1.Institute of Applied Physics and Computational Mathematics, Beijing 100094, China 2.Graduate School of China Academy of Engineering Physics, Beijing 100088, China
Fund Project:Project supported by the National Natural Science Foundation of China (Grant Nos. 12005023, 11775032, 11875094) and the Foundation of President of China Academy of Engineering Physics (Grant No. YZJJLX2019013)
Received Date:30 November 2020
Accepted Date:31 January 2021
Available Online:25 June 2021
Published Online:05 July 2021
Abstract:The veracity of a low temperature plasma model is limited by the accuracy of the electron transport coefficient, which can be solved by simulating the electron transport process. When simulating the transport properties of electrons, there are a variety of approaches to dealing with the scattering of electrons and energy partition between the primary-electrons and secondary-electrons after electron-neutral particles’ collision. In this paper used is a model based on the Monte Carlo collision method to investigate the influence of scattering method and energy partition method on the electron transport coefficient. The electron energy distribution function, electron mean energy, flux mobility and diffusion coefficients, as well as the Townsend ionization coefficients are calculated in the hydrogen atom gas under a reduced electric field from 10 to 1000 Td. The calculation results show that the influence of the isotropic scattering assumption on the electron transport coefficients increases with reduced electric field increasing. However, even under a relatively low reduced electric field (10 Td), the calculated mean energy, flux mobility, and flux diffusion coefficient of electrons under the assumption of anisotropic scattering are 39.68%, 17.38% and 119.18% higher than those under the assumption of the isotropic scattering. The different energy partition methods have a significant influence on the electron transport coefficient under a medium-to-high reduced electric field (> 200 Td). Under a high electric field, the mean energy, flux mobility and flux diffusion coefficient calculated by the equal-partition method (the primary and secondary electrons equally share the available energy) are all less than the values from the zero-partition method (the energy of secondary-electrons is assigned to zero). While the change of Townsend ionization coefficient with reduced electric fields shows a different trend. The electron transport coefficient obtained by the Opal method lies between the values from the equal-partition method and the zero-partition method. In addition, considering the anisotropic scattering, the influence of energy partition method on the transport coefficient is higher than that under the assumption of isotropic scattering. This study shows the necessity of considering the anisotropic electron scattering for calculating the electron transport coefficient, and special attention should be paid to the choice of energy partition method under a high reduced electric field. Keywords:electron scattering/ energy partition/ transport coefficient/ Monte Carlo collision
设置背景氢原子气体的压强为105 Pa, 温度为293.15 K. 忽略空间电荷对场的影响, 约化电场($ E/{n}_{0} $, $ {n}_{0} $为氢原子气体密度)范围为10—1000 Td. 模拟时间步长$\Delta t = 5 \times {10^{ - 14}}\;{\rm{s}}$, 电子初始能量为1 eV, 速度满足麦克斯韦分布. 使用非固定电子个数的方法, 初始时刻模拟电子个数为50000. 电子散射有各向同性和各向异性两种, 能量分配有均分法、零分法和Opal法三种, 其中Opal法中的能量分配系数近似取B = 8 eV[26]. 电子和氢原子弹性和电离碰撞中电子散射采用相同方式, 碰撞的截面数据从IAEA数据库[27]中获得, 如图1所示. 图 1 电子与氢原子弹性和电离碰撞的截面数据 Figure1. Elastic and ionization cross sections between electron and hydrogen atoms.
-->
3.1.MCC模型和BOLSIG+软件结果对比
BOLSIG+软件[28]基于两项近似法求解电子玻尔兹曼方程, 是一种广泛使用的电子输运系数求解工具. 在MCC模型中设置电子散射为各向同性, 并在BOLSIG+软件使用与之相同的截面数据和能量分配方式(均分法). 对比二者计算得出的电子能量分布函数及输运系数, 来检验本模型的准确性. 本模型和BOLSIG+软件在约化电场为10, 100, 500和1000 Td时的电子能量分布函数如图2所示. 随着电场增大, 高能电子的概率随之增加, 曲线整体向右平移. 在约化电场为10和100 Td时, 二者得到的电子能量分布函数整体符合, 当约化电场增大到500 Td后, 则呈现出较大差异. 在高能尾部, 本模型对应的EEDF有一定的波动, 这是MCC方法模拟中粒子数目有限, 对高能电子取样不足所致. 由于波动出现在EEDF量级很小的地方, 因此对电子输运系数的计算影响不大. 图 2 MCC模型和BOLSIG+软件在约化电场为10, 20, 100和1000 Td计算出的EEDF对比(电子散射各向同性, 能量均分). 实线和虚线分别代表MCC模型和BOLSIG+软件的计算结果 Figure2. Comparison of EEDF calculated by MCC model and BOLSIG+ software with isotropic scattering and equally energy partition under the reduced fields of 10, 20, 100 and 1000 Td. Dashed lines from MCC, and solid lines from BOLSIG+ software.
电子平均能量、电离系数、通量迁移率和通量扩散系数如图3所示. 可以看出, 电子的平均能量、通量迁移率和通量扩散系数与电场正相关, 而电离系数随着电场增加, 呈先增后减的变化趋势. 原因是当电场增强时, 电子获得更多能量, 沿电场方向速度增加, 迁移率和扩散系数变大. 而当电子能量很大时, 电离碰撞截面变小(如图1), 导致电离碰撞概率降低, 因此在高电场下电离系数减小. 对比图2在约化电场小于200 Td时, 两种方法的计算结果相符合. 随着电场增加, 二者差距逐渐增大. 当约化电场为1000 Td时, 本模型计算的汤森电离系数比BOLSIG+软件计算的结果低12.83%, 而通量迁移率、电子平均能量和扩散系数则分别高7.57%, 12.28%和86.04%. 高电场下的差异是由于电子速度分布呈强烈的各向异性, 两项近似的有效性不再满足, 导致BOLSIG+软件的计算结果有很大误差[9]. 文献[4]中, 当约化电场大于200 Td后, 也表明两项近似计算得到的电子通量迁移率和通量扩散系数偏小, 与图3(b)和图3(d)显示的结果一致. 图 3 MCC模型和BOLSIG+软件计算出的电子输运系数对比(电子散射各向同性, 均分法) (a)电子平均能量; (b)电离系数; (c)横向扩散系数; (d)迁移系数. 橙色实线和绿色虚线分别为使用BOLSIG+软件和MCC法计算的结果 Figure3. Comparison of electron transport coefficients calculated by MCC model and BOLSIG+ software with isotropic scattering and equally energy partition: (a) Mean energy; (b)Townsend ionization coefficients; (c) flux diffusion coefficients; (d) flux mobility. Dashed-orange lines from MCC, and solid-green lines from BOLSIG+ software.
23.2.电子散射对输运系数的影响 -->
3.2.电子散射对输运系数的影响
当电子能量很高时其散射呈强烈的各向异性, 随着电子能量降低逐渐趋于各向同性[23]. 为了研究不同电子散射对电子输运系数的影响, 根据2.4节的模拟参数, 采用Opal分配方式, 依次设置电子散射为各向同性和各向异性, 模拟电子的输运过程. 统计约化电场10, 20和100 Td时的EEDF和电子沿电场方向速度(${v_z}$)的概率函数(EVPF, 由${\rm{d}}{v_z}$内的电子个数${\rm{d}}{N_z}$和总电子数${N_{\rm{e}}}$之比$ {\rm{d}}{N}_{z}/({N}_{\rm{e}}{\rm{d}}v_{z})$来确定), 如图4所示. 图4(a)是两种散射方式下的EEDF, 对高能量区域的电子, 在10, 20和100 Td时, 各向异性散射条件下的EEDF均高于各向同性. 图4(b)是电子沿电场方向速度的概率分布函数, 近似满足高斯分布. 对比同电场下不同散射方式的EVPF, 可以发现考虑各向异性散射在高电子速度的概率更大. 在约化电场为100 Td时, EVPF已开始偏离高斯分布, 因而此时的电子速度分布函数不再满足各向同性, 这也说明了两项近似法仅适用低电场的求解. 图 4 不同散射条件下电子能量分布函数和沿电场方向的速度概率函数在约化电场为10, 20和100 Td时的值 (a)电子能量分布函数; (b)沿电场方向电子速度概率函数. 图中实线为各向异性散射, 虚线为各向同性散射 Figure4. (a) Electron energy distribution function and (b) the probability function of velocity along the electric field direction with different scattering for reduced fields 10, 20 and 100 Td. Dashed lines from isotropic scattering, and solid lines from anisotropic scattering.
由图4可知, 即使在低电场下, 不同电子散射方式对电子能量分布函数和速度概率函数均有一定的影响, 因此各向同性的假设会造成电子输运系数的计算误差. 不同散射下的电子平均能量、通量扩散系数、电离系数和通量迁移率随约化电场变化的关系如图5所示. 图5(a), 和图5(c)分别为电子的平均能量和通量扩散系数, 二者具有相似的变化趋势. 在各向异性散射下, 电子以前向散射为主[23], 相比各向同性可以从电场中获得更多的能量, 即${\langle \varepsilon \rangle _{{\rm{ani}}}} > {\langle \varepsilon \rangle _{{\rm{iso}}}}$. 由于电子通量扩散系数的值随电子平均能量的增加而增大[25], 因此电子散射各向异性时扩散系数比各向同性时大. 根据图4(a)中两种散射方式在10 和20 Td的EEDF, 各向异性散射的高能电子更多, 所以低电场时, 各向异性散射仍有较大的平均电子能量和通量扩散系数. 图 5 电子散射各向同性和各向异性条件下电子输运系数和平均能量的计算结果对比 (a)电子平均能量; (b)电离系数; (c)横向扩散系数; (d)迁移系数. 图中方块代表各向异性, 圆点代表各向同性 Figure5. Comparison of electron transport coefficients calculated assuming isotropic and anisotropic scattering: (a) Mean energy; (b) townsend ionization coefficients; (c) flux diffusion coefficients; (d) flux mobility. The orange rectangle and green circle represent the results assuming the anisotropic and isotropic scattering, respectively.
根据2.4节的模拟参数, 计算不同能量分配模式下的电子输运系数, 得到如图7所示的结果. 可以看出, 不同的能量分配关系对电子平均能量、电离系数、通量迁移率和通量扩散系数的影响主要体现在中高电场区域. 图 7 不同散射和能量分配方式下电子输运系数和平均能量的计算结果对比 (a)电子平均能量; (b)电离系数; (c)横向扩散系数; (d)迁移系数. 图中方块、圆点和三角依次代表均分、Opal和零分法. 实线和虚线分别代表各向异性和各向同性散射 Figure7. Comparison of electron transport coefficients calculated with different scattering and energy partition methods: (a) Mean energy; (b) Townsend ionization coefficients; (c) flux diffusion coefficients; (d) flux mobility. The rectangle, circle, and triangles represent the calculation results using the equal-division, Opal and zero-division method, respectively. Dashed/solid lines are the results assuming the isotropic/anisotropic scattering.
表1约化电场为1000 Td时, 均分法和零分法在不同电子散射下电子输运系数的差异 Table1.Difference between the electron transport coefficients using the equal and zero-division method, assuming anisotropic and isotropic scattering, respectively.