University of Southern California Dissertations and Theses

Void evolution in vacuum bag-only prepregs

(USC Thesis Other)

Void evolution in vacuum bag-only prepregs

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

VOID EVOLUTION IN VACUUM BAG-ONLY PREPREGS

by

Wei Hu

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfilment of the
Requirements of the Degree
DOCTOR OF PHILOSOPHY
(MATERIALS SCIENCE)

December 2019
I

Acknowledgements
This work would not have been possible without the support of my colleagues, friends and
family and I owe all of them my most heartfelt thanks.

I would first like to acknowledge my advisor, Professor Steven Nutt for the opportunities
he provided, his constructive guidance and constant support over the past six years. I have been
extremely fortunate to study under the guidance of Prof. Nutt. His wealth of research experience
and advice has provided me with the tools necessary in scientific research. I have learned how to
approach a scientific problem and to arrive at a good solution. Additionally, the guidance and
feedback I received on writing and presentation skills have helped improve my ability to
communicate more effectively.

It has been a privilege to be a member of M. C. Gill Composites Center. I thank Dr. Lessa
K Grunenfelder, who helped me start my research at the beginning of my doctorate work. She
provided me with valuable assistance, feedbacks and scientific discussions. I would also like to
thank Dr. Timotei Centea. He always took time to discuss with me and provided new perspective
and insights on my research that lead my work to the next level. I would also like to thank all my
colleagues at M.C. Gill Composites Center who have always been supportive and kind.

Last but certainly not least, I would like to thank my family without whom this would not
have been possible. I would like to thank my parents for their unconditional love and support
throughout my life. I would like to thank my husband, Hengxiao, for standing by my side and
helping me to always see the bright side.
II

Table of Contents
Acknowledgements ...................................................................................................................................................... I
Table of Contents ........................................................................................................................................................II
List of Tables ............................................................................................................................................................... V
List of Figures ........................................................................................................................................................... VI
Abstract ..................................................................................................................................................................... IX
Chapter 1 Introduction ............................................................................................................................................... 1
1.1 Motivation ................................................................................................................................................. 1
1.2 Problem statement ..................................................................................................................................... 2
1.3 Scope of dissertation ................................................................................................................................. 3
Chapter 2 Literature Review ...................................................................................................................................... 4
2.1 Vacuum bag-only prepregs........................................................................................................................ 4
2.2 Vacuum bag-only processing .................................................................................................................... 8
2.2.1 Consolidation .................................................................................................................................. 10
2.2.2 Resin flow and compaction............................................................................................................. 10
2.2.3 Air evacuation ................................................................................................................................. 12
2.3 Voids in composite processing ................................................................................................................ 15
2.3.1 Void formation ............................................................................................................................... 16
2.3.2 Effects of processing parameters on voids ...................................................................................... 17
2.3.3 Void characterization ...................................................................................................................... 20
2.3.4 Effects of voids on mechanical properties ...................................................................................... 24
2.4 Conclusions ............................................................................................................................................. 25
Chapter 3 In Situ Monitoring of Air Removal in Vacuum Bag-Only Processing of Unidirectional Prepregs .. 29
3.1 Introduction ............................................................................................................................................. 29
3.2 Experimental ........................................................................................................................................... 32
3.2.1 Materials ......................................................................................................................................... 32
3.2.2 In situ observation of VBO cure ..................................................................................................... 32
3.2.3 Void content ................................................................................................................................... 36
3.2.4 Partial cure sample preparation ....................................................................................................... 36
3.2.5 Water solubility in epoxy resin ....................................................................................................... 37
3.3 Results and Discussion ............................................................................................................................ 37
3.3.1 Air removal in unidirectional prepregs during VBO cure .............................................................. 37
3.3.2 Tow impregnation during cure ....................................................................................................... 40
3.3.3 Void removal mechanisms ............................................................................................................. 43
3.3.4 Bubble mobility .............................................................................................................................. 47
III

3.4 Conclusions ............................................................................................................................................. 50
Chapter 4 Effects of Material and Process Parameters on Void Evolution in Unidirectional Prepreg during
Vacuum Bag-only Cure ............................................................................................................................................. 53
4.1 Introduction ............................................................................................................................................. 53
4.2 Experimental procedures ......................................................................................................................... 55
4.2.1 Test matrix ...................................................................................................................................... 55
4.2.2 Resin film infiltration during cure .................................................................................................. 57
4.2.3 Resin moisture content during cure ................................................................................................ 58
4.3 Results and discussion ............................................................................................................................. 58
4.3.1 Vacuum hold effects ....................................................................................................................... 58
4.3.2 Bag pressure effects ........................................................................................................................ 61
4.3.3 Moisture effects .............................................................................................................................. 62
4.3.4 Film thickness change during cure ................................................................................................. 64
4.3.5 Moisture content in resin during cure ............................................................................................. 65
4.4 Model development ................................................................................................................................. 67
4.4.1 Model framework ........................................................................................................................... 67
4.4.2 Diffusivity ....................................................................................................................................... 68
4.4.3 Water concentration in resin during cure ........................................................................................ 68
4.4.4 Water concentration at void surfaces .............................................................................................. 70
4.4.5 Model correlation with experimental data ...................................................................................... 72
4.5 Conclusions ............................................................................................................................................. 74
Chapter 5 Void Evolution during Vacuum Bag-Only Processing of Fabric Prepregs ........................................ 76
5.1 Introduction ............................................................................................................................................. 76
5.2 Experimental procedure .......................................................................................................................... 77
5.2.1 Materials and test matrix ................................................................................................................ 77
5.2.2 Model development ........................................................................................................................ 78
5.2.6 Degree of impregnation during cure ............................................................................................... 82
5.3 Results and discussion ............................................................................................................................. 83
5.3.1 Inter-ply air removal during VBO cure .......................................................................................... 83
5.3.2 Effects of reduced vacuum on void evolution during cure ............................................................. 86
5.3.3 Effects of reduced vacuum on air evacuation ................................................................................. 88
5.3.4 Effects of reduced vacuum on bubble expansion ............................................................................ 90
5.3.5 Effects of reduced vacuum on tow impregnation ........................................................................... 91
5.3.6 The interaction mechanisms – inter-ply voids and intra-tow air ..................................................... 93
5.3.7 Effects of room temperature vacuum hold on void evolution during cure ..................................... 95
5.3.7 Effects of moisture on void evolution during cure ......................................................................... 97
IV

5.4 Conclusions ............................................................................................................................................. 99
Chapter 6 Cure Optimization – Effects of Debulk Temperature on Air Removal during Vacuum Bag-Only
Prepregs .................................................................................................................................................................... 101
6.1 Introduction ........................................................................................................................................... 101
6.2 Experimental ......................................................................................................................................... 103
6.2.1 Materials ....................................................................................................................................... 103
6.2.2 Permeability measurements .......................................................................................................... 104
6.2.3 Tow impregnation model .............................................................................................................. 107
6.2.4 Degree of impregnation ................................................................................................................ 110
6.3. Results and discussion .......................................................................................................................... 110
6.3.1 Inter-ply air evacuation ................................................................................................................. 110
6.3.2 Through-thickness permeability during heated debulk ................................................................. 113
6.3.3 In-plane permeability during pre-cure dwell ................................................................................ 117
6.3.4 Discussion ..................................................................................................................................... 118
6.4. Conclusions .......................................................................................................................................... 121
Chapter 7 Contributions and Future Work .......................................................................................................... 123
7.1 Contributions ......................................................................................................................................... 123
7.2 Broader implications ............................................................................................................................. 126
7.3 Future work ........................................................................................................................................... 126
References ................................................................................................................................................................ 128

V

List of Tables
Table 2- 1 In-plane and transverse permeability at room temperature reported in the literature [6]. .......................... 15
Table 2-2 A summary of commonly used void characterization techniques [80]. ....................................................... 24
Table 2-3 Overview of literature on voids in out-of-autoclave prepregs ..................................................................... 28

Table 4-1 Parameters included in this study. ............................................................................................................... 57
Table 4- 2 Parameter values and units used in this model. .......................................................................................... 71

Table 5-1 Cure kinetics and viscosity model parameters. ........................................................................................... 80
Table 5-2 Summary of model parameter values .......................................................................................................... 82

Table 6-1 Summary of model parameter values ........................................................................................................ 109

VI

List of Figures
Figure 2-1 Microstructure of uncured unidirectional VBO prepregs. ............................................................................ 5
Figure 2-2 A typical cure cycle for VBO prepreg processing [15]. ............................................................................... 8
Figure 2-3 Schematic of vacuum bag assembly [24]. .................................................................................................... 9
Figure 2-4 Schematic of the OoA prepreg consolidation process [6]. ......................................................................... 10
Figure 2-5 (a) Ultrasound scans of resin impregnation as a function of time at 80°C [10]; (b) images of resin
impregnation of ST95-RC 200T prepreg at 55°C [26]. ............................................................................................... 11
Figure 2-6 X-ray micrographs pre-treated with different cure cycle [27]. ................................................................... 12
Figure 2-7 Void in a composite laminate. .................................................................................................................... 15
Figure 2-8 (a) Average macro- and micro- porosity versus process conditions; (b) Micrographs of 5320/8HS and
5320/UD laminates [54]; (c) Void content versus relative humidity [55]. .................................................................. 18
Figure 2-9 Porosity as a function of out-time in a VBO prepreg [60]. ........................................................................ 20
Figure 2-10 Micrograph of voids in composite prepregs. ............................................................................................ 21
Figure 2-11 C-scan showing areas with different void contents: (a) cure pressure is 0.0 MPa, (b) cure pressure is 0.6
MPa, (c) presentation of grey level and its porosity [5]. .............................................................................................. 22
Figure 2-12 Voids in a composite with UD carbon/epoxy plies characterized by micro-CT, displayed in (a) top, (b)
3D, (c) front, and (d) side view [79]. ........................................................................................................................... 23

Figure 3-1Micrograph and schematic of inter-ply voids. ............................................................................................ 29
Figure 3- 2 (a) Schematic of in situ observation set-up, (b) Dimensions of perforated resin film, (c) Micrograph of
perforated resin film. ................................................................................................................................................... 35
Figure 3- 3 Cure cycle with partial cure sampling points marked with black squares. ................................................ 37
Figure 3- 4 Time-lapse images of void evolution in unidirectional prepregs during cure. Images were taken at a) initial
state, before vacuum applied, b) 20min, c) 36min, d) 86min, e) 8h and f) 12h. .......................................................... 40
Figure 3-5 Micrographs of tow impregnation in UD prepreg laminates during cure process. Laminates were (a) at
initial state after lay-up, cured to (b) 60°C, (c) 80°C and (d) 90°C. ............................................................................ 42
Figure 3-6 Thickness change of an 8-ply UD laminate during cure. ........................................................................... 43
Figure 3- 7 The relationship between void content, resin properties, and tow impregnation. The three stages of void
evolution are denoted with roman numerals. ............................................................................................................... 44
Figure 3- 8 Surface of unidirectional prepreg, showing resin rich and resin starved regions (arrows). Dry regions on
the surface of the prepreg enable through-thickness gas transport to an air evacuation pathway. ............................... 45
Figure 3- 9 Water solubility in epoxy (Cytec 5320-1) resin as a function of degree of cure. ...................................... 47
Figure 3- 10 (a) Temperature and resin viscosity profile versus time, indicating three time intervals used to analyze
bubble mobility, (b) correlation of bubble velocity with aspect ratio. ......................................................................... 50

Figure 4-1 Void evolution in unidirectional prepregs during room temperature vacuum hold and cure. (a) initial state,
before vacuum applied, (b) after 24-hour RT vacuum hold, (c) 30 min (~82°C), (d) 90 min, (e) 4 h and (f) 12 h. .... 60
VII

Figure 4-2 Void content as a function of time during room temperature and elevated cure. ....................................... 60
Figure 4-3 Void content as a function of time under reduced vacuum level. .............................................................. 62
Figure 4-4 Effects of moisture on void removal during cure. Images were taken at the end of cure for prepregs humidity
conditioned at (a) 1% RH, (b) 40% RH (baseline, as-received), (c) 99% RH for UD prepregs, respectively. ............ 63
Figure 4-5 void content as a function of (a) time for three cases; (b) initial moisture content of prepreg. .................. 64
Figure 4-6 (a) Film thickness as a function of time, (b) micrographs of resin film thickness during cure process. Top:
laminate was at 80°C (the beginning of stage II), bottom: at 90 min into the cure cycle (the end of stage II). Resin film
thickness was measured as the distance between the red parallel lines. ...................................................................... 65
Figure 4-7 Prepreg moisture content during cure. ....................................................................................................... 67
Figure 4-8 Moisture content in uncured prepreg as a function of relative humidity exposure. ................................... 70
Figure 4-9 The model prediction and experimental data of void size as a function of time for prepregs conditioned
under (a) 1% RH, (b) 40% RH (baseline), and (c) 99% RH, and (d) the effects of adjustment parameter on bubble
shrinkage (with an initial bubble size of 0.1 mm
3
). ..................................................................................................... 74

Figure 5-1 Void evolution in fabric prepregs during cure. Images were taken at (a) initial state, before vacuum applied,
(b) 10min, (c) 36 min, (d) 1h, (e) 6 h and (f) 12 h into the cure. ................................................................................. 85
Figure 5-2 (a) Void content, temperature, resin viscosity, and tow impregnation as a function of time. (b) Micrograph
of uncured prepreg surface. ......................................................................................................................................... 85
Figure 5-3 Void evolution under reduced vacuum conditions ..................................................................................... 88
Figure 5-4 Void content as a function of time (a) during room temperature vacuum hold and (b) during cure .......... 89
Figure 5-5 Effects of deficient vacuum applied at different stages during cure on bubble expansion ........................ 91
Figure 5-6 (a) Model predictions of degree of impregnation as a function of time (b) Model prediction vs. experimental
data. ............................................................................................................................................................................. 92
Figure 5-7 Cross-sections of laminates cured at (a) 70% vacuum, (b) 80% vacuum, (c) 100% vacuum, and (d) 100%
vacuum with sealed edges. .......................................................................................................................................... 93
Figure 5-8 Interactions between inter-ply air bubbles and intra-tow air. (a-c) Intra-tow air diffused into inter-ply air
bubbles (representative air bubbles were circled in red) , images were taken from prepreg cured at 70% vacuum at (a)
45 min, (b) 1.5h, (c) 8h; (d-f) air bubbles emerged from the pinholes, images were taken from prepreg cured at 80%
vacuum at (a) 50 min, (b) 52 min, (c) 55 min into the cure cycle; (d-f) elongated air bubbles developed alongside fiber
tows. Images recorded from prepreg cured at 70% vacuum at (d) 1.5h, (b) 4h, and (f) 8h. ........................................ 95
Figure 5-9 Effects of room temperature vacuum hold on air removal. Images were taken at (a) initial state, (b) after
12 h RT vacuum hold (c) after elevated cure at 93°C for 12 h. ................................................................................... 97
Figure 5-10 Effects of moisture on void removal during cure. Images were taken at the end of cure for prepregs
humidity conditioned at (a) 1% RH, (b) 40% RH (baseline, as-received), (c) 99% RH. ............................................. 98
Figure 5-11 (a) Void content as a function of time; (b) bulk porosity as a function of moisture content. ................... 99

Figure 6-1 Comparison of part quality cured using RT debulk and super-ambient dwell [125]. .............................. 102
Figure 6- 2 Schematic of in-plane permeability test set-up. ...................................................................................... 105
VIII

Figure 6-3 Schematic of through-thickness permeability test set-up. ........................................................................ 107
Figure 6-4 Void evolution during pre-cure dwell of PW prepregs: (a) initial state, (b) after 4-h RT vacuum hold, (c)
after 4-h 60°C vacuum hold; and for pre-cure dwell of UD prepregs: (d) initial state, (e) after 4-h RT vacuum hold, (f)
after 4-h 60°C. ........................................................................................................................................................... 112
Figure 6-5 Void content as a function of time during debulk for (a) PW and (b) UD laminates ............................... 112
Figure 6-6 Transverse permeability as a function of time at (a) room temperature; (b) 40°C; (c) 50°C; (d) 60°C; (e)
70°C for PW laminates; (f) Transverse permeability as a function of temperature for UD laminates. ..................... 116
Figure 6-7 (a) Measured in-plane permeability as a function of degree of impregnation. In-plane permeability versus
time during debulk for (b) PW and (c) UD laminates. The cure cycle used in the model was 2°C/min to designated
temperatures and hold for four hours. ........................................................................................................................ 118
Figure 6-8 Comparison air evacuation time between room temperature vacuum hold and 60°C vacuum hold. Air
evacuation time vs distance from edge breathing. Laminate length is 2 times of the distance, assuming edge breathing
is applied on all sides and in-plane permeability is the same on all sides.................................................................. 121

IX

Abstract
Vacuum bag-only (VBO) processing allows the manufacturing of high-performance
composite in a more flexible and cost-effective way by using atmospheric pressure alone to
consolidate composite parts in conventional ovens. However, in the absence of high consolidation
pressure, such as that imparted by traditional autoclave processing, VBO-cured parts are more
susceptible to voids, which are known to have a negative effect on mechanical properties. To
promote air evacuation, VBO prepregs feature a partially impregnated microstructure, consisting
of dry fiber tows surrounded by resin-rich regions. Those dry fiber tows provide gas evacuation
pathways during the early stages of processing. However, the mechanisms of air evacuation and
void formation during VBO processing are not well understood. The purpose of this work is to
improve the understanding of void formation and evolution during VBO processing, specifically
for those gas-induced voids.
Gas-induced voids are typically located in the resin-rich regions between prepreg plies.
The primary sources of gas-induced voids are air entrapped during layup and evolved moisture.
Due to the scarcity of experimental techniques to effectively characterize inter-ply entrapped air
during processing, insights into the removal of this inter-ply air under vacuum bag cure conditions
is limited. In this study, an in situ visualization technique was developed to dynamically observe
inter-ply air removal during the cure of an VBO prepreg. Laminates were laid up and bagged on
the interior side of a transparent oven window, thereby re-creating a standard VBO cure
environment and allowing direct observation in situ. A perforated resin film with controlled pore
size and distribution was placed between the glass window and the first prepreg ply, intentionally
introducing entrapped air into the lay-up. In this configuration, air is entrapped in resin rich region
X

mimicking the conditions encountered by an air bubble trapped at the mid-plane of a stack of
prepreg plies.
The technique was first used to investigate mechanisms of air removal and void evolution
in unidirectional prepreg. Bubble transport and the progression of gas evacuation was monitored
throughout the VBO cure of the UD prepregs. Void content as a function of time and temperature
was determined. Prepreg impregnation was also tracked by the inspection of laminate cross-
sections prepared at different times during the cure cycle. Cure kinetics and resin infiltration during
cure were predicted using existing models. Based on the relationships between void content, resin
properties, and tow impregnation as a function of time, a three-stage air removal mechanism was
established. These three stages include (I) air evacuation through dry fiber pathways, (II) bubble
expansion after prepreg saturation and moisture diffusion, and (III) bubble shrinkage due to
evolving thermochemical conditions and resin properties. Furthermore, bubble morphology and
bubble motion were also investigated. A positive correlation was observed between the rate of
evacuation and bubble elongation.
After that, we performed a parametric study to determine the mechanisms of inter-ply void
evolution in UD prepregs VBO cure, and to identify key factors that affect inter-ply air removal.
We employed the same in situ visualization setup for direct, real-time observation of air removal
for prepregs during cure under different conditions. Processing parameters including room
temperature vacuum hold, reduced vacuum and moisture levels were investigated. Results showed
that laminates cured with a 24-hour room temperature vacuum hold exhibit similar bubble behavior
and void content with the laminate without vacuum hold, indicating that room temperature vacuum
hold is not an effective way for the removal of inter-ply air in UD prepregs. Reduced vacuum
XI

quality (80% vacuum) had negligible effects on part quality, while an increase in moisture content
of the laminate notably increased void content.
Resin film thickness change was investigated by the inspection of the cross-sections of
partially cured laminates. The results showed that resin film thickness decreased from ~ 50 µm to
~10 µm at the end of cure due to resin infiltration. However, despite film thinning effects, air
bubble size (porosity) increased ~36% during Stage II, indicating that bubble expansion did occur
and was an important mechanism of inter-void evolution during VBO processing. Prepreg
moisture content was also tracked by the inspection of laminate water content at different times
during the cure cycle, and the data was combined with a diffusion-based analytical model to predict
void size and to improve the understanding of void evolution mechanisms. Results indicated that
moisture content of the laminate decreased markedly as cure progressed, providing insights into
bubble behavior (expansion and shrinkage) observed during cure. The modified model predictions
aligned with experimental data, especially during the second stage, confirming that the observed
void growth results from moisture diffusion.
Later, the in situ monitoring technique was employed to observe inter-ply air evolution
during vacuum bag-only cure of fabric prepregs. Because of the intrinsic surface topography of
woven fabric and prepreg, large amount of air is inevitably entrapped between adjacent prepreg
plies during layup. The in situ monitoring data showed the three-stage process of air removal was
observed during the cure of fabric prepregs, indicating that the key air removal mechanisms in
fabric prepregs remained the same. However, during the first stage, air evacuation was
significantly faster than that in UD prepreg due to the macro-pores (pinholes) located at the
intersections of fiber tows.
XII

A parametric study was also conducted to understand the effects of processing parameters
on air removal in fabric prepregs. The results showed that room temperature vacuum hold can
effectively evacuate inter-ply air in fabric prepregs. Besides, because of the complete evacuation
of inter-ply air, no bubble formation or expansion was observed throughout the subsequent
elevated cure. Moisture effects were also examined and an increasing in void content with
increasing moisture was observed.
A detailed investigation of the effects of reduced vacuum on inter-ply air removal in fabric
prepregs and the underlying void formation mechanisms was performed. Observations showed
that reduced vacuum levels resulted in inefficient inter-ply air evacuation during Stage I, a more
rapid bubble expansion rate and formation of new air bubbles during Stages II and III. These last
two factors governed the generation of high levels of inter-ply porosity in the final parts. Tow
impregnation during cure at reduced vacuum conditions was investigated. Both model predictions
and experimental data showed that resin infiltration was impeded due to the presence of air in
intra-tow regions. However, the final parts showed that tows were fully impregnated in all cases,
indicating that the entrapped intra-tow air migrated to inter-ply regions during cure. The
interactions between intra-tow and inter-tow air bubbles during cure at deficient vacuum
conditions were revealed. Findings led to the conclusion that air remaining in intra-tow regions
contributed more to the increase of inter-ply voids than the reduction in consolidation pressure
associated with reduced vacuum.
Finally, the effects of heated debulk on air evacuation was investigated. The above research
indicated air removal prior to cure is critical for limiting porosity during vacuum bag-only (VBO)
processing. However, room temperature debulk is inefficient, especially for large and/or complex
parts. Thus, the effects of pre-cure dwell temperature on air evacuation in both plain weave (PW)
XIII

and unidirectional (UD) prepregs were investigated. The in situ observation showed that increase
dwell temperature promotes inter-ply air evacuation. Through-thickness permeability increased
with increasing temperature and decreased with number of plies. The in-plane permeability
decreased during pre-cure dwell at elevated temperature due to tow impregnation. The findings
provide guidelines for cure cycle optimization. For PW laminates, air evacuation during debulk at
60°C was more rapid than at room temperature (RT) debulk if laminate length/ thickness ratio >
5.5. For UD prepregs, a pre-cure dwell at 60°C can be more efficient than RT debulk for laminates
(< 1 mm), although thicker UD laminates (> 8 plies) showed no detectable transverse gas flow,
even at 60°C.
Overall, the work presented here provides an in situ monitoring method that can be widely
applicable to investigated air removal mechanisms under various conditions. The insights gained
using this method leads to an improved understanding of the mechanisms of void formation and
evolution in VBO processed parts. The results of this investigation provide scientific basis for
process optimization and the development of comprehensive models for VBO prepreg processing.

1

Chapter 1 Introduction
1.1 Motivation
Composite materials offer a new paradigm for aerospace manufacturing. Carbon fiber
reinforced polymer (CFRPs) composites have many advantages over traditional metallic material
such as lightweight, higher specific mechanical (i.e. strength to weight ratio) properties, fatigue
and corrosion resistance, and design versatility [1]. With these properties, CFRPs offer
opportunities for improved fuel efficiency and reduced greenhouse gas emissions. Besides,
traditionally, large metallic parts are built up from numerous smaller pieces using fasteners which
can account for as much as 50% of the total airframe cost, while composites manufacturing reduces
the need for joining operation by allowing manufacturing of larger monolithic parts, which can
lead to significant reduction of time and expense.
In recent decades, composites applications in aerospace industry have expanded rapidly,
from small secondary structure to more demanding flight-control surfaces, empennage assemblies,
fuselage and wings. For example, the Boeing 787 Dreamliner aircraft are composed of 50%
composites material by weight (80% composite material by volume) and 20% aluminum. In
comparison, the previous 777 model was composed of only 12% composite material and 50%
aluminum by weight [2]. Additionally, the demand for composites use in aircrafts is expected to
quadruple in the next decade. However, such pervasive use comes into conflict with the limitations
of traditional manufacturing methods.
Currently, autoclave processing is the main processing method for the production of
structural composites in the aerospace industry. The composite parts typically begin with layers
of prepregs - carbon fiber beds pre-impregnated with uncured resin, which are stacked on a tool to
form a laminate driven by end-use application. Then the laminate is enclosed in a vacuum bag
2

assembly and placed in a pressure vessel called autoclave for processing [3]. During autoclave
processing, the pressure within the bag is decreases from the ambient towards vacuum while the
pressure within the autoclave is increased to 5-6 atm, which compacts the laminates to the desired
thickness and shape and suppresses porosity within an acceptable range.
Autoclave processing is a robust and well understood manufacturing method [1,4,5].
However, autoclave processing is also expensive and inefficient. Limitations of autoclave
processing include high capital investment, high operating cost, poor energy efficiency, and long
cycle time. Besides, part size is restricted by the inner diameter of autoclave vessel whereas large
autoclaves must sometimes be inefficiently used for small parts. As it will be impossible for
autoclave processing to meet market demands in a cost-effective way, there is significant interest
in alternative processes that allow the production of comparable quality parts in a more efficient
and cost-effective manner outside of the autoclave.
1.2 Problem statement
In order to meet the growing demand, a new generation of out-of-autoclave (OoA)
processing method has been introduced. Vacuum bag-only (VBO) processing is one such approach,
which allows the manufacturing of composite parts using conventional, unpressurized ovens rather
than pressurized autoclaves. By taking the process out of autoclave, faster production with lower
capital and processing cost are possible. Besides, some high-pressure induced defects such as
honeycomb core crush can also be avoided. Question remains, however, as to the ability to produce
void-free primary structural parts out-of-autoclave and the true economic benefit of VBO cure [6].
As the maximum pressure differential during VBO processing is only 0.1 MPa (1 atm),
VBO composite parts are more susceptible to defects such as voids, which are known to have a
negative effect on the mechanical properties. Therefore, a thorough understanding of void
3

formation and evolution mechanism and the relation between material properties and processing
parameters on defect formation is required.
1.3 Scope of dissertation
Without the “safe guard” of high pressure to suppress void into solution, VBO cured parts
can sometimes exhibit high porosity, which is detrimental to mechanical properties. However, the
mechanisms of void formation and evolution are not well understood. The goal of this work is to
develop an improved understanding of void formation and development during the cure of VBO
prepregs and to establish science-based void mitigation strategies.
The main body of this thesis progresses as follows. The present chapter introduces and
motivates the topic. Chapter 2 reviews relevant background, theory, literature, and identifies areas
where further study is needed. Chapter 3- 6 present the main findings of the thesis. As a starting
point, an in situ monitoring technique was developed to observe the inter-ply void evolution during
cure. The details of the experimental set up is shown in Chapter 3. The inter-ply void formation
and removal in unidirectional (UD) prepregs is also discussed, as well as bubble behavior and air
removal mechanisms in Chapter 3. Chapter 4 examines the influence of various process parameters
on void evolution in UD prepregs and developed a model to predict bubble size during cure. In
Chapter 5, a study of void evolution in fabric prepregs and the effects of processing parameters on
void removal is presented. Chapter 6 investigates the effects of dwell temperature on air evacuation.
The in-plane and through-thickness permeability are measured at various temperature, along with
comparison between vacuum hold at room temperature and intermediate temperatures (40°C -
70°C). Chapter 7 summarizes the key contributions and further work.

4

Chapter 2 Literature Review
Prepregs are composite materials in which the reinforcement fiber/fabric is pre-
impregnated with a thermoplastic or thermoset resin matrix. Prepreg is the most common form of
raw material used in manufacturing of high performance structural composites due to the
controlled resin content and good fiber alignment. Prepregs are usually fabricated by hot-melt
impregnation by applying the desired quantity of resin as a film to one or both sides of the fiber
bed, and using heated compression roller to achieve the desired degree of impregnation. Individual
prepreg layers are typically between 0.1 mm and 1 mm thick with resin content of approximately
30% to 40% by weight.
2.1 Vacuum bag-only prepregs
Autoclave processing uses high applied pressure to suppress voids and keep volatiles
dissolved in the resin to make void-free parts. During vacuum bag-only processing, however, the
maximum pressure gradient is only 1 atm, which is not sufficient to suppress voids and volatiles.
Thus, the key requirement for VBO processing of low porosity parts is the removal of gas
entrapped in the prepregs before gelation. The design of VBO prepreg is inspired by the TLP (thick
laminate prepreg) technology [7,8]. In the 1980s, thick autoclave-cured parts required time-
consuming debulking to achieve low porosity. Thorfinnson and Biermann first investigated the
relationship between prepreg impregnation and porosity. The result indicated that dry areas in
prepreg plies allow the evacuation of entrapped air, vaporized moisture and volatiles [8]. After
over a decade, the first generation VBO prepregs were designed and first introduced by Repecka
and Boyd [9], utilizing TLP technology.
VBO prepregs feature a partially impregnated microstructure, typically containing resin
rich areas at the top and bottom surfaces while leaving a dry fiber channels at the center of each
5

prepreg plies, shown in Fig. 2-1. The unimpregnated dry regions, called “engineered vacuum
channels” or “EVaCs”, allow gas transport and removal towards the laminate boundaries.

Figure 2-1 Microstructure of uncured unidirectional VBO prepregs.
The properties of resin system influence the void formation and evolution of VBO prepreg
microstructure during cure. VBO prepregs use addition-cured thermoset resins, which do not off-
gas during the cure process [10,11]. Epoxy-based resins are commonly used in aerospace industry
due to its superior thermal and mechanical properties and ease of processing [12–14]. VBO prepreg
resin systems consist of reactive components (epoxy resins and cure agents) and additives
(tougheners, accelerators, and flame retardants). The irreversible reaction between epoxy and
amine-based cure agent is step growth polymerization. Epoxy groups react with primary (1°) and
secondary (2°) amines and anilines, producing hydroxy groups to form larger molecules, and
eventually, forming a cross-linked network. Resin behavior is primarily affected by the
temperature profile imposed. This reaction can happen slowly at room temperature and can be
much faster at elevated temperature. The average out-time for VBO prepreg is 14-20 days while
the shelf life is usually 1 year. A typical cure cycle for VBO prepreg is shown in Fig. 2-2 [15],
which includes two heat-up ramps and isothermal dwells. The first hold at 121°C allows the epoxy
Dry fibers (EVaCs)
Resin rich area
6

and amine reaction, while the additional post cure at 177°C allows the etherification to form the
3-D network that maximizes glass transition temperature (Tg).
Degree of cure and resin viscosity are two key factors that affect resin flow and compaction
behavior during cure. Initially, resin is in liquid state and as the polymerization occurs, it will
eventually change to a solid state. The degree of cure (0 ≤ α ≤ 1) represents the extent of the
polymerization reaction, which can be characterized using differential scanning calorimetry (DSC)
by monitoring the heat of reaction as a function of time under different temperature conditions.
The rate of curing reaction dα/dt, is a function of both temperature T and α. Kratz et al. [16]
combined the nth order rate equations [17], the Kamal-Sourour model [18] and Cole model [19,20],
to develop a semi-empirical cure kinetic models for commercial materials. The equation is
expressed as follows:
𝑑𝛼
𝑑𝑡
= 𝐾 1
𝛼 𝑚 1
( 1 − 𝛼 )
𝑛 1
+
𝐾 2
𝛼 𝑚 2
( 1 − 𝛼 )
𝑛 2
1 + exp ⁡ ( 𝐷 ( 𝛼 − ( 𝛼 𝐶 0
+ 𝛼 𝐶𝑇
𝑇 ) ) )
(2-1)
where Ki is the Arrhenius temperature dependency:
𝐾 𝑖 = 𝐴 𝑖 exp (
− 𝐸 𝐴𝑖
𝑅𝑇
) (2-2)
In these equations, D is a diffusion constant; 𝛼 𝐶 0
is the critical degree of cure at absolute
zero; 𝛼 𝐶𝑇
is the latter’s increases with temperature; Ai, mi and ni are constants; R is the universal
gas constant; 𝐸 𝐴𝑖
is the activation energy of the resin;
The evolution of resin viscosity µ is governed by two competing effects: the thermal effects
and the crosslinking of the resin. Increasing temperature lowers viscosity due to a higher molecular
mobility, but accelerates cure as well, while increasing crosslinking of resin results in a higher
viscosity as longer molecular chain lowers the mobility. Fig. 2-2 shows a typical evolution of
viscosity during cure. Initially, resin viscosity remains at a relatively high state at room
7

temperature. As temperature increases, the thermal effect is predominant and the resin viscosity
decreases, while the degree of cure increases. When the temperature approaches the isothermal
dwell, the degree of cure becomes predominant, resin viscosity gradually increases until gelation.
Experimentally, resin viscosity can be obtained by thermo-rheological tools (e.g. parallel
plate rheometers) both under isothermal and dynamic temperature conditions. The commonly used
resin viscosity model was originally developed by Castro and Macosko [21] and was later modified
by Khoun et al. [22]:
𝜇 = ⁡ 𝜇 1
+ 𝜇 2
(
𝛼 𝑔𝑒 𝑙 𝛼 𝑔𝑒 𝑙 − 𝛼 )
𝐴 + 𝐵𝛼 + 𝐶 𝛼 2
(2-3)
where µi is the Arrhenius temperature dependency:
𝜇 𝑖 = 𝐴 𝜇𝑖
exp (
𝐸 𝜇𝑖
𝑅𝑇
) (2-4)
Dibenedetto Model are commonly used to predict the glass transition temperature:
𝑇 𝑔 − 𝑇 𝑔 0
𝑇 𝑔 ∞
− 𝑇 𝑔 0
=
𝜆𝛼
1 − ( 1 − 𝜆 ) 𝛼 (2-5)
In the above equations, ⁡ 𝛼 𝑔𝑒 𝑙 is the degree of cure at gelation; A, B, C and 𝐴 𝜇𝑖
are constants;
𝐸 𝜇𝑖
is the viscosity activation energy; 𝑇 𝑔 is the glass transition temperature corresponding to a
degree of cure α; 𝑇 𝑔 ∞
and 𝑇 𝑔 0
are the glass transition temperatures for the fully cured and uncured
resin, respectively; and λ is a shape parameter used as a fitting constant. These constants can be
determined using rheometers and the numerical values of all the constants for CYCOM 5320-1
resins are provided in [23].
This semi-empirical model shows that resin viscosity (and thus resin flow during cure) is
highly dependent on the cure cycle temperature. In principal, the cure cycle must be designed to
maximize flow time at low viscosity (< 100 Pa) to allow resin infiltrates the dry fiber tows.
8

Figure 2-2 A typical cure cycle for VBO prepreg processing [15].
Compared to the literature on resin properties, researches on fiber beds used in VBO
prepregs are scarce as the fiber properties and fiber bed architectures are usually standardized [6].
The carbon reinforcement used in composites are either in the form of unidirectional (UD) tapes
stitched using nylon threads in the transverse direction or fabrics bundled by tows with various
architectures [1]. UD tapes give a higher strength in the fiber direction while fabric prepregs are
easier to handle. The diameter of a single carbon fiber is approximately 7 µm and a tow may consist
of 3K to 24K individual fibers.
2.2 Vacuum bag-only processing
Vacuum bag-only processing generally consist of three manufacturing steps: lay-up,
vacuum bagging and cure. Lay-up is a process to align multiple prepreg plies on a tool in a specific
order and orientation to form a laminate, which can be done both manually and automated. Several
issues many arise during lay-up, including air entrapment between plies and room temperature
resin polymerization, both of which may be detrimental to final part quality.
9

Vacuum bagging encloses the laminate within an assembly of consumables designed to
ensure adequate cure. A traditional vacuum bag assembly is shown in Figure 2-3, which consists
of a metallic tool plate, covered with a fluorinated ethylene propylene (FEP) release film or coated
with a release agent to prevent resin adhesion; the laminate; a second release film, sometimes
perforated to allow resin and/or gas transfer; breather cloths that allow gas transfer and pressure to
equilibrate within the bag; edge breathing dams that maintain the laminate in place while allowing
gas to escape through the sides, and an impermeable vacuum bag with a vacuum valve that creates
a pressure differential with the exterior.

Figure 2-3 Schematic of vacuum bag assembly [24].
Finally, the vacuum bag assembly is placed in a conventional oven, where temperature and
pressure cycles are applied to cure the laminate. A traditional temperature cycle for an aerospace
epoxy consists of a combination of heating ramps (0.5 °C/min to 3°C/min) and isothermal dwells
(93°C-121°C) with dwell times on the order of hours, and sometime a free-stand post cure at
~177°C [15]. Curing is the key stage of manufacturing as it includes the thermochemical changes
within the resin and the defect-governing consolidation phenomena within the laminate.
10

Due to the special structure of VBO prepregs, the phenomena occur during VBO
processing are quite complex, which involve consolidation, air evacuation, resin flow and
compaction [6].
2.2.1 Consolidation
Figure 2-4 shows the consolidation process of a 2-ply VBO prepreg laminate. Initially, the
laminate consists of dry fiber tow regions surrounded by resin rich regions. During the lay-up, air
can be trapped in both intra-tow dry fiber regions and between prepreg plies. At the beginning of
VBO cure, when vacuum is applied, fiber bed is compacted, while fiber tows remain partially dry,
air can be evacuated via edge-breathing through the dry fiber tows. Once temperature is raised,
resin gradually infiltrates the dry fiber tows ideally to produce a void-free composite parts.

Figure 2-4 Schematic of the OoA prepreg consolidation process [6].
2.2.2 Resin flow and compaction
Resin infiltration, fiber bed compaction and the evolution of resin flow during cure have
been studied extensively using various methods. Thomas et al. [10,25] investigated the through-
thickness resin infiltration using ultrasound under both isothermal and dynamic conditions (Figure
2-5a). Cender et al. [26] monitored resin flow of a one-sided prepreg (resin film on one side of the
prepreg) during vacuum bag-only processing over a transparent tool and developed an analytical
11

model to characterize the dual length scale permeability (Fig. 2-5b). Both results indicated that
resin flow occurs over dual length- and time- scales, first flowing into larger inter-tow macro-space
regions, then into smaller intra-tow micro pores.

Figure 2-5 (a) Ultrasound scans of resin impregnation as a function of time at 80°C [10]; (b)
images of resin impregnation of ST95-RC 200T prepreg at 55°C [26].
Tow impregnation in the cross-section direction were also investigated. Centea et al. [27]
and Farhang et al. [28] partially cured the laminates to different stages and then used micro-CT
and optical microscope to observe the resin impregnation state at the cross-section of laminates.
Both results showed the partially dry microstructure of the prepreg at the initial state and identified
that tow impregnation is the key phenomenon during the cure process. Hu et al. [29] and Farhang
et al. [30] also measured thickness change during cure to understand the fiber bed compaction and
the resin infiltration process. The in situ thickness measurement revealed two major decreases in
laminate thickness, one of which was mainly due to the fiber bed compaction while the other one
was attributed to tow impregnation.
12

Figure 2-6 X-ray micrographs pre-treated with different cure cycle [27].
2.2.3 Air evacuation
Air evacuation and gas flow in VBO processing has been studied in the context of prepreg
permeability. Permeability is a material property that describes the capacity of a porous medium
to accommodate gas/fluid flow, which depends on geometrical characteristic such as porosity,
tortuosity and surface area [31]. Air permeability has been studied extensively in the field of soil
science, oil, and gas extraction, and was latter applied to composites field, where fiber bed is the
porous medium in nature. The intrinsic permeability is determined solely by the solid matrix and
is independent of fluid flow properties [32]. However, VBO prepregs is a fibrous structure partially
impregnated with resin and resin infiltrate into dry fiber tows during cure. The effective
permeability of VBO prepregs is thus affected by fiber arrangement, cure cycle, pressure applied,
flow properties and directions, etc.
13

Permeability can be measured using both steady state and non-steady state (falling pressure)
methods. The non-steady state measurement is commonly used for low permeability materials
where steady state cannot be achieved. Permeability, K, is determined by the 1-D Darcy’s law,
which describes the relationship between volumetric flow rate, Q, and the pressure gradient across
the sample at constant temperature: [33]
Q = ⁡ −
𝐾𝐴
𝜇 𝑑𝑃
𝑑𝑥
(2-6)
where K is the permeability scalar in the flow direction, which has a unit of m
2
, A is the cross-
section area of the sample, µ is the viscosity of the fluid, and dP/dx is the pressure gradient at
position x. Gas flow experiments with low gas pressure or low permeability media can
overestimate flow velocity due to the slip phenomenon known as the Klinkenberg effect [34].
Studies [35] demonstrated that slip flow is more pronounced when the porous media has a pore
size of 10 um, the same order of magnitude as the pores in prepregs. The Klinkenberg effect is
expressed [36]:
𝐾 𝑔 = 𝐾 𝑙 ( 1 +
𝑏 𝑃 ) (2-7)
where Kg is the gas permeability at a given pressure P, Kl is the intrinsic or liquid permeability,
and b is a constant known as the Klinkenberg factor. As Klinkenberg effects is included in the
effective permeability, it can be neglected if the intrinsic permeability is not required.
Prepregs are anisotropic, thus the gas flow measurements are measured in both in-plane
[33,37–39] and transverse [40–42] directions shown in Table 1. In-plane permeability is
commonly measured using steady-state method while transverse permeability is measured using
falling pressure method as the permeability in the through-thickness direction is significantly lower.
Louis et al. [39] measured the in-plane and transverse permeability of a VBO prepreg (MTM 45-
14

1/5HS) at room temperature and the results showed that in-plane permeability (~ 10
-14
m
2
) is 4
orders of magnitude greater than transverse permeability (~ 10
-18
m
2
). The data also indicated that
transverse permeability shows non-Darcian behavior, which would eventually decrease to zero
when ply number increases. Arafath et al. [33] measured the in-plane permeability of Toray 3900
UD tape and the results indicated that unidirectional materials had greater in-plane permeability in
the fiber direction (0°) than the transvers direction (90°), at 10
-14
m
2
vs 10
-15
m
2
, respectively, while
for fabric prepregs, comparable in-plane permeability values were obtained in both directions.
Permeability during process has also been reported in literature. Fahrang et al. [28]
determined the in-plane permeability of MTM 45-1 prepreg during cure and the results showed
that in-plane permeability decreased to zero as the tows were impregnated by resin. Tavares et al.
[42] and Kratz et al. [40] measured the through-thickness permeability of VTM 264 and Cycom
5320, respectively, at both room temperature and cure conditions using falling pressure method.
Both results show an increase in transverse permeability with increasing temperature due to
decreased resin viscosity. These studies indicate that air evacuation capacity during processing is
based on both resin and prepreg properties.
Models were developed to describe gas transport in prepregs. Arafath and Fernlund et al.
[43] derived equations based on Darcy’s Law and the ideal gas law to predict the time required to
evacuate a given mass fraction of gas in one dimension:
𝑡 =
𝜇 𝑃 0
∅ 𝐿 2
𝐾 [ −
1
0 . 9
ln (
𝑚 𝑚 0
) ]
1
0 . 6
(2-8)
where L can be the length to breathing edge (half the laminate length), m/m0 is the mass fraction
of gas remaining in the laminate, P0 is pressure, µ is gas viscosity and ϕ is the porosity. The
equation shows that evacuation time increases quadratically with part length and decreases linearly
15

with permeability, indicating that air evacuation time increases dramatically when part size
increases.
Table 2-1 In-plane and transverse permeability at room temperature reported in the literature [6].
Direction Resin system Fiber bed Permeability (m
2
) Reference
In-plane MTM 45-1 5HS 3.3 × 10
-14
[28,39]
5320 5HS 8.5 × 10
-15
[40]
5320 PW 7.3 × 10
-15
[40]
Toray 3900 UD 2.0× 10
-14
[33]
Transverse MTM 45-1 5HS 1.0 × 10
-18
[39]
5320 5HS 1.6 × 10
-17
[40]
5320 PW 6.4 × 10
-17
[40]
VTM264 UD 5.0 × 10
-18
[42]
Toray 3900 PW 3.5× 10
-16
[33]

2.3 Voids in composite processing
Voids are the most studied manufacturing defects, as it is a critical factor restricting
composite acceptance for aerospace applications. Voids are defined as empty spaces that are not
filled by resin or fiber within a composite part shown in (Figure 2-7). For aerospace applications,
the acceptable void content is typically less than 1-2 %. Autoclave processing utilizes high pressure
to alleviate porosity, however, VBO pressure is not sufficient to suppress air during process. Thus,
void formation and evolution under VBO processing requires a more thorough understanding.

Figure 2-7 Void in a composite laminate.
16

2.3.1 Void formation
The mechanisms of void formation and growth during cure in VBO prepregs are not yet
fully understood. The primary sources of voids in VBO prepregs are air entrapment during lay-up,
volatiles released from resin during cure, and insufficient resin flow during cure [33]. Depending
on void sources, voids can be classified into two categories: gas-induced voids (due to the first
two sources) and flow-induced voids (due to the last source) [6]. Because of the special structure
of VBO prepregs, air inherently exists in the dry fiber regions. Besides, air can be trapped between
prepreg plies during lay-up. If entrapped air is not removed before gelation, it will remain as voids
in the cured composite parts. Gas-induced voids can also form due to the evolution of volatiles
during cure. As VBO prepregs are formulated to have little volatiles, moisture absorbed from air
becomes the main source [44]. During composite processing, if the released moisture remains in
the prepreg system, bubbles will form within the resin, and potentially, result in a void within the
final part. As water vapor pressure increases sharply with increasing temperature, when gas
pressure of a bubble exceeds its surrounding resin pressure, it will expand, otherwise, it will shrink.
Moisture-induced void growth models [45–48] have developed to predict the formation
and evolution of voids. Most of these models are based on Kardos’ pioneering work [45–50],
which relies on the assumption that voids grow via water transfer from an infinite resin medium
to discrete bubbles [45]. This model was developed initially to predict void nucleation and growth
for autoclave processing and was later applied to out-of-autoclave processing. As it did not account
for the surface tension of the resin, the predicted void size increases unrealistically with
temperature. Wood and Bader included surface tension in their framework, but concluded that its
effect was noticeable only for small air bubbles at the early stage of void nucleation [46]. Ledru et
al. coupled visco-mechanical effects and a diffusion model to refine void size prediction [47]. The
17

coupled model more accurately predicted the final void size but still overestimated the value by
orders of magnitude compared to experimental data, a discrepancy which is likely due to the
assumptions of the constant infinite concentration of moisture in the resin and constant diffusivity
during cure [48].
Flow-induced voids generally result from insufficient impregnation of dry fiber tows
before resin gelation. Studies indicated that cure cycle temperature and resin initial degree of cure
are the two key factors that affect resin impregnation process during cure [51,52]. The combination
of fiber-dense tows, low dwell temperature and high initial degree of cure can lead to incomplete
impregnation and pervasive micro-voids [6].
2.3.2 Effects of processing parameters on voids
The effects of processing parameters have been studied extensively, most of which focus
on the understanding of the correlations between processing parameters and porosity in the final
part. Processing pressure and moisture are two most studied parameters. Olivier et al. [53]
investigated the influence of autoclave pressure on void content of composites and concluded that
void-free parts could be produced when pressure exceeded 0.7 MPa and lower cure pressure would
lead to the formation of ellipsoidal voids.
Centea and Hubert [54] investigated the effects of reduced ambient pressure, reduced
vacuum quality and restricted air evacuation on the part quality of VBO-cured laminates using
both PW and UD prepregs. The results showed that reduced ambient pressure results in quasi-
linear increase in macro-voids in resin-rich regions due to a lower ratio of resin pressure to void
pressure. However, while reduced vacuum quality also resulted in comparable macro-voids, it also
led to micro-porosity because of the existence of air.
18

Grunenfelder and Nutt [55] investigated the relationship between moisture and void
content by exposing prepreg to elevated ambient relative humidity conditions. The results showed
that void content increased exponential with increasing moisture weight percentage in the resin
during VBO processing, while the autoclave cure laminates exhibited low porosity for all humidity
levels.

Figure 2-8 (a) Average macro- and micro- porosity versus process conditions; (b) Micrographs
of 5320/8HS and 5320/UD laminates [54]; (c) Void content versus relative humidity [55].
Fernlund et al. [56,57] studied on the influence of ambient relative humidity, vacuum
quality, debulk time and part length on porosity and determined that increased ambient moisture,
reduced vacuum levels, shorter room temperature vacuum hold times and larger part sizes would
increase void content in composite parts attributed to either lager and more entrapped air bubbles
or more difficult air evacuation. They also found for a long part, no porosity was observed if the
laminate was exposed to 0% relative humidity or had a long debulk (24h), which confirmed that
19

moisture and entrapped air were two dominant causes of gas-induced voids. Anderson and Altan
[58] also investigated the coupled effects of moisture and (above ambient) pressure. The results
led to similar conclusion and an enhanced void formation model was proposed.
Flow-induced voids, as mentioned before, are primarily caused by high initial degree of
cure (out-time). Kim et al. [23,59] investigated the out-time effects on resin cure kinetics and
viscosity by using DSC and in situ dielectric method and a cure kinetic model was developed to
predict degree of cure and viscosity with any given temperature and out-time profiles. The results
showed that long out-time would result in high minimum viscosity. Centea and Hubert [51]
developed a simple analytical model for tow impregnation and used the model in a parametric
study to evaluate the effects of material and processing parameters on tow impregnation. The
results indicated that tow impregnation is strongly dependent on the resin viscosity profiles,
therefore long prepreg out-time would impede resin infiltration.
Grunenfelder and Nutt [60] pre-conditioned prepregs with zero to eight weeks of out-time
and compared the cured laminate quality after a single temperature cycle, and the same model was
used to predict the degree of impregnation during cure. Both model and experiments showed that
voids began to form in the intra-tow regions when the out-time exceeds 21 days (manufacturer
specified out-life). Lucas et al. [61] also examined the effects of out-time on porosity in laminates
produced form Cycom 5320 system and the results showed an increasing in porosity with out-time.
Centea et al. [62] later found that rapid and high temperature cure cycles promoted faster
tow impregnation and could mitigate or even eliminate flow-induced micro-porosity due to long
out-time. Besides, Hamill et al. [63] investigated the effects of out-time on surface porosity and
concluded that out-time (within out-time range) reduced prepreg tack and can thus improve the air
evacuation capacity and reduced surface porosity.
20

Figure 2-9 Porosity as a function of out-time in a VBO prepreg [60].
2.3.3 Void characterization
Different methods have been developed for void characterization of composites from
simple density determination methods to advanced X-ray CT, depending on the needs/ requirement
of the study. A summary of commonly used void characterization techniques are listed in Table 2-
2. The density method [64,65] is the only standard method for porosity measurement in polymeric
composites. According to ASTM D2734 [66], void content can be calculated by the following
equation:
𝑉 𝑣 = 100 − 𝜌 𝑐 𝑚 (
𝑊 𝑟 𝜌 𝑟 +
𝑊 𝑓 𝜌 𝑓 ) (2-9)
Where Vv is void content, 𝜌 𝑐 𝑚 is measured composite density, W is the weight percentage, ρ is the
density, and the subscripts r and f stand for resin and fiber. This method is simple and fast but can
be used only when void content is required as no morphological information would be offered with
this method.
21

Optical [30,67,68] and electron microscopy are widely used method for the investigation
of the microstructure of composite materials. Void content is usually determined by the area
fraction of voids to the total area of the cross-sections. This method is well-established and
provides high resolution images including information of void size (2D), shape, location and
distribution shown in Figure 2-10. However, the sample pretreatment (cutting, mounting and
polishing) is time-consuming. Besides, due to the soft nature of partially cured prepreg, the
pretreatment could alter surface morphology of the sample.

Figure 2-10 Micrograph of voids in composite prepregs.
Ultrasonic imaging [5,10,69] is the most widely used method in industry for the evaluation
of voids in composites. Ultrasonic method is a nondestructive method and sometimes portable thus
can be used for in-service inspection in industry. Based on ultrasonic wave attenuation and velocity,
this method is usually used for planar samples using either through-transmission mode (traveling
through) or back scattering mode (reflecting). A typical ultrasonic image can provide information
of void content, distribution, planar location, etc., shown in Fig. 2-11. The main drawbacks,
however, are their low resolution (hundreds of microns) and thus low accuracy. Besides, coupling
agents (e.g. water or oil) are required between the material and transducer when conducting tests,
which would have a negative effect for materials that are sensitive to moisture.
22

Figure 2-11 C-scan showing areas with different void contents: (a) cure pressure is 0.0 MPa, (b)
cure pressure is 0.6 MPa, (c) presentation of grey level and its porosity [5].
Micro computed tomography (micro-CT) is a high resolution (several microns) and semi
non-destructive method that produces a 3D presentation of an object by creating cross-sectional
tomographic images. Based on the difference in the absorption and scattering of X-ray by different
materials, micro-CT allows the analysis of the internal structure of an object [70]. Micro-CT
provides rich 3D information on void size, shape, location and distribution shown in Fig. 2-12,
thus it is been more and more widely used in research on defect characterization of composites in
recent decades [27,40,71–73]. Micro-CT is a unique technique to investigate void morphologies.
Studies found that typically voids located in intra-laminar regions are elongated micro-voids while
those located in the resin-rich regions are generally spherical [44]. The main limitations of this
method include long scan and processing times, relatively small sample sizes, and high cost.
The void characterizations method described above are mostly employed on composites
that are fully cured. Recently, transparent glass tooling surfaces have been employed to observe
air entrapment, distribution, and surface porosity in situ [74–77]. For example, Cender et al. [26]
23

used a transparent acrylic table and a CCD camera to monitor the dual scale flow of resin within
fabric prepregs. Gangloff et al. [78] used the same method to investigate resin flow and bubble
motion in fabric prepregs under constant pressure and temperature conditions. Finally, Hamill et
al. [24] also investigated the formation and evacuation of surface porosity of prepregs using a glass
plate and camera.

Figure 2-12 Voids in a composite with UD carbon/epoxy plies characterized by micro-CT,
displayed in (a) top, (b) 3D, (c) front, and (d) side view [79].

24

Table 2-2 A summary of commonly used void characterization techniques [80].
Technique Measurable characteristics Advantages Disadvantages
Density • Void content
• Easy and inexpensive
• Quick
• Standard
• Destructive
• Only void content
Microscopy
• Void content
• 2D shape
• Size
• Location/distribution
• Easy and inexpensive
• 2D morphology
• Relatively quick
• Destructive
• Section-biased
• Location-biased
Ultrasonic
• Void content
• Planar size
• Location/distribution
• Non-destructive
• Portable
• In-service inspection
• Accuracy
• Time consuming
Micro-CT
• Void content
• Size
• 3D shape
• Location/distribution
• Semi non-destructive
• 3D analysis
• Accuracy
• Expensive
• Time consuming
Transparent
Tool
• Void content
• 2D size and shape
• Location/distribution
• Mobility
• In situ
• Bubble motion
• Only surface porosity
• May deviate from
real cure

2.3.4 Effects of voids on mechanical properties
The main interest in characterization and analysis of voids in composite materials is
because of their negative effects on the mechanical properties, particularly those matrix-dominated
mechanical properties. Due to the complex, coupled effects of void size, shape, distribution and
location, the (micro) mechanisms of the influence of voids on mechanical properties are still not
fully established. One of the mostly studied properties is the effects of void on interlaminar shear
stress (ILSS), as it is dominated by the matrix and the fiber/matrix interface. Cantwell et al. [81]
found that every 1% increases in void content resulted in ~ 6% of (almost linear) reduction in ILSS.
Tang et al. [82] also found similar results that when void content increased ~ 7%, the short-beam
25

shear modulus reduced to ~ 50%. Kousourakis et al. [83] investigated the influence of void size
and shape and observed a linear decreases in ILSS with increasing diameter of the elliptical voids.
The studies on the influence of void content on tensile properties showed that the
longitudinal tensile modulus reduced negligible with increasing void content [84] while the
transverse modulus [85] and out-of-plane modulus was more sensitive to voids [86]. Varna et al.
reported ~ 15% decrease in transverse tensile modulus with a void content of 5%. Gürdal et al.
showed that 1% increase in void content resulted in ~ 10% decrease in the out-of-plane modulus.
The effects of void morphology was also investigated and the results showed that width-to-height
aspect ratio have more significant effects on modulus than the length-to-width ratio [87].
The effects of void content on compressive properties has not been widely studied. Some
studies [88] found a negligible influence of void on compressive modulus, while other studies [89]
observed a linear reduction with increasing void content. [90] found that the reduction in
compressive properties is mainly because the presence of free surfaces facilitates fiber kinking/
buckling effects. The influence of void content and/or morphology on other mechanical properties
such as flexural [5,91], fracture toughness [92] and fatigue behavior [93] have also been studied.
Besides, more advanced characterization and calculation techniques such as micro-CT and FEA
modeling have been employed to increase the understand of the influence of void on mechanical
properties.
2.4 Conclusions
The interest in the shift of composite manufacturing from autoclave processing to more
time- and cost- efficient out-of-autoclave processing is increasing due to the demand of high-
performance composites in aerospace industry. Literature on vacuum bag-only processing are
growing, especially after 2005. Table 2-3 shows an overview of literature published on
26

voids/porosity and air evacuation, which leads to several conclusions about the existing knowledge
of vacuum bag-only processing of composite materials. First, the microstructure of vacuum bag-
only prepreg has been examined and confirmed [27,30,55,77]. VBO prepregs consist of partially
dry fiber tows, acting as air evacuation pathways, and surrounded by resin films. Second, a general
understanding of air evacuation, tow impregnation and compaction phenomena has been
established. Several models have been developed to predict degree of impregnation during process
and the formation of flow-induced voids has been investigated [51,60,94]. Finally, the
relationships between key processing parameters (including pressure, moisture, out-time, etc.) and
the final part quality has been established [54,60,74,95].
However, knowledge gaps still exist, especially on the void formation and evolution during
VBO processing. First, most existing research on void characterization only investigate the final
part quality, although some more recent research investigated the tow impregnation and void
evolution during cure by interrupted cure [30,54]. The formation and growth/shrinkage of
individual air bubbles during cure has not been investigated, as well as bubble mobility,
morphology and distribution during cure. Second, the effects of processing parameters on bubble
migration and evolution has not been studied. Literature showed that deficient vacuum level, high
initial degree of cure, low dwell temperature, and/or high humidity environment would increase
porosity of the cured composite parts. However, limited knowledge is available on how processing
parameters affect void evolution during cure. Finally, more precise void growth models are desired
as most current models overestimate void size during cure.
The overall goal of this thesis is to gain a better understanding of void evolution during
VBO processing, thus to provide science-based void reduction strategies and a basis for
comprehensive void models. To achieve this goal, the following objectives were proposed:
27

1. Develop a robust and effective method that is able to observe void evolution during the
cure of VBO prepregs.
2. Determine inter-ply air removal mechanisms and develop models to predict void growth
during VBO processing.
3. Evaluate, through experiments and modelling, the effects of the key processing parameters
on void formation and development during VBO processing.
4. Evaluate the effects of heated debulk on air removal, so as to guide process optimization.
The following four chapters describe the research through which the above objectives are
achieved.

28

Table 2-3 Overview of literature on voids in out-of-autoclave prepregs

1
Literature published after 2013 when this project started.
Principal
Investigator
Void characterization Tow impregnation Permeability (during process)
Effects of processing
parameters on porosity
During
process
After
process
In
situ
During
process
After
process
In-plane
Transverse
During
process
After
process
In
situ
Continuous Interrupted
Nutt [55,60,95] [25] [60] [55,60,95]
Hubert [54] [27,51] [27,51] [40] [54]
Fernlund [30]
1
[74]
1
[30]
1
[30]
1
[39,96] [39,96] [30]
1
[56] [74]
1

Advani [78]
1
[97] [97] [37]
Hinterhӧlzl [77]
1

Michaud [42]
29

Chapter 3 In Situ Monitoring of Air Removal in Vacuum Bag-Only Processing
of Unidirectional Prepregs
3.1 Introduction
VBO prepregs are partially impregnated by design, featuring dry fiber pathways at ply mid-
planes. The dry fiber pathways, termed engineered vacuum channels (EVaCs), enhance air
transport and removal during early stages of VBO cure (Fig. 3-1), before being saturated by
surrounding resin at high temperature. Voids in composites are generally classified into two
categories according to the source of the voids: flow-induced and gas-induced [6]. Flow-induced
voids, also known as tow voids, are located within the dry fiber bundles [51,97,98]. EVaCs
constitute an initial network of porosity in the prepreg. In the absence of ideal material and
processing conditions, resin infiltration into the tows can cease before full saturation. Thus, the
initial porosity can sometimes lead to residual flow-induced voids. Flow-induced voids have been
observed in cured laminates if resin content or resin pressure was insufficient [54], or if the resin
viscosity profile did not allow sufficient flow prior to gelation [60].

Figure 3-1Micrograph and schematic of inter-ply voids.
The focus of the current study is gas-induced voids, which are formed primarily due to
entrapped air during layup or volatiles evolving during cure. Gas-induced voids are
morphologically different from flow-induced voids, being located, in cured parts, primarily within
30

the resin-rich regions around tows and between prepreg plies.[6] Studies have indicated that initial
inter-ply air entrapment correlates with the surface topology and tack of prepregs.[24] During cure,
entrapped gas bubbles will expand or collapse if the gas pressure within the void is greater or less
than the surrounding resin pressure, respectively. With access to an EVaC (or other flow channel),
entrapped gasses can escape the laminate through edge breathing mechanisms. However, if access
to an air removal pathway is occluded, or if the air is not fully removed prior to gelation of the
resin, voids will remain as defects in the finished part. For example, gases located in inter-ply
regions may not reach EVaCs, and must undergo a more complex removal process. For these
reasons, producing void-free parts requires a thorough understanding of gas-induced void
formation and evolution.
The scarcity of experimental techniques to effectively characterize inter-ply entrapped air
has led to limited insight into the evolution and transport of gas-induced voids during processing.
The common void characterization techniques and their advantages and disadvantages have been
introduced in Section 2.3.3. While ultrasound can be conveniently automated and is non-
destructive, the resolution of ultrasound is relatively low (~hundreds of µm) [99]. The drawbacks
of micro-CT include long scan and processing times, relatively small sample sizes, and high cost
[27]. Observation of laminate cross-sections at different time intervals during VBO processing
with light microscopy provides valuable insights into the microstructure of the laminate at specific
points within the processing cycle [30]. However, the soft nature of the partially cured polymer
matrix can introduce artifacts by alteration of void morphology during sample preparation or
removal from curing conditions. Most importantly, none of the methods described above provide
in situ monitoring of inter-ply void evolution during VBO cure.
31

Transparent glass tooling surfaces have been employed in previous studies to observe air
entrapment, distribution, and surface porosity in situ [74–77]. However, most previous studies
focused on void evolution and flow in woven fabric prepregs, and did not replicate the curing
environment of VBO processing, i.e., a vacuum bagged laminate cured in an air circulating oven.
In this study, laminates were laid up and bagged on the interior side of a transparent oven window,
thereby re-creating a standard VBO cure environment while at the same time allowing direct
observation in situ. Additionally, a perforated resin film with controlled pore size and distribution
was placed between the glass window and the first prepreg ply, intentionally introducing entrapped
air into the layup. With these modifications, air evacuation and entrapment between unidirectional
(UD) prepreg plies was investigated throughout the VBO cure process.
In this chapter, we perform a case study to highlight the utility of the in situ method.
Specifically, the study was undertaken to increase understanding of inter-ply void formation and
removal mechanisms in unidirectional VBO prepregs. A thorough understanding of gas-induced
void formation and evolution during VBO processing is needed to guide the design of more
efficient and robust prepreg formats, and to optimize cure processes to minimize porosity in
finished parts. The in situ observation technique developed for this work yielded new insights into
bubble migration, expansion, and removal during VBO cure. Resin impregnation during cure was
also tracked, and links between void evolution phenomena, resin properties, and tow impregnation
were revealed. Finally, experimental observations were used to establish a three-stage void
removal mechanism for unidirectional VBO prepregs.
32

3.2 Experimental
3.2.1 Materials
Experiments were performed using a carbon fiber/ epoxy prepreg formulated for vacuum
bag only cure. The prepreg consisted of a toughened epoxy (CYCOM 5320-1, Solvay, USA) and
a unidirectional tape (IM7 12K, 145 g/m
2
) with 33% resin content by weight. The manufacturer’s
recommended cure cycle for the material at the time of the research specified cure at 93 ºC for 12
hours or 121 ºC for 3 hours. The lower temperature cycle was used in this study, with a ramp rate
of 2 °C/min. Micrographs of the initial condition of unconsolidated prepreg were obtained using
an SEM (JEOL JSM-6610). Prepregs were laid up and accrued out-time at room temperature for
six months, allowing the resin to vitrify prior to sectioning. Both cross-sectional views and surface
images of samples without conductive coating were captured using a back-scatter electron detector
at an accelerating voltage of 20 kV.
Neat resin films were also utilized in this work. The films (areal weight of 91.3±4.7 g/m
2
,
thickness of 49.5±6.5 µm) were composed of the same resin as the prepreg. To create perforated
resin films, holes were punched manually into the film using a coring tool with a diameter of 0.25
mm.
3.2.2 In situ observation of VBO cure
The primary sources of gas-induced voids in composites are air entrapped during the layup
and evolved gases (notably moisture). Entrapped air bubbles are typically located in the resin-rich
regions between prepreg plies, as shown in Fig. 1. To accurately reproduce the conditions for air
removal and gas entrapment described in the introduction, and to allow monitoring of multiple
bubbles in the same field of view, an analogous configuration in which air bubbles are surrounded
33

by resin must be created. This type of artificial inter-ply zone can be produced by incorporating a
perforated neat resin film into the layup.
In this study, a perforated resin film containing holes of controlled size and distribution
was laid up against a glass oven window, and prepreg plies were subsequently laid onto the film.
In this way, air bubbles were intentionally introduced into the lay-up and surrounded by resin
everywhere except at the bubble-glass interface. In general, the configuration used approximated
the conditions surrounding an internal void located at the mid-plane of a stack of prepreg plies.
Furthermore, incorporating voids of known diameter via a perforated film enabled quantitative
measurements of porosity with controlled initial characteristics, as well as visualization of trapped
air migration. In a parallel set of experiments, the potential effects of the bubble-glass interface on
bubble behavior were investigated by creating bubbles fully enclosed in resin. This condition was
achieved by inserting an additional, non-perforated film between the glass tool plate and the
perforated film. Results from these experiments indicated bubble behavior was the same,
regardless of which configuration was used.
The effects of pore size on air removal was investigated. Four bubble sizes (ϕ = 0.25, 0.5,
1.5, 3 mm) were created and bubble behavior during cure was investigated. The results showed an
increase in void content with increasing bubble size and this effect is more pronounced in UD
prepregs than in fabric prepregs. As air bubbles trapped during cure are usually small, the pore
size of 0.25 mm was used in the rest of this study.
To produce test panels, prepreg plies were cut to 127 mm × 127 mm. To visualize voids,
air pockets were introduced via perforated resin films containing holes with a diameter of 0.25 mm
and spacing between holes of 2 mm (Figs. 3-2(b) and (c)). These resin films were 38 ×38 mm.
Each test was repeated 3 times. The moisture content of the as-received prepregs was 0.10 ± 0.01%.
34

Fig. 3-2(a) shows the experimental setup. Laminates were laid up and bagged vertically on
the interior side of a transparent window in a programmable air-circulating oven (Thermal
Products Solutions Blue M). The glass window was treated with liquid release agent (FreKote
770-NC) prior to layup. A layer of perforated resin film was positioned against the glass window.
Four layers of prepreg were stacked on top of the perforated resin film, and the assembly was then
vacuum bagged using standard consumables. The consumable arrangement consisted of a layer of
non-perforated Teflon release film on the bag-side laminate surface, edge-breathing dams made of
vacuum sealant tape wrapped with fiberglass boat cloth, and finally one layer of breather cloth
followed by a vacuum bag. The placement of the resin film and prepregs was controlled to achieve
similar initial void contents (12% to 15%) for each test. Six thermocouples were used to monitor
temperature. Five thermocouples were placed against the Teflon release film, one at each of the
four corners of the laminate and the last at the center, while the sixth thermocouple was located
against the interior surface of the glass window. Time-lapse videos were recorded from outside the
oven, using a portable microscope (Dino-Lite Premier2 Digital Microscope). Video was collected
throughout the cure cycle at a magnification of 20 ×.
Gravity effects can be neglected in this study for the following reasons. First, the Bond
number, Bo, was estimated to be 0.02. Bond number is a dimensionless number that characterizes
the ratio of gravitational force to surface tension.

2
gL
Bo



= (3-1)
where ∆ρ is density difference between epoxy resin (=1310 kg/m3)[15] and air bubbles
(=1.2 kg/m3),[100] g is gravitational acceleration (=9.8m/s2), L is the characteristic length (the
bubble diameter ~0.25mm was used), and σ is surface tension (= 0.035 N/m).[51] A value of Bo
35

<<1 implies a weak dependence on gravitational force. Secondly, bubble movements during cure
were observed to be horizontal (and along the fiber direction) in all tests. Accordingly, resin
movement was also likely to occur in the horizontal direction. Furthermore, no laminate thickness
gradients were observed in the vertical direction after cure, indicating that gravity-driven bulk resin
flow did not occur.

Figure 3- 2 (a) Schematic of in situ observation set-up, (b) Dimensions of perforated resin film,
(c) Micrograph of perforated resin film.
36

3.2.3 Void content
Void content as a function of time was calculated for each test panel using image analysis
software (ImageJ). Ten frames, recorded at different points throughout the cure cycle, were chosen
from the time-lapse videos, and the voids present in the image were manually selected. The images
were then converted to binary, with voids represented in black, and the remaining area in white. A
percent value for porosity was determined by dividing the number of black pixels by the total
number of pixels in the image.
3.2.4 Partial cure sample preparation
To better understand fiber bed compaction, resin impregnation, and the relationship of both
factors to void evolution, laminates were prepared by interrupting the prepreg cure cycle at
different points, as shown in Fig. 3-3. Each laminate prepared by partial cure processing consisted
of 8 plies of prepreg 127 mm × 127mm with a unidirectional [0]8 layup. To preserve the
morphology of the laminates at each point of interest, panels were removed from the oven at the
desired point in the cure cycle while maintaining vacuum and rapidly quenched with liquid
nitrogen to prevent further resin flow. The skin temperature of each laminate was tracked using
thermocouples.
Partially cured laminates were then placed in an ammonia environment at room
temperature for 10 days. Ammonia vapor reacts with epoxy at a low temperature, acting as a curing
agent to achieve a hard and stiff structure [101]. This technique enables the microstructure of
samples to be assessed at various points in the cure cycle without altering the internal structure.
Following ammonia cure, samples were sectioned at the center of each panel, mounted and
polished using a series of graded abrasive papers. Polished sections were imaged using a stereo
microscope (Keyence VH-Z100R).
37

0.0 0.5 1.0 10 12
20
40
60
80
100
Temperature (°C)
Time (h)

Figure 3- 3 Cure cycle with partial cure sampling points marked with black squares.
3.2.5 Water solubility in epoxy resin
To support hypothesized explanations of void growth and collapse during the cure process,
the effect of degree of cure on the water solubility of epoxy resin was investigated. Neat resin
samples were prepared to different degree of cure values following a cure kinetics model
previously developed for the same resin system [23]. Next, samples were humidity-conditioned
under 99% (± 1%) relative humidity for 24 hours at room temperature (21 ± 2°C). 99% relative
humidity was achieved in a sealed container with saturated K2SO4 solution. Modulated-DSC
measurements (TA Instruments Q2000) were performed to verify the degree of cure of each sample
after humidity conditioning. Moisture content of each sample was measured by Fischer titration
using a coulometric titrator (Mettler Toledo C20 with D0308 drying oven).
3.3 Results and Discussion
3.3.1 Air removal in unidirectional prepregs during VBO cure
The actual means of inter-ply void suppression are not well understood, although three
possible removal mechanisms are generally considered possible. The first is gas transport through
38

the inter-ply zone (Fig. 3-1). Because prepreg is not flat, gaps are created between adjacent plies
during the layup process. These gaps, a consequence of the inherent surface roughness of the
prepreg, provide a potential pathway for in-plane air transport, which can occur as continuum gas
flow (if the gap is empty) or as bubble migration through viscous resin (if the gap is resin-filled).
The second potential evacuation route is along the engineered vacuum channels at the center of
each prepreg ply. Gas initially entrapped within the pores that form the EVaCs can easily flow to
the laminate boundaries. Conversely, gas entrapped between plies must migrate short distances in
the through-thickness direction to reach the dry fiber tows. Once the gas reaches an EVaC, it can
be removed from the laminate, in-plane, via edge breathing. The third evacuation route is out-of-
plane transport through the thickness of the laminate. However, for unidirectional prepregs,
literature has shown that the through-thickness permeability is up to three orders of magnitude less
than the in-plane permeability, and can essentially preclude through-thickness air transport.[33,40]
In this study, most bubble transport occurred in the in-plane direction, in agreement with prior
observations, although concurrent through-thickness transport is possible as well, as discussed
later.
To observe inter-ply air evacuation and entrapment during the VBO cure process, a
custom-built experimental setup was used, incorporating in situ visual observation. Bubble
transport and the progression of gas evacuation during out-of-autoclave cure of unidirectional
prepregs were monitored using the in situ observation technique. Select still images from the
recorded video of a representative experiment are presented in Fig. 3-4. During layup, air is
initially trapped both in the intentionally created (artificial) pores (Fig. 3-4(a) circled in red) as
well as between the perforated resin film and the first ply of prepreg (white regions in Fig. 3-4(a)).
The naturally formed air pockets are randomly located around the artificial pores. When vacuum
39

is applied at the beginning of the cure cycle, the visible area of natural pores (and therefore the
area of the white regions) gradually shrink. The artificial pores also slightly decrease in size, but
the shape and position of the pores remain stable.
At low temperatures, inter-ply air evacuation in unidirectional prepregs is a slow process,
because the resin viscosity is high and there is no direct route for air flow to a breathing edge.
Close inspection of natural pores formed between the prepreg and the perforated resin film
indicates that small regions of trapped air are more likely to be evacuated as the cure cycle
progresses (Fig. 3-4(a-c)). The gas transport process, not surprisingly, occurs more easily at
elevated temperature as a result of reduced resin viscosity. The images in Figs. 3-4(b-d) show that
as the temperature increases, bubble size continues to decrease and even artificial air bubbles
migrate at ~ 70°C (see video). All movement of artificial air bubbles studied here occurred in-
plane, along the fiber direction, and toward the laminate edges. The driving force for bubble
movement is assumed to be the pressure differential supplied by the vacuum source. Throughout
the cure cycle, some air bubbles were observed to coalesce, forming one larger bubble. At 80°C,
a point roughly 30 minutes into the cure cycle, most of the naturally formed air pockets were no
longer visible, and the overall void content reached a minimum value (Fig. 3-4(c)).
Following this point in the cure cycle, an unanticipated phenomenon was observed. Beyond
80°C, as the temperature ramp continued and the dwell at 93°C began, existing air bubbles
increased in size. Additionally, some small bubbles grew from the residual naturally entrapped air.
This bubble expansion continued for approximately one hour. After the expansion, the morphology
of the bubbles also changed, with initially round bubbles elongating to different extents. Bubble
expansion, in general, indicates that the internal gas pressure within the bubble exceeds the
surrounding resin pressure and the resistance to growth created by surface tension. Following the
40

growth phase, bubbles gradually shrank until resin gelation occurred. Prior to completion of the
cure cycle, most bubbles were removed, resulting in a nearly void-free laminate.

Figure 3- 4 Time-lapse images of void evolution in unidirectional prepregs during cure. Images
were taken at a) initial state, before vacuum applied, b) 20min, c) 36min, d) 86min, e) 8h and f)
12h.
3.3.2 Tow impregnation during cure
While in situ monitoring provides insights into inter-ply air removal, tow impregnation is
more readily apparent in cross sections of partially cured prepregs. For this reason, the morphology
and impregnation of EVaCs were investigated via partial processing of a sequence of laminates.
Representative micrographs from different stages of the cure cycle are presented in Fig. 3-5. The
as-laid-up state of a stack of unidirectional prepreg plies, prior to compaction and cure, is shown
in Fig. 3-5(a). Within each ply, dry fiber tow areas are surrounded by resin-rich regions on both
sides. Large, elongated gaps are present between plies as a result of air entrapped during laminate
preparation. When heat and pressure are applied, the progression of fiber bed compaction and resin
infiltration begins. At 60°C (Fig. 3-5(b)), the fiber tows remain partially dry, while the fiber bed
41

is compacted and the inter-ply voids have reduced in size and become round. At 80°C when bubble
expansion is observed, dry fiber tows are almost fully impregnated. For this material system, full
impregnation is complete between 80°C (Fig. 3-5(c)) and 90°C (Fig. 3-5(d)).
Resin infiltration (tow impregnation) has a complex effect on entrapped air removal. Resin
flow can enhance air removal, as air bubbles trapped in resin-rich regions are able to migrate with
the resin flow front to dry fiber tows, from which in-plane evacuation can occur. However, resin
infiltrating the dry fiber tows can also impede air removal by occluding the air evacuation
pathways. Prepreg impregnation also affects local pressure and moisture concentration, both of
which evolve with time, and can affect void nucleation and/or growth in a complex manner.
The thickness change of laminate during cure also provide a proxy for fiber bed compaction
and resin impregnation. For example, the thickness of an 8-ply UD laminate decreased from 1.52
± 0.02 mm to 1.21 ± 0.02 mm (Fig. 3-6). The laminate thickness varies during the process as a
results of void reduction, resin infiltration, and chemical shrinkage. At early stage of a cure cycle,
when temperature is relatively low and fiber tow remain partially dry, thickness change of laminate
is primarily attributed to inter-ply air evacuation, local resin redistribution in the inter-ply space,
and fiber bed compaction. As temperature goes up, resin gradually infiltrate the dry fiber tows, the
main reduction of laminate thickness is due to resin impregnation.

42

Figure 3-5 Micrographs of tow impregnation in UD prepreg laminates during cure process.
Laminates were (a) at initial state after lay-up, cured to (b) 60°C, (c) 80°C and (d) 90°C.
43

As laid up 60°C 70°C 80°C 90°C
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
Thickness (mm)

Figure 3-6 Thickness change of an 8-ply UD laminate during cure.
3.3.3 Void removal mechanisms
The relationships between resin properties, tow impregnation and bubble behavior were
investigated to gain an improved understanding of air removal mechanisms in UD prepregs. Resin
viscosity and degree of cure were calculated using published viscosity and cure kinetics models
[23], while degree of impregnation was quantified by measuring the ratio of visible dry fiber tow
area to total ply area in polished cross-sections of partially processed samples. Void content was
determined by analyzing time-lapse videos from in situ cure monitoring experiments (with three
replicates).
The results of these experiments are summarized in Fig. 3-7, with all variables tracked as
a function of time, and correlated to temperature (Fig. 3-7(a)). Fig. 3-7(b) shows the evolution of
void content during the cure cycle. Initial void content, representing both naturally trapped air and
artificially introduced air bubbles, was approximately 14%. In the first 30 minutes of the cure cycle,
during the temperature ramp, void content decreased from 14% to ~2.5%. At this point, near the
end of the temperature ramp, void growth was repeatedly and consistently observed (without
44

nucleation of new voids), producing a ~2% increase in porosity (Fig. 3-7(b)). This void growth
was followed by a period of gradual decrease. The void content reached an equilibrium state after
approximately 7 hours at the cure temperature of 93° C. The final void content in all test panels
was less than 1%.
0 2 4 6 8 10 12
20
40
60
80
100
0 2 4 6 8 10 12
0
4
8
12
16
0 2 4 6 8 10 12
10
1
10
2
10
3
10
4
10
5
10
6
0 2 4 6 8 10 12
0.0
0.2
0.4
0.6
0.8
1.0
Temperature (°C)
Trial 1
Trial 2
Trial 3

Void Content (%)
Viscosity (Pa.S)
Gelation
Time (h)
Degree of
impregnation (-)
(a)
(b)
(c)
(d)
 
0.0
0.2
0.4
0.6
0.8
1.0
Degree of cure (-)

Figure 3- 7 The relationship between void content, resin properties, and tow impregnation. The
three stages of void evolution are denoted with roman numerals.
45

Examining void content as a function of time, the evolution of inter-ply air bubbles in UD
prepregs during cure can be divided into three stages: (I) air evacuation, (II) bubble expansion, and
(III) bubble shrinkage. These stages are denoted with Roman numerals and dashed lines in Fig 3-
7. During air evacuation, resin viscosity remains relatively high (Fig. 3-7(c)), and fiber tows are
partially dry (Fig. 3-7(d)), facilitating air flow through the prepreg to breathing edges. Air naturally
entrapped between the first ply of prepreg and the perforated resin film is gradually evacuated via
either resin flow or resin-starved regions on the surface of the prepreg (see Fig. 3-8). The
artificially induced air bubbles, in contrast, remain stationary because they are separated from the
air evacuation channels by the resin-rich surface of the prepreg. As temperature increases, resin
viscosity decreases, and large air bubbles begin to migrate in-plane, along the fiber direction.

Figure 3- 8 Surface of unidirectional prepreg, showing resin rich and resin starved regions
(arrows). Dry regions on the surface of the prepreg enable through-thickness gas transport to an
air evacuation pathway.
Bubble expansion occurs during Stage II, which begins at approximately 80°C, slightly
after bubble movement initiates. At this point in the cure cycle, resin viscosity is near minimum,
46

and dry-fiber gas evacuation pathways are almost fully sealed off, as shown in Figs. 3-5(d) and 3-
5(d). Full tow impregnation represents a significant change in the microstructure of the laminates,
potentially affecting gradients in pressure and moisture concentration. Data in Fig. 3-7 show that
these changes coincide with bubble expansion.
Bubble expansion/shrinkage depends on the equilibrium between the gas pressure within
the bubble Pvoid, the surrounding hydrostatic resin pressure Presin, and surface tension effects γLV. If
Pvoid > Presin + 2 γLV /Rvoid, a bubble will grow. During Stage II, both Presin and Pvoid are likely to
change due to full tow impregnation. For example, Pvoid is driven by moisture diffusion from the
resin, which now occurs solely towards bubbles (versus towards the bubbles and the flow front).
Moreover, diffusion is temperature-accelerated and water vapor pressure in the void increases
exponentially with temperature [45]. Full tow impregnation also represents a change in pressure
boundary conditions within the resin. Resin pressure is difficult to measure during prepreg cure,
because the resin is viscous, and because any sensor must form intimate contact with the resin
without interacting with the fibers. Preliminary efforts using a melt pressure sensor embedded
within a metallic tool plate and connected to the resin using a transfer fluid indicate that the resin
pressure remains between 50 – 100 kPa during processing. Studies have shown that during prepreg
processing, the internal vapor pressure in a void can exceed this value [45,55,77]. The interplay
between these factors is complex, but the data clearly show that bubble expansion in Stage II
coincides with full impregnation (Figs. 3-7(b) and (d)).
During the third and final stage of void evolution, bubbles shrink, and the dry fiber tows
are fully impregnated with resin. Meanwhile, the resin viscosity is increasing to a final plateau as
cross-linking completes. Because the tows are fully impregnated at this point in the cure cycle, gas
evacuation via EVaCs to a breathing edge is no longer possible. The decrease in void size, therefore,
47

is attributed to changes in moisture absorption in the resin as a function of degree of cure. Fig. 3-
8 shows moisture content as a function of degree of cure for the resin system studied here. The
data indicate that water solubility increases with degree of cure, a finding consistent with past
experimental studies in which the phenomenon is explained as a result of attractive forces between
water molecules and polar groups formed during epoxy cross-linking [102]. As the degree of cure
increases in the final stage of processing, water previously released into growing voids once again
dissolves into the resin, resulting in the final stage of bubble shrinkage. During this stage, bubbles
continued to migrate in-plane towards the edges. While this void migration could lead to a decrease
in void content, bubble movement did not affect void size measurements.
0.0 0.2 0.4 0.6 0.8 1.0
1.0
1.5
2.0
2.5
3.0
3.5
4.0
Moisture content (%)
Degree of cure (-)

Figure 3- 9 Water solubility in epoxy (Cytec 5320-1) resin as a function of degree of cure.
3.3.4 Bubble mobility
During Stage II, air bubbles in the resin were observed migrating in-plane, along the fiber
direction, towards the laminate edge. To clarify the physics of bubble migration, ten bubbles were
selected and tracked for quantitative analysis. These ten bubbles were independent but in close
48

proximity, ensuring that the local environments (i.e., local pressure/pressure gradient, resin flow,
etc.) were comparable. The morphology of each bubble was measured using a best-fit ellipse
(ImageJ), providing the bubble area, length (l, where l=2a and a is the half-major axis length) and
width (w, where w=2b and b is the half-minor axis). The average bubble velocity was determined
by the displacement of the center of each bubble over time. Bubble velocity was calculated over
three separate five-minute time periods during Stage II, labeled A, B and C, as shown in Fig. 3-
10(a).
The relationship between bubble velocity and resin viscosity, bubble size, and bubble
aspect ratio was analyzed. No obvious correlations between bubble velocity, resin viscosity, and
bubble size were observed. However, the bubble velocity showed a strong correlation to the aspect
ratio of the best-fit ellipse. Indeed, Fig. 3-10(b) shows that bubbles with larger aspect ratios
exhibited higher migration rates, as well as that the effects of bubble aspect ratio became less
pronounced as cure progressed. During bubble migration, the bubble tended to grow larger, and
became more elongated. Elongation can increase the pressure gradient across the bubble via
buoyancy, driving faster bubble migration. Conversely, surface tension also increases as the bubble
surface area increases, impeding bubble migration [78]. Because of these competing factors,
bubble velocity increases only when the increasing buoyancy force overcomes the increasing
surface tension.
Previously, authors have used the Hele-Shaw model to estimate bubble motion in a viscous
fluid [75,78,103]. In a Hele-Shaw cell, the moving bubble is assumed to travel in a Newtonian
fluid between two parallel plates with an infinitely small gap. For small bubbles in a Hele-Shaw
cell, the ratio of average bubble velocity to average resin velocity (U/V) depends on the aspect
ratio (a/b) of the bubble [78].
49

1
Ua
Vb
=+

(3-2)
The average fluid velocity V is given by [78]

2
12
h dP
V
dx 
−
=

(3-3)
where µ is resin viscosity, h is the channel height and dP/dx is the pressure gradient in the
resin.
For the case studies in this article, the physics of bubble transport are more complex than
those captured by this model, although the equations above provide useful insights into potential
factors affecting bubble migration. In Hele-Shaw flow, the channel height is assumed to be
constant. For our case, this channel height is related to the distance between the impermeable tool
plate and the fiber bed, or the instantaneous thickness of the perforated resin film. During
experiments, this channel height decreased as resin infiltrated the dry fiber tows, thereby reducing
bubble velocity. Moreover, as the degree of impregnation increased, the pressure gradient in the
resin also decreased, as the resin pressure equilibrated within the fiber bed. Thus, the effect of
bubble aspect ratio on bubble velocity diminished as time progressed. After reaching minimum
viscosity, the cure reaction led to increasing non-Newtonian behavior within the resin. These
deviations from the expected flow regime are expected to be minor during Stage II, since changes
in measured viscosity were small.
The observations described above emphasize that inter-ply bubble migration during
prepreg cure can be affected by multiple factors, including bubble morphology, resin viscosity,
and pressure gradient. These factors are interrelated and exert mutual influence in a complex
manner. Furthermore, these factors are generally difficult to determine due to stochastic variability
50

in materials and processes (for example, this study included observations that the migration of
some bubbles can be significantly impeded by a single off-axis fiber lying on the tow perimeter).
Overall, the results demonstrate that bubble flow models previously used by others provide a useful
baseline from which to explain experimental observations, but may not fully capture the
complexity of the physics involved and the mechanisms that govern defect formation during cure.
Figure 3- 10 (a) Temperature and resin viscosity profile versus time, indicating three time
intervals used to analyze bubble mobility, (b) correlation of bubble velocity with aspect ratio.
3.4 Conclusions
In this work, an in situ observation technique is described and used to monitor air removal
during the cure of out-of-autoclave carbon fiber/epoxy prepregs. The technique accurately captures
the phenomena that occur between the plies within a laminate during VBO cure. Simulating the
conditions of air removal during VBO cure is accomplished by incorporating a perforated resin
film into the lay-up to enable not only visualization of void evolution throughout cure, but also
quantitative measurements. Additionally, cure is carried out in an air circulating oven with a
standard vacuum bag configuration, reproducing a realistic cure environment. The method can be
51

utilized to ascertain mechanisms of inter-ply air removal (as in this work), as well as in parametric
studies to address key aspects of void formation and evolution in out-of-autoclave prepregs.
The method demonstrated here was applied to the analysis of inter-ply void evolution
during cure in unidirectional prepregs. Data obtained from in situ measurements were combined
with tow impregnation data obtained through cross-sectional analysis of partially processed
laminates. A comprehensive explanation of inter-ply air removal was established by inspecting
void evolution and prepreg morphology in both in-plane and cross-sectional views. The
explanation included a three-stage air removal mechanism based on void behavior, resin properties,
and tow impregnation which, to our knowledge, has not been previously observed or reported in
the literature. The resulting insights support and are consistent with our current understanding of
prepreg consolidation processes, and thus confirm that the method proposed here captures the key
phenomena governing inter-ply void evolution. The observations exhibit some of the trends
predicted by simple models, but highlight the complexity of the physics governing bubble mobility
and migration. In principle, this understanding can be used to optimize the cure of prepregs under
VBO conditions and modify cure cycles to facilitate void reduction. Porosity is the major defect
type arising during VBO processing, in the absence of elevated consolidation pressure.
Consequently, the development of science-based void reduction strategies can support the ongoing
shift from autoclave processing towards lower-cost, higher-efficiency methods.
Insights gained through in situ observation of process phenomena contribute to an
improved understanding of the mechanisms at play in composite manufacturing. For decades,
composite processing has been improved incrementally through a predominantly trial-and-error
approach. The development of in situ process diagnostics and the ability to monitor phenomena in
real time can provide valuable insights and understanding that will lead to improvements in both
52

process efficiency and part quality. The technique presented here was applied to void evolution in
UD prepreg to demonstrate proof of concept. The method, however, is widely applicable, and is
presently being used to assess void formation mechanisms with woven fiber architectures, novel
prepreg formats, and non-traditional cure cycles.
53

Chapter 4 Effects of Material and Process Parameters on Void Evolution in
Unidirectional Prepreg during Vacuum Bag-only Cure
4.1 Introduction
Because of the low applied pressure, VBO processing is more sensitive to sub-optimal
manufacturing conditions, and VBO-cured parts can exhibit unacceptable levels of porosity,
degrading mechanical properties [5,104–106]. Thus, science-based materials and process
optimization strategies must be developed to ensure successful cure. However, despite extensive
use of prepregs, the fundamental mechanisms of void formation and porosity evolution during
VBO processing are not yet well understood. VBO prepregs feature a partially impregnated
microstructure (by design), consisting of dry fiber tows and resin-rich regions [27]. The dry tows
form an interconnected network of high-permeability regions, enabling in-plane gas transport.
Thus, in principle, air entrapped in the laminate can be evacuated via laminate edges, reducing the
likelihood of gas-induced voids. Subsequently, once the laminate is heated, the resin rheology is
designed to achieve full saturation of the dry fiber areas, precluding flow-induced voids and
(ideally) yielding a defect-free part.
In practice, it is usually difficult to fully eliminate voids, because the mechanisms
governing air evacuation and void evolution are complex. Multiple factors that affect part quality
have been reported in the literature. For example, the temperature profile and vacuum quality
during cure strongly affect part quality [44,51,54]. Increased temperature reduces the resin
viscosity at the onset of heated cure, facilitating resin infiltration into dry fiber tows, while also
accelerating polymerization and crosslinking. Pressure gradients between ambient, bag, resin, and
void pressures drive resin and air flow, and govern fiber bed consolidation [6]. For VBO prepreg
54

cure, bag vacuum is the major controllable factor affecting air evacuation and fiber bed compaction.
Centea and Hubert assessed the impact of reduced ambient pressure, reduced vacuum and
restricted air evacuation on consolidation rate and laminate quality [54]. The results showed that
non-ideal pressure conditions led to specific, recognizable microstructural features.
At the material level, VBO processing is most sensitive to out-time (resin age) [60] and
dissolved moisture [55,56,58]. Out-time leads to undesirable resin polymerization/crosslinking
and increasing viscosity, eventually precluding full infiltration of dry regions, but can be
counteracted by modifying the thermal cycle. Moisture is a primary volatile released during VBO
cure. Resin absorbs moisture from the ambient – for example, when exposed to uncontrolled
environmental conditions during storage and layup. Grunenfelder and Nutt compared the effects
of relative humidity on porosity in both autoclave and out-of-autoclave processing [55]. Their
results showed that void content increased with increasing moisture content in VBO parts, while
the autoclave-processed parts remained void-free [55].
To date, studies have linked material and process parameter choices with cured part
characteristics. However, the precise phenomena associated with void evolution (nucleation,
growth/dissolution, and/or migration) remain only partly understood because of the difficulty of
real-time monitoring. Moreover, most studies have focused on flow-induced porosity located
within dry fiber tows [26,27,54,60] or on gas-induced void formation [77,78], while inter-ply gas
removal has received comparatively little attention.
In Chapter 3, we reported an in situ monitoring technique to investigate the formation and
removal mechanisms of gas-induced voids in unidirectional VBO prepregs [107]. This method
provided new insights into bubble migration, expansion, and removal during VBO cure. A three-
stage air removal mechanism was established by correlating void behavior with resin properties
55

and tow impregnation. These three stages include (I) air evacuation within a partially-impregnated
prepreg, (II) bubble expansion after prepreg saturation, and (III) bubble shrinkage due to evolving
thermochemical conditions and resin properties (details discussed in [107]). In this chapter, we use
the same in situ visualization method to clarify (via parametric variations) the effects of major
material and process factors on void evolution during VBO cure.
The present work aims to correlate and understand the relationships between material
properties, process parameters, and bubble behavior during cure, and to identify science-based
void reduction approaches. To achieve these goals, first, we conducted a parametric study to
understand the effects of key processing parameters on inter-ply air transport and removal in
unidirectional prepregs using in situ monitoring. In particular, we investigated the effects of resin
film infiltration and prepreg moisture content change during cure to provide insight and
understanding of void evolution mechanisms. The data were then used to modify the diffusion-
based void growth model to predict void size during cure. Finally, the model predictions were
compared with experimental data to improve the understanding of void growth/dissolution
mechanisms.
4.2 Experimental procedures
4.2.1 Test matrix
The same material was used in this study. The prepreg consisted of a unidirectional tape
(IM7 12K, 145 g/m
2
) and a toughened epoxy resin (CYCOM 5320-1, Solvay). Neat resin film
(CYCOM 5320-1, Solvay) was also used, with areal weight ~ 92 g/m
2
and thickness ~50 µm. The
temperature cycle used to cure the laminates was 93°C for 12 h, with an average ramp rate of
~2°C/min.
56

The same custom-built experimental setup introduced in Chapter 3 was used to observe
interply air entrapment and removal during VBO cure. A perforated resin film with holes of
controlled size and distribution was introduced into the lay-up to mimic the condition of air bubbles
trapped between prepreg plies. The perforated resin film was laid up against the glass window of
an oven, followed by four layers of prepreg plies and standard consumables shown in Figure 3-1.
Each laminate measured 127 × 127 mm and consisted of four plies stacked [0/90]2, while the
perforated resin film was 38 × 38 mm.
Factors considered in this study include both manufacturing parameters and material
properties, shown in Table 4-1. A 24 h vacuum hold at room temperature was performed before
cure to investigate air evacuation under ambient conditions and its effects on inter-ply air removal
during the subsequent cure with elevated temperature. The effects of vacuum quality were
investigated by fabricating laminates using full vacuum (corresponding to an absolute bag pressure
of ~2.3 kPa) and 80% vacuum (corresponding to an absolute bag pressure of ~21.3 kPa). The bag
pressure was monitored throughout cure using a pressure sensor.
Prepregs were conditioned in humid conditions to determine the effects of moisture
absorbed by the resin on inter-ply void evolution. Prepreg plies were conditioned for 24 h at room
temperature within saturated-salt humidity chambers (per [108]) with relative humidity levels of
70%, 80% and 99%, confirmed by digital humidity sensors. All samples conditioned for 24 h
reached ~99% of equilibrium moisture content. Prepreg plies were also dehydrated for 24 h in a
desiccator with a constant relative humidity of 1%. The moisture level of each sample was
measured by Fischer titration using a coulometric titrator (Mettler Toledo C20 with D0308 drying
oven). V oid evolution during cure was investigated using in situ monitoring for all conditions.
57

Within each experiment, only one parameter was varied. A baseline was fabricated under
full vacuum without room-temperature vacuum hold and humidity conditioning. The baseline test
was repeated 3 times, while all the other cases were repeated 2 times. Each laminate cured under
different conditions was compared to the baseline.
Table 4-1 Parameters included in this study.
Parameters Range
Vacuum hold 0, 24 h at 25°C
Reduced vacuum 80%, 100% vacuum
Humidity condition 1%, 40% (as-received), 70%, 80%, 99% RH

4.2.2 Resin film infiltration during cure
To better evaluate bubble size change and to complement visual data of the prepreg surface,
resin film thickness was tracked during cure using cross-sectional analysis. Laminates were
prepared by interrupting the baseline cure cycle at selected points, removing the part from the oven
while maintaining vacuum, and rapidly quenching to room temperature to prevent further resin
flow. Each laminate consisted of 4 plies of UD prepreg and a layer of resin film, same as during
in situ monitoring. The partially cured laminates were placed in an ammonia environment at room
temperature for one week, to achieve a hard and stiff structure [101]. In this way, the morphology
of the laminates at each point of interest was preserved. Cross-section samples were then imaged
after sectioning and polishing using a stereomicroscope (Keyence VH-Z100R), and the thickness
of the resin-rich regions was measured using the same microscope.
58

4.2.3 Resin moisture content during cure
The moisture content of the prepreg resin during cure was monitored by curing prepregs to
different points of the cure cycle, removing from the oven, and rapidly quenching to room
temperature while maintaining vacuum. Once room temperature was reached, laminates were
removed from the vacuum bag, stored in sealed bags, and titration tests were performed within 30
min. Each laminate consisted of four plies of prepregs and was cured using the same cure cycle
used for in situ testing.
4.3 Results and discussion
4.3.1 Vacuum hold effects
A 24 h vacuum hold was performed at room temperature before cure to investigate the
effects on interply air removal. The evolution of the visible resin-rich prepreg surface is shown in
Fig. 4-1. Initially, air was trapped both in the artificial pores created with the coring tool (Fig. 4-
4(a), circled in black), and between the perforated resin film and the first prepreg ply (white regions
in Fig. 4-1(a)). After the vacuum hold, the white regions (gaps between prepreg and resin film
formed by uneven surface topologies) largely disappeared because of compaction, while the
artificial pores did not change appreciably in position and size (Fig. 4-1(b)). After 24 h vacuum
hold, heat was applied to cure the laminates. As temperature increased, resin began to flow, and
air further evacuated (Fig. 4-1(c)). At about 80°C, air bubbles began to increase in size (Figs. 4-
1(c)&(d)), then gradually shrank until resin gelation occurred (Figs. 4-1(e)&(f)).
Fig. 4-2(a) shows that during the first two-hour room temperature vacuum hold, the total
void content decreased from ~13% to ~7% (red curve). Subsequently, evacuation slowed down,
and a final void content of ~4% was achieved after 24 h. These observations indicate that in
unidirectional prepreg, the removal of interply air at room temperature is relatively slow,
59

particularly after the early reduction of the pore network. Fig. 4-2(b) compares the void content of
prepregs that underwent vacuum hold (red curve) with that of baseline prepreg (without RT
vacuum hold) during heated cure. As reported previously [107], void evolution during cure can be
divided into three stages, denoted with Roman numerals (Fig. 4-2(b)). All three stages were
observed and occurred roughly at the same time as the baseline cure, indicating that the dominant
mechanisms of each stage were unchanged. During Stage I, as fiber tows remain partially dry, air
evacuates gradually (a) via resin starved regions on the surface of the prepreg or (b) via advection
by resin flow [75]. The void content at the end of Stage I was only slightly less than that of the
baseline, indicating that interply air evacuation at high temperature is more efficient. During Stage
II, void content increased by ~1.5%, slightly less than that of baseline (~2%). This onset of void
growth coincides with full tow impregnation, and we hypothesize that moisture diffusion from the
resin (to the bubbles) causes bubble expansion [107]. Subsequently, void content decreased
gradually throughout the rest of the cure (Stage III). Prior work showed that water solubility in
resin increased with degree of cure [107], and water content in the laminate decreased during cure
(discussed later), both of which drive moisture diffusion from voids to resin. Despite the lower
initial void content in the sample with a RT vacuum hold, porosity levels throughout the cure were
nearly identical to the baseline (Fig. 4-2(b)), indicating that a RT vacuum hold does not effectively
mitigate interply porosity in unidirectional prepregs. Also, room-temperature vacuum holds may
be effective for evacuation of intra-ply air, which is not entrapped in resin but can be time-
consuming to remove for large parts [33].
60

Figure 4-1 Void evolution in unidirectional prepregs during room temperature vacuum hold and
cure. (a) initial state, before vacuum applied, (b) after 24-hour RT vacuum hold, (c) 30 min
(~82°C), (d) 90 min, (e) 4 h and (f) 12 h.

Figure 4-2 Void content as a function of time during room temperature and elevated cure.
61

4.3.2 Bag pressure effects
The effect of vacuum quality on void evolution in UD prepregs is shown in Fig. 4-3. The
plots show that at 80% vacuum, interply air evacuation during Stage I remained efficient, as the
void content at the end of Stage I decreased to ~2.5% in both cases. Subsequently, at ~ 80°C,
bubble expansion began in both cases, marking Stage II. However, the duration of Stage II (bubble
expansion) at 80% vacuum was ~ 30 min longer than under full vacuum, and the void content at
the end of Stage II was ~50% greater than that of the baseline. These observations indicate that
compromised vacuum conditions do not substantially affect air evacuation (Stage I), although they
influence post-removal growth of bubbles (Stage II). The decrease in consolidation pressure
resulting from reduced vacuum, and the associated change in local resin pressure, cause this
divergent behavior: gas bubbles expand if internal gas pressures exceed the surrounding resin
pressure and surface tension. Finally, two hours into the cure cycle, air bubbles decreased in size.
Note that the final void content under both conditions was below 1%. As discussed before, bubble
dissolution depends both on the moisture content and on water solubility in the resin. As the initial
moisture content and the temperature profile were similar in both cases, the moisture content
profile during cure also should be similar. Although void content after Stage II increased, the resin
absorbed most of the residual moisture at the end of cure.
62

Figure 4-3 Void content as a function of time under reduced vacuum level.
4.3.3 Moisture effects
The effects of moisture on void removal were investigated by conditioning prepreg under
selected relative humidity levels for 24 h prior to cure. Fig. 4-4 shows images of the cured prepreg
laminates that were pre-conditioned to three moisture levels prior to cure. Both the size and number
of voids increased with moisture levels. More bubble coalescence was observed in prepreg
conditioned at 99% RH, resulting in much larger air bubbles in the cured laminate than the other
two cases.
Fig. 4-5 shows the void content of prepregs during cure. Prepreg conditioned at 99% RH
(wet prepreg) exhibited the most porosity, while prepreg dehydrated at 1% RH (dry prepreg)
contained the lowest void content throughout the cure (Fig. 4-5(a)). All the humidity conditioned
prepregs manifest the same three stages identified earlier. During Stage I, visible void content
decreased in all cases, reaching different levels by the end of Stage I. Porosity at the interface with
wet prepreg was more than double that of the baseline case, while the void content in dry prepreg
63

was slightly less than the baseline. During Stage II, air bubbles in wet prepregs expanded for ~1.5
hrs (compared with 1 h for baseline). Titration tests showed that the initial moisture content of wet
prepregs was ~ 0.4%, 4 times that of the baseline (~ 0.1%), and over 6 times that of dry prepregs
(~ 0.06%). As the resin moisture content increased, the concentration gradient of water between
the voids and resin increased, resulting in more diffusion time, allowing more water vapor to
migrate into voids before equilibrium. Note that while the moisture levels in the prepreg appear
small when expressed as a weight percentage, they represent potentially large volume fractions of
water vapor. Finally, during Stage III, the final void content of wet prepregs was ~ 4 times greater
than the baseline, although bubble shrinkage in wet prepreg was the most pronounced. A linear
correlation between void content and prepreg initial moisture content was observed, as shown in
Fig. 4-5(b). Thus, because the initial water concentration in resin is proportional to the square of
the relative humidity of environment, the working environment of prepregs is crucial for interply
void reduction.

Figure 4-4 Effects of moisture on void removal during cure. Images were taken at the end of cure
for prepregs humidity conditioned at (a) 1% RH, (b) 40% RH (baseline, as-received), (c) 99%
RH for UD prepregs, respectively.
64

Figure 4-5 void content as a function of (a) time for three cases; (b) initial moisture content of
prepreg.
4.3.4 Film thickness change during cure
To simulate inter-ply bubble evolution, our method relies on introducing a perforated resin
film at the tool/laminate interface. When heat and pressure are applied, resin viscosity decreases,
resin begins to infiltrate the dry fiber tows, and the resin film thickness decreases due to penetration
into the prepreg. Fig. 4-6 shows the change in film thickness during cure, measured using image
analysis (see red lines in Fig. 4-6(b)). The average initial film thickness was ~50 µm, decreased
from ~39 µm to ~30 µm during Stage II, and decreased to ~10 µm at the end of cure. Film thickness
change under reduced vacuum conditions (80% vacuum) was also measured, yielding nearly
identical results. The thinning of the resin film may in the present case result in flattening of the
artificial bubbles, increasing the visible surface area and thus the measured void contents.
Although the perforated resin film is an artificial construct, the phenomenon of bubble distortion
(flattening) is analogous to conditions experienced by large bubbles in the resin. To determine the
effects of film thinning on void removal mechanisms, an estimation of bubble volume was
performed by multiplying void content by the corresponding film thickness before and after the
65

observed bubble expansion. The estimation revealed that, even when corrected for film thinning,
the total volume of air bubbles (thus porosity) increased ~36% during Stage II. Thus, bubble
expansion does indeed occur after early-stage evacuation, and is an important mechanism of
interply void evolution.

Figure 4-6 (a) Film thickness as a function of time, (b) micrographs of resin film thickness
during cure process. Top: laminate was at 80°C (the beginning of stage II), bottom: at 90 min
into the cure cycle (the end of stage II). Resin film thickness was measured as the distance
between the red parallel lines.
4.3.5 Moisture content in resin during cure
Epoxy resin absorbs moisture from the environment during prepregging, storage, and lay-
up, and the absorbed moisture can lead to void nucleation and gas-induced porosity during cure,
particularly under VBO conditions [55]. While the effects of moisture exposure before cure on the
initial moisture content of prepreg have been reported [45,50,58], the evolution of water
concentration in resin during cure has not been addressed. Fig. 4-7 shows the change in prepreg
66

moisture content during cure. The black curve represents the moisture content of as-received
prepreg, while the red curve pertains to materials conditioned at 99% RH for 24 h. The moisture
content of both prepregs decreased as cure progressed, particularly during the first two hours. After
two hours, the moisture content of the as-received prepreg stabilized, while that of humidity-
conditioned prepreg continued to decrease slowly until the end of the cure. The largest decrease in
moisture content occurred 30-60 minutes into the cure in both cases, when the temperature was
high, and fiber tows were not fully impregnated, indicating that water vapor was evacuating via
breathing edges. Moisture content decreased by ~70% and by ~90% at the end of the cure for the
as-received and humidity-conditioned prepreg, respectively, indicating that higher initial moisture
content resulted in a higher rate of water extraction. Epoxies retain absorbed water molecules either
by forming hydrogen bonds with hydrophilic groups or by accommodating water molecules in the
free volume of the polymer [109,110]. Long exposures to humid environments will increase the
numbers of water molecules in epoxy in both forms, leading to a higher moisture content at the
end of cure (as well as a higher rate of decrease at the onset of cure). Multiple factors affect the
water concentration profile during cure, including cure temperature, heating rate, prepreg size, and
resin properties. We highlight here that water concentration in the resin decreases during cure, a
factor that will contribute to bubble shrinkage during stage III. This finding was used to modify
the diffusion-controlled model for the specific resin system and cure cycle used here, discussed
next.
67

Figure 4-7 Prepreg moisture content during cure.
4.4 Model development
4.4.1 Model framework
The void evolution model employed here is based on the diffusion-controlled void growth
model of Kardos et al. [45]. The theoretical basis of the model relies on the assumption that void
growth is caused by moisture diffusion from surrounding resin. Considering a spherical gas bubble
of radius R trapped at the center of an infinite pseudo-homogenous resin medium with an initial
water concentration of Cbulk, and denoting the water concentration near the bubble surface as Cvoid,
the water concentration gradient at the bubble surface r = R, according to Fick’s second law, is
[46]:
(4-1)
where D is the diffusion coefficient of water in the resin, which can be defined with a
temperature-dependent Arrhenius law.
𝜕𝐶
𝜕𝑟
|
𝑟 = 𝑅 =
𝐶 𝑏 𝑢𝑙𝑘 − 𝐶 𝑣𝑜𝑖𝑑 𝑅 [ 1 +
𝑅 ( 𝜋𝐷 𝑡 )
1
2
⁄
]
68

By expressing the rate of mass transfer using Fick’s first law and gas density ρ, the
differential equation for the bubble radius becomes:
(4-2)
According to Eq. (4-2), the diffusion process can favor either void growth or dissolution,
depending on the solubility of moisture in the resin and the concentration gradient between the
resin and the bubble surface [46,55]. To simplify the equation, the transient term can be neglected,
based on studies showing that transient behavior dominates only the first few seconds of void
growth [46].
4.4.2 Diffusivity
Because the specific diffusivity for the selected resin system is unknown, the diffusion
coefficient used in this model is obtained from Kardos’ model, which is similar in order of
magnitude to the diffusivity of water for a range of uncured epoxy resins [42,55,111]:
(4-3)
4.4.3 Water concentration in resin during cure
The initial water concentration in resin, Cinitial is determined by the relative humidity of the
environment RH and the water solubility S in resin.
(4-4)
𝑑𝑅 𝑑𝑡 =
𝐷 ( 𝐶 𝑏 𝑢𝑙𝑘 − 𝐶 𝑣𝑜𝑖𝑑 )
𝜌𝑅
[ 1 +
𝑅 ( 𝜋𝐷 𝑡 )
1
2
⁄
]
𝐷 = 0 . 175 e x p ( −
2817
𝑇 )
𝐶 𝑖𝑛𝑖𝑡𝑖𝑎 𝑙 =
𝑆 𝜌 𝑟 100
( 𝑅𝐻 )
2

69

where ρr is the density of uncured resin. Water solubility S was obtained from a parabolic
fit of experimental data (Fig. 4-8) measured using Fischer titration. Substituting into Eq. 4-4, the
initial water concentration in resin becomes:
(4-5)
In previous studies, the water concentration in the resin was assumed to be constant
throughout cure [45–47], potentially resulting in an overestimation of final void size. The
measurements presented in Fig. 8 show that water concentration decreased sharply during cure as
temperature increased. To more accurately describe the resin water concentration during cure and
to better predict void size, moisture content as a function of time was fit to an exponential
expression. The bulk water concentration during cure became:
92
1.23 10 (exp( 1.176 / 3600) 0.3012)
bulk
C RH t
−
=   − + For 40% RH
92
1.66 10 (exp( 0.8398 / 3600) 0.1171)
bulk
C RH t
−
=   − + For 99% RH (4-6)
The moisture content change for prepregs conditioned at 1% RH was not measured here
but is assumed to be the same as that of as-received prepregs (40% RH).
𝐶 𝑖𝑛𝑖𝑡𝑖𝑎 𝑙 = 1 . 7 × 10
− 9
( 𝑅𝐻 )
2

70

0 20 40 60 80 100
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Moisture Content (%)
Relative Humidity (%)
S = 1.29 x 10
-4

Figure 4-8 Moisture content in uncured prepreg as a function of relative humidity exposure.
4.4.4 Water concentration at void surfaces
The relative humidity at the void surface is determined by the ratio of partial pressure of
water in the bubble 𝑃 𝐻 2
𝑂 to the saturated vapor pressure 𝑃 𝐻 2
𝑂 ∗
. The water concentration at the void
surface is
𝐶 𝑣𝑜𝑖𝑑 = ⁡
𝑆 𝜌 𝑟 100
( 100
𝑃 𝐻 2
𝑂 𝑃 𝐻 2
𝑂 ∗
)
2
(4-7)
For model simplification, the void is assumed to be an air-water vapor mixture initially
containing dry air only. The partial pressure of water within the void is then
𝑃 𝐻 2 𝑂 = ( 1 − 𝜒 𝑎 𝑖𝑟 ) 𝑃 𝑔 (4-8)
where 𝜒 𝑎 𝑖𝑟 is the mole fraction of air in the gaseous void mixture, and Pg is the total
pressure in the gaseous void, equal to the hydrostatic resin pressure plus surface tension at
equilibrium. In this model, the hydrostatic resin pressure was assumed to be the same as the
pressure applied onto the prepreg, and surface tension was not considered. The saturated water
71

vapor pressure is given by the Clausius-Claperyron relationship [112], and the water concentration
at void surface becomes:
(4-9)
where 𝑃 𝐻 2
𝑂 ∗
( 𝑇 𝑟𝑒𝑓
) is the saturated vapor pressure at a reference temperature, ∆H is the
water enthalpy of vaporization, which is considered constant over the temperature range 25-100°C,
and Rg is the ideal gas constant. Because the water partial pressure cannot exceed the saturated
water vapor pressure, if 𝑃 𝐻 2
𝑂 ≥ 𝑃 𝐻 2
𝑂 ∗
( 𝑇 ) , Eq. 4-9 becomes:
𝐶 𝑣𝑜𝑖𝑑 =
𝑆 𝜌 𝑟 10 0
( 100 )
2
(4-10)
The gas density within the void mixture can be calculated by the following equation:
(4-11)
where MH2O and Mair are the molecular weights of water and air, respectively.
The values and units of the constants are given in Table 4-2. A forward-Euler linear solver was
used to solve all the above equations, and the time step (∆t = 10
-2
s) was adapted to ensure the
stability of the computation.
Table 4-2 Parameter values and units used in this model.
Symbol S ρr Rg ∆H P*H2O(Tref) Tref MH2O Mair
Value 1.29E-04 1.31E-03 8.314 40650 0.98 373 18 28
Unit - g/mm
3
J/mol•K J/mol atm K g/mol g/mol
Source Experiment [15] [113] [48] [48] [48] [113] [113]
𝐶 𝑣𝑜𝑖𝑑 =
𝑆 𝜌 𝑟 100
(
100 𝑃 𝐻 2
𝑂 𝑃 𝐻 2
𝑂 ∗
( 𝑇 𝑟 𝑒 𝑓 ) e x p [ −
𝛥𝐻
𝑅 𝑔 (
1
𝑇 −
1
𝑇 𝑟𝑒𝑓
) ]
)
2

𝜌 𝑔 = ( 1 − 𝜒 𝑎 𝑖𝑟 )
𝑀 𝐻 2 𝑂 𝑃 𝑔 𝑅 𝑔 𝑇 + 𝜒 𝑎 𝑖𝑟 𝑀 𝑎 𝑖𝑟 𝑃 𝑔 𝑅 𝑔 𝑇
72

4.4.5 Model correlation with experimental data
Sample selection. To compare experimental data with model predictions, representative
bubbles were selected to most closely replicate the assumptions of the model. Those bubbles
selected were artificial bubbles that were sessile and had little interaction with other bubbles
throughout the cure. Bubbles were approximated as short cylinders, and the volume of each bubble
was calculated by multiplying the visible surface area (measured using ImageJ) with the thickness
of the resin film at that time. The resin film thickness during cure was measured at only four
intervals, and interim film thickness values were estimated by assuming that the thickness
decreased linearly during each time interval. For each condition, 8 - 10 bubbles were selected, and
bubble size was plotted as symbols in Fig. 4-9. Large scatter in bubble size measurements is
frequently noted during analysis of porosity in composites (e.g.,[72,106]). In this study, dispersion
can be attributed to inherent variability in initial bubble size caused by entrapment or entrainment,
differences in local environmental conditions around each bubble during Stage I evacuation, and
sample preparation steps (e.g., backing paper removal), which can deform bubbles. Average
bubble size is shown in Fig. 10 as the blue dashed line. During Stage I, the average bubble size
decreased or remained constant, while during Stages II and III, the bubble size increased and
decreased, respectively, consistent with the overall evolution of void content.
The solid blue line in Fig. 4-9 represents the model prediction of void size evolution. The
initial void size was set to be same as the initial average bubble size in all cases. As shown in Fig.
4-9, the modified model predictions capture void expansion and shrinkage phenomena, and the
predicted void size is the same order of magnitude as measured values. However, the model
considers only mass transfer between voids and epoxy resin, and is unable to account for other
void evolution phenomena, particularly those occurring during Stage I. Another iteration of the
73

model calculation was performed, setting the void size at 30 minutes (80°C) equal to the measured
value for that time, yielding the solid red lines in Fig. 4-9. Note that at ~30 minutes, the dry fiber
tows are nearly fully impregnated [107], and thus mass transfer occurs solely between air bubbles
and resin. With this correction, model predictions for Stage II match the measured values for the
same period, indicating that void growth is caused by water diffusion from resin to air bubbles.
When bulk water concentration is time-dependent bulk water concentration, the model can
also be used to predict the onset of bubble shrinkage at the appropriate time. However, in all cases,
the predicted decrease in void size underestimates the measured rate and extent of shrinkage during
Stage III. This discrepancy may arise because the model does not account for changes in water
solubility. Note that water solubility in resin increases with increasing degree of cure, a result of
attractive forces between water molecules and polar groups in epoxy [102,107]. Once hydrogen
bonding forms, removing water molecules from resin requires higher activation energy [109]. In
other words, as cure progresses, it becomes more difficult for water molecules to diffuse between
resin and voids. Also note that water diffusivity was assumed to be constant throughout the cure
in this model. However, as degree of cure increases, the polymer network packs more tightly, and
the free space in resin decreases, impeding water molecule transfer between resin and air bubbles
[48,114]. An adjustment parameter α was introduced to understand these effects on void shrinkage:
α = ⁡
𝑛 𝑓𝑟 𝑒 𝑒 𝑛 𝑡 𝑜 𝑡 𝑎𝑙 ⁡ ⁡
(4-12)
where nfree is the number of water molecules that can freely diffuse between air bubbles
and the resin, and ntotal is the total number of water molecules dissolved in resin.
74

Fig. 4-9(d) plots estimates of void shrinkage with different  ratios. Currently, the actual
value of 𝛼 is difficult to determine. However, the parametric study shows that as the ratio decreases,
bubble shrinkage becomes more pronounced.

Figure 4-9 The model prediction and experimental data of void size as a function of time for
prepregs conditioned under (a) 1% RH, (b) 40% RH (baseline), and (c) 99% RH, and (d) the
effects of adjustment parameter on bubble shrinkage (with an initial bubble size of 0.1 mm
3
).
4.5 Conclusions
Parametric studies were conducted to demonstrate the effects of basic material and
processing parameters on interply air removal. The in situ visualization technique used to monitor
void evolution during vacuum bag-only cure reliably captures key phenomena associated with
interply void evolution. In all tests, a three-stage process of void evolution was observed. A room-
temperature vacuum hold prior to cure was not effective in removing interply air bubbles in high-
viscosity resins. Reduced vacuum quality did not meaningfully influence gas evacuation, but did
75

increase the bubble growth rate and time. However, for the lab-scale cases presented here, final
porosity remained similar to that of the baseline, indicating that the final interply porosity largely
depends on the ability of the resin (matrix) to dissolve water molecules during Stage III. Moisture
produced the most detrimental effects on interply porosity - void contents increased linearly with
absorbed moisture content. Thus, to reduce porosity, composite part manufacturers would be
prudent to limit moisture absorption, (more so than focusing on vacuum quality, although both are
important). Furthermore, resin formulations would benefit from greater hydrophobicity.
Prepreg moisture content decreased significantly during cure. Time-dependent moisture
concentration equations were applied to a diffusion-based void growth model to compare observed
trends to those predicted by established theory. The modified model, to the best of our knowledge,
is the first to predict both bubble growth and shrinkage during isobaric cure of out-of-autoclave
prepregs. Model predictions of void growth during Stage II were consistent with experimental
measurements, confirming that the observed void expansion during cure resulted from moisture
diffusion. Bubble shrinkage, while visible, was underpredicted by the model, reflecting the
likelihood that not all physical mechanisms governing this process were taken into account.
A widespread shift from autoclave processing towards out-of-autoclave processing will
rely heavily on the ability to limit and control porosity. Insights gained through the parametric
studies and analytical model described here improve our understanding of void evolution
mechanisms in out-of-autoclave composites manufacturing, and also provide a basis for the
development of science-based void reduction strategies. Most literature on VBO prepreg
processing has argued that air evacuation is the most important part of processing. However, our
findings suggest that the phenomena that take place after full impregnation (Stages II and III) can
be equally important.
76

Chapter 5 Void Evolution during Vacuum Bag-Only Processing of Fabric
Prepregs
5.1 Introduction
The previous two chapters showed the development of an in situ monitoring method that
enables direct observation of void evolution during the cure of VBO prepregs. By incorporating a
perforated resin film between a glass tool plate and a stack of prepreg plies, we introduced air
bubbles with controlled size and distribution into the layup, mimicking the conditions surrounding
an internal void located in the resin-rich regions between prepreg plies. This technique can provide
valuable insight into void evolution and transport during processing. The air removal mechanisms
and the effects of key processing parameters have been established for unidirectional prepregs.
In this chapter, the focus of this study is void evolution during the cure of fabric prepregs.
Fabric prepregs exhibit a periodic pattern of hillocks and valleys corresponding to the overlaps
and underlaps of tows in the weave, while UD prepregs have more uniform surface topology,
shown in Fig. 5-1. Thus the mechanisms of air removal during cure might be different. First, air is
less likely to be trapped during lay-up for UD prepregs due to the even surface. Second, as the
macro-pores exists in the fabric prepregs, air might be evacuated more effectively in fabric
prepregs. The effects of processing parameters on air evacuation in fabric prepregs might also
deviate from that in UD prepregs. Centea and Hubert assessed the effects of three pressure-related
process deficiencies on consolidation and part quality [54]. They showed that the process
deficiencies led to specific void content levels, distribution and morphologies, and were more
pronounced in woven fabric prepregs.
77

The present work aims to improve understanding of void evolution in fabric prepregs and
to clarify the effects of processing parameters (reduced vacuum in particular) on void evolution,
including the underlying mechanisms. We employed the same in situ monitoring method to
investigate mechanisms of inter-ply void evolution in fabric prepregs. First, void evolution during
cure at standard conditions was studied to establish inter-ply air removal mechanisms. Then, the
effects of room temperature vacuum hold, reduced vacuum, and moisture on void evolution were
investigated. Later, the effects of vacuum hold on each stage of void evolution were investigated
to understand void formation and removal mechanisms during each stage. Tow impregnation
during cure at reduced vacuum conditions was also studied, and a tow impregnation model was
modified to provide insights into the effects of reduced vacuum on resin infiltration. Finally, based
on the observations and model predictions, the interactions between intra-tow air and inter-ply air
at reduced vacuum conditions were addressed.
5.2 Experimental procedure
5.2.1 Materials and test matrix
The prepreg used in this chapter consisted of an eight-harness satin (8HS) fabric (T650-35,
3K) and a toughened epoxy resin (CYCOM 5320-1, Solvay). Neat resin film (CYCOM 5320-1,
Solvay) was also used, with areal weight ~ 92 g/m
2
and thickness ~50 µm. The single-dwell cure
cycle was 93°C for 12 h, with an average ramp rate of ~2°C/min. To investigate the effects of
vacuum quality on void evolution, laminates were fabricated using full vacuum, 80% vacuum
(corresponding to an absolute bag pressure of ~20.3 kPa) and 70% vacuum (corresponding to an
absolute bag pressure of ~30.4 kPa). The bag pressure was monitored throughout cure using a
pressure sensor or a vacuum gauge.
78

A 24-hour vacuum hold at 25°C was performed before cure to investigate the effects of
vacuum hold on air evacuation. Prepregs were conditioned in different humid ambient to determine
the effect of moisture content of the resin on void evolution. Prepregs plies were conditioned for
24h on a tray within a humidity chamber filled with saturated K2SO4 solution. The saturated K2SO4
solution provided a constant relative humidity level of 99%, confirmed by a digital humidity sensor.
Prepregs plies were also dehydrated for 24h in a desiccator with a constant relative humidity level
of 1%. Moisture level of each sample was measured by Fisher titration using a coulometric titrator
(Mettler Toledo C20 with D0308 drying oven). The void evolution during cure under different
conditions was investigated using in-situ monitoring technique described in Chapter II. Void
content as a function of time was measured for each in situ monitoring test panel. Ten
representative images were selected from the time-lapse video for subsequent analysis. The area
where the prepreg was not in contact with the perforated resin film was considered as voids. Void
content was determined as the ratio of the area of voids to the total area.
5.2.2 Model development
Tow impregnation is a key flow process occurring during VBO prepreg cure. Models have
been developed to predict the flow kinetics, in light of various material properties and process
parameters [51,60,98,115]. In this study, a simple approach previously developed by Centea et al.
[51] was modified to capture the effects of vacuum conditions on tow impregnation. The model is
based on Darcy’s Law for the low Reynolds number infiltration of a viscous fluid within a porous
medium and mass continuity:
𝑣 ̅ = −
𝐾 ̿
𝜇 ( 1 − 𝑉 𝑠 )
∇ 𝑃 (5-1)
∇ ∙ 𝑣 ̅ = 0 (5-2)
79

In Eqs. (5-1) and (5-2), 𝑣 ̅ is the average velocity of the fluid within the pores; 𝐾 ̿
is the
permeability tensor of the medium; µ is the dynamic viscosity of the fluid; Vs is the volume fraction
of the solid; and P is the fluid pressure. The framework can be simplified by assuming that (a) the
dry fiber bed regions ahead of the resin flow front are compressed to a constant volume fraction
Vf, (b) the tows are circular, and (c) the flow front is axisymmetric within the cross-section and
uniform along the tow length. With the radius of the tow being Rtow, the radius of the resin flow
front being Rf, and the corresponding resin pressure boundary conditions being P∞ and Pf ,
respectively, Eqs. (5-1) and (5-2) were combined to obtain the following expression for the resin
flow front velocity:
𝑣 𝑓 =
𝑑 𝑅 𝑓 𝑑𝑡
= −
𝐾 𝜇 ( 1 − 𝑉 𝑓 )
(
1
𝑅 𝑓 ( 𝑃 𝑓 − 𝑃 ∞
)
𝑙𝑛 (
𝑅 𝑓 𝑅 𝑡 𝑜 𝑤 )
) (5-3)
By defining a degree of impregnation β (0 ≤ β ≤ 1), Eq. (5-3) can be normalized to obtain
the tow impregnation model:
𝛽 = 1 −
𝑅 𝑓 𝑅 𝑡 𝑜 𝑤 = 1 − √
𝐴 𝑓 𝐴 𝑡 𝑜 𝑤 (5-4)
𝑑𝛽
𝑑𝑡
=
𝐾 𝜇 𝑅 𝑡 𝑜 𝑤 2
( 1 − 𝑉 𝑓 )
(
𝑃 ∞
− 𝑃 𝑓 ( 1 − 𝛽 ) 𝑙𝑛 (
1
1 − 𝛽 )
) (5-5)
Eq. (5-5) can be used to predict the evolution of the degree of tow impregnation of the
prepreg for any time-temperature cycle, provided the following parameters are determined: the
evolution of the resin viscosity µ; the tow volume fraction and geometry; the pressure boundary
condition P∞ and Pf , and the tow transverse permeability, K.
The dynamic viscosity µ was obtained using the predictive models developed by Kim et al.
[23]. This model describes the evolution of the degree of cure α (0 ≤ α ≤ 1) and the resin viscosity:
80

𝑑𝛼
𝑑𝑡
= ∑ 𝐾 𝑖 𝛼 𝑚 𝑖 ( 1 − 𝛼 )
𝑛 𝑖 + ∑
𝐾 𝑗 𝛼 𝑚 𝑗 ( 1 − 𝛼 )
𝑛 𝑗 1 + exp ⁡ ( 𝐷 𝑗 ( 𝛼 − ( 𝛼 𝐶 0 , 𝑗 + 𝛼 𝐶𝑇 , 𝑗 𝑇 ) ) )
𝑗 = 2 , 4 𝑖 = 1 , 3
(5-6)
𝐾 𝑛 = 𝐴 𝑛 𝑒 𝑥𝑝 ( −
𝐸 𝐴 , 𝑛 𝑅𝑇
) where n = i, j (5-7)
𝜇 = 𝜇 1
+ 𝜇 2
(
𝛼 𝑔𝑒 𝑙 𝛼 𝑔𝑒 𝑙 − 𝛼 )
𝐴 + 𝐵 𝛼 𝑑 + 𝐶 𝛼 𝑒 (5-8)
𝜇 𝑖 = 𝐴 𝜇 𝑖 𝑒 𝑥𝑝 (
𝐸 𝜇 𝑖 𝑅𝑇
) , i = 1, 2 (5-9)
In Eqs. (5-6) and (5-7), Kn is the Arrhenius temperature dependent term, An is the Arrhenius
constant, EA,n is the activation energy, mi and ni are reaction order-based fitting constants, Dj is the
diffusion constant, T is the temperature, αC0 is the critical degree of cure at absolute zero, αCT
accounts for the increase in critical degree of cure with temperature. In Eqs. (5-8) and (5-9), µi is
the Arrhenius dependent viscosity component, Aµi is the Arrhenius constant, Eµi is the viscosity
activation energy, αgel is the degree of cure at gelation, and A, B, C, d and e are fitting constants.
The numerical values of all constants for the resin system (CYCOM 5320-1) are provided in Table
5-1 (from [23]).
Table 5-1 Cure kinetics and viscosity model parameters.
Cure kinetics model Viscosity model
A 1 (s
-1
) 1.48 × 10
7
A 1 (s
-1
) 6.39 × 10
7
A µ1 (Pa s) 4.52 × 10
-9

E A1/R (K) 1.02 × 10
4
E A1/R (K) 8.94 × 10
3
E µ1/R (K) 7.59 × 10
3

m 1 0.17 m 1 1.65 A µ2 (Pa s) 1.73 × 10
-14

n 1 19.3 n 1 16.6 E µ1/R (K) 1.24 × 10
4

A 2 (s
-1
) 8.3 × 10
4
A 2 (s
-1
) 9.8 × 10
4
α gel 0.66
E A1/R (K) 8.54 × 10
3
E A1/R (K) 7.1 × 10
3
A 14.1 × 10
-6

m 2 0.70 m 2 1.66 B 53.7
81

n 2 0.87 n 2 3.9 C -44.96
D 2 97.4 D 2 63.3 d -0.13
α C0,2 -1.6 α C0,2 -0.60 e -0.11
α CT,2 (K
-1
) 5.7 × 10
-3
α CT,2 (K
-1
) 3.0 × 10
-3

Tow properties were determined according to the method described in Ref. [51]. A fiber
volume fraction of Vf = 0.74 was used for the 8HS prepregs. The major and minor diameters of
the elliptic tows were 2.44 mm and 0.33 mm, measured from an average of 20 tow cross-sections.
These values were converted to an equivalent circular tow radius of 0.116 using the relation
proposed by Van West et al. [116], who reported that circles and ellipses with equal hydraulic radii
have the same fill times. The tow permeability K was a constant and was obtained from [60]. The
values of all the constants are listed in Table 5-2.
The pressure boundary condition P∞ at r = Rtow is assumed to be the atmospheric pressure,
while the pressure boundary condition at Pf at r = Rf is assumed to be the difference between the
gas pressure entrapped within the tow Pgas and the capillary pressure Pc:
𝑃 𝑓 = 𝑃 𝑔𝑎 𝑠 − 𝑃 𝑐 (5-10)
Here, because the test panels were small and flat, the initial gas pressure Pgas within the
tow was assumed to be the same as the bag pressure Pvac. Thus, Pgas = 0 throughout the cure under
perfect vacuum conditions, while under reduced vacuum conditions, gas pressure during cure is
more complex, and can be affected by the rate of air transport out of the part, temperature, and
degree of impregnation. Although the exact evolution of Pgas was unknown, in principle, air within
fiber tows can be driven out by the infiltrating resin until no continuous pathways remain in the
tows.
82

To a first approximation, the evolution of Pgas was separated into two stages by introducing
a critical degree of tow impregnation βc. We assumed that before the degree of impregnation
reaches βc, air within the tow can evacuate instantaneously, a constant vacuum pressure condition
Pgas = Pvac can be applied at the resin front, while once the degree of impregnation exceeds βc, gas
inside the tow can no longer be evacuated (i.e., the mass of the gas inside the tow remains constant).
Upon reaching βc, Pgas can be updated using the ideal gas law. Here, βc = 0.8 was used, because
experiments showed that the effective in-plane permeability decreased by two orders of magnitude
when the degree of impregnation reached 0.8 (from 4.44E-14 m
2
to 4.49E-16 m
2
for this material).
With all parameters defined, Eq. (5-5) can be solved using a forward-Euler linear solver over small
time steps (∆t = 1s) to predict the evolution of β during cure.
Table 5-2 Summary of model parameter values
Parameters Value Ref.
Vf 0.74 [51]
Rtow 0.116 mm Experimental
βi 0.16 Experimental
K 6 × 10
-16
m
2
[60]
Pc 1.76 × 10
5
Pa [51]

5.2.6 Degree of impregnation during cure
To compare the model predictions with experimental data, laminates were partially
processed to selected points during the temperature ramp of the prepreg cure cycle (75°C, 80°C
and 85°C) under different vacuum conditions. Subsequently, panels were removed from the oven
and rapidly quenched to room temperature to prevent further resin flow. The partially cured
83

laminates were then cold-cured in an ammonia environment at room temperature for a week
(following the protocol described by Howard [101]) to achieve a hard and stiff structure while
preserving the morphology of the laminates at each point of interest. Samples were sectioned at
the center of each panel, polished, and inspected using a stereo microscope (Keyence VH-Z100R).
For each sample, 20 individual tow cross-sections were manually selected. The visible dry fiber
tow area Af, as well as the total area of the fiber tow Atow, was measured using ImageJ. The degree
of impregnation β was obtained by Eq. (5-4).
5.3 Results and discussion
5.3.1 Inter-ply air removal during VBO cure
The evolution of air entrapped in the resin-rich inter-ply regions is shown in Fig. 5-1.
Initially, air was trapped both in the artificial pores (Fig. 5-1(a), circled in red), and between the
perforated resin film and the first prepreg ply (Fig. 5-1(a), white regions). The naturally trapped
air pockets in 8HS prepreg were patterned and continuous, corresponding to the woven fiber-bed
architecture; most of the trapped air accumulated in the gaps between the resin film and depressed
regions of the fabric. Once vacuum was applied, air evacuated rapidly through the pinholes at the
intersections of warp and weft tows. Ten minutes after the application of vacuum, the majority of
both artificial bubbles and naturally trapped air were removed (Fig. 5-1(b)), indicating that most
inter-ply air can be evacuated effectively at a relatively low temperature. As temperature increased,
air bubbles began to migrate alongside fiber tows towards the pinholes, achieving further air
evacuation. At ~80°C, only a few small air bubbles remained, and the overall void content declined
to a minimum value (Fig. 5-1(c)). Following this point in the cure cycle, in accordance with void
evolution in UD prepregs [107], the remaining air bubbles increased in size for ~ 25 minutes. Then,
the bubbles gradually shrank throughout the remainder of the cure cycle. Most air bubbles were
84

removed by the end of the cure cycle, resulting in a nearly void-free laminate (Fig. 5-1(f)).
Fig. 5-2a shows the relationship between resin properties, tow impregnation, and void
content. As reported previously [107], void evolution in UD prepregs can be divided into three
stages based on the correlation of bubble behavior, resin properties and tow impregnation . For
fabric prepregs studied here, all three stages were observed, indicating that the dominant
mechanisms of void evolution remained unchanged. The three stages, denoted with Roman
numerals in Fig. 5-2(a), include (I) air evacuation, (II) bubble expansion and (III) bubble shrinkage.
During Stage I, as resin viscosity remains relatively high, and fiber tows are partially saturated,
inter-ply air is evacuated from the resin-starved regions at the intersections of fiber tows. Fig. 5-
2(b) shows the distribution of resin on the uncured prepreg surface. The open slits (dark areas)
located around almost every intersection of fiber tow provides pathways for air to quickly evacuate
through dry fiber tows to a breathing edge.
During Stage II, resin viscosity is near a minimum, and dry fiber tows are almost fully
impregnated. The increase in void size during this stage is attributed to moisture diffusion from
the prepreg resin to the air bubbles [107]. Because water vapor pressure increases exponentially
with increasing temperature, when the water vapor pressure within the voids exceeds the local
surrounding resin pressure, voids grow [77]. Finally, during Stage III, the occurrence of bubble
shrinkage is attributed to the increase of water solubility with increasing degree of cure and the
decrease in total water content in the prepreg as cure proceeds [107]. Both phenomena are expected
to reverse the direction of moisture diffusion from air bubbles to resin.
85

Figure 5-1 Void evolution in fabric prepregs during cure. Images were taken at (a) initial state,
before vacuum applied, (b) 10min, (c) 36 min, (d) 1h, (e) 6 h and (f) 12 h into the cure.

Figure 5-2 (a) Void content, temperature, resin viscosity, and tow impregnation as a function of
time. (b) Micrograph of uncured prepreg surface.
86

5.3.2 Effects of reduced vacuum on void evolution during cure
Fig. 5-3(a) shows void content of prepregs during cure under reduced vacuum conditions.
For all tests, the initial void content was ~ 50%. At the end of Stage I, the average void content of
prepregs at 70% and 80% vacuum decreased to ~ 1.2% and ~ 0.9%, respectively. Both porosity
levels were slightly greater than that of the control panel (~ 0.7%), indicating that air evacuation
in fabric prepregs became less efficient under reduced vacuum conditions. At ~ 80°C, bubble
expansion occurred roughly in the same manner as in the control panel. However, the duration of
Stage II (bubble expansion) at 80% vacuum extended to 4 hours, and by the end of Stage II, the
void content increased to ~ 1.2% (30% greater than control panel). Similar behavior was observed
in UD prepregs, although the duration of Stage II was only ~ 30 minutes longer than the control
[117]. The increased void content was attributed to the decrease in consolidation pressure
difference, as well as the associated change in local resin pressure resulting from reduced vacuum
[98].
At 70% vacuum, surprisingly, Stage III (bubble shrinkage) was not observed throughout
the cure cycle. A minimum value was reached at the end of Stage I, after which the void content
increased steadily until gelation. These observations indicate that further reducing vacuum level
will strongly affect the post-removal growth/ shrinkage mechanisms (Stage II and Stage III). The
final void content of prepreg at 80% vacuum was ~1% (10 times greater than control panel), while
that at 70% vacuum was ~ 2%.
To obtain a clearer picture of void evolution, the average bubble size and number of
bubbles during cure were investigated for all three vacuum conditions (see Fig. 5-3(b)). Because
the naturally trapped air pockets were interconnected at the outset, the initial size and number of
bubbles were not calculated, and Stage I was not included. Stages II and III were divided based on
87

the void evolution of the control panel. The plot shows that both size and number of bubbles
increased with decreasing vacuum at most of the junctures measured throughout the cure. At full
vacuum, bubble size nearly doubled during Stage II, then decreased slightly during Stage III, while
the number of bubbles decreased throughout the cure. At 80% vacuum, bubble size increased
steadily until 4 hours into the cure cycle, while the number of bubbles gradually decreased
throughout the cycle. At 70% vacuum, the bubble size increased markedly from ~ 45 minutes to 1
hour into the cure cycle. The number of bubbles also decreased rapidly during the same period,
indicating that air was evacuating, most likely due to a delay in resin infiltration. After that, both
size and number of bubbles increased gradually until gelation, indicating that new bubbles formed
during Stage III, which was not observed at full vacuum.
Based on these observations, the effects of reduced vacuum on inter-ply void evolution
include (1) less efficient inter-ply air evacuation during Stage I due to the lower pressure gradient
between inter-ply regions and vacuum source, (2) increased bubble expansion time and expansion
rate during Stage II, attributed to the decrease in consolidation pressure differential and local resin
pressure, and (3) formation of new air bubbles during Stage III, indicating a new source of voids
at reduced vacuum conditions (discussed later). In addition, reduced vacuum can impede tow
impregnation due to the resistance of residual air in dry fiber tows.
88

Figure 5-3 Void evolution under reduced vacuum conditions
5.3.3 Effects of reduced vacuum on air evacuation
To investigate the effects of reduced vacuum on air removal during Stage I and the effects
of inefficient initial air evacuation on void evolution, a 2-hour vacuum hold (unheated) was
performed prior to cure at 70% and full vacuum. Fig. 5-4(a) shows void content as a function of
time during room temperature vacuum hold. During the first 15-minute vacuum hold at full and
70% vacuum, the void content decreased from ~ 50%, to ~ 8% and ~ 15%, respectively. At the
end of the vacuum hold, however, the void content in both cases decreased to a similar value (~
3%), indicating that although the inter-ply air removal process was slower at reduced vacuum,
most inter-ply air bubbles could be evacuated with sufficient vacuum hold time.
Fig. 5-4(b) shows the void content during the subsequent heated cure. As temperature
increased, additional air evacuation occurred in both cases. At the end of Stage I, void contents for
the prepregs cured at full-vacuum and at 70% vacuum decreased to ~ 0.4% and to ~ 0.7%,
respectively. Both void levels were less than the control panel cured at full vacuum without an
unheated vacuum hold. The void content in prepreg cured at full vacuum reached a plateau during
Stage II, then gradually decreased. Further observation of the in situ video showed that the
89

remaining air bubbles expanded slightly during Stage II and shrank gradually during Stage III, and
no bubble formation was observed after Stage I. However, the void content of the prepreg at 70%
vacuum increased steadily and exhibited a marked increase in bubble size, and new bubbles formed
as well. The final void content of the prepreg at 70% vacuum was ~ 1.5%, much greater than the
control panel.
The findings indicate that while the unheated vacuum hold under deficient vacuum can
reduce the amount of initial inter-ply air to the same extent as that at full vacuum, the vacuum hold
has little influence on reducing defect contents in the final laminates. In other words, the inefficient
evacuation of inter-ply air bubbles during Stage I is not the major cause of the high final void
content at reduced vacuum conditions. We hypothesize that the final inter-ply void content is
determined by residual air remaining in the prepreg assembly (both inter-tow and intra-tow) when
the prepreg is fully saturated.

Figure 5-4 Void content as a function of time (a) during room temperature vacuum hold and (b)
during cure
90

5.3.4 Effects of reduced vacuum on bubble expansion
As discussed in Section 5.3.2, the longer bubble expansion time and higher expansion rate
observed at reduced vacuum levels were hypothesized to result from the reduced consolidation
pressure difference. Specifically, these observations were attributed to the associated change in
local resin pressure as bubbles expand when the internal gas pressure exceeds the surrounding
resin pressure and surface tension. To test this hypothesis, two experiments were performed by
controlling vacuum level during different stages of cure. As shown in Fig. 5-5, the vacuum history
for the red curve was full vacuum during the Stage I - the vacuum level was changed at the end of
Stage I (red arrow), when temperature reached 80°C, followed by 70% vacuum during the
remainder of the cure. Compared to a control panel cured at full vacuum (the blue curve), the
bubble expansion rate for the red curve was slightly greater, and the final void content was also
greater, confirming that reduced vacuum level affects the bubble expansion process. However, all
three stages were manifest in the red curve, and the final void content was much less than that at
70% vacuum. The findings indicate that reducing vacuum level during Stages II and III does not
substantially affect the bubble expansion/ shrinkage mechanisms.
In contrast, the black curve in Fig. 5-5 shows a vacuum history that began with 70%
vacuum during the Stage I, followed by full vacuum. After increasing to full vacuum (red arrow),
void contents continued to fall for about 20 minutes, then increased steadily for the remainder of
the cure cycle. Although full vacuum was applied during Stages II and III, the final void content
was similar to that at 70% vacuum, indicating that moisture vaporization was not the sole source
of bubble expansion. Failure to maintain a high vacuum quality at the beginning of the cure could
cause significantly increased porosity.
91

Figure 5-5 Effects of deficient vacuum applied at different stages during cure on bubble
expansion
5.3.5 Effects of reduced vacuum on tow impregnation
To understand the influence of reduced vacuum on intra-tow air entrapment and resin
infiltration, a model was modified (described in Section 5.2.4), and the model results were
compared to experimental data. Fig. 5-6(a) shows the model predictions for degree of impregnation
as a function of temperature and time at different vacuum levels. The model predictions show that
tow impregnation rate decreases with decreasing vacuum level. The measured data (shown in Fig.
5-6(b)) also show that the infiltration process decelerates as vacuum level decreases. The model
predictions match the experimental data during the tow impregnation process, indicating that the
model captures the key effects of reduced vacuum on tow impregnation.
Model predictions also show that for prepreg cured at 70% and 80% vacuum, resin
infiltration ceases before full tow impregnation due to the presence of intra-tow entrapped air.
However, cross-sections of the cured laminates (Fig. 5-7) show that fiber tows were fully
92

impregnated in all three vacuum conditions, contrary to the model predictions. One possible reason
for the difference is that air is fully evacuated before tow impregnation is completed in all three
conditions. However, this explanation is unlikely to be valid at reduced vacuum conditions, as tow
impregnation during cure is not uniform. When fiber tows are highly impregnated and dry fiber
regions become isolated, air can no longer be evacuated. A second (more likely) possible
explanation of the difference between model predictions and experimental data is that trapped
intra-tow air migrates into inter-ply regions during cure, which the model does not account for.
Figs. 5-7 (a-c) show that inter-tow voids increased with decreasing vacuum level, an indication of
a potential interaction between intra-tow air and inter-ply voids. To confirm this hypothesis, an
additional panel was cured at 100% vacuum with sealed edges to ensure air was trapped in the dry
fiber tows during cure. Fig. 8d shows that in this case, most fiber tows were still fully impregnated,
and large voids appeared in inter-ply regions, supporting the assertion that air trapped in the dry
fiber tows can migrate to inter-ply regions as resin infiltrates the tows.

Figure 5-6 (a) Model predictions of degree of impregnation as a function of time (b) Model
prediction vs. experimental data.
93

Figure 5-7 Cross-sections of laminates cured at (a) 70% vacuum, (b) 80% vacuum, (c) 100%
vacuum, and (d) 100% vacuum with sealed edges.
5.3.6 The interaction mechanisms – inter-ply voids and intra-tow air
Based on the in situ tests described above, three main interactions between inter-ply air
bubbles and the trapped intra-tow air were identified (Fig. 5-8). Figs. 5-8 (a-c) show that at 70%
vacuum, inter-ply air bubble size increased steadily throughout the cure, starting from the end of
Stage I. This observation confirms that the higher expansion rate and longer expansion time at
reduced vacuum conditions are caused not only by moisture diffusion, but also the coalescence of
intra-tow and inter-ply air bubbles.
Figs. 5-8 (d-f) show that a bubble emerged from a tow intersection (arrow) at 50 minutes
into the cure, and developed into a large air bubble within 5 minutes. This observation was not an
isolated case, but was observed repeatedly in both 70% and 80% vacuum conditions, despite the
small field of view of the microscope. Moreover, those bubbles appeared at roughly the same time
that tow impregnation was entering the plateau region of the model predictions, i.e., when the resin
flow front could no longer infiltrate remaining dry fiber regions. This change in boundary
condition could strongly affect the microstructure of the laminate, potentially changing the local
pressure gradient and causing air bubbles near pinholes to migrate from fiber tows to resin-rich
94

regions. Studies have shown that there is an increase in gas mobility when resin viscosity decreases
to its minimum [42,118], which promotes air migration. Furthermore, air migration can reduce
surface tension of air bubbles when elongated shapes, typically constrained by fiber arrays, adopt
spherical shapes in resin-rich regions.
Figs. 5-8 (g-i) show an elongated air bubble formed alongside adjacent fiber tows in
prepreg cured at 70% vacuum. Unlike the second type of bubble formation, the onset of this
expansion occurred at ~ 90 minutes into cure, and expansion was a slow process, lasting several
hours. Furthermore, this kind of air bubbles was observed only in prepregs cured at 70% vacuum.
As the vacuum level further decreased, fiber tows underwent less compaction, and the trapped
intra-tow air gradually migrated to inter-ply regions directly through fiber tows.
95

Figure 5-8 Interactions between inter-ply air bubbles and intra-tow air. (a-c) Intra-tow air
diffused into inter-ply air bubbles (representative air bubbles were circled in red) , images were
taken from prepreg cured at 70% vacuum at (a) 45 min, (b) 1.5h, (c) 8h; (d-f) air bubbles
emerged from the pinholes, images were taken from prepreg cured at 80% vacuum at (a) 50 min,
(b) 52 min, (c) 55 min into the cure cycle; (d-f) elongated air bubbles developed alongside fiber
tows. Images recorded from prepreg cured at 70% vacuum at (d) 1.5h, (b) 4h, and (f) 8h.
5.3.7 Effects of room temperature vacuum hold on void evolution during cure
A 24-hour room temperature vacuum hold was performed prior to cure to investigate the
removal of inter-ply air bubbles in fabric prepregs. Figure 5-9 shows the images of surface of the
prepreg before and after debulk, indicating that after 24-hour vacuum hold, almost all the air
96

bubbles are removal. In fact, air evacuation in 8HS prepreg during room temperature debulk is
very effective. 30 minutes after vacuum applied, void content of 8HS prepreg reduced from ~ 50%
to about 6%, after which void content gradually decreased with a lower rate. An interesting
phenomenon was observed during debulk – that was some air bubbles remained relatively stable
for hours, then suddenly got removed. This is likely due to the resin redistribution during vacuum
hold. Although resin viscosity remains relatively high at room temperature, the pressure
differential near the pinholes would drive some resin flow to create new pathways for air bubbles.
At the end of the vacuum hold, the void content dropped to 0.5%.
After 24-hour debulk, laminates were heated using the standard cure cycle and bubble
migration throughout the cure cycle was documented. At the beginning of the cure cycle, the
remaining air bubble migrated to the pinhole for further evacuation with the help of elevated
temperature (decreased viscosity). At the end of stage I, all the remaining air was removed in 8HS
prepreg. During the remainder of the cure, no new bubbles were formed and thus no bubble
expansion and/or shrinkage was observed. Another test was performed with prepregs humidity
conditioned with 99% relatively humidity before 24 room debulk. Similar result was observed,
indicating the coupled effects of entrapped air and moisture. Those existing air bubbles act as
nucleation site for moisture. During VBO processing, moisture in the resin is unlikely to nucleate
to form new voids but only diffuses into existing air pockets.
97

Figure 5-9 Effects of room temperature vacuum hold on air removal. Images were taken at (a)
initial state, (b) after 12 h RT vacuum hold (c) after elevated cure at 93°C for 12 h.
5.3.7 Effects of moisture on void evolution during cure
The effects of moisture on void removal in fabric prepregs were investigated by
conditioning prepreg under selected relative humidity levels for 24 h prior to cure. Fig. 5-10 shows
images of the cured prepreg laminates that were pre-conditioned in 1% RH, 40% RH and 99% RH
environments. The results showed that prepregs dried at 1% RH (dry prepreg) was void-free while
both the size and number of voids in prepregs conditioned at 99%RH (wet prepreg) exhibited the
highest.
Fig. 5-11(a) shows the void content of prepregs during cure. The initial void content of the
wet prepregs was ~ 3% higher than that in the other two cases, which is likely due to the increased
tack of wet prepregs. At the end of Stage I, porosity at the interface with wet prepreg was ~ 30%
higher than that of the other two cases. During Stage II, bubble expansion continued for 4 hours,
much longer than the baseline (at 40% RH), resulting in a ~ 2% void content at the end of Stage
II. After that, void content in the wet prepreg decreases slightly due to the increase in water
solubility in resin.
98

For dry prepreg, however, Stage II was not observed throughout the cure. After Stage I, a
nearly porosity-free laminate was achieved, and the remaining small air bubble kept shrinking
throughout the rest of the cure. The titration test showed the moisture content of the wet 8HS
prepreg is about 4 time higher than prepreg under standard conditions and is 6 time higher than
that of dry prepreg. As mentioned in previous chapters, void growth is due to the diffusion of water
from the surrounding resin and the driving forces are temperature and concentration gradient. As
the moisture content in the resin increases, the water concentration differential between the void
and resin increases, resulting in more water vapor to diffuse into voids.
Fig. 5-11 (b) shows the relationship between bulk porosity of 8HS laminate and prepreg
initial moisture content. The results indicate a linear relationship between laminate bulk porosity
and moisture when the moisture content exceeds certain critical value.

Figure 5-10 Effects of moisture on void removal during cure. Images were taken at the end of
cure for prepregs humidity conditioned at (a) 1% RH, (b) 40% RH (baseline, as-received), (c)
99% RH.
99

Figure 5-11 (a) Void content as a function of time; (b) bulk porosity as a function of moisture
content.
5.4 Conclusions
Degradation of composite mechanical properties often correlates with porosity, as well as
the shape, size and locations of the voids. Failure typically initiates from large voids located in
resin-rich inter-ply regions, especially for matrix dominant properties [119,120]. A thorough
understanding of the formation mechanisms and evolution of inter-ply voids is key to building a
basis for science-based defect reduction strategies. The work presented in this chapter provides
new insights into void evolution in fabric prepregs during VBO cure of prepregs, and the effects
of key processing parameters on inter-ply void formation.
This study highlights the new insight gained in the air evacuation mechanisms under
reduced vacuum conditions. Most studies have argued that the increase in void content under
deficient vacuum conditions is a result of decreased consolidation pressure. However, we conclude
that the sharp increase in both size and number of inter-ply air bubbles under such conditions arises
because of the remaining intra-tow air, rather than the reduced consolidation pressure. Our results
also show that even full vacuum is not sufficient to prevent the migration of intra-tow air towards
100

the inter-ply regions if air remains in the prepreg system at the end of Stage I. These new insights
emphasize strict control of vacuum quality, especially during the early stage of cure. In addition,
they also highlight the need for sufficient air evacuation prior to tow saturation. Air evacuation in
VBO prepregs (with conventional formats) can be especially challenging for large parts and/or
those with complex geometries, as it relies almost entirely on edge-breathing. To increase the
efficiency of air evacuation in VBO prepregs, modified prepreg formats with shorter breathe-out
pathways may be helpful to mitigate this issue [121,122].
Note that in our previous work, interactions between intra-tow and inter-tow air were not
observed during the cure of unidirectional prepregs at 80% vacuum [117], indicating that these
interactions are also affected by the fiber bed architecture. The inherent waviness of woven fiber
tows and the presence of resin-rich pinholes could promote migration of intra-tow air due to the
potential change in local pressure gradient. Thus, prepregs with less tow intersections such as
unidirectional tape and spread tow fabric might be less susceptible to vacuum deficiencies.
A simple tow impregnation model was modified and used in this study to predict the effects
of reduced vacuum on resin infiltration. The model captures the delay phenomenon of resin
infiltration but failed to predict the final degree of impregnation because it did not account for
interactions between inter-tow and intra-tow air. Furthermore, as the purpose of the model in this
study is to improve the understanding of resin infiltration at different vacuum levels, the critical
degree of impregnation represents only a first approximation. To provide more precise predictions
of tow impregnation during cure cycles, experimental data for in-plane permeability at different
degrees of impregnation will be required.
101

Chapter 6 Cure Optimization – Effects of Debulk Temperature on Air Removal
during Vacuum Bag-Only Prepregs
6.1 Introduction
Because of the partially impregnated format of VBO prepregs, air is initially present within
tows (intra-tow dry fiber regions), as well as between plies (inter-ply resin-rich regions) [30]. Dry
fiber regions provide high-permeability pathways for gas transport within the plane of the plies,
promoting air evacuation via edge breathing [27]. However, when air is trapped between plies
during lay-up, air bubbles must migrate in the through-thickness direction (for a short distance) to
reach in-plane evacuation pathways within the tows [107]. The results from previous chapters
showed that air evacuation prior to gelation is critical to achieve low porosity in laminates
produced by VBO processing of prepregs. Complete evacuation of inter-ply air prior to Stage I not
only reduces void size and number of voids in the final parts but also mitigates the negative effects
of moisture on void evolution.
In practice, air removal is usually accomplished by a room temperature vacuum hold
(debulk), typically hours, before initiating high-temperature cure. While edge-breathing is readily
achieved for small and flat laminates, pathways are often occluded in large and/or complex
geometries [123]. For large and/or complex parts, prepreg manufacturers generally recommend
debulks for 16 hours or more, which becomes the longest and thus rate-limiting step for VBO
production of large parts [124]. For some common geometric constraints, such as embedded ply
drops, edge-breathing can be problematic and/or virtually impossible.
A recent study [125] reported that similar quality was achieved in laminates cured using a
four-hour debulk at 50°C versus those cured with a 16-hour room temperature debulk prior to
102

heated cures shown in Fig 6-1. They concluded that debulking at an intermediate temperature
(50°C – 60°C), a so-called “super-ambient” dwell (SAD), reduced the cure cycle time without
compromising laminate quality. However, the mechanisms operative in super-ambient dwell were
not provided.

Figure 6-1 Comparison of part quality cured using RT debulk and super-ambient dwell [125].
The increase in debulk temperature has a complex effect on air removal. On one hand, the
decrease in resin viscosity increases resin flow, which generally enhances air removal. On the
other hand, increased resin infiltration can also occlude pathways for air evacuation. The interplay
between these factors introduces added complexity to the process. Thus, a clear understanding of
the effects of pre-cure dwell temperature on air evacuation must be acquired to establish science-
based guidelines for cure cycle optimization.
Permeability describes gas (and liquid) transport through a porous medium, and is thus an
important process variable affecting air evacuation. The effective gas permeability of composites
103

is affected by both material and process parameters [40]. In-plane permeability of prepregs has
been studied extensively [33,37–39], albeit mostly at room temperature. However, the initial in-
plane permeability values for typical VBO prepregs are on the order of 10
-14
m
2
[33,39]. In contrast,
the through-thickness permeability is generally 3-4 orders of magnitude less than the in-plane
permeability [33,40]. Kratz, et al. [40] and Tavares, et al. [42] both investigated the through-
thickness permeability during cure and reported an increase in through-thickness permeability at
high temperature due to the decrease in resin viscosity. However, no observations have been
reported on the effects of debulk temperature on air evacuation, and no comparison is available to
identify which air evacuation pathway (in-plane or transverse) is more effective during this period.
In this chapter, the removal of inter-ply air during heated debulk was investigated using an
in situ monitoring method. The effective permeability in both in-plane and through-thickness
directions was investigated to understand the effects of temperature on gas transport. Tow
impregnation during heated debulk was predicted using a model developed by Centea et al [51].
The relationships between gas permeability, resin properties, and tow impregnation as functions
of time and temperature are established. Cure optimization guidelines are provided by a
comparison of evacuation times at different debulk conditions.
6.2 Experimental
6.2.1 Materials
The prepregs selected for this study featured a toughened epoxy resin (CYCOM 5320-1,
Solvay) and two carbon fiber beds: IM7/12K unidirectional (UD) tape with fiber areal weight of
145 g/m
2
and a resin content of 33% by weight, and a T650-35 3K plain weave (PW) with areal
weight 196 g/m
2
and a resin content of 36% by weight. Neat resin film (CYCOM 5320-1) was also
used, with areal weight ~ 92 g/m
2
and thickness ~50 µm.
104

6.2.2 Permeability measurements
Gas flow can be used to characterize permeability of a porous medium. The relationship
between the flow rate and permeability is commonly expressed by 1-D Darcy’s law:
Q = ⁡ −
𝐾𝐴
𝜇 𝑑𝑃
𝑑𝑥
(6-1)
where K is the permeability scaler in the flow direction, A is the cross-sectional area of the
sample, dP/dx is the pressure gradient at position x, and µ is gas viscosity.
6.2.2.1 In-plane permeability test
In-plane permeability was measured using a steady-state air flow test. According to Eq. (1),
gas permeability can be measured by applying a constant flow rate of fluid through a porous
medium and measuring the pressure drop across the length of the sample. Assuming the gas
follows the ideal gas law, the flow rate at constant temperature is:
Q = ⁡ −
𝐾𝐴
2 𝐿𝜇
𝑃 2
2
− 𝑃 1
2
𝑃 1
(6-2)
where P1 is the pressure at the vent side, P2 is the pressure at the vacuum side, and L is the
sample length.
The experimental set-up is shown in Fig. 6-2 based on the work by Arafath et al. [33,39].
The laminate was bagged with three isolated compartments. The right compartment (vacuum side)
was used to apply vacuum to the laminate, while the left compartment (vent side) was connected
to a mass flow rate controller (MC-10SCCM, Alicat Scientific) to allow air flow into the laminate.
The center compartment contained the laminate, which was sealed on all sides with sealant tape,
allowing flow only in the in-plane direction. In this way, by imposing a known volumetric flow
rate using the flow rate controller and measuring the pressure on both sides, permeability was
calculated using Eq. (6-2).
105

During the pre-cure heated dwell, resin viscosity drops, and resin gradually infiltrates the
dry fiber tows. Consequently, a steady state flow cannot be achieved, hampering measurement of
permeability. Thus, rather than attempting to measure the in-plane permeability during dwell,
laminates were first partially processed to different degrees of impregnation according to the tow
impregnation model [51], then cooled to room temperature at each point of interest to avoid further
resin flow. Permeability measurements were then conducted on such laminates at room
temperature. The length and width of samples were L = 50.8 mm and W = 101.6 mm, while the
thickness of each sample was measured before each permeability test using a micrometer. For UD
prepregs, each sample consisted of 12 plies of prepreg stacked [0]12, while for PW prepregs, each
sample consisted of 8 plies stacked [0/90]4s. To obtain an average effective permeability value,
three samples (replicates) were measured for each degree of impregnation with a minimum of three
flow rate trials per sample.

Figure 6- 2 Schematic of in-plane permeability test set-up.
6.2.2.2 Through-thickness permeability test
Through-thickness permeability was measured using a “falling pressure” method, which is
frequently used in cases of low gas permeability where the steady-state flow test cannot be applied.
106

A custom test fixture, shown schematically in Fig. 6-3, was used for the experiments based on the
work of Tavares et al. [40,42] . A reservoir with a known volume was added to the set-up. When
vacuum is applied, the reservoir pressure will decrease if flow occurs through the porous media,
provided no leaks are present within the fixture. Using Eq. (6-1), the 1-D laminar flow of
compressible air through a porous medium at isothermal conditions is described by
−
𝐾𝐴 𝑃 𝐵 𝑎 𝑔 𝐿𝜇 𝑉 𝐶 𝑜 𝑟𝑒 𝑡 = ln [
( 𝑃 𝐶 𝑜 𝑟𝑒 ( 0 ) + 𝑃 𝐵 𝑎 𝑔 ) ( 𝑃 𝐶 𝑜 𝑟𝑒 ( 𝑡 ) − 𝑃 𝐵 𝑎 𝑔 )
( 𝑃 𝐶 𝑜 𝑟𝑒 ( 0 ) − 𝑃 𝐵 𝑎 𝑔 ) ( 𝑃 𝐶 𝑜 𝑟𝑒 ( 𝑡 ) + 𝑃 𝐵 𝑎 𝑔 )
] (6-3)
where PBag is the pressure at the bag side, PCore is the pressure at the honeycomb core side,
t is time, and VCore is the volume of the core (6.37 × 10
-4
m
3
). The air viscosity µ is a function of
temperature, which can be updated using the following equation:
𝜇 = 𝜇 0
(
𝑇 0
+ 𝐶 𝑇 + 𝐶 ) (
𝑇 𝑇 0
)
3 2 ⁄
(6-4)
where 𝜇 0
= 1.83 × 10
-5
Pa s at T0 = 293 K, and C is the Sutherland Constant, 117 K. Plotting
the left-hand side versus time yields a straight-line plot, the slope of which can be used to determine
the effective transverse permeability of the prepreg.
Prepregs were cut to 127 ×127 mm and were placed over the reservoir, supported by
aluminum honeycomb core. All four edges of the samples were sealed with sealant tape to prevent
in-plane gas flow, and a perforated release film was placed on top of the sample to allow air
evacuation only in the through-thickness direction, followed by breather cloth and vacuum bag.
Vacuum was drawn in the bag to compact the laminates and create a pressure difference between
core and bag. The evolution of pressure in the cavity was monitored over time using a pressure
transducer (Omegadyne, Inc) and data acquisition software (LabView, National Instruments).
Measurements were recorded for 4-, 8-, 16-, and 24-ply PW laminates and 1-, 2-, 4-, and 8-ply UD
107

laminates at isothermal temperatures (20°C, 40°C, 50°C, 60°C, and 70°C) for four hours, and
temperature was monitored throughout the tests using thermocouples.

Figure 6-3 Schematic of through-thickness permeability test set-up.
6.2.3 Tow impregnation model
VBO prepreg are partially impregnated consisting of a dry fiber bed located between two
continuous resin film. During cure, resin gradually infiltrates those dry fiber regions (ideally) to
form a void-free laminate. Thus, tow impregnation is a key flow process occurring during VBO
prepreg cure. Models have been developed to predict the flow kinetics, in light of various material
properties and process parameters [51,60,98,115]. In this study, a simple approach previously
developed by Centea et al. [51] was used to estimate the impregnation times at any given cure
cycles for both UD and PW laminates.
The model is based on Darcy’s Law for the low Reynolds number infiltration of a viscous
fluid within a porous medium and mass continuity:
𝑣 ̅ = −
𝐾 ̿
𝜇 ( 1 − 𝑉 𝑠 )
∇ 𝑃 (6-5)
∇ ∙ 𝑣 ̅ = 0 (6-6)
108

In Eqs. (6-5) and (6-6), 𝑣 ̅ is the average velocity of the fluid within the pores; 𝐾 ̿
is the
permeability tensor of the medium; µ is the dynamic viscosity of the fluid; Vs is the volume fraction
of the solid; and P is the fluid pressure. The framework can be simplified by assuming that (a) the
dry fiber bed regions ahead of the resin flow front are compressed to a constant volume fraction
Vf, and (b) the flow front is axisymmetric within the cross-section and uniform along the tow length.
For PW prepregs, the tows were assumed to be circular with a radius of Rtow. With the
radius of the resin flow front being Rf, and the corresponding resin pressure boundary conditions
being P∞ and Pf , respectively, Eqs. (6-5) and (6-6) were combined to obtain the following
expression for the resin flow front velocity:
𝑣 𝑓 =
𝑑 𝑅 𝑓 𝑑𝑡
= −
𝐾 𝜇 ( 1 − 𝑉 𝑓 )
(
1
𝑅 𝑓 ( 𝑃 𝑓 − 𝑃 ∞
)
𝑙𝑛 ( 𝑅 𝑓 / 𝑅 𝑡 𝑜 𝑤 )
) (6-7)
By defining a degree of impregnation β (0 ≤ β ≤ 1), Eq. (6-7) can be normalized to obtain
the tow impregnation model:
𝛽 = 1 −
𝑅 𝑓 𝑅 𝑡 𝑜 𝑤 = 1 − √
𝐴 𝑓 𝐴 𝑡 𝑜 𝑤 (6-8)
𝑑𝛽
𝑑𝑡
=
𝐾 𝜇 𝑅 𝑡 𝑜 𝑤 2
( 1 − 𝑉 𝑓 )
(
𝑃 ∞
− 𝑃 𝑓 ( 1 − 𝛽 ) 𝑙𝑛 ( 1 / ( 1 − 𝛽 ) )
) (6-9)
For UD prepregs, tow impregnation was approximated with a linear, one-directional
infiltration model as the micrographs show that the dry tows are planar, and that infiltration occurs
mainly in the transverse direction [95]. Thus, the resin flow velocity can be simplified to the
following form
𝑣 𝑥 =
𝐾 𝑥 𝜇 ⁡ ( 1 − 𝑉 𝑓 )
[
𝑃 ∞
− 𝑃 𝑓 𝑥 𝑓 ] (6-10)
where the subscript x denotes the direction of flow, P∞ is the resin pressure outside the tow
(x = 0) and Pf is the pressure at the flow front (x = xf). Eq. (6-10) can be nondimensionalized by
109

defining a degree of tow impregnation β = xf / L, where L is the total flow length. The rate of tow
impregnation can be expressed as

𝑑𝛽
𝑑𝑡
=
𝐾 𝑥 𝜇 ⁡ ( 1 − 𝑉 𝑓 )
[
𝑃 ∞
− 𝑃 𝑓 𝛽 𝐿 2
] (6-11)
Eqs. (6-9) and (6-11) can be solved using a simple forward Euler numerical method to
predict the evolution of the degree of tow impregnation of PW and UD prepregs for any time-
temperature cycle, provided the following parameters are determined: the evolution of the resin
viscosity µ; the tow volume fraction and geometry; the pressure boundary condition P∞ and Pf ,
and the tow transverse permeability, K.
The dynamic viscosity µ of the resin was determined using the predictive models
developed by Kim et al. [23]. Tow properties of PW were determined according to the method
described in Ref. [51]. A fiber volume fraction of Vf = 0.71 was used for the PW prepregs. By
measuring the major and minor diameters of the elliptic tows, an equivalent circular tow radius of
0.09 mm was obtained using the relation proposed by Van West et al. [116], who reported that
circles and ellipses with equal hydraulic radii have the same fill times. The tow permeability K of
PW was a constant and was obtained from [60]. The constant for UD prepreg were obtained from
Ref. [95]. The values of all the constants are listed in Table 6-1.

Table 6-1 Summary of model parameter values
Parameters
Values
PW UD
Vf 0.74 0.55
Rtow / L 0.116 mm 0.075 mm
βi 0.25 0.3
K 6 × 10
-16
m
2
1 × 10
-15
m
2

110

Pc 1.76 × 10
5
Pa N/A

6.2.4 Degree of impregnation
To determine the degree of impregnation for the in-plane permeability samples, a coupon
was processed alongside each sample. The coupons were removed from the oven with the samples
and quenched to room temperature to prevent further resin flow. The partially cured laminates
were then cold-cured in an ammonia environment at room temperature for a week [101]. This
treatment preserved the morphology of the laminates at each point of interest and yielded a hard
and stiff structure. Samples were sectioned at the center, polished, and inspected using a stereo
microscope (Keyence VH-Z100R).
6.3. Results and discussion
6.3.1 Inter-ply air evacuation
The effects of debulk temperature on inter-ply air evacuation are shown in Fig. 6-4. Initially,
air was trapped both in the artificial pores and between the perforated resin film and the first
prepreg ply (Fig. 6-4a white regions). These bubbles were situated primarily in the gaps between
the resin film and depressed regions of fabric due to the natural fiber crimp of PW prepreg. Once
vacuum was applied, bubbles disappeared rapidly through the pinholes at the intersections of warp
and weft tows. After a 4-hour debulk, the inter-ply air was completely evacuated, both at 20°C and
at 60°C (Figs. b-c).
In UD prepregs (Fig. 6-4(d)), although the initial entrapped air was much less prevalent
than in PW laminates. After debulk for 4 hours at room temperature (RT), the white regions
decreased slightly in size, while the larger artificial pores did not change in both size and position
(Fig. 6-4e). However, after a 4-hour debulk at 60°C, most naturally trapped air bubbles had
111

disappeared, and some of the artificial air bubbles decreased in size, indicating that debulk at 60°C
was more effective in reducing inter-ply porosity in UD prepregs.
Fig. 6-5(a) shows void content as a function of time in PW prepreg. The initial void content
was ~ 35 % due to the naturally trapped air. Though the air evacuation rate was slightly lower
during the RT vacuum hold, in both cases, void content decreased sharply to ~ 0.3% within the
first hour. UD prepregs exhibited a lower initial void content of ~ 13%. However, the evacuation
of inter-ply air was much slower in UD prepregs. The void content decreased to ~ 6% and ~ 3%
after a 4-hour debulk at 20°C and 60°C, respectively.
As mentioned before, inter-ply air bubbles must migrate a short distance to the dry fiber
tows to reach evacuation pathways. PW prepregs exhibit large open pores at the intersections of
the tow bundles (Fig. 6-4(b) circled in red), affording a grid of pathways for efficient evacuation
of inter-ply air. In contrast, UD prepregs feature resin film that is more uniformly distributed on
the surface, and air can be evacuated only via some resin-starved regions [107]. As debulk
temperature increases, resin viscosity decreases, facilitating resin flow, which promotes air
evacuation through those resin-starved regions. However, the larger, artificial air bubbles require
further resin flow to allow air bubbles to migrate towards evacuation pathways [107].
112

Figure 6-4 Void evolution during pre-cure dwell of PW prepregs: (a) initial state, (b) after 4-h
RT vacuum hold, (c) after 4-h 60°C vacuum hold; and for pre-cure dwell of UD prepregs: (d)
initial state, (e) after 4-h RT vacuum hold, (f) after 4-h 60°C.

Figure 6-5 Void content as a function of time during debulk for (a) PW and (b) UD laminates
113

6.3.2 Through-thickness permeability during heated debulk
The effects of debulk temperature and number of plies on the through-thickness
permeability of PW and unidirectional laminates are shown in Figs. 6-6(a-e). These plots show
transverse permeability of PW laminates as a function of time at different debulk temperatures.
Initially, permeability decreases with increasing ply number. For a 4-ply laminate, the permeability
was ~ 1 × 10
-14
m
2
, while for a 24-ply laminate, the permeability decreased by ~ 50 times to ~ 2 ×
10
-16
m
2
. The decrease arises because through-thickness gas flow occurs predominately via open
pores at tow bundle intersections (Fig.6-4(b)). As ply number increase, these open channels in the
laminate do not align, resulting in decreased permeability.
After the 4-hour RT debulk, the transverse permeability decreased by ~ 2-3 orders of
magnitude in all cases (Fig. 6-6(a)). The reduction in permeability was attributed to fiber bed
compaction, ply nesting, and the redistribution of resin and air pathways [42]. When the debulk
temperature increased to 40°C, resin viscosity decreased ten-fold (from 6.5 × 10
4
to 4.7 × 10
3
Pa·s),
while the transverse permeability decreased by an order of magnitude in the first hour, then
increased slightly for 4-ply and 8-ply laminates, while for 16-ply and 24-ply laminates, the
permeability fluctuated during the remainder of the debulk (Fig. 6-6(b)).
As resin viscosity decreases and resin mobility increases, new air pathways can form,
provided the gas pressure gradient overcomes the resin resistance. As ply number increases,
creating continuous flow channels becomes more difficult, decreasing permeability. Delay time is
another indication of this process. Delay time refers to the time between the start of vacuum hold
and the onset of the pressure decrease in the honeycomb core [40]. A delay in pressure drop in PW
permeability tests was observed only at 40°C. For an 8-ply laminate, the average delay time was
~ 6 minutes, while for a 24-ply laminates, the delay time increased to ~32 minutes.
114

When resin viscosity decreased to ~ 1500 Pa·s during the 50°C debulk, the permeability of
PW prepreg increased slightly for all cases, but continued to decrease with increasing ply number
(thickness), as shown in Fig. 6-6(c_. However, during 60°C and 70°C dwells, the permeability of
4-, 8-, and 16-ply laminates increased and stabilized at a similar value (~ 2 × 10
-14
m
2
) throughout
the dwell. In contrast, the 24-ply laminates exhibited a permeability value one order of magnitude
lower. This finding, while somewhat unexpected, can be understood because when resin viscosity
is sufficiently low (< 1000 Pa·s), the resistance of resin to air flow decreases to a minimum, and
when a pressure gradient is present, air evacuation pathways can be easily created. Air permeability
is independent of number of plies within a thickness range (< ~ 4 mm). However, as ply number
increases, the driving force for air evacuation (pressure gradient per thickness) decreases, and air
evacuation pathways become more tortuous, causing air permeability eventually to decrease.
The effects of debulk temperature on transverse permeability of UD prepegs exhibited clear
differences from PW prepregs. First, because each pressure decay test required 30 minutes to
several hours for UD laminates due to low permeability, the transverse permeability was plotted
as a function of temperature rather than time (Fig. 6-6(f)). UD laminates exhibited much lower
permeability than PW laminates. For a 1-ply laminate, the average permeability at room
temperature was 1.1× 10
-17
m
2
, ~ 3 orders of magnitude less than that of a 4-ply PW laminate.
As the number of UD plies increased, transverse permeability decreased sharply. For a 2-
ply and 4-ply laminate, the initial permeability at room temperature decreased to 4.8 × 10
-19
m
2

and 1.0 × 10
-19
m
2
, respectively, two orders of magnitude lower than that of the 1-ply laminate.
For an 8-ply laminate, permeability was effectively zero (undetectable). The sharp decrease in gas
permeability with increasing number of plies occurred because gas flow was dominated by a few
resin-starved areas, and as ply number increased, these air pathways were occluded.
115

Permeability increased with increasing temperature in all cases. However, as ply number
increased, higher debulk temperature was required to achieve an increase in permeability. The air
permeability began to increase when the debulk temperature reached 60°C for a 4-ply laminate,
while for an 8-ply laminate, no pressure drop was observed until the temperature reached 70 °C.
The increase in transverse permeability facilitated through-thickness gas evacuation, and
contributed to the more efficient inter-ply air evacuation described in Section 6.3.1.

116

Figure 6-6 Transverse permeability as a function of time at (a) room temperature; (b) 40°C; (c)
50°C; (d) 60°C; (e) 70°C for PW laminates; (f) Transverse permeability as a function of
temperature for UD laminates.
117

6.3.3 In-plane permeability during pre-cure dwell
Fig. 6-7(a) shows the in-plane permeability as a function of degree of impregnation (DOI).
The initial in-plane permeability of PW laminates (~ 6 × 10
-14
m
2
with DOI = 0.25) was slightly
greater than that of UD laminates (~ 2× 10
-14
m
2
with a DOI = 0.3). However, for both PW and
UD laminates, in-plane permeability decreased continuously with increasing degree of
impregnation, as resin infiltrated and occluded gas evacuation pathways. When the degree of
impregnation reached ~ 0.8, the in-plane permeability decreased to 2.3 × 10
-15
and 3.7 × 10
-16
m
2
,
for PW and UD, respectively, ~ 1-2 orders of magnitude less than the initial state. The permeability
of fully impregnated laminates was also measured, although a steady state was not achieved with
the lowest flow rate, indicating a permeability less than 1× 10
-18
m
2
.
The permeability data were then correlated to predictions of a tow impregnation model
developed by Centea et al. [51,95]. The model was used to generate in-plane permeability during
dwell for both PW and UD prepregs (Figs. 6-7(b-c)). During the RT vacuum hold, the laminate
exhibited the greatest in-plane permeability, which remained constant throughout the debulk.
However, when the debulk temperature increased, permeability decreased with time, as resin
gradually infiltrated the dry fiber pathways. The dashed vertical lines mark when the laminate was
“impermeable” in the in-plane direction (out of measurement range). When the debulk temperature
was >60°C, both PW and UD laminates became impermeable within ~ 1 hour.
118

Figure 6-7 (a) Measured in-plane permeability as a function of degree of impregnation. In-plane
permeability versus time during debulk for (b) PW and (c) UD laminates. The cure cycle used in
the model was 2°C/min to designated temperatures and hold for four hours.
6.3.4 Discussion
The permeability data in the previous sections showed that transverse permeability did not
increase significantly until the debulk temperature reached 60°C in UD prepregs. The transverse
permeability at 60°C was ~ 3-4 orders of magnitude less than the in-plane permeability. To
compare the gas removal efficiency of RT and 60°C vacuum holds, models can be used to estimate
the time required for air evacuation in both cases. Multiple models have been developed to predict
the evacuation of entrapped air in prepregs, and most of them relying on Darcy’s Law [37,38,43].
Combining Darcy’s law and the continuity equation, flow in the porous medium can be expressed:
119

⁡∇ ∙ ( 𝑃 𝐾 𝜇 ∇ 𝑃 ) = ∅
𝜕𝑃
𝜕𝑡
(6-12)
where P is pressure, K is the permeability, µ is gas viscosity and ∅ is the initial porosity of
the laminate. Considering a two-dimensional problem with K in the X and Z direction, Eq. (6-12)
becomes [38]:

𝐾 𝑥 𝜇 𝜕 ( 𝑃 𝜕𝑃
𝜕𝑥
)
𝜕𝑥
+
𝐾 𝑧 𝜇 𝜕 ( 𝑃 𝜕𝑃
𝜕𝑧
)
𝜕𝑧
= ∅
𝜕𝑃
𝜕𝑡
(6-13)
As the permeability, Kx and Kz, changes during heated or RT debulk, solving the two-
dimensional equation requires extensive calculation. The problem was therefore simplified to one-
dimensional by assuming that in-plane air evacuation is the sole pathway during RT debulk, while
air evacuation occurs exclusively via through-thickness pathways during heated (60°C) debulk.
These assumptions are used for estimation because during RT debulk, the transverse permeability
of 2- and 4-ply UD laminates were 5-6 orders of magnitude less than the in-plane permeability and
decreased with increasing time. Besides, during a 60°C debulk, tow impregnation reaches full
saturation in about an hour. The solution for one-dimensional gas transport can be expressed as
[43]:
𝑡 =
𝜇 𝑃 0
∅ 𝐿 2
𝐾 [ −
1
0 . 9
ln (
∅
𝐹 ∅
) ]
1
0 . 6
(6-14)
In this equation, permeability K can be in-plane, Kx , or through-thickness, Kz , and L can
be the length to breathing edge (half the laminate length) or the thickness of the laminate,
accordingly. Initial porosity ∅ was calculated from the cross-section micrographs of uncured UD
laminates, while ∅
𝐹 , the final porosity, was assumed to be 1%. With all the constants determined,
the vacuum hold times required for RT and for 60°C were estimated for different part sizes (L)
using Eq. (6-14).
120

Fig. 6-8 shows the vacuum hold times required for room temperature and 60°C air
evacuation. Transverse permeability data for 2- and 4-ply laminates were used to estimate the air
evacuation time required for 4- and 8-ply laminates, assuming that perforated release film was
applied to both sides of the laminates. The solid black line shows air evacuation time during RT
debulk as a function of distance to a laminate edge (half the laminate length, or the edge breathing
distance). The solid blue and red lines show the time required for air evacuation at 60°C for 4- and
8-ply laminates. The figure shows that if the laminate length exceeds 48 mm for a 4-ply laminate,
or 460 mm for an 8-ply laminate, debulking at 60°C is more efficient than at room temperature.
The delay time should also be considered to estimate air evacuation time, as pressure drop (i.e. air
evacuation) does not occur instantaneously upon application of vacuum (discussed in Section
6.3.2), and as number of plies increases, the delay time increases. The dashed lines showed that
the laminate critical length increases to 184 mm and 910 mm for a 4- and 8-ply laminate,
respectively, when taking delay time into account. The plot also indicates that for thin UD
laminates, a vacuum hold at 60°C prior to cure can be more effective when the laminate length
exceeds a threshold value. However, note that as the number of plies increases, the through-
thickness permeability decreases sharply and can eventually go to zero (no gas flow in the through-
thickness direction), even at elevated temperature.
For PW prepreg debulking at 60°C, the transverse permeability of a 4-, 8-, and 16-ply
laminates consistently reached ~ 2×10
-14
m
2
, a level comparable to the initial in-plane permeability
(~ 6×10
-14
m
2
), while the transverse permeability for a 24-ply laminate exhibited slightly lower
permeability (~2 ×10
-15
m
2
). The evacuation efficiency for in-plane and transverse gas removal
can be compared using the following equation, derived from Eq. (6-14):
121

𝑡 𝑥 𝑡 𝑧 = (
𝐿 𝑥 𝐻 𝑧 )
2
∙
𝐾 𝑧 𝐾 𝑥 (6-15)
Eq. 8 indicates that provided Lx/Hz > 5.5, a heated debulk at 60°C is more effective than a
RT debulk. In practice, laminate length is almost always 2-3 orders of magnitude greater than
laminate thickness, so a 60°C debulk will greatly increase the efficiency of air evacuation in PW
laminates.

Figure 6-8 Comparison air evacuation time between room temperature vacuum hold and 60°C
vacuum hold. Air evacuation time vs distance from edge breathing. Laminate length is 2 times of
the distance, assuming edge breathing is applied on all sides and in-plane permeability is the
same on all sides.
6.4. Conclusions
The effects of debulk temperature on inter-ply air removal and gas permeability of VBO
prepregs were investigated. Results showed that debulk temperature (resin viscosity) and fiber bed
architecture are critical factors for through-thickness air transport. Reducing resin viscosity
promotes inter-ply air removal via through-thickness gas transport while concurrently impeding
the in-plane gas transport due to increasing tow impregnation. In situ monitoring observations
122

confirmed that inter-ply air evacuation during 60°C debulk was more effective than room
temperature debulk. This method can be applied more widely to assess mechanisms of void
formation and migration, as well as revealing bubble behavior under different cure conditions and
with different prepreg formats.
This study provides new insights into air transport in prepregs at intermediate temperatures,
which can be used to guide development of more effective cure cycles for prepregs under VBO
conditions. Woven prepregs have large openings in the fiber bed, and thus increased debulk
temperature greatly enhances through-thickness gas transport. Results showed that for PW
laminates, if laminate length/ thickness ratio > 5.5, a debulk at 60°C can be applied to reduce total
processing time. However, in UD prepregs, transverse air evacuation relies on random resin-
starved regions, which are easily occluded. Thus, while increasing vacuum hold temperature can
benefit for thin UD laminates (< 8 plies ~ 1 mm), the transverse permeability of UD prepregs will
eventually decrease to zero when the number of plies increases (> 8 plies, in this case). Thus,
heated debulk is not an effective means to achieve air evacuation prior to cure in most laminates
comprised of UD prepregs. To increase transverse air evacuation efficiency in UD prepregs,
prepreg formats with discontinuous resin distributions can be a more effective approach [121,122].

123

Chapter 7 Contributions and Future Work
7.1 Contributions
The projects investigated in this thesis have led to several contributions to the knowledge
on the VBO prepreg processing that have not been reported before. Overall, the work presented
here provides an improved understanding of the mechanisms of void formation and evolution in
VBO-cured composite parts and insights into the optimized manufacturing protocols. The specific
contributions can be presented here:
1. A robust in situ monitoring technique for observation of void formation and
evolution during processing was developed. An in situ monitoring method was developed to
capture the phenomena that occur between the plies within a laminate during VBO cure by
incorporating a perforated resin film into and lay-up and curing the laminate on the glass window
of an oven door. This method accurately reproduces the conditions of inter-ply air removal during
VBO cure, enabling not only visualization of void evolution throughout cure, but also quantitative
measurements. This method was utilized to determine air evacuation mechanisms for both UD and
fabric prepregs and to address the effects of key processing parameters for void formation and
evolution in this work. This method can be widely applicable to assess void formation mechanisms
with other conditions. For decades, composite processing has been improved incrementally
through a predominantly trial-and error approach. This in situ monitoring technique provides
valuable insight into the void formation and evolution mechanisms, which offers scientific basis
on void mitigation strategies for VBO processing.
2. A three-stage air removal mechanism during the cure of VBO prepregs was
established. The in situ monitoring employed in this work enables the direct observation of bubble
behavior during cure and track void content as a function of time. Combining the data with tow
124

impregnation data obtained through cross-sectional analysis of partially processed laminates and
cure kinetics predicted by models, a three-stage void removal mechanism was established, which
to our knowledge, has not been previously observed before. The three stages include (I) air
evacuation through dry fiber tows during the early stage of process; (II) bubble expansion due to
moisture vaporization; and (III) bubble shrinkage associated with resin properties. This mechanism
applied to both UD and fabric prepregs and is consistent with our current understanding of prepreg
consolidation processes. This understanding can be used, in principle, to optimize VBO cure
process.
3. The detailed effects of material properties and processing parameters on void
formation and evolution were evaluated and the associated mechanisms were revealed. The
effects of room temperature debulk, vacuum quality and moisture levels on bubble behavior during
VBO cure were assessed using in situ monitoring method. The results highlighted the different
behavior of voids in UD and fabric prepregs. A room temperature vacuum hold prior to cure was
not effective in unidirectional prepregs, but resulted in void-free parts with 8HS prepregs. Reduced
vacuum quality did not meaningfully influence gas evacuation in UD prepregs, but had detrimental
effects on void formation in 8HS prepregs. The void formation mechanism during deficient
vacuum condition were established for 8HS prepregs. We concluded that the sharp increase in both
size and number of inter-ply air bubbles under reduced vacuum conditions is attributed to the
remaining intra-tow air, rather than the reduced consolidation pressure. The results also showed
that void content increased with increasing moisture content in both UD and 8HS prepregs,
indicating that resin formulations would benefit from greater hydrophobicity. These detailed
observations on bubble behavior under various condition provide basis for the development of
science-based void reduction strategies based on fiber bed architecture.
125

4. The effects of heated debulk on air evacuation and void reduction were assessed
and potential process optimizations were proposed based on the fiber bed architecture. The
effects of debulk temperature on inter-ply air removal were investigated. Results showed that inter-
ply air evacuation at 60°C was more efficient than that at room temperature. The gas permeability
measurement at various temperature and degree of impregnation showed that reduce resin
viscosity of VBO prepregs promoted through-thickness gas transport while impeded the in-plane
gas transport due to increasing tow impregnation. The results also lead to the conclusions that
increased debulk temperature greatly reduce total processing time for PW prepregs. However, for
UD prepreg, heated debulk is not an effective means to achieve air evacuation prior to cure in most
conditions except for super-thin laminates.
5. Bubble mobility during cure was evaluated and several models have been modified
depending on experimental data to achieve a more accurate prediction. The relationship
between bubble size and velocity during VBO cure of UD prepregs was established. The
experimental data suggested that Hele-Shaw cell model previously used by others a useful baseline
for bubble mobility in UD prepregs, but also emphasized the complexity of the physics involved
and that defect formation during cure cannot be fully captured by a simple model. The diffusion-
based void growth model was modified by adding the time-dependent moisture concentration
equations obtained from experimental data. The modified model more accurately predicted void
size during cure, and was the first to predict both bubble growth and shrinkage during isobaric cure
of VBO prepregs, although bubble shrinkage was underpredicted due to other physical
mechanisms that were not taken into account. The tow impregnation model was also modified to
predict tow impregnation during reduced vacuum conditions.
126

The above-listed contributions directly resulted in scholarly literature on VBO processing,
in the form of journal articles [107,126] and conference papers [29,117,127]. In addition, two
manuscripts based on the work presented in Chapters 5 and 6 were submitted to Polymer
Composites and Journal of Composites Materials for publication, respectively and is currently in
review.
7.2 Broader implications
The widespread ongoing shift from autoclave processing towards out-of-autoclave
processing relies heavily on the ability to replace the process robustness that autoclave pressures
provide, and the increase in economic value. As the demand of composite in aerospace industry is
expanding, cost- and time- effective manufacturing technologies such as VBO processing are
required to meet the growing demand. However, the removal of high pressure adds challenges in
the VBO processing to produce low-porosity (< 1%) composites parts. Careful selections of
materials and processing parameters are required to minimize porosity, which relies on a thorough
understanding of void formation and evolution during VBO processing. The projects completed
here provided new insights into mechanisms of air removal at various condition for VBO prepregs
and proposed science-based void reduction strategies.
7.3 Future work
Recommended future works include:
1. Bubble mobility during VBO cure of UD prepregs. A strong correlation between
bubble velocity and the aspect ratio of the best-fit ellipse of a bubble was established in this work.
However, the driving force of bubble migration was not fully understood. The observation of in
situ video showed that air bubbles migrated to different directions in different tests, although
always along fiber direction. Bubble migration seems to largely depend on the local environment
127

of a bubble. A detailed understanding of the factors (such as fiber orientation and resin flow) that
affect bubble migration can help the manufactures better design the materials that promotes inter-
ply air evacuation in UD prepregs.
2. Scale up to industrial relevant structures. The tests presented in this study were from
lab scale, flat and thin laminate (mostly 127 mm ×127 mm with 4 plies). The results provide a
fundamental understanding of void formation and evolution mechanisms. However, discrepancy
may exist in the large/ thick/ complex industrial parts. Conditions that occur commonly in general
scale up operation such as longer breathing and flow distance, uneven heating, longer out-time
should be examined.
3. Development of comprehensive cure optimization models. In Chapter 5, the modified
tow impregnation model for the prediction of degree of impregnation under reduced vacuum
conditions was developed by assuming a critical degree of impregnation value after which air can
no longer be evacuated. With the air permeability data in Chapter 6, a more precise model could
be developed to account for the effects of remain air in dry fiber tows on tow impregnation. In
addition, the two-dimensional air evacuation equation in Chapter 6 might be solved with FE
analysis, provided a detailed permeability profile during cure is obtained.
4. Investigation of other resin systems. The prepregs and resin film used in this study
have the same resin system (CYCOM 5320-1) from Solvay. It is necessary to evaluate the
generality of the results gained in this work.

128

References
[1] Campbell FC. Manufacturing Technology for Aerospace Structural Materials. Elsevier,
London, UK; 2006.
[2] Hale J. Boeing 787 From Ground Up. Boeing Comer Aeromagazine 2012:9.
[3] Tavares SS, Michaud V, Månson JAE. Assessment of semi-impregnated fabrics in
honeycomb sandwich structures. Compos Part A Appl Sci Manuf 2010;41:8–15.
doi:10.1016/j.compositesa.2009.09.005.
[4] Vanlandingham MR, Eduljee RF, Gillespie JW. Relationships between Stoichiometry,
Microstructure, and Properties for Amine-Cured Epoxies. J Appl Polym Sci 1999;71:699–
712.
[5] Liu L, Zhang BM, Wang DF, Wu ZJ. Effects of cure cycles on void content and mechanical
properties of composite laminates. Compos Struct 2006;73:303–9.
doi:10.1016/j.compstruct.2005.02.001.
[6] Centea T, Grunenfelder LK, Nutt SR. A review of out-of-autoclave prepregs - Material
properties, process phenomena, and manufacturing considerations. Compos Part A Appl Sci
Manuf 2015;70:132–54. doi:10.1016/j.compositesa.2014.09.029.
[7] Thorfinnson B, Biermann TF. Production of void free composite parts without debulking.
31st Int. SAMPE Symp., 1986, p. 480–90.
[8] Thorfinnson B, Biermann TF. Degree of Impregnation of Prepregs--Effects on Porosity.
Adv Mater Technol 1987:1500–9.
[9] Repecka L, Boyd J. Vacuum-bag-only-curable prepregs that produce void-free parts. 47 th
Int. SAMPE Symp. Exhib. 2002, 2002, p. 1862–74.
[10] Thomas S, Nutt SR. Temperature dependence of tesin flow in a Resin Film Infusion (RFI)
process by ultrasound imaging. Appl Compos Mater 2009;16:183–96. doi:10.1007/s10443-
009-9086-6.
[11] Xu GF, Repecka L, Moritmer S, Peake S, Boyed J. MANUFACTURE OF VOID-FREE
LAMINATES AND USE THEREOF. USOO6391436B1, n.d.
[12] Pascault J-P, Williams RJJ. General Concepts about Epoxy Polymers. Epoxy Polym
2010:1312–27. doi:10.1002/9783527628704.ch1.
[13] Min BG, Stachurski ZH, Hodgkin JH. Cure kinetics of elementary reactions of a diglycidyl
ether of bisphenol A/diaminodiphenylsulfone epoxy resin: 2. Conversion versus time.
Polymer (Guildf) 1993;34:4488–95. doi:10.1016/0032-3861(93)90155-4.
[14] Min BG, Stachurski ZH, Hodgkin JH. Cure kinetics of elementary reactions of a
DGEBA/DDS epoxy resin: 1. Glass transition temperature versus conversion. Polymer
(Guildf) 1993;34:4908–12. doi:10.1016/0032-3861(93)90017-5.
[15] Sovay Cytec Industries Inc. CYCOM® 5320-1 Epoxy Resin System 2015:1–11.
http://cytec.com/sites/default/files/datasheets/CYCOM_977_3.pdf.
129

[16] Kratz J, Hsiao K, Fernlund G, Hubert P. Thermal models for MTM45-1 and Cycom 5320
out-of-autoclave prepreg resins. J Compos Mater 2012;47:341–52.
doi:10.1177/0021998312440131.
[17] Yousefi A, Lafleur PG, Gauvin R. Kinetic studies of thermoset cure reactions: a review.
Polym Compos 1997;18:157–68.
[18] Kamal MR, Sourour S. Kinetics and thermal characterization of thermoset cure. Polym Eng
Sci 1973;13:59–64.
[19] Cole KC, Hechler JJ, Noel D. A new approach to modeling the cure kinetics of epoxy/amine
thermosetting resins. 2. Application to a typical system based on bis [4-(diglycidylamino)
phenyl] methane and bis (4-aminophenyl) sulfone. Macromolecules 1991;24:3098–110.
[20] Cole KC. A new approach to modeling the cure kinetics of epoxy/amine thermosetting
resins. 1. Mathematical development. Macromolecules 1991;24:3093–7.
[21] Castro JM, Macosko CW. KINETICS AND RHEOLOGY OF TYPICAL
POLYURETHANE REACTION INJECTION MOLDING SYSTEMS. Soc Plast Eng Tech
Pap Annu Tech Conf, 38th, Plast Prog Through Process, 1980.
[22] Khoun L, Centea T, Hubert P. Characterization methodology of thermoset resins for the
processing of composite materials -Case study. J Compos Mater 2010;44:1397–415.
doi:10.1177/0021998309353960.
[23] Kim D, Centea T, Nutt SR. Out-time effects on cure kinetics and viscosity for an out-of-
autoclave (OOA) prepreg: Modelling and monitoring. Compos Sci Technol 2014;100:63–
9. doi:10.1016/j.compscitech.2014.05.027.
[24] Hamill L, Centea T, Nutt S. Surface porosity during vacuum bag-only prepreg processing:
Causes and mitigation strategies. Compos Part A Appl Sci Manuf 2015;75:1–10.
doi:10.1016/j.compositesa.2015.04.009.
[25] Thomas S, Bongiovanni C, Nutt SR. In situ estimation of through-thickness resin flow using
ultrasound. Compos Sci Technol 2008;68:3093–8. doi:10.1016/j.compscitech.2008.07.012.
[26] Cender TA, Simacek P, Advani SG. Resin film impregnation in fabric prepregs with dual
length scale permeability. Compos Part A Appl Sci Manuf 2013;53:118–28.
doi:10.1016/j.compositesa.2013.05.013.
[27] Centea T, Hubert P. Measuring the impregnation of an out-of-autoclave prepreg by micro-
CT. Compos Sci Technol 2011;71:593–9. doi:10.1016/j.compscitech.2010.12.009.
[28] Fahrang L, Fernlund G. Void evolution and gas transport during cure in out-of-autoclave
prepreg laminates. Proc. SAMPE 2011 Conf., 2011.
[29] Hu W, Grunenfelder LK, Nutt SR. In-situ observation of void transport during Vacuum
Bag-Only cure. Int SAMPE Tech Conf 2016;2016-January.
[30] Farhang L, Fernlund G. Void and porosity characterization of uncured and partially cured
prepregs. J Compos Mater 2016;50:937–48. doi:10.1177/0021998315583924.
[31] Civan F. Porous media transport phenomena. Wiley Online Library; 2011.
130

[32] Dullien FAL. Porous media: fluid transport and pore structure. Academic press; 2012.
[33] Arafath ARA, Fernlund G, Poursartip A. Gas transport in prepregs: model and permeability
experiments. Proc. 17th Int. Conf. Compos. Mater. July 27-31, 2009, p. 1–9.
[34] Cender TA, Simacek P, Davis S, Advani SG. Gas Evacuation from Partially Saturated
Woven Fiber Laminates. Transp Porous Media 2016;115:541–62. doi:10.1007/s11242-016-
0784-x.
[35] Miguel AF, Serrenho A. On the experimental evaluation of permeability in porous media
using a gas flow method. J Phys D Appl Phys 2007;40:6824.
[36] Bear J. Dynamics of Fluids in Porous Media. New York: American Elsevericer Pubishing
Company 1972.
[37] Cender TA, Simacek P, Advani SG. In-plane gas evacuation of partially impregnated pre-
preg laminates in out of autoclave processing. Int. SAMPE Tech. Conf., Baltimore, MD:
2015.
[38] Levy A, Kratz J, Hubert P. Air evacuation during vacuum bag only prepreg processing of
honeycomb sandwich structures : In-plane air extraction prior to cure. Compos Part A Appl
Sci Manuf 2015;68:365–76. doi:10.1016/j.compositesa.2014.10.013.
[39] Louis B, Hsiao K, Fernlund G. Gas permeability measurements of out of autoclave prepreg
MTM45-1/CF2426A. SAMPE Conf 2010 May 17– 20, Seattle, WA; 2010 2010.
[40] Kratz J, Hubert P. Anisotropic air permeability in out-of-autoclave prepregs: Effect on
honeycomb panel evacuation prior to cure. Compos Part A Appl Sci Manuf 2013;49:179–
91. doi:10.1016/j.compositesa.2013.02.013.
[41] Kratz J, Centea T, Hubert P. Air Permeability of Out-Of-Autoclave Honeycomb Structures
During Processing. Proc 26th Annu Tech Conf Am Soc Compos / 2nd Jt US-Canada Conf
Compos 2011.
[42] Tavares SS, Michaud V, Månson JAE. Through thickness air permeability of prepregs
during cure. Compos Part A Appl Sci Manuf 2009;40:1587–96.
doi:10.1016/j.compositesa.2009.07.004.
[43] Fernlund G, Wells J, Fahrang L, Kay J, Poursartip A. Causes and remedies for porosity in
composite manufacturing. IOP Conf Ser Mater Sci Eng 2016;139. doi:10.1088/1757-
899X/139/1/012002.
[44] Agius SL, Magniez KJC, Fox BL. Cure behaviour and void development within rapidly
cured out-of-autoclave composites. Compos Part B Eng 2013;47:230–7.
doi:10.1016/j.compositesb.2012.11.020.
[45] Kardos J, Duduković M, Dave R. Void growth and resin transport during processing of
thermosetting - Matrix composites - Epoxy Resins and Composites IV - Advances in
Polymer Science. Adv Polym Sci 1986;80:101–23. doi:10.1007/3-540-16423-5_13.
[46] Wood R, Bader G. Void control for polymer-matrix composites (1): theoretical and
experimental methods for determining the growth and collapse of gas bubbles 1994;5:139–
47.
131

[47] Ledru Y, Bernhart G, Piquet R, Schmidt F, Michel L. Coupled visco-mechanical and
diffusion void growth modelling during composite curing. Compos Sci Technol
2010;70:2139–45. doi:10.1016/j.compscitech.2010.08.013.
[48] Préau M, Hubert P. Effects of processing conditions on bondline void formation in vacuum
bag only adhesive bonding : Modelling , validation and guidelines. Int J Adhes Adhes
2018;80:43–51. doi:10.1016/j.ijadhadh.2017.10.004.
[49] de Parscau du Plessix B, Le Corre S, Jacquemin F, Lefebure P, Sobotka V. Improved
simplified approach for the prediction of porosity growth during the curing of composites
parts. Compos Part A Appl Sci Manuf 2016;90:549–58.
doi:10.1016/j.compositesa.2016.08.024.
[50] Mohseni S, Fernlund G, Lane M. Cure cycle design to suppress moisture-driven bubble
growth in polymer composites. J Compos Mater 2018;52:1821–32.
doi:10.1177/0021998317734393.
[51] Centea T, Hubert P. Modelling the effect of material properties and process parameters on
tow impregnation in out-of-autoclave prepregs. Compos Part A Appl Sci Manuf
2012;43:1505–13. doi:10.1016/j.compositesa.2012.03.028.
[52] Kim D, Centea T, Nutt SR. In-situ cure monitoring of an out-of-autoclave prepreg: Effects
of out-time on viscosity, gelation and vitrification. Compos Sci Technol 2014;102:132–8.
doi:10.1016/j.compscitech.2014.07.027.
[53] Olivier P, Cottu JP, Ferret B. Effects of cure cycle pressure and voids on some mechanical
properties of carbon/epoxy laminates. Composites 1995;26:509–15.
[54] Centea T, Hubert P. Out-of-autoclave prepreg consolidation under deficient pressure
conditions. J Compos Mater 2014;48:2033–45. doi:10.1177/0021998313494101.
[55] Grunenfelder LK, Nutt SR. Void formation in composite prepregs - Effect of dissolved
moisture. Compos Sci Technol 2010;70:2304–9. doi:10.1016/j.compscitech.2010.09.009.
[56] Kay J, Fernlund G. Process conditions and voids in out of autoclave prepregs. Proc. SAMPE
Tech. Conf., Baltimore, Maryland May 21-24: 2012, p. 12.
[57] Mobuchon C, Keulen C, Hsiao K, Fernlund G, Poursartip A. Development of a structured
approach to porosity management in composites manufacturing. Proc. 44th SAMPE Tech.
Conf. Seattle, Washingt. June 2-5, 2014, p. 12.
[58] Anderson JP, Altan MC. Formation of Voids in Composite Laminates : Coupled Effect of
Moisture Content and Processing Pressure. Polym Compos 2015;36:376–84.
doi:10.1002/pc.
[59] Kim D, Centea T, Nutt SR. Modelling and monitoring of out-time and moisture absorption
effects on cure kinetics and viscosity for an out-of-autoclave (OoA) prepreg. Compos Sci
Technol 2017;138:201–8. doi:10.1016/j.compscitech.2014.05.027.
[60] Grunenfelder LK, Centea T, Hubert P, Nutt SR. Effect of room-temperature out-time on
tow impregnation in an out-of-autoclave prepreg. Compos Part A Appl Sci Manuf
2013;45:119–26. doi:10.1016/j.compositesa.2012.10.001.
132

[61] Lucas S, Howard S, Senger J. Vacuum bag only processing: Improving prepreg out-time
and porosity for large composite structure. Proc. SAMPE 2010 Conf. Soc. Adv. Mater.
Process Eng., 2010, p. 1–16.
[62] Centea T, Kratz J, Hubert P. The effect of out-time and cure cycle on the consolidation of
out-of-autoclave prepregs. Proc. 11th Int. Conf. Flow Process. Compos. Mater, Auckland,
New Zeal., 2012.
[63] Hamill L, Centea T, Nilakantan G, Nutt S. Surface porosity in out-of-autoclave prepreg
processing: Causes and reduction strategies. Int SAMPE Tech Conf 2014.
[64] Ghiorse SR. A comparison of void measurement methods for carbon/epoxy composites.
ARMY LAB COMMAND WATERTOWN MA MATERIAL TECHNOLOGY LAB; 1991.
[65] Ghiorse SR. Effect of void content on the mechanical properties of carbon/epoxy laminates.
SAMPE Q 1993;24:54–9.
[66] Astm D. Standard test method for void content of reinforced plastics. West Conshohocken
ASTM Int 1999.
[67] Paciornik S and, d’Almeida JRM. Measurement of void content and distribution in
composite materials through digital microscopy. J Compos Mater 2009;43:101–12.
[68] Handbook MS. Composite Materials Handbook. Volume 3, Polymer Matrix Composites,
Guidelines for Characterization of Structural Materials, Department of Defense Handbook,
USA. MILHDBK-17-1F, June 2002; n.d.
[69] Zhan-Sheng Guo, Ling Liu, Bo-Ming Zhang, Shanyi Du. Critical Void Content for
Thermoset Composite Laminates. J Compos Mater 2009;43:1775–90.
doi:10.1177/0021998306065289.
[70] Mehdikhani M, Gorbatikh L, Verpoest I, Lomov S V. Voids in fiber-reinforced polymer
composites: A review on their formation, characteristics, and effects on mechanical
performance. J Compos Mater 2018:002199831877215. doi:10.1177/0021998318772152.
[71] Léonard F, Stein J, Wilkinson A, Withers P. 3D Characterisation of Void Distribution in
Resin Film Infused Composites. Conf Ind Comput Tomogr 2012;17:69–76.
[72] Sisodia SM, Garcea SC, George AR, Fullwood DT, Spearing SM, Gamstedt EK. High-
resolution computed tomography in resin infused woven carbon fibre composites with voids.
Compos Sci Technol 2016;131:12–21. doi:10.1016/j.compscitech.2016.05.010.
[73] Cooper D, Turinsky A, Sensen C, Hallgrimsson B. Effect of voxel size on 3D micro-CT
analysis of cortical bone porosity. Calcif Tissue Int 2007;80:211–9. doi:10.1007/s00223-
005-0274-6.
[74] Wells J, Kay J, Poursartip A, Lane M, Fernlund G. Surface and Bulk Porosity in Out-of-
Autoclave Prepregs. Proc. Int. Conf. Compos. Mater. (ICCM-20). Copenhagen, Denmark
July 19-24, 2015, p. 12.
[75] Gangloff JJ, Daniel C, Advani SG. A model of two-phase resin and void flow during
composites processing. Int J Multiph Flow 2014;65:51–60.
doi:10.1016/j.ijmultiphaseflow.2014.05.015.
133

[76] Bloom LD, Napper M a., Ward C, Potter K. On the evolution of the distribution of entrapped
air at the tool/first ply interf‘ace during lay-up and debulk. Adv Manuf Polym Compos Sci
2015;1:36–43. doi:10.1179/2055035914Y.0000000005.
[77] Helmus R, Kratz J, Potter K, Hubert P, Hinterholzl R. An experimental technique to
characterize interply void formation in unidirectional prepregs. J Compos Mater
2017;51:579–91. doi:10.1177/0021998316650273.
[78] Gangloff JJ, Cender TA, Eskizeybek V, Simacek P, Advani SG. Entrapment and venting of
bubbles during vacuum bag prepreg processing. J Compos Mater 2017;51:2757–68.
doi:10.1177/0021998316676325.
[79] Little JE, Yuan X, Jones MI. Characterisation of voids in fibre reinforced composite
materials. NDT E Int 2012;46:122–7.
[80] Mehdikhani M, Gorbatikh L, Lomov S V. An overview of research on voids in fiber-
reinforced composites from formation to mechanical effects 2017.
[81] Judd NCW, Wright WW. Voids and their effects on the mechanical properties of
composites- an appraisal. Sampe J 1978;14:10–4.
[82] Tang J-M, Lee WI, Springer GS. Effects of cure pressure on resin flow, voids, and
mechanical properties. J Compos Mater 1987;21:421–40.
[83] Kousourakis A, Mouritz AP, Bannister MK. Interlaminar properties of polymer laminates
containing internal sensor cavities. Compos Struct 2006;75:610–8.
[84] Farouk A, Langrana NA, Weng GJ. Modulus prediction of a cross‐ply fiber reinforced
fabric composite with voids. Polym Compos 1992;13:285–94.
[85] Varna J, Joffe R, Berglund LA, Lundström TS. Effect of voids on failure mechanisms in
RTM laminates. Compos Sci Technol 1995;53:241–9.
[86] Gurdal Z, Tomasino AP, Biggers SB. Effects of processing induced defects on laminate
response-Interlaminar tensile strength 1991.
[87] Huang H, Talreja R. Effects of void geometry on elastic properties of unidirectional fiber
reinforced composites. Compos Sci Technol 2005;65:1964–81.
[88] Stone DEW, Clarke B. Ultrasonic attenuation as a measure of void content in carbon-fibre
reinforced plastics. Non-Destructive Test 1975;8:137–45.
[89] Cinquin J, Triquenaux V, Rousne Y. Porosity influence on organic composite material
mechanical properties. Proc. 16th Int. Conf. Compos. Mater. Japan, 2007, p. 11.
[90] Kosmann N, Karsten JM, Schuett M, Schulte K, Fiedler B. Determining the effect of voids
in GFRP on the damage behaviour under compression loading using acoustic emission.
Compos Part B Eng 2015;70:184–8.
[91] Hayashi T, Takahashi J. Influence of void content on the flexural fracture behaviour of
carbon fiber reinforced polypropylene. J Compos Mater 2017;51:4067–78.
[92] Huang Y, Varna J, Talreja R. Statistical methodology for assessing manufacturing quality
related to transverse cracking in cross ply laminates. Compos Sci Technol 2014;95:100–6.
134

[93] Prakash R. Significance of defects in the fatigue failure of carbon fibre reinforced plastics.
Fibre Sci Technol 1981;14:171–81.
[94] Farhang L, Fernlund G. Out-of-Autoclave Prepreg Laminates. 18Th Int. Conf. Compos.
Mater., 2011.
[95] Centea T, Peters G, Hendrie K, Nutt SR. Effects of thermal gradients on defect formation
during the consolidation of partially impregnated prepregs. J Compos Mater 2017;51:3987–
4003. doi:10.1177/0021998317733317.
[96] Hsiao K. GAS TRANSPORT AND WATER VAPOURIZATION IN OUT-OF-
AUTOCLAVE PREPREG LAMINATES by A THESIS SUBMITTED IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS. 2012.
[97] Cender TA, Gangloff Jr JJ, Simacek P, Advani SG. Void reduction during out-of-autoclave
thermoset prepreg composite processing. Soc Adv Mater Process Eng Tech Conf 2014.
[98] Helmus R, Centea T, Hubert P, Hinterholzl R. Out-of-autoclave prepreg consolidation:
Coupled air evacuation and prepreg impregnation modeling. J Compos Mater
2016;50:1403–13. doi:10.1177/0021998315592005.
[99] Teagle PR. The quality control and non-destructive evaluation of composite aerospace
components. Composites 1983;14:115–28. doi:10.1016/S0010-4361(83)80007-X.
[100] Davis RS. Equation for the Determination of the Density of Moist Air (1981/91). Metrologia
1992;29:67. doi:10.1088/0026-1394/29/1/008.
[101] Howard S. Evaluating Consolidation Dynamics - Study of Key Consolidation Parameters
in a Vacuum Bag Only (VBO) Cure. Proc. SAMPE 2016 Tech. Conf., Long Beach, CA:
2016.
[102] Jianmao T, Springer GS. Effects of Cure and Moisture on the Properties of Fiberite 976
Resin. J Compos Mater 1988;22:2–14.
[103] Sir B, Taylor G, Saffman PG. A note on the motion of bubbles in a Hele-Shaw cell and
porous medium. Quart Journ Mech Appl Math 1959;XII:265–79.
[104] Huang H, Talreja R. Effects of void geometry on elastic properties of unidirectional fiber
reinforced composites. Compos Sci Technol 2005;65:1964–81.
doi:10.1016/j.compscitech.2005.02.019.
[105] Bowles KJ, Frimpong S. Void Effects on the Interlaminar Shear Strength of Unidirectional
Graphite-Fiber-Reinforced Composites. J Compos Mater 1992;26:1487–509.
doi:10.1177/002199839202601006.
[106] Hernández S, Sket F, González C, Llorca J. Optimization of curing cycle in carbon fiber-
reinforced laminates: Void distribution and mechanical properties. Compos Sci Technol
2013;85:73–82. doi:10.1016/j.compscitech.2013.06.005.
[107] Hu W, Grunenfelder LK, Centea T, Nutt S. In-situ monitoring of void evolution in
unidirectional prepreg. J Compos Mater 2018;52:2847–58.
doi:10.1177/0021998318759183.

135

[108] Greenspan L. Humidity fixed points of binary saturated aqueous solutions. J Res Natl Bur
Stand Sect A Phys Chem 1977;81A:89–96. doi:10.6028/jres.081A.011.
[109] Zhou J, Lucas JP. Hygrothermal effects of epoxy resin. Part I: The nature of water in epoxy.
Polymer (Guildf) 1999;40:5505–12. doi:10.1016/S0032-3861(98)00790-3.
[110] Adamson MJ. Thermal expansion and swelling of cured epoxy resin used in graphite/epoxy
composite materials. J Mater Sci 1980;15:1736–45. doi:10.1007/BF00550593.
[111] MAGGANA C, PISSIS P. Water Sorption and Diffusion in an Epoxy Resin System. J
Polym Sci Part B-Polymer Phys 1999;37:1165–82.
[112] Skogestad S. Chemical and Energy Process Engineering, Chapter 7: entropy and
Equilibirum. CRC Press; 2009, p. 179–182. n.d.
[113] Masterton WL, Hurley CN. Chemistry: Principles and Reactions. Cengage Learning; 2015.
[114] Chatterjee A, Jr. JWG. Moisture Absorption Behavior of Epoxies and Their S2 Glass
Composites. J Appl Polym Sci 2008;108:3942–51. doi:10.1002/app.28076.
[115] Kourkoutsaki T, Comas-Cardona S, Binetruy C, Upadhyay RK, Hinterhoelzl R. The impact
of air evacuation on the impregnation time of Out-of-Autoclave prepregs. Compos Part A
Appl Sci Manuf 2015;79:30–42. doi:10.1016/j.compositesa.2015.08.034.
[116] West BPVAN, Pipes RB. The Consolidation of Commingled Thermoplastic 1971.
[117] Hu W, Nutt S. In-situ Montoring of Air Removal During Vacuum Bag-Only (VBO)
Processing. 21st Int. Conf. Compos. Mater., Xi’an, 20-25th, 2017: n.d.
[118] Kratz J. Transport Phenomena in Vacuum Bag Only Prepreg Processing of Honeycomb
Sandwich Panels. McGill University, 2013.
[119] Wisnom MR, Reynolds T, Gwilliam N. Reduction in Interlaminar Shear Strength by
Discreate and Distributed Voids 1996;3538:93–101.
[120] Nikishkov Y, Seon G, Makeev A. Structural analysis of composites with porosity defects
based on X-ray computed tomography 2014. doi:10.1177/0021998313494917.
[121] Grunenfelder LK, Dills A, Centea T, Nutt S. Effect of prepreg format on defect control in
out-of-autoclave processing. Compos Part A Appl Sci Manuf 2017;93:88–99.
doi:10.1016/j.compositesa.2016.10.027.
[122] Schechter SGK, Centea T, Nutt SR. Polymer film dewetting for fabrication of out-of-
autoclave prepreg with high through-thickness permeability. Compos Part A Appl Sci
Manuf 2018;114:86–96. doi:10.1016/j.compositesa.2018.08.002.
[123] Ma Y, Centea T, Nutt SR. Defect Reduction Strategies for the Manufacture of Contoured
Laminates Using Vacuum BAG-Only Prepregs 2017. doi:10.1002/pc.
[124] Schlimbach J, Ogale A. Out-of-autoclave curing process in polymer matrix composites.
Woodhead Publishing Limited; 2012. doi:10.1533/9780857096258.3.435.
[125] Rigard C. Process Selection and Optimization for Out of Autoclave Prepreg Structures. Proc.
CAMX, Dallas, TX: 2015.
136

[126] Hu W, Centea T, Nutt S. Effects of material and process parameters on void evolution in
unidirectional prepreg during vacuum bag-only cure 2019. doi:10.1177/0021998319864420.
[127] Hu W, Nutt S. Effects of “ Super - Ambient ” Dwell on Air Removal in Vacuum Bag - Only
Prepregs. SAMPE Student Res. Symp. C., Anaheim, CA: 2019.

Abstract (if available)

Abstract Vacuum bag-only (VBO) processing allows the manufacturing of high-performance composite in a more flexible and cost-effective way by using atmospheric pressure alone to consolidate composite parts in conventional ovens. However, in the absence of high consolidation pressure, such as that imparted by traditional autoclave processing, VBO-cured parts are more susceptible to voids, which are known to have a negative effect on mechanical properties. To promote air evacuation, VBO prepregs feature a partially impregnated microstructure, consisting of dry fiber tows surrounded by resin-rich regions. Those dry fiber tows provide gas evacuation pathways during the early stages of processing. However, the mechanisms of air evacuation and void formation during VBO processing are not well understood. The purpose of this work is to improve the understanding of void formation and evolution during VBO processing, specifically for those gas-induced voids. ❧ Gas-induced voids are typically located in the resin-rich regions between prepreg plies. The primary sources of gas-induced voids are air entrapped during layup and evolved moisture. Due to the scarcity of experimental techniques to effectively characterize inter-ply entrapped air during processing, insights into the removal of this inter-ply air under vacuum bag cure conditions is limited. In this study, an in situ visualization technique was developed to dynamically observe inter-ply air removal during the cure of an VBO prepreg. Laminates were laid up and bagged on the interior side of a transparent oven window, thereby re-creating a standard VBO cure environment and allowing direct observation in situ. A perforated resin film with controlled pore size and distribution was placed between the glass window and the first prepreg ply, intentionally introducing entrapped air into the lay-up. In this configuration, air is entrapped in resin rich region mimicking the conditions encountered by an air bubble trapped at the mid-plane of a stack of prepreg plies. ❧ The technique was first used to investigate mechanisms of air removal and void evolution in unidirectional prepreg. Bubble transport and the progression of gas evacuation was monitored throughout the VBO cure of the UD prepregs. Void content as a function of time and temperature was determined. Prepreg impregnation was also tracked by the inspection of laminate cross-sections prepared at different times during the cure cycle. Cure kinetics and resin infiltration during cure were predicted using existing models. Based on the relationships between void content, resin properties, and tow impregnation as a function of time, a three-stage air removal mechanism was established. These three stages include (I) air evacuation through dry fiber pathways, (II) bubble expansion after prepreg saturation and moisture diffusion, and (III) bubble shrinkage due to evolving thermochemical conditions and resin properties. Furthermore, bubble morphology and bubble motion were also investigated. A positive correlation was observed between the rate of evacuation and bubble elongation. ❧ After that, we performed a parametric study to determine the mechanisms of inter-ply void evolution in UD prepregs VBO cure, and to identify key factors that affect inter-ply air removal. We employed the same in situ visualization setup for direct, real-time observation of air removal for prepregs during cure under different conditions. Processing parameters including room temperature vacuum hold, reduced vacuum and moisture levels were investigated. Results showed that laminates cured with a 24-hour room temperature vacuum hold exhibit similar bubble behavior and void content with the laminate without vacuum hold, indicating that room temperature vacuum hold is not an effective way for the removal of inter-ply air in UD prepregs. Reduced vacuum quality (80% vacuum) had negligible effects on part quality, while an increase in moisture content of the laminate notably increased void content. ❧ Resin film thickness change was investigated by the inspection of the cross-sections of partially cured laminates. The results showed that resin film thickness decreased from ~ 50 µm to ~10 µm at the end of cure due to resin infiltration. However, despite film thinning effects, air bubble size (porosity) increased ~36% during Stage II, indicating that bubble expansion did occur and was an important mechanism of inter-void evolution during VBO processing. Prepreg moisture content was also tracked by the inspection of laminate water content at different times during the cure cycle, and the data was combined with a diffusion-based analytical model to predict void size and to improve the understanding of void evolution mechanisms. Results indicated that moisture content of the laminate decreased markedly as cure progressed, providing insights into bubble behavior (expansion and shrinkage) observed during cure. The modified model predictions aligned with experimental data, especially during the second stage, confirming that the observed void growth results from moisture diffusion. ❧ Later, the in situ monitoring technique was employed to observe inter-ply air evolution during vacuum bag-only cure of fabric prepregs. Because of the intrinsic surface topography of woven fabric and prepreg, large amount of air is inevitably entrapped between adjacent prepreg plies during layup. The in situ monitoring data showed the three-stage process of air removal was observed during the cure of fabric prepregs, indicating that the key air removal mechanisms in fabric prepregs remained the same. However, during the first stage, air evacuation was significantly faster than that in UD prepreg due to the macro-pores (pinholes) located at the intersections of fiber tows. ❧ A parametric study was also conducted to understand the effects of processing parameters on air removal in fabric prepregs. The results showed that room temperature vacuum hold can effectively evacuate inter-ply air in fabric prepregs. Besides, because of the complete evacuation of inter-ply air, no bubble formation or expansion was observed throughout the subsequent elevated cure. Moisture effects were also examined and an increasing in void content with increasing moisture was observed. ❧ A detailed investigation of the effects of reduced vacuum on inter-ply air removal in fabric prepregs and the underlying void formation mechanisms was performed. Observations showed that reduced vacuum levels resulted in inefficient inter-ply air evacuation during Stage I, a more rapid bubble expansion rate and formation of new air bubbles during Stages II and III. These last two factors governed the generation of high levels of inter-ply porosity in the final parts. Tow impregnation during cure at reduced vacuum conditions was investigated. Both model predictions and experimental data showed that resin infiltration was impeded due to the presence of air in intra-tow regions. However, the final parts showed that tows were fully impregnated in all cases, indicating that the entrapped intra-tow air migrated to inter-ply regions during cure. The interactions between intra-tow and inter-tow air bubbles during cure at deficient vacuum conditions were revealed. Findings led to the conclusion that air remaining in intra-tow regions contributed more to the increase of inter-ply voids than the reduction in consolidation pressure associated with reduced vacuum. ❧ Finally, the effects of heated debulk on air evacuation was investigated. The above research indicated air removal prior to cure is critical for limiting porosity during vacuum bag-only (VBO) processing. However, room temperature debulk is inefficient, especially for large and/or complex parts. Thus, the effects of pre-cure dwell temperature on air evacuation in both plain weave (PW) and unidirectional (UD) prepregs were investigated. The in situ observation showed that increase dwell temperature promotes inter-ply air evacuation. Through-thickness permeability increased with increasing temperature and decreased with number of plies. The in-plane permeability decreased during pre-cure dwell at elevated temperature due to tow impregnation. The findings provide guidelines for cure cycle optimization. For PW laminates, air evacuation during debulk at 60°C was more rapid than at room temperature (RT) debulk if laminate length/ thickness ratio > 5.5. For UD prepregs, a pre-cure dwell at 60°C can be more efficient than RT debulk for laminates (< 1 mm), although thicker UD laminates (> 8 plies) showed no detectable transverse gas flow, even at 60°C. ❧ Overall, the work presented here provides an in situ monitoring method that can be widely applicable to investigated air removal mechanisms under various conditions. The insights gained using this method leads to an improved understanding of the mechanisms of void formation and evolution in VBO processed parts. The results of this investigation provide scientific basis for process optimization and the development of comprehensive models for VBO prepreg processing.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Defect control in vacuum bag only processing of composite prepregs

PDF

Fabrication and analysis of prepregs with discontinuous resin patterning for robust out-of-autoclave manufacturing

PDF

Efficient manufacturing and repair of out-of-autoclave prepreg composites

PDF

Material and process development and optimization for efficient manufacturing of polymer composites

PDF

In situ process monitoring for modeling and defect mitigation in composites processing

PDF

Vacuum-bag-only processing of composites

PDF

Sustainable manufacturing of out-of-autoclave (OoA) prepreg: processing challenges

PDF

Processing and properties of phenylethynyl-terminated PMDA-type asymmetric polyimide and composites

PDF

In situ process analysis for defect control during composites manufacturing

PDF

Developing efficient methods for the manufacture and analysis of composites structures

PDF

Chemical recycling of amine/epoxy composites at atmospheric pressure

Asset Metadata

Creator Hu, Wei (author)

Core Title Void evolution in vacuum bag-only prepregs

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Materials Science

Publication Date 12/10/2020

Defense Date 10/16/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag composite,defect control,manufacturing,OAI-PMH Harvest,out-of-autoclave,process optimization,vacuum bag-only,void formation

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Nutt, Steven (committee chair), Grunenfelder, Lessa K (committee member), Gupta, Satyandra Kumar (committee member)

Creator Email huweiusc@gmail.com,weihu@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-251347

Unique identifier UC11673361

Identifier etd-HuWei-8052.pdf (filename),usctheses-c89-251347 (legacy record id)

Legacy Identifier etd-HuWei-8052.pdf

Dmrecord 251347

Document Type Dissertation

Rights Hu, Wei

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA