- Email: [email protected]
Methodology for optimization of the curing cycle of paste adhesives
International Journal of Adhesion & Adhesives 40 (2013) 112–119
Contents lists available at SciVerse ScienceDirect
International Journal of Adhesion & Adhesives journal homepage: www.elsevier.com/locate/ijadhadh
Methodology for optimization of the curing cycle of paste adhesives A. Sa´nchez Cebria´n n, M. Zogg, P. Ermanni Centre of Structure Technologies, ETH Z¨ urich, Leonhardstrasse 27, 8092 Z¨ urich, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 1 September 2012 Available online 11 September 2012
This contribution, carried out in the frame of the European Joint Technology Initiative ‘Clean Sky’, presents the results of a research program investigating the influence of fast curing on the quality of epoxy based paste adhesives. Today, the curing of paste adhesives is typically carried out following the supplier’s recommendations. In order to reduce cycle time, save costs and energy resources, paste adhesives could be cured at higher temperature. To ensure a bonded joint quality with maximum mechanical performance, the limitation of this temperature increase is studied. This study shows the effects of the use of high temperatures in the curing process, which can lead to a degradation of the adhesive system due to the increase of void content, decreasing the mechanical performance in the paste adhesive as well as in the bonded joint. The goal of this research is to find fast and robust processing of paste adhesives and to develop a methodology to determine the maximum curing temperature possible. Different properties of the adhesive are investigated, including different thermal analysis techniques, optical and mechanical testing of the pure adhesive. Additionally, state of the art qualification of paste adhesives, single lap shear testing, is considered. In this study, a novel method to control the quality of the cured paste adhesives is defined based on the analysis of the pure cured paste adhesive, not influenced by the adherent quality, by measuring the void content and its effects on the bonded joint. & 2012 Elsevier Ltd. All rights reserved.
Keywords: Paste adhesive Non-destructive testing Cure Certification
1. Introduction In this research contribution, a methodology to control the quality of paste adhesives with a high temperature curing process is investigated. The background of this research is to accelerate the curing cycle of the paste adhesive by increasing the temperature of the process without affecting the mechanical performance of the joint. The paste adhesive system used is the LMB 6687-1/LME 10049-3 from Huntsman Advanced Materials. The curing process is performed without pressure [1]. Samples are cured under different curing cycles, obtaining products with different properties that are analyzed. An acceleration of the supplier’s recommended curing cycle is possible without decreasing mechanical performance. For industrial applications, acceleration of the curing process is usually not applied as the mechanical performance of the joint can be affected if too high temperatures are applied [2]. For this reason, effects of an accelerated curing process are not considered [3]. The goal of this research is to investigate a methodology to control the quality of the paste adhesive when high temperatures are used in the accelerated curing process. This quality control is established through the
n
Corresponding author. Tel.: þ41 44 632 71 72; fax: þ 41 44 633 11 25. E-mail address: [email protected] (A. Sa´nchez Cebria´n).
0143-7496/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijadhadh.2012.09.002
analysis of different properties of the adhesive in a range of samples with different degrees of degradation. One of the main indicators of degradation of paste adhesives is an increase of the void content. An increased void content leads to a reduction of the mechanical performance of the paste adhesive. During the mixing process some air is entrapped [4]. If the paste adhesive is cured at higher temperature, the air inside expands, increasing the volume of the voids thus affecting the mechanical properties of the joints [5,6]. The quality of adhesives is mechanically tested by strength tests such as shear and peel tests [7]. Non-Destructive Inspection (NDI) techniques such as ultrasonic testing are also used to detect defects such as voids [8]. Maximum void content can be found in literature for composite panels and limitations about positioning of voids in the edge in order to avoid delamination problems [9]. However, there are no explicit considerations for bonding systems [10]. The detection of void content in paste adhesives is limited by the minimum defect size detected by NDI techniques. If the voids are small enough they may not be detected by NDI [11]. Today, there are no reliable methods to assess quality of a bonded joint [12]. This contribution focuses on the experimental analysis of this effect, observing how the increment of curing temperature affects the void content and how this increment of porosity affects different mechanical properties of the paste adhesive, and to find a methodology to assess the quality of the adhesive.
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
The samples produced in this research are completely cured applying different curing cycles, with temperatures from 80 1C (as recommended by the supplier) to 200 1C (which causes a clear thermal degradation on the adhesive). Physical and mechanical properties of these samples are then measured. The tests considered in this study include thermal analysis techniques, e.g. Differential Scanning Calorimetry (DSC) and Thermogravimetric analysis (TGA), as well as optical microscopy to observe the evolution of different physical properties of cured epoxy paste adhesive when the curing temperature is changed. Additionally, mechanical testing is considered by the measurement of flexural properties of the pure adhesive in three point bending test and Dynamic mechanical analysis (DMA) and single lap shear test with Carbon fiber Reinforced Polymers (CFRP) bonded systems. All the testing is firstly considered by separate in order to analyze the benefits and inconveniences of each test. From the information obtained from each test, recommendations for suitable qualification of paste adhesives are defined. The acceleration of the curing process of a paste adhesive is considered and a methodology to control the quality considering thermal degradation is established
2. Curing kinetics modeling theory Firstly, the curing cycles used to produce the samples are defined. The reference curing cycle from the supplier is composed of a gelling stage of 2 h at room temperature, followed by a curing process at 80 1C for 4 h. The rest of the curing profiles chosen have shorter cycle times but higher temperatures with a risk of thermal degradation. The definition of the curing temperatures is carried out by using the kinetics model of the paste adhesive, which gives the relation between temperature, time and curing degree for this paste adhesive system. The theoretical model of the chemical reaction is measured following Arrhenius relation, based on the n-th order kinetics [13], in the Eq. (1) f ðaÞ ¼ ð1aÞn
ð1Þ
This reaction can be written in the following form by combining an Arrhenius relation [14] da E ¼ k0 exp ð1aÞn ð2Þ RT dt The model depends on three variables: the degree of cure a [-], temperature T [K] and time t [s], and three parameters which characterize the reaction: k0 [-], which is the specific rate
113
constant at temperature T [K], the activation energy E [KJ/mol] and the reaction order n [-]. Once the parameters of the chemical reaction are defined, it is possible to predict which will be the degree of cure for a certain temperature applied for a given time. The equation with the degree of cure depending on temperature and time is shown in Eq. (3).
a ¼ 1 1ð1nÞ z t exp
1 1n E RT
ð3Þ
Eq. (3) can be solved by using multiple linear regression of the general form:z ¼ a þbx þ cy where the two basic parameters on Eq. (2), the reaction rate ddta [1/s], and the curing degree a, are determined from the DSC measurements. Three parameters (z, which is the Arrhenius frequency factor [1/s], E and n) are calculated by the software of the DSC, Pyris. Material obeying n-th order has the maximum rate of heat evolution at the beginning of the reaction [15]. This theoretical modeling for the kinetics characterization is carried out by a single dynamic heating measurement in the DSC of a non-cured sample of the paste adhesive. To minimize errors, this measurement is done 8 times and the average is used to measure the relation between temperature, time and degree of cure. Results are shown in Fig. 1. The curing times are chosen following the theoretic model, having always more than 90% theoretical degree of cure. It must be considered that samples will have an extra gelling stage not considered in the theoretical model. Therefore, the real degree of cure will be slightly higher than the theoretical one. This should reach more than 95%, as recommended by supplier. The curing profiles for the samples are summarized in Table 1.
Table 1 Summary of curing conditions for tested samples. Temperature (1C)
Time (min)
Theoretical curing degree (%)
80 100 120 140 160 180 200
240 60 60 45 30 15 10
93.2 91.3 95.6 97.1 97.8 97.8 98.2
Fig. 1. Kinetics of curing reaction of LMB 6687-1/LME 10049-3 from Huntsman Advanced Materials.
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
114
3. Material and methods Preparation of the samples is described in previous work [16], where some of the tests carried out are detailed. The experiments performed can be divided into physical and mechanical testing. Experiments carried out to study the change in physical properties include DSC and TGA analysis as well as the analysis of size and quantity of voids, carried out with microscopy techniques. With DSC analysis, the degree of cure of the paste adhesive is determined in order to validate the curing process carried out in the different samples analyzed in this study. DSC is used to validate if the measured degree of cure in the samples is higher than 95%. TGA analysis is carried out to measure the degradation of the components of the paste adhesive. The equipment in this case heats up a sample until high temperatures, measuring the weight during the process. In this experiment, the evaporation of particles due to the application of high temperatures is measured and the temperature limit can be assessed. This test is carried out three times. Components are analyzed by separate to observe where the degradation is starting and to observe the behavior of both components by separate as well as the mixed component. Optical microscopy techniques are used to measure the increment of void content in the paste adhesive when it is heated with a higher temperature. Void content of the samples as well as the average diameter and number of voids are measured by microscopy techniques (Leica DM RXA). Ten images are taken from different areas of each sample and they are post-processed with the software (Leica QWin), measuring the void content and the diameter of ten different voids for every image. The void content observed by optical testing is validated with the measurement of the density of all the samples used in the different mechanical tests carried out; DMA analysis, 3 point bending test of pure adhesive and single lap shear of bonded samples. DMA analysis measures dynamic mechanical properties of the material and its evolution in a temperature range; in this case, using a 3-point bending test. These results are expected to provide similar information on the mechanical performance from mechanical testing e.g. three point bending test on the standard testing machine, confirming the effect of degradation on mechanical properties of the adhesive. Additionally, the DMA analysis provides information about the effect of a higher curing temperature on the glass transition temperature (Tg) of the adhesive. Mechanical testing of the pure adhesive is carried out following ISO 178 for measurement of flexural properties of plastics in 3-point bending, on a standard testing machine Zwick 1474. Flexural modulus and flexural strength are measured in order to confirm the tendency measured in the DMA with smaller samples and to show the influence of the adherents, as used in single lap shear test. Finally, a single lap shear test, which is typically used as the state of the art qualification of bonded joints, is considered. Out of autoclave CFRP prepreg system (resin system MTM 44-1s, fibers Sigmatex CF 5804A) from Umeco, is used to manufacture the
adherents. Laminate plates are manufactured following the recommendations from the supplier. Surfaces of the adherents are treated by a manual sanding process (sand size P100); which has been previously demonstrated to improve mechanical performance of the joint [17]. The plates are also thoroughly cleaned first with acetone, then with tap water and finally with de-ionized water. Then, the CFRP plates are dried in a forced convection oven for 2 h at 65 1C. Once the surface treatment is complete, the bonding process is carried out by applying the adhesive and curing with the different curing cycles defined in 2. Finally, the samples are tested in simple lap shear test (SLS) following EN 2243-1 for structural adhesives. A summary of all the information necessary to prepare samples and carry out the experiments is summarized in Table 2.
4. Results and discussion 4.1. Analysis of physical properties Table 3 summarizes the degree of cure of the different samples measured with the DSC, considering the measured total released energy of the curing reaction, equal to 320 J/g. As it is shown, all the samples are cured more than 95% as expected. Therefore, the curing process can be considered complete for all the samples. TGA analysis is also considered. Results for separate components can be seen in Fig. 2. TGA shows the onset of the hardener at 124 1C and in the resin at 191 1C, but a loss of mass can be observed in the hardener before 120 1C, meaning that samples heated with this temperature or greater will have a higher percentage of voids. This effect is combined with the expansion of the voids trapped in the dosing and mixing process resulting that samples heated with temperatures higher than 100 1C will have more and larger voids which will affect the mechanical performance. Microscopy is used to observe voids inside of the paste adhesive and to measure the void content inside the frame that appears in all the images. Figs. 3–6 show examples of each sample measured. Void content of all the samples are summarized in Fig. 7.
Table 3 Curing degree of samples measured by DSC. Temperature (1C)
Time (min)
Released energy (J/g)
Curing degree (%)
80 100 120 140 160 180 200
240 60 60 45 30 15 10
10.5 7.6 7.6 8.4 11.4 12.5 3.1
96.7 97.6 97.6 97.4 96.4 96.1 99.0
Table 2 Summary of experimental parameters. Experiment
Samples
Sample size/weight
Initial conditions
Final conditions
Heating rate (K/min)
DSC DMA
1 1
1 min at 0 1C 5 min at 0 1C
1 min at 300 1C 5 min at 120 1C
10 3
TGA 3-pt bending SLS Optical testing
1 6 5 10
1–5 mg Rectangular 18 mm*8 mm*2 mm Force 110 mN7 100 mN (1 Hz) 5–15 mg Rectangular 40 mm*25 mm*2 mm 100 mm*25 mm*1 mmOverlap: 12.5 mm Rectangular 20 mm*15 mm*2 mm
1 min at 40 1C – – –
1 min at 250 1C – – –
1 – – –
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
Fig. 2. TGA analysis for epoxy and hardener.
Fig. 3. Samples cured at 80 1C (Avg. void content 1.6%) and 100 1C (Avg. void content 1.4%).
Fig. 4. Samples cured at 120 1C (Avg. void content 2.1%) and 140 1C (Avg. void content 21.4%).
Fig. 5. Samples cured at 160 1C (Avg. void content 33.5%) and 180 1C (Avg. void content 60.5%).
115
116
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
Fig. 9. Average diameter values of voids measured in samples cured with different temperatures. Fig. 6. Samples cured at 200 1C (Avg. void content 75.1%).
Fig. 10. Relative density measured in samples for mechanical testing. Fig. 7. Porosity values for samples cured with different temperatures. Table 4 Summary of optical measurements. Temperature (1C)
Time (min)
Average voids diameter (mm)
Average number of voids (-)
Void content (%)
Density (%)
80 100 120 140 160 180 200
240 60 60 45 30 15 10
49 720 60 730 100 730 230 780 260 7130 340 7240 270 7180
8.4 73 9.3 73 11 72 17 74 – – –
1.67 0.5 1.47 0.5 27 1 217 6 347 12 617 17 757 12
97.6 98.7 95.1 80.7 61.2 41.0 25.1
Fig. 8. Void content values for samples cured between 80 and 120 1C.
In order to observe better the values of void content of samples cured at lower temperatures, between 80 1C and 120 1C, Fig. 8 is shown. The results show a low void content for the samples cured at 120 1C or less when compared with the rest of the samples. However, an increase of void content is observed in samples cured at 120 1C. These results can be compared with TGA analysis, where it can be observed that when the hardener is heated with more than 100 1C, a loss of mass effect appears producing the evaporation of particles in the hardener. This fact can explain the big difference in the porosity levels between 120 1C and 140 1C, when evaporation effect is higher. Comparing the images in Fig. 5, it can be observed that not only do the voids increase the size, but also the quantity increases due to the evaporation of the component. The average diameter of the voids is also determined by measuring ten different voids for each sample. The density of the samples used in mechanical testing is also measured. Results are shown in Figs. 9 and 10 and summarized in Table 4. The results show a similar tendency as the size of the voids, having a clear increase of diameter for samples cured with more
than 100 1C. The number of voids increases slightly in the samples cured with higher temperatures from 80 1C to 120 1C and increases strongly for samples cured at 140 1C. As it can also be observed, the void content values are approximately inverted values of density, meaning a homogeneous distribution of the voids in the different samples. 4.2. Mechanical testing A 3-point bending test is carried out in the DMA for the different samples measuring the storage modulus from 0 1C to 120 1C. Curves of all the samples are shown in Fig. 11. It is observed a lower performance in samples cured with higher temperature. In order to evaluate results and compare them for all the samples, storage modulus at room temperature (20 1C) as well as the Tg are measured and shown in Figs. 11 and 12 and summarized in Table 5. It can be observed that the storage modulus at 20 1C decreases within the samples cured at higher temperature. The Tg remains constant. The increase in performance of the sample cured at 200 1C compared with the sample cured at 180 1C can be explained by the high degree of degradation of both samples.
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
117
Fig. 11. Storage modulus measured with DMA in a range from 0 1C to 120 1C.
Fig. 12. Storage modulus from DMA at 20 1C.
Fig. 14. Flexural strength measured in samples cured with different temperatures.
Table 5 Storage modulus at 20 1C and Tg values from DMA.
Table 6 Summary of results in 3 point bending test.
Temperature (1C)
Time (min)
Storage mod. at 20 1C (Pa)
Tg (1C)
80 100 120 140 160 180 200
240 60 60 45 30 15 10
4.4Eþ 08 3.4Eþ 08 2.9Eþ 08 2Eþ 08 1Eþ 08 4Eþ 07 1Eþ 08
112 113 117 120 122 110 120
Fig. 13. Flexural modulus measured in samples cured with different temperatures.
Temperature (1C) Time (min) Flex. modulus (MPa) Flexural strength (MPa) 80 100 120 140 160 180 200
240 60 60 45 30 15 10
1100 7120 1190 7110 1060 775 650 7120 475 7170 195 7110 115 720
417 4 447 2 417 2 257 6 157 6 67 3 37 1
Results of -point bending tests are shown in Figs. 13 and 14 and then summarized in Table 6. The results of the 3-point bending test show that the mechanical performance decreases with curing at high temperatures. A decrease of performance can be observed for samples cured with more than 120 1C. The tendency is similar to the DMA results, but the results of ISO 178 show high values for samples cured with 120 1C or less. Finally, a SLS test is carried out under EN 2243-1 for structural adhesives using CFRP adherents. Results and fracture modes are characterized and summarized in Fig. 15 and Table 7. Results show a decrease of shear strength when the curing temperature is increased having an adhesive failure mode in samples that are highly degraded. Samples cured with 80 1C and 100 1C show adherent failure, meaning that the adhesive is stronger than the
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
118
Fig. 15. SLS test for CFRP bonded samples.
Table 7 Results of single lap shear test. Temperature (1C)
Time (min)
Shear strength (MPa)
Fracture mode
80 100 120 140 160 180 200
240 60 60 45 30 15 10
24 72.9 22 72 16.3 71 15.5 72 6.0 71 6.1 70.5 6.6 71
Adherent Adherent Cohesive Cohesive Adhesive Adhesive Adhesive
adherent and the quality of the paste adhesive is well cured. Samples with cohesive failure show that the adherence of the paste adhesive is correct, but that the mechanical properties of the joint are lower than with adherent failure, meaning a certain degradation of the adhesive for samples cured with 120 1C or more. Finally, samples with adhesive failure show a poor adhesion between parts meaning that curing process is not correct due to the high temperatures applied. In this study, lower values are obtained due to the low quality of the adherent, which usually delaminates at about 20–25 MPa. Samples cured at 100 1C or less show a high shear performance until the fracture of the adherent.
4.3. Methodology to minimize curing time Today in industry the curing process of an adhesive is carried out following recommendations from the supplier, which are not optimized for energy consumption. By increasing the curing temperature, which is typically held constant during the process, an acceleration of the curing process is possible. The limitation of this temperature comes from the void formation when high temperatures are applied to the adhesive, which influences the quality of the joint reducing its mechanical performance. The purpose of this study is to analyze the change of different mechanical and physical properties of the adhesive when high temperatures are applied and define a methodology for limiting the curing temperature which minimizes the curing time ensuring good quality on the adhesive. For the definition of the methodology, first the requirements must be considered. As can be observed from the experimental results, most of the results show a similar tendency, but a detailed analysis must be carried out to set the maximum curing temperature. In order to analyze the results of the applied methods, four parameters are considered. The size of the sample gives information about use in real applications; methods that need a small amount of adhesive are more suitable because access to the bondline can be sometimes difficult.
Dependency on other materials is undesirable because they can influence the assessment of the adhesive quality. Sensitivity to predict degradation is needed because, as happened with some of the methods used in this contribution, if the sensitivity is too small, the line that separates a valid paste adhesive from another one slightly degraded is difficult to assess accurately. The possibility to qualify by giving a minimum value used not for only specific paste adhesive but in general is also desirable to generalize the methodology. After considering these parameters analyzing the experiments applied to the adhesive and its use to assess the adhesive performance it can be observed that: DSC is used to measure curing degree, but does not give any further information valid for qualification of adhesives. DMA measurements do not show a clear point where degradation is started and do not differ between curing at 100 1C, where the paste adhesive has good quality, and curing at 120 1C, when the paste adhesive is degraded. By this technique results cannot be compared with other adhesives because mechanical properties can differ without meaning degradation. TGA analysis cannot explain the entire decrease of performance because evaporation of particles can only explain formation of new voids and not the expansion of voids trapped in the mixing process and also does not give the possibility of qualification, but its application results important to assess the limit temperature for the curing process. Mechanical and density measurements require big samples to be tested, which are not always possible to acquire from real applications, where the access to the bondline can be limited, and in the case of 3 point bending there is not a clear difference between curing at 100 1C and 120 1C. Single lap shear testing, despite being today’s state of the art measuring technique, depends on the quality of the adherent used, which can influence on the results. Finally, optical analysis is found to be the most restrictive indicator to assess the bonding quality, having the possibility to analyze the quality of the bonding system with a small sample. This method restricts the use of curing at 120 1C for primary structures, having a void content higher than 2%, and voids which are found to clearly increase in size compared with samples cured with lower temperatures.
In order to measure the sensitivity of the measurement of void content, values for heating with lower temperatures are compared with results of single lap shear tests and shown in Fig. 16. As shown in the results, curing at 120 1C produces a void content higher than 2% which will lead to a lower mechanical performance. The high deviation of the results indicates that there are areas where the evaporation of particles has started.
Fig. 16. Void content compared with single lap shear test results.
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
Fig. 17. Results referenced to supplier’s curing at 80 1C.
Observing all the experiments carried out, some of them show more clearly a decrease of performance on samples cured using temperatures higher than 100 1C. In order to observe better the behavior of the adhesive at lower temperatures, from 80 1C to 120 1C, results of all the experiments are shown divided by the values of the curing reference at 80 1C and then compared in Fig. 17. All the test show a decrease of performance on samples cured with 1201 compared with samples cured at 100 1C. Optical testing by the analysis of voids as well as single lap shear test show more clearly a decrease of performance on samples heated with 120 1C compared with the reference curing at 80 1C. After discussing and comparing the results regarding the requirements, the methodology to assess the limit curing temperature of a paste adhesive can be defined. Four techniques are selected within the techniques applied in this study: TGA, DSC, single lap shear test and optical testing. TGA analysis of a non-cured paste adhesive sample determines the temperature where the evaporation of particles begins that will lead to the increment of porosity and a decrease of mechanical performance by measuring the onset point. TGA analysis of the mixed adhesive can be carried out, but by the analysis of the single components of the paste adhesive shows which component is evaporating at first. Once the onset is calculated, this temperature is used as a first assessment for the upper limit for the manufacturing of samples, which are produced covering all the range of temperatures from the recommended by the supplier until this onset point. After producing the samples by heating at different temperatures, DSC must be carried out to ensure the complete curing of samples. Then mechanical testing assesses the temperature limit where the performance of the adhesive is affected. Recommended mechanical testing is today’s state of the art, single lap shear. The limit temperature for the curing process is selected by observing when the mechanical performance of the different samples decreases. It is important to observe after the tests the fracture mode of the samples; an adherent fracture is desirable in the case of CFRP adherent. If the fracture mode is cohesive, it is possible that particles in the adhesive are evaporating and the temperature applied is too high. In the case of adhesive fracture, the degradation of the adhesive is clearly reached. Additionally, optical testing is recommended to measure the void content in order to validate the results from TGA.
5. Conclusions The optimal curing cycle for the paste adhesive under study is 100 1C for one hour, which reduces 75% the curing time if comparing with recommended curing conditions. For this curing cycle, it is proved that mechanical conditions are maintained.
119
Curing cycles of temperatures higher than 120 1C will produce thermal degradation on the paste adhesive. When low temperatures are applied, the void content is low. At low temperatures the voids come from the air trapped during the mixing process. Void content increases when samples are heated with higher temperatures due to the evaporation of some chemicals. The size of the void will also increase when higher temperatures are used, decreasing the mechanical properties of the paste adhesive. Decrease of performance when using higher temperatures was observed by all the tests carried out in this study. The state of the art for qualification of adhesives in aerospace industry, single lap shear test, showed adherent fracture for samples cured between 80 1C and 100 1C and a weaker cohesive failure in samples cured with higher temperature, and resulting in a decrease of mechanical performance due to the porosity increment. The goals of this contribution are to define a methodology to assess the limit temperature for the curing process paste adhesives and to assess the quality after curing independently from the adherent of the joint. By thermal analysis with TGA and DSC, simple lap shear and optical testing this limit temperature can be set accurately. Analysis of void content is an accurate method to assess the quality of a cured paste adhesive, by a simple optical study with the microscope. This technique can be used as complement for single lap shear tests, being especially useful because it does not depend on the quality of any other material, like the adherent in destructive techniques, and requires small samples. In this study it is shown that void diameter and void content of the paste adhesive can be precisely measured and represent the quality of the paste adhesive under study, when NDI techniques cannot be applied due to the small size of the voids.
References [1] Wegman R, Tullos T. Handbook of adhesive bonded structural repair. first ed.. New Jersey: Noyes Publications; 1992. [2] Troughton M. Handbook of plastics joining. second ed.. New York: William Andrew Inc.; 2008. [3] Armstrong K, Cole W, Bevan G. Care and repair of advanced composites. second edition. Warrendale: SAE International; 2005. [4] Duncan B, Girardi M, Read B. Measurement of basic mechanical properties for design. first edition. New York: Crown Publishing; 1994. [5] Wu X, Hao H. The effect of adhesive curing condition on bonding strength in auto body assembly. J Manuf Sci Eng 2005;127:411–9. [6] Katnam KB, Stevenson JPJ, Standley WF, Buggy M, Young TM. Tensile strength of two part epoxy paste adhesives: influence of mixing technique and micro void formation. Int J Adhes Adhes 2011;31:666–73. [7] Mylevarapu P, Woldesenbet E. Non-destructive characterization of bondline in composite adhesive joints. J Adhes Sci Technol 2006;20(7):647–60. [8] Dillard D, Pocius A. The mechanics of adhesion. first edition Blacksburg: Elsevier Science B.V.; 2002. [9] Lin L, Zhan X, Chen J, Mu Y, Li X. A novel random void model and its application in predicting void content of composites based on ultrasonic attenuation coefficient. Appl. Phys. A 2010, http://dx.doi.org/10.1007/s00339-010-6061-x. [10] Department of Defense. Composite materials handbook volume 3. Polymer matrix composite materials usage, design, and analysis. Department of Defense, MIL-HDBK-17-3F; 2002. [11] Davis M. Aircraft airworthiness and sustainment conference. Brisbane, AU; 2011. [12] European cooperation for space standardization. Space engineering: adhesive bonding handbook. European cooperation for space standardization, ECSS-EHB-32-21A; 2011, p. 149–159. [13] White J. Rubber processing. first edition Cincinnati: Hanser & Gardner Publications Inc.; 1995. [14] Dave´ R, Loos A. Processing of composites. first ed. Cincinnati: Hanser & Gardner Publications Inc.; 2000. [15] Levenspiel O. Chemical reaction engineering. third ed. New Jersey: John Wiley & Sons; 1999. [16] Sa´nchez A, Zogg M, Ermanni P. ICCM 18, 18th International conference on composite materials composites and reinforced plastics. Jeju, KO; 2011. [17] Sa´nchez A, Lucchini S, Zogg M, Ermanni P. SAMPE EUROPE, 32nd International technical conference & forum. Paris; 2011, p. 393–400.
Contents lists available at SciVerse ScienceDirect
International Journal of Adhesion & Adhesives journal homepage: www.elsevier.com/locate/ijadhadh
Methodology for optimization of the curing cycle of paste adhesives A. Sa´nchez Cebria´n n, M. Zogg, P. Ermanni Centre of Structure Technologies, ETH Z¨ urich, Leonhardstrasse 27, 8092 Z¨ urich, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 1 September 2012 Available online 11 September 2012
This contribution, carried out in the frame of the European Joint Technology Initiative ‘Clean Sky’, presents the results of a research program investigating the influence of fast curing on the quality of epoxy based paste adhesives. Today, the curing of paste adhesives is typically carried out following the supplier’s recommendations. In order to reduce cycle time, save costs and energy resources, paste adhesives could be cured at higher temperature. To ensure a bonded joint quality with maximum mechanical performance, the limitation of this temperature increase is studied. This study shows the effects of the use of high temperatures in the curing process, which can lead to a degradation of the adhesive system due to the increase of void content, decreasing the mechanical performance in the paste adhesive as well as in the bonded joint. The goal of this research is to find fast and robust processing of paste adhesives and to develop a methodology to determine the maximum curing temperature possible. Different properties of the adhesive are investigated, including different thermal analysis techniques, optical and mechanical testing of the pure adhesive. Additionally, state of the art qualification of paste adhesives, single lap shear testing, is considered. In this study, a novel method to control the quality of the cured paste adhesives is defined based on the analysis of the pure cured paste adhesive, not influenced by the adherent quality, by measuring the void content and its effects on the bonded joint. & 2012 Elsevier Ltd. All rights reserved.
Keywords: Paste adhesive Non-destructive testing Cure Certification
1. Introduction In this research contribution, a methodology to control the quality of paste adhesives with a high temperature curing process is investigated. The background of this research is to accelerate the curing cycle of the paste adhesive by increasing the temperature of the process without affecting the mechanical performance of the joint. The paste adhesive system used is the LMB 6687-1/LME 10049-3 from Huntsman Advanced Materials. The curing process is performed without pressure [1]. Samples are cured under different curing cycles, obtaining products with different properties that are analyzed. An acceleration of the supplier’s recommended curing cycle is possible without decreasing mechanical performance. For industrial applications, acceleration of the curing process is usually not applied as the mechanical performance of the joint can be affected if too high temperatures are applied [2]. For this reason, effects of an accelerated curing process are not considered [3]. The goal of this research is to investigate a methodology to control the quality of the paste adhesive when high temperatures are used in the accelerated curing process. This quality control is established through the
n
Corresponding author. Tel.: þ41 44 632 71 72; fax: þ 41 44 633 11 25. E-mail address: [email protected] (A. Sa´nchez Cebria´n).
0143-7496/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijadhadh.2012.09.002
analysis of different properties of the adhesive in a range of samples with different degrees of degradation. One of the main indicators of degradation of paste adhesives is an increase of the void content. An increased void content leads to a reduction of the mechanical performance of the paste adhesive. During the mixing process some air is entrapped [4]. If the paste adhesive is cured at higher temperature, the air inside expands, increasing the volume of the voids thus affecting the mechanical properties of the joints [5,6]. The quality of adhesives is mechanically tested by strength tests such as shear and peel tests [7]. Non-Destructive Inspection (NDI) techniques such as ultrasonic testing are also used to detect defects such as voids [8]. Maximum void content can be found in literature for composite panels and limitations about positioning of voids in the edge in order to avoid delamination problems [9]. However, there are no explicit considerations for bonding systems [10]. The detection of void content in paste adhesives is limited by the minimum defect size detected by NDI techniques. If the voids are small enough they may not be detected by NDI [11]. Today, there are no reliable methods to assess quality of a bonded joint [12]. This contribution focuses on the experimental analysis of this effect, observing how the increment of curing temperature affects the void content and how this increment of porosity affects different mechanical properties of the paste adhesive, and to find a methodology to assess the quality of the adhesive.
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
The samples produced in this research are completely cured applying different curing cycles, with temperatures from 80 1C (as recommended by the supplier) to 200 1C (which causes a clear thermal degradation on the adhesive). Physical and mechanical properties of these samples are then measured. The tests considered in this study include thermal analysis techniques, e.g. Differential Scanning Calorimetry (DSC) and Thermogravimetric analysis (TGA), as well as optical microscopy to observe the evolution of different physical properties of cured epoxy paste adhesive when the curing temperature is changed. Additionally, mechanical testing is considered by the measurement of flexural properties of the pure adhesive in three point bending test and Dynamic mechanical analysis (DMA) and single lap shear test with Carbon fiber Reinforced Polymers (CFRP) bonded systems. All the testing is firstly considered by separate in order to analyze the benefits and inconveniences of each test. From the information obtained from each test, recommendations for suitable qualification of paste adhesives are defined. The acceleration of the curing process of a paste adhesive is considered and a methodology to control the quality considering thermal degradation is established
2. Curing kinetics modeling theory Firstly, the curing cycles used to produce the samples are defined. The reference curing cycle from the supplier is composed of a gelling stage of 2 h at room temperature, followed by a curing process at 80 1C for 4 h. The rest of the curing profiles chosen have shorter cycle times but higher temperatures with a risk of thermal degradation. The definition of the curing temperatures is carried out by using the kinetics model of the paste adhesive, which gives the relation between temperature, time and curing degree for this paste adhesive system. The theoretical model of the chemical reaction is measured following Arrhenius relation, based on the n-th order kinetics [13], in the Eq. (1) f ðaÞ ¼ ð1aÞn
ð1Þ
This reaction can be written in the following form by combining an Arrhenius relation [14] da E ¼ k0 exp ð1aÞn ð2Þ RT dt The model depends on three variables: the degree of cure a [-], temperature T [K] and time t [s], and three parameters which characterize the reaction: k0 [-], which is the specific rate
113
constant at temperature T [K], the activation energy E [KJ/mol] and the reaction order n [-]. Once the parameters of the chemical reaction are defined, it is possible to predict which will be the degree of cure for a certain temperature applied for a given time. The equation with the degree of cure depending on temperature and time is shown in Eq. (3).
a ¼ 1 1ð1nÞ z t exp
1 1n E RT
ð3Þ
Eq. (3) can be solved by using multiple linear regression of the general form:z ¼ a þbx þ cy where the two basic parameters on Eq. (2), the reaction rate ddta [1/s], and the curing degree a, are determined from the DSC measurements. Three parameters (z, which is the Arrhenius frequency factor [1/s], E and n) are calculated by the software of the DSC, Pyris. Material obeying n-th order has the maximum rate of heat evolution at the beginning of the reaction [15]. This theoretical modeling for the kinetics characterization is carried out by a single dynamic heating measurement in the DSC of a non-cured sample of the paste adhesive. To minimize errors, this measurement is done 8 times and the average is used to measure the relation between temperature, time and degree of cure. Results are shown in Fig. 1. The curing times are chosen following the theoretic model, having always more than 90% theoretical degree of cure. It must be considered that samples will have an extra gelling stage not considered in the theoretical model. Therefore, the real degree of cure will be slightly higher than the theoretical one. This should reach more than 95%, as recommended by supplier. The curing profiles for the samples are summarized in Table 1.
Table 1 Summary of curing conditions for tested samples. Temperature (1C)
Time (min)
Theoretical curing degree (%)
80 100 120 140 160 180 200
240 60 60 45 30 15 10
93.2 91.3 95.6 97.1 97.8 97.8 98.2
Fig. 1. Kinetics of curing reaction of LMB 6687-1/LME 10049-3 from Huntsman Advanced Materials.
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
114
3. Material and methods Preparation of the samples is described in previous work [16], where some of the tests carried out are detailed. The experiments performed can be divided into physical and mechanical testing. Experiments carried out to study the change in physical properties include DSC and TGA analysis as well as the analysis of size and quantity of voids, carried out with microscopy techniques. With DSC analysis, the degree of cure of the paste adhesive is determined in order to validate the curing process carried out in the different samples analyzed in this study. DSC is used to validate if the measured degree of cure in the samples is higher than 95%. TGA analysis is carried out to measure the degradation of the components of the paste adhesive. The equipment in this case heats up a sample until high temperatures, measuring the weight during the process. In this experiment, the evaporation of particles due to the application of high temperatures is measured and the temperature limit can be assessed. This test is carried out three times. Components are analyzed by separate to observe where the degradation is starting and to observe the behavior of both components by separate as well as the mixed component. Optical microscopy techniques are used to measure the increment of void content in the paste adhesive when it is heated with a higher temperature. Void content of the samples as well as the average diameter and number of voids are measured by microscopy techniques (Leica DM RXA). Ten images are taken from different areas of each sample and they are post-processed with the software (Leica QWin), measuring the void content and the diameter of ten different voids for every image. The void content observed by optical testing is validated with the measurement of the density of all the samples used in the different mechanical tests carried out; DMA analysis, 3 point bending test of pure adhesive and single lap shear of bonded samples. DMA analysis measures dynamic mechanical properties of the material and its evolution in a temperature range; in this case, using a 3-point bending test. These results are expected to provide similar information on the mechanical performance from mechanical testing e.g. three point bending test on the standard testing machine, confirming the effect of degradation on mechanical properties of the adhesive. Additionally, the DMA analysis provides information about the effect of a higher curing temperature on the glass transition temperature (Tg) of the adhesive. Mechanical testing of the pure adhesive is carried out following ISO 178 for measurement of flexural properties of plastics in 3-point bending, on a standard testing machine Zwick 1474. Flexural modulus and flexural strength are measured in order to confirm the tendency measured in the DMA with smaller samples and to show the influence of the adherents, as used in single lap shear test. Finally, a single lap shear test, which is typically used as the state of the art qualification of bonded joints, is considered. Out of autoclave CFRP prepreg system (resin system MTM 44-1s, fibers Sigmatex CF 5804A) from Umeco, is used to manufacture the
adherents. Laminate plates are manufactured following the recommendations from the supplier. Surfaces of the adherents are treated by a manual sanding process (sand size P100); which has been previously demonstrated to improve mechanical performance of the joint [17]. The plates are also thoroughly cleaned first with acetone, then with tap water and finally with de-ionized water. Then, the CFRP plates are dried in a forced convection oven for 2 h at 65 1C. Once the surface treatment is complete, the bonding process is carried out by applying the adhesive and curing with the different curing cycles defined in 2. Finally, the samples are tested in simple lap shear test (SLS) following EN 2243-1 for structural adhesives. A summary of all the information necessary to prepare samples and carry out the experiments is summarized in Table 2.
4. Results and discussion 4.1. Analysis of physical properties Table 3 summarizes the degree of cure of the different samples measured with the DSC, considering the measured total released energy of the curing reaction, equal to 320 J/g. As it is shown, all the samples are cured more than 95% as expected. Therefore, the curing process can be considered complete for all the samples. TGA analysis is also considered. Results for separate components can be seen in Fig. 2. TGA shows the onset of the hardener at 124 1C and in the resin at 191 1C, but a loss of mass can be observed in the hardener before 120 1C, meaning that samples heated with this temperature or greater will have a higher percentage of voids. This effect is combined with the expansion of the voids trapped in the dosing and mixing process resulting that samples heated with temperatures higher than 100 1C will have more and larger voids which will affect the mechanical performance. Microscopy is used to observe voids inside of the paste adhesive and to measure the void content inside the frame that appears in all the images. Figs. 3–6 show examples of each sample measured. Void content of all the samples are summarized in Fig. 7.
Table 3 Curing degree of samples measured by DSC. Temperature (1C)
Time (min)
Released energy (J/g)
Curing degree (%)
80 100 120 140 160 180 200
240 60 60 45 30 15 10
10.5 7.6 7.6 8.4 11.4 12.5 3.1
96.7 97.6 97.6 97.4 96.4 96.1 99.0
Table 2 Summary of experimental parameters. Experiment
Samples
Sample size/weight
Initial conditions
Final conditions
Heating rate (K/min)
DSC DMA
1 1
1 min at 0 1C 5 min at 0 1C
1 min at 300 1C 5 min at 120 1C
10 3
TGA 3-pt bending SLS Optical testing
1 6 5 10
1–5 mg Rectangular 18 mm*8 mm*2 mm Force 110 mN7 100 mN (1 Hz) 5–15 mg Rectangular 40 mm*25 mm*2 mm 100 mm*25 mm*1 mmOverlap: 12.5 mm Rectangular 20 mm*15 mm*2 mm
1 min at 40 1C – – –
1 min at 250 1C – – –
1 – – –
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
Fig. 2. TGA analysis for epoxy and hardener.
Fig. 3. Samples cured at 80 1C (Avg. void content 1.6%) and 100 1C (Avg. void content 1.4%).
Fig. 4. Samples cured at 120 1C (Avg. void content 2.1%) and 140 1C (Avg. void content 21.4%).
Fig. 5. Samples cured at 160 1C (Avg. void content 33.5%) and 180 1C (Avg. void content 60.5%).
115
116
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
Fig. 9. Average diameter values of voids measured in samples cured with different temperatures. Fig. 6. Samples cured at 200 1C (Avg. void content 75.1%).
Fig. 10. Relative density measured in samples for mechanical testing. Fig. 7. Porosity values for samples cured with different temperatures. Table 4 Summary of optical measurements. Temperature (1C)
Time (min)
Average voids diameter (mm)
Average number of voids (-)
Void content (%)
Density (%)
80 100 120 140 160 180 200
240 60 60 45 30 15 10
49 720 60 730 100 730 230 780 260 7130 340 7240 270 7180
8.4 73 9.3 73 11 72 17 74 – – –
1.67 0.5 1.47 0.5 27 1 217 6 347 12 617 17 757 12
97.6 98.7 95.1 80.7 61.2 41.0 25.1
Fig. 8. Void content values for samples cured between 80 and 120 1C.
In order to observe better the values of void content of samples cured at lower temperatures, between 80 1C and 120 1C, Fig. 8 is shown. The results show a low void content for the samples cured at 120 1C or less when compared with the rest of the samples. However, an increase of void content is observed in samples cured at 120 1C. These results can be compared with TGA analysis, where it can be observed that when the hardener is heated with more than 100 1C, a loss of mass effect appears producing the evaporation of particles in the hardener. This fact can explain the big difference in the porosity levels between 120 1C and 140 1C, when evaporation effect is higher. Comparing the images in Fig. 5, it can be observed that not only do the voids increase the size, but also the quantity increases due to the evaporation of the component. The average diameter of the voids is also determined by measuring ten different voids for each sample. The density of the samples used in mechanical testing is also measured. Results are shown in Figs. 9 and 10 and summarized in Table 4. The results show a similar tendency as the size of the voids, having a clear increase of diameter for samples cured with more
than 100 1C. The number of voids increases slightly in the samples cured with higher temperatures from 80 1C to 120 1C and increases strongly for samples cured at 140 1C. As it can also be observed, the void content values are approximately inverted values of density, meaning a homogeneous distribution of the voids in the different samples. 4.2. Mechanical testing A 3-point bending test is carried out in the DMA for the different samples measuring the storage modulus from 0 1C to 120 1C. Curves of all the samples are shown in Fig. 11. It is observed a lower performance in samples cured with higher temperature. In order to evaluate results and compare them for all the samples, storage modulus at room temperature (20 1C) as well as the Tg are measured and shown in Figs. 11 and 12 and summarized in Table 5. It can be observed that the storage modulus at 20 1C decreases within the samples cured at higher temperature. The Tg remains constant. The increase in performance of the sample cured at 200 1C compared with the sample cured at 180 1C can be explained by the high degree of degradation of both samples.
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
117
Fig. 11. Storage modulus measured with DMA in a range from 0 1C to 120 1C.
Fig. 12. Storage modulus from DMA at 20 1C.
Fig. 14. Flexural strength measured in samples cured with different temperatures.
Table 5 Storage modulus at 20 1C and Tg values from DMA.
Table 6 Summary of results in 3 point bending test.
Temperature (1C)
Time (min)
Storage mod. at 20 1C (Pa)
Tg (1C)
80 100 120 140 160 180 200
240 60 60 45 30 15 10
4.4Eþ 08 3.4Eþ 08 2.9Eþ 08 2Eþ 08 1Eþ 08 4Eþ 07 1Eþ 08
112 113 117 120 122 110 120
Fig. 13. Flexural modulus measured in samples cured with different temperatures.
Temperature (1C) Time (min) Flex. modulus (MPa) Flexural strength (MPa) 80 100 120 140 160 180 200
240 60 60 45 30 15 10
1100 7120 1190 7110 1060 775 650 7120 475 7170 195 7110 115 720
417 4 447 2 417 2 257 6 157 6 67 3 37 1
Results of -point bending tests are shown in Figs. 13 and 14 and then summarized in Table 6. The results of the 3-point bending test show that the mechanical performance decreases with curing at high temperatures. A decrease of performance can be observed for samples cured with more than 120 1C. The tendency is similar to the DMA results, but the results of ISO 178 show high values for samples cured with 120 1C or less. Finally, a SLS test is carried out under EN 2243-1 for structural adhesives using CFRP adherents. Results and fracture modes are characterized and summarized in Fig. 15 and Table 7. Results show a decrease of shear strength when the curing temperature is increased having an adhesive failure mode in samples that are highly degraded. Samples cured with 80 1C and 100 1C show adherent failure, meaning that the adhesive is stronger than the
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
118
Fig. 15. SLS test for CFRP bonded samples.
Table 7 Results of single lap shear test. Temperature (1C)
Time (min)
Shear strength (MPa)
Fracture mode
80 100 120 140 160 180 200
240 60 60 45 30 15 10
24 72.9 22 72 16.3 71 15.5 72 6.0 71 6.1 70.5 6.6 71
Adherent Adherent Cohesive Cohesive Adhesive Adhesive Adhesive
adherent and the quality of the paste adhesive is well cured. Samples with cohesive failure show that the adherence of the paste adhesive is correct, but that the mechanical properties of the joint are lower than with adherent failure, meaning a certain degradation of the adhesive for samples cured with 120 1C or more. Finally, samples with adhesive failure show a poor adhesion between parts meaning that curing process is not correct due to the high temperatures applied. In this study, lower values are obtained due to the low quality of the adherent, which usually delaminates at about 20–25 MPa. Samples cured at 100 1C or less show a high shear performance until the fracture of the adherent.
4.3. Methodology to minimize curing time Today in industry the curing process of an adhesive is carried out following recommendations from the supplier, which are not optimized for energy consumption. By increasing the curing temperature, which is typically held constant during the process, an acceleration of the curing process is possible. The limitation of this temperature comes from the void formation when high temperatures are applied to the adhesive, which influences the quality of the joint reducing its mechanical performance. The purpose of this study is to analyze the change of different mechanical and physical properties of the adhesive when high temperatures are applied and define a methodology for limiting the curing temperature which minimizes the curing time ensuring good quality on the adhesive. For the definition of the methodology, first the requirements must be considered. As can be observed from the experimental results, most of the results show a similar tendency, but a detailed analysis must be carried out to set the maximum curing temperature. In order to analyze the results of the applied methods, four parameters are considered. The size of the sample gives information about use in real applications; methods that need a small amount of adhesive are more suitable because access to the bondline can be sometimes difficult.
Dependency on other materials is undesirable because they can influence the assessment of the adhesive quality. Sensitivity to predict degradation is needed because, as happened with some of the methods used in this contribution, if the sensitivity is too small, the line that separates a valid paste adhesive from another one slightly degraded is difficult to assess accurately. The possibility to qualify by giving a minimum value used not for only specific paste adhesive but in general is also desirable to generalize the methodology. After considering these parameters analyzing the experiments applied to the adhesive and its use to assess the adhesive performance it can be observed that: DSC is used to measure curing degree, but does not give any further information valid for qualification of adhesives. DMA measurements do not show a clear point where degradation is started and do not differ between curing at 100 1C, where the paste adhesive has good quality, and curing at 120 1C, when the paste adhesive is degraded. By this technique results cannot be compared with other adhesives because mechanical properties can differ without meaning degradation. TGA analysis cannot explain the entire decrease of performance because evaporation of particles can only explain formation of new voids and not the expansion of voids trapped in the mixing process and also does not give the possibility of qualification, but its application results important to assess the limit temperature for the curing process. Mechanical and density measurements require big samples to be tested, which are not always possible to acquire from real applications, where the access to the bondline can be limited, and in the case of 3 point bending there is not a clear difference between curing at 100 1C and 120 1C. Single lap shear testing, despite being today’s state of the art measuring technique, depends on the quality of the adherent used, which can influence on the results. Finally, optical analysis is found to be the most restrictive indicator to assess the bonding quality, having the possibility to analyze the quality of the bonding system with a small sample. This method restricts the use of curing at 120 1C for primary structures, having a void content higher than 2%, and voids which are found to clearly increase in size compared with samples cured with lower temperatures.
In order to measure the sensitivity of the measurement of void content, values for heating with lower temperatures are compared with results of single lap shear tests and shown in Fig. 16. As shown in the results, curing at 120 1C produces a void content higher than 2% which will lead to a lower mechanical performance. The high deviation of the results indicates that there are areas where the evaporation of particles has started.
Fig. 16. Void content compared with single lap shear test results.
´nchez Cebria ´n et al. / International Journal of Adhesion & Adhesives 40 (2013) 112–119 A. Sa
Fig. 17. Results referenced to supplier’s curing at 80 1C.
Observing all the experiments carried out, some of them show more clearly a decrease of performance on samples cured using temperatures higher than 100 1C. In order to observe better the behavior of the adhesive at lower temperatures, from 80 1C to 120 1C, results of all the experiments are shown divided by the values of the curing reference at 80 1C and then compared in Fig. 17. All the test show a decrease of performance on samples cured with 1201 compared with samples cured at 100 1C. Optical testing by the analysis of voids as well as single lap shear test show more clearly a decrease of performance on samples heated with 120 1C compared with the reference curing at 80 1C. After discussing and comparing the results regarding the requirements, the methodology to assess the limit curing temperature of a paste adhesive can be defined. Four techniques are selected within the techniques applied in this study: TGA, DSC, single lap shear test and optical testing. TGA analysis of a non-cured paste adhesive sample determines the temperature where the evaporation of particles begins that will lead to the increment of porosity and a decrease of mechanical performance by measuring the onset point. TGA analysis of the mixed adhesive can be carried out, but by the analysis of the single components of the paste adhesive shows which component is evaporating at first. Once the onset is calculated, this temperature is used as a first assessment for the upper limit for the manufacturing of samples, which are produced covering all the range of temperatures from the recommended by the supplier until this onset point. After producing the samples by heating at different temperatures, DSC must be carried out to ensure the complete curing of samples. Then mechanical testing assesses the temperature limit where the performance of the adhesive is affected. Recommended mechanical testing is today’s state of the art, single lap shear. The limit temperature for the curing process is selected by observing when the mechanical performance of the different samples decreases. It is important to observe after the tests the fracture mode of the samples; an adherent fracture is desirable in the case of CFRP adherent. If the fracture mode is cohesive, it is possible that particles in the adhesive are evaporating and the temperature applied is too high. In the case of adhesive fracture, the degradation of the adhesive is clearly reached. Additionally, optical testing is recommended to measure the void content in order to validate the results from TGA.
5. Conclusions The optimal curing cycle for the paste adhesive under study is 100 1C for one hour, which reduces 75% the curing time if comparing with recommended curing conditions. For this curing cycle, it is proved that mechanical conditions are maintained.
119
Curing cycles of temperatures higher than 120 1C will produce thermal degradation on the paste adhesive. When low temperatures are applied, the void content is low. At low temperatures the voids come from the air trapped during the mixing process. Void content increases when samples are heated with higher temperatures due to the evaporation of some chemicals. The size of the void will also increase when higher temperatures are used, decreasing the mechanical properties of the paste adhesive. Decrease of performance when using higher temperatures was observed by all the tests carried out in this study. The state of the art for qualification of adhesives in aerospace industry, single lap shear test, showed adherent fracture for samples cured between 80 1C and 100 1C and a weaker cohesive failure in samples cured with higher temperature, and resulting in a decrease of mechanical performance due to the porosity increment. The goals of this contribution are to define a methodology to assess the limit temperature for the curing process paste adhesives and to assess the quality after curing independently from the adherent of the joint. By thermal analysis with TGA and DSC, simple lap shear and optical testing this limit temperature can be set accurately. Analysis of void content is an accurate method to assess the quality of a cured paste adhesive, by a simple optical study with the microscope. This technique can be used as complement for single lap shear tests, being especially useful because it does not depend on the quality of any other material, like the adherent in destructive techniques, and requires small samples. In this study it is shown that void diameter and void content of the paste adhesive can be precisely measured and represent the quality of the paste adhesive under study, when NDI techniques cannot be applied due to the small size of the voids.
References [1] Wegman R, Tullos T. Handbook of adhesive bonded structural repair. first ed.. New Jersey: Noyes Publications; 1992. [2] Troughton M. Handbook of plastics joining. second ed.. New York: William Andrew Inc.; 2008. [3] Armstrong K, Cole W, Bevan G. Care and repair of advanced composites. second edition. Warrendale: SAE International; 2005. [4] Duncan B, Girardi M, Read B. Measurement of basic mechanical properties for design. first edition. New York: Crown Publishing; 1994. [5] Wu X, Hao H. The effect of adhesive curing condition on bonding strength in auto body assembly. J Manuf Sci Eng 2005;127:411–9. [6] Katnam KB, Stevenson JPJ, Standley WF, Buggy M, Young TM. Tensile strength of two part epoxy paste adhesives: influence of mixing technique and micro void formation. Int J Adhes Adhes 2011;31:666–73. [7] Mylevarapu P, Woldesenbet E. Non-destructive characterization of bondline in composite adhesive joints. J Adhes Sci Technol 2006;20(7):647–60. [8] Dillard D, Pocius A. The mechanics of adhesion. first edition Blacksburg: Elsevier Science B.V.; 2002. [9] Lin L, Zhan X, Chen J, Mu Y, Li X. A novel random void model and its application in predicting void content of composites based on ultrasonic attenuation coefficient. Appl. Phys. A 2010, http://dx.doi.org/10.1007/s00339-010-6061-x. [10] Department of Defense. Composite materials handbook volume 3. Polymer matrix composite materials usage, design, and analysis. Department of Defense, MIL-HDBK-17-3F; 2002. [11] Davis M. Aircraft airworthiness and sustainment conference. Brisbane, AU; 2011. [12] European cooperation for space standardization. Space engineering: adhesive bonding handbook. European cooperation for space standardization, ECSS-EHB-32-21A; 2011, p. 149–159. [13] White J. Rubber processing. first edition Cincinnati: Hanser & Gardner Publications Inc.; 1995. [14] Dave´ R, Loos A. Processing of composites. first ed. Cincinnati: Hanser & Gardner Publications Inc.; 2000. [15] Levenspiel O. Chemical reaction engineering. third ed. New Jersey: John Wiley & Sons; 1999. [16] Sa´nchez A, Zogg M, Ermanni P. ICCM 18, 18th International conference on composite materials composites and reinforced plastics. Jeju, KO; 2011. [17] Sa´nchez A, Lucchini S, Zogg M, Ermanni P. SAMPE EUROPE, 32nd International technical conference & forum. Paris; 2011, p. 393–400.