Investigate the sealing and integrity of the laser welded X-ray counterpart collection device

doi:10.21203/rs.3.rs-3118780/v1

Download PDF

Research Article

Investigate the sealing and integrity of the laser welded X-ray counterpart collection device

https://doi.org/10.21203/rs.3.rs-3118780/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Observation of X-ray counterpart produced by the merger of binary compact star system is of great significance for studying their physical mechanisms. Based on the application requirements of the X-ray counterpart collection device, ring welded samples of 316L stainless steel and 4J29 Kovar alloy were investigated using a continuous wave laser. The effects of heat input on the weld formation, microstructure, mechanical properties, and sealing performance were studied. It was shown that because of the different heat distribution inside and outside the ring weld, the internal profile of the weld cross-section was closer to the circular shape and the columnar crystal growth near the fusion line was longer. Specially, the epitaxial growth was found on the Kovar alloy side. However, the element distribution of the weld was uniform and its crystal structure was austenitic phase. When the heat input was greater than 11.88 J/mm, the tensile strength of the weld was higher than that of the Kovar alloy. The tensile failure locations were at the Kovar alloy BM zone, except for samples with heat input less than 7.92 J/mm. However, when the heat input was greater than 37.50 J/mm, the weld had better sealing performance, and the minimum leak rate could reach 4×10^− 10 Pažmm³/s.

Laser welding

Kovar alloy

Stainless steel

Microstructure

Sealing performance

The X-ray detection technique can observe X-ray counterparts from the merger of binary compact star system in space, and the sealing and integrity of the X-ray counterpart collection device have a direct impact on the experimental data observed by X-ray detection techniques. The hermetic bonding of the device usually involves joining of copper tubing for ventilation, beryllium windows, and Kovar alloy to steel. The welding of the Kovar alloy to the steel component is an important part of the detector base, and the quality of this dissimilar metal joint directly affects the sealing and integrity of the device. It is a crucial and difficult part in the device manufacturing process. Since the device is placed in space, the sealing reliability of the Kovar alloy and steel weld determines whether the X-ray counterpart collection device can work normally and stably.

The Kovar alloy is a Fe-Ni-Co alloy with a very low thermal expansion coefficient close to that of inorganic materials such as glass and ceramics. Due to this character, the alloy is widely used in the joining of different metals, such as steel to glass or ceramics [1–3]. 316L stainless steel has excellent properties such as high strength, hardness, and corrosion resistance, and is commonly used in structures with demand strength [4, 5].

Because of the high welding heat input, severe post-weld deformation, and poor welding stability of arc welding, the dissimilar metal joints are prone to generate defects such as coarse grains, cracking, and brittle phase precipitation, resulting in decay on the mechanical properties of the joints [6–8]. Compared to arc welding, laser welding has low heat input, high energy density, profound heating and cooling rate, and narrow heat affected zone (HAZ), which is beneficial to dissimilar metals welding [9–11]. And laser welding is an effective method for producing hermetically sealed devices [12]. In order to meet the sealing manufacturing requirements of the X-ray counterpart collection device, laser welding process was used to join dissimilar metal components between Kovar and steel.

There are still some deficiencies in laser welding of Kovar alloy and stainless steel. Akbari Mousavi et al. [13, 14] studied pulsed laser welding of AISI 420 stainless steel and Kovar alloy, where the composition in the weld was not uniformly mixed, resulting in a non-uniform distribution of Cr. The offset of the laser beam towards the steel side could increase the crack sensitivity of the weld and increased the hardness of the weld. Hassan et al [15] investigated the microstructure of welded joints between the Kovar alloy and 304L dissimilar steel and the effect of welding heat input on the joint properties. The results showed that solidification cracks and incomplete mixing existed in the weld and increasing the heat input would result in increased thermal stresses as well as number of cracks. In addition, when the heat input was less than 15 J/mm, although the cracks were eliminated, the non-mixing phenomenon was more obvious.

It is easy to cause cracks and poor fusion of dissimilar metals due to short laser input time in one cycle when using pulsed laser to weld dissimilar metals. While using continuous wave laser can effectively reduce these defects. Cui et al. [16] used continuous laser welding of 304 stainless steel, obtained weld without cracks and elemental segregation. And the microhardness of the weld was increased due to the strengthening of fine grain. Rong et al. [17] used laser double-sided welding of 316L T-joints, found that the weld was consisted of austenite and a small amount of ferrite. Due to the heat treatment of the weld during the second pass laser welding, the hardness of the weld was greater than that of the first pass weld, and the hardness at the top of the weld was higher than that at the bottom. It is shown from the comparison between laser and laser-TIG welding of 316L(N) stainless steel [18] that hardness of the base material (BM) of the laser-TIG weld was higher than that of the laser weld because of higher heat input. From the above research, it can be seen that pulsed laser welding of Kovar alloy and steel has more defects such as cracks and lack of fusion, which deteriorates the mechanical properties of the weld. The use of continuous laser welding of Kovar alloy and steel is expected to reduce the occurrence of defects, especially for the defects of non-fusion, so as to obtain products that meet the requirements of the X-ray counterpart collection device.

In this paper, the sealing welding process of the X-ray counterpart collection device was studied, the 4J29 Kovar alloy and 316L stainless steel were welded using a continuous wave laser. The effects of welding heat input on weld appearance, microstructure, mechanical properties, and sealing performance were studied to reveal the relationship between weld metal elements and microstructure, as well as the effect of process parameters on the mechanical properties and sealing performance of the weld. The purpose of this paper is to explore the optimal process window to provide experimental support for the sealing performance of the X-ray counterpart collection device.

2.1 Materials

The BMs were 316L stainless steel and 4J29 Kovar alloy. Their chemical compositions were shown in Table 1. 316L stainless steel had a hollow disc structure with an outer diameter of 45 mm, an inner diameter of 17 mm, and a thickness of 1.5 mm, while 4J29 Kovar alloy had a disc structure with a diameter of 35 mm and a thickness of 0.5 mm.

2.2 Welding process

The Kovar alloy was stacked on the steel in a lap joint in this study. The two metals were welded using a laser beam that penetrated the Kovar alloy at the top surface. The path of the weld was a ring pattern with a diameter of 28 mm. The schematic diagram is shown in Fig. 1(a). A clamp was designed to hold the upper and lower discs and not interfere with the laser head movement. The body of the clamp resembles a circular tube, and the sample is fixed to the clamp using a pressure plate, and an auxiliary ring is used to ensure the concentricity of the two discs. The laser equipment was an IPG YLR 500-MM-CC laser and a Han's Laser WFF500-PT continuous laser control system. The maximum power of the laser was 500 W, the fiber diameter was 50 µm, and the fiber length was 10 m. The shielding gas was 99.99% Argon gas with flow rate of 20 L/min. The surface of the samples was polished with 2000 grit sandpapers before welding, the oil was removed by acetone, and the water was removed by ethanol. The process parameters were shown in Table 2, and welding tests were carried out under the conditions of the specific laser power and the welding speed, respectively.

2.3 Characterization

The weld metallographic samples were selected from the middle of the weld, ensuring that the cross-section and the ring tangent line were vertical. The metallographic samples were polished using 500, 1000, and 2000 grit sandpaper, and then polished using 2.5 µm diamond suspension. The cross-sectional metallography was observed by Carl Zeiss Microscopy GmbH optical microscope. The microstructure was observed using a ZEISS EVO Scanning Electron Microscopy (SEM) with an accompanying Bruker XFlash 6130 for X-ray Energy Dispersive Spectrometer (EDS) analysis. The Vickers hardness tester was loaded with 0.1 Kg and the loading time was 15 s. The tensile test used an electronic universal testing machine (SUNS, UTM5105X) with a special tensile test clamp. As shown in Fig. 1(b), the steel side was placed in the upper part, and the Kovar alloy was placed in the hole of the clamp. A cylinder was used to exert pressure on the Kovar alloy through the hole in the middle of the steel. With the help of this special structure, the pressure was ensured to act on the weld, which was transformed into the tensile force on the weld.

2.4 Sealing performance test

Considering the sealing requirements of the X-ray counterpart collection device, the sealing performance of the welded sample is the most important indicator of the reliability of the process. The tested samples were sequentially polished by 500, 1000, 2000 and 3000 grit sandpapers, and then polished using 2.5 µm diamond suspension to remove surface scratches. The device for the sealing performance test was shown in Fig. 2. The equipment used for the test was the KYKY ZQJ-530 helium mass spectrometer leak detector. The steel surface was attached to a steel cap and rubber ring to ensure the seal during pumping, and then the cap was connected to the leak detector through a pipe. Ethanol was applied between the sample and the rubber ring to prevent gas from entering through the surface. Repeat the test three times for each sample and take the average value. The leak detector was turned on and pumped until the pressure at the leak detection port reached 5 Pa and the leak rate was stable. The sample was then filled with 99.999% pure He to create a helium atmosphere until the leak rate reached stability for the second time. The principle of the helium mass spectrometer is to use He as a leak detector gas and to calculate the leak rate of the sample by detecting the He in the extracted gas. Filling the sample with He allows accurate testing of the leak rate, minimizes errors, and aids in determining the location of the leakage.

2.5 Temperature field simulation

To simulate the temperature changes during the welding process, a laser rotary Gauss heat source model was used as the laser heat source [19].

$${{\text{q}}_{laser}}=\frac{{9{\alpha _{abs}}Q}}{{\pi {R_0}^{2}H(1 - {e^3})}}\exp \left( {\frac{{ - 9({x^2}+{y^2})}}{{{R_0}^{2}\log (H/z)}}} \right)$$

Where, ${\alpha _{abs}}$is the total fraction of laser energy absorbed by the keyhole wall, can be obtained from literature [20]. is laser heat source height. When the molten pool is partially penetrated, is equal to the keyhole depth. ${R_0}$is the effective radius of the laser heat source, whose value is 41.5 µm.

The thermophysical properties of 316L stainless steel were obtained from literature [21] and the thermophysical properties of 4J29 Kovar alloy were obtained according to the elemental composition of the alloy using the thermophysical properties calculation software. The specific parameters were shown in Table 3. The temperature distribution on the weld surface as well as the molten pool shape and temperature distribution during the welding process were simulated. Combined with the experimental data, the effect of heat input on the weld formation can be better explained.

Table 1

Chemical composition of 4J29 Kovar alloy and 316L stainless steel (wt.%).
Alloy	Ni	Co	Cr	Mn	Mo	Si	S	P	C	Fe
4J29	28.06	15.00	-	0.40	-	0.20	0.02	0.02	0.03	Bal.
316L	10.17	-	16.78	1.19	1.81	0.51	0.03	0.04	0.03	Bal.

Table 2

Welding process parameters.
Sample	Power (W)	Welding speed (mm/s)	Heat input (J/mm)
1	475	10	47.50
2	475	20	23.75
3	475	40	11.88
4	475	60	7.92
5	475	80	5.94
6	425	10	42.50
7	375	10	37.50
8	325	10	32.50
9	275	10	27.50

Table 3

Thermophysical properties of 316L and 4J29.
Nomenclature	State	316L	4J29
Density (kgm^− 3)		8900	8100
Specific heat capacity (J/kgžK)	Solid	462 + 0.134T	420 + 0.15T
Specific heat capacity (J/kgžK)	Liquid	755	770
Thermal conductivity (W/mžK)	Solid	9.428 + 0.01571T	7.625 + 0.01673T
Thermal conductivity (W/mžK)	Liquid	12.41 + 0.003279T	-1.88 + 0.0209T
Solidus temperature (K)		1788	1725
Liquidus temperature (K)		1808	1740
Viscosity (kg/mžs)		0.004	0.003
Latent heat of fusion (J/kg)		260000	264000

3.1 Welding macroscopic geometry

To study the effect of laser power and welding speed on the weld bead, two sets of tests were conducted under a constant laser power and welding speed, respectively. The welding process parameters were shown in Table 2. The weld cross-sectional morphology and the curves of penetration depth and width obtained from the tests were shown in Figs. 3 and 4. Three locations on the weld cross-section were selected for dimensional measurements: the surface width (taken from the top position), the cross width (taken from the two BM contact locations), and the penetration depth of the weld. Figure 3(a) showed the curves of penetration depth and width at 475 W. It was showed that the surface width, the cross width, and the penetration depth of the weld decreased with increasing welding speed. The apparent trend of variation was observed at 10–40 mm/s speed, especially for the cross width, which decreased from 1257.4 µm to 399.16 µm. Figure 3(b)-(f) showed the cross-sectional morphology under different welding speeds, respectively. Figure 4(a) showed the penetration depth and width curves at 10 mm/s. As the laser power increased, the surface width, the cross width, and the penetration depth decreased gradually, and the distribution was approximately linear with a similar slope of the curve.

It was observed that as the welding speed increased, the geometry of the cross-section gradually developed from a "wine-cup" shape to a "nail" shape. And the penetration depth and width with the change of laser power showed the same effect. When the welding speed and laser power were changed, the welding heat input was also altered. As the welding speed increased, the residence time of the keyhole became shorter, resulting in a decrease in heat input and a smaller molten pool. This caused the weld geometry to resemble the shape of the keyhole, with reduced width and penetration depth, ultimately forming a "nail" shape. Moreover, higher welding speed made the formation of porosity defects more likely. On the other hand, as the laser power decreased, the heat input also decreased, leading to a smaller molten pool. The decrease in heat input, combined with consistent keyhole duration, resulted in reduced penetration depth and width, gradually shaping the molten pool into a smaller "wine-cup" form.

The weld cross width represented the bonding area of the weld, and generally, a larger bonding area indicated better weld performance. To some extent, it could reflect the weld's performance. From Figs. 3 and 4, it could be seen that the cross width increased with the increase in welding heat input. The weld cross width for variations in laser power was generally greater than that for variations in welding speed. This indicated a direct proportional relationship between cross width and heat input. Therefore, for the selection of process parameters, it was suitable to use the experimental conditions of samples 1, 6, and 7, which had higher heat input.

Because of the dissimilar heat accumulation and heat dissipation during the welding process, the welding conditions inside and outside of the weld were different, resulting in differences in the morphology and grain growth conditions. As displayed in Figs. 3 and 4, there were significant differences on both sides of the welds, especially under low welding speed conditions.

3.2 Microstructure and element distribution

Heat input can affect grain distribution and growth, thereby influencing the quality of welds. The microstructure and elemental distribution are better indicators of weld quality and can explain changes in performance. In the experiment, due to the welding path, there were significant differences in heat input between the inner and outer sides of the weld, which had a more pronounced impact on the microstructure. Sample 1 (475W, 10mm/s), which had the largest difference between the two sides of the weld in Fig. 3, was selected for the microstructure analysis. Figure 5(a) showed microstructure at the inside of steel and Kovar alloy on the outside of the weld, with the upper being the Kovar alloy and the lower being the steel. It was well known that the growth direction of columnar crystal was opposite to the temperature gradient. Therefore, the columnar crystal growth was also almost perpendicular to the fusion line, and grew from the weld boundary to the center of the weld during the solidification process.

Columnar crystal grew directly from the Kovar alloy BM could be observed. According to previous studies [13, 22], if the crystal structures of the BM and the weld were identical, the BM grains were able to serve as a substrate for the nucleation of the weld grains. When the molten pool was solidified, the liquid metal was more inclined to nucleate and grow by following the grain orientation of the BM, resulting in the same grain orientation of the weld and the BM, a situation called epitaxial growth. The Kovar alloy and the weld were Face Centered Cubic Crystals (FCC), which showed a tendency of epitaxial growth during the solidification process. As shown in Fig. 5(b), it was showed that the weld grains grew based on the Kovar alloy at the fusion line. Figure 5(c) showed a local magnified image of the weld on the steel side, and the weld and BM were separated by a line where the grains were regrown on the fusion line. Compared to pulsed laser welding, continuous wave laser welding can effectively improve the poor fusion of dissimilar metals in the weld, which helps the uniform distribution of elements and reduces the possibility of cracking [14, 15].

Figure 6 showed the microstructure of the inside of the weld. The fusion line on this side exhibited a circular profile, which reflected a different form from the linear profile on the outside of the weld. Due to the higher heat input during welding, the temperature and residence time of molten pool were bigger, resulting in wider weld. Furthermore, comparing Fig. 5 and Fig. 6, the epitaxial growth phenomenon was more obvious due to the increased heat input. Higher heat input and slower cooling rate provided suitable conditions for columnar crystal growth [4]. The length of the columnar crystal on the inside was significantly larger than that on the outside.

Figure 5(d)-(i) showed elements distribution of Fe, Co, Cr, Ni, Mn, and O on the outside of the weld, respectively. Ni and Co were mainly present in the Kovar alloy, and Cr and O were mainly present in the steel. Fe was uniform distributed on both the BM and the weld sides, while Co and Ni were predominantly distributed on the Kovar alloy side. Cr, Mn, and O were evenly distributed in the weld, with the most Cr content located in the steel BM and less at the fusion line. It should be noted that the content of element O in the welds was minimal and almost undetected.

Figure 7 showed the local magnified image of the central region of the weld, the red marker was the EDS measurement position. Equiaxed crystals and columnar crystals existed in the center of the weld. EDS local surface scan measurements were performed on the equiaxed crystals and columnar crystals, respectively, and the results were shown in the right figure of Fig. 7. According to the Schaeffler diagram [14] (Fig. 8), Ni equivalent (Ni_eq) and Cr equivalent (Cr_eq) of P1 were 20.18 and 9.44, respectively, and were judged as austenitic phase (γ). The Ni_eq and Cr_eq of P2 were 21.03 and 9.07, respectively, and were also judged as austenitic phase (γ). The austenitic phase was the main phase in the weld.

3.3 Microhardness

Microhardness is an essential indicator of the mechanical properties of a weld as it reflects the extent of thermal influence during the welding process. In order to accurately assess the microhardness of each region within the weld, microhardness tests were conducted on the 4J29 Kovar alloy side and the 316L stainless steel side of sample 1 (475W, 10mm/s). Figure 9 illustrated the location of the microhardness test, which was indicated by the yellow dashed line.

The microhardness curves were depicted by the white and red curves, representing the Kovar alloy side and the steel side respectively. The 4J29 side exhibited a hardness curve with an average hardness of approximately 168.95 HV in the BM. Notably, there was a significant decrease in microhardness, which could be attributed to the diffusion of metal elements within the weld, leading to a reduced content as evidenced by the EDS results in Fig. 7. Moreover, the heat transferred from the molten pool caused grain coarsening in the HAZ, resulting in a substantial decrease in hardness.

Furthermore, the hardness of the weld exhibited a decreasing trend from the outer region towards the inner region. On the steel side, the average hardness of the BM was around 215.33 HV. In the HAZ, the hardness displayed a minor downward trend. This could be attributed to the lower diffusion of heat into the HAZ from the laser beam, along with the lower grain roughness as illustrated in Fig. 5(a). The grain size in the HAZ resembled that of the BM, resulting in minimal variation in hardness. Additionally, there was a noticeable reduction in hardness within the weld, which could be associated with the movement of elements from the weld towards the entire weld, thereby reducing the Cr content in this specific area compared to the BM.

3.4 Tensile strength

The detector works in space with one atmospheric pressure inside and a vacuum environment outside. The weld is subjected to a large tensile force as a connection, so a high tensile strength is a guarantee for the normal operation of the device. In order to simulated the tensile force of the weld, a special clamp and tensile test method were designed. The tensile test used a special clamp and stretching method, resulting in two different force modes on the sample. The upper and lower sides of the weld were subjected to tensile forces and belonged to the tensile test. However, the Kovar alloy was subjected to tensile shear, which was a tensile shear test, as Fig. 2 showed. The results of the tensile test were shown in Fig. 10. Sample 1 (475W, 10mm/s) had the highest tensile force of 6065 N and sample 5 (475W, 80mm/s) had the lowest tensile force of 3577 N. The two modes of force resulted in two failure cases, which were discussed separately. Sample 4 (475W, 60mm/s) and sample 5 had lower tensile forces, and the samples fractured at the weld. The failure locations of sample 5 were shown in Fig. 10(c), which indicated that the tensile strength of the weld exceeded that of the Kovar alloy BM. The difference in heat input at different process parameters gave the weld a different morphology, while the higher heat input ensured a stronger performance of the weld.

The fracture surface morphology was shown in Fig. 11. Figure 11(a)-(c) were the fracture surface morphology and local magnification of the Kovar alloy side, and Fig. 11 (d)-(f) were the fracture surface morphology and local magnification of the steel side. There were obvious dimples with uniform size and distribution, showing the characteristics of ductile fracture [23]. On the Kovar alloy side, some of the surfaces showed pitted fractures, and some surfaces appeared smoothed, showing the brittle fracture characteristics [24, 25]. Therefore, the failure mode of the sample was judged as a mixed ductile and brittle fracture mode. And as seen in Figs. 3 and 4, the wider cross width and uniform distributed dimples might be the reason for the better tensile properties. In Fig. 9, the microhardness of the outside weld was lower than that of the inside, and the lower hardness generally demonstrated greater toughness [26]. The toughness and brittle fracture characteristics in Fig. 11(a) also reflected significant differences in the microhardness of different regions.

The tensile force of the samples under the remaining conditions was 6000 N approximately. The fracture locations were on the Kovar alloy BM, as illustrated in Fig. 10(b). The fracture surface morphology of the Kovar alloy BM of sample 1 was depicted in Fig. 12, there were dimples along the shear direction in a layered tearing pattern. It was obvious tensile shear fracture characteristics and could be judged as a ductile fracture. This was consistent with the phenomena observed by Dancette et al. [27] and Chao et al. [28], where shear dimples extended along the direction of shear loading. In this situation, the weld was the weak position of the sample, in Fig. 3(e) and (f), the cross widths of the weld were only 307.43 µm and 263.00 µm, which were much smaller than that of other samples, resulting in poor mechanical properties and tensile fracture location in the weld. The effect of welding heat input on tensile properties was reflected by the tensile strength and fracture morphology. When the heat input was higher, the weld had better bonding, larger penetration depth and width, uniform microstructure distribution, and obtained a stronger tensile strength with fracture location in the Kovar BM. When the heat input was lower, the weld bonding properties were worse, the tensile strength was lower, and the fracture was located in the weld.

3.5 Sealing performance

The X-ray counterpart collection device will carry out detection work in space, where the external environment of the detector is in a vacuum and the internal environment is filled with Xe and CO₂ as protective gases, with an internal pressure of 1 Bar. Therefore, the detector needs to withstand a pressure difference close to one atmosphere between the inside and outside in space, and ensure that the protective gas inside the detector does not leak during operation. But the side with higher gas density will transfer to the side with lower density through diffusion, permeation, and leakage [29]. In a vacuum, it is inevitable for gas to leak from the sealing device. However, considering the usage conditions of the device, a certain leakage rate is permissible to ensure the normal performance of the device. According to the design requirements of the detector, the internal pressure of the device should not be less than 0.9 Bar during the two years’ service period to maintain normal operation.

In space, where the environmental pressure is P₁ and the internal pressure of the device is P₂, the gas leak rate needs to meet:

$${\text{L}}=\frac{{({P_2} - {P_1}){\text{V}}}}{{\text{T}}}$$

Here, V represents the volume of the detector, and T represents time.

When the leakage rate reaches the design standard of 1×10^− 9 Pažmm³/s, it can ensure the normal operation of the detector during use. Weld is the first element to ensure sealing and must meet the design standard. In order to detect the sealing performance of welds, it is necessary to perform testing on the welds to ensure that the gas leakage rate is less than 1×10^− 9 Pažmm³/s. Helium is a small molecular gas with a volume much smaller than that of Xe and CO2 molecules, making it more likely to leak through welding defects. In addition, the content of helium in the air is only 5 ppm, which can minimize interference from environmental gases and gas adsorbed by the material on the test results. To simulate the operating conditions of the detector, a vacuum environment is created on one side of the weld, while the other side is in a helium environment at normal pressure. The helium mass spectrometer leak detector is connected to the vacuum chamber to detect the helium leaked through the weld and determine if the leakage rate meets the usage requirements.

The results of the sealing performance test were shown in Fig. 13. The height of the bar in the figure indicated the leak rate of the sample under the conditions of the corresponding laser power and welding speed, and the lower bar indicated the better sealing performance of the sample. From the figure, it was showed that all samples meet less than 1×10^− 9 Pažmm³/s leak rate index, However, the sealing performance of the weld showed a different trend with the change of process parameters. Sample 1 (475W, 10mm/s) had the best sealing performance, reaching 4.0×10^− 10 Pažmm³/s. As the laser power decreased, the sealing performance gradually decreased, and the maximum leak rate reached 5.2×10^− 10 Pažmm³/s. And with the increase in welding speed, the sealing performance appeared to increase and then decrease, sample 3 (475W, 40mm/s) had the largest leak rate, also reached 5.2×10^− 10 Pažmm³/s. Synthetically, as shown in the process parameters in Table 2, welds with heat input greater than 37.50 J/mm (sample1, sample6, sample7) had good sealing performance, while welds with heat input below this value had higher leak rates.

The sealing performance was directly related to the defects such as cracks and porosity and the fusion of the two materials. As the weld cross-sections shown in Figs. 3 and 4, only fewer pores were observed in the weld without cracks, and adequate fusion could be seen in Fig. 5. The result of fewer defects was the sound sealing performance obtained in this test.

4.1 The effect of heat input on the weld

The difference in the laser absorption rate of the material is reflected in the variation of the heat input during the welding process, which directly affects the quality of the weld. The effect of heat input on the weld can be clearly observed from the results in the previous section. This effect is reflected not only on both sides of the weld in the same parameter sample, but also in the samples with different parameters.

Firstly, the ring weld path was adopted in the welding. Because of the dissimilar heat accumulation and heat dissipation during the welding process, the welding conditions inside and outside of the weld were different, resulting in differences in the morphology and grain growth conditions. As shown in Figs. 3–4, the significant differences existed on both sides of the welds. Moreover, the laser dwell time was shorter under high welding speed conditions. The time of heat accumulation was too short, the heat input was insufficient, and the differences between the two sides of the weld were ambiguity.

Visualization of the effect of heat input on weld formation was characterized by means of temperature field simulation. As illustrated in Fig. 14(a), the temperature field on the surface of the sample was distributed as a ring shape, and there was a tendency for the heat to accumulate inside the ring. Figure 14(b) showed the cross-section of the molten pool, with the molten pool in the dark green area. Because of the ring path welding, the laser speed on the inside was slower than on the outside, so the laser heat input was higher on the inside and the heat dissipation conditions were worse than on the outside. As a result, the inside side absorbed more laser energy and the molten pool generates a larger temperature gradient, which showed in the cross section of the weld that the temperature gradient on the right side (which is the inside of the weld) was larger than that on the outside (which is the outside of the weld), and the temperature gradient was greater, so the melting area on the right side was greater. Figure 14(c) showed the simulated molten pool profile with the temperature distribution on the 4J29 side and 316L side of the molten pool, respectively, as shown in Fig. 14(d). The temperature distribution profile indicated that the temperature on the right side of the molten pool was higher than the left side at the equidistant position from the center of the molten pool. This resulted in a difference in the shape of the inside and outside of the weld.

Secondly, heat input also affected the growth of microstructure. Figure 14 (e) showed that the outside profile of the weld was nearly straight with shorter growth of columnar crystals from the fusion line. In contrast, the inside profile of the weld, with a longer columnar crystal and the growth direction perpendicular to the weld profile, was similar to a circular, as seen in Fig. 14(f). The shape of the weld and columnar crystal growth were directly relevant to the laser heat input. The heat dissipation conditions on the outside of the weld were excellent and the heat was not easily accumulated, resulting in a straight profile similar to the shape of the keyhole. It was difficult to release heat inside the weld, and then heat accumulated during welding, resulting in a circular profile, which also provided suitable conditions for the growth of columnar crystals. The grain growth is related to the temperature gradient (G) and growth rate (R, associated with the cooling rate). The greater the values of G/R, the more favorable the nucleation and growth of columnar crystals [30]. The right side of the molten pool had a greater temperature gradient and smaller cooling rate. Therefore, the value of G/R was larger than the left side, and the growth conditions of columnar crystals were enhanced.

The microstructure of the weld was demonstrated in Figs. 5 and 6. Due to the higher heat input during welding, the temperature and dwell time of molten pool were bigger, resulting in wider weld. Furthermore, higher heat input and slower cooling rate provided suitable conditions for columnar crystal growth [4]. The length of the columnar crystal in the inside was significantly larger than that in the outside, and the growth direction was also perpendicular to the fusion line and toward the center of the circle.

Finally, the tensile test results under different heat input conditions revealed the following. Sample 4 and sample 5 had smaller welding heat inputs of 7.92 J/mm and 5.94 J/mm, respectively, which were significantly lower than the heat input used in other welding conditions. Consequently, the welds had smaller cross widths, resulting in subtle failures. It should be noted that the heat input of sample 3 was only 11.88 J/mm, slightly higher than that of sample 4, yet the cross width of the weld was 399.16 µm, and the maximum tensile force reached 5879.40 N, surpassing sample 4. Thus, there existed a critical heat input value between 11.88 J/mm and 7.92 J/mm that enabled the weld to achieve high tensile strength and caused the failure position to occur in the Kovar alloy base metal zone. From a sealing performance perspective, more stringent requirements were placed on the welding process parameters, with a heat input of over 37.50 J/mm necessary to ensure a weld with satisfactory performance. In summary, a heat input of ≥ 37.50 J/mm is recommended to ensure the samples possess sound tensile strength within the parameters of this experiment.

4.2 The factors affecting the sealing performance

The maximum leakage rate of the sample was 5.2×10^− 10 Pažmm³/s and the minimum leakage rate was 4.0×10^− 10 Pažmm³/s, both of which met the use standard. But the causes of gas leakage need to be further explored in favor of improving the process and enhancing the sealing performance of the weld.

Pores are common defects in laser welding, and their formation can be classified into two main types [31–33]. The first type is caused by excessive heat input during laser welding, leading to the formation of a keyhole. The keyhole, influenced by the metal vapor inside and the forces of gravity and surface tension on the molten pool, remains in a quasi-stable state and undergoes periodic collapses. As a result, residual gas trapped in the keyhole forms bubbles. The second type is attributed to the presence of hydrogen (H) within and on the surface of the base metal (BM). As the metal transitions from a liquid to a solid state, the concentration of dissolved hydrogen significantly decreases. Consequently, when the molten pool cools, hydrogen precipitates and accumulates, forming bubbles.

To investigate the cause of the leakage of the weld, sample 7 was cut along the longitudinal direction of the weld and the longitudinal section was observed, as presented in Fig. 15. Figure 15(a) showed the macroscopic morphology of sample 7, and Figs. 15(b-f) displayed the morphology of five longitudinal sections taken along the welding direction. It was observed that no cracks were found in the weld, but a large number of pores existed in the longitudinal section, mainly distributed in the middle and bottom of the weld, and the shape of the pores was nearly circular. The number of pores showed a tendency to increase along the welding direction, with a significant increase in the number of pores in the position of Figs. 15 (e-f). The higher number of pores present in the weld might have been the reason for the larger leakage rate of sample 7. In particular, as shown in Fig. 15(f), pores were also present in the middle of the weld. The dense distribution of pores tended to allow He to leak, leading to an increased leakage rate.

This paper focuses on the welding of the sealing device for the X-ray counterpart collection device. Laser welding was used to join 316L stainless steel and 4J29 Kovar alloy, and the effect of heat input on the weld morphology, microstructure, mechanical properties, and sealing performance was investigated. The following conclusions were drawn.

1. A ring-shaped welding path was employed during the testing, and the simulation results showed that the inside of the weld received more heat input than the outside, causing a difference in profile shape.

2. The continuous laser welding improved element distribution in the weld and produced an austenitic crystal structure. Grain epitaxial growth was observed on the Kovar alloy side.

3. The weld tensile test yielded two modes depending on the heat input. For heat inputs greater than 11.88 J/mm, the maximum force of 6065 N and failure location occurred in the Kovar alloy BM zone. Conversely, for heat inputs less than 7.92 J/mm, the failure location was in the weld, and the tensile load was smaller.

4. The operating environment of the device was simulated by creating a vacuum, and the helium leak rate was tested to evaluate the welds' sealing performance. The causes of leak were discussed. Results showed that a heat input greater than 37.50 J/mm resulted in both good sealing performance and improved mechanical properties, making it the optimal process parameter.

Acknowledgements

The work was supported by the National Key R&D Program of China (Grant No. 2021YFE0203500); Guangxi Key Laboratory for Relativistic Astrophysics, Nanning, China; the Special Fund for Local Scientific and Technological Development guided by the Central Government (Grant No. GKZY21195029); Innovation Project of Guangxi Graduate Education (Grant No. JGY2021001). And the leak rate measurements were performed by Xiwen Liu in Guangxi Key Laboratory for Relativistic Astrophysics.

Author contribution

Guangjie Xia: Writing - original draft, Investigation, Validation. Jun Zhou: Conceptualization, Writing - review & editing. Lugui He: Data curation. Canyang Chen: Data curation. Jie Jia: Software. Wei Huang: Data curation. Yu Long: Supervision, Conceptualization.

Competing interests: The authors declare no competing interests.

Tsann-Shyi C, Hsien-Lung T (2007) Wetting and sealing of interface between 7056 Glass and Kovar alloy, Mater. Chem. Phys 104 (2-3) 472-478.
Mai T A, Spowage A C (2004) Characterisation of dissimilar joints in laser welding of steel–kovar, copper–steel and copper–aluminium, Mater. Sci. Eng., A 374 (1-2) 224-233.
Fang Y J, Jiang X S, Mo D F, et al (2019) Microstructure and mechanical properties of the vacuum diffusion bonding joints of 4J29 kovar alloy and 316L stainless steel using pure cobalt interlayer, Vacuum 168 108847.
Ragavendran M, Vasudevan M (2020) Laser and hybrid laser welding of type 316L(N) austenitic stainless steel plates, Mater. Manuf. Processes 35 (8) 922-934.
Yan S H, Meng Z, Chen B, et al (2022) Prediction of temperature field and residual stress of oscillation laser welding of 316LN stainless steel, Opt. Laser Technol 145 107493.
Gao M, Zeng X Y, Yan J, et al (2008) Microstructure characteristics of laser–MIG hybrid welded mild steel, Appl. Surf. Sci 254 (18) 5715-5721.
Yan J, Gao M, Zeng X Y (2010) Study on microstructure and mechanical properties of 304 stainless steel joints by TIG, laser and laser-TIG hybrid welding, Opt. Lasers Eng 48 (4) 512-517.
Lima A S, Nascimento A M, Abreu H F G, et al (2005) Sensitization evaluation of the austenitic stainless steel AISI 304L, 316L, 321 and 347, J. Mater. Sci 40 139– 144.
Mazar Atabaki M, Ma J, Yang G, et al (2014) Hybrid laser/arc welding of advanced high strength steel in different butt joint configurations, Mater. Des 64 573-587.
Yang W X, Xin J J, Fang C, et al (2019) Microstructure and mechanical properties of ultra-narrow gap laser weld joint of 100 mm-thick SUS304 steel plates, J. Mater. Process. Technol 265 130-137.
Hu Y F, Wu L B, Zhou P S, et al (2021) Fiber laser welding of Ti–6Al–4V to Inconel 718 bimetallic structure via Cu/Ta multi-interlayer, Vacuum 192 110461.
Zhang J F, Liu S X, Zhang Y J, et al (2018) Formation mechanism of sealing edge pores for vacuum glazing using laser brazing technique, Vacuum 147 1-7.
Baghjari S H, AkbariMousavi S A A (2104) Experimental investigation on dissimilar pulsed Nd:YAG laser welding of AISI 420 stainless steel to kovar alloy, Mater. Des 57 128-134.
Baghjari S H, Gholambargani M, Akbari Mousavi S A A (2018) Application of the Pulsed Nd: YAG Laser Welding to Investigate the Effect of Laser Beam Position on Weld Characteristics of AISI 420 Stainless Steel to Kovar Alloy, Lasers Manuf. Mater. Process 6 (1) 14-25.
Ardakani H.A., Naffakh-Moosavy H (2019) The effect of pulsed Nd:YAG laser welding parameters on defects of Kovar to AISI 304L dissimilar joint, Opt. Laser Technol 118 62-68.
Cui C Y, Cui X G, Ren X D, et al (2013) Microstructure and microhardness of fiber laser butt welded joint of stainless steel plates, Mater. Des 49 761-765.
Rong Y M, Huang Y, Zhang G J, et al (2016) Laser beam welding of 316L T-joint: microstructure, microhardness, distortion, and residual stress, Int. J. Adv. Manuf. Technol 90 (5-8) 2263-2270.
Ragavendran M, Vasudevan M (2021) Effect of Laser and Hybrid Laser Welding Processes on the Residual Stresses and Distortion in AISI Type 316L(N) Stainless Steel Weld Joints, Metall. Mater. Trans. B 52(4) 2582-2603.
Wu D S, Hua X M, Huang L J, et al (2018) Numerical simulation of spatter formation during fiber laser welding of 5083 aluminum alloy at full penetration condition, Opt. Laser Technol 100 157-164.
Lu F G, Li X B, Li Z G, et al (2015) Formation and influence mechanism of keyhole-induced porosity in deep-penetration laser welding based on 3D transient modeling, Int. J. Heat Mass Transfer 90 1143-1152.
Kan X F, Yin Y J, Yang D C, et al (2021) Micro pool characteristics of 316L and the influence of sulfur during SLM, Opt. Laser Technol 142 107136.
Liang J J, Liu Y S, Li J G, et al (2019) Epitaxial growth and oxidation behavior of an overlay coating on a Ni-base single-crystal superalloy by laser cladding, J. Mater. Sci. Technol 35 (2) 344-350.
Xue J L, Peng P, Guo W, et al (2021) HAZ Characterization and Mechanical Properties of QP980-DP980 Laser Welded Joints, Chin. J. Mech. Eng 34 (1) 1-14.
Almeida T D, Clarke R T G, Souza C J H, et al (2021) The effect of laser welding on microstructure and mechanical properties in heavy-gage press hardening steel alloys, Mater. Sci. Eng., A 821 141341.
Mostaan H, Saeedpour P, Ahmadi H, et al (2021) Laser welding of dual-phase steels with different silicon contents: Phase evolutions, microstructural observations, mechanical properties, and fracture behavior, Mater. Sci. Eng., A 811 140974.
Chen S H, Huang J H, Ma K, et al (2012) Influence of a Ni-foil interlayer on Fe/Al dissimilar joint by laser penetration welding, Mater. Lett 79 296-299.
Dancette S, Fabrègue D, Massardier V, et al (2012) Investigation of the Tensile Shear fracture of Advanced High Strength Steel spot welds, Eng. Fail. Anal 25 112-122.
Chao Yuh J (2003) ltimate Strength and Failure Mechanism of Resistance Spot Weld Subjected to Tensile, Shear, or Combined Tensile/Shear Loads, J. Eng. Mater. Technol 125 125-132.
Guo F Z (2020) Vacuum science and technology, second ed., Liaoning, Dalian University of Technology Press.
Hu Y N, Wu S C, Chen L (2019) Review on failure behaviors of fusion welded high-strength Al alloys due to fine equiaxed zone, Eng. Fract. Mech 208 45-71.
Meng W, Xu Z L, Ma Q S, et al (2019) Pulse fiber laser welding of AISI 321-AISI 405 stainless steel thick plates butt joints, J. Mater. Process. Technol 271 214-225.
Zhang C, Yu Y, Chen C, et al (2020) Suppressing porosity of a laser keyhole welded Al-6Mg alloy via beam oscillation, J. Mater. Process. Technol 278 116382.
Han X H, Yang Z B, Ma Y, et al (2020) Porosity distribution and mechanical response of laser-MIG hybrid butt welded 6082-T6 aluminum alloy joint, Opt. Laser Technol 132 106511.

Download PDF

Reviewers agreed at journal
05 Sep, 2023
Reviewers invited by journal
17 Jul, 2023
Editor invited by journal
10 Jul, 2023
Editor assigned by journal
10 Jul, 2023
First submitted to journal
27 Jun, 2023

You are reading this latest preprint version

Investigate the sealing and integrity of the laser welded X-ray counterpart collection device

Status:

Version 1

Abstract

Figures

1. Introduction

2. Experimental procedure

2.1 Materials

2.2 Welding process

2.3 Characterization

2.4 Sealing performance test

2.5 Temperature field simulation

3. Results

3.1 Welding macroscopic geometry

3.2 Microstructure and element distribution

3.3 Microhardness

3.4 Tensile strength

3.5 Sealing performance

4. Discussion

4.1 The effect of heat input on the weld

4.2 The factors affecting the sealing performance

5. Conclusion

Declarations

References

Status:

Version 1