Ultrasonic non-destructive evaluation study of molecular diffusion bonding of thin copper-aluminum electrode sheets

The weld quality of copper and aluminum thin electrode sheets in molecular diffusion bonding was non-destructively evaluated using ultrasonic resonance techniques. During the welding process, the intermediate layer material nickel diffuses into the molecules of both the copper sheet and aluminum sheet, resulting in the formation of a solid solution phase layer. This leads to a 5-layer structure in the welded body. If there are defects in this solid solution phase layer, it can cause mutations in the ultrasonic resonance signals within the weld body. In order to characterize the weld quality between copper and aluminum sheets, an acoustic attenuation coefficient was introduced. The ultrasonic resonance signals within the weld body of copper and aluminum thin electrode sheets were analyzed under four different welding states. Experimental testing revealed significant differences in acoustic attenuation coefficients among these different welding states. A smaller acoustic attenuation coefficient indicates better welding quality. Therefore, by setting a reasonable threshold for this coefficient, it is possible to effectively evaluate the welding quality of molecular diffusion bonding between copper and aluminum thin electrode sheets.


Introduction
Lithium batteries serve as the primary power source for electric vehicles, storing and releasing energy to propel the vehicle [1].The copper-aluminum battery pole piece is a crucial conductive component of the power battery that directly impacts critical technical indicators such as electric vehicle range, charging speed, and safety [2].The thickness of the copper and aluminum sheets in the battery electrode is about 1.2 mm respectively.Copper exhibits excellent electrical conductivity, while aluminum demonstrates superior heat dissipation performance.By employing molecular diffusion bonding technology to connect the copper and aluminum sheets, the respective advantages of these two materials can be fully utilized.There exists a 10 μm thick layer of nickel between the welded surfaces of copper and aluminum sheets.After molecular diffusion bonding, nickel diffuses and reacts between the two metals to form a solid solution phase layer [3,4].In the actual production process, the thickness of the intermediate layer material exceeding 10 μm presents dual challenges: firstly, it poses difficulties in achieving proper spraying; secondly, it elevates the risk of flaking off the nickel layer due to excessive stress, thereby compromising weld body strength.Conversely, if the intermediate layer is too thin, insufficient molecular diffusion gradient hampers effective molecular diffusion and prevents formation of a reliable solid solution phase layer, consequently impacting weld body strength [5][6][7].In the industrial production process, welding interface pollution (such as the presence of oil or impurities) or unqualified welding parameters (including welding temperature, pressure, and time) can result in incomplete molecular diffusion reaction on the surface of the weld joint, which significantly impacts the lifespan and performance of battery packs [8].
Currently, the non-destructive testing methods for molecular diffusion bonding primarily focus on ultrasound, electromagnetic, and radiographic techniques.Among these, ultrasonic detection encompasses measurements such as ultrasonic pulse reflection and transmission.S S Kumar et al investigated the water immersion pulse reflection method to molecularly diffusion weld two surfaces of a 1 mm thick nickel sheet to two copper columns with a diameter of 40 mm and a height of 20 mm, respectively, and realized the molecular diffusion bonding by calculating the interfacial echo reflection coefficient (the reflection coefficient is defined as the ratio of the amplitude of the wave reflected from a defect to that reflected from the rearmost end of the test sample) and correlating the echo reflection coefficient with the destructive test quality evaluation [9][10][11].However, considering that the copper-aluminum battery pole piece in this study are thin and the thickness of the intermediate nickel layer is only 10 μm, it is pretty difficult to calculate the reflection coefficients due to the mixing of the reflection echo signals at the defects with the echo signals at the back wall of the samples, which makes the calculation of the reflection coefficients quite difficult.Obviously, this method is not applicable.Y Zhu et al investigated the relationship between the transmission coefficient and weld strength using the water immersion ultrasonic transmission method.The average value of the transmission influence coefficient was compared with the weld strength profile obtained from mechanical experiments to evaluate the quality of the weld body [12].However, the difference in acoustic impedance between copper and nickel in this study is small, which can lead to a decrease in the transmission coefficient and thus amplify the inaccuracy of the experimental results.J L Rose used ultrasonic guided wave technology to inspect the diffusion welded structures of titanium alloys.Based on the results of the acoustic beam perpendicular incidence experiments, a three-layer model was proposed, and its dispersion curves and wave structures were analyzed.Two features of Lamb waves are extracted in the diffusion welded structure of titanium alloy plate: the spectral peak ratio and wave mode frequency shift (wave mode frequency shift).The results show that the extracted features are very sensitive to the state of diffusion bonding and can be used to evaluate the quality of interface bonding [13][14][15].However, Lamb wave is not sensitive to the welding quality of discrete small areas, and the data processing is relatively complex, which requires high requirements for inspectors.S Y Dong et al prepared dissimilar steel weld samples using 20 steel and stainless steel and fabricated artificial defects according to the standard, using the change of electromagnetic eddy current detection signal to reflect the weld quality of the weld body [16].In this study, the magnetic permeability and electrical conductivity of copper, nickel, and aluminum are very different.The change in the electromagnetic eddy current detection signal caused by the change in material will mask the change in the eddy current signal caused by defects, and the effect of detecting the weld quality of copper and aluminum thin electrode sheets using the electromagnetic method is not good.P Kapranos et al used a microfocus x-ray instrument for radiographic detection of titanium diffusion joints.They successfully detected defects up to 1 mm in size with a sensitivity of up to 0.025 mm utilizing an image intensifier and a digital image processor finally on display [17].Although the sensitivity of radiographic inspection is high, the radiation harms the human body, the operator needs to be specially trained, and the maintenance and operation cost is high.
With the wide application of machine learning, P. S. Mantra et al used a full factorial design methodology for experimental planning by selecting several process parameters such as welding pressure and welding time, developed a regression model to predict and simulate the weld strength of copper-aluminum thin plates using Artificial Neural Networks (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS), and used tensile and peeling loads to validation to ensure the accuracy of the results [18][19][20].Despite the results achieved by this method, there are significant application limitations in practical engineering testing as the welding quality is affected by numerous process parameters, and the chosen model inputs directly affect the training results.
The nondestructive evaluation means of the molecular diffusion bonding quality of copper and aluminum battery pole pieces in the industry is based on destructive tensile inspection, together with the metallographic inspection after random dissection, which is not able to realize the quality control of all products [21][22][23].In addition, the detection results of electromagnetic, x-ray or industrial CT and other technical means also have significant errors and cannot effectively evaluate the welding quality of copper and aluminum battery electrodes, which has a severe impact on the battery range of electric vehicles and battery safety.Therefore, there is an urgent need for a reliable nondestructive means of molecular diffusion bonding quality of copper and aluminum battery electrodes to be evaluated.
Due to the high-pressure, high-temperature process influence of the copper-aluminum battery electrode welding process, even if the copper-aluminum part of the welded area did not form a stable and reliable molecular diffusion of the solid solution phase layer, the different metal layers are still maintained between the close fit, the direct use of ultrasonic reflection/transmission signal amplitude method to assess the quality of the welded surface is not applicable, the amplitude of the received waveforms differed in a tiny way, which was difficult to distinguish.This study proposes a nondestructive evaluation method of molecular diffusion bonding quality of multilayer thin electrode sheets based on ultrasonic resonance technology.Taking copper and aluminum thin battery electrode sheets as experimental inspection objects, a water-immersed ultrasonic inspection system is established, the ultrasonic resonance signals of the electrode sheets with four different welding areas are respectively collected, and the acoustic attenuation coefficients are set as the characteristic parameter of the weld interfaces to judge the welding quality.Tensile testing machine was used to carry out tensile tests on different welded electrode sheets.Combined with metallographic analysis, the measured maximum drawing force and effective diffused solid phase layer area were correlated with ultrasonic testing results.The results show that the ultrasonic resonance method can effectively evaluate the molecular diffusion bonding quality of copper-aluminum electrode sheets.

Materials and processes
The thickness of the welding material copper sheet and aluminum sheet respectively is 1.2 mm, and the welding surface is a rectangle of 8x16mm, as shown in figure 1.In the experiment, a 10 μm thick nickel layer was first   sprayed on a 1.2 mm thick copper sheet, and then the surface of the aluminum base material was sandpapered to 1200# to remove the oxide film.Then, a 3 wt% NaOH solution was used to further etch the welded surfaces of the aluminum parent material after rinsing the corroded surfaces with water.Finally, the copper-nickel, aluminum parent material for ultrasonic cleaning (cleaning agent first acetone and then alcohol), hair dryer blow dry with a clean specimen tape wrapped, and as soon as possible the specimen placed in a vacuum furnace at room temperature and normal pressure storage [24].Subsequently, diffusion bonding is performed.The welding material is characterized by the parameters shown in table 1.Standard welding process parameters: vacuum degree is 1.5 × 10 −3 Pa, holding time is 30 min, welding temperature is 615 °C, welding pressure is 1Mpa.By controlling different welding areas, four samples with different welding qualities were prepared, as shown in figure 2.

Experimental equipment and instruments
A water-immersed ultrasonic inspection system was set up to collect ultrasonic signals from copper and aluminum diffusion bonding electrodes using the transmission method (ultrasonic transmitting and receiving probes were placed on both sides of the sample respectively).Transmitting probe to choose the center frequency of 5 MHz water immersion line focusing probe (On the one hand, the selection of line focusing probe emission can prevent the ultrasonic scattering of the probe, which is easy to cause interference at the welding edge; On the  other hand, the detection area can be reduced after line focusing to obtain more sound energy and improve the detection resolution); receiving probe for the straight probe (straight probe can effectively receive ultrasound transmission signals).The effective emission diameter of the probe wafer is 8 mm, which covers the entire welding height of the pole piece in the vertical direction (too large a diameter will cause ultrasonic signal interference in non-welded areas, while too small a diameter will not cover the entire welding height).The width of the acoustic beam after probe focusing is 3.2 mm, which is 1/5 of the total width of the welding, the vertical distance between the transmitting probe and the aluminum sheet is 25 mm, and the vertical distance from the receiving probe to the copper sheet is 8 mm, as shown in figure 3. The transmitting and receiving probes are strictly concentric and uniformly installed on a movable slider, which is driven by a stepper motor through a spiral screw.Stepping motor drive probes are to adjust the probe accurately offset distance each time the probe horizontal moving distance for the total width of the weld 1 / 6 to ensure that if part of the detection area overlaps, there will be no leakage detection.The tested pole piece is fixed vertically installed, located in the middle of the two ultrasonic probes, and the welding surface is facing the probes.The schematic diagram of the ultrasonic detection experimental setup is shown in figure 4.
Tensile experiments on welded joints can evaluate the tensile strength of welded joints, determine whether the welded joints are safe and reliable, and then verify the accuracy of the evaluation of the ultrasonic resonance method.Electronic tensile testing machine was used to carry out tensile experiments on the welded joints of copper and aluminum pole pieces, the upper and lower two collets were clamped on the end of the copper piece and the end of the aluminum piece, and the maximum tensile load of different copper and aluminum pole pieces was tested under the condition that the loading speed was 2 mm min −1 .

Experimental detection principle 4.1. Principles of ultrasound resonance
Ultrasound propagates in a particular medium or cavity, and when the conditions of the following equation are met (thickness is an integer multiple of half wavelength), standing waves are generated within the specimen, resulting in resonance phenomena.
Where: N is a positive integer (1, 2, 3•••);d is the thickness of the specimen;l is the wavelength of the ultrasound in the specimen; f 0 is the natural frequency of the specimen; -- f f m m 1 are two adjacent resonant frequencies;c is the velocity of sound of the examined specimen.
Thickness d and wavelength l are the critical parameters of ultrasonic resonance.The higher the ultrasonic frequency, the higher the resonance frequency, theoretically more effective identification of desoldering defects.However, the higher frequency ultrasonic attenuation is more severe than obtaining more effective resonance information.

Principle of sound attenuation coefficient measurement
In the case of resonance, the quality function is defined as [25,26]: Where: w is the angular frequency of the resonance point;Eis the total energy;DE is the energy lost per second; w D is the angular frequency of the bandwidth between the half-power resonance points.
The propagation time of ultrasound in the sample: Where: d is the sample's thickness;v is the sample material's sound velocity.According to Parseval's theorem, the energy lost per second by ultrasound: Where: a is the attenuation constant.
Carrying equation (4) into (2) provides: Then the relationship between the attenuation constants a and w D is as follows: Expression for the velocity of sound: Carrying equation ( 7) into (6) provides: Where: According to the lambert-beer theorem, it follows that: From equations (8) and (9), the acoustic attenuation coefficient can be calculated as follows: Where: Df is the interval between adjacent resonant frequencies, and Df is numerically equal to the resonant frequency f .D ¢ f is the bandwidth between the half-power resonance points, which is the bandwidth corresponding to 1 2 of the peak of the spectral signal.
The initial thickness of the intermediate nickel layer is 10 μm, which is much smaller than the wavelength of the acoustic wave, and the change in its thickness is negligible, so d can be considered constant.
Defining 20/d in equation (10) as a constant C, the state of the weld between the copper and aluminum sheets is characterized by a redefined acoustic attenuation coefficient b¢:

Experimental results and discussions
No. 1-7 samples are completely welded samples (100% welded area), No. 8-14 samples are 80% welded area samples, No. 15-21 samples are 50% welded area samples, and No. 22-28 samples are 30% welded area samples.Ultrasonic testing of samples 1-28, respectively, each sample to measure the centerline of the weld surface of the six positions of the acoustic attenuation coefficient value, and then its average value, from left to right for the measurement of points 1-6, as shown in figure 3. Differences in diffusion bonding states lead to differences in the absorption of ultrasound, and the attenuation of resonance signals in the weld body can be analyzed to effectively characterize the welding between copper and aluminum sheets.Ultrasonic waves are incident from the aluminum end and pass through the aluminum layer/aluminum-nickel layer/nickel layer/nickel-copper layer/copper layer in sequence.In the five-layer metal structure, the parameters of each layer (e.g., wavelength, thickness, Young's modulus, etc) affect the generation of resonance.Because the diffusion state of the nickel-copper and nickelaluminum molecules is complex to determine, their contribution to the ultrasonic resonance of the whole welded body cannot be calculated accurately.However, the welded body's ultrasonic resonance law trend is consistent.Figure 5 shows the ultrasonic resonance waveforms of the experimental samples.A fast Fourier transform is applied to the ultrasound signal, as shown in figures 6(a)-(d).The ultrasonic resonance frequency in the pole piece is 4.848 MHz (It can be inferred that when the thickness of the nickel layer is changed but the resonance conditions listed in equation (1) of 3.1 are satisfied, the acoustic wave will still resonate in the weld body, only the resonance frequency will change).Table 2 shows that the sound attenuation coefficient of the sample with 100% welded area is 0.58 to 0.65; the sound attenuation coefficient of the sample with 80% welded area is 0.69 to 0.89; the sound attenuation of the sample with 50% welded area is 0.95 to 1.00; and the sound attenuation coefficient of 30% welded area is 1.20 to 1.39.The average values of sound attenuation coefficient and maximum pulling force measured for different samples with different weld areas can be seen in figure 7: the sound attenuation coefficient gradually increases with decreasing weld quality, and the pulling force gradually decreases.Analysis of the reasons: When there are defects in the welding, the modulus of elasticity and Poisson's ratio of the organization changes, resulting in changes in its vibration mode so that the ultrasonic resonance bandwidth is narrowed or shifted by the formula  (11) can be derived from the acoustic attenuation coefficient value to be larger; welded interface due to the molecular diffusion of the more intact, energy attenuation is low, by the theory of ultrasonic resonance, the higher the energy of the higher the average resonance frequency is higher, and therefore its resonance bandwidth broadens, the value of the acoustic attenuation coefficient is relatively small.According to the new energy copper-aluminum row welding technology requirements (the first part of the material and process has a detailed description) and acceptance criteria: welding tensile strength is not less than the aluminum base material, the effective welding area shall not be less than 50% of the total welding area.Combined with the mechanical tensile experiment, when the tension is greater than 1.4KN, the strength of the welding area is greater than the strength of the aluminum material, the fracture occurs in the aluminum, as shown in figure 8(a); when the tension is lower than 1.4 KN, the strength of the welding area is lower than the strength of the base material, the copperaluminum overlap from the weld surface to occur separation, as shown in figure 8(b).
In conclusion, the welding area of 50% and the sound attenuation coefficient of 1 are set as the thresholds for determining the welding quality of copper and aluminum poles, and the welded poles with a welding area of more than 50% and a sound attenuation coefficient of less than one are defined as qualified samples.
To validate the accuracy of the experimental findings, the samples were vertically sectioned at the scanning points and examined under an electron microscope at a magnification of 200 times.Figure 9(a) illustrates a wellwelded sample, while figure 9(b) depicts a sample exhibiting oil defects.The upper yellow portion represents copper, whereas the lower white portion corresponds to aluminum.In between, there is a sequential arrangement of copper-nickel layer, nickel layer, and nickel-aluminum layer from top to bottom.Within the defective samples, detailed presence of copper-nickel solid solution layer can be observed; however, localized generation of nickel-aluminum solid solution layer occurs due to contamination by oil, which aligns with the test results.
In this experiment, the water immersion transmission method was used.Given the small size and simple shape of the thin copper-aluminum electrode sheet, the sample was completely submerged in water, which improves the resolution and sensitivity of the detection, reduces the blind spot, and also reduces the wear and  tear of the probes, but this also takes into account the discharge of water from the sink, which complicates the experimental operation and is not suitable for some samples that need to be kept dry.In addition, the system requires the transmitting and receiving probes to be tightly concentric, and precise mechanical control is needed to ensure that they remain aligned with the object under test.

Conclusions
In this study, based on the principle of ultrasonic resonance, ultrasonic inspection of the welded body of copper and aluminum thin electrode sheets in different welding states, combined with mechanical tensile experiments, the following conclusions were drawn: (1) The ultrasonic signal attenuation in the weld body of copper and aluminum thin electrode sheets with different welding conditions has significant differences, and the ultrasonic resonance frequency in the thin electrode sheets is 4.848 MHz.
(2) The sound attenuation coefficient gradually increases with decreasing weld quality, and the pull-out force gradually decreases.
(3) Welded pole pieces with a more than 50% welded area and a sound attenuation coefficient of less than one can be judged as qualified samples.
(4) The sliced sample exists as a layer structure, and the well-welded sample has five layers with no clear boundaries.In defective samples, the copper-nickel solid solution phase layer exists in fine detail, while the nickel-aluminum solid solution phase layer, due to oil contamination, is only locally generated.

Figure 1 .
Figure 1.Dimensions of the welding area of the sample.

Figure 3 .
Figure 3. Schematic of ultrasonic propagation for pole piece detection.

Figure 4 .
Figure 4. Schematic diagram of ultrasonic testing experimental setup.

Figure 6 .
Figure 6.Signal spectra of four different welding area pole pieces.

Figure 7 .
Figure 7. Average value of acoustic attenuation coefficient and average value of maximum tensile force for different welding areas.

8 .
Tensile fracture diagram of the sample.

Table 1 .
Characteristic parameters of welding materials.

Table 2 .
Sample test results for different welding areas.