DOT/FAA/AR-06/28 The Effect of Peel-Ply Surface

Office of Aviation Research

and Development
Preparation Variables
Washington, DC 20591
on Bond Quality

August 2006

Final Report

This document is available to the public through the National

Technical Information Service (NTIS), Springfield, Virginia 22161.

U.S. Department of Transportation

Federal Aviation Administration
 NOTICE

This document is disseminated under the sponsorship of the U.S.

Department of Transportation in the interest of information exchange.
The United States Government assumes no liability for the contents or
use thereof. The United States Government does not endorse products
or manufacturers. Trade or manufacturer's names appear herein
solely because they are considered essential to the objective of this
report. This document does not constitute FAA certification policy.
Consult your local FAA aircraft certification office as to its use.

This report is available at the Federal Aviation Administration William J.

Hughes Technical Center’s Full-Text Technical Reports page:
actlibrary.tc.faa.gov in Adobe Acrobat portable document format
(PDF).
 Technical Report Documentation Page
1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.
DOT/FAA/AR-06/28
4. Title and Subtitle 5. Report Date

THE EFFECT OF PEEL-PLY SURFACE PREPARATION VARIABLES ON August 2006

BOND QUALITY 6. Performing Organization Code

7. Author(s) 8. Performing Organization Report No.

Brian Flinn and Molly Phariss

Performing Organization Name and Address 10. Work Unit No. (TRAIS)
Department of Material Science and Engineering
University of Washington 11. Contract or Grant No.
Seattle, WA 98103
12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered

U,S. Department of Transportation Final Report

Federal Aviation Administration 14. Sponsoring Agency Code

Office of Aviation Research and Development ANM-115N

Washington, DC 20591
15. Supplementary Notes
The Federal Aviation Administration Airport and Aircraft Safety R&D Division COTR was Peter Shyprykevich.
16. Abstract

To further understand the effect of the variables present in surface preparation on the durability of cobonded and secondary-
bonded composite joints, surface analysis coupled with mechanical testing and fractography were used to analyze samples
prepared using peel-ply removal as the sole surface preparation technique. Laminates were made from aerospace carbon fiber
prepreg and bonded with two different film adhesives. Nylon and polyester peel plies, and siloxane-coated release fabrics, were
investigated to determine the effect of peel-ply material on bond quality and to examine why some peel-ply/adhesive systems are
incompatible. Varying weaves of nylon and polyester peel plies were used to investigate the effect of peel-ply texture (weave) on
bond quality. The moisture content of polyester peel plies was varied to examine its effect on bond quality. Laminate surfaces
and peel plies were analyzed after peel-ply removal through scanning electron microscopy (SEM) and x-ray photoelectron
spectroscopy (XPS); the results of which were correlated to Mode I double cantilever beam strain energy release rates to
determine bond quality. Uncoated polyester peel plies were easily removed from laminate surfaces after curing and produced
good bonds with both adhesives for all textures (GIC> 850 J/m2, cohesive failure). Super Release Blue-coated polyester peel ply
created surfaces that bonded very poorly in both cases (GIC< 94 J/m2, adhesion failure), the result of the transfer of the peel-ply
siloxane coating to the composite surface. Nylon peel plies were more difficult to remove from the laminate and, in the coarser
weaves, could not be removed without damaging the laminates. Laminate surfaces prepared with nylon peel plies bonded well
with AF555 (GIC> 750J/m2, cohesive failure). Laminate surfaces prepared with nylon peel plies bonded poorly with MB1515-3
(GIC<150 J/m2, adhesion failure). This may be explained by the transfer of nylon to the prepared surface, which was found during
SEM and XPS analysis. Peel-ply texture or peel-ply moisture content had no significant effect on fracture energy or mode of
failure.

17. Key Words 18. Distribution Statement

Adhesive bonding, Surface preparation, Carbon fiber This document is available to the public through the National
composites, Wettability Technical Information Service (NTIS), Springfield, Virginia
22161
19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Price
Unclassified Unclassified 35

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

 TABLE OF CONTENTS

Page

EXECUTIVE SUMMARY ix

1. INTRODUCTION 1

1.1 Motivation 1
1.2 Related Previous Work 1
1.3 Primary Preparation Technique 2

2. EXPERIMENTAL APPROACH 3

2.1 Sample Preparation 3

2.2 Characterization 7

2.2.1 X-Ray Photoelectron Spectroscopy 7

2.2.2 Scanning Electron Microscopy 8
2.2.3 Contact Angle Measurement 8

2.3 Mechanical Testing and Fractography 9

3. RESULTS AND DISCUSSION 9

3.1 Effect of Peel-Ply Material 9

3.2 Effect of Peel-Ply Texture 16
3.3 Effect of Peel-Ply Moisture 21
3.4 Determination of Surface Energy by Contact Angle 21

4. CONCLUSIONS 25

5. REFERENCES 25

iii
 LIST OF FIGURES

Figure Page

1 Photomicrograph of Cured Part Cross Section 2

2 Four Possible Fracture Modes in Peel Ply Removal 3

3 Scanning Electron Microscopy Micrographs of PFG Peel-Ply Fabrics Used

in This Study 5

4 Double Cantilever Beam Specimen and Panel Geometry and Dimensions, Plan View 6

5 Peel-Ply Removal With no Observable Laminate Damage 7

6 Examples of Laminates Damaged During Removal of Some Nylon Peel Plies 7

7 Double Cantilever Beam Test Results of Laminates Prepared With Polyester or SRB
Peel Ply and Bonded With MB1515-3 10

8 Mode I Strain Energy Release Rate of Laminates Bonded With AF555 or MB1515-3 11

9 Scanning Electron Microscopy of Composite Surface After Removal of Polyester,

Nylon, and SRB 12

10 Polyester, Nylon, and SRB Fabric After Removal From Laminate 13

11 Nylon-Prepared Composite Surface 13

12 X-Ray Photoelectron Spectroscopy Composition Scan Spectra of Control Laminate

Surface, Polyester, Nylon, and SRB Peel Plies 14

13 High-Resolution C (1s) Spectra of Epoxy Control, Polyester, Nylon, and SRB 15

14 Laminate Surfaces After Removal of Peel Plies 17

15 Peel Plies After Removal From Cured Laminate 18

16 Scanning Electron Microscopy Image of Removed PFG Peel-Ply 60001 and 41661
Illustrating Damage to Peel-Ply Filaments in the Latter 19

17 Average GIC of Laminates Prepared With Various Peel Plies and Bonded With
AF555 or MB1515-3 20

18 Average GIC of Laminates Prepared With 60001 Peel Ply Conditioned at Various
Humidity Levels and Bonded With AF555 21

iv
19 Linear Kaelble Plot for Surface Energy Measurement From Contact Angle Data on
BMS8-276 Laminate Prepared With PFG 60001 Peel Ply 22

20 Wettability Envelopes for BMS8-276 Laminates Prepared With PFG Polyester

Peel Plies 23

21 Wettability Envelopes for BMS8-276 Laminates Prepared With PFG Nylon

Peel Plies 24

22 Wettability Envelopes for BMS8-276 Laminates Prepared With PFG Nylon

Peel Ply 51879 and Polyester Peel Ply 60001 24

v
 LIST OF TABLES

Table Page

1 Weave Characteristics of PFG Peel-Ply Fabrics Used in This Study 4

2 Fracture Surfaces of Tested Laminates 10

3 Scanning Electron Microscopy Image of Adhesive Side of Fracture Surfaces From

DCB Specimens 12

4 Relative Concentrations of Elements Found on Laminate Surfaces by XPS 14

5 Carbon Bonding in Samples After Peel-Ply Removal 15

6 Summary of Average GIC Data and Failure Mode for Laminates Prepared With
Various Peel Plies and Bonded With AF555 and MB1515-3 20

7 Contact Angle Measurements and Surface Energies for a Variety of Fluids Measured
on BMS8-276 Laminates Prepared With Different Peel Plies 22

vi
 LIST OF ACRONYMS

DCB Double cantilever beam

DMSO Dimethyl sulfoxide
PFG Precision Fabrics Group
RH Relative humidity
SEM Scanning electron microscopy
SRB Super Release Blue
XPS X-ray photoelectron spectroscopy

vii/viii
 EXECUTIVE SUMMARY

The understanding of composite bonding lags behind that of metals. Although techniques to
achieve good bond strength in composites have been developed, the role of surface-preparation
techniques, such as peel-ply removal at the fundamental level, is poorly understood. Peel-ply
removal is widely used for bonding composite primary structures, but is a technique with several
variables which the effect on bond quality has not been fully investigated.

To further understand the effect of the variables present in surface preparation on the durability
of cobonded and secondary-bonded composite joints, surface analysis, coupled with mechanical
testing and fractography, were used to analyze samples prepared using peel-ply removal as the
sole surface preparation (no abrasion). Nylon and polyester peel plies, and siloxane-coated
release fabric, were investigated to determine the effect of peel-ply material on bond quality and
to examine why some peel-ply/adhesive systems are incompatible. To investigate the effect of
peel-ply texture (weave) on the quality of cobonded and secondary-bonded composite joints,
laminates were prepared using nylon and polyester peel plies with fine to rough weaves. The
effect of peel-ply moisture content was investigated by preparing laminates with polyester peel
plies with different moisture contents. The laminates were made from aerospace carbon fiber
prepreg and bonded with two different film adhesives. Laminate surfaces and peel plies were
analyzed after peel-ply removal through scanning electron microscopy (SEM) and x-ray
photoelectron spectroscopy (XPS); the results of which were correlated to Mode I double
cantilever beam strain energy release rates to determine bond quality. Finally, fractography was
performed with SEM to determine the mode of fracture.

Uncoated polyester peel plies were easily removed from laminate surfaces after curing and
produced good bonds with both adhesives for all textures (GIC> 850 J/m2, cohesive failure).
Super Release Blue-coated polyester peel ply created surfaces that bonded very poorly in both
cases (GIC< 94 J/m2, adhesion failure), the result of the transfer of the peel-ply siloxane coating
to the composite surface. Nylon peel plies were more difficult to remove from the laminate and,
in some cases, could not be removed without damaging the laminates. Finer weaves were easier
to remove than coarser weaves. Laminate surfaces prepared with nylon peel plies bonded well
with AF555 (GIC> 750J/m2, cohesive failure). Laminate surfaces prepared with nylon peel plies
bonded poorly with MB1515-3 (GIC<150 J/m2, adhesion failure). This may be explained by the
transfer of nylon to the prepared surface, which was found during SEM and XPS analysis. Peel-
ply texture or peel-ply moisture content had no significant effect on fracture energy or mode of
failure.

ix/x
1. INTRODUCTION.

1.1 MOTIVATION.

Joints represent one of the greatest challenges in the design of structures in general and of
composite structures in particular. Joints entail interruptions in the geometry of a structure and
often material discontinuities. The latter almost always produce localized areas of high stress.

In principle, adhesive joints are structurally more efficient than mechanically fastened joints;
therefore, they offer the potential for structural weight saving because they provide better
opportunities for eliminating stress concentrations. For example, advantage can be taken of
ductile response of the adhesive to reduce stress peaks [1]. Mechanically fastened joints tend to
use the available material inefficiently. However, adhesive joints tend to lack structural
redundancy and are highly sensitive to manufacturing deficiencies, including poor bonding
technique, poor fit of mating parts, and sensitivity of the adhesive to temperature and
environmental effects such as moisture [2]. Assurance of bond quality has been a continuing
problem in adhesive joints. While ultrasonic and x-ray inspection may reveal gaps in the bond,
no technique currently available can guarantee that a bond that appears to be intact does, in fact,
have adequate load transfer capability.

Most joint design principles are irrelevant if adhesion between the adherends is poor. Integrity
of a bond hinges on strong primary chemical bonding and, to a lesser degree, mechanical factors
[3]. Surface preparation is the most important factor because it can remove contamination from
the environment or left from processing and can create chemically active surfaces [1]. It is,
therefore, of the utmost importance that surfaces be properly prepared to create strong bonds.
Surface preparation and bonding techniques have been well developed for metal-to-metal
bonding; however, this is not the case for composite-to-composite or composite-to-metal
bonding. Techniques to achieve good bond strength in composites have been developed, but
there is not a fundamental understanding of the role of surface preparation techniques at the
atomic level.

1.2 RELATED PREVIOUS WORK.

Sanding, grit blasting, and peel-ply removal are three common methods used to prepare
composite surfaces for adhesive-bonding [4]. The effect of surface roughness and material
removal on bond quality has been investigated for abrasive surface preparation techniques [2 and
5]. Bossi showed an increase in strain energy release rate in grit-blasted samples as surface
roughness increased, and a corresponding change in failure mode from interfacial (adhesion) to
cohesive and finally interlaminar [5]. No published work has been found with respect to the
effect of surface texture from peel ply prepared on bond quality. The moisture content of
prepregs, laminates, and adhesives has been shown to have a strong, negative effect on bond
quality [6]. Quality control procedures are commonly used to ensure low-moisture content in
prepregs, laminates, and adhesives. Hart-Smith has mentioned that moisture absorbed in nylon
peel plies has been identified as a contributing problem [6]. Armstrong vaguely references
research by J.R.J Wingfield on the effect of peel-ply moisture, but no publications were found
[7]. The authors have not been able to find any published research that investigated the effect of

1
wet polyester peel ply on surface preparation, or any manufacturers’ specifications with respect
to moisture content or storage requirements for peel ply.

1.3 PRIMARY PREPARATION TECHNIQUE.

Peel ply is added as the last layer in the lay-up of the composite part to be bonded. During cure,
the epoxy in the part becomes viscous and flows into gaps in the peel ply. This can be seen in
the cross-section of the Precision Fabrics Group (PFG), Inc., 40000 nylon peel ply on a cured
laminate shown in figure 1. The peel ply is removed from the surface immediately before
bonding.

FIGURE 1. PHOTOMICROGRAPH OF CURED PART CROSS SECTION

The characteristics of a surface created by peel ply removal are directly influenced by how the
peel ply separates from the laminate. The possible types can be seen in figure 2, and are either
(1) the fracture of the epoxy resin between the peel ply and the underlying carbon fibers (dark
blue) or (2) interfacial fracture between the peel-ply fabric fibers and the epoxy matrix (pink). In
severe cases of poor surface preparation, the peel-ply fibers may fracture and leave material on
the composite surface (3 green) or there may be interlaminar failure in the composite (4
turquoise). In the first type, a fresh epoxy surface is created that should be chemically active and
easily bonded. Though ideal, this is rarely the only mode of fracture present. In the second type,
the chemistry of the surface created may be affected by the nature of the peel-ply material
surface. Peel-ply coatings or fiber surface treatments may be transferred to the surface to be
bonded and affect the future bond [8]. The third type may occur if the bond between the peel ply

2
and epoxy is stronger than the peel-ply fibers and the fourth if the interlaminar strength of the
laminate is low or the peel ply is removed incorrectly.

FIGURE 2. FOUR POSSIBLE FRACTURE TYPES IN PEEL PLY REMOVAL

Additionally, peel plies that may work for one resin or adhesive system may be ineffective with
others. Although trial and error in industrial practice has shown what combinations create strong
bonds, there remains a lack of understanding of the mechanisms occurring on a fundamental
level. The goal of this research is, therefore, to further understand the effect of peel-ply fabric
characteristics on laminate surfaces being prepared for bonding with film adhesives through
surface analysis coupled with mechanical testing and fractography.

2. EXPERIMENTAL APPROACH.

2.1 SAMPLE PREPARATION.

Unidirectional ten-ply laminates [0]10T were laid up using 0.30 by 0.50 m, 176oC cure aerospace
carbon fiber/epoxy prepreg strips and covered with a layer of polyester or nylon peel- ply fabric,
and then vacuum bagged. Laminates were produced using nine different peel-ply fabrics from
PFG listed in table 1. The texture and weave of the as-received plies can be seen in figure 3.

3
 TABLE 1. WEAVE CHARACTERISTICS OF PFG PEEL-PLY FABRICS USED
IN THIS STUDY

Warp Fill Peel Ply

Filament Precision (ends/mm) (picks/mm) Thickness
Material Fabrics Code (ends/in.) (ends/in.) (mm) Comments
Polyester 60001 2.75 (70) 1.97 (50) 0.13-0.15 BMS8-308
Polyester 60001 SRB 2.75 (70) 1.97 (50) 0.13-0.15 Siloxane Coated
Polyester 60004 VLP 4.72 (120) 2.32 (59) 0.13-0.15 Calendered
Polyester 60004 4.72 (120) 2.32 (59) 0.11- 0.14
Polyester 60005 3.54 (90) 2.28 (58) 0.15-0.18
Nylon 6,6 51789 6.30 (160) 4.06 (103) 0.11-0.14 Very Fine
Nylon 6,6 52008 3.98 (101) 3.23 (82) 0.10-0.13
Nylon 6,6 50000 2.36 (60) 1.97 (50) 0.17-0.19 Twill weave
Nylon 6,6 40000 2.99 (76) 2.01 (51) 0.19-0.22
Nylon 6,6 41661 2.36 (60) 1.97 (50) 0.17-0.19 Coarse

The peel plies were chosen to produce laminate surfaces for bonding with a wide range of
surface finishes. All peel plies were stored at ambient conditions (20oC and 50% +10% relative
humidity (RH)). In addition, one commonly used polyester peel ply (60001) was selected for
conditioning at various temperatures and humidity levels prior to laminate fabrication to
investigate if moisture in the peel ply affects the surface to be bonded. The 60001 peel-ply
samples were weighed and then placed in environmental chambers at 25oC-90% RH and 60oC-
90% RH. The mass change of the peel-ply samples was measured at 0.5, 1, 2, 4, and 18 hours.
One sample of peel ply was placed in a drying oven at 60oC for 18 hours. The sample
conditioned at 60oC-90% RH gained 25% mass after 0.5 hours and then stabilized. The sample
conditioned at 25oC-90% RH did not exhibit any measurable change in mass. Laminates were
fabricated using the wet and dry peel plies within 1 hour of removal from conditioning chamber.

4
 Polyester Peel-Ply Fabrics Nylon Peel-Ply Fabrics

60001

51789
60001SRB

52008
60004VLP

50000
60004VLP

40000
60005

41661

FIGURE 3. SCANNING ELECTRON MICROSCOPY MICROGRAPHS OF PFG PEEL-PLY

FABRICS USED IN THIS STUDY

5
All laminates underwent a standard cure cycle of 176oC for 2 hours under 0.58 MPa in an
autoclave with full vacuum maintained throughout the cure cycle. The ends of the laminates
were then removed for surface characterization with peel ply intact. After trimming the laminate
panels to dimensions specified in figure 4, peel plies were removed immediately before
application of epoxy-based film adhesives MB1515-3 or AF555. A strip of nonporous
fluorinated ethylene propylene release film was placed between the halves to be bonded to create
an initial debond for fracture toughness testing as seen in figure 4. Bonded samples were cured
using the same conditions as the laminates.

FIGURE 4. DOUBLE CANTILEVER BEAM SPECIMEN AND PANEL GEOMETRY AND

DIMENSIONS, PLAN VIEW

All polyester peel plies were easily removed from the laminates with no damage to laminate or
peel-ply filaments observed, as shown in figure 5. Nylon peel plies were difficult to remove
from the cured laminates and required much greater force to peel away the fabric. Despite
several attempts by experienced technicians, PFG nylon peel plies coded 52008, 50000, and
41661 could not be removed without causing significant damage to the laminates, as shown in
figure 6, Interlaminar fracture occurred during peel ply removal leaving carbon fiber attached to
the nylon peel plies. Only laminates from which the peel ply could be removed without damage
to laminate were approved for bonding. Hence, there is no GIC data for samples prepared with
52008, 50000, or 41661 peel plies. Double cantilever beam (DCB) specimens were machined
from the coupon after bonding, as shown in figure 4.

6
 FIGURE 5. PEEL-PLY REMOVAL WITH NO OBSERVABLE LAMINATE DAMAGE

FIGURE 6. EXAMPLES OF LAMINATES DAMAGED DURING REMOVAL OF SOME

NYLON PEEL PLIES

2.2 CHARACTERIZATION.

2.2.1 X-Ray Photoelectron Spectroscopy.

When Surface Science Instruments M-Probe spectrometer was used for the x-ray photoelectron
spectroscopy (XPS) analyses of the laminate surface in a high vacuum, the environment pressure
did not exceed 1.3x10-6 Pa. The peel-ply material was removed from laminates by peeling back
from the center of the short edge of the coupons immediately prior to analysis. A cured laminate
was split and the untreated surface scanned as a control. Survey scans and high-resolution XPS
spectra over the C (1s) peak were recorded in three spots per sample to find chemical
composition using nominal pass energies of 150 and 50 eV, respectively. Spectra were obtained
from elliptical spots with major axes of approximately 1.7 mm and minor axes of approximately
0.4 mm, to depths of between 10 and 20 atom layers of the sample surface. The samples were
nonconducting; therefore, a low-energy electron load-gun set at ~4.0 eV was used for charge

7
neutralization of the samples. All spectra measurements were corrected for the charging by
setting the C (1s) peak to 285.0 eV.

2.2.2 Scanning Electron Microscopy.

Scanning electron microscopy (SEM) is a technique in which the emission of secondary

electrons from a sample is analyzed to create magnified images of a surface. Insulating samples
must be covered in a thin conductive coating before analysis. An electron beam of an energy
ranging from a few keV to 50 keV is emitted from a cathode filament with high-melting point
and low-vapor pressure, such as tungsten or lanthanum hexaboride LaB6, an anode, or via field
emission. The electron beam is focused by two successive condenser lenses into a beam with a
very fine spot size (~ 5 nm), passes through the objective lens, and is deflected by pairs of
scanning coils either linearly or in a raster fashion over a rectangular area of the sample surface.
As the primary electrons strike the surface, they are inelastically scattered by atoms in the sample
and cause emission of electrons that are then detected to produce an image. The most common
imaging mode monitors low-energy (<50 eV) secondary electrons. SEM gives images with good
three-dimensional contrast because the surface area, which is relatively small for a flat surface
but increases for a steep surface, that is exposed to the primary beam determines the brightness
of the signal.

The spatial resolution of the SEM falls somewhere between 3 and 20 nm and depends on the size
of the electron spot, which is determined by magnetic electron-optical system that produces the
scanning beam. The resolution is also limited by the extent of the material that interacts with the
electron beam. The SEM cannot image down to the atomic scale, because the spot size and
interaction volume are very large compared to the distances between atoms. The SEM can,
however, image a comparatively large area of the specimen, image bulk materials, and has a
variety of analytical modes available for measuring the composition and nature of the
specimen [9].

In this study, SEM was used to characterize the morphology of the peel plies before and after use
and the laminate surface created by peel-ply removal, yielding visual topographic results. In
addition, fractography was performed using SEM. Postfracture images show whether specific
features acted as crack initiation sites, and they can distinguish between interfacial failure and
thin-layer cohesive failure. SEM can also expose if there are small patches of interfacial failure
in a primarily cohesive failure (or vice versa).

2.2.3 Contact Angle Measurement.

As is clear from a review of adhesion theory, contact angle measurements are a very powerful
tool in understanding surfaces. They are also very simple to perform. A liquid drop is placed
(not dropped in order to decrease the gravitational effect) on a surface that does not swell or
react. Very viscous liquids are given some time to equilibrate. A goniometer is then used to
measure the angle between the surface and the drop, with the crosshairs aimed at the drop edge
[10]. At the University of Washington, contact angles are measured with a Ramé-Hart Tilting
Contact Angle Goniometer Model 100-00115 with an overall magnification of 23 times and
working distance 57 mm. The measurement is dependent on the direction it is taken. The
advancing angle of a drop will be greater than the receding angle. This is known as contact

8
angle hysteresis, and is caused by nonhomogeneous-surface chemistry, surface roughness, and
molecular rearrangement in the solid from the liquid (or vice versa). All measurements reported
here are advancing angle. The fluids used were deionized water, dimethyl sulfoxide (DMSO),
glycerol, and ethylene glycol. Contact-angle measurements were taken 5-10 seconds after the
drop was placed on the laminate surface. Ten measurements were taken on each laminate using
the Kaelble approach as discussed by Bossi, et al. [5] surface. Using the Kaelble approach as
discussed by Bossi, et al. [5], surface energies and wettability envelopes were calculated by
fitting.

2.3 MECHANICAL TESTING AND FRACTOGRAPHY.

A DCB test was performed according to ASTM D5528-01 standards using a servo-hydraulic test
frame with a 200N-load cell and 25-mm/min ramp rate on bonded specimens to determine GIC
values. Generally used to evaluate composite parts, this test was used in this study as a measure
of joint bond quality. Ten samples prepared with each peel ply were tested. Fractography was
performed using SEM. Postfracture images show whether specific features acted as crack
initiation sites, and they can distinguish between interfacial failure and thin-layer cohesive
failure. SEM can also expose if there are small patches of interfacial failure in a primarily
cohesive failure (or vice versa).

3. RESULTS AND DISCUSSION.

3.1 EFFECT OF PEEL-PLY MATERIAL.

To examine the effect of peel-ply material on bond quality, the results from Mode I DCB
fracture toughness test of 60001, 52008, and 60001 Super Release Blue (SRB)-prepared surfaces
were compared. The DCB test produced a load displacement plot, data for crack extension, a
failure mode, and a corresponding critical-strain energy release rate. The load displacement
plots were very indicative of the crack events that occurred during the DCB test, as shown in
figure 7.

In both load displacement plots, the initial linear portion of the curve shows where the specimen
is being loaded through the precracked region up to the peak of the curve where a new crack
begins. The initial slopes are dissimilar due to differing initial crack lengths. The laminates
cured with an SRB-release ply displayed steady crack growth, as indicated by the nearly linear
trend of decreasing load. The polyester-prepared specimen displayed stable crack growth up to a
point upon which rapid (nonsteady) crack growth occurred, as indicated by the instantaneous
drop in load. The linear portion tracing back to zero load and displacement shows the unloading
of the specimen. Based on the combination of the DCB plots and the fracture surfaces shown in
table 2, it seems that stable crack propagation predominated in all samples. Three types of
failure were observed: (1) adhesion failure at the adhesive/adherend bondline, (2) cohesive
failure in the adhesive layer, and (3) a mix of cohesive and interlaminar failure. Types of failure
were confirmed by SEM.

9
 35.0

30.0
Polyester

SRB
25.0

Load (N) 20.0

15.0

10.0

5.0

0.0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0
Displacement (mm)

FIGURE 7. DOUBLE CANTILEVER BEAM TEST RESULTS OF LAMINATES PREPARED

WITH POLYESTER OR SRB PEEL PLY AND BONDED WITH MB1515-3

TABLE 2. FRACTURE SURFACES OF TESTED LAMINATES

Polyester Prepared Nylon Prepared SRB Prepared

AF555

Failure Type Cohesive Cohesive & Interlaminar Adhesion

MB1515-3

Failure Type Cohesive Adhesion Adhesion

Sample width = 12.7

10
 1000.0
910.7 909.6
900.0
812.3
800.0

700.0

600.0

AF 555
500.0
MB 1515-3

400.0

300.0

200.0
122.1
93.9 86.0
100.0

0.0

52008 60001 SRB

(Nylon) (Polyester) (Polyester)
Peel Ply

FIGURE 8. MODE I STRAIN ENERGY RELEASE RATE OF LAMINATES BONDED

WITH AF555 OR MB1515-3

The measure of bond quality for this purpose is the critical-strain energy release rate of the
bonded laminate, determined by Mode I DCB testing. As shown in figure 8, surfaces prepared
with SRB displayed consistently poor bonding. Samples prepared with polyester peel ply had
the best consistent bond quality. A dramatic change in fracture energy was observed when
nylon-prepared surfaces were bonded with MB1515-3 rather than AF555. The fracture type also
changed from cohesive (AF555) to adhesion (MB1515-3), as shown in table 2. The type of
failure (cohesive/interlaminar) seen in samples prepared with polyester peel ply and nylon peel
ply bonded with AF555 is desirable. The adhesion (interfacial) failure seen in the other samples
is unacceptable.

The interfacial (adhesion) failure of samples prepared with nylon and MB1515-3 or with SRB is
reflected in their very low GIC values. The fracture surfaces of the bonded laminates examined
in the SEM can be seen in table 3. The examples of adhesion failure are taken of the cured
adhesive side of the fracture, and show very clearly that the texture of the peel-ply transfers
directly to the surface of the sample. This suggests that peel-ply texture may also be a factor in
bond quality.

To investigate the mechanisms behind varying bond quality, SEM was used to examine the
laminate surfaces after peel-ply removal and before bonding, as can be seen in figure 9. At this
level, the surface quality of all of the samples looked very similar with the resin surface
replicating the peel ply weave pattern. From what can be seen by SEM, all three samples seem
to have at least acceptable surface preparation. There appears to be simple interfacial failure
between the peel-ply filaments and the laminate in all three cases, and some epoxy fracture in
between.

11
 TABLE 3. SCANNING ELECTRON MICROSCOPY IMAGE OF ADHESIVE SIDE OF
FRACTURE SURFACES FROM DCB SPECIMENS

Polyester Prepared Nylon Prepared SRB Prepared

AF555

Cohesive Cohesive Adhesion

MB1515-3

Cohesive Adhesion Adhesion

(a) (b) (c)

FIGURE 9. SCANNING ELECTRON MICROSPCOPY OF COMPOSITE SURFACE AFTER

REMOVAL OF (a) POLYESTER, (b) NYLON, AND (c) SRB

Examination of the peel plies (in figure 10) that have been removed from the surfaces confirms
epoxy impregnation in the gaps where warp and fill meet and, in some cases, between the
filaments—the source of the freshly fractured epoxy surfaces.

A significant finding was that, in the nylon and SRB samples, the peel ply showed evidence of
damage. There were nicks and strips exfoliated from the filaments. The nylon-prepared surface
in figure 11 reveals corresponding blemishes and small fibers, suggesting that parts of the peel
ply are being left behind on the surface. In the case of the SRB, the siloxane coating is probably
what has been transferred.

12
 (a) (b) (c)

FIGURE 10. (a) POLYESTER, (b) NYLON, AND (c) SRB FABRIC AFTER REMOVAL
FROM LAMINATE

FIGURE 11. NYLON-PREPARED COMPOSITE SURFACE

The transfer of material from the peel plies was investigated through XPS analysis. Figure 12
shows the spectra collected during the survey scan of the composite surfaces for composition in
comparison with a spectrum of an epoxy control, while table 4 gives the amounts of each
element shown by the peaks.

13
 (a) (b)

(c) (d)

FIGURE 12. X-RAY PHOTOELECTRON SPECTROSCOPY COMPOSITION SCAN

SPECTRA OF (a) CONTROL LAMINATE SURFACE, (b) POLYESTER, (c) NYLON
AND (d) SRB PEEL PLIES

TABLE 4. RELATIVE CONCENTRATIONS OF ELEMENTS FOUND ON LAMINATE

SURFACES BY XPS

Peel Ply C O N Si S
Epoxy Possible
Control 81.1 12.9 5.0 Trace Trace
Possible
Polyester 74.8 22.5 1.8 Trace 0.0
Nylon 76.1 13.3 10.3 Trace 0.0
SRB 66.5 15.6 2.6 2.9 Trace

The high-carbon content in all samples arose from the epoxy matrix and carbon fibers. The
polyester-prepared surfaces seem to have been substantially oxygenated, which has been shown
to increase bond quality [2]. The sulfur seen in the polyester and SRB samples was most likely
the cross-linking agent in the epoxy of the prepreg or perhaps toughening agents in the epoxy
matrix. The low-fracture toughness of the SRB peel-ply specimen was explained by the
presence of the significant Si contamination (most likely in the form of siloxane from the peel
ply coating) exposed in the survey scan. The high amount of nitrogen found on the surface
prepared with nylon peel ply and additional molecular information were examined more closely
through high-resolution scans of the C (1s) region (figure 13). Table 5 shows the respective
amounts of the species given by these peaks.

14
 (a) (b)

(c) (d)

FIGURE 13. HIGH-RESOLUTION C (1s) SPECTRA OF (a) EPOXY CONTROL, (b)

POLYESTER, (c) NYLON, AND (d) SRB

TABLE 5. CARBON BONDING IN SAMPLES AFTER PEEL-PLY REMOVAL

Binding Energy
Peel Ply Species (eV) %
CC/CH 285 83.8
Epoxy Control
CO/CN 286.4 16.2
CC/CH 285 74.9
Polyester CO 286.7 15.6
C=O 289.1 9.5
CC/CH 285 74.3
Nylon CO/CN 286.2 14.6
Amide (NC=0) 287.9 11
CC/CH 285 79.1
SRB CO 286.9 11.4
C=O 289.2 9.4

15
The amide group found on the nylon sample confirmed that some of the nylon had been
transferred to the laminate surface. This may be the key to the poor bonding of the sample
prepared with nylon peel ply and MB1515-3.

Based on the SEM and XPS results, it seemed that there are variations of the second form of
fracture (discussed in the introduction) when peel ply is removed from a laminate surface. The
polyester peel ply did not transfer any detectable material onto the resin surface. It was also
likely that with polyester the composite resin was indeed fracturing at the interface between
filament and resin. This interfacial fracturing was an important distinction from what occurred
with the nylon and SRB peel plies. The nylon and SRB peel plies also have a superficial fracture
type 2 as well. However, in contrast to the polyester, they have transferred material to the resin
surface. The composite resin and nylon filaments reacted during curing to create a bond strong
enough to rip some of the nylon off during peel-ply removal. The SRB filaments have a siloxane
coating to facilitate removal from the cured composite, and some of this coating was left behind
during removal from the cured composite. Due to the proprietary nature of adhesives, it was
difficult to determine why AF555 bonded well and MB1515-3 bonded poorly to amide
contaminated surfaces.

3.2 EFFECT OF PEEL-PLY TEXTURE.

The texture of the peel-ply fabric directly controlled the laminate surface topography when the
predominate mode of fracture during peel-ply removal was along the interface between the peel-
ply filaments and the epoxy matrix (type 2 in figure 2). This was observed in all laminates
fabricated with polyester peel plies, as shown in figure 9. During the cure cycle, the epoxy
flowed around the filaments in the peel-ply fabric and subsequently gelled and cured. When the
peel ply was removed, a negative of the peel-ply fabric was left in the epoxy matrix above the
carbon fibers in the laminate. This is clearly illustrated by comparing mating surfaces of the peel
plies in figure 3 and the laminate surfaces after peel-ply removal in figure 9. On the macroscopic
scale, similar results were observed in laminates fabricated with PFG nylon peel plies 52008 and
40000 that were removed without observable damage to laminates. The PFG nylon peel plies
with coarser weaves (50000 and 41661) adhered to the laminates during cure and upon removal,
interlaminar fracture occurred and carbon fibers were removed from the laminate, as shown in
figure 14.

The peel plies removed from the laminates were also examined in the SEM and are shown in
figure 15 where both the peel-ply filaments and the fracture epoxy matrix are visible. The
filaments in the polyester peel plies appear similar to the filaments in the as-received peel plies
shown in figure 3. The filaments in the nylon peel plies have a more tortured appearance; the
surfaces are not clean and smooth, but rather ragged and torn. A higher-magnification view of
the removed peel-ply filaments is shown in figure 16.

16
 Polyester Peel-Ply Fabrics Nylon Peel-Ply Fabrics

60001

51789
60004

52008
60004 VLP

50000
60005

40000
41661

THIS SPACE BLANK

FIGURE 14. LAMINATE SURFACES AFTER REMOVAL OF PEEL PLIES

(Arrows note areas of damage.)

17
 Polyester Peel-Ply Fabrics Nylon Peel-Ply Fabrics

60001

51789
60004

52008
60004 VLP

50000
60005

40000
41661

THIS SPACE BLANK

FIGURE 15. PEEL PLIES AFTER REMOVAL FROM CURED LAMINATE

(Arrows note damaged peel-ply filaments or carbon fibers from laminate surface.)

18
 FIGURE 16. SCANNING ELECTRON MICROSCOPY IMAGE OF REMOVED PFG
PEEL-PLY 60001 AND 41661 ILLUSTRATING DAMAGE TO PEEL-PLY FILAMENTS
IN THE LATTER

The measured fracture energy, or strain energy release rate, GIC, of the bonded samples
(averaged from the 10 tested samples from each peel ply) is presented in table 6 and figure 17 as
a function of peel-ply material and weave (surface texture) and adhesive. The bond quality of
laminates produced with nylon peel plies was very sensitive to the adhesive used. AF555 bonded
well to the nylon prepared surfaces; where as MB1515-3 did not. This is reflected in both the
GIC values and the mode of fracture shown in table 6. A dramatic decrease in fracture energy
was observed when nylon-prepared surfaces were bonded with MB1515-3 and the fracture mode
changed from cohesive (AF555) to adhesion (MB1515-3).

Due to the difficulty in removing three of the five nylon peel plies from the laminates, there is
insufficient data to draw conclusions on the effect of peel-ply surface texture using nylon peel
plies. In the authors’ opinion, the effect of surface texture would be minor, if any, when the
fracture is cohesive (within the adhesive layer). Surface texture may be more important when
fracture occurs along the adhesive/laminate interface or between the lamina.

19
 TABLE 6. SUMMARY OF AVERAGE (FROM 10 SAMPLES EACH) GIC DATA AND
FAILURE MODE FOR LAMINATES PREPARED WITH VARIOUS PEEL PLIES AND
BONDED WITH AF555 AND MB1515-3

Specimen ID GIC Standard

No. (J/m2) Deviation Failure Mode
40000 AF 745 75 Cohesive/Interlaminar
40000 MB 209 58 Adhesion
52008 AF 911 80 Cohesive/Interlaminar
52008 MB 122 15 Adhesion
60001 AF 910 35 Cohesive
60001 MB 812 35 Cohesive
60004 AF 956 202 Cohesive
60004 MB 856 37 Cohesive
60004 VLP AF 863 49 Cohesive
60004 VLP MB 914 140 Cohesive
60005 AF 904 38 Cohesive
60005 MB 657 277 Cohesive
1200.0

1000.0 955.5
910.7 914.4 904.4 909.6
812.3
800.0 772.4
745.0 743.8
656.7
GIC (J/M2)

AF555
600.0
MB1515-3

400.0 346.9

200.0
122.1

0.0
52008 40000 60004 60004 60005 60001 60001 60001
VLP Wet As Rec Dry
Peel Ply

FIGURE 17. AVERAGE GIC OF LAMINATES PREPARED WITH VARIOUS PEEL PLIES
AND BONDED WITH AF555 OR MB1515-3
(Peel ply weaves coarsen from right to left.)

20
3.3 EFFECT OF PEEL-PLY MOISTURE.

The effect of moisture in the 60001 polyester peel ply on GIC is presented in figure 18. All
samples failed in a cohesive manner with respectable GIC values. The slightly higher value of
GIC for laminates prepared with as-received 60001 peel ply has not been explained; this may be
an experimental anomaly. Based on the cohesive fracture mode of all the samples and
bracketing of the as-received GIC data by the wet and dry data, it appeared that peel-ply moisture
had little if any effect on the bonding of the cured laminates with AF555. Other adhesives may
be more sensitive to moisture than AF555, and other peel plies/laminate systems may be affected
by precure moisture in the peel ply.
1000.0
909.6
900.0

800.0 772.4
743.8

700.0

600.0
GIC (J/m 2)

500.0

400.0

300.0

200.0

100.0

0.0
60001 Wet 60001 As Rec 60001 Dry
Peel Ply

FIGURE 18. AVERAGE GIC OF LAMINATES PREPARED WITH VARIOUS PEEL PLIES
AND BONDED WITH AF555 OR MB1515-3

3.4 DETERMINATION OF SURFACE ENERGY BY CONTACT ANGLE.

The effect of peel-ply material and texture on surface energy of the laminate surface was
determined by examining the surface energies and wetting envelopes of the prepared surfaces.
The known polar and dispersive components of the surface tension (γlp and γld) of ethylene
glycol, deionized water, and DMSO [3] and the measured contact angles (θ) on each surface
were used to construct Kaelble plots. The square of the slope of a linear fit of these plots is the
polar component, γsp, and the square of the intercept is the polar component, γsd, of the surface
energy of the laminate. The contact angle data obtained with glycerol was not used in
constructing the Kaelble plots due to experimental difficulties associated with bubbles forming
during the measurements. An example of a Kaelble plot is shown in figure 19 for BMS8-276
laminate prepared with PFG60001 peel ply. The results of contact angle measurements and
calculated surface energies are summarized in table 7. The total surface energy γtot, is very
similar between the surfaces, however there are significant differences in the dispersive and polar

21
components of the surface energy between the polyester and nylon peel-ply-prepared surfaces.
The polyester peel-ply-prepared surfaces have a greater dispersive component, while the nylons
have a larger polar component of surface energy.

16.0
γ l(1+cost)/(2*sqrt(γ l ))

y = 5.4864x + 3.7178
14.0
p

DMSO
12.0
10.0
8.0 Ethylene
6.0
Glycol

4.0 D.I. water

2.0
0.0 γsd = 30.3 mJ/m2, γsp=13.7 mJ/m2, γtot=44.0 mJ/m2
0.0 0.5 1.0 1.5 2.0 2.5
d p
sqrt(γ l /γ l )

FIGURE 19. LINEAR KAELBLE PLOT FOR SURFACE ENERGY MEASUREMENT

FROM CONTACT ANGLE DATA ON BMS8-276 LAMINATE PREPARED WITH
PFG 60001 PEEL PLY

TABLE 7. CONTACT ANGLE MEASUREMENTS AND SURFACE ENERGIES FOR A

VARIETY OF FLUIDS MEASURED ON BMS8-276 LAMINATES PREPARED WITH
DIFFERENT PEEL PLIES

DI H2O DMSO Ethylene Glycol Gylcerol γsd γsp γs total

Peel Ply (degrees) (degrees) (degrees) (degrees) (mJ/m2) (mJ/m2) (mJ/m2)
40000 51.3 12.9 27.6 61.5 23.0 24.0 47.0
51879 56.3 8.0 20.8 59.6 25.0 20.3 45.3
60001 63.4 6.8 27.0 56.1 30.3 13.7 44.0
60004 76.5 20.1 36.0 80.4 36.0 5.8 41.8
60004 VLP 71.5 19.0 45.8 76.7 33.6 7.3 40.9
60005 73.7 5.7 30.7 77.5 37.2 6.8 44.0

Wettability envelopes were calculated from the dispersive and polar components of the surface
energy data in table 7. Fluids with surface energies inside the envelope will wet spontaneously
and those outside the envelope will not. It is important for adhesives to wet out the substrate.
Therefore, the wettability envelopes are a potential method to determine if a surface is suitable
for bonding. The wettability envelopes for BMS8-276 laminates prepared with four polyester
peel plies from PFG are shown in figure 20. There is some slight variation in the wettability
envelopes, especially between the laminates prepared 60001 and the other three peel plies. The
60001-prepared surface has a greater polar component than the 60004, 60004VLP or 60005. The
wettability envelopes for the laminate prepared with nylon peel plies (40000 and 51879) are

22
shown in figure 21. The laminate surface prepared with 40000 had a slightly greater polar
component than the 51879-prepared surface, and the dispersive components were very similar.
There were significant differences in the wettability envelopes between nylon- and polyester-
prepared surfaces, as is evident in figure 22 where the envelopes for nylon 51879 and polyester
60001 are shown. For reference, typical values of surface energies for epoxy adhesives are
shown by the shaded circle. The precise surface energies of the adhesives used in this study
were not available from the literature or the manufacturers. Epoxy adhesives clearly fall within
the envelope for polyester-prepared surfaces, but are on the boundary for the nylon surfaces.
This may offer a possible explanation of the poor bonding with MB1515-3 on nylon peel-ply-
prepared surfaces. The surface energy of MB1515-3 may be slightly outside that wettability
envelope. It is important to note that contact angles by themselves were not sufficient to predict
poor bonding with nylon-prepared surfaces. Nylon-prepared surfaces had lower contact angles
with some fluids, including water, than polyester peel-ply prepared surfaces.

45
40
35
30 VLP
25 60005
(mJ/m 2)

20 60004
15 60001
Dγ 1 10
5
0
0 10 20 30
γlp (mJ/m 2)

FIGURE 20. WETTABILITY ENVELOPES FOR BMS8-276 LAMINATES PREPARED

WITH PFG POLYESTER PEEL PLIES
(Left to right at the nose: 60004, 60004VLP, 60005, and 60001.)

23
 40
35
30
γ l (mJ/m )
2

25
52006
20
40000
15
d

10
5
0
0 10 20 30 40
p 2
γ l (mJ/m )

FIGURE 21. WETTABILITY ENVELOPES FOR BMS8-276 LAMINATES PREPARED THE

PFG NYLON PEEL PLIES

FIGURE 22. WETTABILITY ENVELOPES FOR BMS8-276 LAMINATES PREPARED

WITH PFG NYLON PEEL PLY 51879 AND POLYESTER PEEL-PLY 60001
(Left to right at the nose: 60001 and 51879.)

24
4. CONCLUSIONS.

The results of mechanical tests demonstrated a significant difference in bond quality when joints
were prepared with varying peel ply materials and adhesives.

All four weaves of uncoated polyester peel ply were easily removed from laminate surfaces after
curing and created surfaces that bonded well using either adhesive. The surfaces that were
generated were in all cases interfacial between the peel-ply fabric and the epoxy resin matrix. X-
ray photoelectron spectroscopy (XPS) surface characterization of the surface generated with
polyester peel did not detect any anomalous materials on the surface as were found on nylon.
The bond failure was cohesive with average GIC values of 908 and 809 J/m2 for adhesives AF555
and MB1515-3, respectively.

The 60001 Super Release Blue (siloxane-coated polyester) release films created very poor
bonding surfaces with both adhesives, the result of the transfer of the peel-ply siloxane coating to
the composite surface.

Laminate surfaces prepared with nylon peel plies bonded well with AF555 (all GIC> 750J/m2;
cohesive failure). Laminate surfaces prepared with nylon peel plies bonded poorly with
MB1515-3 (average GIC< 165 J/m2; adhesion failure). An explanation for this difference was
found using scanning electron microscopy and XPS. The filaments of the removed nylon peel
ply were shredded and small fiber fragments were found on the corresponding composite surface
after peel-ply removal. MB1515-3 must have a chemical composition that does not bond well to
epoxy surfaces contaminated with amides. AF555 is tolerant of some level of amide
contamination. Nylon peel plies were more difficult to remove from the laminate and, in some
cases, could not be removed without damaging the laminates. Finer weaves were easier to
remove than coarser weaves.

Single contact angle measurements and surface energies were not sufficient to predict bond
quality. Nylon-prepared peel-ply surfaces had lower contact angles and hence higher surface
energies, but produced unsatisfactory results when bonded with MB1515-3. Using wettability
envelopes, epoxy adhesives were found to be on the border of acceptable wetting for these
surfaces.

Peel-ply texture or peel-ply moisture content had no significant effect on fracture energy or mode
of failure for the adhesives and laminate system investigated.

5. REFERENCES.

1. Hart-Smith, L., et al., “Some Observations on the Analysis of Inplane Matrix Failures in
Fibrous Composite Laminates,” in 10th DoD/NASA/FAA Conference on Fibrous
Composites in Structural Design, 1993.

2. Bardis, J. and Kedward, K., “Effects of Surface Preparation on the Long-Term Durability
of Adhesively Bonded Composite Joints,” DOT/FAA/AR-03/53, September 2004.

3. Kinloch, A.J., Adhesion and Adhesives, Chapman and Hall, 1987.

25
4. Davis, M.J. and Bond, D., “Principles and Practices of Adhesive Bonded Structural Joints
and Repairs,” International Journal of Adhesion and Adhesives, 19 (3): 1999, p. 91.

5. Bossi, R., et al., “Composite Surface Preparation QA for Bonding,” International SAMPE
Symposium and Exhibition, 50 (New Horizons for Materials and Processing
Technologies), 2005. p. 2156.

6. Hart-Smith, L.J., “A Peel-Type Durability Test Coupon to Assess Interfaces in Bonded,

Co-Bonded, and Co-Cured Composite Structures,” International Journal of Adhesion and
Adhesives, 19 (2-3), 1999 p. 181.

7. Armstrong, K/B., “Effect of Absorbed Water in CFRP Composites on Adhesive

Bonding,” International Journal of Adhesion and Adhesives, 16 (1), 1996, p. 21.

8. Hart-Smith, L.J., et al., “Surface Preparation of Fibrous Composites for Adhesive

Bonding or Painting”, Douglas Service Magazine, First Quarter (12), 1984, p. 12.

9. Goldstein, J.I., et al., Scanning Electron Microscopy and X-Ray Microanalysis: A Text for
Biologists, Materials Scientists, and Geologists, 2 ed., New York: Plenum Press, 1992.

10. Kitazaki, Y. and Hata, T., “Surface-Chemical Criteria for Optimum Adhesion.3.
Variability of Critical Surface Tension(Gamma-C) and Its Choice,” Journal Of Adhesion,
4 (2), 1972, p. 123.

26