Citation
Composite repair utilizing bond exchange reaction polymers

Material Information

Title:
Composite repair utilizing bond exchange reaction polymers
Creator:
Egan, Kyle W.
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English

Thesis/Dissertation Information

Degree:
Master's ( Master of science)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Mechanical Engineering, CU Denver
Degree Disciplines:
Mechanical engineering
Committee Chair:
Rorrer, Ron
Committee Members:
Yu, Kai
Carpenter, Dana

Notes

Abstract:
The use of a bond exchange reaction epoxy in carbon fiber reinforced composites is investigated for use in the repair of composites. Three ply composite specimens were created with an inter-lamina defect using PTFE strips during the curing process. The specimen was repaired by applying pressure and a temperature of 180° C. The flexural stiffness and strength was tested using a three point bend test. The repaired specimens are shown to retain stiffness and strength up to 99% and 95% respectively that of a specimen without a defect. The experiments examined the relationship between time and pressure on the effectiveness of the repair process. A parametric study was completed to determine the optimal pressure and time duration of repair for a delaminated bond exchange reaction epoxy composite. The study revealed an increase in strength and stiffness of the composite as time and pressure were increased. The study shows an increase in strength compared to other methods used for composite repair, while not increasing weight or dimensions of the composite. The method discussed in this paper is shown to recover 95% strength, proving the efficacy of the method as well as use of bond exchange reaction epoxy within composites for composite repair.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
Copyright Kyle W. Egan. Permission granted to University of Colorado Denver to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item has the following downloads:


Full Text
COMPOSITE REPAIR UTILIZING BOND EXCHANGE REACTION POLYMERS
by
KYLE W. EGAN
B.S., Northern Arizona University, 2015
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Mechanical Engineering
2018


This thesis for the Master of Science degree by
Kyle W. Egan has been approved for the Mechanical Engineering Program by
Ron Rorrer, Chair Kai Yu
Dana Carpenter
Date: May 12, 2018
11


Egan, Kyle W. (M.S., Mechanical Engineering)
Composite Repair Utilizing Bond Exchange Reaction Polymer Thesis directed by Associate Professor Ron Rorrer
ABSTRACT
The use of a bond exchange reaction epoxy in carbon fiber reinforced composites is investigated for use in the repair of composites. Three ply composite specimens were created with an inter-lamina defect using PTFE strips during the curing process. The specimen was repaired by applying pressure and a temperature of 180° C. The flexural stiffness and strength was tested using a three point bend test. The repaired specimens are shown to retain stiffness and strength up to 99% and 95% respectively that of a specimen without a defect. The experiments examined the relationship between time and pressure on the effectiveness of the repair process. A parametric study was completed to determine the optimal pressure and time duration of repair for a delaminated bond exchange reaction epoxy composite. The study revealed an increase in strength and stiffness of the composite as time and pressure were increased. The study shows an increase in strength compared to other methods used for composite repair, while not increasing weight or dimensions of the composite. The method discussed in this paper is shown to recover 95% strength, proving the efficacy of the method as well as use of bond exchange reaction epoxy within composites for composite repair.
The form and content of this abstract are approved. I recommend its publication.
Approved: Ron Rorrer
in


ACKNOWLEDGEMENTS
I would like to thank the employees at National Institute of Standards and Technology who were involved in conducting testing for this research project as well as providing scanning electron microscopy images: [Applied Chemicals and Materials Division, Andrew Slifka, and May Martin], Without their professional help and input, this research could not have been successfully completed.
IV


TABLE OF CONTENTS
CHAPTER
I. The Growing Need for Composite Repair..................................................1
Literature Review......................................................................2
Epoxy Refill Repair...................................................................2
Matrix Removal........................................................................2
Scarf Joints..........................................................................4
Overlay Patch Repair..................................................................5
Bond Exchange Reaction Epoxies........................................................6
II. Preliminary Studies...................................................................9
Defect Repair.........................................................................9
Single Lamina Manufacturing and Parametric Study......................................9
Curing Effect Revealed...............................................................12
Solution Injection Repair Method.....................................................13
III. Parametric Study Simulating Delaminated Composite Repair............................17
Methods..............................................................................17
Materials...........................................................................17
Epoxy Formulation...................................................................18
Pre-preg Manufacturing..............................................................18
Composite Specimen Manufacturing....................................................19
Vacuum Bagging Process..............................................................21
v


Dry Repair Process.......................................................................22
Testing Procedure........................................................................23
Failure Mode Analysis....................................................................24
Data analysis............................................................................25
Results.....................................................................................26
Applied Pressure of Repair...............................................................26
Time Duration of Repair..................................................................28
Failure Modes of Pressure and Time Tests.................................................30
Optimal Pressure and Time Duration of Repair.............................................33
Discussion..................................................................................34
Conclusion..................................................................................35
RI I I RI N( I S..............................................................................37
vi


CHAPTER I
THE GROWING NEED FOR COMPOSITE REPAIR
On January 16, 2013 the Federal Aviation Administration grounded the Boeing 787, costing the company an estimated $600 million dollars.[1] The plane was grounded for three months as a solution to a battery fire problem was found. This event illustrates just how important a quick turnaround is when it comes to repairing aircraft components. The Boeing 787 has more composites per weight than any other material in the aircraft. [2] Damage to composite components can render them nonfunctional. This is especially true for any part that is structural to the airframe or aerodynamics. Composites are widely being utilized in industries ranging from fashion and sporting goods to automobiles and large commercial aircraft. Composites are seeing a boom in the automotive industry where they are no longer used only in exotic supercars. As composite materials become more affordable and necessary to decrease vehicle weight, the automotive industry will being to incorporate them into their vehicles structures. General Motors Company is now utilizing carbon fiber to create its truck beds rather than steel. The use of carbon fiber in the 2019 GMC Sierra has decreased the weight of the truck bed by 62 pounds over a standard steel truck bed, leading to a more fuel efficient vehicle.[3] Toyota and Honda already utilize composites in their mid-size trucks which accounted for over 200,000 sales in 2017 alone.[4, 5] These sales figures are impressive on their own but are just a fraction of the total truck market in the United States. A total of 2.37 million full-size trucks were sold in 2017, most of those being the Ford F-150.[5] When a large American truck manufacturer such as Ford or Chevrolet finally adopts composite material for their truck beds, a low cost a quick composite repair system will be required. A cost effective repair process that returns a damaged part to a functional percentage of its original strength is necessary for the future of composites.
1


LITERATURE REVIEW
Epoxy Refill Repair
Current methods of composite repair utilizes varying forms of epoxy injection or patch repair. The technique of epoxy refill repair makes use of low viscosity adhesives or epoxies applied directly to the damaged area of the composite. The epoxy can be injected to the damaged area as in Bauer [6] or applied like the original epoxy process. This injection process allows the technician to methodically apply the exact amount of epoxy as well as being able to apply it to the exact place of damage on the composite. Is the case of Bauer [6] the epoxy was injected between lamina. The other end of the spectrum of refill repair spectrum is used by Kuchler [7] by applying epoxy to the damaged area using the Seeman composite resin infusion molding process (SCRIMP) method. The carbon fiber is placed in a vacuum bag that pulls resin into the bag through vacuum hoses which then infuses the carbon fiber to cure into a laminate. This process can be applied locally to selected area by securing a vacuum tight seal around the damaged area before resin is infused. Kuchler [7] shows that this process is a viable technique to infuse a damaged area of a composite with new resin. However this research is limited in its scope because it does not test the effectiveness of the infused epoxy to recover the original strength of the composite. Kuchler [7] falls short of a researching a complete process of matrix removal, resin injection, and finally testing of composite. Without the quantification of composite strength testing after resin infusion, it is unknown whether the SKRIMP process can be viable as a method to repair a damaged composites. Kuchler [7] choses to instead focus on the process of matrix removal.
Matrix Removal
The process of matrix removal is necessary in the repair of a damaged area of a composite. If broken fragments of the matrix are often left inside the damaged area of the composite the strength recovery in that area will be compromised. Fracture or damage to the resin within a composite can
2


lead to failure of the composite. The carbon fiber and epoxy matrix that make up a composite are difficult to separate due to the resin thermoset properties. A thermoset polymer does not have a melting temperature due to the crosslinking polymer chains that makeup the epoxy. When put under high heat the epoxy will only degrade but will say intact, therefore thermal separation of the materials is unattainable. Due to the difficulty of this process other methods such as acid washes and supercritical water baths have been the only known method to separate the matrix from the fiber without damaging the fiber. [8] The use of harsh chemicals as a solvent leads to many ecological and economic problems. In response to this Kuchler [7] has developed a process that uses infrared radiation instead of chemicals to remove the matrix from the fiber. The process begins by applying a thin layer of C^Ch powder on the damaged area of the composite. A glass nonwoven fiber is placed on top of this to shield the undamaged area from radiation. Finally an IR lamp is placed over this and run at 380 W for 20 minutes. The matrix degrades and through chemical reaction with the CT2O3 powder is removed from the fibers. In this process the fibers strength is reduced by 10%. The process is shown to be a viable way to separate the matrix from the fiber. The removal can be localized relatively easily because the removal only occurs where the power is located. The precision available with this technique makes it easier to only remove the damaged matrix. In other methods such as those that use acids to dissolve the matrix the entire composite sample must be submerged in the acid the dissolve the matrix. Therefore the process from Kuchler\ 1 | is valuable in that it can be localized to the damaged area. It is still limited however by its extensive use of electrical energy to irradiate the composite as well as its waste product of CT2O3 powder. The process can also only penetrate the composite to dissolve the matrix to a depth of approximately two millimeters. There is still a need to develop a viable method of matrix removal that does not create waste or consume vast amounts of energy.
3


Scarf Joints
The composite repair method used in aviation utilizes a machining process to cut out the damaged area exposing a clean surface.[9] The void is then fdled with composite patches and recured. These patches are either pre-cured composite-epoxy lamina or clean carbon fiber weave patches. The pre-cured patch is sealed and resin transfer method (RTM) used to fill the area with epoxy to create a laminate. The patch can also be a clean carbon fiber weave, which is cured into a laminate using RTM. RTM is similar to SCRIMP method discussed earlier. RTM is used to ensure a desired volume fraction of epoxy to carbon fiber. While the scarf method is effective at removing damaged fibers and matrix material there are a few negatives associated with this process. The new patches placed into the scarf are not carrying the same load as the original fibers before the damage occurred, this leads to a weaker part. The patch cannot distribute the load as effectively as the original because the new and original fibers are separated by a layer of epoxy. The epoxy cannot withstand the forces that the fiber can and so a weak point in the part has just been manufactured. The shape of the scarf and patch is another variable to consider. A scarf can either be stepped as seen in Figure 1 or straight.
Straight taper for v-
scarf repair Scarf angle
Stepped taper for stepped scarf repair
Figure 1 Straight taper and Stepped taper scarf joints used in laminate composite repair[9]
A stepped tapper scarf creates local stress concentrations of two or three times that of straight tapered scarf joints. [10] Studies have shown stress concentrations are strongly correlated to ply thickness, adhesive thickness, taper angle, stacking sequence, overply layup, and overply lap
4


length. [10] The tapered scarf joint is used more widely than the stepped scarf due to its performance. Patch shape is another variable to consider in the scarf repair process. Multiple studies that have examined the effect of patch shape of the repair process differ on the efficiency of different shapes depending on the loading and other factors associated with patch repairs. The recovery strength of a laminate with patch repair is dependent on multiple factors that makes it difficult to determine an optimum patch for all conditions. These factors are fiber orientation, number of layers, stacking sequence, loading condition of the part, material used, and over-ply among others. With these factors considered studies have shown that for a simple over-ply patch with no scarf joint and simple tension loading, a rectangular over-ply patch is best for epoxy strength.[11] This same study however concedes that a rectangular patch has increased stress intensity at the crack compared to an oval or octagon shape. [11] Another study supports the use of a octagon shaped patch as an optimum shape. [12] This study also indicated that the stacking sequence contributed to the effectiveness of the patch. A rectangular shape was optimal for sequence 0/±45/90 and 0/90/±45 while an oval shape was optimal for sequence ±45/0/90 and 90/0/±45.[12]
Scarf repairs need to be implemented accurately and this makes this a time consuming and technical task. The technician completing the repair must be experienced not only in composites but also in machining. For these reasons the scarf joint is mostly used for structural or thick composite laminates.
Overlay Patch Repair
Overlay patch repair utilizes multiple other techniques to create a strong repair. In Shams [13] a scarf cut is made in the damaged area, then is injected with epoxy, then a patch of carbon fiber is laid over the scarf area. This overlay patch can be one to three layers and is meant to resist the bending associated with the weakened damaged area. These patches increase the strength of the composite as well as distributing the stress around the area of damage. Without this external patch on the outside of the damaged area, the epoxy/resin inside the void does not add enough strength to deem
5


the composite repaired. The overlay patch of carbon fiber fixes this issue, resulting in a considerable amount of strength recovered, but it also adds thickness and some weight. The added thickness can contribute to a change in aerodynamics which is not viable for aircraft. This gap in the composite where the damaged composite has been cut and replaced with epoxy is a large weakness in the composite. The overlay patch is a way of strengthening that void of fiber. While the overlay patch can be an effective way of increasing strength recovery and distributing the load around the damaged area there are still unknown such as patch shape as previously mentioned.[11, 12]
Bond Exchange Reaction Epoxies
All of the aforementioned repair methods have one common flaw. These repair methods rely on new epoxy bonding to the original epoxy of the part. The problem with this process of epoxy to epoxy bonding are that this does not create covalent bonds in the epoxy matrix but relies on secondary bonding for stress transfer. When an epoxy cures it is covalently bonded together. The reasoning for this are the thermosets crosslinking. The available crosslinking reaction sites have all been used during the original curing process. The polymer structure of the epoxy is relied on to fill and cling to micro defects on the original epoxy in conjunction with secondary bonding to increase adhesion. These secondary bonds are much weaker than the covalent bonds of the original epoxy. The weak bonds are part of the reason that repairing composite structures to their original strength is difficult to accomplish. Thermosets lack crosslinking between the polymer chains which allows them to move past each other at elevated temperature; allowing flow of the material. Thermoplastics on the other hand do have the ability to be reformed and change shape when heat is applied. Thermoplastics are also much weaker than a thermoset because of the lack of crosslinking. Strides to implement thermoplastics into composites have been taken and are most successful in 3D printing. Thermoplastics can be easily 3D printed due to its ability to flow at elevated temperatures. Studies have found that while adding carbon fiber particles to the thermoplastic when printing increases the modulus, there is only a slight increase.[14] For these reasons a thermoset is ideal for composite
6


laminates. What is needed to improve on current repair methods is a matrix material that crosslinks like a thermoset but can also reconnect the crosslinking polymers when damaged.
A new class of polymer called bond exchange reaction (BER) or covalent adaptive network (CAN) polymers are currently being researched. This type of polymer blurs the line between thermoset and thermoplastic. BER polymers/epoxies utilize a covalent adaptive network that allows the epoxy to re-crosslink severed chains.[15] With applied heat, the epoxy can reconnect these covalent bonds that have been previously broken(see Figure 2). The ability of BER epoxies to repair a damaged surface have been previously demonstrated.[15, 16]
In these studies both recyclability of the polymer and pressure free solvent assisted repair are demonstrated. The first study demonstrated the ability to use BER epoxy in a composite lamina repair. The surface of a single BER epoxy composite was damaged by scratching it and then repair using powdered BER epoxy and solvent. The repaired lamina was able to recover 97% of the fiber modulus compared to a lamina which was not damaged. [15] The study demonstrated the positive uses for BER epoxy as a surface repair epoxy as well as a recyclable matrix by utilizing solvent dissolution. This experiment showed that utilizing BER epoxy could create a near 100% closed loop process for manufacturing, recycling, and repurposing materials in composites. The later solvent assisted pressure free method was shown to repair a specimen from two fractures pieces into on piece using ethylene glycol. The study showed a recovery of 74% of the original modulus, demonstrating the BER epoxy’s particularly unique ability of solvent dissolution. [16] Compared to other thermosets BERs are unique in that they can be dissolved by a solvent and then solidified as the solvent is evaporated. This ability makes BER epoxies applicable to more than one method or repair and an ideal candidate epoxy for composite repair.
7


Figure 2 Diagram showing the covalent adaptive network (CAN) of the BER epoxy [16]
8
y y


CHAPTER II
PRELIMINARY STUDIES
Preliminary research was completed to develop a greater understanding of how the BER epoxy would perform as a composite epoxy. To study this, multiple studies were completed and the results documented to assist in the final research study.
Defect Repair
An initial study of a specimen that simulated a repair of a delaminated specimen was completed as a proof of concept. A laminate made from three layers of carbon fiber reinforced BER epoxy was created. A strip of PTFE was placed between layers 1-2 and 2-3, through the width of the laminate. The PTFE strips were laid in line with each other to create a void through the width of the sample. The laminate was cured using a vacuum bagging process discussed in Chapter HE The PTFE strips, of length 45 mm, were then removed from the specimen to create a void between the lamina, simulating a delaminated part. The specimens were repair using the Dry test method detailed in Chapter HE A pressure of 500 kPa was applied for two hours, at a temperature of 180° C. The specimens were then tested utilizing a 3-point bend method. The results of this study was recovery of 86% stiffness. This result gave proof that with some fine tuning of the repair process a significant recovery of strength and stiffness of a part could be achieved.
Single Lamina Manufacturing and Parametric Study
The BER epoxy also lends itself to new manufacturing techniques that have not been previously possible. The covalent adaptive network of the BER epoxy can be exploited to not only repair damaged areas but also to manufacture laminates for single cured lamina. Single lamina carbon fiber/BER epoxy sheets are cured and stored. These sheets of cured lamina can then be shipped to any destination without the use of refrigeration or fear of degradation of the epoxy. The cured sheets can then be cut to a desired shape and used in conjunction with other sheets to create a laminate. The technician would simply need to apply a distributed pressure while heating the laminate to 180° C for
9


a few hours. The manufacturing process would need no additional epoxy. This process could also permit the technician to apply patches to surfaces in an efficient and timely manner.
To test this idea, a single flat lamina of carbon fiber/BER epoxy was created using a vacuum bagging process. This single lamina was cut into specimen dimensions of 55 mm by 15 mm. The single lamina specimens were placed on top of one another to create a 3 ply laminate specimen of dimension 55 mm by 15 mm. These laminate specimens were then repaired using applied pressure and heat in the form of a heated platen press. Two parameters were tested for this study, pressure and duration of repair. To examine the effects of pressure on a repaired specimen, specimens were repaired at pressures of 500, 625, 750, 875, and 1000 kPa for a duration of one hour. To examine the effects of duration of the repair, specimens were repaired at 500 kPa for a duration of 1, 2, 3, and 4 hours. All tests were performed at a temperature of 180° C.
The results of this parametric study proved that there was indeed a relationship between the time of the repair, pressure applied during the repair and the recovered strength and stiffness of the laminate. The stiffness and strength of the repaired specimens increased as pressure and time was increased. The relationship however was not linear and seemed to taper off at the high pressures and longer times. For the time study, there was only a slight difference between specimens repaired for 3 hours and specimens repaired for 4 hours(see Figure 3). The pressure application study revealed a similar relationship, with specimens repaired at 875 kPa and 1000 kPa exhibiting almost identical strength and stiffness recovery(see Figure 4). The recommendation for repair of specimens using this method was then determined to be an applied pressure of 875 kPa at a temperature of 180° C for a time of 3 hours. This recommendation was determined to reduce the amount of resources while still maximizing the strength and stiffness recovery.
10


Figure 3 Stress(strain) chart of specimens repaired for 1, 2, 3, and 4 hours at 500 kPa
Figure 4 Stress(strain) chart of specimens repaired for 500, 625, 750, 875, and 1000 kPa for 1
hour
11


However this study did not meet the goal of fully repairing single sheets into a solid laminate. The main reason that this study did not achieve this goal was due to the complexities with manufacturing a single sheets of lamina. The sheets were rough and in some places the carbon fiber was void of any epoxy. These voids would drastically reduce the recoverable strength and stiffness of laminates. This study was abandoned to work on an effective manufacturing process. The manufacturing process that was developed as a result, gives a desired surface texture, even distribution of epoxy, and repeatable results. The process is detailed in Chapter III Methods section.
Curing Effect Revealed
The previous study revealed a curing effect of the epoxy that made determining the exact parameters for a repair method difficult to determine. The problem arose from curing the laminates at 130° C and performing repairs at 180° C. This variation in temperature resulted in a specimen that was not fully cured. Because the laminates were not fully cured, the laminates experienced additional curing when repaired, which created an additional variable to track during testing. To remove this unknown variable from further testing, specimens were cured at 180° C for 6, 8, and 10 hours. The specimens’ stiffness and strength were then tested to determine if a significant change could be determined between 6, 8, and 10 hours. This would also show that a specimen cured at 180° C for 6 hours and then further repaired at 180° C for 2-4 hours would not exhibit a dramatic change due only to temperature exposure.
The results of this study revealed that specimens cured at 180° C did not exhibit a curing effect like that of specimens cured at 130° C. Specimens cured for 6 and 8 hours had similar stiffness and strength with slight drop in stiffness and strength at 10 hours. This slight drop is most likely due to some epoxy degradation. This study revealed a more reliable method for manufacturing specimens for repeatable results. All future specimens were then cured at 180° C for 6 hours and repaired at the same temperature.
12


Solution Injection Repair Method
Another benefit of the BER epoxy is that it can be dissolved using ethylene glycol as a solvent. The solvent breaks down the polymer chain. This reaction only occurs at a temperature of 180° C. Ethylene glycol has a boiling point of approximately 198° C so the ethylene glycol evaporate as the BER epoxy re-crosslinks, eventually leaving a repaired BER epoxy sample. Previous works have examined the use of solvent assisted repair of the BER epoxy.[16] The same idea of solvent assisted repair for composites using BER epoxy was the next logical step. Solvent would be injected into a defect area on a specimen and then the specimen would then be placed in an oven at 180° C to evaporate the solvent and promote the re-crosslinking of the epoxy.
Laminate specimens with a defect through the width of the specimen, as described in Chapter III, were created. Solvent was injected into the void of the specimen then placed in a vacuum bag for repair. The vacuum bag was chosen instead of simply placing the specimens in an oven because pressure was desired to fully repair the voids. It was thought that if pressure was not applied to the specimen during repair that the solvent would simply evaporate and not repair the void. Specimens was placed in the oven for 30 minutes before a vacuum was pulled on the bag to make sure the solvent was not squeezed out of the void. For those 30 minutes the solvent could break down the polymer chains on the surface of the void creating a tacky epoxy surface on either side of the void. Once the vacuum was pulled and pressure applied, the two tacky surfaces would be pressed together and allowed to re-crosslink. The specimens were then left in the oven for another hour and a half to allow the solvent to evaporate and the specimen to fully repair. The results of the repair yielded a specimen that had deformed during the process(see Figure 5). The center of the specimen where the solvent was injected into the void was wavy and not uniform compared to its original shape. The nylon fabric used in the vacuum bagging process also became attached to the specimens. The pure solvent method was then abandoned because of these results.
13


Figure 5 Specimens after solvent injection
A different injection method was developed to take the place of solvent injection. The idea of a solvent/epoxy solution was developed. The solution would carry both epoxy and solvent into to defect area. A solution mixture was thought to incorporate the best of the epoxy injection and solvent injection methodologies without the downsides of either. The epoxy injection method's downside being that the epoxy resin would need to be kept in a refrigerated state to ensure the epoxy does not degrade. The epoxy resin would therefore take a considerable amount of equipment to utilize in the field. The downside of solvent injection was discussed previously. The solution injection would only require a heating source which is already necessary for repair. The process for the solution injection method is as follows. Epoxy resin is cured in sheets approximately 5 mm thick. The cured epoxy sheets are cut into small particles then placed in a beaker. Solvent is then poured into the beaker in the appropriate mass ratio of cured epoxy-solvent. The beaker is sealed and heated to 180° C to promote the dissolution of cured epoxy. The solution is heated until a consistent solution mixture is obtained (4-6 hours). The transformation from cured epoxy to epoxy-solvent solution can be seen in Figure 6.
14


Figure 6 Pictures of manufacturing epoxy-solvent solution. A) Cured epoxy particles B) Cured epoxy particles with ethylene glycol solvent added before heating C) Epoxy-solvent solution after heating to 180° C
A small study was conducted to determine the best ratio of cured epoxy-to-solvent by mass. Ratios of cured epoxy-to-solvent of 1:3, 1:2, 1:1,2:1,3:1, and 6:1 were created. It was found that at higher concentrations of solvent such as 1:3, 1:2, and 1:1, the epoxy simply separated after being heated (see Figure 7). Ratios with higher concentrations of cured epoxy stayed in solution. The ratio of 6:1 cured epoxy-to-solvent by mass was chosen for a few reasons. The first reason was that it created the most consistent solution in terms of regular dispersion of epoxy into solvent solution. The second reason was that the solution would re-cure quicker because it had a lower concentration of solvent that needed to be evaporated from the laminate.
15


Figure 7 Solvent and epoxy solution at various ratios showing separation of solvent and epoxy
at higher concentrations of solvent
With a ratio for the solution chosen, a study to determine the efficacy of this method was completed. The same three ply laminate with inter-lamina defect created with the use of PTFE strips used in previous experiments were used to test the solution repair method. The solution mixture was warmed to approximately 100° C to soften the solution to be able to work with it better. The softened solution was then applied to the void area of the defect specimens with a small flat metal tool. The specimens were then placed in a vacuum bag for range of times. The correct time for the repair process needed to be determined to find the maximum recovery strength of the laminate. The times chosen were 2, 4, 6, 8, and 10 hours. The specimens were then tested under three-point bend to determine the strength and stiffness of the specimen. The results of the tests show an average increase in strength and stiffness as the time to re-cure is increased. At a maximum time of 10 hours to repair the defect, the specimen on average recovered 105% stiffness and 94% strength. This result is to be expected because the process physically adds more epoxy that covalently bonds to the defect areas, pulling them back together and forming a cohesive material. However this experiments did have some anomalies that could not be explained such as a 6 hour specimen on average recovering more stiffness
16


than an 8 hour specimen. The 10 hour specimens were also slightly thinner than the others as well as the control specimen which means that these results cannot be compared without further research. Overall the method proved to be an effective method for repair and warrants further research.
One downside of epoxy injection is the added weight associated with the additional epoxy. The solution injection method explored is also at risk of the same downside but the results of the tests show an increase in weight of each specimen of around 6.5%. This small increase in weight is a good result considering that half of the specimen was being injected with the solution.
CHAPTER III
PARAMETRIC STUDY SIMULATING DELAMINATED COMPOSITE REPAIR
A parametric study of repair conditions for a BER composite, with inter-lamina defect, was completed. The effects of pressure and time during the repair process were investigated. This study revealed the optimum pressure and time to repair a delaminated BER composite. These optimum pressure and time of the repair was combined into one method of repair. This method is shown to be the optimal repair method for dry repair. This optimally repaired specimen is then compared to a control set of specimens. The dry method is shown to be as effective as current repair methods utilizing current engineering epoxies. The purpose of this study is to determine the efficacy of composite repair utilizing BER epoxy by completing a parametric study.
METHODS
Materials
Epoxy synthesis materials such as metal catalyst Zn(Ac)2, diglycidyl ether of bisphenol A (DGEBA), and fatty acid Pripol 1040 were supplied by Sigma-Aldrich Inc. A 3K plain weave carbon fiber was ordered from Fibre Glast Development Corporation. The material has a thickness of 0.305 mm and a weight of 198 g/m2. High temp release film from Fibre Glast Development Corporation,
17


item number 1782. Nylon release peel ply from Fibre Glast Development Corporation, item number 582. Breather fabric from Plasticare.
Epoxy Formulation
The epoxy was synthesized utilizing the same techniques outlined by Yu.[15] Fatty acid was mixed with metal catalyst at a ratio of 10 to 0.89 grams. The mixture was then placed in a vacuum oven at a temperature of 140° C until no bubbles from the mixture were observed(approximately 3 hours). DGEBA was then added to the mixture at a ratio of 10 grams fatty acid to 5.67 grams DGEBA. This was thoroughly mixed together for later use in pre-preg manufacturing.
Pre-preg Manufacturing
Specimens were created using pre-preg carbon fiber that was manufactured as follows. BER epoxy was heated to approximately 100° C to soften the epoxy. Woven neat carbon fiber was laid flat on plastic sheet. The warm epoxy was then poured over the carbon fiber, a squeegee was used to spread the epoxy evenly. The carbon fiber was flipped over and the process repeated in the other side of the fiber to completely and evenly coat and impregnate the carbon fiber with epoxy. The pre-preg was wrapped in plastic sheet and placed flat in a freezer for later use (see Figure 8). The pre-preg had an initial mass fraction ratio of 1:1 carbon fiber-to-epoxy. Specimens post cure had a mass fraction of carbon fiber of 69-74% fiber mass.
18


Figure 8 Manufacturing steps for making pre-preg carbon fiber/BER epoxy A) neat carbon fiber B) neat carbon fiber with epoxy C) neat carbon fiber with epoxy covered with plastic sheeting D) epoxy spread evenly through carbon fiber
Composite Specimen Manufacturing
Three pre-preg sheets of identical dimensions were cut to later create the laminate. The sheets were laid on top of one another, three layers thick. To create a specimen with inter-lamina defects, a 30 mm wide strip of PTFE was placed across the width of the composite between the first and second sheet as well as the second and third sheet. The PTFE strips were laid in line with each other to create a uniform defect through the middle of the width of the specimen. This was then cured using vacuum bagging technique detailed below. The composite containing both defect and non-defect areas were cut into individual specimens to be tested. These cured composites samples were cut into specimens approximately 60 mm x 14 mm. The diagram in Figure 9 shows how the PTFE strips were placed in the laminate before the cure process. The photo in Figure 10 show what the laminate looks like after the cure process, before it is cut into individual specimens.
19


Figure 9 Diagram of the manufacturing process of the defect specimens
A) Top view of defect laminate with PTFE strip shown by stripe pattern runs through the laminate length.
C) (4X zoom) Cross-section of defect specimen showing PTFE strip (light grey) placement within specimen.


Figure 10 Image of defect composite laminate with PTFE strips on nylon fabric
Vacuum Bagging Process
The composite is placed between two nylon peel ply to ensure epoxy distribution throughout the surface of the composite. The sample is then placed between two non-perforated release films. The film is non-perforated to keep as much epoxy within the specimen as possible. The sample is then placed on a flat metal plate to create a flat specimen, which is then covered with a layer of breather fabric and placed in the vacuum bag(see Figure 11).
21


Figure 11 Picture of laminate in vacuum bag before cure
Dry Repair Process
The repair process utilizes pressure and heat to complete a repair. To achieve a constant temperature and pressure an MTS machine equipped with a thermal chamber and compression platens was used(see Figure 12). Three specimens with inter-lamina defects were stacked directly on top of one another with a thin sheet of PTFE separating the specimens to prevent contact between specimens. The method of stacking specimens was used to assure that all the specimens received the same pressure and temperature. The curing temperature was 180°C. The pressures applied to the specimens were 500, 750, 1000, 1250, and 1500 kPa. Time duration of each of these repairs was one hour. Only the pressure of the repair was altered. For tests that would examine the effect of time duration of the repair a constant pressure of 500 kPa was applied to each variation of time duration. The time of repairs were 1, 2, 4, and 6 hours.
22


Figure 12 Picture of compression platens in oven chamber utilized in the dry method repair
Testing Procedure
The composite specimens were tested in three point bending on an MTS 858 Mini Bionix II machine. A 100 N load cell was used. Load and displacement of the crosshead were measured. The crosshead rate was set to 1 mm/min as designated by ASTM standard D7264. Per ASTM section 8.2 a support span-to-thickness ratio of 60:1 was chosen. The average specimen thickness was 0.635 mm The span length was set to 38 mm to achieve the 60:1 ratio desired(see Figure 13). The test was run
23


until a crosshead extension of 15 mm was completed. The reasoning for this exact extension limit was to achieve enough data to calculate both stiffness maximum stress of the specimen.
6543210123456
Centimeters
Figure 13 Image of MTS 858 mini bionix ii with three-point bend test fixture with specimen
after fracture
Failure Mode Analysis
Failure analysis of the specimens were conducted to evaluate the effectiveness of the repair procedure. The failure modes were analyzed using the format in the ASTM standard used for testing and data analysis [1]. Each specimen was analyzed during and after testing to determine the failure mode. The mode of failure was determined by three characteristics shown in Figure 14.
24


Second Character
First Character
Failure Mode Code
Tension T
Compression C
Buckling B
interlaminar Shear S
Multi-mode M(xyz)
Other O
Failure Area Code
At loading nose A
Between loading noses B
at Support nose S
between Load and support nose L
Unknown U
Third Character
Failure Location Code
Top T
Bottom B
Left L
Right R
Middle M
Various V
Unknown U
Figure 14 Flexure Test Specimen Three-Part Failure Identification Code [1]
Data analysis
Each specimen is analyzed by calculating the strength and stiffness. Stiffness is calculated by determining the slope of the stress(strain) curve between 0.001 and 0.003 strain. Strength of the specimen is calculated by determining the maximum stress experienced by the specimen. Stress experienced by the specimen is calculated using eq. 1 .[17] Maximum strain experienced by the specimen at mid-span is calculated using eq. 2.[17]
1)(T =
3 PL 2 bh2
6 Sh
2)£=ir
where:
a = stress at the outer surface mid-span, MPa
P = applied force, N
L = support span length, mm
b = specimen width, mm
h = specimen thickness, mm
e = strain at outer surface, mm/mm
S = mid-span deflection, mm
25


RESULTS
The strength and stiffness for each specimen were cataloged. The three specimens for each test condition were compared and standard deviation for stiffness and strength calculated. The average stiffness and strength along with standard deviation are shown in Table 1. The average stiffness and strength for each test condition were compared to other variations of either pressure of time duration.
Table 1 Average stiffness and strength of each sample set
Average Stiffness (MPa) Std. dev (MPa) Average Strength (MPa) Std. dev (MPa)
Time Tests (hours)
1 16265.51 5323.62 87.32 25.57
2 20944.30 154.26 130.16 17.19
4 28439.40 79.48 131.44 13.08
6 25380.39 3041.60 127.54 4.63
Pressure Tests (kPa)
500 16265.51 5323.62 87.32 25.57
750 19935.73 1330.53 87.04 9.07
1000 25030.82 2559.62 98.89 13.09
1250 26059.21 2994.36 126.11 2.52
1500 24172.94 2273.47 124.04 11.59
Applied Pressure of Repair
The average stiffness and strength of each variation of applied pressure during repair are plotted as a function of pressure in Figures 15 and 16 respectively. Average stiffness of the specimen increases as applied pressure is increased, until 1250 kPa. The average stiffness of repaired specimens decreases at a pressures higher than 1250 kPa. Specimens repaired at a pressure of 1500 kPa are lower than that of 1000 kPa or 1250 kPa. While average stiffness of 1000, 1250, and 1500 kPa specimens are different; a t-test showed that the values are not statistically different.
26


35000
30000
25000
20000
& 15000
10000
5000
250
500 750 1000 1250 1500 1750
Pressure Applied (kPa)
Figure 15 Plot of stiffness of repaired specimens for each variation in applied pressure
The average strength of the specimens in Figure 16 increases to a maximum value at 1250 kPa. The strength looks as though it is going to continue increasing, but decreases when a pressure of 1500 kPa is used for repair. The decrease seen in Figure 15 is also seen in Figure 16 at a pressure of 1500 kPa. The maximum average stiffness and strength of a repaired specimen occurred at an applied pressure of 1250 kPa.
27


160
20
0 --------------------------------------------------------------------------------------------
250 500 750 1000 1250 1500 1750
Pressure Applied (kPa)
Figure 16 Plot of strength of repaired specimens for each variation in applied pressure
Time Duration of Repair
The average stiffness and strength of each variation of duration of repair are plotted as a function of time in Figures 17 and 18 respectively. The average stiffness of specimen increases as the duration of the repair is increased. Average stiffness increases linearly, then peaks at four hours, and then decreases at six hours. In Figure 17, the average strength of specimens for each dataset increases as the duration of the repair is increased, then plateaus after four hours of repair until the six hour mark. At six hours of repair the average strength of the specimens slightly decrease when compared to the four hour mark. The average strength of each data set in Figure 17 is shown to increase in a
28


slightly linear pattern and then decrease linearly as well. The maximum average stiffness and strength of a repaired specimen occurs at 6 hours of repair duration.
Duration of Repair (hours)
Figure 17 Plot of stiffness of repaired specimens for each variation in time duration of repair
29


160
140
120
100
£ 80 p
co
60
40
20
1 2 3 4 5 6
Duration of Repair (hours)
7
Figure 18 Plot of strength of repaired specimens for each variation in time duration of repair
Failure Modes of Pressure and Time Tests
The examination of the failure mode of each specimen is critical to understanding the effectiveness of the repair parameters. The failure modes of each specimen in the pressure and time tests are shown in Table 2-3. The tables are filled out using the characterization guide set forth in the ASTM standard and Figure 14. The specimens either failed due to interlaminar shear, buckling of the top lamina, or a combination of both of these. Figure 20 shows an SEM image of buckling fracture of a specimen repaired at 1000 kPa.
30


Table 2 Failure modes of each specimen of the pressure study
Failure Mode
Pressure (kPa) A B C
500 SLM SLB SLM
750 M(S,B)LM M(S,B)LM SLM
1000 BST OST BST
1250 BST BST BST
1500 BST BST SLM
Table 3 Failure Modes of each specimen of the time study
Failure mode
Time (hours) A B C
1 SLM SLB SLM
2 BST SLM BST
4 BST BST M(S,B)LT
6 BST BLT BST
Analysis of the tables show that at pressures less than 1000 kPathe specimen would delaminate first at the edges of the repaired defect and grow in length toward the center of the specimen. The same failure mode also occurred on time specimens that were repaired for less than 4 hours. Specimens that experienced pressures greater than or equal to 1000 kPa during repair would remain rigid and only flex then fracture under buckling, although one 1500 kPa specimen failed due to interlaminar shear and delamination. The modes of failure can be seen in Figure 19.
31


1.0mm'
Figure 19 Failure modes of A) buckling fracture and B) delamination
3/29/2018 HV WD Mag X: -6.9 mm Tilt 7:59:29 AM 15.0 kV 10.1 mm 50x Y: 0.2 mm -0.1 °
Figure 20 SEM Image of specimen repaired at 1000 kPa for one hour, showing buckling of
lamina (Photo provided by May Martin)
32


Optimal Pressure and Time Duration of Repair
With these results the a sample set of specimens repaired for 4 hours with an applied pressure of 1250 kPa were created and tested. A representative optimal repair specimen and representative control specimen are plotted on a stress as a function of strain chart in Figure 21. The average stiffness and strength of the optimal repair specimen are 69.4% and 85.8% of the average control specimens respectively. A second set of specimens was also repaired under the same conditions and tested. The average stiffness and strength of the optimal repair specimen are 74% and 86% of the average control specimens respectively.
0 0.005 0.01 0.015 0.02 0.025 0.03
Strain (lnin/mm)
Figure 21 Stress(strain) plot of representative samples of a control and optimal repair specimen
33


DISCUSSION
This study showed that pressure and time do have a measureable effect on the epoxy and therefore the repair method process. The time study revealed a maximum stiffness and strength occurred at time of four hours and a pressure of 500 kPa, resulting in a recovery of stiffness and strength of 99% and 95% respectively. This result is as expected because more bonds will be reconnected as a function of time. The decrease in stiffness and strength in specimens repaired for six hours is likely due to epoxy degradation. A temperature of 180°C is orders of magnitude the 30°C Tg of the epoxy. This high heat for an extended period of time is likely to cause damage and degradation to the epoxy structure. This plateau where the maximum number of bonds are reconnected without epoxy degradation should be explored in further studies. The pressure study revealed a maximum stiffness at 1250 kPa for a duration of one hour, instead of the expected 1500 kPa. This surprising result could be explained by epoxy fractures(see Figure 22). It is likely that above a certain experienced pressure the epoxy begins to fracture under this load. The resulting fractured epoxy creates more unwanted defects and therefore does not behave in the same manner as it did at a lower repair pressure.
An optimal set of parameters for repair was determined to be an applied pressure of 1250 kPa for a duration of 4 hours, although further tests would reveal that this was not an optimal set of parameters. The resulting repair method was show to be an ineffective method of repair resulting in a recovery of 69.4% stiffness and 85.8% strength. Repeated tests showed similar results. The parameters used for this optimal repair were the cause for the sharp drop in recoverable stiffness and strength. The assumption before testing began was that the optimal pressure and time would create the optimal repair, but this study did not show that. Instead the high pressure of 1250 kPa and extended period of time of four hours likely contributed to damaging the epoxy. It is shown that high pressures damage the epoxy by fracture in Figure 22. Extended durations repair are also shown to decrease the recoverable strength and stiffness of a specimen, likely due to epoxy degradation. The
34


results of this study revealed a relationship between pressure of repair and duration of repair. Further research into this relationship is necessary to fully understand the ideal parameters for composite repair.
Figure 22 SEM Picture of specimen repaired at 1500 kPa, showing epoxy fractures (Photo provided by May Martin-NIST)
CONCLUSION
The results of this research indicates that BER epoxy is a viable composite epoxy that allows for repairs. With a recoverability of nearly 100% stiffness and strength, a component can be repaired without complicated machining or chemical processes. The method of repair dictated in this paper is shown to be a viable and possibly more efficient repair method when compared to present methods such as epoxy injection. This method can be easily exploited for use in the field or even as a manufacturing method. This repair process can be used for quick turnaround of damaged components
35


or parts on vehicles and structures. To further corroborate these findings an epoxy injection repair process should be explored in future works. One disadvantage of this particular BER epoxy is its low Tg of 30°C. Future research into BER epoxies should examine those BER epoxies with a higher Tg. BER epoxy allows for the possibility to repair surface and subsurface defects with components in place on structures and vehicles in the field.
36


REFERENCES
[1] A. S. Tim Hepher. (2013). The Dreamliner Debacle Has Already Cost Boeing $600 Million. Available: http://www.businessinsider.com/dreamliner-trouble-has-cost-boeing-600-million-2013-4
[2] T. N. Y. Times. (2013). The Jump to a Composite Plane. Available: http://www.nvtimes.com/interactive/2013/07/29/business/The-Jump-to-a-Composite-Plane.html
[3] R. ZumMallen. (2018). 2019 CMC Sierra 1500 Adds Carbon Fiber Box. Available: https://www.trucks.com/2018/03/01/2019-gmc-sierra-150Q/
[4] D. White. (2005). New trucks from Toyota and Honda get it Right - with composite beds. Available: https://www.compositesworld.com/columns/new-trucks-from-tovota-and-honda-get-it-right—with-composite-beds
[5] J. Hirsch. (2018). Popularity of Pickup Trucks Drive 2017 Auto Sales. Available: https://www.trucks.com/2018/01/04/2Q17-pickup-trucks-auto-sales/
[6] M. T. Amy Bauer, Kristine Obusek, Mufit Akinc, "Bisphenol E cyanate ester as a novel resin for repairing BMI/carbon," Elsevieer, no. Polymer, 2013.
[7] E. S. Kristin Kuchler, Rolf-Dieter Hund, Olaf Diestel, Martin Kirsten, Chokri Cherif, "Local repair procedure for carbon-fiber-reinforced plastics by refilling with thermoset matrix," Journal of Applied Polymer Science, 2013.
[8] H. K. Paolo Feraboli, Bonnie Wade, Federico Gasco, Luciano DeOto and Attilio Masini, "Recyclability and reutilization of carbon fiber facbric/epoxy composites " Journal of Composite Materials 2011.
[9] L. F. M. D. S. K.B. Katnam, T.M.Young, "Bonded repair of composite aircraft structures: A review of scientific challenges and opportunities," Elsevier, no. Progress in Aerospace Sciences, pp. 26-42, 2013.
37


[10] H. Bendemra, P. Compston, and P. J. Crothers, "Optimisation study of tapered scarf and stepped-lap joints in composite repair patches," Composite Structures, vol. 130, pp. 1-8, 2015/10/15/2015.
[11] F. Benyahia, A. Albedah, and B. Bachir Bouiadjra, "Analysis of the adhesive damage for different patch shapes in bonded composite repair of aircraft structures,"
Materials & Design (1980-2015), vol. 54, pp. 18-24, 2014/02/01/2014.
[12] M. Kashfuddoja and M. Ramji, "Design of optimum patch shape and size for bonded repair on damaged Carbon fibre reinforced polymer panels," Materials & Design
(1980-2015), vol. 54, pp. 174-183, 2014/02/01/2014.
[13] R. F. E.-H. Seyedmohammad S. Shams, "Overlay patch repair of scratch damage in carbon fiber/epoxy laminated composites," Elsevier, no. Composites: Part A, pp. 148-156, 2013.
[14] F. Ning, W. Cong, J. Qiu, J. Wei, and S. Wang, "Additive manufacturing of carbon fiber reinforced thermoplastic composites using fused deposition modeling," Composites Part B: Engineering, vol. 80, pp. 369-378, 2015/10/01/2015.
[15] Q. S. Kai Yu, Martin L. Dunn, Tiejun Wang, and H. Jerry Qi, "Carbon Fiber Reinforced Thermoset Composite with Near 100% Recyclability," Advenced Functional Materials, 2016.
[16] Q. Shi, K. Yu, M. L. Dunn, T. Wang, and H. J. Qi, "Solvent Assisted Pressure-Free Surface Welding and Reprocessing of Malleable Epoxy Polymers," Macromolecules, vol. 49, no. 15, pp. 5527-5537, 2016/08/09 2016.
[17] "D7264-Standard Test Method for Flexural Properties of Polymer matrix Composite Materials," 2017.
38


Full Text

PAGE 1

COMPOSITE REPAIR UTILIZING BOND EXCHANGE REACTION POLYMERS by K YLE W. E GAN B.S., Northern Arizona University, 2015 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for t he degree of Master of Science Mechanical Engineering 2018

PAGE 2

ii This thesis for the Master of Science degree by Kyle W. Egan has been approved for the Mechanical Engineering Program by Ron Rorrer , Chair Kai Yu Dana Carpenter Date: May 12, 20 18

PAGE 3

iii Egan, Kyle W. (M.S., Mechanical Engineering) Composite Repair Utilizing Bond Exchange Reaction Polymer Thesis directed by Associate Professor Ron Rorrer ABS T R ACT The use of a bond exchange reaction epoxy in carbon fiber reinforced composites is inve stigated for us e in the repair of composites. Three ply c omposite specimens were created with an inter lamina defect using PTFE strips during the curing process. The specimen was repair ed by applying pressure and a temperature of 180 ° C . The flexural stiff ness and strength was tested using a three point bend test. The repaired specimens are shown to retain stiffness and strength up to 9 9 % and 95% respectively that of a specimen without a defect. The experiments examined the relationship between time and pre ssure on the effectiveness of the repair process. A parametric study was completed to determine the optimal pressure and time duration of repair for a delaminated bond exchange reaction epoxy composite. The study revealed an increase in strength and stiffn ess of the composite as time and pressure were increased. The study shows an increase in strength compared to other methods used for composite repair, while not increasing weight or dimensions of the composite. The method discussed in this paper is shown t o recover 95% strength, proving the efficacy of the method as well as use of bond exchange reaction epoxy within composites for composite repair. The form and content of this abstract are approved. I recommend its publication. Approved: Ron Rorrer

PAGE 4

iv A CKNOWLEDGEMENTS I would like to thank the employees at National Institute of Standards and Technology who were involved in conducting testing for this research project as well as providing scanning electron microscopy images: [ Applied Chemicals and Materia ls Division , Andrew Slifka, and May Martin ]. Without their professional help and input, this research could not have been successfully completed.

PAGE 5

v TABLE OF CONTENTS CHAPTER I. The Growing Need for Composite Repair ................................ ................................ ......................... 1 Literature Review ................................ ................................ ................................ ............................. 2 Epoxy Refill Repair ................................ ................................ ................................ ........................ 2 Matrix Removal ................................ ................................ ................................ .............................. 2 Scarf Joints ................................ ................................ ................................ ................................ ..... 4 Overlay Patch Repair ................................ ................................ ................................ ...................... 5 Bond Exchange Reaction Epoxies ................................ ................................ ................................ .. 6 II. Preliminary Studies ................................ ................................ ................................ .......................... 9 Defect Repair ................................ ................................ ................................ ................................ .. 9 Single Lamina Manufacturing and Parametric Study ................................ ................................ ..... 9 Curing Effect Revealed ................................ ................................ ................................ ................ 12 Solution Injection Repair Metho d ................................ ................................ ................................ . 13 III . Parametric Study Simulating Delaminated Composite Repair ................................ ...................... 17 Methods ................................ ................................ ................................ ................................ ........ 17 Materials ................................ ................................ ................................ ................................ ..... 17 Epoxy Formulation ................................ ................................ ................................ ..................... 18 Pre pre g Manufacturing ................................ ................................ ................................ .............. 18 Composite Specimen Manufacturing ................................ ................................ .......................... 19 Vacuum Bagging Process ................................ ................................ ................................ ........... 21

PAGE 6

vi Dry Repair Process ................................ ................................ ................................ ..................... 22 Testing Procedure ................................ ................................ ................................ ....................... 23 Failure Mode Analysis ................................ ................................ ................................ ................ 24 Data analysis ................................ ................................ ................................ .............................. 25 Results ................................ ................................ ................................ ................................ ............. 26 Applied Pressure of Repair ................................ ................................ ................................ .......... 26 Time Duration of Repair ................................ ................................ ................................ .............. 28 Failure Modes of Pressure and Time Tests ................................ ................................ .................. 30 Optimal Pressure and Time Duration of Repair ................................ ................................ ........... 33 Discussion ................................ ................................ ................................ ................................ ....... 34 Conclusion ................................ ................................ ................................ ................................ ....... 35 REFERENCES ................................ ................................ ................................ ................................ ... 37

PAGE 7

1 CHAPTER I THE GROWING NEED FOR COMPOSITE RE PAIR On January 16, 2013 the Federal Aviation Administration grounded the Boeing 787, costing the company an estimated $600 million dollars . [1] The plane was grounded for three months as a solution to a battery fire problem wa s found. This event illustrates just how important a quick turnaround is when it comes to repairing aircraft components. The Boeing 787 has more composites per weight than any other material in the aircraft . [2] Dam age to composite components can render them nonfunctional. This is especially true for any part that is structural to the airframe or aerodynamics. Composites are widely being utilized in industries ranging from fashion and sporting goods to automobiles an d large commercial aircraft. Composites are seeing a boom in the automotive industry where they are no longer used only in exotic supercars. As composite materials become more affordable and necessary to decrease vehicle weight, the automotive industry wil l being to incorporate them into their vehicles structures. General Motors Company is now utilizing carbon fiber to create its truck beds rather than steel . The use of carbon fiber in the 2019 GMC Sierra has decreased the weight of the truck bed by 62 poun ds over a standard steel truck bed, leading to a more fuel efficient vehicle . [3] Toyota and Honda already utilize composites in their m id size trucks which account ed for over 200,000 sales in 2017 alone . [4, 5] These sales figures are i mpressive on their own but are just a fraction of the total truck market in the United States. A total of 2.37 million ful l size trucks were sold in 2017, most of those being the Ford F 150 . [5] When a large American truck manufacturer such as Ford or Chevrolet finally adopt s composite material for their truck beds , a low cost a quick composite repair system will be required. A cost effective repair process that returns a damaged part to a functional percentage of it s original strength is necessary for the future of composites.

PAGE 8

2 LITERATURE REVIEW Epoxy Refill Repai r Current methods of composite repair utilizes varying forms of epoxy injection or patch repair. The technique of epoxy refill repair makes use of low viscosity adhesives or epoxies applied directly to the damaged area of the composite. The epoxy can be i njected to the damaged area as in Bauer [6] or applied like the original epoxy process. This injection process allows the technician to methodically apply the e xact amount of epoxy as well as being able to apply it to the exact place of damage on the composite. Is the case of Bauer [6] the epoxy was injected between la mina. The other end of the spectrum of refill repair spectrum is used by Kuchler [7] by applying epoxy to the damaged area using the Seeman composite resin infusion molding process (SCRIMP) method. The carbon fiber is placed in a vacuum bag that pulls re sin into the bag through vacuum hoses which then infuses the car bon fiber to cure into a laminate . This process can be applied locally to selected area by securing a vacuum tight seal around the damaged area before resin is infused. Kuchler [7] shows that this process is a viable technique to infuse a damaged area of a composite with new resin. However this research is limited i n its scope because it does not test the effectiveness of the infused epoxy to recover the original strength of the composite. Kuchler [7] falls short of a researching a complete process of matrix removal, resin injection, and finally testing of composite. Without the quantification of composite strength testing afte r resin infusion, it is unknown whether the SKRIMP process can be viable as a method to repair a damaged composites. Kuchler [7] choses to instead focus on the process of matrix removal. Matrix Removal The process of matrix removal is necessary in the repair of a damaged area of a composite. If broken fragments of th e matrix are often left inside the damaged area of the composite the strength recovery in that area will be compromised. Fracture or damage to the resin within a composite can

PAGE 9

3 lead to failure of the composite. The carbon fiber and epoxy matrix that make up a composite are diffic ult to separate due to the resin thermoset properties. A thermoset polymer does not have a melting temperature due to the crosslinking polymer chains that makeup the epoxy. When put under high heat the epoxy will on ly degrade but wil l say intact, therefore thermal separation of the materials is unattainable. Due to the difficulty of this process other methods such as acid washes and supercritical water baths have been the only known method to separate the matrix from the fiber without damaging the fiber . [8] The use of harsh chemicals as a solvent leads to many ecological and economic pro blems. In response to this Kuchler [7] has developed a process that uses infrared radiation instead of chemicals to remove the matrix from the fiber. The process begins by applying a thin layer of Cr 2 O 3 powder on the damaged area of the composite. A glass nonwoven fiber is placed on top of this to shield the undamage d area from radiation. Finally an IR lamp is placed over this and run at 380 W for 20 minutes. The matrix degrades and through chemical reaction with the Cr 2 O 3 powder is removed from the fibers. In this process the fibers strength is reduced by 10%. The pr ocess is shown to be a viable way to separate the matrix from the fiber. The removal can be localized relatively easily because the removal only occurs where the power is located . The precision available with this technique makes it easier to only remove t he damaged matrix. In other methods such as those that use acids to dissolve the matrix the entire composite sample must be submerged in the acid the dissolve the matrix. Therefore the process from Kuchler [7] is valuable in that it can be localized to the damaged area. It is still limited however by its extensive use of electrical energy to ir radiate the composite as well as its waste product of Cr 2 O 3 powder. The process can also only penetrate the composite to dissolve the matrix to a depth of approximately two millimeters. There is still a need to develop a viable m ethod of matrix removal that does not create waste or consume vast amounts of energy.

PAGE 10

4 Scarf Joints The composite repair method used in aviation utilizes a machining process to cut out the damaged area exposing a clean surface . [9] The void is then filled with composite patches and re cured. These patches are either pre cur ed composite epoxy lamina or clean carbon fiber weave patches . Th e pre cured patch is sealed a nd resin transfer method (RTM) used to fill the area with epoxy to create a laminate. The patch can also be a clean carbon fiber weave, which is cured into a lami nate using RTM . RTM is s imilar to SCRIMP method discussed earlier . RTM is used to ensure a desired volume fraction of epoxy to carbon fiber. While the scarf method is effective at removing damaged fibers and matrix material there are a few negatives associ ated with this process. The new patches placed into the scarf are not carrying the same load as the original fibers before the damage occurred, this leads to a weaker part. The patch cannot distribute the load as effectively as the original because the new and original fibers are separated by a layer of epoxy. The epoxy cannot withstand the forces that the fiber can and so a weak point in the part has just been manufactured. The shape of the scarf and patch is another variable to consider . A scarf can eithe r be stepped as seen in Figure 1 or straight. Figure 1 Straight taper and Stepped taper scarf joints used in laminate composite repair [9] A stepped tapper scarf creates local stress concentrations of two or three times that of straight tapered scarf joint s . [10] Studies have shown stress concentrations are strongly correlated to ply thickness, adhesive thickness, taper angle, stacking sequence, overpl y layup, and overply lap

PAGE 11

5 length. [ 10 ] The tapered scarf joint is used more widely than the stepped scarf due to its performance. Patch shape is another variable to consider in the scarf repair process. Multiple studies that have examined the effect of patch shape of the repair process differ on the efficiency of differe nt shapes depending on the loading and other factors associated with patch repairs . The recovery strength of a laminate with patch repair is dependent on multiple factors that makes it difficult to determine an optimum patch for all conditions. These facto rs are fiber orientation, number of layers, stacking sequence, loading condition of the part, material used, and over ply among others. With these factors considered studies have shown that for a simple over ply patch with no scarf joint and simple tension loading, a rectangular over ply patch is best for epoxy strength . [11] This same study however concedes that a rectangular patc h has increas ed stress intensity at the crack compared to an oval or octagon shape . [11] Another study support s the use of a octagon shaped patch as an op timum shape . [12] This study also indicated that the stacking sequence contributed to the effectiveness of the patch. A rectangular shape was optimal for sequence 0/ ± 45 /90 and 0/90/ ± 45 while an oval shape was optimal for sequence ± 45/0/90 and 90/0/ ± 45 . [12] Scarf repairs need to be implemented accurately and this makes this a time con suming and technical task. The technician completing the repair must be experienced not only in composites but also in machining. For these reasons the scarf joint is mostly used for structural or thick composite laminates. Overlay Patch Repair Overlay p atch repair utilizes multiple other techniques to create a strong repair. In Shams [13] a scarf cut i s made in the damaged area, then is injected with epoxy, then a patch of carbon fiber is laid over the scarf area. This overlay patch can be one to three layers and is meant to resist the bending associated with the weakened damaged area . These patches inc rease the strength of the composite as well as distributing the stress around the area of damage. Without this external patch on the outside of the damaged area, the epoxy/resin inside the void does not add enough strength to deem

PAGE 12

6 the composite repaired. T he overlay patch of carbon fiber fixes this issue, resulting in a considerable amount of strength recovered, but it also adds thickness and some weight. The added thickness can contribute to a change in aerodynamics which is not viable for aircraft. This g ap in the composite where the damaged composite has been cut and replaced with epoxy is a large weakness in the composite. The overlay patch is a way of strengthening that void of fiber. While the overlay patch can be an effective way of increasing strengt h recovery and distributing the load around the damaged area there are still unknown such as patch shape as previously mentioned . [11, 12] Bond Exchange Reaction Epoxies All of the aforementioned repair methods hav e one common flaw. These repair methods rely on new epoxy bonding to the original epoxy of the part. The problem with this process of epoxy to epoxy bonding are that this does not create covalent bonds in the epoxy matrix but relies on secondary bonding fo r stress transfer . When an epoxy cures it is covalently bonded together. The reasoning for this are the thermosets crosslinking. The available crosslinking reaction sites have all been used during the original curing process . The polymer structure of the e poxy is relied on to fill and cling to micro defects on the original epoxy in conjunction with secondary bonding to increase adhesion . These secondary bonds are much weaker than the covalent bonds of the original epoxy. The weak bonds are part of the reaso n that repairing composite structures to their original strength is difficult to accomplish. Thermosets lack crosslinking between the polymer chains which allows them to move past each other at elevated temperature; allowing flow of the material . Thermopla stics on the other hand do have the ability to be reformed and change shape when heat is applied. Thermoplastics are also much weaker than a thermoset beca use of the lack of crosslinking. Strides to implement thermoplastics into composites have been taken and are most successful in 3D printing. Thermoplastics can be easily 3D printed due to its ability to flow at elevated temperatures. Studies have found that while adding carbon fiber particles to the thermoplastic when printing increases the modulus, ther e is only a slight increase . [14] For these reasons a thermoset is ideal for composite

PAGE 13

7 laminates. What is needed to improve on curr ent repair methods is a matrix material that crosslinks like a thermoset but can also reconnect the crosslinking polymers when damaged. A new class of polymer called bond exchange reaction (BER) or covalent adaptive network (CAN) polymers are currently be ing researched. This type of polymer blurs the line between thermoset and thermoplastic. BER polymers/epoxies utilize a covalent adaptive network that allows the epoxy to re crosslink severed chains . [15] With applied heat , the epoxy can reconnect these covalent bonds that have been previously broken (see Figure 2) . The ability of BER epoxies to repair a damaged surface have be en previously demonstrated . [15, 16] In these studies both recyclability of the polymer and pressure free solvent assisted repair are demonstrated. The fir st study demonstrated the ability to use BER epoxy in a composite lamina repair. The surface of a single BER epoxy composite was damaged by scratching it and the n repair using powdered BER epoxy and solvent. The repaired lamina was able to recover 97% o f the fiber modulus compared to a lamina which was not damaged. [15] The study demonstrated the positive uses for BER epoxy as a surface repair epoxy as well as a recyclable matrix by utilizing solvent dissolution . This experiment showed that utilizing BER epoxy could create a near 100% closed loop process for manufacturing, recycling , and r epurposing materials in composites. Th e later solvent ass isted pressure free method was sh own to r epair a specimen from two fractures pieces into on piece using ethylene glycol. The study showed a recovery of 74% of the original modulus, demonstrating the BER epoxy s particularly unique ability of solvent dissolution. [16] Compared to other thermosets BERs are unique in t hat they can be dissolved by a solvent and then solidified as the solvent is evaporated. This ability makes BER epoxies applicable to more than one method or repair and an ideal candidate epoxy for composite repair.

PAGE 14

8 Figure 2 Di agram showing the covalent adaptive network (CAN) of the BER epoxy [16]

PAGE 15

9 CHAPTER II PRELIMINARY STUDIES Preliminary resea rch was completed to develop a greater understanding of how the BER epoxy would perform as a composite epoxy. To study this, multiple studies were completed and the results documented to assist in the final research study. Defect Repair An initial study of a specimen that simulated a repair of a delaminated specimen was completed as a proof of concept. A laminate made from three layers of carbon fiber reinforced BER epoxy was created. A strip of PTFE was placed b etween layers 1 2 and 2 3, through the width of the laminate. The PTFE strips were laid in line with each other to create a void through the width of the sample. The laminate was cured using a vacuum bagging p rocess discussed in C hapter III. The PTFE strips , of length 45 mm, were then removed from th e specimen to create a void between the lamina, simulating a delaminated part. The specimens were repair using the Dry test method detailed in Chapter III . A pressure of 500 kPa was applied for two hours, at a temperature of 180 ° C. The specimens were then tested utilizing a 3 point bend method. The results of this study was recovery of 86% stiffness. This result gave proof that with some fine tuning of the repair process a significant recovery of strength and stiffness of a part could be achieved. Single Lamina Manufacturing and Parametric Study The BER epoxy also lends itself to new manufacturing techniques that have not been previous ly possible. The covalent adaptive network of the BER epoxy can be exploited to not only repair damaged areas but also to manufacture laminates for single cured lamina. Single lamina carbon fiber/BER epoxy sheets are cured and stored. These sheets of cured lamina can then be shipped to any destination without the use of refrigeration or fear of degradation of the epoxy. The c ured sheets can then be cut to a desired shape and used in conjunction with other sheets to create a laminate. The technician would simply need to apply a distributed pressure while heating the laminate to 180 ° C for

PAGE 16

10 a few hours. The manufacturing process would need no additional epoxy. This process could also permit the technician to apply patches to surfaces in an efficient and timely manner. To test this idea, a single flat lamina of carbon fiber/BER epoxy was created using a vacuum bagging process. Thi s single lamina was cut into specimen dimensions of 55 mm by 15 mm. The single lamina specimens were placed on top of one another to create a 3 ply laminate specimen of dimension 55 mm by 15 mm. These laminate specimens were then repaired using applied pre ssure and heat in the form of a heated platen press. Two parameters were tested for this study, pressure and duration of repair. To examine the effects of pressure on a repaired specimen, specimens were repaired at pressures of 500, 625, 750, 875, and 1000 kPa for a duration of one hour. To examine the ef fects of duration of the repair, specimens were repaired at 500 kPa for a duration of 1, 2, 3, and 4 hours. All tests were performed at a temperature of 180 ° C. The results of this parametric study proved that there was indeed a relationship between the time of the repair, pressure applied during the repair and the recovered strength and stiffness of the laminate. The stiffness and strength of the repaired specimens increased as pressure and time was increa sed. The relationship however was not linear and seemed to taper off at the high pressures and longer times . For the time study, there was only a slight difference between specimens repaired for 3 hours and specimens repaired for 4 hours (see Figure 3) . The pressure application study revealed a similar relationship, with specimens repaired at 875 kPa and 1000 kPa exhibiting almost identical strength and stiffness recovery (see Figure 4) . The recommendation for repair of specimens using this method was then de termined to be an applied pressure of 875 kPa at a temperature of 180 ° C for a time of 3 hours. This recommendation was determined to reduce the amount of resources while still maximizing the strength and stiffness recovery.

PAGE 17

11 Figure 3 Stress(strain) chart of specimens repaired for 1, 2, 3, and 4 hours at 500 kPa Figure 4 Stress(strain) chart of specimens repaired for 500, 625, 750, 875, and 1000 kPa for 1 hour 0 10 20 30 40 50 60 0 0.005 0.01 0.015 0.02 0.025 0.03 Stress (MPa) Strain (mm/mm) 1hr-A 1hr-B 2hr-A 2hr-B 3hr-A 3hr-B 4hr-A 4hr-B 0 10 20 30 40 50 60 0 0.005 0.01 0.015 0.02 0.025 0.03 Stress (MPa) Strain (mm/mm) 500kPa-A 500kPa-B 625kPa-A 625kPa-B 750kPa-A 750kPa-B 875kPa-A 875kPa-B 1000kPa-A 1000kPa-B

PAGE 18

12 However this study did not meet the goal o f fully repairing single sheets into a solid laminate. The main reason that this study did not achieve this goal was due to the complexities with manufacturing a single sheets of lamina. The sheets were rough and in some places the carbon fiber was void of any epoxy. These voids would drastically reduce the recoverable st rength and stiffness of laminates. This study was abandoned to work on an effective manufacturing process. The manufacturing process that was developed as a result, gives a desired surface texture, even distribution of epoxy, and repeatable results. The process is detailed in Chapter III Methods section. Curing Effect Revealed The previous study revealed a curing effect of the epoxy that made determining the exact parameters for a repair me thod difficult to determine. The problem arose from curing the laminates at 130 ° C and performing repairs at 180 ° C. This variation in temperature resulted in a specimen that was not fully cured. Because the laminates were not fully cured, the laminates ex perienced additional curing when repaired , which created a n additional variable to track during testing . To remove this unknown variable from further testing, specimens were cured at 180 ° C for 6, 8 , and 10 hours. The stiffness and strength were then tested to determine if a significant change could be determined between 6, 8, and 10 hours. This would also show that a specimen cured at 180 ° C for 6 hours and then further repaired at 180 ° C for 2 4 hours would not exhibit a dramatic change due onl y to temperature exposure. The results of this study revealed that specimens cured at 180 ° C did not exhibit a curing effect like that of specimens cured at 130 ° C. Specimens cured for 6 and 8 hours had similar stiffness and strength with slight drop in s tiffness and strength at 10 hours. This slight drop is most likely due to some epoxy degradation. This study revealed a more reliable method for manufacturing specimens for repeatable results. All future specimens were then cured at 180 ° C for 6 hours and repaired at the same temperature.

PAGE 19

13 Solution Injection Repair Method Another benefit of the BER epoxy is that it can be dissolved using ethylene glycol as a solvent. The solvent breaks down the polymer chain. This reaction only occurs at a temperature of 18 0 ° C. Ethylene glycol has a bo iling point of approximately 198 ° C so the ethylene glycol evaporate as the BER epoxy re crosslinks, eventually leaving a repaired BER epoxy sample . Previous works have examined the use of solvent assisted repair of the BER ep oxy. [16] The same idea of solvent assisted repair for composites using BER epoxy was the next logical st ep. Solvent would be injected into a defect area on a specimen and then the specimen would then be placed in an oven at 180 ° C to evaporate the solvent and promote the re crosslinking of the epoxy. Laminate specimens with a defect through the width of the specimen, as descr i bed in Chapter III , were created. Solvent was injected into the void of the specimen then placed in a vacuum bag for repair. The vacuum bag was chosen instead of simply placing th e specimens in an oven because pressure was desired to fully repair the void s. It was thought that if pressure was not applied to the specimen during repair that the solvent would simply evaporate and not repair the void. S pecimen s was placed in the oven for 30 minu tes before a vacuum was pulled on the bag to make sure the solvent was not squeezed out of the void . For those 30 minutes the solvent could break down the polymer chains on the surface of the void creating a tacky epoxy surface on either side of the void. Once the vacuum was pulled and pressure applied, the two tacky sur faces would be pressed together and allowed to re crosslink. The specimens were then left in the o ven for another hour and a half to allow the solvent to evaporate and the specimen to fully repair. The results of the repair yielded a specimen that had defo rmed during the process (see Figure 5 ) . The center of the specimen where the solvent was injected into the void was wavy and not uniform compared to its original shape. The n ylon fabric used in the vacuum bagging process also became attached to the specimen s. The pure solvent method was then abandoned because of these results.

PAGE 20

14 Figure 5 Specimens after solvent injection A different injection method was developed to take the place of solvent injection. The idea of a solvent/epoxy solution was developed. The solution would carry both epoxy and solvent into to defect area. A solution mixture was thought to incorporate the best of the epoxy injection and solvent injection methodologies without the downsides of either. The epoxy inject being that the epoxy resin would need to be kept in a refrigerated state to ensure the epoxy does not degrade. The epoxy resin would therefore take a considerable amount of equipment to utilize in the field. The downside of solvent in jection was discussed previously. The solution injection would only require a heating source which is already necessary for repair. The process for the solution injection method is as follows. Epoxy resin is cured in sheets approximately 5 mm thick. The cu red epoxy sheets are cut into small particles then placed in a beaker. Solvent is then poured into the beaker in the appropriate mass ratio of cured epoxy solvent. The beaker is sealed and heated to 180 ° C to promote the dissolution of cured epoxy. The sol ution is heated until a consistent solution mixture is obtained (4 6 hours). The transformation from cured epoxy to epoxy solvent solution can be seen in Figure 6 .

PAGE 21

15 Figure 6 Pictures of manufacturing epoxy solvent solution. A) Cu red epoxy particles B) Cured epoxy particles with ethylene glycol solvent added before heating C) Epoxy solvent solution after heating to 180 ° C A small study was conducted to determine the best ratio of cured epoxy to solvent by mass. Ratios of cured epoxy to solvent of 1:3, 1:2, 1:1, 2:1, 3:1, and 6:1 were created. It was found that at higher concentrations of solvent such as 1:3, 1:2, and 1:1 , the epoxy simply separated after being heated (see Figure 7 ) . Ratios with higher concentrations of cured epo xy stayed in solution. The ratio of 6:1 cured epoxy to solvent by mass was chosen for a few reasons. The first reason was that it created the most consistent solution in terms of regular dispersion of epoxy into solvent solution. The second reason was that the solution would re cure quicker because it had a lower concentration of solvent that needed to be evaporated from the laminate.

PAGE 22

1 6 Figure 7 Solvent and epoxy solution at various ratios showing separation of solvent and epoxy at higher concentrations of solvent With a ratio for the solution chosen, a study to determine the efficacy of this method was completed. The same three ply laminate with inter lamina defect created with the use of PTFE strips used in previous experiments w ere used to test the solution repair method. The solution mixture was warmed to approximately 100 ° C to soften the solution to be able to work with it better. The softened solution was then applied to the void area of the defect specimens with a small flat metal tool. The specimens were then placed in a vacuum bag for range of times. The correct time for the repair process needed to be determined to find the maximum recovery strength of the laminate. The times chosen were 2, 4, 6, 8, and 10 hours. The speci mens were then tested under three point bend to determine the strength and stiffness of the specimen. The results of the tests show an average increase in strength and stiffness as the time to re cure is increased. At a max imum time of 10 hours to repair t he defect, the specimen on average recovered 105% stiffness and 94% strength. This result is to be expected because the process physically adds more epoxy that covalently bonds to the defect areas, pulling them back together and forming a cohesive material . However this experiments did have some anomalies that could not be explained such as a 6 hour specimen on average recovering more stiffness

PAGE 23

17 than an 8 hour specimen. The 10 hour specimens were also slightly thinner than the others as well as the control s pecimen which means that these results cannot be compared without further research. Overall the method proved to be an effective method for repair and warrants further research. One downside of epoxy injection is the added weight associated with the addit ional epoxy. The solution injection method explored is also at risk of the same downside but the results of the tests show an increase in weight of each specimen of around 6.5%. This small increase in weight is a good result considering that half of the sp ecimen was being injected with the solution. CHAPTER III PARAMETRIC S TUDY S IMULATING D ELAMINATED C OMPOSITE R EPAIR A parametric study of repair conditions for a BER composite , with inter lamina defect , was completed . The effects of pressure and time durin g the repair process were investigated . This study revealed the optimum pressure and time to repa ir a delaminated BER composite. These optimum pressure and tim e of the repair was combined into one method of repair. This method is shown to be the optima l re pair method for dry repair. This optimally repaired specimen is then compared to a contro l set of specimens . The dry method is shown to be as effective as current repair methods utilizing current engineering epoxies . The purpose of this study is to determi ne the efficacy of composite repair utilizing BER epoxy by completing a parametric study. ME THODS Material s Epoxy synthesis materials such as metal catalyst Zn(Ac) 2 , diglycidyl ether of bisphenol A (DGEBA), and fatty acid Pripol 1040 were supplied b y Sigm a Aldrich Inc. A 3K plain weave carbon fiber was ordered from Fibre Glast Development Corporation. The material has a thickness of 0.305 mm and a weight of 198 g/m 2 . High temp release film from Fibre Glast Development Corporation,

PAGE 24

18 item number 1782. N ylon r elease peel ply from Fibre Glast Development Corporation, item number 582. Breather fabric from Plasticare. Epoxy Formulation The epoxy was synthesized utilizing the same techniques outlined by Yu . [15] Fatty acid was mixed with metal catalyst at a ratio of 10 to 0 .89 grams. The mixture was then placed in a vacuum oven at a temperature of 140 ° C until no bubbles from the mixture were observed(approximately 3 hours). DGEBA was then added to the mixture at a ratio of 10 grams fatty acid to 5.67 grams DGEBA. This was thoroughly mixed together for later us e in pre preg manufacturing. Pre preg Manufacturing Specimens were crea ted using pre preg carbon fiber that was manufacture d as follows. BER epoxy was heated to approximately 100 ° C to soften the epoxy. Woven neat carbon fiber was laid flat on plastic sheet. The warm epoxy was then poured over the carbon fiber , a squeegee was used to spread the epoxy evenly. Th e carbon fiber was flipped over and the process repeated in the other side of the fiber to completely and evenly coat and impregnate the carbon fiber with epoxy. The pre preg was wrapped in plastic sheet and placed flat in a freezer for later use (s ee Figu re 8 ) . The pre preg had an initial mass fraction ratio of 1:1 carbon fiber to epoxy . Specimens post cure had a mass fraction of carbon fiber of 69 74% fiber mass.

PAGE 25

19 Figure 8 Manufacturing steps for making pre preg carbon fiber/BE R epoxy A) neat carbon fiber B) neat carbon fiber with epoxy C) neat carbon fiber with epoxy covered with plastic sheeting D) epoxy spread evenly through carbon fiber Composite Specimen Manufacturing Three pre preg sheets of identical dimensions were cut to later create the laminate. The sheets were laid on top of one another, three layers thick. To create a specimen with inter lamina defects, a 30 mm wide strip of PTF E was placed across the width of the composite between the first and second sheet as well as the second and third sheet. The PTFE strips were laid in line with each other to create a uniform defect through the middle of the width of the specimen. This was then cured using vacuum bagging technique detailed below. The composite containing both d efect and non defect areas were cut into individual specimens to be tested. These cured composites samples were cut into specimens approximately 60 mm x 14 mm. The diagram in Figure 9 shows how the PTFE strips were placed in the laminate before the cur e pr ocess. The photo in Figure 10 show what the laminate looks like after the cure process , before it is cut into individual specimens .

PAGE 26

20 Figure 9 Diagram of the manufacturing process of the defect specimens

PAGE 27

21 Figure 10 Image of defect composit e laminate with PTFE strips on n ylon fabric Vacuum Bagging Process The composite is placed between two nylon peel ply to ensure epoxy distribution throughout the surface of the composite. The sample is then placed betwe en two non perforated release films. The film is non perforated to keep as much epoxy within the specimen as possible. The sample is then placed on a flat metal plate to create a flat specimen, which is then covered with a layer of breather fabric and plac ed in the vacuum bag (see Figure 11 ) .

PAGE 28

22 Figure 11 Picture of laminate in vacuum bag before cure Dry Repair Process The repair process utilizes pressure and heat to complete a repair. To achieve a constant temperature and pressure a n MTS machine equipped with a thermal chamber and compression platens was used (see Figure 12) . Three s pecimens with inter lamina defects were stacked directly on top of one another with a thin sheet of PTFE separating the specimens to prevent contact betwe en specimens. The method of stacking specimens was used to assure that all the specimens received the same pressure and temperature . The curing temperature was 180°C. The pressures applied to the specimens were 500, 750, 1000, 1250, and 1500 kPa. Time dura tion of each of these repairs was one hour. Only the pressure of the repair was alt ered. For tests that would examine the effect of time duration of the repair a constant pressure of 500 kPa was applied to each varia tion of time duration. The time of repai rs were 1, 2, 4, and 6 hours.

PAGE 29

23 Figure 12 Picture of compression platens in oven chamber utilized in the dry method repair Testing Procedure The composite specimens wer e tested in three point bending on an MTS 858 Mini Bionix II m achine . A 100 N load cell was used. L oad and displacement of the crosshead were measured. The crosshead rate was set to 1 mm/min as designated by ASTM standard D7264. Per ASTM section 8.2 a support span to thickness ratio of 60:1 was chosen. The average sp ecimen thickness was 0.635 mm . The span length was set to 38 mm to achieve the 60:1 ratio desired (see Figure 13) . The test was run

PAGE 30

24 until a crosshead extension of 15 mm was completed . The reasoning for this exact extension limit was to achieve enough data t o calculate both stiffness max imum stress of the specimen. Figure 13 Image of MTS 858 mini bionix ii with three poi nt b end test fixture with specimen after fracture Failure Mode Analysis Failure analysis of the specimens were c onducted to evaluate the effectiveness of the repair procedure. The failure modes were analyzed using the format in the ASTM standard used for testing and data analysis [1] . Each specimen was analyzed during and after testing t o determine the failure mode. The mode of failure was determined by three characteristics shown in Figure 1 4 .

PAGE 31

25 Figure 14 Flexure Test Specimen Three Part Failure Identification Code [1] Data analysis Each sp ecimen is analyzed by calculating the strength and stiffness. Stiffness is calculated by dete rmining the slope of the stress ( strain ) curve between 0 .001 and 0 .003 strain. Strength of the specimen is calculated by determining the max imum stress expe rienced by the specimen. Stress experienced by the specimen is calculated using eq.1 . [17] Max imum strain experienced by the specimen at mid span is calculated using eq. 2 . [17] where: = stress at the outer surfac e mid span , MPa P = applied force, N L = support span length, mm b = specimen width, mm h = s pecimen thickness, mm = strain at outer surface, mm/mm = mid span deflection, mm

PAGE 32

26 RESULTS The strength and stiffness for each specimen were cataloged . The three specimens for each test condition were compared and standard deviation for stiffness and st rength calculated . The average stiffness and strength along with standard deviation are shown in Table 1. The average stiffness and strength for each test condition were compared to other variations of either pressure of time duration. Table 1 Average stiffness and strength of each sample set Average Stiffness (MPa) Std. dev (MPa) Average Strength (MPa) Std. dev (MPa) Time Tests (hours) 1 16265.51 5323.62 87.32 25.57 2 20944.30 154.26 130.16 17.19 4 28439.40 79.48 131. 44 13.08 6 25380.39 3041.60 127.54 4.63 Pressure Tests (kPa) 500 16265.51 5323.62 87.32 25.57 750 19935.73 1330.53 87.04 9.07 1000 25030.82 2559.62 98.89 13.09 1250 26059.21 2994.36 126.11 2.52 1500 24172.94 2273.47 124.04 11.59 Applied Pr essure of Repair The average stiffness and strength of each variation of applied pressure during repair are plotted as a function of pressure in Figures 1 5 and 1 6 respectively. A verage stiffness of the specimen increases as applied pressure is increased, u ntil 1250 kPa. The average stiffness of repaired specimens decreases at a pressure s higher than 1250 kPa . Specimens repaired at a pressure of 1500 kPa are lower than that of 1000 kPa or 1250 kPa. While average stiffness of 1000, 1250, and 1500 kPa specimen s are different; a t test showed that the values are not statistically different.

PAGE 33

27 Figure 15 Plot of stiffness of repaired specimens for each variation in applied pressure The average strength of the specimens in Figure 1 6 i ncreases to a maximum value at 1250 kPa . The strength looks as though it is going to continue increasing , but decreases when a pressure of 1500 kPa is used for repair . The decrease seen in F igure 15 is also seen in Figure 1 6 at a pressure of 1500 kPa. The maximum average stiffness and strength of a repaired specimen occurred at an applied pressure of 1250 kPa. 0 5000 10000 15000 20000 25000 30000 35000 250 500 750 1000 1250 1500 1750 Stiffness (MPa) Pressure Applied (kPa)

PAGE 34

28 Figure 16 Plot of strength of repaired specimens for eac h variation in applied pressure Time Duration of Repair The av erage stiffness and strength of each variation of duration of repair are plotted as a function of time in Figures 1 7 and 18 respectively . The average stiffness of specimen increase s as the duration of the repair is increased. Average stiffness increases li nearly , then peaks at four hours , and then decreases at six hours . In Figure 1 7 , the average strength of specimens for each dataset increase s as the duration of the repair is increased , then plateaus after four hours of repair until the six hour mark . At s ix hours of repair the average strength of the specimens slightly decrease when compared to the four hour mark. The average streng th of each data set in Figure 1 7 is shown to increase in a 0 20 40 60 80 100 120 140 160 250 500 750 1000 1250 1500 1750 Strength (MPa) Pressure Applied (kPa)

PAGE 35

29 slightly linear pattern and then decrease linearly as well . The max imum average stiffness and strength of a repaired specimen occurs at 6 hours of repair duration. Figure 17 Plot of stiffness of repaired specimens for each variation in time duration of repair 0 5000 10000 15000 20000 25000 30000 0 1 2 3 4 5 6 7 Stiffness (MPa) Duration of Repair (hours)

PAGE 36

30 Figure 18 Plot of strength of repaired specimens for each variation in time duration of repair F ailure Modes of Pressure and Time Tests The examination of the failure mode of each specimen is critical to understanding the effectiveness of the repair parameters. T he failure modes of each specimen in the pressure and time tests are shown in Table 2 3. The tables are filled out using the characterization guide set forth in the ASTM standard and Figure 1 4 . The specimens either failed due to inter laminar shear , bucklin g of the top lamina, or a combination of both of these. Figure 20 shows an SEM image of buckling fracture of a specimen repaired at 1000 kPa. 0 20 40 60 80 100 120 140 160 0 1 2 3 4 5 6 7 Strength (MPa) Duration of Repair (hours)

PAGE 37

31 Table 2 Failure modes of each specimen of the pressure study Failure Mode Pressure (kPa ) A B C 500 SLM SLB SLM 750 M(S,B)LM M(S,B)LM SLM 1000 BST OST BST 1250 BST BST BST 1500 BST BST SLM Table 3 Failure Modes of each specimen of the time study Failure mode Time (hours) A B C 1 SLM SLB SLM 2 BST SLM BST 4 B ST BST M(S,B)LT 6 BST BLT BST Analysis of the table s show that a t pressures less than 1000 kPa the specimen would delaminate first at the edges of the repaired defect and grow in length toward the center of the specimen. The same failure mode also occur red on time specimens that were repaired for less than 4 hours. Specimens that experienced pressures greater than or equal to 1000 kPa during repair would remain rigid and only flex then fracture under buckling , although one 1500 kPa specimen failed due to interlaminar shear and delamination . The modes of failure can be seen in Figure 19.

PAGE 38

32 Figure 19 Failure m odes of A ) buckling fracture and B ) delamination Figure 20 SEM Image of specimen repaired at 1000 kPa for one hour, showing buckling of lamina (Photo provided b y May Martin)

PAGE 39

33 Optimal Pressure and Time Duration of Repair With these results the a sample set of specimens repaired for 4 hours with an applied pressure of 1250 kPa were created and tested . A representative optimal repair specimen and representative cont rol specimen are plotted on a stress as a function of strain chart in Figure 21 . The average stiffness and strength of the optimal repair specimen are 69.4 % and 85.8 % of the average c ontrol specimens respectively. A second set of specimens was also repaire d under the same conditions and tested. The average stiffness and strength of the optimal repair specimen are 74 % and 86 % of the average control specimens respectively . Figure 21 Stress(strain) plot of representative samples of a control and optimal repair specimen 0 20 40 60 80 100 120 140 160 0 0.005 0.01 0.015 0.02 0.025 0.03 Stress (MPa) Strain (mm/mm) 4 hours at 1250 kPa Control

PAGE 40

34 DISCUSSION This study showed that pressure and time do have a measureable effect on the epoxy and therefore the repair method process. The time study revealed a m aximum stiffnes s and strength occurred at time of four hours and a pressure of 500 kPa , resulting in a recovery of stiffness and strength of 99% and 95% respectively. This result is a s expected because more bonds will be reconnected as a function of time. The decrease in stiffness and strength in specimen s rep aired for six hours is likely due to epoxy degradation. A temperature of 180°C is orders of magnitude the 30 °C T g of the epoxy. This high heat for an extended period of time is likely to cause damage and degradation to the epoxy structure. This plateau whe re the maximum number of bonds are reconnected without epoxy degradation should be explored in further studies. The pressure study revealed a m aximum stiffness at 1250 kPa for a duration of one hour, instead of the expected 1500 kPa. This surprising result co uld be explained by epoxy fractures (see Figure 2 2 ) . It is likely that above a certain experienced pressure the epoxy begins to fracture under this load. The resulting fractured epoxy creates more unwanted defects and therefore does not behave in the sam e manner as it d id at a lower repair pressure. An optimal set of parameters for repair was determined to be a n applied pressure of 1250 kPa for a duration of 4 hours , although further tests would reveal that this was not an optimal set of parameters . The resulting repair method was show to be an in effective method of repair resulting in a recovery of 69.4% stiffn ess and 85.8 % strength. Repeated tests showed similar results. The parameters used for this optimal repair were the cause for the sharp drop in recoverable stiffness and strength. The assumption before testing began was that the optimal pressure and time w ould create the optimal repair, but this study did not show that. Instead the high pressure of 1250 kPa and extended period of time of four hours likely contributed to damaging the epoxy. It is shown that high pressures damage the epoxy by fracture in Figu re 22. Extended durations repair are also shown to decrease the recoverable streng th and stiffness of a specimen, likely due to epoxy degradation. The

PAGE 41

35 results of this study revealed a relationship between pressure of repair and duration of repair. Further research into this relationship is necessary to fully understand the ideal parameter s for composite repair. Figure 22 SEM Picture of specimen repaired at 1500 kPa, showing epoxy fractures (Photo provided by May Martin NIST ) CONCLUS ION The results of this research indicates that BER epoxy is a viable composite epoxy that allows for repairs. With a recoverability of nearly 100% stiffness and strength , a component can be repaired without complicated machining or chemical processes . The method of repair dictated in this paper is shown to be a viable and possibly more efficient repair method when compared to present methods such as epoxy injection . This method can be easily exploited for use in the field or even as a manufacturing method . This repair process can be used for quick turnaround of damaged components

PAGE 42

36 or pa rts on vehicles and structures. To further corroborate these findings an epoxy injection repair process should be explored in future works. One disadvantage of this particular BER epoxy is its low T g of 30 °C . Future research into BER epoxies should examine those BER epoxies with a higher T g . BER epoxy allows for the possibility to repair surface and subsurface defects with components in place on structures and vehicles in the field.

PAGE 43

37 REFERENCES [1] A. S. Tim Hepher. (2013). The Dreamliner Debacle Has Already Cost Boeing $600 Million . Available: ht tp://www.businessinsider.com/dreamliner trouble has cost boeing 600 million 2013 4 [2] T. N. Y. Times. (2013). The Jump to a Composite Plane . Available: http://www.nytimes.com/interactive/2013/07/29/business/The Jump to a Composite Plane.html [3] R. ZumMallen. (2018). 2019 GMC Sierra 1500 Adds Carbon Fiber Box . Available: https: //www.trucks.com/2018/03/01/2019 gmc sierra 1500/ [4] D. White. (2005). New trucks from Toyota and Honda get it Right with composite beds . Available: https://www.compositesworld.com/columns/new trucks from toyota and honda get it right --with composite beds [5] J. Hirsch. (2018). Popularity of Pickup Trucks Drive 2017 Auto Sales . Available: https://www.trucks.com/2018/01/04/2017 pickup trucks auto sales/ [6] M. T. Amy Bauer, Kristine Obusek, Mufit Akinc, "Bisphenol E cyanate ester as a novel resin for repairing BMI/carbon," Elsevieer, no. Polymer, 2013. [7] E. S. Kristin Kuchler, Rolf Dieter Hund, Olaf Diestel, Martin Kirsten, Chokri Cherif, "Local repair procedure for carbon fiber reinforced plastics by refilling with thermoset matrix," Journal of Applied Polymer Science, 2013. [8] H. K. Paolo Feraboli, Bonnie Wade, Federico Gasco, Luciano DeOto and Attilio Masini, "Recyclability and reutilization of carbon fiber facbric/epoxy composites " Journal of Composite Materials 2011. [9] L. F. M. D. S. K.B. Katnam, T.M.Young, "Bonded repair of composi te aircraft structures: A review of scientific challenges and opportunities," Elsevier, no. Progress in Aerospace Sciences, pp. 26 42, 2013.

PAGE 44

38 [10] H. Bendemra, P. Compston, and P. J. Crothers, "Optimisation study of tapered scarf and stepped lap joints in composite repair patches," Composite Structures, vol. 130, pp. 1 8, 2015/10/15/ 2015. [11] F. Benyahia, A. Albedah, and B. Bachir Bouiadjra, "Analysis of the adhesive damage for different patch shapes in bonded composite repair of aircraft structures," Materials & Design (1980 2015), vol. 54, pp. 18 24, 2014/02/01/ 2014. [12] M. Kashfuddoja and M. Ramji, "Design of optimum patch shape and size for bonded repair on damaged Carbon fibre reinforced polymer panels," Materials & Design (1980 2015), vol. 54, pp. 174 183, 2014/02/01/ 2014. [13] R. F. E. H. Seyedmohammad S. Shams, "Overlay patch repair of scratch damage in carbon fiber/epoxy laminated composites," Elsevier, no. Composites: Part A, pp. 148 156, 2013. [14] F. Ning, W. Cong, J. Qiu, J. Wei, an d S. Wang, "Additive manufacturing of carbon fiber reinforced thermoplastic composites using fused deposition modeling," Composites Part B: Engineering, vol. 80, pp. 369 378, 2015/10/01/ 2015. [15] Q. S. Kai Yu, Martin L. Dunn, Tiejun Wang, and H. Jerry Qi, "Carbon Fiber Reinforced Thermoset Composite with Near 100% Recyclability," Advenced Functional Materials, 2016. [16] Q. Shi, K. Yu, M. L. Dunn, T. Wang, and H. J. Qi, "Solvent Assisted Pressure Free Surface Welding and Reprocessing of Malleable Epoxy Polymers," Macromolecules, vol. 49, no. 15, pp. 5527 5537, 2016/08/09 2016. [17] "D7264 Standard Test Method for Fl exural Properties of Polymer matrix Composite Materials," 2017.