2025 Volume 34 Issue 6
Article Contents

Yanpeng Wang, Shichao Wu, Peifeng Fan. General relaxation model for a homogeneous plasma with spherically symmetric velocity space[J]. Chinese Physics B, 2025, 34(6): 065201. doi: 10.1088/1674-1056/adc2de
Citation: Yanpeng Wang, Shichao Wu, Peifeng Fan. General relaxation model for a homogeneous plasma with spherically symmetric velocity space[J]. Chinese Physics B, 2025, 34(6): 065201. doi: 10.1088/1674-1056/adc2de

General relaxation model for a homogeneous plasma with spherically symmetric velocity space

  • Received Date: 20/01/2025
    Accepted Date: 11/03/2025
    Available Online: 01/06/2025
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Figures(2)

Article Metrics

Article views(640) PDF downloads(1) Cited by(0)

Access History

General relaxation model for a homogeneous plasma with spherically symmetric velocity space

Abstract: A kinetic moment-closed model (KMCM), derived from the Vlasov–Fokker–Planck (VFP) equation with spherically symmetric velocity space, is introduced as a general relaxation model for homogeneous plasmas. The closed form of this model is presented by introducing a set of new functions called R function and R integration. This nonlinear model, based on the finitely distinguishable independent features (FDIF) hypothesis, enables the capture of the nature of the equilibrium state and non-equilibrium state. From this relaxation model, a general temperature relaxation model is derived when the velocity space exhibits spherical symmetry, and the general characteristic frequency of temperature relaxation is presented.

1.   Introduction
  • The evolution of fusion plasmas over time can be effectively depicted by the Vlasov–Fokker–Planck[1,2] (VFP) equation in combination with the Maxwell equations. Nevertheless, except for a few specific cases such as thermodynamic equilibrium,[3,4] the VFP equation typically exhibits significant nonlinearity.[5] Solving the VFP equation, either analytically or numerically, often requires certain assumptions to simplify the equation.[6,7] Furthermore, numerical solutions to the VFP equation are generally classified into two approaches: direct discretization methods[8,9] and moment methods.[6,10]

    The moment methods typically transform the VFP equation into a set of nonlinear equations, which often lack closure[5,6] and demand some assumptions to close the system. Traditionally, the near-equilibrium assumption is the widely adopted and effective approach in plasma physics, such as Grad’s moment theory[10] derived from the VFP equation based on Hermite polynomial expansion (HPE), as well as the traditional low-order moment theories which are based on the Chapman–Enskog expansion[11–13] (essentially a Taylor expansion), including the magnetohydrodynamic equations, two-fluid equations, and so forth.[14] However, the near-equilibrium assumption enforces the expansions of the distribution function into an orthogonal series around a local Maxwellian, also leading to inadequate convergence in highly non-Maxwellian system,[7] including both the Grad and Chapman–Enskog approaches.

    In Refs. [3,6,9], another assumption, namely, finitely distinguishable independent features (FDIF) hypothesis, has been introduced to enclose the moment equations. Based on FDIF hypothesis, a novel framework[15] including a meshfree approach[9] and a moment approach[3,6] is provided for addressing the nonlinear simulation in fusion plasmas. In this framework, spherical harmonics expansion (SHE) is utilized for the angular coordinate in velocity space, and the King function expansion (KFE) method is employed for the speed coordinate, which has been demonstrated to be a moment convergent method.[9] Reference [3] has presented a relaxation model for homogeneous plasmas in the scenario with shell-less spherically symmetric velocity space. In this study, a further development of the relaxation model for general spherically symmetric velocity space is presented, which may incorporate shell structures.[3,16] Based on this general nonlinear relaxation model, a general temperature relaxation model will also be presented when the velocity space exhibits spherical symmetry.

    The following sections of this paper are organized as follows. Section 2 provides an introduction to the VFP equation and its spectrum form. Section 3 deliberates on the general relaxation model in cases where the velocity space exhibits spherical symmetry, including the closure relations. Then, a general temperature relaxation model is derived from the provided relaxation model in Section 4. Finally, a summary of this work is presented in Section 5.

2.   Theoretical formulation
  • The relaxation of Coulomb collision in fusion plasmas can be described by the VFP[1,2] equation for the velocity distribution functions, denoted as f = f(r, v,t) for species a, in physics space r and velocity space v, it reads

    Here, symbols ma and qa respectively denote the mass and charge of species a. Vectors E and B respectively represent the electric field intensity and magnetic flux density. Function ℭ is the Coulomb collision operator of species a, which encompasses both its self-collision effect and the mutual collision effect between species a and background species, as described by

    where Ns denotes the total number of plasmas species and ℭa,b represents the mutual FPRS collision operator[3,17,18] between species a and species b, given by

    Here, Γa,b=4π(qaqb4πε0ma)2lnΛa,b, the mass ratio mM = ma/mb. Here, symbols mb and qb respectively represent the mass and charge of species b. Parameters ε0 and lnΛa,b correspond to the dielectric constant of vacuum and the Coulomb logarithm.[19] Function F = F(r,vb,t), denoting the background distribution function of species b. The Rosenbluth potentials are integral functions of F, which can be expressed as follows:

    This paper exclusively focuses on the general scenario where the velocity space exhibits spherical symmetry in a spatially homogeneous plasma. It can be proved[6] that the spatial convection terms (second term), electric and magnetic field effect terms (third term) in Eq. (1) will vanish exactly when velocity space exhibits spherical symmetry. Therefore, the VFP equation reduces to

    Hereafter, the symbol r is disregarded in this homogeneous plasmas scenario.

  • By employing the spherical harmonics expansion[20,21] (SHE) within spherical-polar coordinate v(v, θ,ϕ), the distribution function of species a can be expanded as

    where v = |v|, μ = cosθ, and Ylm represents the spherical harmonic[22] without the normalization coefficient, Nlm=2l+14π(l−m)!(l+m)!. Then, the VFP spectrum equation[3] in spherically symmetric velocity space can be derived from Eq. (6). For order of l, it is

    where δl0 denotes the Kronecker symbols. Equation (8) indicates that harmonics with order l ≥ 1 or m ≠ 0 will vanish due to the spherical symmetry of the velocity space.[23] For convenience, the normalized speed is adopted as the dependent variable, for species a denoted as v^=v/vath. Moreover, when velocity space exhibits spherical symmetry, Ylm→Pl. In this paper, the superscript m is omitted because its value is always zero. For instance, we utilize fl instead of fl0. The amplitude fl is zero when l ≥ 1 and f0 is non-negative, which can be computed through the inverse transformation of SHE, and reads

    The lth-order amplitude of FPRS collision operator[3] of species a in Eq. (8) can be expressed as

    The parameters na = ∫vfdv and vath respectively denote the number density and thermal velocity of species a, which will be further elaborated in Subsection 3.1. The lth-order normalized FPRS collision operator of species a is defined as

    Similarly, the parameters nb and vbth denote the number density and thermal velocity of species b. The lth-order normalized amplitude of FPRS collision operator for species a colliding with species b can be represented as[3]

    where v^ab=vabthv^, the thermal velocity ratio vabth = vath/vbth and the coefficients

    In Eq. (12), the normalized amplitude of the distribution function can be calculated as follows:

    The normalized amplitudes of the background distribution function and its relative Rosenbluth potentials can be formulated as follows:

    The symbols Ii,0 and Ji,0 denote the functionals of F^0(v^b,t), similar to Shkarofsky et al.,[3,18] and read

    where the normalized speed of species b, v^b=vb/vbth. The definitions given by Eqs. (18) and (19) do not include the coefficient 4π. This coefficient arises from the application of spherical-polar coordinates in velocity space and is included in Eq. (12). Similarly, the Jacobian (v^b)2 in Eqs. (18) and (19) also originates from the utilization of the spherical-polar coordinates.

3.   General relaxation model for homogeneous plasmas
  • By introducing a new set of functions named R (given in Appendix A) and R integration (given in Appendix B), and applying the FDIF hypothesis introduced in Ref. [3], a general relaxation model for homogeneous plasmas with spherically symmetric velocity space, including shell structures,[3,16] will be presented based on the VFP spectrum equation (8).

  • In the spherical-polar coordinate system, the transport equations can be readily derived[3] by multiplying 4πmavj+2dv to both sides of the VFP spectrum equation (8), integrating over interval [0, ∞), thereby resulting in the (j,0)th-order kinetic moment evolution equation (KMEE)

    where j ∈ ℕ and the mass density ρa = mana. That is the transport equation for the case with spherically symmetric velocity space. In the above equation, ℳj,0 represents the (j,0)th-order kinetic moment[3] given by

    Symbol ℛ^j,0 denotes the (j,0)th-order normalized kinetic dissipative force

    where j ≥ −2. Function ℳ^j,0 represents the (j,0)th-order normalized kinetic moment.

    Especially when j = 0 and j = 2, the kinetic moments (21) can be simplified as follows:

    where Ka = (ma/2) ∫vv2fdv, representing the total energy of species a. Thus, the transport equation (20) represents the mass and energy conservation equations[3] respectively, which are stated as

    The average velocity is zero, which is dependent on the amplitude of the distribution function when l = 1. Therefore, the momentum conservation equation for scenarios where velocity space exhibits spherical symmetry will be

    Due to the symmetries[24] during the Coulomb collision between two species, the normalized FPRS collision operator (12) ensures the conservation of mass, momentum, and energy theoretically

    These are the conservation equations when the velocity space exhibits spherical symmetry.

    By employing the relations ρa (t) = mana and Ka (t) = (3/2)naTa, expressions for the number density and temperature can be derived in cases where the velocity space exhibits spherical symmetry, and read

    Subsequently, the thermal velocity can be determined according to the relation vath(t)=2Ta/ma, given by

    As stated in the introduction and demonstrated in Ref. [3], the aforementioned transport equation (20) is characterized by inherent nonlinearity and a lack of closure. This is ascribed to two primary factors: (i) the nonlinear interdependence among kinetic moments of different orders, (ii) the nonlinear dependence of the normalized kinetic dissipative forces on the kinetic moments. The closure strategies for these two aspects are provided in the following section.

  • Unlike the conventional approaches which usually simplify the nonlinear relations based on the near-equilibrium assumption,[10] the set of transport equations has been truncated based on the FDIF hypothesis[3] in (j) space. By employing the King function expansion (33) method, we establish the closure relations based on the FDIF hypothesis for kinetic moments with characteristic parameter equations (details in Subsection 3.2.2), as well as for the kinetic dissipative forces ℛ^j,0 with a kinetic dissipative force closure relation (details in Subsection 3.2.3). Furthermore, a kinetic moment-closed model is presented for homogeneous plasmas in the presence of spherical symmetry and shell structures in velocity space.

  • The spherical symmetry of the velocity space does not necessarily indicate that the system is in a thermodynamic equilibrium state. In this general scenario, Ref. [3] introduces a new function, which is referred to as zeroth-order King function as follows:

    where ι and σ represent the expectation and deviation of the King function. Under the FDIF hypothesis, by applying the King function expansion (KFE) technology,[3] the amplitude function f^0 can be approximated as

    where NKa ∈ ℕ+. Here, n^ar=nar/na, u^ar=uar/vath, and v^athr=vathr/vath, representing the characteristic parameters of the rth sub-distribution of f^0. Similarly, the normalized amplitudes of the background distribution function of species b can be approximated as

    The FDIF hypothesis indicates that NKa is a finite-size number in the zeroth-order King mixture model (KMM0) represented by Eq. (33). Let KMM0 denote the plasmas being in a quasi-equilibrium state,[15] which is a special case of non-equilibrium. Particularly, when ∑r(|u^ar|2)≡0, KMM0 reduces to the Maxwellian mixture model (MMM), indicating that the plasmas are in a shell-less quasi-equilibrium state.[3] The convergence of KMM0 can be established by using Wiener’s Tauberian theorem[25,26] as referenced in Ref. [3], and it has been demonstrated to have a moment convergence in Ref. [9]. When species a is in states of quasi-equilibrium with NKa = 1, the amplitude functions described by KMM0 as functions of v^, along with various normalized average velocities u^a, are depicted in Fig. 1. In order to examine the details when u^a>1, the amplitude functions are multiplied by a factor of (1+u^a2).

  • Under FDIF hypothesis, by substituting KMM0 (33) into the definition of normalized kinetic moments (21) and simplifying the result, the characteristic parameter equations (CPEs) are obtained when the velocity space exhibits spherical symmetry, namely,

    where j ≥ −2 and 1F1 (a,b,z) represents the Kummer confluent hypergeometric 1F1 function, and the coefficient

    Especially when j is even, the aforementioned CPEs can be reformulated into a more concise and efficient form

    where j∈{(2jp−2)|jp∈ℕ+}. The coefficient

    and Cβj/2 represents the binomial coefficient.

    The CPEs (35) with various values of j constitute a set of nonlinear algebraic equations. Generally, these equations typically involve a total of 3NKa unidentified parameters, which can be determined by solving the well-posed CPEs with 3NKa known normalized kinetic moments M^j,0. This approach is delineated in Refs. [3,9]. From the perspective of spectral analysis (where the King function serves as a non-orthogonal basis), distinct King functions represent different spectral components, while the characteristic parameters are quantities in the spectral space (j) rather than the physical quantities. This study focuses on offering the closure relations for transport equation (20), providing a kinetic moment-closed model for case with spherically symmetric velocity space.

  • To enclose the kinetic dissipative forces arising from the collision operator (12), one can rephrase Eq. (22) as

    where

    Let Eq. (39) represent the kinetic dissipative force closure relation.

    The analytical expression of the kinetic dissipative force closure relations can be derived by substituting KMM0 (33) into FPRS collision operator as described by Eq. (12), and then applying Eq. (40). With the introduction of a new set of functions, including the R function (details in Appendix A), the normalized kinetic dissipative forces between species a and species b can be formulated as

    where

    The terms on the right-hand side of the above equation correspond to the R integration, as defined in Appendix B.

    Especially, the utilization of the outcomes of the R function when the velocity space exhibits spherical symmetry without shell structure,[3] as represented by Eqs. (A8)–(A15) (wherein ∑r(|u^ar|2)≡0 and ∑s(|u^bs|2)≡0), equation (41) will reduce to the form presented in Ref. [3], which reads

    Here, 2F1(a,b,c,z) denotes the Gauss hypergeometric 2F1 function[22] with variable z.

  • Under FDIF hypothesis and by employing the KFE (33) method for the speed coordinate, the evolution of the plasmas system can be described by the transport equation (20), incorporating the analytical expression (41) of the normalized kinetic dissipative forces between species a and species b. The conservation equations represented by Eqs. (24)–(26), the characteristic parameter equations (35) and the kinetic dissipative force closure relations (39) serve as the constraint equations of the transport equation (20). The combination of these equations constitutes a set of nonlinear equations, which will be referred to as kinetic moment-closed model (KMCM) for the scenarios in which velocity space exhibits spherical symmetry. The advantages of KFE, and subsequently, the benefits of KMCM are outlined in Refs. [3,15].

    In particular, it should be noted once again that Ref. [6] strictly proves that the plasmas will be spatially homogeneous when the velocity space exhibits spherical symmetry, derived from the general VFP equation (1). Therefore, it is a natural process to extend the presented transport theory to the situation with general velocity space (scenarios of nonhomogeneous plasmas). When the velocity space of nonhomogeneous plasmas is axisymmetric, the corresponding transport theory has been developed and the collisionless part, described by the corresponding KMEE, is given in Ref. [6]. Meanwhile, the corresponding KMCM is also provided with the general closure relations. Furthermore, the collision part of the transport theory with axisymmetric velocity space based on the semi-anisotropic FPRS operator[23] will be published later, which can easily be extended to the form with nonlinear FPRS operator (3). Moreover, this transport theory can be further extended for general nonhomogeneous plasmas with definitions of associated King function, which will be provided in the subsequent research. Review [15] presents the proposed unified framework for constructing the overall picture of the King-type models in accordance with the different velocity space symmetries, as well as the strict theoretical application scope and the relationship among different specific models.

4.   Model application
  • One potential application of the aforesaid general relaxation model is to establish a temperature relaxation model for homogeneous plasmas with general spherically symmetric velocity space. When j = 0, the mass conservation equation can be determined by substituting relation represented by Eq. (23) into the transport equation (20), which reads

    The above equation utilizes the fact that ℛ^ab0,0≡0 for all elastic collisions given by Eq. (27).

    When j = 2, the energy conservation equation can be derived by substituting the relation M2,0(t)=2Ka (23) into the transport equation (20), as follows:

    Neglecting relativistic effects, the application of the relation Ka=32naTa and mass conservation (44) results in a temperature relaxation equation

    The general characteristic frequency of temperature relaxation

    Equation (46) serves as a general temperature relaxation model for the scenario where velocity space exhibits spherical symmetry. Generally, νTa shows a nonlinear dependence on Ta and Tb. The solution to the above equation (46) will take the following form only when νTa has weak dependence on Ta:

    where Ta0 = Ta (t = 0), which will be determined by the initial condition.

    The previous investigation[3] has revealed that when the near-equilibrium condition is fulfilled during two-species collisions, the temperature relaxation equation (46) can be explicitly reformulated as the traditional zeroth-order Braginskii heat transfer model,[3,27] assuming that both species are in close proximity to their respective thermodynamic equilibrium states. This zeroth-order model ignores the deviation of distribution function from the Maxwellian one and can be expressed as

    where νTab represents the characteristic frequency of temperature relaxation as given by Huba,[19] and is also derived from the shell-less KMCM.[3] This characteristic frequency is quoted as follows:

    However, the function ℛ^ab2,0 (41) is generally a nonlinear function of mM, vabth and characteristic parameters of species a and species b. By comparing Eq. (49) with Eq. (46), the characteristic frequency of temperature relaxation in the presence of spherical symmetry in velocity space can generally be expressed as

    The characteristic frequency of temperature relaxation described by Huba[19] is clearly a specific instance of the one described by Eq. (51). However, when Ta is very close to Tb for a quasi-equilibrium state plasma, the representation of νTab by Eq. (51) may numerically tend towards infinity. In this situation, equation (49) should be replaced by the original temperature relaxation model represented by Eq. (46).

    It is more evident to observe the distinctions in the physics pictures between the general temperature relaxation model (46) and the zeroth-order Braginskii heat transfer model (49), particularly in the simple scenario where the velocity space exhibits spherical symmetry without shell structure.[3] Under the traditional near-equilibrium assumption, if Ta is very close to Tb, the plasmas system tends towards the thermodynamic equilibrium state in the zeroth-order Braginskii heat transfer model. However, under FDIF hypothesis in the presented general temperature relaxation model, the plasmas system may still be in a shell-less quasi-equilibrium state.[3]

    Homogeneous models, such as zeroth-order Braginskii model, typically have two levels of application: a non-theoretically rigorous one (such as the experimental zero-order estimation for temperature relaxation) and a theoretically rigorous one (where the velocity space strictly satisfies the symmetry of the model). Meanwhile, although the authors endeavor to construct and promote a theoretically rigorous application, the presented theory can also be applied in a non-rigorous manner. As an alternative model, a numerical example is provided in the next section, offering an intuitive comparison between the above-mentioned general temperature relaxation model and the zeroth-order Braginskii model.

  • In general, the presented KMCM provides a nonlinear nonequilibrium description of the kinetic moments by using characteristic parameters as intermediate parameters in the (j,s) spectrum space, serving as a general relaxation model for homogeneous plasmas with spherically symmetric velocity space. For numerically solving this nonlinear model, the transport equation (20) can be solved[9] by utilizing a Runge–Kutta solver, such as the trapezoidal scheme.[28] The CPEs (35) can be addressed[9] through a least squares method[29] (LSM). As mentioned in Ref. [3] and demonstrated in a 0D-2V situation,[9] a general optimization algorithm, including self-adaptive NKa and collection of j, must be adopted for the solution of this nonlinear system. This issue will be addressed and published in the near future as a topic for the general optimization algorithm.

    In this work, the optimization schemes L01jd2 and L01jd2NK provided in Ref. [9] are employed to solve the CPEs (37) respectively, determining the characteristic parameters in the (j,s) spectrum space. The trapezoidal scheme[28] is utilized to solve the KMEE (20) while the standard explicit Runge–Kutta[28] method of order 4 is used for solving the zeroth-order Braginskii model. We provide the case of two-species thermal equilibration between species a and species b. For example, a collision process of deuterium and tritium with ma = 2[mp], mb = 3[mp], qa = qb = 1[e] and initial parameters na = nb = 1 × 1020m−3, ua = ub = 0, Ta = 10 [Tk], Tb = 20 [Tk], NKa = NKb = 1. Here, the symbols mp, e and Tk denote the mass of proton, elementary charge and practical unit Tk = keV. Symbol τ0 is the characteristic time, which is set to the initial temperature relaxation time in the zeroth-order Braginskii model, τ0=1/νTab(t=0). Theoretical values for the temperature at the equilibrium state can be obtained by utilizing the conservation equation for energy

    where k represents the kth time level. The initial total energy is Ks0=45[Tk] and the final temperatures of the thermal equilibrium state should be T∞ = 15[Tk].

    The temperatures of the zeroth-order Braginskii model (49) and the general temperature relaxation model (46) with the utilization of L01jd2 and L01jd2NK are depicted as functions of time in Fig. 2. In this case, a fixed timestep is employed with Δt = 2−5[τ0] and the maximum number of the King function in each species, NK = max (NKa, NKb), is set to 2 in the L01jd2NK scheme. The results of the L01jd2 scheme strictly adhere to the solution of the zeroth-order Braginskii model, as indicated by the relative disparity of temperatures, ΔT=|Tak−Tbk|/(Tak+Tbk) in the lower subplot of Fig. 2. This is because, in the L01jd2 scheme, n^as is set to be constant for each species,[9] meaning the maximum numbers of King function equal to the initial ones all the time. That is the same as the zero-order Braginskii model in this case where NKa = NKb = 1. However, the results obtained by L01jd2NK with NK = 2 exhibit the same qualitative attenuation trend as the zero-order Braginskii model, but with distinct attenuation rates. It should be noted that more accurate results need to be verified for their correctness on a larger scale, such as NK → 10 which means more high-order kinetic moments will be convergent.[9] This demands more robust optimization algorithms, which will be accomplished in our subsequent work.

5.   Conclusion
  • In this paper, a general relaxation model, namely, a kinetic moment-closed model (KMCM) for homogeneous plasmas with general spherically symmetric velocity space is proposed. This model consists of the transport equation and normalized kinetic dissipative forces between two species, along with incorporated constraint equations including the conservation equations, characteristic parameter equations (CPEs) and kinetic dissipative force closure relation. By employing the KMM0 which is based on the finitely distinguishable independent features (FDIF) hypothesis, and introducing a set of special functions named as R function and R integration, the closure relations of kinetic moments and kinetic dissipative forces can be presented in a closed form. Furthermore, in cases with shell-less velocity space, this closure relation for kinetic dissipative forces reduces to the simple version presented in our previous work, which is expressed in terms of the Gaussian hypergeometric 2F1 function.

    The presented general relaxation model accurately captures both the near-equilibrium state and the far-from-equilibrium states of homogeneous plasmas with spherically symmetric velocity space. These findings offer a new semi-analytical, semi-numerical approach for solving the VFP equation. The general temperature relaxation model is also achieved, and the general characteristic frequency of temperature relaxation in the presence of spherical symmetry in velocity space can be explicitly determined. Furthermore, this general relaxation model can serve as valuable benchmarks for nonlinear statistical physics applications such as fusion plasmas and solar plasmas.

    In upcoming research, we aim to address the general optimization algorithm for solving this nonlinear system, and extend these findings to encompass scenarios involving axisymmetric velocity space. Subsequently, this nonlinear model can be employed to investigate the classical physics phenomena such as Landau damping and Bernstein instability driven by the shell-type ion velocity distribution.

Appendix A: R function
  • When n = α = 0, the R function can be expressed in terms of Kummer confluent hypergeometric 1F1 function,[22] given by

    where

    When n = 0 and α = ±1, the R function will be

    Similarly, the other R function, jR0(γ,ι1,σ1,ι2,σ2;n, ±1) can also be expressed in terms of the Kummer confluent hypergeometric 1F1 function when n = 2 or n = ±1.

  • When ι1 = ι2 ≡ 0 for ∀ (n, α), the function jR0(γ,ι1,σ1,ι2,σ2;n,α) can be expressed in terms of Gamma function,[22] and reads

    Function jR0(γ,ι1,σ1,ι2,σ2;n) for ∀n can also be expressed in terms of Gaussian hypergeometric 2F1 function,[22] as follows:

    where σr = γσ1/σ2.

Figure (2)  Reference (29)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return