Interface - American Chemical Society

Article pubs.acs.org/Langmuir

Molecular Behavior at Buried Epoxy/Poly(ethylene terephthalate) Interface Chi Zhang, John N. Myers, and Zhan Chen* Department of Chemistry, University of Michigan, 930 North University Avenue, Ann Arbor, Michigan 48109, United States S Supporting Information *

ABSTRACT: Epoxies are widely used as main components in packaging underfills for microelectronics. Their strong adhesion to different substrate materials is an important factor for the functioning of electronic devices. Amines are commonly used cross-linking agents for epoxides. However, the molecular mechanisms of epoxide−amine mixture adhesion to substrate materials remain unclear. In this research we investigated the adhesion mechanism of epoxide−amine mixtures at poly(ethylene terephthalate) (PET) interfaces using attenuated total-internal reflection Fourier transform infrared (ATR-FTIR) spectroscopy and sum frequency generation (SFG) vibrational spectroscopy. Results show that both epoxide and amine could diffuse into the PET film. They could also dissolve or modify the PET film at the interphase region. In the process of epoxy curing on PET, epoxide molecules could cross-link with the modified PET film, providing strong adhesion. This hypothesis was further confirmed by adding reactive and nonreactive silanes to the epoxies and measuring the adhesion strengths of such mixtures to PET. The reactive silanes could cross-link with the system, showing good adhesion, while the nonreactive silane prevented sufficient cross-linking, showing poor adhesion. This research developed an in-depth insight for molecular behaviors at the epoxy/PET interface which helped clarify the related adhesion mechanism.

1. INTRODUCTION Epoxy materials are widely used as coatings, adhesives, and composites in electronics, marine applications, aerospace materials, biology, etc.1,2 They have high thermal and chemical resistance, as well as good mechanical properties, due to the three-dimensional cross-linked structure after curing. Amines are commonly used epoxy curing agents, also known as epoxy hardeners.1,2 A primary amine group reacts with an epoxide group through an addition reaction and produces a hydroxyl group and a secondary amine group. The secondary amine group can further react with an additional epoxide group in the system.1,2 Although the details of the epoxy curing chemistry are well understood, the molecular adhesion mechanisms of epoxy to various other materials remain unclear. It is necessary to obtain in-depth understanding of adhesion mechanism at interfaces between substrate materials and epoxies because the failure at such interfaces is a general and problematic issue in the microelectronic industry.3 Such interfacial studies can help to establish further insight in interfacial chemistry and can also potentially help develop adhesives with better performance for many applications. Adhesion mechanisms are related to the interfacial molecular structures of the adhesives and the substrates. It is difficult to investigate such buried interfacial structures in situ. The most commonly used method to study an interface is to break the interface and study the two fractured surfaces using various © 2014 American Chemical Society

surface-sensitive spectroscopic techniques such as X-ray photoelectron spectroscopy (XPS),4 infrared (IR) spectroscopy, and Raman spectroscopy, or microscopic techniques such as atomic force microscopy (AFM)5,6 and transmission electron microscopy (TEM).7 Breaking an interface may greatly alter its molecular structure; therefore, this method may not be able to provide accurate structural information on the buried interface. It is necessary to develop methodologies which can probe buried interfaces in situ to accurately examine the interfacial structures. Optical spectroscopic techniques can be used to probe buried interfaces in situ if the input beam(s) can reach the interfaces. Both IR and Raman vibrational spectroscopies can be applied to characterize molecular structures of a material by examining intrinsic vibrational transitions of the molecules. There are surface-sensitive IR and Raman spectroscopies. For example, surface-enhanced Raman spectroscopy (SERS)8,9 can probe surfaces or interfaces, utilizing the surface-enhanced plasmon effect. However, SERS requires a rough metal surface and thus is difficult to be applied to study many different interfaces formed by dielectric materials. Attenuated total-internal reflection Fourier transform infrared (ATR-FTIR) spectroscoReceived: June 8, 2014 Revised: September 25, 2014 Published: September 27, 2014 12541

dx.doi.org/10.1021/la502239u | Langmuir 2014, 30, 12541−12550

Langmuir

Article

py10,11 can probe interfacial molecules, utilizing the evanescent wave of the IR beam. Although not intrinsically interface specific (i.e., molecules in the bulk which are inside the evanescent wave penetration depth in an ATR-FTIR experiment can also contribute an IR signal), ATR-FTIR spectroscopy is still widely used for adhesion-related studies because it can provide important information on the molecular behaviors of thin films. Second-order nonlinear optical spectroscopic techniques, especially sum frequency generation (SFG) vibrational spectroscopy, have been extensively used to study buried interfaces in situ.12−16 The SFG selection rule indicates that it is an intrinsic surface- or interface-selective technique which makes it unique to study surfaces or interfaces.16−25 SFG has been applied to study solid/gas (e.g., solid material surfaces in air),26−28 solid/liquid,29−32 liquid/liquid,33−35 and liquid/ vapor36−38 as well as solid/solid19,24,39,40 interfaces. The bulk of most thermoplastics is isotropic and will not generate an SFG signal. At a buried interface, the centrosymmetry can be broken and SFG signals containing interfacial molecular structural information can be generated and detected. This advantage makes SFG an important technique for studying buried adhesive interfaces including silicone and epoxy adhesives.41−48 The bulk of cross-linked epoxy resin is usually composed of randomly oriented epoxide−amine units, generating negligible SFG signals. Comparably, SFG signals from epoxy interfaces with various substrates might be resolvable and thus could be used to characterize molecular structures of these interfaces.43,49 We performed extensive research on the epoxy/poly(ethylene terephthalate) (PET) interface. PET is widely used as a barrier material in industry and as a material for thin film substrates in solar cells and electronic circuit boards.49 Epoxies have been used on PET substrates as packaging materials, and amines are used as cross-linking agents for epoxies.50,51 Understanding the adhesion mechanisms of epoxy adhesives to PET is crucial in developing effective polymer adhesives for advanced microelectronic applications and beyond. Buried interfaces between PET and a model epoxy with and without the incorporation of various silane molecules have been investigated previously in our group.42 In addition, we examined the buried interfacial structures between polymers and two commercial epoxy materials.42 The results indicate that interfaces with ordered methyl groups have weak adhesion, and interfaces with disordered molecular structures have strong adhesion. In this study, further investigations were carried out on the epoxy PET adhesion mechanism. Two model epoxides were cured on a PET surface using a model amine. We first applied ATR-FTIR to examine the interactions between the PET film and epoxides and/or amines. We also used SFG as an interface-specific technique to reveal the molecular structures at the interface area to help clarify the interfacial molecular behaviors and the adhesion mechanisms. Furthermore, interfacial molecular structures and adhesion were modified by reactive and nonreactive silanes in the two epoxy systems. A corresponding adhesion mechanism was proposed based on our experimental data and analysis.

solution. The polymer in solution was then spin coated on silicon, fused silica, and CaF2 right angle prisms (as solid supports). The silicon prisms were obtained from Chengdu-yasi Optoelectronics (Chengdu, China) for the ATR-FTIR experiments. The IR grade fused silica and CaF2 prisms were purchased from Alto Photonics Inc. for the SFG experiments. All silicon and silica prisms were soaked in a concentrated sulfuric acid bath saturated with potassium dichromate overnight at 60 °C before use. Then the prisms were rinsed with ultrapure water and dried with nitrogen gas. Air plasma (performed in a PE-50, Plasma Etch, Inc.) was used to further clean the prisms for 2 min. CaF2 prisms were cleaned using toluene, rinsed with ultrapure water, dried, and plasma cleaned for 2 min. The PET-d4 or PET films on all the prisms were prepared by spin coating the polymer solution on the prisms at 2500 rpm for 30 s. PET film thickness was measured as ∼58 nm using a Dektak 3 surface profilometer. Two epoxides, (1,4-butanediol)diglycidyl ether (BDDGE) and (1,4cyclohexanedimethanol)diglycidyl ether (CDDGE), were purchased from Sigma-Aldrich. Both epoxides were at liquid phase at room temperature, and they were cured using diethylenetriamine (DETA), which was also obtained from Sigma-Aldrich. The epoxy:amine ratio used in the experiment was calculated considering that each epoxide ring reacts with one amino-hydrogen atom. An additional 10 wt % amine was used in the epoxy mixture, taking into account the amine evaporation. The epoxide and amine were mixed homogeneously using a vortex mixer. The epoxy curing was carried out in an oven at 50 °C for 2 h. Then the samples were stored at room temperature for ∼12 h. In the mechanical adhesion test, two PET blocks were adhered together using epoxies with and without silanes. In the contact angle measurement, cured epoxy on substrate was peeled off to obtain the resulting surface for study. All silanes used in this experiment were purchased from Sigma-Aldrich, including (3-glycidoxypropyl)trimethoxysilane (γ-GPS), (3-aminopropyl)trimethoxysilane (ATMS), and octadecyltrimethoxysilane (OTMS). Silanes were incorporated into the epoxy−amine mixtures at 1.5 wt %. The mixtures were also well mixed and stored for 1 h before use. Chemical structures of the materials used in this work are shown in Figure 1.

Figure 1. Chemical structures of the materials used in this work: (a) poly(ethylene terephthalate) (PET), (b) poly(ethylene terephthalate) with deuterated aliphatic chain (PET-d4), (c) (1,4-butanediol)diglycidyl ether (BDDGE), (d) (1,4-cyclohexanedimethanol)diglycidyl ether (CDDGE), (e) diethylenetriamine (DETA), (f) (3glycidoxypropyl)trimethoxysilane (γ-GPS), (g) (3-aminopropyl)trimethoxysilane (ATMS), (h) octadecyltrimethoxysilane (OTMS). The ATR-FTIR experiment was carried out using a Nicolet 6700 FTIR spectrometer (Thermo Scientific Inc.). The sample chamber was purged with nitrogen gas to reduce IR absorption from water vapor and carbon dioxide. The experimental geometry used for the ATRFTIR experiment is shown in Figure 2a. A homemade holder was used to support the silicon prism in the FTIR instrument. An unpolarized IR beam was used for the ATR-FTIR study. The details of our narrow-band frequency scanning SFG experimental setup has been published previously.52 The SFG experimental geometry is shown in Figure 2b. In our study, the

2. MATERIALS AND METHODS PET (Mw ≈ 30 000) pellets were obtained from Scientific Polymer Products Inc. PET with a deuterated aliphatic chain (PET-d4 Mw ≈ 72 000) was purchased from Polymer Source Inc. PET or PET-d4 samples were dissolved in 2-chlorophenol (Sigma-Aldrich) to form a 2 wt % 12542


Langmuir

Article

vibrational peak assignments in this study are listed in Supporting Information Table S1. 3.1. Epoxide/PET Interface. The ATR-FTIR spectra of the spin-coated PET thin film (∼58 nm) on silicon substrate are shown in Figure 3 (black). A strong IR absorption at 1722 cm−1 was observed and could be assigned to the CO stretching vibration. Weak C−H stretching vibrational peaks could also be observed in the frequency range of 2800−3100 cm−1. No C O vibration was found in the other compounds used in the study such as epoxides (Supporting Information, section 2) or amines. Therefore, the peak at 1722 cm−1 could be used to monitor the change and the existence of the PET film. When the PET film contacted the BDDGE epoxide, a stronger C−H signal was observed due to the contribution from BDDGE in the IR penetration depth (Figure 3a, red). Furthermore, the CO peak intensity from PET decreased. Because no specific chemical reaction was expected between PET and BDDGE, the decrease of the PET signal was attributed to the solvation of PET by BDDGE (the dissolved PET that diffused out of the IR penetration depth could not be detected by ATR-FTIR). An interphase region with the presence of both PET and BDDGE was likely formed after their contact. Although BDDGE could dissolve part of the PET film, some PET features still remained and the spectrum was stable. After 2 h contact, the BDDGE was removed using a large amount of acetone (PET is not soluble in acetone) and the remaining surface was further washed using water and dried using nitrogen gas. The ATR-FTIR spectrum from the remaining film was collected and is shown in Figure 3a (blue). Almost no C−H stretching signal could be resolved, and a weaker CO stretching signal was detected (weaker than that of the original PET film). This suggests that BDDGE was washed away and that it dissolved part of the PET film, but a thinner PET film still remained on the silicon substrate surface. The partial removal of the PET film by BDDGE was also proved by film thickness measurement using a surface profilometer, as shown in Supporting Information. A similar study was also carried out for the CDDGE−PET interaction. The difference between the two cases was that the contact of CDDGE by PET showed less change of the PET CO signal (as shown in Figure 3b), indicating less solvation of PET by CDDGE. This result was also proved by PET film thickness measurement (Supporting Information). We placed PET pellets in pure BDDGE and CDDGE liquids for a week at the room temperature, respectively. Significant solvation of PET by BDDGE was observed while the solvation of PET by CDDGE was not as significant.

Figure 2. Experimental geometries used in this study: (a) ATR-FTIR spectroscopy and (b) SFG spectroscopy.

pulse energies of the IR and visible beams were ∼100 μJ and ∼30 μJ, respectively. The pulse widths for both beams were ∼20 ps. The visible beam diameter at the interface was ∼400 μm. All SFG spectra were collected with 5 cm−1 per step for IR frequency scan. All SFG spectra in this research were collected using the polarization combination of s-polarized SFG signal, s-polarized visible, and ppolarized IR, which usually gives the strongest SFG signals at interfaces. A mechanical adhesion test was performed to compare adhesion properties when reactive and nonreactive silanes were added to the epoxy system. The test was carried out using a commercial adhesion testing instrument (Instron 5544). PET blocks were obtained from Small Parts Inc. and were cut to pieces of smaller size (surface area: ∼24 × 10 mm2, thickness: ∼8 mm). Then they were sanded, washed using ultrapure water, and dried in air. Two PET block pieces were glued together using epoxies and were cured in an oven under the conditions mentioned previously. Then the glued test piece was pulled apart while the adhesion strength was monitored. The adhesion test geometry was a 180° shear test. The adhesion test data shown in this work, in megaPascal (MPa) units, were calculated by dividing the breaking force by the effective adhesion area. The breaking force was taken as the maximum force applied before a sudden decrease which occurred due to the adhesive failure. Water contact angle measurements were performed using a commercial contact angle goniometer (KSV Cam 101) to study the hydrophobicity of the fractured PET/epoxy interface after curing. Ultrapure water (Milli-Q water 18.2 MΩ·cm) was used in the measurements.

3. RESULTS AND DISCUSSION ATR-FTIR and SFG spectroscopic studies were performed at the epoxide/PET, amine/PET, and epoxy/PET interfaces. Silane molecules were also added to the epoxy system for interfacial structural modification. For convenience, all the

Figure 3. ATR-FTIR spectra of (a) PET (black), PET contacting BDDGE epoxide (red), PET washed with acetone and water after contacting BDDGE epoxide for 2 h (blue), (b) PET (black), PET contacting CDDGE epoxide (red), PET washed with acetone and water after contacting CDDGE epoxide for 2 h (blue). 12543


Langmuir

Article

Figure 4. SFG spectra collected from (a) BDDGE/blank CaF2 interface (black square) and BDDGE/PET-d4 (deposited on CaF2) interface (red circle), (b) BDDGE/blank silica interface (black square) and BDDGE/PET-d4 (deposited on silica) interface (red circle), (c) CDDGE/blank CaF2 interface (black square) and CDDGE/PET-d4 (deposited on CaF2) interface (red circle), and (d) CDDGE/blank silica interface (black square) and CDDGE/PET-d4 (deposited on silica) interface (red circle).

Figure 5. (a) ATR-FTIR spectra of PET (black), PET contacting DETA (red), PET after contacting DETA for 40 min and washing with water (blue); (b) time-dependent ATR-FTIR spectra of PET in contact with DETA.

This indicates that epoxide molecules possibly diffused into the PET-d4 film and segregated at the PET-d4/prism interfacial area, generating an SFG spectrum similar to that of the epoxide/prism (without PET-d4) interface. Although the PETd4 aromatic groups are ordered on the surface in air,53 no substantial ordering of such groups can be detected at the BDDGE/PET-d4 interface. This may be due to the epoxide solvation of the polymer film. To further confirm the BDDGE diffusion through PET, we used silica prisms as a solid support to perform the SFG experiment. The SFG spectrum collected from the BDDGE/ blank silica prism interface is quite different from that of the BDDGE/blank CaF2 prism, as shown in Figure 4b (black square). The epoxide ring peak at 3000 cm−1 became stronger while other peaks became weaker. Interestingly, the SFG spectrum collected from the BDDGE/PET-d4 (on silica) interface is quite similar to that from the BDDGE/silica prism interface but very different from that of the BDDGE/ PET-d4 (on CaF2 prism) interface. This indicates the diffusion of BDDGE through the PET film and the aggregation of BDDGE at the PET/prism interface. Additionally, although the PET-d4 surface generates a strong CO SFG signal in air, when contacting BDDGE, the CO peak at 1725 cm−1 disappeared (not shown), indicating the loss of order of such groups. This suggests that the PET-d4/BDDGE interphase area had no sharp boundary, generating no SFG signal. A similar study was performed on the CDDGE epoxide. As shown in Figure 4c,d, SFG spectra collected from the CDDGE/ substrate interface are similar for the same substrate with and without PET-d4 deposited but are quite different for different substrates. These observations imply that CDDGE molecules

SFG spectra of PET and PET-d4 in air have been published previously.53,54 For PET, both aliphatic and aromatic C−H stretching signals at 2960 and 3075 cm−1 were observed. For PET-d4, only the aromatic C−H stretching signal at 3075 cm−1 was detected.53,54 The PET surface in air also generated a strong SFG signal from CO stretching centered at 1725 cm−1.53 However, no SFG signal was detected from the PET/ substrate interface.53 Because epoxides tend to dissolve PET films, it is highly possible that epoxide molecules can diffuse into the films. SFG can be used to study such possible diffusion: If epoxide molecules diffuse into a PET film, they could further aggregate at the PET/prism substrate interface, generating an SFG signal; on the other hand, because of the PET solvation and epoxide diffusion, there should be no sharp boundary between PET and epoxide, likely generating no SFG signal. To monitor the behavior of interfacial epoxides without spectral interference with PET, we used PET-d4 in the following SFG experiments. Additionally, different prism substrates were used, including CaF2 and silica, because the diffused epoxide may have different structures while in contact with different substrates, which can be used to confirm the epoxide diffusion behavior. The interfacial SFG signal from a BDDGE/blank CaF2 prism interface was first collected as shown in Figure 4a (black square). The peaks centered at around 2850 and 2950 cm−1 were attributed to various methylene groups in the epoxide molecule; the peak at around 3000 cm−1 could be assigned to the C−H group in the epoxide ring.55 The SFG spectrum collected from the BDDGE/PET-d4 (on CaF2) interface is shown in Figure 4a (red circle), which is very similar to the spectrum collected from the BDDGE/blank CaF2 interface. 12544


Langmuir

Article

peaks. The peaks around 2820−2850 cm−1 and 2930−2960 cm−1 are assigned to the C−H symmetric and asymmetric stretching of the N−CH2 group in DETA, respectively. The peak centered at 3300 cm−1 is from the N−H stretching in DETA. However, the relative intensities of these three peaks can be similar or different in different spectra: they are similar when collected from the same prism substrate (regardless of having PET-d4 or not) and are quite different when collected from interfaces of PET-d4 on different prism substrates. Similar to the previous epoxide diffusion study, here we conclude that DETA can diffuse into PET-d4, generating an SFG signal from the PET-d4/prism interfaces. No SFG signal was likely generated from the PET-d4/DETA interphase because the chemical reaction between the PET and DETA resulted in no clear boundary between them. No SFG CO stretching peak was detected after contacting DETA with PET-d4 (not shown), indicating the loss of order of these CO groups due to the chemical reaction and diffusion. The amine diffusion could be further verified by the interactions of DETA molecules with a silane self-assembled monolayer grown on a silica substrate. The details of such an experiment are discussed in Supporting Information, section 3. 3.4. Epoxy/PET Interface. In the real adhesion applications, epoxies (epoxide−amine mixtures) are used instead of epoxide or amine alone. We contacted PET-d4 with the BDDGE−DETA mixture and the CDDGE−DETA mixture, respectively, and studied the resulting interfaces using ATRFTIR and SFG. When the BDDGE−DETA mixture contacted PET, the detected ATR-FTIR spectrum more closely resembled that collected from the BDDGE/PET case (Figure 7a and Figure 3a) and was different from the DETA/PET case (Figure 5a). This is because the major component in the mixture was BDDGE instead of DETA. After curing, the CO vibrational stretching intensity slightly decreased (Figure 7a). It is possible that more PET was dissolved and degraded in the curing process. When we separated the cured epoxy from the substrate and tested the resulting substrate surface, the CO vibrational signal intensity further decreased. This indicates that when we separated the cured epoxy from the substrate, the interphase between PET and the BDDGE−DETA mixture would break. Small PET CO signal could still be detected on the remaining substrate surface, indicating a small amount of PET remaining on the surface. For the CDDGE−DETA mixture case, similar results were found. However, as compared to the BDDGE−DETA mixture case, less PET was dissolved before and after curing the epoxy (Figure 7b). After breaking the interphase and removing the cured epoxy chunk, we examined the resulting prism surface. It was found that the CO signal intensity also decreased. Compared to the BDDGE−DETA case, a stronger CO signal remained after breaking the interphase, indicating that more PET was left on the substrate. Both epoxide and amine tend to diffuse into the PET-d4 film and adopt some order at the PET-d4/silica prism interface; the PET-d4 and epoxide or amine tend to have no clear boundary that could generate a detectable SFG signal. Therefore, we believe that there is no distinct interface when PET-d4 is in contact with the epoxy mixture due to their interactions and the diffusion of epoxy. The SFG results show that before curing the mixture, some signals, likely from the amine molecules, dominate the SFG spectra (Figure 8a,e, where only silica prisms were used as solid supports for PET-d4). The spectral

could diffuse into the PET film. The CDDGE/PET-d4 interphase area likely generated no SFG signal. 3.2. Amine/PET Interface. An ATR-FTIR study was also carried out to clarify PET−amine interactions. After contacting DETA with a PET film deposited on a silicon prism, strong C− H and N−H stretching peaks were detected but little change in the PET CO signal was seen in the IR spectrum (Figure 5a, red). A time-dependent ATR-FTIR study showed that the PET signal decreased only slightly over time, as shown in Figure 5b. It has been reported that DETA can react with PET by breaking the chemical bonds in the polymer backbone and by grafting N−H and O−H groups to the polymer.56−58 Therefore, such a signal change might be mainly caused by the reaction between PET and DETA. This reaction process might only slightly decrease the CO signal at 1722 cm−1 during contact but made the film much easier to be removed by breaking the polymer backbone. We removed DETA after the contact and washed the PET surface using water (the contact time was 40 min). The PET CO ATR-IR signal showed a significant decrease (Figure 5a, blue). In this case, a substantial amount of PET polymers were broken into smaller parts and removed from the surface after washing, with a smaller amount of PET molecules remaining on the substrate. This hypothesis was further proved by PET film thickness measurement before and after the amine contact, which is shown in Supporting Information. After the DETA was washed away with water, no amine spectral feature could be detected in the ATR-FTIR spectrum (Figure 5a, blue) because only a small amount of amine might diffuse into and remain in the PET matrix after washing. Furthermore, although the chemical reaction could graft N−H and O−H groups to the polymer, the polymer film was too thin and the amount of product too little to be detected by ATR-FTIR. These results suggest that a large amount of PET molecules reacted with DETA (after the surface was washed, the PET signal greatly decreased) with little disturbance of the PET film (if DETA only contacted PET without washing, little signal decrease was observed). It is likely that DETA can diffuse into the PET film and react with PET in the bulk. The diffusion of DETA into the PET film can be confirmed by SFG study. We compared the SFG spectra collected from the DETA/CaF2 prism, DETA/PET-d4 (on CaF2 prism), DETA/silica prism, and DETA/PET-d4 (on silica prism) interfaces, as shown in Figure 6. All spectra contain three

Figure 6. SFG spectra collected from (a) DETA/CaF2 interface (black square) and DETA/PET-d4 (deposited on CaF2) interface (red circle), and (b) DETA/silica interface (black square) and DETA/PET-d4 (deposited on silica) interface (red circle). 12545


Langmuir

Article

Figure 7. ATR-FTIR spectra collected from (a) PET (black), PET contacting uncured BDDGE−DETA mixture (red), PET contacting cured BDDGE−DETA mixture (blue), the resulting PET flim after breaking the PET/BDDGE−DETA mixture interface (green), and (b) PET (black), PET contacting uncured CDDGE−DETA (red), PET contacting cured CDDGE−DETA (blue), the resulting PET film after breaking the PET/ CDDGE−DETA mixture interface (green).

Figure 8. SFG spectra collected when PET-d4 was contacted by (a) uncured or (b) cured BDDGE−DETA mixtures with γ-GPS (red) or ATMS (blue) or without silane (black) incorporated; (c) time-dependent SFG spectra of PET-d4 contacted with BDDGE−DETA mixture incorporated with OTMS silane before curing the sample. The time interval between spectra is 15 min; (d) SFG spectra of PET-d4 contacted with BDDGE− DETA mixture incorporated with OTMS silane after curing; SFG spectra collected when PET-d4 was contacted by (e) uncured or (f) cured CDDGE−DETA mixtures with γ-GPS (red) or ATMS (blue) or without silane (black) incorporated; SFG spectra of PET-d4 contacted with CDDGE−DETA mixture incorporated with OTMS silane (g) before and (h) after curing the sample. Here only silica prisms were used as solid supports.

structure, very likely due to sufficient interfacial diffusion, crosslinking, or reaction which could lead to strong adhesion.42 To generalize such a conclusion and develop further understanding of the adhesion mechanism, in this work we incorporated reactive and nonreactive silanes into the epoxy systems and performed an SFG study and mechanical adhesion test on these samples. Silanes are small molecules and are widely used as addition promoters to change adhesion properties and other mechanical properties of adhesives. In this work we used three types of silanes, γ-GPS, ATMS, and OTMS, the molecular structures of which are shown in Figure 1. All silanes have the same functional methoxy head groups connected to a silicon atom. Methoxy groups can react with hydroxyl groups and link silane to a system releasing methanol molecules. In the epoxy/PET system, methoxy headgroups can react with hydroxyl groups on PET grafted by the amine. Additionally, γ-GPS has an epoxide

features are similar to those shown in Figure 6b but are quite different from those in Figure 4b and 4d. This implies that the amine primarily determines the structure of the PET-d4/silica substrate interface after diffusion of the epoxy mixture through PET-d4. After curing, the SFG spectra collected were similar to those collected before curing, but the signal intensities were comparably weaker (Figure 8b,f). Again we believe that these spectral features were likely generated from the PET-d4/silica prism interface. No SFG signal was likely generated from the PET-d4/epoxy interphase area due to the extensive chemical reaction, diffusion, and cross-linking in this area. This is an important feature for good adhesion.42 3.5. Silane Incorporation into the Epoxy System. A strong SFG interfacial signal detected from the polymer epoxy system has been shown to indicate insufficient interfacial crosslinking or reaction which could lead to weak adhesion.43 On the other hand, a weak SFG signal indicated a disordered interfacial 12546


Langmuir

Article

end group, which can react with amine. ATMS has an amino end group, which can react with epoxide. Therefore, both γGPS and ATMS can cross-link with the system. However, OTMS has a methyl end group, which cannot react with any functional groups in the system. Adding OTMS to the system could disrupt the interfacial cross-linking network. Silanes were incorporated into the BDDGE−DETA mixtures at 1.5 wt % before curing. Then the interfacial structures of the mixtures in contact with PET-d4 (deposited on silica prisms) were examined using SFG at room temperature. As shown in Figure 8a,b, for the γ-GPS and ATMS cases, the interfacial structures were similar to the case where silanes were not incorporated. This means that such silanes did not affect interfacial structures of BDDGE−DETA mixtures at the PETd4 interface. We also detected an N−H signal centered at 3300 cm−1 for all cases before curing, and the signal disappeared after curing (Supporting Information, section 4), indicating the reaction of amine N−H groups in the curing process. For the OTMS case, the interfacial signal in the C−H range was quite different. Shown in Figure 8c, a time-dependent SFG signal change could be detected at the interface when OTMS was incorporated into the mixture. As time increased, silane methoxy headgroups or the methylene backbone (2840−2850 cm−1) and methyl end group (2880 and 2940 cm−1) gradually ordered at the interface. A strong silane signal could be observed after curing the epoxy (shown in Figure 8d). The change of interfacial structure was monitored for 75 min. This time-dependent interfacial silane ordering change was not observed for the γ-GPS and ATMS cases. Because γ-GPS and ATMS are reactive silanes, they can react and cross-link with the epoxy system. However, OTMS has a nonreactive end group and thus could form an ordered layer at the interface, generating a strong SFG signal. We also incorporated silanes into the CDDGE−DETA mixtures and contacted the mixtures with PET-d4 on silica prisms. Similarly, for the addition of γ-GPS and ATMS, the interfacial structures were quite similar to that without introducing the silanes (shown in Figure 8e,f). A very weak signal could be detected at such interfaces before and after curing. For the OTMS case, before curing the CDDGE−DETA mixture at the PET-d4 interface, some signals from the silane could be observed. The SFG spectra at the interface were monitored for 2 h at room temperature. Unlike the BDDGE epoxide case, here the OTMS signal did not change with time before curing. The viscosity of CDDGE epoxide is higher than that of the BDDGE epoxy. Silanes may require higher energy to move and reorient in the CDDGE system. This may explain why, after curing the CDDGE−DETA mixture with OTMS, SFG signals from silane were weaker than that from the BDDGE case but still increased as compared to that from the uncured condition (shown in Figure 8d,g,h). We believe that OTMS tends to form an ordered hydrophobic layer at the substrate/PET-d4/epoxy interface. Such a silane layer is also likely to be present at the PET epoxy interphase under the following adhesion testing condition (when no silica or CaF2 substrate is used). This layer is not adequately cross-linked with the epoxy system and is easier to break as compared to other cross-linked interfaces. Therefore, weak adhesion is expected when OTMS is incorporated into the epoxy system while strong adhesion is expected for the other two silane cases. The adhesion test was carried out to measure the adhesion strength between PET and epoxy. As shown in Figure 9 for both epoxies, when γ-GPS and ATMS were incorporated into

Figure 9. Mechanical adhesion test results of PET blocks adhered by (a) BDDGE−DETA mixture and (b) CDDGE−DETA mixture with and without γ-GPS, ATMS, and OTMS incorporated into the system.

the system, adhesion strength of the epoxy-silane mixture to PET was measured to be similar to the case without silane incorporated. Differently, when OTMS was incorporated into the epoxy, the adhesion strength was significantly reduced. This result supports that nonreactive silanes might segregate to the epoxy/PET interphase, forming a silane layer that reduces the interfacial cross-linking. It is also shown in Figure 9 that after the same curing condition, CDDGE−DETA tends to have much stronger adhesion to PET than BDDGE−DETA. This may be due to different reaction rates of epoxies with DETA. In the curing experiment, we found that CDDGE had a higher reaction rate with DETA as compared to BDDGE. Therefore, CDDGE should also have a higher reaction rate with grafted N−H groups on the PET surface, resulting in better interfacial cross-linking and providing stronger adhesion. Additionally, comparing panels b, d, f, and h of Figure 8, we found that after curing, CDDGE−DETA/PET-d4/substrate system tends to generate a weaker SFG signal as compared to the BDDGE− DETA/PET-d4/substrate system. On the basis of the previous hypothesis that a weaker SFG signal indicates stronger adhesion,42 it is reasonable that CDDGE−DETA/PET has stronger adhesion. 3.6. Adhesion Mechanism between PET and Epoxy. On the basis of the above observations, we proposed a scheme for adhesion between PET and the epoxy, as shown in Scheme 1. It is known that amine can modify the PET surface structure.56−58 Amino and hydroxyl functional groups are introduced to PET during the amine−PET interaction.56−58 We showed that amine could diffuse into PET and react with the polymer bulk at the interphase area. Therefore, in this area the PET backbone tends to be broken with N−H and O−H groups tethered. At the initial contact of epoxy with PET, this reaction is triggered. In the curing process carried out at a higher temperature, this reaction is accelerated. Epoxide could also form an interphase with PET by dissolving part of the polymer. Therefore, in this interphase region, amine, epoxide, and PET are all present. In the epoxy curing process, the grafted N−H and O−H functional groups on PET would react with epoxide groups to cross-link epoxy with PET. The reaction between the N−H group and the epoxide group should dominate such cross-linking reactions, providing strong 12547


Langmuir

Article

Scheme 1. Interfacial Molecular Behavior between Epoxy and PET

It also shows the feasibility to combine ATR-FTIR and SFG spectroscopies to noninvasively study buried adhesive interfaces.

adhesion between epoxy and PET. Reactive silanes could sufficiently cross-link with the epoxy−PET system through their end groups, providing similar strong adhesion as the case without silane. On the other hand, nonreactive silane OTMS could not sufficiently cross-link with the system and tends to form a layer at the interface, giving weak adhesion. The silane proportion in the epoxy was low (1.5 wt %). Therefore, the OTMS silane layer might not cover the entire epoxy polymer interphase, leaving part of the area well cross-linked. This explains why, for the OTMS case, the adhesion strength was reduced as compared to other cases, but certain adhesion strength still remained (Figure 9). According to Scheme 1, amine tends to modify PET by grafting amino and hydroxyl groups for further epoxide reaction. After curing, additional hydroxyl groups may still remain due to their weak reaction with epoxide groups. Therefore, the cured epoxy/PET interface should be more hydrophilic as compared to the original PET surface. In packaging, it is well-known that water molecule can diffuse into the epoxy/substrate interface and delaminate the epoxy material from the substrate under high humidity or at high temperature.59,60 It is highly possible that the interfacial reaction during the epoxy curing process makes the interface more hydrophilic and thus allows water to easily diffuse through. This was proved by the water contact angle measurements of the fractured PET/epoxy interfaces after curing, which showed significantly reduced hydrophobicity. The details of the experimental results are shown in Supporting Information, section 6. Such reduction in hydrophobicity of the PET/epoxy interface could help to explain the water affinity to the interface and the poor water resistance of epoxy adhesives.48

■

ASSOCIATED CONTENT

S Supporting Information *

Peak assignments in the study, ATR-FTIR results for BDDGE and CDDGE epoxides in contact with bare silicon substrate, PET film thickness measurements, SFG and ATR-FTIR studies of silane diffusion into PET films, SFG spectra for the N−H stretching frequency range before and after curing the epoxy, and water contact angle measurements of the fractured interfaces. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: 734-615-4189. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This research is funded by Semiconductor Research Corporation (contract no. 2012-KJ-2282). REFERENCES

(1) May, C. A. Epoxy Resins: Chemistry and Technology; CRC Press: Boca Raton, FL, 1988. (2) Petrie, E. M. Epoxy Adhesives Formulations; McGraw-Hill: New York, 2006. (3) Lin, Y.; Zhong, J. A Review of the Influencing Factors on Anisotropic Conductive Adhesives Joining Technology in Electrical Applications. J. Mater. Sci. 2008, 43, 3072−3093. (4) Occhiello, E.; Morra, M.; Morini, G.; Garbassi, F.; Johnson, D. On Oxygen Plasma-Treated Polypropylene Interfaces with Air, Water, and Epoxy Resins. II. Epoxy Resins. J. Appl. Polym. Sci. 1991, 42, 2045−2052. (5) Binnig, G.; Quate, C. F.; Gerber, C. Atomic Force Microscope. Phys. Rev. Lett. 1986, 56, 930−933. (6) Stamm, M. Polymer Interfaces on a Molecular Scale: Comparison of Techniques and Some Examples. In Macromolecules: Synthesis, Order and Advanced Properties; Springer: New York, 1992; pp 357−400. (7) Muller, D. A. Structure and Bonding at the Atomic Scale by Scanning Transmission Electron Microscopy. Nat. Mater. 2009, 8, 263−270. (8) Moskovits, M. Surface-Enhanced Spectroscopy. Rev. Mod. Phys. 1985, 57, 783−826. (9) Moskovits, M. Surface-Enhanced Raman Spectroscopy: A Brief Retrospective. J. Raman Spectrosc. 2005, 36, 485−496. (10) Harrik, N. Internal Reflection Spectroscopy; Interscience: New York, 1967. (11) Mirabella, F., Jr. Internal Reflection Spectroscopy. Appl. Spectrosc. Rev. 1985, 21, 45−178. (12) Williams, C. T.; Beattie, D. A. Probing Buried Interfaces with Non-Linear Optical Spectroscopy. Surf. Sci. 2002, 500, 545−576.

4. CONCLUSION In this work, we studied the adhesion mechanism between epoxy and PET at a molecular level using SFG and ATR-FTIR spectroscopy. The results show that epoxides can dissolve part of the PET film during their contact and can diffuse into the PET film. Amine can diffuse into and modify the structure of PET. Both epoxide and amine tend to change PET surface structures at initial contact. The reaction between the epoxides and the grafted amino functional groups occurs during the curing process and could provide good interfacial cross-linking at the epoxy/PET interphase. This cross-linking event tends to give good interfacial adhesion. This hypothesis was further tested by coupling reactive and nonreactive silane molecules to the epoxy mixtures. Nonreactive silane OTMS tends to segregate at the interphase and prevents good cross-linking, leading to weak adhesion. Adhesion testing results further support the hypothesis. This research provides an in-depth understanding of the adhesion mechanism between PET and epoxy materials at a molecular level, which could help to develop advanced adhesives and adhesion-promoting materials. 12548


Langmuir

Article

Sensitive Amide I and Amide III Spectral Signals. J. Am. Chem. Soc. 2014, 136, 1206−1209. (33) Eisenthal, K. B. Liquid Interfaces Probed by Second-Harmonic and Sum-Frequency Spectroscopy. Chem. Rev. 1996, 96, 1343−1360. (34) Miranda, P. B.; Shen, Y. R. Liquid Interfaces: A Study by SumFrequency Vibrational Spectroscopy. J. Phys. Chem. B 1999, 103, 3292−3307. (35) Beaman, D. K.; Robertson, E. J.; Richmond, G. L. Ordered Polyelectrolyte Assembly at the Oil−Water Interface. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 3226−3231. (36) Perry, A.; Ahlborn, H.; Space, B.; Moore, P. B. A Combined Time Correlation Function and Instantaneous Normal Mode Study of the Sum Frequency Generation Spectroscopy of the Water/Vapor Interface. J. Chem. Phys. 2003, 118, 8411−8419. (37) Perry, A.; Neipert, C.; Kasprzyk, C. R.; Green, T.; Space, B.; Moore, P. B. A Theoretical Description of the Polarization Dependence of the Sum Frequency Generation Spectroscopy of the Water/Vapor Interface. J. Chem. Phys. 2005, 123, 144705. (38) Jubb, A. M.; Hua, W.; Allen, H. C. Environmental Chemistry at Vapor/Water Interfaces: Insights from Vibrational Sum Frequency Generation Spectroscopy. Annu. Rev. Phys. Chem. 2012, 63, 107−130. (39) Beattie, D. A.; Fraenkel, R.; Winget, S. A.; Petersen, A.; Bain, C. D. Sum-Frequency Spectroscopy of a Monolayer of Zinc Arachidate at the Solid−Solid Interface. J. Phys. Chem. B 2006, 110, 2278−2292. (40) Lu, X. L.; Shephard, N.; Han, J. L.; Xue, G.; Chen, Z. Probing Molecular Structures of Polymer/Metal Interfaces by Sum Frequency Generation Vibrational Spectroscopy. Macromolecules 2008, 41, 8770− 8777. (41) Vázquez, A. V.; Holden, B.; Kristalyn, C.; Fuller, M.; Wilkerson, B.; Chen, Z. Surface and Buried Interfacial Structures of Epoxy Resins Used as Underfills Studied by Sum Frequency Generation Vibrational Spectroscopy. ACS Appl. Mater. Interfaces 2011, 3, 1640−1651. (42) Zhang, C.; Hankett, J.; Chen, Z. Molecular Level Understanding of Adhesion Mechanisms at the Epoxy/Polymer Interfaces. ACS Appl. Mater. Interfaces 2012, 4, 3730−3737. (43) Vázquez, A. V.; Shephard, N. E.; Steinecker, C. L.; Ahn, D.; Spanninga, S.; Chen, Z. Understanding Molecular Structures of Silanes at Buried Polymer Interfaces Using Sum Frequency Generation Vibrational Spectroscopy and Relating Interfacial Structures to Polymer Adhesion. J. Colloid Interface Sci. 2009, 331, 408−416. (44) Ahn, D.; Dhinojwala, A. Sum Frequency Generation Vibrational Spectroscopy of Silicone Surfaces & Interfaces. Silicone Surface Sci. 2012, 23−58. (45) Zhang, C.; Shephard, N. E.; Rhodes, S. M.; Chen, Z. Headgroup Effect on Silane Structures at Buried Polymer/Silane and Polymer/ Polymer Interfaces and Their Relations to Adhesion. Langmuir 2012, 28, 6052−6059. (46) Zhang, C.; Chen, Z. Quantitative Molecular Level Understanding of Ethoxysilane at Poly(dimethylsiloxane)/Polymer Interfaces. Langmuir 2013, 29, 610−619. (47) Zhang, C.; Chen, Z. Probing Molecular Structures of Poly(dimethylsiloxane) at Buried Interfaces in Situ. J. Phys. Chem. C 2013, 117, 3903−3914. (48) Myers, J. N.; Zhang, C.; Lee, K.-W.; Williamson, J. M.; Chen, Z. Hygrothermal Aging Effects on Buried Molecular Structures at Epoxy Interfaces. Langmuir 2013, 30, 165−171. (49) Nomura, K.; Ohta, H.; Takagi, A.; Kamiya, T.; Hirano, M.; Hosono, H. Room-Temperature Fabrication of Transparent Flexible Thin-Film Transistors Using Amorphous Oxide Semiconductors. Nature 2004, 432, 488−492. (50) Minges, M. L. Electronic Materials Handbook: Packaging; CRC Press: Boca Raton, 1989; Vol. 1. (51) Lee, M. C.; Ho, T. H.; Wang, C. S. Synthesis of Tetrafunctional Epoxy Resins and Their Modification with Polydimethylsiloxane for Electronic Application. J. Appl. Polym. Sci. 1996, 62, 217−225. (52) Zhang, C.; Wang, J.; Khmaladze, A.; Liu, Y.; Ding, B.; Jasensky, J.; Chen, Z. Examining Surface and Bulk Structures Using Combined Nonlinear Vibrational Spectroscopies. Opt. Lett. 2011, 36, 2272−2274.

(13) Wilson, P. T.; Briggman, K. A.; Wallace, W. E.; Stephenson, J. C.; Richter, L. J. Selective Study of Polymer/Dielectric Interfaces with Vibrationally Resonant Sum Frequency Generation via Thin-Film Interference. Appl. Phys. Lett. 2002, 80, 3084−3086. (14) Sovago, M.; Vartiainen, E.; Bonn, M. Observation of Buried Water Molecules in Phospholipid Membranes by Surface SumFrequency Generation Spectroscopy. J. Chem. Phys. 2009, 131, 161107. (15) Chen, Z. Investigating Buried Polymer Interfaces Using Sum Frequency Generation Vibrational Spectroscopy. Prog. Polym. Sci. 2010, 35, 1376−1402. (16) Zhang, C.; Myers, J. N.; Chen, Z. Elucidation of Molecular Structures at Buried Polymer Interfaces and Biological Interfaces Using Sum Frequency Generation Vibrational Spectroscopy. Soft Matter 2013, 9, 4738−4761. (17) Shen, Y. R. Surface Properties Probed by Second-Harmonic and Sum-Frequency Generation. Nature 1989, 337, 519−525. (18) Cremer, P. S.; Su, X.; Shen, Y. R.; Somorjai, G. A. Ethylene Hydrogenation on Pt (111) Monitored in Situ at High Pressures Using Sum Frequency Generation. J. Am. Chem. Soc. 1996, 118, 2942− 2949. (19) Gautam, K. S.; Schwab, A. D.; Dhinojwala, A.; Zhang, D.; Dougal, S. M.; Yeganeh, M. S. Molecular Structure of Polystyrene at Air/Polymer and Solid/Polymer Interfaces. Phys. Rev. Lett. 2000, 85, 3854−3857. (20) Shultz, M. J.; Baldelli, S.; Schnitzer, C.; Simonelli, D. Aqueous Solution/Air Interfaces Probed with Sum Frequency Generation Spectroscopy. J. Phys. Chem. B 2002, 106, 5313−5324. (21) Lambert, A. G.; Davies, P. B.; Neivandt, D. J. Implementing the Theory of Sum Frequency Generation Vibrational Spectroscopy: A Tutorial Review. Appl. Spectrosc. Rev. 2005, 40, 103−145. (22) Wang, H. F.; Gan, W.; Lu, R.; Rao, Y.; Wu, B. H. Quantitative Spectral and Orientational Analysis in Surface Sum Frequency Generation Vibrational Spectroscopy (SFG-VS). Int. Rev. Phys. Chem. 2005, 24, 191−256. (23) Ma, G.; Allen, H. C. DPPC Langmuir Monolayer at the Air− Water Interface: Probing the Tail and Head Groups by Vibrational Sum Frequency Generation Spectroscopy. Langmuir 2006, 22, 5341− 5349. (24) Lu, X. L.; Li, D. W.; Kristalyn, C. B.; Han, J. L.; Shephard, N.; Rhodes, S.; Xue, G.; Chen, Z. Directly Probing Molecular Ordering at the Buried Polymer/Metal Interface. Macromolecules 2009, 42, 9052− 9057. (25) Quast, A. D.; Curtis, A. D.; Horn, B. A.; Goates, S. R.; Patterson, J. E. Role of Nonresonant Sum-Frequency Generation in the Investigation of Model Liquid Chromatography Systems. Anal. Chem. 2012, 84, 1862−1870. (26) Wei, X.; Zhuang, X.; Hong, S. C.; Goto, T.; Shen, Y. SumFrequency Vibrational Spectroscopic Study of a Rubbed Polymer Surface. Phys. Rev. Lett. 1999, 82, 4256−4259. (27) Jena, K. C.; Covert, P. A.; Hall, S. A.; Hore, D. K. Absolute Orientation of Ester Side Chains on the PMMA Surface. J. Phys. Chem. C 2011, 115, 15570−15574. (28) Calchera, A. R.; Curtis, A. D.; Patterson, J. E. Plasma Treatment of Polystyrene Thin Films Affects More Than the Surface. ACS Appl. Mater. Interfaces 2012, 4, 3493−3499. (29) Bain, C. D. Sum-Frequency Vibrational Spectroscopy of the Solid/Liquid Interface. J. Chem. Soc., Faraday Trans. 1995, 91, 1281− 1296. (30) Hsu, P. Y.; Dhinojwala, A. Contact of Oil with Solid Surfaces in Aqueous Media Probed Using Sum Frequency Generation Spectroscopy. Langmuir 2012, 28, 2567−2573. (31) Thompson, C. M.; Carl, L. M.; Somorjai, G. A. Sum Frequency Generation Study of the Interfacial Layer in Liquid-Phase Heterogeneously Catalyzed Oxidation of 2-Propanol on Platinum: Effect of the Concentrations of Water and 2-Propanol at the Interface. J. Phys. Chem. C 2013, 117, 26077−26083. (32) Ye, S.; Li, H.; Yang, W.; Luo, Y. Accurate Determination of Interfacial Protein Secondary Structure by Combining Interfacial12549


Langmuir

Article

(53) Loch, C. L.; Ahn, D.; Chen, C.; Wang, J.; Chen, Z. Sum Frequency Generation Studies at Poly(ethylene Terephthalate)/Silane Interfaces: Hydrogen Bond Formation and Molecular Conformation Determination. Langmuir 2004, 20, 5467−5473. (54) Loch, C. L.; Ahn, D.; Chen, Z. Sum Frequency Generation Vibrational Spectroscopic Studies on a Silane Adhesion-Promoting Mixture at a Polymer Interface. J. Phys. Chem. B 2006, 110, 914−918. (55) Carrasco, F.; Lacorte, T.; Briceño, K. Fourier Transform IR and Differential Scanning Calorimetry Study of Curing of Trifunctional Amino-Epoxy Resin. J. Appl. Polym. Sci. 2005, 98, 1524−1535. (56) Avny, Y.; Rebenfeld, L. Chemical Modification of Polyester Fiber Surfaces by Amination Reactions with Multifunctional Amines. J. Appl. Polym. Sci. 1986, 32, 4009−4025. (57) Nissen, K.; Stevens, M.; Stuart, B.; Baker, A. Characterization of PET Films Modified by Tetraethylenepentamine (TTEPA). J. Polym. Sci., Part B: Polym. Phys. 2001, 39, 623−633. (58) Bech, L.; Meylheuc, T.; Lepoittevin, B.; Roger, P. Chemical Surface Modification of Poly(ethylene terephthalate) Fibers by Aminolysis and Grafting of Carbohydrates. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 2172−2183. (59) Zhou, J.; Lucas, J. P. Hygrothermal Effects of Epoxy Resin. Part I: The Nature of Water in Epoxy. Polymer 1999, 40, 5505−5512. (60) Lam, D.; Wang, J. Hygrothermal Delaminations in Lead-Free Solder Reflow of Electronic Packages. J. Electron Mater. 2007, 36, 226−231.

12550
