betaFeedback

WorksPhysics and Astronomy

New method for the solution of the two-body Dirac equation for the positronium bound states

Abstract

A new theoretical method is developed to solve the two-body bound-state Dirac equation for positronium. Only Coulomb potential was included in the Dirac Hamiltonian. It is shown that the two-body Dirac Hamiltonian can be written in the Hermitian matrix form of the 4×\times4 size and diagonalized in the momentum-state representation. Numerical results for the energy spectrum of the para- and ortho-positronium ground states performed within the variational method using the harmonic oscillator basis functions are in good agreement with a high-precision finite-element method of T.C. Scott et al. After the Fourier transformation into the coordinate-state representation the bound state wave functions of the para-Ps and ortho-Ps do not contain any singularity at the origin in contrast to the method mentioned above. The weights of the large-small and small-large components of the ground state wave functions are estimated to be of order 10−6^{-6}, while the weight of the small-small component is of order 10−12^{-12}.

Introduction

A new generation of the high sensitivity and multi-photon total-body positron emission tomography systems opens perspectives for clinical applications of positronium (Ps) in medicine [1]. On the other hand, Ps is the most important object for the precision tests of the violation of the discrete symmetries such as CC (charge conjugation), PP (spatial parity), TT (time reversal) and their combinations [2, 3, 4, 5, 6, 7, 8]. The spin-singlet bound states of the electron e−e^{-} and its antiparticle, the positron e+e^{+} present para-positronium (pp-Ps), at the same time the spin-triplet bound states present the ortho-positronium (oo-Ps).

The most important spectroscopic properties of the positronium, such as energy-level intervals are reproduced with a high precision at order α6\alpha^{6} within the non-relativistic QED [9, 10, 11, 12, 13] based on the solutions of the Schrödinger bound-state equation. The explicit relativistic invariance built into QED is essential for particle physics but leads to unnecessary complications when QED is applied to Coulombic bound states, which deals with the energy scales small compared to mec2m_{e}c^{2} [2].

The two-body Dirac equation (TBDE) for positronium was studied in Ref. [14] using only the Coulomb interaction potential. In Ref. [15] the sixteen-component, no-pair Dirac-Coulomb-Breit equation, derived from the Bethe-Salpeter equation, is solved in a variational procedure using Gaussian-type basis functions for positronium, muonium, hydrogen atom, and muonic hydrogen. Another type of relativistic Schrödinger equation was obtained and solved based on the Pauli reduction of the TBDE [16, 17].

Recent studies of properties of anomalous bound states within the TBDE formalism [18] have demonstrated that there are still many open questions for the structure of positronium. It was shown that the radial components of the Positronium bound-state wave-function [14] contain a singularity near the origin and differ from their Pauli approximations. This difference is due to mixing of the ground-state wave function with the anomalous bound-state wave functions near the origin. The solutions of the Bethe-Salpeter equation do not show any mixing with anomalous states and do not have a singularity near the origin [19].

The question is whether a bound state solution of the two-body Dirac equation with a pure Coulomb interaction potential can be free of mixing with an anomalous state. The aim of present work is to study this question within an alternative method. We solve the TBDE for the positronium bound states within a new theoretical method. The method is based on the Hermitian Dirac Hamiltonian in the form of 4×44 \times4 matrix, each element of which represents a four-component operator. The solution of the TBDE for the positronium state with total momentum JJ and its projection MM with the spin-orbital couplings J⃗=l⃗+S⃗\vec{J} = \vec{l} + \vec{S} and S⃗=S⃗1+S⃗2\vec{S} = \vec{S}_1 + \vec{S}_2, S1=S2=1/2S_1 = S_2 = 1/2, can be expressed in the form of a column of four components. Each of the large-large (LL), large-small (LS), small-large (SL) and small-small (SS) components of the eigen function represents a four-component spinor. The radial parts of these components are expressed as linear combination of the basic functions. This form of the Hermitian Dirac Hamiltonian contains a term σ⃗^p⃗^\hat{\vec{\sigma}}\hat{\vec{p}}, responsible for the coupling of the main large-large component with the large-small and small-large components. Numerical calculations have been performed within the variational method on the harmonic oscillator basic functions. The matrix elements of the most important operators including σ⃗^p⃗^\hat{\vec{\sigma}}\hat{\vec{p}}, are calculated in momentum state representation and are all real numbers in contrast to the coordinate space representation. As a results we come to the system of algebraic equations which is solved numerically [20, 21, 22, 23].

In § 0.2 we give the main formalism of the method. § 0.3 deals with numerical results and the conclusions are presented in the last Section.

Solution of the two-body bound state Dirac equation

Two-body Dirac equation for positronium bound states

The Dirac Hamiltonian for the two-body bound system interacting via Coulomb potential (ℏ=c=1\hbar= c = 1) is written in the form [14]

H^=α⃗1×I4(2)⋅p⃗1+I4(1)×α⃗2⋅p⃗2+β1×I4(2)m1+I4(1)×β2m2+I16V(r12),(1)\hat{H} = \vec{\alpha}_1 \times I_4^{(2)} \cdot\vec{p}_1 + I_4^{(1)} \times\vec{\alpha}_2 \cdot\vec{p}_2 + \beta_1 \times I_4^{(2)}m_1 + I_4^{(1)} \times\beta_2m_2 + I_{16}V(r_{12}), \tag*{(1)}

where the Dirac matrices

α⃗=βγ⃗=(0σ⃗σ⃗0),β=γ0=(I200−I2),(2)\vec{\alpha} = \beta\vec{\gamma} = \begin{pmatrix} 0 & \vec{\sigma} \\ \vec{\sigma} & 0 \end{pmatrix}, \qquad \beta= \gamma^0 = \begin{pmatrix} I_2 & 0 \\ 0 & -I_2 \end{pmatrix}, \tag*{(2)}

and InI_n is the unit n×nn \times n matrix operator, V(r12)V(r_{12}) is the Coulomb potential. Index 1 (2) of the operator means that this operator acts on the wave function of the first (second) particle. In order to write the Hamiltonian in the relative coordinates we introduce the radius-vectors of the center of mass R⃗\vec{R} and the relative-motion r⃗\vec{r} of the positron-electron two-body system:

R⃗=12(r⃗1+r⃗2),r⃗=r⃗1−r⃗2.\vec{R} = \frac{1}{2}(\vec{r}_1+\vec{r}_2), \quad\vec{r} = \vec{r}_1-\vec{r}_2.

With the assumption P⃗R=0\vec{P}_{R}=0, one can write the total Dirac Hamiltonian of the relative motion in terms of the relative momentum p⃗\vec{p} and coordinate r⃗\vec{r}:

H^={α⃗1×I4(2)−I4(1)×α⃗2}⋅p⃗+β1×I4(2)m1+I4(1)×β2m2+I16V(r).(3)\hat{H} = \left\{\vec{\alpha}_1 \times I_4^{(2)}-I_4^{(1)} \times\vec{\alpha}_2\right\}\cdot\vec{p}+\beta_1 \times I_4^{(2)}m_1+I_4^{(1)}\times\beta_2m_2+I_{16}V(r). \tag*{(3)}

Using the standard representation of the Dirac gamma matrices, the kinetic part of the Hamiltonian can be written in the form of symmetric matrix:

{α⃗1×I4(2)−I4(1)×α⃗2}⋅p⃗=(0−I2(1)×σ⃗2⋅p⃗σ⃗1×I2(2)⋅p⃗0−I2(1)×σ⃗2⋅p⃗00σ⃗1×I2(2)⋅p⃗σ⃗1×I2(2)⋅p⃗00−I2(1)×σ⃗2⋅p⃗0σ⃗1×I2(2)⋅p⃗−I2(1)×σ⃗2⋅p⃗0).(4)\left\{\vec{\alpha}_1 \times I_4^{(2)}-I_4^{(1)} \times\vec{\alpha}_2\right\}\cdot\vec{p}= \begin{pmatrix} 0 & -I_2^{(1)}\times\vec{\sigma}_2\cdot\vec{p} & \vec{\sigma}_1\times I_2^{(2)}\cdot\vec{p} & 0 \\ -I_2^{(1)}\times\vec{\sigma}_2\cdot\vec{p} & 0 & 0 & \vec{\sigma}_1\times I_2^{(2)}\cdot\vec{p} \\ \vec{\sigma}_1\times I_2^{(2)}\cdot\vec{p} & 0 & 0 & -I_2^{(1)}\times\vec{\sigma}_2\cdot\vec{p} \\ 0 & \vec{\sigma}_1\times I_2^{(2)}\cdot\vec{p} & -I_2^{(1)}\times\vec{\sigma}_2\cdot\vec{p} & 0 \end{pmatrix}. \tag*{(4)}

Then the total Hermitian Dirac Hamiltonian can be written in the form of 4×44 \times4 matrix, each element of which represents the 4×44 \times4 matrix operator acting on the four-component spinor:

H^=(δ1+I4V(r12)−I2(1)×σ⃗2⋅p⃗σ⃗1×I2(2)⋅p⃗00−I2(1)×σ⃗2⋅p⃗δ2+I4V(r12)0σ⃗1×I2(2)⋅p⃗σ⃗1×I2(2)⋅p⃗0−δ2+I4V(r12)−I2(1)×σ⃗2⋅p⃗0σ⃗1×I2(2)⋅p⃗−I2(1)×σ⃗2⋅p⃗−δ1+I4V(r12)),(5)\hat{H}= \begin{pmatrix} \delta_1+I_4V(r_{12})-I_2^{(1)}\times\vec{\sigma}_2\cdot\vec{p} & \vec{\sigma}_1\times I_2^{(2)}\cdot\vec{p} & 0 & 0 \\ -I_2^{(1)}\times\vec{\sigma}_2\cdot\vec{p} & \delta_2+I_4V(r_{12}) & 0 & \vec{\sigma}_1\times I_2^{(2)}\cdot\vec{p} \\ \vec{\sigma}_1\times I_2^{(2)}\cdot\vec{p} & 0 & -\delta_2+I_4V(r_{12}) & -I_2^{(1)}\times\vec{\sigma}_2\cdot\vec{p} \\ 0 & \vec{\sigma}_1\times I_2^{(2)}\cdot\vec{p} & -I_2^{(1)}\times\vec{\sigma}_2\cdot\vec{p} & -\delta_1+I_4V(r_{12}) \end{pmatrix}, \tag*{(5)}

where the diagonal elements read δ1=I2(1)×I2(2)(m1+m2)\delta_1=I_2^{(1)}\times I_2^{(2)}(m_1+m_2), δ2=I2(1)×I2(2)(m1−m2)\delta_2=I_2^{(1)}\times I_2^{(2)}(m_1-m_2), V(r12)=−αrV(r_{12})=-\frac{\alpha}{r}. The identity operators I2(1)I_2^{(1)}, I2(2)I_2^{(2)} of the 2×22 \times2 size act only on the spin function of the first and second particles, respectively. The above form of the Hermitian Dirac Hamiltonian matrix differs from the non-Hermitian matrix Hamiltonian of Ref. [14]. However, they coincide with each other in the 16×1616 \times16 matrix form.

It can be proven that the operators −I2(1)×σ2⋅p^-\mathbf{I}_{2}^{(1)} \times\boldsymbol{\sigma}_{2} \cdot\hat{\mathbf{p}} and σ1×I2(2)⋅p^\boldsymbol{\sigma}_{1} \times\mathbf{I}_{2}^{(2)} \cdot\hat{\mathbf{p}} of the Dirac Hamiltonian Eq.(5), where p^=p/p\hat{\mathbf{p}} = \mathbf{p}/p is the unit momentum vector, while acting on the spherical tensor YℓSJM(ξ⃗1,ξ⃗2;p^)\mathcal{Y}_{\ell S}^{JM}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\mathbf{p}}), can change the spin SS and the orbital momentum ℓ\ell of the system by one:

(−I2(1)×σ2⋅p^)Y(S1,S2)SℓJM(ξ⃗1,ξ⃗2;p^)=(−1)ℓ+ℓ±∑h,S′[h]([S][S′])12{S1S2SℓJh}{S1S2S′ℓ±Jh}Y(S1,S2)S′ℓ±JM(ξ⃗1,ξ⃗2;p^),(6)\left(-\mathbf{I}_{2}^{(1)} \times\boldsymbol{\sigma}_{2} \cdot\hat{\mathbf{p}}\right)\mathcal{Y}_{(S_{1},S_{2})S\ell}^{JM}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\mathbf{p}}) = (-1)^{\ell+\ell^{\pm}}\sum_{h,S'}[h]([S][S'])^{\frac{1}{2}}\left\{\begin{array}{ccc} S_{1} & S_{2} & S \\ \ell& J & h \end{array}\right\}\left\{\begin{array}{ccc} S_{1} & S_{2} & S' \\ \ell^{\pm} & J & h \end{array}\right\}\mathcal{Y}_{(S_{1},S_{2})S'\ell^{\pm}}^{JM}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\mathbf{p}}), \tag*{(6)}
(σ1×I2(2)⋅p^)Y(S1,S2)SℓJM(ξ⃗1,ξ⃗2;p^)=(−1)ℓ+ℓ±+1∑h,S′(−1)S+S′[h]([S][S′])12{S1S2SℓJh}{S1S2S′ℓ±Jh}Y(S1,S2)S′ℓ±JM(ξ⃗1,ξ⃗2;p^).(7)\left(\boldsymbol{\sigma}_{1} \times\mathbf{I}_{2}^{(2)} \cdot\hat{\mathbf{p}}\right)\mathcal{Y}_{(S_{1},S_{2})S\ell}^{JM}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\mathbf{p}}) = (-1)^{\ell+\ell^{\pm}+1}\sum_{h,S'}(-1)^{S+S'}[h]([S][S'])^{\frac{1}{2}}\left\{\begin{array}{ccc} S_{1} & S_{2} & S \\ \ell& J & h \end{array}\right\}\left\{\begin{array}{ccc} S_{1} & S_{2} & S' \\ \ell^{\pm} & J & h \end{array}\right\}\mathcal{Y}_{(S_{1},S_{2})S'\ell^{\pm}}^{JM}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\mathbf{p}}). \tag*{(7)}

where ℓ±=ℓ±1\ell^{\pm}=\ell\pm1, S′=0S'=0 and/or 11, and ℓ⃗±+S⃗′=J⃗\vec{\ell}^{\pm}+\vec{S}'=\vec{J}. The last equations Eq.(6) and Eq.(7) contain the most important result for the structure of positronium within the two-body Dirac equation formalism.

Variational method on a harmonic-oscillator basis

The solution of the two-body Dirac equation for the positronium state with total momentum JJ and its projection MM, where J⃗=ℓ⃗+S⃗\vec{J}=\vec{\ell}+\vec{S} and S⃗=S⃗1+S⃗2\vec{S}=\vec{S}_{1}+\vec{S}_{2},

(H^−E)ΨJM(ξ⃗1,ξ⃗2;r⃗)=0.(8)\left(\hat{\mathbf{H}}-E\right)\Psi^{JM}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r})=0. \tag*{(8)}

can be expressed as a column of four functions

ΨJM(ξ⃗1,ξ⃗2;r⃗)=(ΨLL(ξ⃗1,ξ⃗2;r⃗)ΨLS(ξ⃗1,ξ⃗2;r⃗)ΨSL(ξ⃗1,ξ⃗2;r⃗)ΨSS(ξ⃗1,ξ⃗2;r⃗)).(9)\Psi^{JM}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r})= \left( \begin{array}{c} \Psi_{LL}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r})\\ \Psi_{LS}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r})\\ \Psi_{SL}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r})\\ \Psi_{SS}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r}) \end{array} \right). \tag*{(9)}

where each of the large-large (LL), large-small (LS), small-large (SL) and small-small (SS) components are four-component spinors and expressed as linear combination of the harmonic-oscillator basic functions with corresponding spin-angular parts. For the fixed total momentum JJ, possible spin-orbital couplings include terms with ℓ=J−1\ell= J - 1, ℓ=J\ell= J, and ℓ=J+1\ell= J + 1, and spins S=0S = 0 and S=1S = 1. From Eq.(6) and the form of the Hamiltonian of Eq.(5) one can find that the main LL- and the last SS- components of the eigen-state contain contributions from the same spin-orbital (ℓ,S)(\ell,S) channels and differ from the LS- and SL- components containing contributions from channels (ℓ±,S′)(\ell^{\pm},S^{\prime}), whose orbital components differ by one. In particular, in the ground state of the para-Ps with Jπ=0+J^{\pi} = 0^{+} the LL- and SS- components consist of only (ℓ,S)=(0,0)(\ell,S) = (0,0) spin-orbital channel, while the LS- and SL-components contain contributions only from the (ℓ±,S′)=(1,1)(\ell^{\pm},S^{\prime}) = (1,1) channel. In the case of the ortho-Ps with Jπ=1+J^{\pi} = 1^{+}, the main LL- and smallest SS-components contain contributions from the (ℓ,S)=(0,1)(\ell,S) = (0,1) and (2,1)(2,1) spin-orbital channels, while the LS- and SL-components contain contributions from the (ℓ±,S′)=(1,0)(\ell^{\pm},S^{\prime}) = (1,0) and (1,1)(1,1) channels.

For the solution of the Dirac equation Eq.(8) a column of probe wave functions of Eq.(9) is expanded in a complete set of orthonormal harmonic oscillator states with which we calculate the matrix elements of the Hamiltonian of Eq.(5).

The harmonic oscillator basis functions ΨnℓSJM(ξ⃗1,ξ⃗2;r⃗)\Psi^{JM}_{n\ell S}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r}) with the variational oscillator parameter a0a_{0}:

ΨnℓSJM(ξ⃗1,ξ⃗2;r⃗)=Rnℓ(a0;r)⋅YℓSJM(ξ⃗1,ξ⃗2;r⃗^),(10)\Psi^{JM}_{n\ell S}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r}) = R_{n\ell}(a_{0};r) \cdot\mathcal{Y}^{JM}_{\ell S}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\vec{r}}), \tag*{(10)}

where YℓSJM(ξ⃗1,ξ⃗2;r⃗^)\mathcal{Y}^{JM}_{\ell S}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\vec{r}}) is the spherical tensor YℓSJM(ξ⃗1,ξ⃗2;r⃗^)={[χS1(ξ⃗1)×χS2(ξ⃗2)]S×Yℓ(r⃗^)}MJ\mathcal{Y}^{JM}_{\ell S}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\vec{r}}) = \left\{\left[\chi^{S_{1}}(\vec{\xi}_{1}) \times\chi^{S_{2}}(\vec{\xi}_{2})\right]_{S} \times\mathbf{Y}_{\ell}(\hat{\vec{r}})\right\}_{M}^{J}, with S1=S2=12S_{1} = S_{2} = \frac{1}{2}. The radial part of the basis function is presented as

Rnℓ(a0;r)=[2(n!)a03Γ(n+ℓ+32)]12(ra0)ℓexp⁡(−r22a02)Lnℓ+12(r2/a02).(11)R_{n\ell}(a_{0};r) = \left[\frac{2(n!)}{a_{0}^{3}\Gamma(n+\ell+\frac{3}{2})}\right]^{\frac{1}{2}}\left(\frac{r}{a_{0}}\right)^{\ell}\exp\left(-\frac{r^{2}}{2a_{0}^{2}}\right)L_{n}^{\ell+\frac{1}{2}}(r^{2}/a_{0}^{2}). \tag*{(11)}

where Lnℓ+12(r2)L_{n}^{\ell+\frac{1}{2}}(r^{2}) are the Laguerre polynomials:

Lnℓ+12(r2)=∑k=0n(−1)kk!Γ(n+ℓ+32)(n−k)!Γ(k+ℓ+32)⋅r2k.L_{n}^{\ell+\frac{1}{2}}(r^{2}) = \sum_{k=0}^{n}\frac{(-1)^{k}}{k!}\frac{\Gamma(n+\ell+\frac{3}{2})}{(n-k)!\Gamma(k+\ell+\frac{3}{2})}\cdot r^{2k}.

Accordingly, the harmonic oscillator basis functions ΨnℓSJM(ξ⃗1,ξ⃗2;r⃗)\Psi^{JM}_{n\ell S}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r}) are normalized to unity.

The Fourier transformation of the coordinate-space basis functions is the same state multiplied by a phase [20, 21]:

ΨnℓSJM(p⃗)=(2π)−3/2∫d3rΨnℓSJM(r⃗)exp⁡(−ip⃗r⃗)=(−i)2n+ℓRnℓ(a0;p)⋅YℓSJM(ξ⃗1,ξ⃗2;p⃗^)(12)\Psi_{n\ell S}^{JM}(\vec{p})=(2\pi)^{-3/2}\int d^{3}r\Psi_{n\ell S}^{JM}(\vec{r})\exp(-i\vec{p}\vec{r})=(-i)^{2n+\ell}R_{n\ell}(a_{0};p)\cdot\mathcal{Y}_{\ell S}^{JM}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\vec{p}}) \tag*{(12)}

where the harmonic oscillator wave functions in momentum representation is defined by the equation

Rnℓ(a0;p)=[2(n!)a03Γ(n+ℓ+32)]12(pa0)ℓexp⁡(−p2a022)Lnℓ+12(p2a02).(13)R_{n\ell}(a_{0};p)=\left[\frac{2(n!)a_{0}^{3}}{\Gamma(n+\ell+\frac{3}{2})}\right]^{\frac{1}{2}}(pa_{0})^{\ell}\exp\left(-\frac{p^{2}a_{0}^{2}}{2}\right)L_{n}^{\ell+\frac{1}{2}}(p^{2}a_{0}^{2}). \tag*{(13)}

The matrix elements of the Hamiltonian Eq. (5) are calculated in the momentum state representation [20, 21], because in this case the overlaps are all real, in contrast to the coordinate space representation, where the contribution of the operators −I2(1)×σ⃗2⋅p⃗^-\mathbf{I}_{2}^{(1)}\times\vec{\sigma}_{2}\cdot\hat{\vec{p}} and σ⃗1×I2(2)⋅p⃗^\vec{\sigma}_{1}\times\mathbf{I}_{2}^{(2)}\cdot\hat{\vec{p}} are purely imaginary. Therefore, in momentum space, the four-component spinors ΨLL\Psi_{LL}, ΨLS\Psi_{LS}, ΨSL\Psi_{SL}, ΨSS\Psi_{SS} of Eq. (9) are expanded in a complete set of harmonic oscillators with corresponding coefficients according to

ΨLL(ξ⃗1,ξ⃗2;p⃗)=∑nℓSanℓS∣nℓSJM⟩p-space\Psi_{LL}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{p})=\sum_{n\ell S}a_{n\ell S}|n\ell SJM\rangle_{\mathrm{p\text{-}space}}
ΨLS(ξ⃗1,ξ⃗2;p⃗)=∑nℓSbnℓS∣nℓSJM⟩p-space\Psi_{LS}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{p})=\sum_{n\ell S}b_{n\ell S}|n\ell SJM\rangle_{\mathrm{p\text{-}space}}
ΨSL(ξ⃗1,ξ⃗2;p⃗)=∑nℓScnℓS∣nℓSJM⟩p-space\Psi_{SL}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{p})=\sum_{n\ell S}c_{n\ell S}|n\ell SJM\rangle_{\mathrm{p\text{-}space}}
ΨSS(ξ⃗1,ξ⃗2;p⃗)=∑nℓSdnℓS∣nℓSJM⟩p-space\Psi_{SS}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{p})=\sum_{n\ell S}d_{n\ell S}|n\ell SJM\rangle_{\mathrm{p\text{-}space}}

with

∣nℓSJM⟩p-space=Rnℓ(a0;p){{χS1(ξ⃗1)×χS2(ξ⃗2)}S×Yℓ(p⃗^)}JM.|n\ell SJM\rangle_{\mathrm{p\text{-}space}}=R_{n\ell}(a_{0};p)\left\{\left\{\chi^{S_{1}}(\vec{\xi}_{1})\times\chi^{S_{2}}(\vec{\xi}_{2})\right\}_{S}\times Y_{\ell}(\hat{\vec{p}})\right\}_{JM}.

A minimization of the energy is performed with the oscillator parameter a0a_{0}. After the inverse Fourier transformation, we again come to the solution of the Dirac equation in the coordinate space:

ΨLL(ξ⃗1,ξ⃗2;r⃗)=∑nℓSi2n+ℓanℓSRnℓ(a0;r)YℓSJM(ξ⃗1,ξ⃗2;r⃗^)\Psi_{LL}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r})=\sum_{n\ell S}i^{2n+\ell}a_{n\ell S}R_{n\ell}(a_{0};r)\mathcal{Y}_{\ell S}^{JM}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\vec{r}})
ΨLS(ξ⃗1,ξ⃗2;r⃗)=∑nℓSi2n+ℓbnℓSRnℓ(a0;r)YℓSJM(ξ⃗1,ξ⃗2;r⃗^)\Psi_{LS}(\vec{\xi}_{1},\vec{\xi}_{2};\vec{r})=\sum_{n\ell S}i^{2n+\ell}b_{n\ell S}R_{n\ell}(a_{0};r)\mathcal{Y}_{\ell S}^{JM}(\vec{\xi}_{1},\vec{\xi}_{2};\hat{\vec{r}})
ΨSL(ξ⃗1,ξ⃗2;r⃗)=∑nℓSi2n+ℓcnℓSRnℓ(a0;r)YℓSJM(ξ⃗1,ξ⃗2;r⃗^)\Psi_{SL}(\vec{\xi}_1,\vec{\xi}_2;\vec{r})=\sum_{n\ell S}i^{2n+\ell}c_{n\ell S}R_{n\ell}(a_0;r)Y_{\ell S}^{JM}(\vec{\xi}_1,\vec{\xi}_2;\hat{\vec{r}})
ΨSS(ξ⃗1,ξ⃗2;r⃗)=∑nℓSi2n+ℓdnℓSRnℓ(a0;r)YℓSJM(ξ⃗1,ξ⃗2;r⃗^)\Psi_{SS}(\vec{\xi}_1,\vec{\xi}_2;\vec{r})=\sum_{n\ell S}i^{2n+\ell}d_{n\ell S}R_{n\ell}(a_0;r)Y_{\ell S}^{JM}(\vec{\xi}_1,\vec{\xi}_2;\hat{\vec{r}})

Numerical results

Numerical calculations have been performed with the value of the fine-structure constant α=1/137\alpha= 1/137 for comparison with the high-precision results of the finite-element method [14]. From the form of the Dirac Hamiltonian in Eq. (5) one can find that the LS- and SL-components of the solution of the two-body Dirac equation are identical both for the para-Ps and ortho-Ps. For the ground and excited states of the para-positronium with J=0J = 0 the main large-large (LL) and small-small (SS) components consist of only (ℓ,S)=(0,0)(\ell,S) = (0,0) spin-orbital channel, while the large-small (LS) and small-large (SS) components consist of only (ℓ,S)=(1,1)(\ell,S) = (1,1) channel. In the case of the ortho-Ps, the large-large (LL) and small-small (SS) components contain contributions from the (ℓ,S)=(0,1)(\ell,S) = (0,1) and (2,1)(2,1) spin-orbital channels, while the large-small (LS) and small-large (SL) components contain contributions from (ℓ,S)=(1,0)(\ell,S) = (1,0) and (1,1)(1,1) channels.

In Figure 1 and Figure 2 the radial wave-functions of the large-large (LL), large-small (LS) and small-small (SS) components of the para-Ps and ortho-Ps ground states are presented, respectively. In the upper index of the radial function notations the corresponding spin-orbital channels are indicated as ΨLL(ℓ,S)\Psi_{LL}^{(\ell,S)}, ΨLS(ℓ,S)\Psi_{LS}^{(\ell,S)}, and ΨSS(ℓ,S)\Psi_{SS}^{(\ell,S)}. As can be noted, the behaviors of the radial wave functions of all the components of the para-Ps are close to the main parts of the corresponding components of the ortho-Ps. In particular, the corresponding LL-components are almost identical. The values of the SS-component of the para-Ps are very close to the absolute values of the main SS-component ΨSS(0,1)(r)\Psi_{SS}^{(0,1)}(r) of the ortho-Ps. The LS-components of the ortho-Ps corresponding to different spin-orbital channels ΨLS(1,0)(r)\Psi_{LS}^{(1,0)}(r) and ΨLS(1,1)(r)\Psi_{LS}^{(1,1)}(r) are of the same form, but have opposite signs and differ by about 40% in the absolute values.

Radial wave-functions of the large-large (LL), large-small (LS) and small-small (SS) components of the para-Ps ground state

Figure 1. Radial wave-functions of the large-large (LL), large-small (LS) and small-small (SS) components of the para-Ps ground state.

Image not extracted

Figure 2. Radial wave-functions of the large-large (LL), large-small (LS) and small-small (SS) components of the ortho-Ps ground state.

Another important result is that all the large-large, large-small , small-large and small-small components of the ground state wave functions for the para-Ps and ortho-Ps do not have any singularity around zero. This results is in contrast with the results of the method, employed in Refs.[14], where the ground state solution of the two-body Dirac equation contains a singularity at the origin due to the mixing with the anomalous bound state. This problem was analyzed recently in details in Refs. [18, 19].

In Table I a convergence of the numerical results of calculations for the energy E/2meE/2m_e, binding energy EbE_b, the sum of the weights of the LS- and SL- components W=∥Ψ(L,S)∥2+∥Ψ(S,L)∥2W = \lVert\Psi_{(L,S)}\rVert^2 + \lVert\Psi_{(S,L)}\rVert^2 in the para- and ortho-positronium ground-states are presented in respect to the maximal oscillator quantum number of basis states nmaxn_{max}. The optimal value of the oscillator length a0=3.3×105 GeV−1a_0 = 3.3 \times10^5\ \mathrm{GeV}^{-1}. As can be seen from the table, all the parameters are well converged with increasing nmaxn_{max}. As noted above, the further increase of nmaxn_{max} results in the collapse of the variational basis. From the table one can see that the final result for the weights of the sum of the large-small and small-large components are identical for the para-Ps and ortho-Ps and estimated to be 6.656E-6. The weight of the small-small component is of order 10−1210^{-12} in the both para-Ps and ortho-Ps ground states.

In Table II the energies of the lowest states of the para- and ortho-Ps are compared with the results of the finite-element method of high accuracy [14]. Of course, our variational method on the harmonic-oscillator basis is not very accurate as the finite-element method

(ℓ,S,J)(\ell, S, J)E/2meE/2m_eE/2meE/2m_e [14]
(0,0,0)(0,0,0)0.999 993 343 356 8430.999 993 340 148 538 880
(0,1,1)(0,1,1)0.999 993 343 356 8920.999 993 340 148 552 498
(1,1,0)(1,1,0)0.999 998 342 584 5490.999 998 335 009 885 854
(1,0,1)(1,0,1)0.999 998 342 513 5430.999 998 335 017 278 391

Table II. Energies of the lowest states of the para- and ortho-Ps in comparison with the results of the finite-element method [14].

Statenmax⁡n_{\max}E/2meE/2m_eEbE_b (eV)WW
1S0^{1}S_{0} (p-Ps)100.99999350565806956.6372035297.467E-6
1S0^{1}S_{0} (p-Ps)200.99999335970229656.7863700166.760E-6
1S0^{1}S_{0} (p-Ps)300.99999334539550956.8009915226.667E-6
1S0^{1}S_{0} (p-Ps)350.99999334389247526.8025276206.659E-6
1S0^{1}S_{0} (p-Ps)370.99999334335684336.8030750346.656E-6
3S1^{3}S_{1} (o-Ps)100.99999350565806946.6372035297.467E-6
3S1^{3}S_{1} (o-Ps)200.99999335970229616.7863700176.756E-6
3S1^{3}S_{1} (o-Ps)300.99999334539550916.8009915226.666E-6
3S1^{3}S_{1} (o-Ps)350.99999334389247456.8025276206.660E-6
3S1^{3}S_{1} (o-Ps)370.99999334335689236.8030749846.656E-6

Table 1. Convergence of the numerical results for E/2meE/2m_e, EbE_b, the sum of weights of the LS- and SL-components in the para- and ortho-positronium ground-states in respect to nmax⁡n_{\max}.

due-to the collapse of the variational basis at some value of nmax⁡n_{\max}. The binding energies of the pp-Ps and oo-Ps ground states in our calculations are estimated to be 6.803075 eV, which should be compared with the values of 6.806403 eV for the para-Ps and 6.806354 eV for the ortho-Ps of Ref.[14]. This means that the precision of the energy calculations is of order 0.003 eV, which is good enough for the application of corresponding wave functions to the study of different processes with para- and ortho-positronium.

Summary

A new theoretical method was developed to solve the two-body Dirac equation for para- and ortho-positronium bound states. Only Coulomb potential was included into the Dirac Hamiltonian. It is shown that the two-body Dirac Hamiltonian can be written in the Hermitian form and solved in the momentum state representation. Numerical results for the energy spectrum of the para- and ortho-positronium lowest bound states performed within the variational method using the harmonic oscillator basis functions are in good agreement with a high-precision finite-element method. It was shown that the para- and ortho-Ps ground state wave functions after the Fourier transformation to the coordinate space representation do not contain any singularity at the origin in contrast to the traditional bound-state solution in the literature. The weights of the sum of the large-small and small-large components were estimated to be of order 10−610^{-6}, while the weight of the small-small components are of order 10−1210^{-12} for the both para- and ortho-positronium ground states.

Acknowledgements

The authors thank D. Baye for valuable discussions and comparison with the results of the Lagrange-mesh method, E. Czerwinski, T.C. Scott and A.M. Rakhimov for useful discussions of the presented results.

References

References

  1. [1]S.D. Bass, S. Mariazzi, P. Moskal, E. Stepien, Rev. Mod. Phys. 95, 021002 (2023).
  2. [2]G.S. Adkins, D.B. Cassidy, J. Perez-Rios, Phys. Rep. 975, 1-61 (2022).
  3. [3]P.A. Vetter, S. J. Freedman, Phys. Rev. A 66, 052505 (2002).
  4. [4]A. Rubbia, Int. J. Mod. Phys. 19, 3961 (2004)
  5. [5]T. Yamazaki, T. Namba, S. Asai, T. Kobayashi, Phys. Rev. Lett. 104, 083401 (2010).
  6. [6]C Bartram and R Henning, J. Phys.: Conf. Ser. 1342, 012106 (2020).
  7. [7]T.-E. Haugen, E. A. George, O. Naviliat-Cuncic, and P. A. Voytas, EPJ Web of Conferences 282, 01003(2023).
  8. [8]P. Moskal, E. Czerwinski, J. Raj et al. Nat. Commun. 15, 78 (2024).
  9. [9]W.E. Caswell, G.P. Lepage, Phys. Lett. B 167, 437 (1986).
  10. [10]A. Czarnecki, K. Melnikov, A. Yelkhovsky, Phys. Rev. Lett. 82, 311 (1999).
  11. [11]K. Pachucki, S.G. Karshenboim, Phys. Rev. Lett. 80, 2101 (1998).
  12. [12]A. Czarnecki, K. Melnikov, A. Yelkhovsky, Phys. Rev. A 59, 4316 (1999).
  13. [13]P. Labelle, Phys. Rev. D 58, 093013 (1998).
  14. [14]T.C. Scott, J. Shertzer, R. A. Moore, Phys. Rev. A 45, 4393 (1992)
  15. [15]D. Ferenc and E. Matyus, Phys. Rev. A bf 107, 052803 (2023).
  16. [16]H.W. Crater, C.Y. Wong, P. Van Alstine, Phys. Rev. D 74, 054028 (2006).
  17. [17]H.W. Crater, Jin-Hee Yoon, Cheuk-Yin Wong, Phys. Rev. D 79, 034011 (2009).
  18. [18]C.W. Patterson, Phys. Rev. A 100, 062128 (2019).
  19. [19]C.W. Patterson, Phys. Rev. A 107, 042816 (2023).
  20. [20]Th. Gutsche and D. Robson, Phys. Lett. B 229, 333 (1989).
  21. [21]Th. Gutsche. A chiral potential model for the light-quark baryons. Ph.D thesis, Florida State University, 249 p. (1987).
  22. [22]E.M. Tursunov, J. Phys. G 36, 095006 (2009).
  23. [23]E.M. Tursunov, and S. Krewald, Phys. Rev. D 90, 074015 (2014).

Paper details

vlnc:W_h78pcd99
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents