Multifunctional MEN-Doped Adhesives: Strengthening, Bond Quality Evaluation, and Variations in Magnetic Signal with Environmental Exposure

Featured Application

Non-destructive evaluation of bond quality using magneto-electric nanoparticles.

Abstract

Adhesive bonding of polymer matrix composites offers various advantages over traditional fasteners, such as a uniform stress state, reduced weight, and delay of composite delamination. However, adhesive bonding has limited implementation due to challenges in the prediction of durability. This work introduces a new method to monitor an adhesively bonded composite joint by dispersing magneto-electric nanoparticles (MENs) into the polymer precursor and monitoring changes in their surface charge density by evaluating the output magnetic signal under an applied magnetic field. Real-time monitoring of the curing process of a polymer adhesive was performed and corroborated via thermal analysis and mechanical testing. Lap shear and end notch flexure testing showed that adding 1 vol% MENs led to a ~23% increase in shear strength and a ~12% increase in mode II critical energy release rates compared to the undoped adhesive. Adding 5 vol% MENs also increased the adhesive’s peak tensile stress by ~8%. Strengthening mechanisms of the doped adhesive were monitored using in situ electron microscopy. A correlation between water ingression and a change in the magnetic moment was observed. Results show the MENs’ potential as a structural health-monitoring tool for a wide range of materials and applications.

1. Introduction

Polymer matrix composites (PMCs) have shown improved properties over traditional structural materials, especially weight-normalized properties, such as specific strength and stiffness [1]. While known for some time, the implementation of PMCs as structural materials is limited by complications in joining complex geometries [2]. Currently, the bonding of composite panels is completed primarily through the use of mechanical fasteners [3]. However, while their use can create a robust joint with the potential to compartmentalize crack growth, the process of drilling holes in a laminate remains a consistent challenge. For example, drilling in a PMC can create edge effects, leading to delamination and interlaminar failure, and also create locations of stress concentrations [3]. Furthermore, traditional mechanical fasteners are mainly manufactured from high-density metals, adding significant weight and complexity (possible issues with corrosion, etc.) [4]. Over-design of the composite panel is often necessary to compensate for these factors, leading to a loss in the overall efficiency [4].

One solution that could mitigate the issues associated with mechanical fasteners is to use adhesively bonded composite joints (ABCJs). Adhesive bonds offer the potential to join complex geometries at a fraction of the weight, simplifying a challenging problem. The implementation of this type of bond also allows for an improved distribution, eliminates the corrosion occurring in metallic fasteners, and reduces the number of parts of the structure. While the potential is vast, the current Federal Aviation Administration certification requires a demonstration that every bond used on a primary structure maintains integrity over the expected aircraft life cycle [5]. The motivation behind this restriction is that unlike mechanical fasteners, adhesive bonds cannot act to delay damage propagation, potentially leading to large-scale catastrophic failure [6]. Therefore, to increase the scope of PMC usage, a robust and exhaustive structural health-monitoring (SHM) tool must be developed to evaluate the quality of ABCJs without reliance on the initiation of damage in the bond [6]. The method should advance beyond the current structural health-monitoring methods (such as an E/M impedance and wave propagation approach [7], acoustic and ultrasonic methods [8], and terrestrial laser scanning [9]) that are limited in resolution and scope and rely on damage initiation.

In this work, the dispersal and integration of magneto-electric nanoparticles (MENs) into an adhesive resin was used to monitor the quality of a polymer adhesive when exposed to an accelerated temperature and a moisture environment. Monitoring was achieved through a coupling of magnetostriction and piezoelectricity in the MENs [10]. This consists in having an electric polarization when a magnetic field is applied, while obtaining magnetic polarization when an electric field is applied. A schematic of this relationship is shown in Figure 1. As a result, the dipole surface charge density of the particle induced an electric field, which caused a change in magnetization that was detected using standard magnetometry techniques [11]. The variation in the obtained magnetic moment was associated with the history of the environmental exposure of the bond. The integration of MENs also provided a multifunctional effect, leading to improved shear, tensile, and flexural performance over the unfilled adhesive. Additionally, the strengthening of the adhesive due to doping was investigated using in situ microscopy and a nanocomposite toughening model. The changes in the surface charge density of MENs were also used to non-destructively evaluate the curing process of the adhesive. In the past, MENs have been used for optical and magnetic applications [12,13] and as materials for biomedicine and drug delivery [14,15,16], but this is the first study to investigate their potential as a multifunctional SHM tool. Future applications of this work could lead to SHM techniques that can evaluate the loading history and damage state (without reliance on damage initiation) for a broad range of materials, loading types, geometries, magnitudes, and applications.

2. Materials and Methods

The carbon fiber composite material selected to manufacture the lap shear and end notch flexure (ENF) samples was a unidirectional prepreg material (T800H) from Toray. Following ASTM D5868-01 [17] and ASTM D7905/7905M [18] standards for the manufacturing process, the lap shear specimens had a 12-layer layup process and the end notch flexure specimens had a 10-layer layup process. A polyester peel ply was placed on both sides of the prepreg layups to prepare the surface for bonding. An autoclave from American Company Co. was used to cure the prepreg layup at a temperature of 177 °C and a pressure of 30 psi for 1 hour. The lap shear specimens were cut to a size of 25.4 mm by 177.8 mm, with a bonded area of 25.4 mm by 25.4 mm, while ENF specimens were manufactured to be 35 mm by 10 mm. The carbon fiber panels were bonded together using 3M’s Scotch-Weld two-part epoxy adhesive EC 2615 B/A. A 2:1 resin-to-hardener ratio was used. For the doped adhesive, MENs, synthesized to consist of a cobalt ferrite (CoFe₂O₄) core and barium titanate (BaTiO₃) shell nanoparticles with a median diameter of ~30 nm [19], were hand-mixed into the hardener component of the epoxy at a 1% volume concentration prior to curing. The fabrication and characterization of the MENs used in this study has also been discussed previously [14,15]. Both ENF and lap shear specimens were placed under vacuum during the adhesive cure in order to minimize void content. Next, the same epoxy adhesive was used to manufacture the tensile-testing dog-bone samples.

The lap shear specimens were tested using the ASTM D5868-01 standard [17]. A tensile tester from MTS (Criterion Model 43) was used to test the samples with a loading rate of 13 mm/min. The dog-bone tensile-testing samples were tested at a rate of 5 mm/min using the same MTS instrument using ASTM D638-03 [20]. ENF testing was performed conforming to ASTM D7905/7905M [19] using an MTI Instruments SEM 1000 micro-load frame in a three-point bend configuration. Testing was performed at a fixed displacement rate of 0.5 mm/min. The supporting span was 50 mm, while the sample was 120 mm in length and 3.5 mm in thickness and contained a 25 mm pre-crack. In situ ENF testing was performed within the chamber of a JEOL JIB-4500 SEM/FIB under vacuum at a pressure of 1.4 × 10⁻⁴ Pa or below after Au coating the sample for 30 s. Uncertainty measurements were obtained using the law of propagation of uncertainty [21]. To obtain the mode II critical energy release rate, G_IIC, in kJ/m² of the ENF samples, Equation (1) was used as follows [22]:

$${\mathrm{G}}_{\mathrm{II}}=\frac{9a^{2}P\delta }{2\beta \left(2l^3+3a^3\right)}$$

(1)

where a is the pre-crack in meters, P is the load in kN, δ is the displacement of the pre-crack, β is the width of the ENF samples in meters, and l is the distance between the two bottom points from the three-point bend test fixture, which is fixed at 0.0165 m.

To study the curing process of the 3M epoxy adhesive, three 8 × 8 × 4 mm samples were manufactured containing 5 vol% of MENs. This volume percentage was used since the manufacturing process with the resources available was time-consuming, but a higher percentage than the previous 1% used for the lap shear and ENF samples was used to obtain a higher signal using the B–H looper setup. The curing and testing process took place in a laboratory environment at room temperature (~22 °C). Signal measurements during the curing process were taken every 2 hours using the B–H looper setup, which is described later in this section.

Using the same adhesive, samples measuring 5 × 5 × 1 mm and containing 0, 5, 10, and 15 vol% of MENs were placed in an environmental chamber at 95% relative humidity and at an elevated temperature of 70 °C. The magnetic signatures were taken prior to environmental exposure. After a 4-week period, the samples were removed and again scanned. All magnetic signatures were taken using a vibrating samples magnetometer, which has a sensitivity of 0.01 µemu [23].

The adhesive’s glass transition temperature was obtained using a Q600 SDT from TA Instruments to determine its evolution as a function of the curing time during the first 24 hours of curing. The tests were run from room temperature (~22 °C) until 250 °C at a heating rate of 5 °C/min. Argon gas was used for the testing.

To monitor the local physical properties of adhesives non-destructively and non-invasively, the magnetic properties of MENs dispersed into these materials were evaluated using a B–H looper setup. Specifically, this setup was used to characterize the local properties of adhesive samples, adhesively bonded composite panels, and bonded mini double-cantilever beam (DCB) samples. The principle of reciprocity states that local values of the magnetic fields B and H are reflected by the system’s induced and applied voltages, respectively.

The application of other techniques, such as vibrating sample magnetometry (VSM), is limited since the sample needs to be attached to a piezo-actuator, thus requiring a specific sample preparation. In contrast, the B–H looper setup allows for the evaluation of the sample’s surface without physically damaging it. This setup takes advantage of the magneto-electric effect to non-invasively evaluate magnetic signature differences at a microenvironment level.

Based on the reciprocity principle, the electromotive force (EMF) or measured signal, ε, in volts is obtained by using Equation (2) as follows:

$$\epsilon =\frac{n∆\phi }{∆t}={\mu }_0 \frac{n∆ \int H_{img}M\left(r\right)dv}{∆t}$$

(2)

where Δφ is the change in the magnetic flux in volt-seconds, Δt is the change in time in seconds, n is the number of turns in a coil, M(r) represents how the magnetization is distributed throughout the sample, and H_img is the normalized reciprocal imaginary field representing the geometry of the B–H setup.

It can be seen in the B–H looper setup in Figure 2 how it consists of a function generator equipped in a lock-in amplifier and three different coils. Two detection coils are located at either side (shown in blue) with the purpose of acquiring a balanced detection. To amplify the magnetic signal from the sample and to cancel any background noise, these two detection coils were connected in series. The source coil located in the middle (shown in orange) was connected to the function generator to generate an AC magnetic field. The lock-in amplifier connected the detection and source coils and was used to amplify only the signal at the detection frequency and phase and cancel out all the noise signals at all the other frequencies. The sensitivity of the setup was reduced down to the microvolt level, and it was limited by the inductance of the whole setup, which is frequency dependent. Fortunately, this is not a fundamental limit, and we can further improve it in the future by increasing the number of balancing coils and optimizing the geometry of the coils through numerical simulations.

To test the curing process using the B–H looper of the adhesive containing MENs, an input voltage of 4 V and a frequency of 1.3 kHz were used, which is the critical frequency. To determine this frequency, the frequency was swept from 100 Hz to 10 kHz, and the value at which the signal due to the MENs reached its maximum was recorded. This frequency value was characterized by the net inductance of the whole system, which is influenced by the sample’s microenvironment [22]. Thus, the inductance is partially determined by the magnetic dynamics of the MENs dispersed in the sample. The critical frequency value slightly increases when the applied field is further increased via the source voltage.

When a voltage was applied to the source coil of the B–H looper and the sample was placed next to the coils, the AC magnetic field generated caused a magnetic flux in the sample containing MENs, due to the surface charge density of the nanoparticles. This magnetic flux was perceived by the detection coils, causing a change in signal. The background signal, which includes the noise factor, prior to putting the sample was subtracted from the recorded signal for every measurement.

3. Results

3.1. Magnetic Moment on MEN-Doped Adhesive

The influence of surface charge on the magnitude of the magnetic moment generated during VSM is shown in Figure 3a for MENs with and without a polymer coating. Results showed a 60% decrease in the magnitude of the magnetic moment when MENs contained a polymer coating as compared to the free particles. Figure 3b shows the difference in the VSM signals generated from the undoped and doped (1 vol% MENs) cure polymer adhesives (EC 2615 B/A, 3M). The doping of the adhesive resulted in an approximate 500% increase in the magnetic moment from approximately 10 µemu to 50 µemu.

The sensing capability of MENs predominantly comes from variations in the surface charge density based on the loading history [24]. For the uncoated and coated individual particles, Figure 4a shows the possible mechanism for a decrease in the surface charge and, as a result, the magnetic moment. As a MEN is coated with polymer chains, free H+ atoms in the chains act to neutralize the surface charge [15], increasing the bond strength between the particle and the polymer as well as lowering the measured magnetic moment, as shown in Figure 3a. Figure 3b shows how adding the nanoparticles to the adhesive caused the magnetic properties of the material to increase.

3.2. Monitoring of the Curing Process Using the B–H Looper

The B–H looper was used to evaluate the magnetic signal of adhesive samples containing 5% MENs to monitor their curing process. The initial magnetic signal (at curing hour zero) of the three samples was measured. Subsequent measurements were recorded, and the percentage increase with respect to the initial signal was obtained. The average percentage increase of the three samples was calculated and is graphed in red in Figure 5. It can be observed that there was an increasing trend in the magnetic signal during the first curing hours, which started converging after approximately 8 h of curing time.

To confirm the magnetic signal results obtained with the B–H looper, tensile testing was performed on the undoped adhesive to study whether a similar type of convergence was observed on the adhesive’s mechanical properties. The peak stress with respect to curing time is plotted in Figure 5 in blue. A similar trend was appreciated, in which there was a rapid increase in tensile strength during the first curing hours and then it started converging at around 8 hours of curing time until it reached ~41 MPa. This trend was also confirmed by the progression of the adhesive’s glass transition temperature (plotted in black) as it cured, which converged at ~137 °C.

3.3. Single Lap Shear

For the single lap shear, Figure 6 shows the relationship between displacement and peak stress at failure for each undoped and doped (1 vol% MENs addition) adhesive sample. The average shear strength for each condition was 25.01 ± 2.45 MPa and 30.79 ± 2.16 MPa, respectively, signifying an increase of 23% for the doped adhesive. This increase in the bond strength is significant as the primary application of the addition of MENs is for sensing aspects and any gains in bond quality are secondary. These samples experienced cohesive failure, and the addition of the nanoparticles did not seem to impact the type of failure in lap shear testing.

3.4. End Notch Flexure (ENF)

Regarding the ENF samples, Figure 7 shows the relationship between displacement and the mode II critical energy release rate G_IIC at failure for each sample that was doped and undoped. The average G_IIC value for each condition was 503 ± 46 J/m² and 564 ± 37 J/m², respectively, showing an increase of 12% for the doped adhesive. The strengthening of adhesive bonds with the addition of hard ceramic nanoparticles (such as Al₂O₃, ZnO, and TiO₂) is well documented [24,25,26,27,28,29]. These samples experienced cohesive failure, and the addition of the nanoparticles did not seem to impact the type of failure in lap shear testing. This study shows how adhesive joints containing MENs show an increase in mechanical performance compared to undoped joints. The increase in the critical energy release rate can be modeled using the formulation discussed in Johnsen et al. [27], in which the G_C of a particle-toughened polymer adhesive can be determined using the G_C of the adhesive (values obtained from ENF testing above) and the change in energy required for plastic zone growth. This energy can be obtained from the mechanical properties of the ceramic (E_MENs = 50 GPa [30]), the volume fraction of the nanoparticles (1 vol%) and voids, and the plastic zone radius of the unmodified polymer. Based on these inputs, the range of the expected critical energy release rate as a function of MEN addition was calculated to be between 546 J/m² and 553 J/m², which slightly under-predicted the experimental G_IIC values found with ENF testing (average of 564 J/m²).

In addition, a JEOL JIB-4500 Dual Beam FIB-SEM was used to take in situ images and a video of the failure of one doped ENF sample, which are included as Figure 8 and Video S1. Imaging shows localized shear loading, plastic zone evolution, and failure of the ABCJ in the crack tip region. As indicated by the blue arrows, localized damage was seen to occur outside the plastic zone in the form of visible microcracking resulting from the distribution of loading away from the pre-crack. In addition, the red arrows identify the initial formation microvoid cluster in the high-shear region. The fracture surfaces of the failed specimen showed similar plastic zone development in the failure and crack tip regions for plain samples and samples containing MENs. The most likely toughening mechanisms, in this case, are stress shielding of the crack tip and localized stress transfer to the stiffer nanoparticle [31].

3.5. Tensile Testing

Table 1. Maximum tensile stress in undoped and doped adhesives with 5 vol% MENs.

Sample Type/Sample #	Peak Tensile Stress (MPa)					Std. Dev. (MPa)	Avg. Stress (MPa)
	1	2	3	4	5
Baseline (2R:1H ratio)	38.09	36.61	36.10	38.44	35.00	1.28	36.85
Doped (5 vol% MENs)	38.18	36.32	42.56	40.58	43.84	2.76	40.30

displays the peak tensile stress on undoped and doped samples. It can be appreciated how dispersing 5 vol% MENs into the 3M adhesive increased its maximum tensile stress compared to the undoped samples, which was 39.5 ± 1.76 MPa and 36.6 ± 1.28 MPa, respectively. This signifies approximately an 8% increase, confirming that adding MENs to the epoxy adhesive does not negatively affect its mechanical properties.

3.6. Environmental Exposure

VSM scans were completed in individual cured adhesive resin samples (non-bonded) with concentrations of 5, 10, and 15 vol% MENs before and after environmental exposure, with the results shown in Figure 9. Before exposure, the doped adhesive showed an increasing magnetic moment with increased concentration (3.7 × 10⁻⁵ emu for 5 vol% MENs, 10.2 × 10⁻⁵ emu for 10.2 vol% MENs, and 15.5 × 10⁻⁵ emu for 15 vol% MENS). After exposure, the signal saturated and was consistent over varying MEN concentrations (magnitude approximately 15.0 × 10⁻⁵ emu).

As illustrated in Figure 4b, water ingression had the opposite effect on the magnetic moment. This can occur since the polarization of the surface charge of MENs can experience an increase due to the ingression of polar H₂O molecules. As the concentration of MENs increases, the magnetic moment can result from both the increased amount of magnetic material and the agglomeration of MENs, resulting in a higher percentage of MENs with uncoated free surfaces. After exposure, the saturation point of the magnetic moment was most likely the result of water ingression being a surface-area-dominated phenomenon.

4. Discussion

Using MENs to non-destructively evaluate adhesives using a magnetic signal can aid with the widespread implementation of adhesive bonding. It has been shown how this method detects magnetic changes that are affected by variations in the sample’s microenvironment, which allows one to measure changes in the material at a localized level. Correlating the curing process of the doped polymer adhesive with the strength of the output magnetic signal, and corroborating it with thermal analysis and tensile testing, shows the potential of this method to non-destructively evaluate materials not only during their lifetime but also during their manufacturing process. In addition, being able to measure changes in the surface charge density of MENs allows for the analysis of materials that undergo harsh environmental conditions, which causes the ingression of water molecules that polarize the surface of MENs.

Ideally, for each specific material, a baseline calibration curve needs to be obtained. Knowing how MENs behave in each material and how the magnetic signal evolves during the curing process will help to know when a specific sample is not curing ideally. This would help to observe whether a sample is incorrectly manufactured or whether the curing parameters are not correct. Future studies will focus on obtaining calibration curves for common adhesives used in the aerospace industry and then determining when specific samples undergo a curing process deviated from the baseline.

It is important to highlight that the tensile, lap shear, and end notch flexure testing shows enhancement in the mechanical properties of the adhesive bond when dispersing MENs into the adhesive. This means that the addition of these nanoparticles to the adhesive does not negatively influence the integrity of the adhesive bond when a low volume percentage is added. This is necessary so that MENs can successfully be used in industries, such as aerospace and automotive, where the integrity of bonds is of primary importance.

5. Conclusions

VSM and the B–H looper setup were used to non-destructively evaluate an adhesive resin doped with magneto-electric nanoparticles. These setups were used to obtain changes in the surface charge density in the nanoparticles of ABCJs with respect to curing time and varying history of environmental exposure. There was a strong correlation between the magnetic signal obtained during the curing process of the doped adhesive and its mechanical and thermal properties. There was also a relationship between the increased levels of exposure and an increase in the measured magnetic moment (up to a saturation point). In addition, mechanical property evaluation revealed an increase in the shear strength, mode II critical energy release rate (G_IIC), and tensile strength of MEN-doped adhesives. With further refinement and development, this technique has the potential to become a non-destructive evaluation tool for damage and quality inspection in both in-field and manufacturing settings.