the thermoplastic matrix in solid form in order
to produce a preform where the reinforcing fibers are not fully wetted by the matrix.
The fully impregnation of the fibers is usually shifted in the final shaping step with
the advantage of a more flexible and drapeable preform. Therefore in case of PA6T/6I
oligomer powder impregnated prepregs, the final laminate impregnation quality can
be only slightly improved during powder impregnation decreasing the powder size
with a consequent drawback on the final laminates mechanical properties.
The post-extrusion in-situ solid state polymerization of the PA6T/6I prepolymers on the carbon fibers after laminate shaping offers several advantages in terms
of processing time, consistent reduction of energy consumption, tooling costs, and
improved mechanical properties of the final polymer matrix. Crystallization, consolidation and shaping of PA6T/6I matrix occur simultaneously in few seconds under
isothermal condition before SSP with a consistent reduction of processing time and
pressure. There is no limit to the achievable molecular weight of the PA6T/6I matrix
system, since shaping occurs before polymerization and there is no need of subsequent processing for the final polymer. High flexural modulus and high Tg matrix can
be produced without catalyst in a hot press within ten minutes at a polymerization
temperature below the melting temperature of the resulting polymer. Additionally,
the low reaction temperature of PA6T/6I SSP restrains side reactions and thermal
degradation leading to higher degree of chain reactions and recyclability of the final
polymer. However, an increase in thickness of the final laminate polymerized by postextrusion SSP was found to adversely affect the mechanical properties probably due
to the increase of the diffusion path around the fibers in the reacting mass of the poly-
8.2 Outlook
159
merization by-product. Therefore, a prerequisite to obtain high performance PA6T/6I
composite of increasing thickness consists in increasing the coefficient of diffusion of
the condensate in the polymerizing matrix to prevent de-polymerization.
8.2
Outlook
In order to achieve full impregnation of the reinforcing fibers, to improve the final
laminate quality and therefore the final composite properties, pre-impregnation processes instead of post shaping impregnation processes should be used for the production of PA6T/6I high performance composites. In order to achieve full impregnation
of the reinforcement fibers and short impregnation time, several melt impregnation
processes aim at reducing the viscosity of the thermoplastic matrix upon forcing the
molten polymer by high shear rate forces into the fiber tows to form a prepreg tape of
different widths. These techniques were found to operate very well with low viscosity
thermoplastics melts at line speed up to 30 m/min. Due to the low melt viscosity of
the PA6I/6T prepolymers under shear forces, higher processing speed and high fiber
volume fractions could be achieved.
Fully impregnated PA6T/6I prepolymer tapes could be used in automated tape laying processes for the production of complex shape parts or filament winding. Due the
post-extrusion in-situ solid state polymerization process, high consolidation and interlaminar adhesion should be achieved. Moreover, due to fully impregnated PA6T/6I
prepolymer tapes, new low pressure forming techniques could be developed and applied to cost-effective produce large continuous fiber thermoplastic composites without high consolidation pressure, heavy steel moulds, heavy presses and high tooling
costs.
Another advantage associate with PA6T/6I prepolymer is the possibility of adding
to the unreacted monomer nano-particles in order to further more engineer the matrix
material. Moreover, the end groups of the PA6T/6I prepolymers could react with
the functional groups of a specific, for the purpose developed coupling agent on the
carbon fibers, thereby significantly improving the adhesion between carbon fiber and
matrix and thus improving the final properties of the composite.
Finally, in order to increase the rate of solid-state polymerization or to increasing the coefficient of diffusion of the condensate in the polymerizing matrix to prevent de-polymerization in the laminate with increasing thickness microwave energy
160
Chapter 8. Concluding remarks and outlook
should be investigated. Many work in literature report the enhancement of solid-state
polycondensation reaction rate by increasing the overall diffusion rate and decreasing
the activation energy for diffusion independently of temperature. Moreover, rapid and
uniform heating, improved physical and mechanical properties are other advantages
associated with microwave processing of composites. These advantages observed in
materials processed using microwave energy are being attributed to "microwave effects" which are particular of this technology and not yet well explained.
Acknowledgements
There are really many people which I would like to thank for their contribution to this
work. In particular I would like to thank:
GEBERT RÜF STIFTUNG for financially supporting the work presented in this
thesis and EMS Chemie A.G. for supplying the materials.
A special thanks to Prof. Dr. Paolo Ermanni, my supervisor, for giving me the
chance to work on a fantastic project, for supporting me during the most challenging
time of my life and for his patience especially at the end of my PhD.
Thanks also to Prof. Dr. Verpoest from K.U. Leuven (Belgium) for accepting
being my co-examinator.
A special acknowledgment to Dr. Klaus Häusler and Dr. Thomas-Bruno Schweizer
for the support and knowledge in the rheological characterization.
My very special thanks to Dr. Stephan Busato and Dr. Gerald Kress, two men of
great knowledge and research capabilities but of immense modesty, for their friendship and their contribution to this work.
My gratitude also goes to the whole group of the Centre of Structure Technologies, in particular to Niccolo’ Pini, Rene Arbter, Gion Barandun, Michael Sauter,
Simon Steiner and Alberto Belloli. I would like to thank them not only for the interesting discussions and their large help, but also for their friendship and their daily
162
Acknowledgements
contribution to an amusing working climate. A special thank also to Gerhard Kuhn
and Hanspeter Eigenmann for supporting me in the lab, to Johannes Hengstler and to
Anke-Christiane Kleint for helping me in many situations.
I would like to thank also following students, who contributed to the result of this
thesis (in strictly alphabetical order): Matthias Bobst, Tobias Huber, David Mächler,
Adrian Mettler, Christian Pitta, Griaznov Sergeji and Nadja Zürcher.
I special thanks to Dr. R.P. Müller, Hans-Ulrich Gerber, Joëlle Zingraff, Karl
Brentrup, Tilo Schimanski, Jérôme Lefèvre for supporting and encouraging me to finalize the thesis.
To my parents my gratitude for their encouragement and affection and for giving
me the opportunity to enjoy an academic education.
To Graziella and Evelino, my second family, a very special thanks for their kindness, their love and their big help.
To Andrea, my love and my husband, my gratitude for his uncommon generosity
and for having always supported me in any possible manner!
To my little son Pietro my gratitude for giving me joy every day.
Publication List
Journal publications
N. Pini, C. Zaniboni,S. Busato, P. Ermanni, Perspectives for Reactive Molding of
PPA as Matrix for High-performance Composite Materials, Journal of Thermoplastic
Composite Materials, Vol. 19, No. 2, 207-216 (2006)
Contributions to Conferences
N. Pini, C. Zaniboni, S. Busato, P. Ermanni, Perspectives for Reactive Molding of
PPA Matrix for High-Performance Composite Materials, FPCM7 2004.
Zaniboni C., Pini N., Busato S., Ermanni P., Reactive Processing of Polyphthalamide as Temperature Resistant Matrix for High Performance Composite Materials,
THEPLAC 2005, 15-16 September 2005.
C. Zaniboni, P. Ermanni, An electrostatic Powder Spray Process for Manufacturing
Polyphthalamide high Performance Composite, FPCM8, 445-451, 2006.
C. Zaniboni, P. Ermanni, Reactive Stamp Forming of Polyphthalamide as Matrix Material in high Performance Composites, ESAFORM 2006, 763-766, 2006.
164
Publication List
C. Zaniboni, P. Ermanni, Reactive Stamp Forming of Carbon Fiber / PPA Composites,
ICCM-16, Kyoto, Japan, 2007.