Electronic transport of Lorentz plasma with collision and magnetic field effects
Lv Chong1, Wan Feng1, Jia Mo-Ran1, Li Zi-Liang1, Sang Hai-Bo1, Xie Bai-Song1, 2, †,
College of Nuclear Science and Technology, Beijing Normal University, Beijing 100875, China
Beijing Radiation Center, Beijing 100875, China

 

† Corresponding author. E-mail: bsxie@bnu.edu.cn

Project supported by the National Natural Science Foundation of China (Grant Nos. 11475026 and 11305010) and the NSAF of China (Grant No. U1530153).

Abstract
Abstract

The electronic transverse transport of Lorentz plasma with collision and magnetic field effects is studied by solving the Boltzmann equation for different electron density distributions. For the Maxwellian distribution, it is shown that transport coefficients decrease as Ω increases, Ω is the ratio of an electron’s magneto-cyclotron frequency to plasma collision frequency. It means that the electrons are possible to be highly collimated by a strong magnetic field. For the quasi-monoenergetic distribution with different widths, it is found that the transport coefficients decrease greatly as ɛ̄ decreases. In particular when the width approaches to zero the transverse transport coefficients are hardly affected by the magnetic field and the minimal one is obtained. Results imply that the strong magnetic field and quasi-monoenergetic distribution are both beneficial to reduce the electronic transverse transport. This study is also helpful to understand the relevant problems of plasma transport in the background of the inertial confinement fusion.

1. Introduction

In the inertial confinement fusion (ICF) study,[13] electron transport is an important but difficult problem. For example, for fast ignition (FI) the laser energy is first deposited into fast electrons in the interaction region, which is usually 50 μm–100 μm away from the dense core, and then the energy flux transports into the dense core. Studies[4] showed that electrons with an energy of 1 MeV–3 MeV are optimal for FI. The transport coefficients are usually required for both target design and magnetohydrodynamic codes. Because of its significance, many theoretical and numerical models have been established in search of electron transport features, such as the Spitzer–Härm model,[5,6] the Braginskii model,[7] and the nonlocal electron transport model,[8] and so on.

In laser–plasma interaction physics, in order to obtain accurate electron transport coefficients one must consider two effects, one is associated with the self-generated magnetic field and the other is the plasma collision.[913] Certainly the electron distribution[1416] has also played a key role. The self-generated magnetic field has been studied in laser interacting with plasma over the period of time.[9,1722] Recently, experimental measurements by Tatarakis et al.[23] showed that the self-generated magnetic field, from a few of MGs (1 Gs = 10−4 T) to several hundred of MG, can be generated. The corresponding electron cyclotron frequency ωBe lies within ∼ 1.8 × 1015 Hz–1.8 × 1017 Hz. Such a strong magnetic field is believed to have an important impact on electron transport.

Besides, the collision effect on the transport of electrons has been considered and studied in the high density region.[2426] For example, Kemp et al.[27] showed that the collision effect cannot be neglected when the number density is beyond ∼ 1023 cm−3, due to the collision frequency being close to the plasma frequency. Usually, the number density can be 1022 cm−3–1026 cm−3 in ICF, and the corresponding electron–ion collision frequency νei is ∼ 1.5 × 1015 Hz–3.8 × 1017 Hz, which is obviously comparable to ωBe. The most previous studies about the electronic transverse transport coefficients were focused on the calculations of the weak collision, i.e. ωBeνei, with a Maxwellian distribution function.[28] However, the combination effect of collision and strong magnetic field for different distributions have not been considered, in particular when ωBe is comparable to νei. Therefore, the electronic transverse transport needs to be studied further in these cases.

Plasma consists of electrons and ions. With the condition of assuming the electron temperature equals the ion temperature and the mass of the electrons is much smaller than that of the ions, the speed of electrons is much faster than the speed of the ions. Therefore, it is feasible assuming that electrons are moving while ions are stationary, so that ions are not affected by the collision and electromagnetic field. Besides, for simplicity we just consider the involving problem for the Lorentz plasma,[2830] in which the electron–electron collision νeee4Ne can be ignored in comparison with the electron–ion collision νeie4ziNe, i.e., νeeνei, where Ne denotes the electron density and zi is the atomic number. In order to control the divergence of electrons, Robinson et al.[31] showed that an electron beam could be collimated well by the self-generated magnetic field in the high-Z material area. Without loss of generality, we can assume that the atomic number is zi ≫ 1.[30,32,33]

In this paper, the electronic transverse transport coefficients, such as electric and thermal conductivity, are studied by solving the Boltzmann transport equation for Lorentz plasma with a collision effect in the presence of a magnetic field. The different electrons distribution functions are also considered. The analytical formula is given and the corresponding numerical results are illustrated for the transport coefficients. It is found that the strong magnetic field and quasi-monoenergetic distribution are both beneficial to reduce the electronic transverse transport.

The paper is organized as follows. Section 2 gives the theoretical frame and analytical formula. Section 3 exhibits the results in the presence of both magnetic field and collision for different distribution functions. The results are also analyzed physically. A brief summary and discussion are given in Section 4.

2. Theoretical frame and analytical formula

Assuming that f(r,υ,t) is the electrons distribution function at position r, velocity υ, and time t, then the stationary transport equation of Lorentz plasma in a magnetic field reads[28]

where B is the magnetic field, E is the electric field, p is the electron momentum, −e is the electron charge, c is the light speed in a vacuum, and C(f) is the electron–ion collision term. Furthermore, we assume that

where f0(r,υ) is the stationary equilibrium distribution and δ f(r,υ,t) is the small perturbation, i.e.,

Then the collision integral can be written as[28]

The electron–ion collision frequency νei(v) is usually given by

where m is electron mass, v is magnitude of electron velocity, L(ei) is scattering factor, ze is charge of the ion and Ne is electron number density. In Eq. (1), we can take E = 0 since E only appears in the required current j as a sum formula E + ∇p/eNe,[28] or it can be seen from the view point of the electric field in the moving reference frame (with speed u) as[29]

Besides, since the longitudinal transport coefficients are independent of the magnetic field, all involved quantities for the plasma can be regarded as only depending on the coordinates in the plane perpendicular to B. With all these considerations, one gets the following starting equation

where ∇ is along the direction perpendicular to B. A remarkable and useful solution of Eq. (4) is

where b = B/B is the unit vector of magnetic field B.

In order to calculate the electronic transport coefficients such as the electric conductivity and thermal conductivity, the current density and the generalized Ohm law as well the heat flux and the generalized heat flow law are need. They are respectively given as (refer to Ref. [28])

and

where j is current density, q is heat flux, T is temperature, P is pressure, σ is electric conductivity, α is the thermoelectric coefficient, κ is thermal conductivity, and the coefficients ℜ, ℵ, and ℓ respectively represent the Hall, Nernst, and Leduc–Righi effects.

Before giving the detailed physical analysis and numerical results, we first see the concrete expressions for the current density and the heat flux, respectively. Substituting Eq. (5) into Eq. (6), one can obtain

where we have introduced an average value defined by

Similarly, substituting Eq. (5) into Eq. (8), the heat flux is

For simplicity the following normalized variables are introduced as follows:

where is the magnitude of electron thermal velocity.

3. Physical analysis and numerical results

In Section 2, we have given the theoretical frame and analytical formula. Let us now turn to calculate the transverse transport coefficients for the different distribution functions.

3.1. Results for the Maxwellian distribution function

In the case of a Maxwellian distribution function, f0 can be written as

By substituting Eqs. (10) and (13) into Eq. (7), one can obtain the electric conductivity

where

is defined, and σ is normalized by in the weak collision.[28,29] Similarly, by substituting Eqs. (10) and (13) into Eq. (9), one can obtain the thermal conductivity

where κ is normalized by .

The σ and κ against Ω are plotted in Fig. 1. First from Fig. 1(a) when Ω → 0, one can see that the collision leads to the enhanced electronic mobility and heat conduction due to the large transverse divergency by the electron–ion collision scattering. With the increase of Ω, σ decreases and recovers to the value σ⊥0 when Ω → ∞ due to the strong magnetic field decreasing the impact of the collision. This is not surprising because the magnetic field reduces the mobility of electrons as the Larmor radius becomes smaller than the collision mean free paths,[32,34] which results in that the mobility of an electron is limited. Similarly, since the magnetic field can suppress heat flow by decreasing the heat-carrying electronic mobility, the κ decreases also with the increase of Ω and recovers to the value κ⊥0 of ultra-strong magnetic field when Ω → ∞, as seen in Fig. 1(b). As the reduction of σ and κ, electronic transverse mobility and heat flow are greatly suppressed.

Fig. 1. Transverse electric conductivity σ (a) and thermal conductivity κ (b) against Ω for the case of Maxwellian distribution. The black dashed line (case 1) and red solid line (case 2) represent the condition of weak collision (ωBeνT, see Refs. [28] and [29]) and the full condition without any limitation to the magnetic field and collision.

It is interesting to see the decreasing degree when the magnetic field increases, for example, in the typical case of ICF (ωBeνei), compared to that without magnetic field, the σ reduces about by 55.27%, however, κ reduces dramatically at about by 99.21%. This means that the spontaneous magnetic field can reduce the influence of collision on the electronic transport, in particular the heat flux. This would lead to the electrons being highly collimated by the magnetic field in the transverse direction.

3.2. Results for the quasi-monoenergetic distribution function

It is worth pointing out that in many cases, for example in some regions of ICF or/and laser interacting with a solid target, the distribution of electrons may be non-Maxwellian, therefore, the electronic transverse transport may exhibit different behavior. Now we consider a typical case of the quasi-monoenergetic distribution. For convenience as well as comparison with the previous Maxwellian case, the electrons distribution function is written as

where ɛ is the full width at half maximum (FWHM) or the quasi-monoenergetic degree, and υT is the virtual “electron thermal speed” which is associated to the electrons peak energy Em through and the corresponding virtual “temperature” Tm = 2Em. Here we assume that the directional speed υT is approximately longitudinal, i.e. υT ≈ (0,0,vT), so that the electric field is still ignored. If the collision is not significant, i.e., when Ω → ∞, one has

where

and ɛ̄ = ɛ/vT are defined and introduced, respectively. However, when the electron–ion collision is included, the transverse electric conductivity would become

where

is defined. It is ready to check that in the limit of Ω → ∞, equation (18) recovers to Eq. (17). Thus, it is convenient to get the following normalized σ by σ⊥0 as

Similarly, for the thermal conductivity we can get respectively

and

where both κ⊥0 and κ are normalized by NeTm/T.

The σ and κ dependent of Ω are shown in Fig. 2 for a quasi-monoenergetic distribution with different energy spread width ɛ. From Fig. 2 one can see that there exist the similar effect of magnetic field on lowering the transverse transport coefficients for a certain ɛ̄ as in the case of Maxwellian distribution. Moreover, it is also found that there exist some delicate discrepancies for different distributions. First the transverse transport coefficients decrease rapidly as ɛ̄ decreases, which means that an electron’s motion and heat flow can be suppressed greatly. Second, for the given ɛ̄, as the magnetic field increases, the reduction effect of the magnetic field on the electric transport in the regime of a strong magnetic field (see the lower right curves of Fig. 2(a)), as well heat transport in the regime of a weak magnetic field (see the upper left curves of Fig. 2(b)), would become weaker and weaker. These results may be attributed to the fact that the distribution is far away from the Maxwellian distribution. It is not surprising to understand the lowering transport dependence of decreasing width. Because the narrower width would lead to the decrease of high-energy electrons, for example, in Ref. [33] it is demonstrated that the heat-flow is dominated by a small number of high energy electrons within 3vT–4vT, which would limit the heat flow greatly. Therefore, it concludes that the smaller the ɛ is, the less amount the transverse transport exhibits.

Fig. 2. Transverse electric conductivity σ (a) and thermal conductivity κ (b) against Ω with different width ɛ. The dashed line represents the condition of ɛ̄ → 0.

On the other hand, it is also noted that when ɛ̄ varies from 0.5–0.9, the conductivity can still be affected by Ω at 104–106. This can be explained by the relation between Larmor radius rb and magnetic field B:

where E is the electron energy. We can see that rb is proportional to E while inversely proportional to B. For ɛ̄ varies from 0.5–0.9, which means that there are more electrons in the higher energy region, so that a stronger magnetic field is required to collimate the transverse transport of electrons.

It is worth noting that for our special interested cases of ICF when ωBeνei, σ and κ are also reduced as the ɛ̄ decreases while this influence becomes weaker and weaker. A more interesting aspect is that the transport is hardly affected by the magnetic field when ɛ̄ → 0 for σ (as seen in the dashed lines in Fig. 2(a)) as well for κ when Ω < 1 (as seen in the dashed lines in Fig. 2(a)). This will be studied further in the following subsection and discussed in the final section.

3.3. Results for the Delta distribution function

In the limit of ɛ → 0, quasi-monoenergetic distribution will recover to the Delta one with almost perfect monoenergetic electron density. Thus it is necessary to discuss this special condition in this subsection. First the distribution function is now written as

As mentioned above, in the limit of Ω → ∞, σ and κ are given as

When electron–ion collision is included, one obtains

where all of the coefficients are normalized by e2Ne/T and NeTm/T, respectively.

In the comparison with each other in two cases (the weak collision, i.e., ωBeνei), σ does not change over all the parameter change region of Ω, and κ also hardly changes at the region of Ω < 1 (as seen in the left red line of Fig. 3(b)). The reason for the conductivity insensitivity to the magnetic field is due to the combinational effects of the Hall term and the thermal pressure term. And it is also seen as the result of υTB. Certainly if the higher order is included or there exists the perpendicular directional velocity, the conductivity would not keep constant to the magnetic field. In particular, for example in our interested case of ICF where ωBeνei, the κ is reduced about one half. On the other hand the region hardly affected on the transverse transport κ is now only limited to the parameter changed region of Ω < 1. When Ω > 1, the κ is dramatically reduced similar to the case of weak collision (ωBeνei),[28] due to the κ being always proportional to 1/Ω2 or 1/(Ω + 1)2.

Fig. 3. Same as in Fig. 1 except for the case of Delta distribution.
4. Summary and discussion

In summary, we have investigated the electronic transverse transport with the collision effect in the presence of a magnetic field for different electrons density distributions. According to our analyses, the transverse transport is not only affected by magnetic field and collision, but also modified by distribution functions.

First, for the Maxwellian one, when Ω < 1, the transverse transport is dominated by a collision effect. But with the competition of magnetic field and collision, σ and κ are obviously reduced due to the Larmor radius becoming smaller than the collision mean free paths and magnetic field effect will be dominant at Ω > 1. Therefore, the magnetic field leads to great suppression of the electronic mobility and heat flux.

Second when the distribution function is far away from the Maxwellian, the transport is different from the classical transport theory. In this case, the transverse transport coefficients are significantly modified by a magnetic field as well as the distribution. On the one hand, similar to the Maxwellian one, all of the transverse transport coefficients decrease with the increase of Ω. On the other hand, the electronic mobility and heat flow are suppressed greatly as ɛ̄ decreases. As a result, the smaller the ɛ is, the less amount the transverse transport exhibits. On the other hand, however, with the decrease of ɛ̄, the magnetic field and collision balances with each other in the region of Ω < 1 so that σ and κ change weaker and weaker. Until to the region of Ω > 1, in which κ is rapidly reduced due to the fact that the magnetic field effect is dominated over the collision effect, which destroys the weak balance.

Finally, in the limit of ɛ → 0, the distribution will tend to be the Delta function with almost perfect monoenergetic density distribution. Under this circumstance, the thermal pressure of affecting σ and the Hall effect of affecting σ are canceled with each other, which causes a perfect balance so that the σ keeps a constant for the whole region of Ω. But this behavior is approximately valid to κ only when Ω < 1 and again the magnetic field effect is dominated over the collision effect when Ω > 1 in which the κ decrease rapidly.

In a word, the results studied in this paper indicate that the magnetic field and proper quasi-monoenergetic distribution can effectively reduce the transverse transport and suppress the transverse divergence of electrons. It would be helpful to understand the transport problem of Lorenz plasma as well the fast electrons transport in ICF background. Besides, we consider the first-order approximation in the process of calculation. High-order small quantities may also slightly affect the results, which we will investigate in the future.

Reference
1LindlJ F1998Inertial Confinement FusionNew YorkSpringer-Verlag
2Atzeni SMeyer-Ter-Vehn J2004The Physics of Inertial FusionOxfordOxford Science Press
3Lindl JAmendt PBerger R LGlendinning S GGlenzer S HHaan S WKauffman R LLanden O LSuter L J2004Phys. plasmas11339
4Key M H2008Phys. plasmas11022701
5Cohen R SSpitzer LRoutly P M1950Phys. Rev.80230
6Spitzer LHärm R1953Phys. Rev.89977
7Braginskii S I1966Rev. Plasma Phys.1205
8Luciani J FMora PVirmont J1983Phys. Rev. Lett.511664
9Stamper J APapadopoulos KSudan R NDean S OMclean E ADawson J M1971Phys. Rev. Lett.261012
10Glenzer S HAlley W EEstabrook K GDe Groot J SHaines M GHammer J HJadaud J PMacGowan B JMoody J DRozmus WSuter L JWeiland T LWilliams E A1999Phys. Plasmas62117
11Zheng W ZZhao BHu G YZheng J2015Acta Phys. Sin.64195201(in Chinese)
12Zhang WHu L QSun Y WDing S YZhang Z JLiu S L2014Chin. Phys. B23105201
13Li KLi Y TZhang JYuan X HXu M HWang Z HZhang J2014Acta Phys. Sin.555909(in Chinese)
14Langdon A B1980Phys. Rev. Lett.44575
15Liu S QGuo H M2011Acta Phys. Sin.60055203(in Chinese)
16Huo W YZeng Q H2015Phys. Plasmas22094503
17Davies J RBell A RTatarakis M1999Phys. Rev. E596032
18Cai H BZhu S PChen MWu S ZHe X TMima K2011Phys. Rev. E83036408
19Bake M AXie B SZhang SWang H Y2013Phys. Plasmas20033112
20Zhu S PHe X TZheng C Y2001Phys. Plasmas8321
21Tabak MHammer JGlinsky M EKruer W LWilks S CWoodworth JCampbell E MPerry M DMason R J1994Phys. Plasmas11626
22Wu F JZhou W MShan L QLi FLiu D XZhang Z MLi B YBi BWu BWang W WZhang FGu Y QZhang B H2014Acta Phys. Sin.63094101(in Chinese)
23Tatarakis MGopal AWatts IBeg F NDangor A EKrushelnick KNorreys P AClark E LZepf MEvans R G2002Phys. Plasmas92244
24Sentoku YMima KTaguchi TMiyamoto SKishimoto Y1998Phys. Plasmas54366
25Sentoku YMima KMiyamoto S1999Fusion Eng. Des.44233
26Sentoku YKemp A J2008J. Comput. Phys.2276846
27Kemp A JSentoku YSotnikov VWilks S C2006Phys. Rev. Lett.97235001
28Lifshitz E MPitaevskii L P1999Physical KineticsBeijingWorld Publishing Corporation227231227–31
29Huang Z QDing E J2008Transport TheoryBeijingScience Press438
30Du J L2013Phys. Plasmas20092901
31Robinson A P LSherlock M2007Phys. Plasmas14083105
32Ridgers C PKingham R JThomas A G R2008Phys. Rev. Lett.100075003
33Bissell J JRidgers C PKingham R J2013New J. Phys.15025017
34Cai H BZhu S PHe X T2013Phys. Plasmas20072701