Abstract
Nanometer-(70‒80 nm) and micrometer-sized (500‒600 nm) rare-earth (RE) oxides (La2O3, Y203) were separately mixed with tungsten powder by a mechanical alloying method. Afterwards, the W-1.5La2O3-0.1Y2O3-0.1ZrO2 (wt%) was prepared by cold isostatic pressing, medium-frequency induction sintering, rotary forging, and drawing. Then we performed tungsten argon arc welding (TIG) under the same welding current for 0.5, 1, and 2 h on the cathode samples containing, separately, nanometer- and micrometer-sized RE oxides. Results show that the sample with nanometer-sized RE oxides exhibits higher working stability during the welding process, and the burning loss is decreased by nearly 85.4%. Moreover, with prolonging the working time, the aggregation degree of RE oxides in different regions of the tip significantly increases. Combined with the temperature simulation by COMSOL Multiphysics, we found that the diffusion activation energy of the second phase is decreased by nearly 34%. This is because the finer second phase effectively controls the evolution of the tungsten matrix structure, thus preserving many grain boundaries as channels and promoting the diffusion of active substances.
Tungsten cathodes are one of the key heat source materials widely used in argon arc welding, plasma welding, spraying, cutting technology, and metallurgical industr
At high temperatures, the used cathode undergoes fast evaporation of the active RE material on the surface. If the RE elements inside the cathode cannot diffuse and migrate to the working surface in a given time, the working stability of cathode is greatlyed affected. Thus, it is necessary to control the growth rate of tungsten grains at high temperatures so that more grain boundaries can be preserved as diffusion channels to improve the diffusion mobility of RE. The finely dispersed second phase can hinder the grain growth of the matrix during high-temperature processing, so the average size of tungsten grains can be significantly reduce
In the present study, we used powder metallurgy to separately add nanometer- and micrometer-sized RE oxides as the second phase to synthesize W-La2O3-Y2O3-ZrO2 with fine and dispersed second-phase particles in the finished cathode. The tungsten argon arc welding (TIG) was performed on the two kinds of prepared cathodes doped with different particle sizes of RE oxides, and the effect of adding RE oxides with different particle sizes on the microstructure and working performance was discussed.
The purity of tungsten powder (Ganzhou Hongfei Tungsten and Molybdenum Company) with the average particle size 1.5 μm and nanometer- and micrometer-sized RE oxides powders (Xuzhou Boguan Welding Company) was 99.9%. Two composite powders were prepared, and the average particle size of micrometer-sized RE oxides and nanometer-sized RE oxides was 500‒600 nm and 70‒80 nm. ZrO2 (Beijing Jia Shi Teng Trading Company) had a purity of 99% and an average particle size of 20 μm.
The tungsten powder, RE oxides, and zirconium hydride were used as raw materials to prepare W-1.5La2O3-0.1Y2O3-0.1ZrO2 (wt%). After ball milling and ultrasonic mixing, the composite powders, with micrometer-sized (W-1) and nanometer-sized (W-2) RE oxides, exhibit irregular polyhedral morphology, as shown in
Fig.1 SEM morphologies of W-La2O3-Y2O3-ZrO2 powder samples: (a) W-1 and (b) W-2
Compared to the W-1 sample with micrometer-sized RE oxides, the W-2 sample with nanometer-sized RE oxides exhi-bits lower porosity, and its bulk density and tap density are greatly improved. The specific results are shown in
| Sample | Powder mixing method | Porosity | Fisher particle size/μm | Bulk density/g·c | Tap density/g·c |
|---|---|---|---|---|---|
| W-1 | Ball-milling | 0.727 | 1.29 | 2.28 | 4.85 |
| W-2 | Ultrasonic-wave | 0.645 | 1.62 | 3.33 | 6.67 |
Fig.2 SEM images of sintered samples: (a) W-1 and (b) W-2
| Point | W | La | Y | Zr | O | C |
|---|---|---|---|---|---|---|
| 1 | - | 82.56 | 3.66 | 5.74 | 8.04 | - |
| 2 | 96.49 | - | - | - | - | 3.51 |
| 3 | 4.86 | 71.49 | 4.42 | 3.23 | 16.00 | - |
| 4 | 8.90 | 76.40 | 3.52 | 2.24 | 8.93 | - |
| 5 | 96.37 | - | - | - | 0.61 | 3.02 |
After multiple passes of swaging, drawing, and high-frequency annealing depending on the state of the sample during the process, a tungsten cathode with a diameter of 3 mm was finally prepared. The microstructure of the cathode sample is shown in
Fig.3 Microstructure of multi-pass machining of 3-mm diameter tungsten cathode
The cathode was sharpened and assembled into a welding torch. The welding device is shown in
Fig.4 Schematic diagram of TIG device (a); detail view (b); preparation of the metallographic sample with the axial section of the tip (c); metallographic sample (d)
As the working temperature of the cathode increases, the cathode tip melts and deforms or even burns out. During the heat transfer from the tip to the cathode body, the burnout area gradually migrates from the cathode tip to the base and becomes larger.
Fig.5 Burnout macromorphologies of the electrode tip with different welding time: (a) W-1, 1 h; (b) W-1, 2 h; (c) W-2, 1 h; (d) W-2, 2 h;
(e) without welding
Fig.6 Mass loss comparison of the TIG-welded samples
Fig.7 Metallographic structures of TIG-welded samples: (a) W-1, welding time 1 h; (b) W-1, welding time 2 h; (c) W-2, welding time 1 h; (d) W-2, welding time 2 h; (e) SEM image of the microstructure of the W-2 sample; (f) EDS mappings of the white square area for W-2 sample marked in Fig.7e
After multi-pass processing in the early stage, various defects and internal stress may form and gradually accumulate inside the material, making it thermodynamically unstable. During the TIG process with the prepared cathode, a high-frequency voltage spark induces discharge between the cathode and the base metal, causing space ionization and gradually increasing the loop current in the discharge range of the arc. In this process, the cathode collides with positive and negative ions to obtain energy so that the surface temperature of the electrode rises rapidly, and the temperature inside the electrode gradually increases. When atoms have sufficient mobility, they migrate to the equilibrium position with lower energy so that the internal stress can be relaxed and the stored energy is gradually released. If the area where evaporation and melting occur due to high temperatures is temporarily ignored, the effect of high temperatures gradually reduces or eliminates the deformation damage caused by the pre-processing of the inner area, so the working process of the electrode can be regarded as a kind of “annealing” treatment. With the release of the stored deformation energy, the matrix changes from a deformed state to a low-energy state and from an unstable state to a stable state.
Considering the metallographic structure, the sample can be roughly divided into three regions. According to its shape and distribution, the second phase in region C is strip-shaped or streamlined; region B is close to the working surface, where the second phase is spherical or quasi-spherical; in region A, which is the closest area to the arc center, the second phase is massively aggregated with an extremely irregular shape, and the migration along the grain boundary can be observed. Temperature gradients in different regions of the tip are important factors affecting the material's evolution. For the low degree of heating in region C, the grains are recovered from fibrous to flat, and still distributed along the machining direction. Due to the high temperature of region B and region A, different degrees of recrystallization and grain growth occur after recovery. To further analyze the evolution of the material, we need to understand the temperature distribution at the tip.
According to the TIG device shown in
Fig.8 Geometric model of TIG welding device
During the arcing process, the temperature inside the tungsten cathode depends on the Joule heat generated by the external direct current, the heat conduction inside the cathode, and the convective heat transfer between the arc and the electrod
The differential equation governing the arc and the cathode is as follows.
Mass conservation equation:
| (1) |
Energy conservation equation:
| (2) |
| (3) |
Momentum conservation equation:
| (4) |
| (5) |
| F=J×B | (6) |
Current-Magnetic field coupling equation:
| Jmf(ec)=–σ(Eec+Emf) | (7) |
| Emf=– | (8) |
| Eec= | (9) |
| (10) |
These equations describe the physical conditions of the arc and the cathode, where ρ is the density, t is the time, u is the velocity vector, Cp is the isobaric heat capacity, T is the absolute temperature, k is the thermal conductivity, Q is the heat source, kB is the Boltzmann constant, q is the electron charge, J is the current density vector, E is the electric field strength vector, Qrad is the total volume radiation coefficient, ι is the unit matrix vector, μ is the dynamic viscosity, F is the Lorentz force, and B is the magnetic flux density vector of the self-induced magnetic field. In the subscript of Jmf(ec), mf represents the magnetic field and ec represents the electric field. σ is the electrical conductivity, A is the magnetic vector potential, and V is the electric potential. Each vector item includes its r-component, phi-component, and z-component.
The experimental parameters required for the input of the model include the electrode size, external welding current (A), the flow rate of protective argon gas (m·
In the geometric model shown in
Fig.9 Calculation results of TIG welding temperature field
In the calculation process, the cathode and the arc temperature rapidly increase within 400‒600 ms after the arc ignition and reach the temperature peak, after which the temperature remains almost constant during the welding process. During the arcing process, the RE elements migrate and diffuse to the surface of the tip, and the amount of migrated RE elements at different positions of the tip are different, which leads to differences in the RE active layer on the surface of the tip, further causing regional differences in surface work function and affecting the electron emission properties of the cathode. High external energy is required to maintain the stable arc combustion when the electron emission capacity is low, so there may be abnormally high or low temperatures in this area. Therefore, during the working process for the same cathode, the temperature of the tip surface does not exhibit a uniform and smooth transition. In actual cases, the temperature of a certain area on the tip surface may suddenly increase, affecting the heat transfer field inside the cathode.
To further revise the model and to approach the real behavior, the factors of the surface work function change are added to the model, referring to Ref.[
Fig.10 Simulation results of local temperature field at the cathode tip
Masao et a
During the welding process, affected by numerous external factors, the welding head fuses to varying degrees, and we cannot get the blown part. Therefore, when calculating the cross-sectional area of the second phase, according to the tip morphology and the metallographic structure of the sample after welding, we selected the area at a certain distance from the tip, showing three particular regions for analysis.
Fig.11 SEM images and schematic diagram of microstructure of the cathode tip: (a‒d) W-2 and (e‒f) W-1
| Welding time/h | Equivalent size of W-1/μm | Equivalent size of W-2/μm | |||||
|---|---|---|---|---|---|---|---|
| Region A | Region B | Region C | Region A | Region B | Region C | ||
| 0 | 0.320 295 | 0.331 894 | 0.322 923 | 0.310 463 | 0.302 122 | 0.315 468 | |
| 0.5 | 0.352 529 | 0.342 841 | 0.323 291 | 4.109 051 | 0.567 767 | 0.289 106 | |
| 1 | 1.910 107 | 1.047 857 | 0.344 485 | 5.654 024 | 0.944 282 | 0.317 265 | |
| 2 | 2.115 835 | 1.124 181 | 0.353 517 | 10.859 804 | 1.381 681 | 0.354 714 | |
During the welding process, the evaporation and consumption of RE elements in the high-temperature area decrease their concentration on the surface, resulting in many crystal defects, such as vacancies, and enhancing the diffusion driving force of RE elements. At the same time, the loss of RE elements at the grain boundary enlarges the grain boundary area and increases the grain boundary energy. Under the com-bined action of these factors, a high concentration gradient is formed between the surface and the interior, and the RE elements diffuse from the interior to the surface. However, it requires a certain energy barrier, which is generally referred to as diffusion energy, Q, and only those atoms with energy higher than the diffusion energy can migrate and diffuse to the cathode surface. It is known that the kinetic energy of atoms follows the Maxwell-Boltzmann distribution, and that temperature greatly influences the diffusion of metal atoms, the migration and growth of the second phase, and the migration of grain boundaries, so the diffusion rate of active substances should be proportional to exp(–Q/kT
| Dt-D0= | (11) |
| K=Aexp(–Q/kT) | (12) |
where t is the working time, Dt is the equivalent diameter of the second phase at t, n is the sensitivity coefficient, A is a constant related to the temperature and the concentration of active substances in the material, and k is the Boltzmann constant. Taking the logarithm on both sides of
| ln(Dt–D0)=nlnt+lnK | (13) |
It can be found that ln(Dt–D0) is proportional to lnt, so
Fig.12 ln(D-D0) -lnt curves of W-1
Fig.13 ln(D-D0) -lnt curves of W-2
| Region | W-1 | W-2 |
|---|---|---|
| A | -38.357 | -18.060 |
| B | -40.802 | -22.662 |
| C | -45.004 | -22.511 |
Taking the logarithm for
| lnK=–Q/kT+lnA | (14) |
Likewise, lnK is proportional to (–1/T). Combined with the temperature of each region obtained from the COMSOL-simulated temperature field, the results of data fitting are shown in
Fig.14 lnK-
345.2% for W-1 and W-2 samples, respectively. The above two regions exhibit no obvious difference from the statistical results, but in region A, the second phase of the W-2 sample is increased by nearly 1200% after 0.5 h of operation, and 3400% after 2 h of operation. The growth of the W-1 sample is only 12.9% and 580.6% after 0.5 and 2 h of operation, respectively.
The stable balance between the diffusion mobility of RE to the cathode surface and the evaporation rate is the key to maintain the high electron emission capacity of the cathode, and grain boundaries are the main channels for diffusion and migration. In region C and or the area farther from the tip, the average equivalent size of the second phase remains approximately 0.3 μm, and the effective pinning action keeps the recovery and recrystallization degree of the matrix structure at a low level, enabling the subsequent migration and diffusion of RE elements. In the A and B regions, the high temperature provides energy for the material evolution. On the one hand, the dispersed second phase increases the total interfacial energy inside the system. To alleviate this unstable state, the second phase tends to coarsen so that large particles grow while small particles dissolv
Region A is the frontier of electron emission, and the difference in the morphology of RE oxides in this region significantly affects the material's performance. As mentioned above, the tip grains of the W-1 sample grow sharply due to the absence and uneven distribution of the second phase, as shown in the white boxed area in
Fig.15 Extra-large grains at the tip of the cathode of W-1
From Fig.
1) W-1.5La2O3-0.1Y2O3-0.1ZrO2 cathode samples were prepared with nanometer- and micrometer-sized rare-earth (RE) oxides (La2O3, Y2O3) and tungsten powder as starting materials. TIG welding was performed on the samples under the same current. During electron emission, the second phase gradually changes its morphology from the initial diffusing appearance to a concentrated area at the emission tip. The variations in size and morphology affect the evolution of different regions (A, B, C) of the material.
2) The second phase at the tip of the sample doped with micrometer-sized RE oxides is rapidly consumed, while the internal active substances cannot be replenished in time, so the secondary recrystallization of the structure is caused, and the reduction of grain boundary worsens the diffusion ability of RE elements.
3) In the sample doped with nanometer-sized RE oxides, the second phase at the tip almost completely occupies the grain boundary position, exhibiting an extremely irregular dendritic morphology, which effectively curbs the appearance of coarse grains while retains a large number of grain boundaries. It creates favorable conditions for the diffusion of RE elements.
4) Combined with the simulated temperature field, it is found that the diffusion activation energy of the second phase is decreased by nearly 34% during the welding process, which improves the electron emission performance of the cathode. As a result, the cathode burn-off is decreased by nearly 85.4%.
References
Kim Youngmoo, Hong Moon-Hee, Lee Sung Ho et al. Metals and Materials International[J], 2006, 12(3): 245 [Baidu Scholar]
Mabuchi M, Okamoto K, Saito N et al. Materials Science & Engineering A[J], 1997, 237(2): 241 [Baidu Scholar]
Daoush W M R, Elsayed A H A, Kady O A G E et al. Metallurgical and Materials Transactions A[J], 2016, 47: 2387 [Baidu Scholar]
Chen Y, Wu Y C, Yu F W et al. International Journal of Refractory Metals and Hard Materials[J], 2007, 26(6): 525 [Baidu Scholar]
Chen Y, Wu Y C, Yu F W et al. Rare Metal Materials and Engineering[J], 2007, 36(5): 822 (in Chinese) [Baidu Scholar]
Lang E, Schamis H, Madden N et al. Journal of Nuclear Materials[J], 2021, 545: 152 613 [Baidu Scholar]
Yang X B, Tang H P, Zhu J L et al. Rare Metal Materials and Engineering[J], 2022, 51(8): 3025 [Baidu Scholar]
Guo X C, Liu Z J, Li H B et al. Rare Metal Materials and Engineering[J], 2021, 50(3): 902 [Baidu Scholar]
Zheng X, Bai R, Wang D H et al. Rare Metal Materials and Engineering[J], 2011, 40(10): 1871 (in Chinese) [Baidu Scholar]
Yang Z M, Wang W, Ogawa Yoji et al. Rare Metal Materials and Engineering[J], 2005, 34(11): 1806 (in Chinese) [Baidu Scholar]
Wang J F, Zuo D W, Zhu L et al. International Journal of Refractory Metals and Hard Materials[J], 2018, 71: 167 [Baidu Scholar]
Zhao M Y, Zhou Z J, Ding Q M et al. International Journal of Refractory Metals and Hard Materials[J], 2015, 48: 19 [Baidu Scholar]
Zong X, Lu Y X. Mining & Processing Equipment[J], 2010, [Baidu Scholar]
38(22): 1 [Baidu Scholar]
Du Y B, Liu Z G, Cui S. Chinese Journal of Rare Metals[J], 2011, 35(3): 378 (in Chinese) [Baidu Scholar]
Raabe Dierk. Physical Metallurgy[M]. Amsterdam: Elsevier, 2014: 2291 [Baidu Scholar]
Zan X, Gu M, Wang K et al. Fusion Engineering and Design[J], 2019, 144: 1 [Baidu Scholar]
Wang F Z, Zhang H, Ding B J et al. Materials Science & Engineering A[J], 2002, 336(1): 59 [Baidu Scholar]
Guo X C, Liu Z J, Li H B et al. Rare Metal Materials and Engineering[J], 2022, 51(5): 1879 [Baidu Scholar]
Zhang H, Yang Z M, Ding B J et al. IEEE Transactions on Components and Packaging Technologies[J], 1999, 22(3): 455 [Baidu Scholar]
Zhang H, Chen X F, Yang Z M et al. Materials Letters[J], 1999, 38(6): 401 [Baidu Scholar]
Zhuge F, Ye Z Z, Wang F Z et al. Materials Letters[J], 2003, [Baidu Scholar]
57(19): 2776 [Baidu Scholar]
Zheng H J, Huang J G, Wang W et al. Electrochemistry Communications[J], 2005, 7(10): 1045 [Baidu Scholar]
Wang F Z, Zhuge F, Zhang H et al. Materials Research Bulle- [Baidu Scholar]
tin[J], 2003, 38(4): 629 [Baidu Scholar]
Hussein Gamal A M, Ismail Hamdy M. Colloids and Surfaces A: Physicochemical and Engineering Aspects[J], 1995, 99(2):129 [Baidu Scholar]
Kim Youngmoo, Lee Kyong Ho, Kim Eun-Pyo et al. International Journal of Refractory Metals and Hard Mater- [Baidu Scholar]
ials[J], 2009, 27(5): 842 [Baidu Scholar]
Wang L Y, Yu Y G, Peng Y et al. Nonferrous Metals Engine- ering[J], 2019, 9(10): 21 [Baidu Scholar]
Liu R, Zhou Y, Hao T et al. Journal of Nuclear Materials[J], 2012, 424(1‒3): 171 [Baidu Scholar]
Liu T, Fan J L, Huang B Y et al. Rare Metal Materials and Engineering[J], 2010, 39(2): 314 (in Chinese) [Baidu Scholar]
Veleva L, Schaeublin R, Battabyal M et al. International Journal of Refractory Metals and Hard Materials[J], 2015, 50: 210 [Baidu Scholar]
Eric Lang, Nathan Madden, Charles Smith et al. International Journal of Refractory Metals and Hard Materials[J], 2018, 75: 279 [Baidu Scholar]
Makoto Fukuda, Akira Hasegawa, Shuhei Nogami et al. Journal of Nuclear Materials[J], 2014, 449: 213 [Baidu Scholar]
Xie Z M, Zhang T, Liu R et al. International Journal of Refractory Metals and Hard Materials[J], 2015, 51: 180 [Baidu Scholar]
Jnsson P G, Szekely J, Choo R T C et al. Modelling and Simulation in Materials Science and Engineering[J], 1994, 2(5): 995 [Baidu Scholar]
Lowke JJ, Morrow R, Haidar J. Journal of Physics D: Applied Physics[J], 1997, 30(14): 2033 [Baidu Scholar]
Lu F G, Yao S, Lou S N et al. Computational Materials Sci- ence[J], 2004, 29(3): 371 [Baidu Scholar]
Massoud Goodarzi, Roland Choo, James M Toguri. Journal of Physics D: Applied Physics[J], 1997, 30: 2744 [Baidu Scholar]
Haidar J. Journal of Applied Physics[J], 1998, 84(7): 3518 [Baidu Scholar]
Zhu P Y, Lowke JJ, Morrow R. Journal of Physics D: Applied Physics[J], 1992, 25(8): 1221 [Baidu Scholar]
Cram L E. Journal of Physics D: Applied Physics[J], 1983, [Baidu Scholar]
16(9): 1643 [Baidu Scholar]
Masao Ushio, Alber A Sadek, Fukuhisa Matsuda. Plasma Chemistry and Plasma Processing[J], 1991, 11(1): 81 [Baidu Scholar]
Wang H T, Fang Z Zak, Hwang Kyu Sup. Metallurgical and Materials Transactions A[J], 2011, 42(11): 3534 [Baidu Scholar]
Burke J E, Turnbull D. Progress in Metal Physics[J], 1952, 3: 220 [Baidu Scholar]
Du Z Y, Lv Y Q, Han Y et al. Tungsten[J], 2021, 2(4): 371 [Baidu Scholar]