betaFeedback

WorksPhysics and Astronomy

Magnetic dipole contribution to the γγ-spectrum from dd-tt collisions

Abstract

One in about a few hundred thousand sub-Coulomb dd-tt collisions is accompanied by the emission of a γγ-photon. The γγ-spectrum from these collisions is dominated by a 16.7 MeV peak, corresponding to the population of the pp-wave 3/2−3/2^- αα-nn ground-state resonance in the E1 transition from an intermediate dd-wave αα-nn state, corresponding to the 5^5He excited state around 16.7 MeV. The strength of this spectrum at the large γγ-energy endpoint decreases fast both due to kinematic factors and due to centrifugal repulsion between neutron and αα-particle at near-zero relative energies. However, no centrifugal repulsion would occur if ss-wave αα-nn states were populated in magnetic dipole transition. Therefore, one can expect that close to the γγ-spectrum endpoint around 17.5 MeV the M1 transition could dominate, leading to additional counts in experimental spectra which could be easily misinterpreted as a background. This work presents calculations of the M1 contribution to the d+t→α+n+γd+t\rightarrow α+n+γ cross section caused both by convection and magnetization currents. While the first contribution was found to be negligible, the one from magnetization depends strongly on the choice of the model for the dd-tt channel scattering wave function. Using dd-tt interaction models from the literature, represented either by a gaussian of by a square well, resulted in unrealistically strong M1 contribution. With constructed in the present work optical folding potential, correctly representing the sizes of deuteron and triton, the M1 contribution near the endpoint is reduced to about 2%\% of the 16.7-MeV E1 peak. Arguments in favor of more detailed calculations, that could potentially change this number, are put forward.

Introduction

The d+t→α+nd+t\rightarrow\alpha+n reaction is the best candidate for successful fusion-energy-exploiting industrial applications. Rarely, for about one in a few hundred thousand events, the dd-tt collisions end up with the emission of a γ\gamma-ray. The γ\gamma-spectrum has a peak at around 16.7 MeV, corresponding to the population of the 3/2−3/2^{-} ground-state resonance in the α\alpha-nn system. Given that these high-energy γ\gamma's could potentially become a basis of a new dd-tt plasma diagnostic tool [1], the study of the d+t→α+n+γd+t\rightarrow\alpha+n+\gamma reaction has attracted significant interest from the experimental nuclear physics community. Experimental studies of this reaction were carried out in cyclotrons through 1960-1990s [2, 3, 4, 5, 6, 7] and since 2010s they are also conducted in inertial confinement fusion (ICF) facilities [8, 9, 10, 11, 12, 13]. The first results on γ\gamma-detection in JET has just been reported in [14, 15].

Until very recently, no theoretical predictions were available for the d+t→α+n+γd+t\rightarrow\alpha+n+\gamma to d+t→α+nd+t\rightarrow\alpha+n branching ratio and the only available modelling tool was the phenomenological R-matrix approach, which was used in [10, 16] for γ\gamma-spectrum predictions on the assumption that spectral features in the α\alpha-nn scattering and the d+t→α+n+γd+t\rightarrow\alpha+n+\gamma reaction are exactly the same. The first theoretical prediction of the low-energy d+t→α+n+γd+t\rightarrow\alpha+n+\gamma reaction has been reported in [17], aimed at understanding both the shape of the γ\gamma-spectrum and its absolute strength. The main challenge encountered in that work was that the widely-used in low-energy nuclear physics approximation for electric dipole transition amplitude, ME1≈∑ieiriM_{E1}\approx\sum_i e_i\mathbf{r}_i, forbids the d+t→α+n+γd+t\rightarrow\alpha+n+\gamma reaction from happening because of the spin conservation rule. In the vicinity of the 3/2+3/2^{+} resonance, determining the main properties of the dd-tt system, the deuteron spin Jd=1J_d=1 and the triton spin Jt=1/2J_t=1/2 couple to the total spin S⃗=J⃗d+J⃗t\vec{S}=\vec{J}_d+\vec{J}_t equal to 3/23/2 while the channel spin in the α\alpha-nn channel is different, being 1/21/2. However, the dd-tt component of the 5He{}^{5}\mathrm{He} five-body wave function is coupled by the nucleon-nucleon tensor interaction to the high-energy dd-wave α\alpha-nn component with the spin S=1/2S=1/2 and it is the E1 γ\gamma-transition between the high- and low-energy S=1/2S=1/2 states in the α\alpha-nn partition of 5He{}^{5}\mathrm{He} that provides the main contribution to the d+t→α+n+γd+t\rightarrow\alpha+n+\gamma reaction mechanism.

If allowed by selection rules, electric dipole transitions are the strongest, with other multipolarities being suppressed. In the case of the d+t→α+n+γd+t\rightarrow\alpha+n+\gamma reaction, the electric quadrupole transitions E2 are several orders of magnitude weaker than the E1 transitions [17]. Usually, magnetic dipole transitions M1 have a similar strength to that the E2 ones. However, for the d+t→α+n+γd+t\rightarrow\alpha+n+\gamma reaction these transitions should be studied more carefully because of the following reason. Since the M1 transition does not conserve the total spin of the system, it can directly populate the final α\alpha-nn state from the dd-tt partition of 5He{}^{5}\mathrm{He} without any need for going through the high-energy dd-wave α\alpha-nn state with S=1/2S=1/2. The populated by M1 final α\alpha-nn state has zero orbital momentum and, therefore, the same (positive) parity that the initial state has and thus will not contribute to the (negative parity) resonant peak featuring in the E1 dtγd t\gamma-spectrum. However, it may contribute to the high-energy endpoint of this spectrum because, unlike in the case of the pp-wave α\alpha-nn scattering, the ss-wave motion is not affected by the centrifugal barrier and the α\alpha-nn wave function does not vanish at small α\alpha-nn separations at near-zero energies.

Also, the strength of the M1 transitions grows with γ\gamma-energy. As the result, some M1-originated counts could appear in the endpoint of the d+t→α+n+γd+t\rightarrow\alpha+n+\gamma spectrum, which could be easily misidentified as background. Subtracting their contribution could induce a systematic error into the γ\gamma-spectrum presented as a measured quantity.

In this paper the M1 contribution to the d+t→α+n+γd+t\rightarrow\alpha+n+\gamma reaction is calculated using a potential model description of the dd-tt and α\alpha-nn channel functions. These calculations are based on the exact M1 operator, which along with ss-wave to ss-wave transitions, caused by magnetization currents, allows convection-induced bremsstrahlung between dd-wave states of the 5He{}^{5}\mathrm{He}. Sect. II presents the M1 formalism for the case of the d+t→α+n+γd+t\rightarrow\alpha+n+\gamma reaction. Sect. III shows numerical results while Sect. IV discusses the findings made in the previous section and draws the conclusions.

Formalism for M1-induced d+t→α+n+γd+t\rightarrow\alpha+n+\gamma transition

The general form of the Mλ\lambda operator describing the interaction of a photon with an AA-body system in first-order perturbation theory is given by the sum of two terms [18]:

MμMλ=MμMλ;conv+MμMλ;magn=eℏmpc(2λ+1)!!(λ+1)kγλ∑j=1A{glj[∇⃗jϕλμ(rj)]⋅L⃗j+12gsj∇⃗j×L⃗jiϕλμ(rj)⋅S⃗j},(1)\begin{aligned} \mathcal{M}^{M\lambda}_{\mu}=\mathcal{M}^{M\lambda;\mathrm{conv}}_{\mu}+\mathcal{M}^{M\lambda;\mathrm{magn}}_{\mu} \\ &=\frac{e\hbar}{m_{p}c}\frac{(2\lambda+1)!!}{(\lambda+1)k_{\gamma}^{\lambda}}\sum_{j=1}^{A}\left\{g_{l_{j}}\left[\vec{\nabla}_{j}\phi_{\lambda\mu}(r_{j})\right]\cdot\vec{L}_{j}+\frac{1}{2}g_{s_{j}}\frac{\vec{\nabla}_{j}\times\vec{L}_{j}}{i}\phi_{\lambda\mu}(r_{j})\cdot\vec{S}_{j}\right\}, \tag*{(1)} \end{aligned}

where ee is the proton charge, mpm_{p} is the proton mass, glj=(12−t3j)g_{l_{j}}=\left(\frac{1}{2}-t_{3j}\right) with t3jt_{3j} being the isospin projection of nucleon jj onto the zz-axis, gsj=(gn+gp)/2+t3j(gn−gp)g_{s_{j}}=(g_{n}+g_{p})/2+t_{3j}(g_{n}-g_{p}) and gng_{n} and gpg_{p} are gyromagnetic factors of the neutron and proton, respectively. Also, L⃗j\vec{L}_{j} is the operator of the orbital momentum of nucleon jj,

L⃗j=−ir⃗j×∇⃗j,(2)\vec{L}_{j}=-i\vec{r}_{j}\times\vec{\nabla}_{j}, \tag*{(2)}

S⃗j\vec{S}_{j} is the spin operator acting on nucleon jj and

ϕλμ(rj)=jλ(kγrj)Yλμ(r^j).(3)\phi_{\lambda\mu}(r_{j})=j_{\lambda}(k_{\gamma}r_{j})Y_{\lambda\mu}(\hat{r}_{j}). \tag*{(3)}

The coordinate r⃗j\vec{r}_{j} of nucleon jj is with respect to the centre of mass of system AA, which is important for such light systems as d+td+t.

The first term of this operator, MμMλ;conv\mathcal{M}^{M\lambda;\mathrm{conv}}_{\mu}, comes from the convection current, arising from the motion of protons, while the second term, MμMλ;magn\mathcal{M}^{M\lambda;\mathrm{magn}}_{\mu}, is due to magnetization current, depending on the spins of all nucleons. The first term conserves the total nuclear spin, and therefore will contribute to bremsstrahlung between the dd-wave states in the α\alpha-nn partition of 5He{}^{5}\mathrm{He}. The second term can couple the ss-wave dd-tt and α\alpha-nn partition. The formal expressions for differential cross section corresponding to convection and magnetization currents are considered below separately. The cross sections are calculated from the modulus squared of the matrix element ⟨Ψf∣MμM1∣Ψi⟩\langle\Psi_{f}|\mathcal{M}^{M1}_{\mu}|\Psi_{i}\rangle between the wave functions Ψi\Psi_{i} and Ψf\Psi_{f} of the initial and final states of the 5He{}^{5}\mathrm{He} system.

Partial wave representation of the wave functions

The wave function Ψf\Psi_{f} in the final channel is taken as an antisymmetrized product of the 4+14+1 partition wave function Φ(41)(1,2,3,4;5)\Phi^{(41)}(1,2,3,4;5) in which nucleons 1, 2, 3 and 4 belong to the α\alpha-particle and nucleon 5 is a neutron:

Ψf(1,2,3,4,5)=AαnΦnα,Mn(41)(1,2,3,4;5),(4)\Psi_{f}(1,2,3,4,5)=\mathcal{A}_{\alpha n}\Phi^{(41)}_{n\alpha,M_{n}}(1,2,3,4;5), \tag*{(4)}

where Ac1c2\mathcal{A}_{c_{1}c_{2}} is the antisymmetrization operator

Ac1c2=(c1!c2!(c1+c2)!)1/2∑Πc1c2ϵΠc1c2,(5)\mathcal{A}_{c_{1}c_{2}}=\left(\frac{c_{1}!c_{2}!}{(c_{1}+c_{2})!}\right)^{1/2}\sum_{\Pi_{c_{1}c_{2}}}\epsilon_{\Pi_{c_{1}c_{2}}}, \tag*{(5)}

that contains all permutations Πc1c2\Pi_{c_{1}c_{2}} between nucleons from two different clusters c1c_{1} and c2c_{2} and the sign ϵΠc1c2\epsilon_{\Pi_{c_{1}c_{2}}} accompanying these permutations, being either ++ or −- depending on whether the number of permutations is even or odd. The partition function Φnα,Mn(41)(1,2,3,4;5)\Phi^{(41)}_{n\alpha,M_{n}}(1,2,3,4;5) depends on the α\alpha-nn momentum kfk_{f} and on the spin projection MnM_{n} of the neutron with it is far from α\alpha-particle. It is represented by a general partial wave expansion of the form

Φnα,Mn(41)(1,2,3,4;5)=1vf4πkfrψα(1,2,3,4)×∑Lf′LfJfMfMf′σnf(LfMf12Mn∣JfMf)(Lf′Mf′12σnf∣JfMf′)×χ12σn(spin)(5)iLf′YLfMf∗(k^f)YLf′Mf′(r^αn)χLf′LfJf(kf,rαn),(6)\begin{aligned} \Phi^{(41)}_{n\alpha,M_{n}}(1,2,3,4;5)=\frac{1}{\sqrt{v_{f}}}\frac{4\pi}{k_{f}r}\psi_{\alpha}(1,2,3,4) \\ &\times\sum_{\substack{L'_{f}L_{f}J_{f}\\M_{f}M'_{f}\sigma_{n_{f}}}}\left(L_{f}M_{f}\tfrac{1}{2}M_{n}|J_{f}M_{f}\right)\left(L'_{f}M'_{f}\tfrac{1}{2}\sigma_{n_{f}}|J_{f}M'_{f}\right) \\ &\times\chi^{(\mathrm{spin})}_{\frac{1}{2}\sigma_{n}}(5)i^{L'_{f}}Y^{*}_{L_{f}M_{f}}(\hat{k}_{f})Y_{L'_{f}M'_{f}}(\hat{r}_{\alpha n})\chi^{J_{f}}_{L'_{f}L_{f}}(k_{f},r_{\alpha n}), \tag*{(6)} \end{aligned}

where vf=ℏkf/μαnv_{f}=\hbar k_{f}/\mu_{\alpha n} is the velocity in the exit α\alpha-nn channel determined by the reduced mass μαn\mu_{\alpha n} and the energy Ef=ℏ2kf2/2μαnE_{f}=\hbar^{2}k_{f}^{2}/2\mu_{\alpha n}, the LfL_{f} and JfJ_{f} are orbital momentum and total angular momentum, respectively, and MfM_{f} and Mf′M'_{f} are their corresponding projections. The Clebsch-Gordan coefficients provide angular momentum conservation, L⃗f+12=J⃗f\vec{L}_{f}+\frac{1}{2}=\vec{J}_{f}. Also in (6), ψα(1,2,3,4)\psi_{\alpha}(1,2,3,4) is the intrinsic antisymmetric α\alpha-particle wave function, χ12σn(spin)(5)\chi^{(\mathrm{spin})}_{\frac{1}{2}\sigma_{n}}(5) is the spin wave function of the neutron, labelled by number 5, with the spin projection σn\sigma_{n}, and χLf′LfJf(kf,rαn)\chi^{J_{f}}_{L'_{f}L_{f}}(k_{f},r_{\alpha n}) is the radial wave function that depends on the neutron coordinate with respect to the centre of mass of the α\alpha-particle, r⃗αn=r⃗n−r⃗α\vec{r}_{\alpha n}=\vec{r}_{n}-\vec{r}_{\alpha}.

The wave function Ψi\Psi_i in the entrance dd-tt channel is represented by two partitions,

Ψi(1,2,3,4,5)=AdtΦdt,MdMt(32)(1,2,3;4,5)+Anα,MnΦnα,Mn(41)(1,2,3,4;5),(7)\Psi_i(1,2,3,4,5)=\mathcal{A}_{dt}\Phi_{dt,M_dM_t}^{(32)}(1,2,3;4,5)+\mathcal{A}_{n\alpha,M_n}\Phi_{n\alpha,M_n}^{(41)}(1,2,3,4;5), \tag*{(7)}

where MdM_d and MtM_t are the projections of the deuteron and triton spins at large separations before the collision. The 3+2 partition is given by

Φdt,MdMt(32)(1,2,3;4,5)=1vdt4πkdtrdt∑Ldt′Si′LdtSiJiMdtMi′MiMJiiLdt(JtMtJdMd∣SiMi)×[ψt(1,2,3)⊗ψd(4,5)]Si′Mi′(LdtMdtSiMi∣JiMJi)×(Ldt′Mdt′Si′Mi′∣JiMJi)YLdtMdt∗(k^dt)YLdt′Mdt′(r^dt)χLdtSdt,Ldt′Sdt′Ji(kdt,rdt),(8)\begin{aligned} \Phi_{dt,M_dM_t}^{(32)}(1,2,3;4,5)={}&\frac{1}{\sqrt{v_{dt}}}\frac{4\pi}{k_{dt}r_{dt}}\sum_{\substack{L'_{dt}S'_{i}L_{dt}S_iJ_i\\M_{dt}M'_iM_iM_{J_i}}}i^{L_{dt}}(J_tM_tJ_dM_d|S_iM_i)\\ &\times[\psi_t(1,2,3)\otimes\psi_d(4,5)]_{S'_iM'_i}(L_{dt}M_{dt}S_iM_i|J_iM_{J_i})\\ &\times(L'_{dt}M'_{dt}S'_iM'_i|J_iM_{J_i})Y^*_{L_{dt}M_{dt}}(\hat{k}_{dt})Y_{L'_{dt}M'_{dt}}(\hat{r}_{dt})\chi^{J_i}_{L_{dt}S_{dt},L'_{dt}S'_{dt}}(k_{dt},r_{dt}), \tag*{(8)} \end{aligned}

with kdt=2μdtEdt/ℏk_{dt}=\sqrt{2\mu_{dt}E_{dt}}/\hbar (μdt\mu_{dt} and EdtE_{dt} are the reduced mass and centre-or-mass energy of the dd-tt system), rdtr_{dt} representing the radius-vector connecting centres of mass of deuteron and triton, and ψt(1,2,3)\psi_t(1,2,3) and ψd(4,5)\psi_d(4,5) being the triton and deuteron wave functions, respectively. The deuteron wave function will be represented below by the ss-wave component only: ψd(rnp)=φd(rnp)Y00(r^np)\psi_d(r_{np})=\varphi_d(r_{np})Y_{00}(\hat{r}_{np}).

The 4+1 partition function is given by

Φdt,MdMt(41)(1,2,3,4;5)=1vi4πkdtrαn∑ψα(1,2,3,4)×χ12σn(spin)(5)(JtMtJdMd∣SdtMdt)(LdtMdtSdtMdt∣JiMJi)×(LiMi12σn∣JiMJi)iLiYLdtMdt∗(k^dt)×YLiMi(r^αn)χLdtSdt,LiJi(ki,rαn),(9)\begin{aligned} \Phi_{dt,M_dM_t}^{(41)}(1,2,3,4;5)={}&\frac{1}{\sqrt{v_i}}\frac{4\pi}{k_{dt}r_{\alpha n}}\sum\psi_\alpha(1,2,3,4)\\ &\times\chi_{\frac{1}{2}\sigma_n}^{(\mathrm{spin})}(5)(J_tM_tJ_dM_d|S_{dt}M_{dt})(L_{dt}M_{dt}S_{dt}M_{dt}|J_iM_{J_i})\\ &\times(L_iM_i\tfrac{1}{2}\sigma_n|J_iM_{J_i})i^{L_i}Y^*_{L_{dt}M_{dt}}(\hat{k}_{dt})\\ &\times Y_{L_iM_i}(\hat{r}_{\alpha n})\chi^{J_i}_{L_{dt}S_{dt},L_i}(k_i,r_{\alpha n}), \tag*{(9)} \end{aligned}

where rαnr_{\alpha n} is the neutron coordinate-vector with respect to the centre of mass of α\alpha and the sum runs on all angular momenta and their projections, {JiMJiLdtMdtSdtMdtσnLαnMαn}\{J_iM_{J_i}L_{dt}M_{dt}S_{dt}M_{dt}\sigma_nL_{\alpha n}M_{\alpha n}\}. The kdtk_{dt} and kik_i in the dd-tt and α\alpha-nn partitions are related by the energy conservation,

ℏ2kdt22μdt+Q=ℏ2ki22μαn(10)\frac{\hbar^2k_{dt}^2}{2\mu_{dt}}+Q=\frac{\hbar^2k_i^2}{2\mu_{\alpha n}} \tag*{(10)}

where the QQ-value is 17.59 MeV. The representation (9) also includes the vi−1/2v_i^{-1/2} factor, where vi=ℏki/μαnv_i=\hbar k_i/\mu_{\alpha n} is the velocity in the α\alpha-nn partition. Because the entrance channel wave function does not have incoming flux in this partition, the asymptotics of χLdtSdt,LiJi(ki,r)\chi^{J_i}_{L_{dt}S_{dt},L_i}(k_i,r) contains outgoing waves only:

χLdtSdt,LiJi(ki,r)→−i2SLdtSdt,LiJiOLi(kir),(11)\chi^{J_i}_{L_{dt}S_{dt},L_i}(k_i,r)\to-\frac{i}{2}S^{J_i}_{L_{dt}S_{dt},L_i}O_{L_i}(k_ir), \tag*{(11)}

where Ol=Gl+iFlO_l=G_l+iF_l, with FlF_l and GlG_l being the regular and irregular Coulomb functions, and the SS-matrix element SLdtSdt,LiJiS^{J_i}_{L_{dt}S_{dt},L_i} describes transitions from the dd-tt channel to the α\alpha-nn channel with quantum numbers LiJiL_iJ_i and Si=1/2S_i=1/2.

Direct and exchange amplitudes

Because of the antisymmetrization requirement the matrix element that determines the transition between initial and final states has direct and exchange parts:

⟨Ψf∣MμMλ∣Ψi⟩=MμMλ;dir(kγ)+MμMλ;ex(kγ),(12)\langle\Psi_f|\mathcal{M}^{M\lambda}_\mu|\Psi_i\rangle=M^{M\lambda;\mathrm{dir}}_\mu(k_\gamma)+M^{M\lambda;\mathrm{ex}}_\mu(k_\gamma), \tag*{(12)}

where

MμMλ;dir(kγ)=⟨Φf(41)(1,2,3,4;5)∣MμMλ(kγ)∣Φi(41)(1,2,3,4;5)⟩+22⟨Φf(41)(1,2,3,4;5)∣MμMλ(kγ)∣Φi(32)(1,2,3;4,5)⟩,(13)M^{M\lambda;\mathrm{dir}}_\mu(k_\gamma)=\langle\Phi_f^{(41)}(1,2,3,4;5)|\mathcal{M}^{M\lambda}_\mu(k_\gamma)|\Phi_i^{(41)}(1,2,3,4;5)\rangle+2\sqrt{2}\langle\Phi_f^{(41)}(1,2,3,4;5)|\mathcal{M}^{M\lambda}_\mu(k_\gamma)|\Phi_i^{(32)}(1,2,3;4,5)\rangle, \tag*{(13)}

and

MμMλ;ex(kγ)=−4⟨Φf(41)(1,2,3,4;5)∣MμMλ(kγ)∣Φi(41)(1,2,3,5;4)⟩+32⟨Φf(41)(1,2,3,4;5)∣MμMλ(kγ)∣Φi(32)(3,4,5;1,2)⟩.(14)M^{M\lambda;\mathrm{ex}}_\mu(k_\gamma)=-4\langle\Phi_f^{(41)}(1,2,3,4;5)|\mathcal{M}^{M\lambda}_\mu(k_\gamma)|\Phi_i^{(41)}(1,2,3,5;4)\rangle+3\sqrt{2}\langle\Phi_f^{(41)}(1,2,3,4;5)|\mathcal{M}^{M\lambda}_\mu(k_\gamma)|\Phi_i^{(32)}(3,4,5;1,2)\rangle. \tag*{(14)}

In the exchange part the nucleons 4 and 5 in the 4+1 component of the initial wave function are swapped while the final state wave function keeps 4 and 5 in the original order. In the 3+2 component of the entrance channel wave function the nucleons 1 and 2 are swapped with 4 and 5.

M1 transition due to convection current

The convection–current-induced part MμMλ;conv\mathcal{M}_{\mu}^{M\lambda;\mathrm{conv}} of the M1 operator does not contain any spin. Therefore, it does not cause any M1 transitions from the strong S=3/2S=3/2 component in the entrance channel to the α\alpha-n final state with the spin S=1/2S=1/2. This term connects the dd-wave S=1/2S=1/2 state of the 4+14+1 partition in the entrance channel with the dd-wave final α\alpha-n state. The motivation behind considering this contribution is that it can add to increase of the d+t→α+n+γd+t\to\alpha+n+\gamma cross section at low γ\gamma-energy contributed to bremsstrahlung, discussed in [17].

The formal expression for the MμMλ;conv\mathcal{M}_{\mu}^{M\lambda;\mathrm{conv}} amplitude can be obtained easily by relating the [∇⃗jϕλμ(rj)]⋅L⃗j\left[\vec{\nabla}_{j}\phi_{\lambda\mu}(\boldsymbol{r}_{j})\right]\cdot\vec{L}_{j} operator in the first term of (1) to the 1i[∇⃗j×L⃗ϕλμ(rj)]⋅∇⃗j\frac{1}{i}\left[\vec{\nabla}_{j}\times\vec{L}\phi_{\lambda\mu}(\boldsymbol{r}_{j})\right]\cdot\vec{\nabla}_{j} operator from the convection-induced EλE\lambda transition amplitude and then using developments from [17]. Using (2) and (3) one obtains

[∇⃗jϕλμ(rj)]⋅L⃗j=−[L⃗jϕλμ(rj)]⋅∇⃗j.(15)\left[\vec{\nabla}_{j}\phi_{\lambda\mu}(\boldsymbol{r}_{j})\right]\cdot\vec{L}_{j}=-\left[\vec{L}_{j}\phi_{\lambda\mu}(\boldsymbol{r}_{j})\right]\cdot\vec{\nabla}_{j}. \tag*{(15)}

Since this operator should be sandwiched between the wave functions with a cluster structure, A=A1+A2A=A_{1}+A_{2}, the operator MμMλ;conv\mathcal{M}_{\mu}^{M\lambda;\mathrm{conv}} could be rewritten in a form similar to that from [19]:

MμMλ;conv=eℏmpcAA1A2(2λ+1)!!(λ+1)kγλ[Z1χ~λμ(A2Akγ,r)+(−)λZ2χ~λμ(A1Akγ,r)]⋅∇⃗r.(16)\mathcal{M}_{\mu}^{M\lambda;\mathrm{conv}}=\frac{e\hbar}{m_{p}c}\frac{A}{A_{1}A_{2}}\frac{(2\lambda+1)!!}{(\lambda+1)k_{\gamma}^{\lambda}}\left[Z_{1}\widetilde{\chi}_{\lambda\mu}\left(\frac{A_{2}}{A}k_{\gamma},r\right)+(-)^{\lambda}Z_{2}\widetilde{\chi}_{\lambda\mu}\left(\frac{A_{1}}{A}k_{\gamma},r\right)\right]\cdot\vec{\nabla}_{r}. \tag*{(16)}

Here, ZiZ_{i} are the charges of the clusters AiA_{i}, r=Rc.m.(A1)−Rc.m.(A2)\boldsymbol{r}=\boldsymbol{R}_{\mathrm{c.m.}}(A_{1})-\boldsymbol{R}_{\mathrm{c.m.}}(A_{2}) is the intercluster coordinate, connecting the centers of mass positioned at Rc.m.(Ai)\boldsymbol{R}_{\mathrm{c.m.}}(A_{i}), and χ~λμ=−L⃗ϕλμ\widetilde{\chi}_{\lambda\mu}=-\vec{L}\phi_{\lambda\mu}.

The direct contribution to the convection part of the MλM\lambda transition amplitude is obtained by using Eq. (32) from Appendix A and making standard Racah algebra manipulations with summation over the angular momentum projections. For α\alpha-n bremsstrahlung (with A2=1A_{2}=1 and Z2=0Z_{2}=0) it results in

⟨Φα,n(41)(1,2,3,4;5)∣MμMλ;conv∣Φd,t(41)(1,2,3,4;5)⟩=eℏmpc(2λ+1)!!(λ+1)kγλ∑LdtSiJiLfJfRLdtSiJiLfJf(Mλ);conv,dir×∑MiMJiMfMJfYLfMf(k^f)YLdtMdt∗(k^dt)×(LfMf12Mn∣JfMJf)(LdtMdtSiMi∣JiMJi)×(λμJiMJi∣JfMJf)(JtMtJdMd∣JiMJi).(17)\begin{aligned} \langle\Phi_{\alpha,n}^{(41)}(1,2,3,4;5)|\mathcal{M}_{\mu}^{M\lambda;\mathrm{conv}}|\Phi_{d,t}^{(41)}(1,2,3,4;5)\rangle ={}&\frac{e\hbar}{m_{p}c}\frac{(2\lambda+1)!!}{(\lambda+1)k_{\gamma}^{\lambda}} \sum_{L_{dt}S_{i}J_{i}L_{f}J_{f}}\mathcal{R}_{L_{dt}S_{i}J_{i}L_{f}J_{f}}^{(M\lambda);\mathrm{conv,dir}}\\ &\times\sum_{M_{i}M_{J_{i}}M_{f}M_{J_{f}}}Y_{L_{f}M_{f}}(\hat{\boldsymbol{k}}_{f})Y_{L_{dt}M_{dt}}^{*}(\hat{\boldsymbol{k}}_{dt})\\ &\times(L_{f}M_{f}\tfrac{1}{2}M_{n}|J_{f}M_{J_{f}})(L_{dt}M_{dt}S_{i}M_{i}|J_{i}M_{J_{i}})\\ &\times(\lambda\mu J_{i}M_{J_{i}}|J_{f}M_{J_{f}})(J_{t}M_{t}J_{d}M_{d}|J_{i}M_{J_{i}}). \tag*{(17)} \end{aligned}
RLdtSiJiLfJf(Mλ);conv,dir=Z1AA1A2(−)λ(4π)3/2vfviL^iJ^ikdtkf∑LiLf′iLi−Lf′L^f′W(λLiJf12;Lf′Ji)(DLiλLf′(1)Iβ;λ(1)(kγ)+DLiλLf′(2)Iβ;λ(2)(kγ)).(18)\mathcal{R}_{L_{dt}S_{i}J_{i}L_{f}J_{f}}^{(M\lambda);\mathrm{conv,dir}}=\frac{Z_{1}A}{A_{1}A_{2}}\frac{(-)^{\lambda}(4\pi)^{3/2}}{\sqrt{v_{f}v_{i}}}\frac{\widehat{L}_{i}\widehat{J}_{i}}{k_{dt}k_{f}}\sum_{L_{i}L_{f}^{\prime}}i^{L_{i}-L_{f}^{\prime}}\widehat{L}_{f}^{\prime}W(\lambda L_{i}J_{f}\tfrac{1}{2};L_{f}^{\prime}J_{i})\left(\mathcal{D}_{L_{i}\lambda L_{f}^{\prime}}^{(1)}I_{\beta;\lambda}^{(1)}(k_{\gamma})+\mathcal{D}_{L_{i}\lambda L_{f}^{\prime}}^{(2)}I_{\beta;\lambda}^{(2)}(k_{\gamma})\right). \tag*{(18)}

and contains the radial integrals

Iβ;λ(1)(kγ)=∫0∞dr rχLfLf′Jf(kf,r)jλ(15kγr)∂∂rχLdtSiLiJi(ki,r)r,(19)I_{\beta;\lambda}^{(1)}(k_{\gamma})=\int_{0}^{\infty}dr\,r\chi_{L_{f}L_{f}^{\prime}}^{J_{f}}(k_{f},r)j_{\lambda}\left(\tfrac{1}{5}k_{\gamma}r\right)\frac{\partial}{\partial r}\frac{\chi_{L_{dt}S_{i}L_{i}}^{J_{i}}(k_{i},r)}{r}, \tag*{(19)}

and

Iβ;λ(2)(kγ)=∫0∞drrχLfLf′Jf(kf,r)jλ(15kγr)χLdtSiLiJi(ki,r).(20)I_{\beta;\lambda}^{(2)}(k_{\gamma})=\int_{0}^{\infty}\frac{dr}{r}\chi_{L_{f}L_{f}^{\prime}}^{J_{f}}(k_{f},r)j_{\lambda}\left(\tfrac{1}{5}k_{\gamma}r\right)\chi_{L_{dt}S_{i}L_{i}}^{J_{i}}(k_{i},r). \tag*{(20)}

In this equations β\beta stands for the set of quantum numbers {LdtSiLiJi,LfLf′Jf;λ}\{L_{dt}S_{i}L_{i}J_{i},L_{f}L_{f}^{\prime}J_{f};\lambda\} and the coefficients DLiλLf′(1)\mathcal{D}_{L_{i}\lambda L_{f}^{\prime}}^{(1)} and DLiλLf′(2)\mathcal{D}_{L_{i}\lambda L_{f}^{\prime}}^{(2)} are given by Eq. (33) and (34) from the Appendix A, respectively. To get formal expressions for the exchange amplitude, which contains contributions from relatively small α\alpha-n separations, the Siegert approximation to the amplitude (17) could be used, in a similar fashion to what was done in [17]. However, as shown below, the direct convection-induced M1 transitions are small enough to not worrying about the exchange effects.

M1 transition due to magnetization current

The magnetization current can connect ss-wave states in partitions 3+23+2 and 4+14+1 with different spin. The intrinsic wave functions ψα\psi_{\alpha} and ψt\psi_{t} of α\alpha-particle and triton in the corresponding two different partitions of 5He{}^{5}\mathrm{He} contain different numbers of nucleons. To deal with this situation, fractional parentage expansion is used here to separate one nucleon from α\alpha-particle in the exit channel:

ψα(1,2,3,4)≈Itp(r⃗pt)[ψt(1,2,3)⊗χp(4)]00+Inh(r⃗nh)[ψh(1,2,3)⊗χn(4)]00.(21)\psi_{\alpha}(1,2,3,4)\approx I_{tp}(\vec{r}_{pt})[\psi_{t}(1,2,3)\otimes\chi_{p}(4)]_{00}+I_{nh}(\vec{r}_{nh})[\psi_{h}(1,2,3)\otimes\chi_{n}(4)]_{00}. \tag*{(21)}

which, according to standard practice, includes isospin Clebsch-Gordan coefficients into the radial parts of the overlap integrals Itp(r⃗pt)=Itp(rpt)Y00(r⃗pt)I_{tp}(\vec{r}_{pt})=I_{tp}(r_{pt})Y_{00}(\vec{r}_{pt}) and Inh(r⃗nh)=Inh(rnh)Y00(r⃗nh)I_{nh}(\vec{r}_{nh})=I_{nh}(r_{nh})Y_{00}(\vec{r}_{nh}). For dd-tt collisions, only the t+pt+p part of this expansion contributes to the M1 amplitude.

Using developments from the Appendix B and Appendix C one can show that the MμM1;magn,dir\mathcal{M}_{\mu}^{M1;\mathrm{magn,dir}} amplitude is given by Eq. (17) with λ=1\lambda=1, Si=3/2S_{i}=3/2 and with RLdtSiJiLfJf(Mλ);conv,dir\mathcal{R}_{L_{dt}S_{i}J_{i}L_{f}J_{f}}^{(M\lambda);\mathrm{conv,dir}} replaced by

RLdtSiJiLfJf(Mλ);magn⁡,dir⁡=−14kγvdtvf(4π)3/2kfkdt(−)Ji+λ−Jf×∑Ldt′Lf′Si′iLdt′−Lf′(−)Ldt′+λ′−Lf′(−)12+Si′2JiJ^iL^f′L^dt×cλλ′ILdt′Lf′(kγ){Ldt′Si′Jiλ′1λLf′12Jf},(22)\begin{aligned} \mathcal{R}^{(M\lambda);\operatorname{magn},\operatorname{dir}}_{L_{dt}S_iJ_iL_fJ_f}=-\frac{1}{4}\frac{k_\gamma}{\sqrt{v_{dt}v_f}}\frac{(4\pi)^{3/2}}{k_fk_{dt}}(-)^{J_i+\lambda-J_f} \\ &\times\sum_{L'_{dt}L'_fS'_i}i^{L'_{dt}-L'_f}(-)^{L'_{dt}+\lambda'-L'_f}(-)^{\frac{1}{2}+S'_i}\sqrt{2J_i}\hat{J}_i\hat{L}'_f\hat{L}_{dt} \\ &\times c_{\lambda\lambda'}\mathcal{I}_{L'_{dt}L'_f}(k_\gamma) \left\{\begin{matrix} L'_{dt}&S'_i&J_i\\ \lambda'&1&\lambda\\ L'_f&\frac{1}{2}&J_f \end{matrix}\right\}, \tag*{(22)} \end{aligned}

where

ILdt′Lf′(kγ)=(85)3∫0∞drαrαnXLfLf′Jf(kf,rα)×∫0∞drdtrdtFLdt′λ′Lf′(rdt,rα)XLdtSiLf′Si′Ji(kdt,rdt)(23)\begin{aligned} \mathcal{I}_{L'_{dt}L'_f}(k_\gamma)=\left(\frac{8}{5}\right)^3\int_0^\infty dr_\alpha r_\alpha n X^{J_f}_{L_fL'_f}(k_f,r_\alpha) \\ &\times\int_0^\infty dr_{dt}r_{dt}F_{L'_{dt}\lambda'L'_f}(r_{dt},r_\alpha)X^{J_i}_{L_{dt}S_iL'_fS'_i}(k_{dt},r_{dt}) \tag*{(23)} \end{aligned}

and FLdt′λ′Lf′(rdt,rα)F_{L'_{dt}\lambda'L'_f}(r_{dt},r_\alpha) is given in the Appendix C. As for the exchange contribution to M1 magnetization amplitude, Appendix D shows that its leading term is zero.

Differential cross section

The differential cross section describing the γ\gamma-spectrum from dd-tt collisions is obtained here following Ref. [20], which results in the expression

dσdEγ=Eγkf2π2ℏ3c3(2Jd+1)(2Jt+1)∑MdMtMn14π∫dk^dtdk^f×∑λμ12λ+1∣AμMλ⟨Ψnα,Mn∣MμMλ∣Ψdt,MdMt⟩∣2(24)\begin{aligned} \frac{d\sigma}{dE_\gamma}=\frac{E_\gamma k_f^2}{\pi^2\hbar^3c^3(2J_d+1)(2J_t+1)}\sum_{M_dM_tM_n}\frac{1}{4\pi}\int d\hat{k}_{dt}d\hat{k}_f \\ &\times\sum_{\lambda\mu}\frac{1}{2\lambda+1}\left|\mathcal{A}^{M\lambda}_\mu\langle\Psi_{n\alpha,M_n}|\mathcal{M}^{M\lambda}_\mu|\Psi_{dt,M_dM_t}\rangle\right|^2 \tag*{(24)} \end{aligned}

with

αMλ=−2π(λ+1)iλ+1kγλλ(2λ+1)(2λ−1)!!.(25)\alpha^{M\lambda}=-\frac{\sqrt{2\pi(\lambda+1)}i^{\lambda+1}k_\gamma^\lambda}{\sqrt{\lambda(2\lambda+1)(2\lambda-1)!!}}. \tag*{(25)}

As usual, the unpolarized cross section contains summation over spin projection MnM_n in the final state and averaging over the spin projections MdM_d and MtM_t of the colliding deuteron and triton in the entrance state. The matrix element ⟨Ψnα,Mn∣MμMλ∣Ψdt,MdMt⟩\langle\Psi_{n\alpha,M_n}|\mathcal{M}^{M\lambda}_\mu|\Psi_{dt,M_dM_t}\rangle is given by the same partial expansion as in Eq. (17) but with RLdtSiJiLfJf(Mλ);conv⁡,dir⁡\mathcal{R}^{(M\lambda);\operatorname{conv},\operatorname{dir}}_{L_{dt}S_iJ_iL_fJ_f} replaced by

RLdtSiJiLfJf(Mλ)(kγ)=RLdtSiJiLfJf(Mλ);conv⁡,dir⁡(kγ)+22RLdtSiJiLfJf(Mλ);magn⁡,dir⁡(kγ),(26)\begin{aligned} \mathcal{R}^{(M\lambda)}_{L_{dt}S_iJ_iL_fJ_f}(k_\gamma)=\mathcal{R}^{(M\lambda);\operatorname{conv},\operatorname{dir}}_{L_{dt}S_iJ_iL_fJ_f}(k_\gamma) \\ &+2\sqrt{2}\mathcal{R}^{(M\lambda);\operatorname{magn},\operatorname{dir}}_{L_{dt}S_iJ_iL_fJ_f}(k_\gamma), \tag*{(26)} \end{aligned}

where the factor of 222\sqrt{2} arises due to the antisymmetrization, see Eq. (13). After integrating over directions of the final state momentum kf\mathbf{k}_f and averaging over the incident momentum direction k^dt\hat{\mathbf{k}}_{dt}, followed by summation over angular momentum projections, one obtains

dσdEγ=e22π2kγkf2mp2c2∑LdtSiJiLfJf2Jf+1(2Jd+1)(2Jt+1)×∑λ1λ(λ+1)∣RLdtSiJiLfJf(Mλ)(kγ)∣2.(27)\begin{aligned} \frac{d\sigma}{dE_\gamma}=\frac{e^2}{2\pi^2}\frac{k_\gamma k_f^2}{m_p^2c^2}\sum_{L_{dt}S_iJ_iL_fJ_f}\frac{2J_f+1}{(2J_d+1)(2J_t+1)} \\ &\times\sum_\lambda\frac{1}{\lambda(\lambda+1)} \left|\mathcal{R}^{(M\lambda)}_{L_{dt}S_iJ_iL_fJ_f}(k_\gamma)\right|^2. \tag*{(27)} \end{aligned}

Numerical results

Numerical calculations for M1 transitions have been performed here for convection-induced bremsstrahlung between two dd-wave states of the α\alpha-nn partition and for the direct transition from the ss-wave dd-tt continuum state to the ss-wave α\alpha-nn state caused by magnetization current. In both cases the final α\alpha-nn wave function has been calculated from the potential model of [21] with Perey-correction for nonlocality of the range of 0.85 fm [22], in the same way as it was done in [17]. The introduction of the Perey factor is needed to simulate the five-body aspect of the α\alpha-nn collision which would force the channel function to be a solution of a nonlocal equation. As for the entrance-state wave functions, they are described below separately for convection and magnetization M1 transitions.

Since all experimental data on d+t→α+n+γd+t\to\alpha+n+\gamma reaction are given as branching ratios to the cross section of the γ\gamma-less reaction d+t→α+nd+t\to\alpha+n, results of all model calculations are also presented by this quantity, denoted as σdtγ(Eγ)/σdt\sigma_{d t\gamma}(E_\gamma)/\sigma_{dt}.

Convection-induced bremsstrahlung

Following [17], to model the intermediate dd-wave α\alpha-nn channel function the α\alpha-nn potential from [21] has been renormalized in this wave to obtain exactly the same phase shifts as those given by ab-initio calculation in [23]. Then the α\alpha-nn channel wave function was constructed by propagating the outgoing-wave-only asymptotic solution of the Schrödinger equation with the renormalized potential into the nuclear potential interior using the Numerov method.

Figure 1 shows the dd-wave bremsstrahlung spectrum calculated for Edt=50E_{dt}=50 keV. Since no resonances are present in the low-energy α\alpha-nn dd-wave scattering, the spectrum has one broad feature only. It also exhibits a sharp increase towards zero energy EγE_\gamma, which gives a divergent contribution to the integrated cross section. This is similar to the sharp increase of the d+t→α+n+γd+t\to\alpha+n+\gamma cross sections in E1 transitions, calculated in [17] and attributed there to unsuitability of the first-order perturbation theory at very small EγE_\gamma. The branching ratio, σdtγ(Eγ)/σdt=5.15×10−8\sigma_{dt\gamma}(E_\gamma)/\sigma_{dt}=5.15\times10^{-8} integrated above the lowest point Eγ=0.85E_\gamma=0.85 MeV is about three orders of magnitude smaller than the E1 value, which according to all measurement and previous theoretical calculations is around 5×10−55 \times10^{-5}. Therefore, convection-induced transitions can be safely neglected. The same conclusion is drawn after several modifications of the intermediate channel α\alpha-nn wave functions in the same fashion as described in [17]. The model sensitivity was not sufficiently high to warrant further investigation of this M1 path.

Figure 1. The differential branching ratio σdtγ(Eγ)/σdt\sigma_{dt\gamma}(E_\gamma)/\sigma_{dt} for convection-current-induced M1 transition for the dd-tt centre-of-mass energy Edt=50E_{dt}=50 keV.

Results for magnetization

In a sharp contrast with convection-induced M1 transition, the magnetization path turned out to be very sensitive to the model choice for the dd-tt scattering wave function. The first calculation, performed here, used the potential model from Dubovichenko et al [24], originally designed for a mirror scattering system, dd-3^{3}He. The dd-3^{3}He potential is represented in this model by one gaussian only, V(r)=−V0e−0.1r2V(r)=-V_{0}e^{-0.1r^{2}} with V0=34.5V_{0}=34.5 MeV. When applied to dd-tt scattering, it gives the position of the 3/2+3/2^{+} resonance at about 50 keV, which is close to the experimental value. With the dd-tt wave function at this energy the maximum of the M1 spectrum is also located around 16.7 MeV, but is almost twice the maximum of the E1 peak, while the integrated branching ratio, 4.1×10−54.1\times10^{-5}, is comparable to that of the E1 transition. This is unrealistically large.

As a next step, the optical model potential from Wu et al [25] was used to generate the dd-tt channel function. This potential is represented by a square well with a range of 3.616 fm both for the real and imaginary parts, with the corresponding real and imaginary well depths of 30 MeV and 49.64 keV, respectively. The corresponding differential branching ratio σdtγ(Eγ)/σdt\sigma_{dt\gamma}(E_\gamma)/\sigma_{dt}, also calculated at 50 keV, is compared in Figure 2 to that obtained with the wave function from Dubovichenko et al. It is significantly smaller for Eγ>7E_\gamma> 7 MeV, with the integrated branching ratio becoming 1.3×10−61.3\times10^{-6}, which is only about 3% of the E1 value. However, in the area of the 17-MeV peak the M1 transition is almost half that of the E1 (see gray dashed line in Figure 3), which still seems to be very unrealistic.

Figure 2. The differential branching ratio σdtγ(Eγ)/σdt\sigma_{dt\gamma}(E_\gamma)/\sigma_{dt} for magnetization-induced M1 transition caused obtained with three models of the dd-tt wave function, described in the main text, for the dd-tt centre-of-mass energy Edt=50E_{dt}=50 keV.

Figure 3. The differential branching ratio σdtγ(Eγ)/σdt\sigma_{dt\gamma}(E_\gamma)/\sigma_{dt} for magnetization-induced M1 transition, calculated at Edt=50E_{dt}=50 keV with optical folding model dd-tt potential from this work, in comparison to the E1 contribution from [17]. The grey dashed line represents M1 calculation with optical model dd-tt potential of Wu et al.

The M1 transition amplitude depends on how well the α\alpha-nn and dd-tt wave functions overlap in the integrand of Eq. (23). Because χLdtSiLf′Ji(ki,rdt)\chi^{J_i}_{L_{dt}S_iL'_f}(k_i,r_{dt}) has a node, ILdt′,λ′Lf′(kγ)\mathcal{I}_{L'_{dt},\lambda'L'_f}(k_\gamma) becomes very sensitive to cancellations between contributions from internal and external regions of the dd-tt wave function. The potential from Dubovichenko et al has unrealistically fast (gaussian) decrease with increasing dd-tt separations, while the potential from Wu et al is just a step function representing a sudden disappearance of the dd-tt interaction beyond a certain point, which could have affected the strength of the M1 transition. The deuteron is a weakly bound object, so that one can expect that its interaction potential with other nuclei could be characterized by a large diffuseness. Therefore, a folding model was employed here to construct the dd-tt potential at 50 keV. A simple deuteron wave function given by the Hulthén model [26] was chosen to calculate the deuteron density, which guarantees its correct asymptotic form. The triton density was presented by a single gaussian with the range of 1.37 fm chosen to provide the triton charge radius of 1.67 fm [27]. The nucleon-nucleon interaction is modelled by the same Hulthén potential, which has a realistic rate of decrease. Then it was assumed that both the real and imaginary parts of the optical potential have the same shape. Their depths have been fitted to provide the dd-tt phase shift of 16 degrees, close to the result from ab-initio calculations in [23], and the reaction cross section of 4.3 b, corresponding to astrophysical SS-factor of 27 MeV b. With this optical folding model the calculated σdtγ(Eγ)/σdt\sigma_{dt\gamma}(E_\gamma)/\sigma_{dt} ratio becomes 7.7×10−57.7\times10^{-5}, which is about 1% of the E1 value integrated above Eγ=10E_\gamma=10 MeV. The differential branching ratios from all three dd-tt interaction models are compared to each other in Figure 2. The minimum around 12 MeV in the folding model calculation is due to the almost complete cancellation between internal and external contributions from the dd-tt wave function at this point. In Figure 3 the M1 contribution calculated in the present optical folding model is compared to the E1 contribution from [17]. It gives about 2% at the main peak but is negligible everywhere below 15 MeV. It starts to dominate the E1 transition above Eγ≳17.5E_\gamma\gtrsim17.5 MeV leading to additional counts in the γ\gamma-spectrum in this area, which potentially could be mistaken for background.

To understand the difference between the three model calculations the wave functions from gaussian model of Dubovichenko et al, square well optical model of Wu et al and the present optical folding model are compared to each other in Figure 4. It is noticeable that the folding model wave function spreads out much further, with the node at a significantly larger dd-tt separation. This leads to a smaller absolute values of the wave function as compared to Dubovichenko et and and Wu et al, which is responsible for the small M1 contribution.

Figure 4. The dd-tt scattering wave function at Edt=50E_{dt}=50 keV calculated with gaussian potential from Dubovichenko et al, square well optical model from Wu et al and the optical folding model of the present paper.

Discussion and Conclusions

The present study of the M1 contribution to the d+t→α+n+γd+t \to\alpha+n+\gamma cross section shows that it is dominated by the direct transition from the ss-wave dd-tt partition of 5He{}^{5}\mathrm{He} to the ss-wave α\alpha-n partition in the final channel. The corresponding cross section has been found to be strongly dependent on the model choice for the dd-tt scattering wave function. In particular, it depends on how fast the dd-tt interaction potential decreases with the dd-tt separation. With the choice of two potential models from the literature, one having a gaussian decrease and another one being just a square well, the M1 transition becomes unrealistically large due to the large values of dd-tt channel function inside the dd-tt interaction. The optical folding potential, constructed in this work, provides a more spatially extended dd-tt interaction, which leads to a further spread and the correspondingly smaller values of the dd-tt wave function within the dd-tt interaction region. This leads to a much smaller M1 cross section.

The strong model dependence of the M1 transition is reminiscent of the situation with theoretical calculations of β\beta-delayed deuteron emission 6He→α+d+e−+νˉe{}^{6}\mathrm{He}\to\alpha+d+e^{-}+\bar{\nu}_{e}. Its measured branching ratio to the main 6Li{}^{6}\mathrm{Li} ground state decay branch is very small, of the order of 10−610^{-6}, however, potential model predictions are two orders of magnitude higher [28]. To bring theoretical predictions down, full three-body calculations of both 6He=α+n+n{}^{6}\mathrm{He}=\alpha+n+n and 6Li=α+n+p{}^{6}\mathrm{Li}=\alpha+n+p must be performed [29, 30], keeping all the quantum numbers included. Such a situation is due to the spatial parts of 6He{}^{6}\mathrm{He} and α+d\alpha+d wave functions being almost orthogonal, which could be achieved in full calculations only.

The operator ∑iσit3i\sum_i \sigma_i t_{3i} that causes the Galow-Teller transition 6He→α+d+e−+νˉe{}^{6}\mathrm{He}\to\alpha+d+e^{-}+\bar{\nu}_{e} is almost identical to the M1 operator of this paper if the long-wave approximation is done. Although all results for d+t→α+n+γd+t\to\alpha+n+\gamma shown here do not use the long-wave approximation, it has been checked numerically that this approximation does not change much the results. A similarity with the β\beta-delayed deuteron emission of 6He{}^{6}\mathrm{He} is also in the three-body nature of the dd-tt system, which could be considered as a strongly-bound triton plus neutron plus proton. The lessons learned from β\beta-decay studies tell us that for reliable calculations of the M1 transition one has to go beyond the two-body potential model description of the dd-tt motion and include into consideration such details as deuteron breakup and polarization in the sub-Coulomb dd-tt scattering. Also, more quantum numbers, such as the dd-wave state admixture in the incoming deuteron, should be included in the calculations.

This work has been motivated by a possibility of having more counts near the endpoint of the γ\gamma-spectrum from dd-tt collisions due to M1 transition to the final near-zero energy ss-wave α\alpha-nn state, which could be favored because of the absence of the centrifugal barrier. A realistic model of the dd-tt interaction suggests that the strength of the spectrum in this area is about 2% only, which would not induce much of a systematical error in the background subtraction. However, to confirm this conclusion, as a minimum, three-body treatment of the dd-tt collisions are required. In the long term, five-body calculations starting from realistic nucleon-nucleon interactions will be needed to clarify this matter.

ACKNOWLEDGMENTS

This work was part-funded by AWE, UK Ministry of Defence ©Crown owned copyright 2024/AWE. Remaining part of funding came from the United Kingdom Science and Technology Facilities Council (STFC) under Grant No. ST/V001108/1.

Appendix

Matrix element of L⃗ϕλμ⋅∇⃗\vec{L}_{\phi_{\lambda\mu}}\cdot\vec{\nabla}

To develop a formal expression for M1 amplitude given by Eq. (17) one needs the following matrix element:

⟨lfmf∣L⃗ϕλμ⋅∇⃗∣limi⟩=⟨lfmf∣L⃗ϕλμ⋅n⃗∣limi⟩∂∂r+⟨lfmf∣L⃗ϕλμ⋅∇⃗Ω∣limi⟩1r.(28)\langle l_fm_f \mid\vec{L}_{\phi_{\lambda\mu}}\cdot\vec{\nabla} \mid l_im_i\rangle= \langle l_fm_f \mid\vec{L}_{\phi_{\lambda\mu}}\cdot\vec{n} \mid l_im_i\rangle\frac{\partial}{\partial r} + \langle l_fm_f \mid\vec{L}_{\phi_{\lambda\mu}}\cdot\vec{\nabla}_{\Omega} \mid l_im_i\rangle\frac{1}{r}. \tag*{(28)}

Using

Lνϕλμ=λ(λ+1)j^λ(kr)∑μ′(λμ1ν∣λμ′)Yλμ′(r^)(29)L_{\nu}\phi_{\lambda\mu} = \sqrt{\lambda(\lambda+1)}\hat{j}_{\lambda}(kr)\sum_{\mu'}(\lambda\mu1\nu\mid\lambda\mu')Y_{\lambda\mu'}(\hat{r}) \tag*{(29)}

and

⟨lfmf∣∇⃗Ω∣limi⟩=(limiλμ∣lfmf)l^f−1Clilf,(30)\langle l_fm_f \mid\vec{\nabla}_{\Omega} \mid l_im_i\rangle= (l_im_i\lambda\mu\mid l_fm_f)\hat{l}_f^{-1}C_{l_il_f}, \tag*{(30)}
Clilf=−(lili+1δlf,li+1+(li+1)liδlf,li−1),(31)C_{l_il_f}=-\left(l_i\sqrt{l_i+1}\delta_{l_f,l_i+1}+(l_i+1)\sqrt{l_i}\delta_{l_f,l_i-1}\right), \tag*{(31)}

one obtains

⟨lfmf∣L⃗ϕλμ⋅∇⃗∣limi⟩=λ^(limiλμ∣lfmf)×14πj^λ(kr)(Dliλlf(1)∂∂r+Dliλlf(2)1r),(32)\langle l_fm_f \mid\vec{L}_{\phi_{\lambda\mu}}\cdot\vec{\nabla} \mid l_im_i\rangle= \hat{\lambda}(l_im_i\lambda\mu\mid l_fm_f) \times\frac{1}{\sqrt{4\pi}}\hat{j}_{\lambda}(kr)\left(\mathcal{D}^{(1)}_{l_i\lambda l_f}\frac{\partial}{\partial r}+\mathcal{D}^{(2)}_{l_i\lambda l_f}\frac{1}{r}\right), \tag*{(32)}

where

Dliλlf(1)=(−)λλ(λ+1)∑L(L010∣li0)(lf0λ0∣L0)W(li1lfλ;Lλ)(33)\mathcal{D}^{(1)}_{l_i\lambda l_f}=(-)^{\lambda}\sqrt{\lambda(\lambda+1)}\sum_L(L010\mid l_i0)(l_f0\lambda0\mid L0)W(l_i1l_f\lambda;L\lambda) \tag*{(33)}

and

Dliλlf(2)=(−)λ+1λ(λ+1)(2λ+1)∑LClilL(lf0λ0∣L0)W(li1lfλ;Lλ).(34)\mathcal{D}^{(2)}_{l_i\lambda l_f}=(-)^{\lambda+1}\sqrt{\lambda(\lambda+1)(2\lambda+1)} \sum_L C_{l_il}L(l_f0\lambda0\mid L0)W(l_i1l_f\lambda;L\lambda). \tag*{(34)}

Spin and isospin matrix elements

The contribution from nucleons inside the triton is characterized by eik⃗γr⃗je^{i\vec{k}_{\gamma}\vec{r}_j}. The coordinate of each nucleon is r⃗j=r⃗j′−r⃗t\vec{r}_j=\vec{r}'_j-\vec{r}_t, where r⃗j′\vec{r}'_j is its position with respect to the triton centre of mass and r⃗t=−25r⃗dt\vec{r}_t=-\frac{2}{5}\vec{r}_{dt} is the position of the latter with respect to the 5He{}^{5}\mathrm{He} centre of mass. Because triton is spatially confined one can assume that eik⃗γr⃗j≈1e^{i\vec{k}_{\gamma}\vec{r}_j}\approx1. In this approximation the M1 magnetization-induced amplitude does not contain spatial variables and, therefore, one has to consider matrix elements in the spin space only. There are three such matrix elements, the first of which is

∑i=1,2,3⟨ψt(1,2,3)∣Sμ,igsi∣ψt(1,2,3)⟩=12(gn+gp)⟨ψt∣Sμ∣ψt⟩+(gn−gp)∑i=1,2,3⟨ψt(1,2,3)∣Sμ,it3i∣ψt(1,2,3)⟩(35)\sum_{i=1,2,3}\langle\psi_t(1,2,3)\mid S_{\mu,i}g_{s_i}\mid\psi_t(1,2,3)\rangle =\frac{1}{2}(g_n+g_p)\langle\psi_t\mid S_{\mu}\mid\psi_t\rangle +(g_n-g_p)\sum_{i=1,2,3}\langle\psi_t(1,2,3)\mid S_{\mu,i}t_{3i}\mid\psi_t(1,2,3)\rangle \tag*{(35)}

To evaluate the second term it is assumed here that the triton wave functions is dominated by the fully antisymmetric in spin-isospin space component [3]S=12MST=12MT\left[3\right]S=\frac{1}{2}M_ST=\frac{1}{2}M_T described by the Young’s tableau [f]=[3][f]=[3] with spin and isospin equal to 1/21/2 and the spin and isospin projections of MSM_S and MTM_T, respectively. Using fractional parentage expansion to extract the mm-th nucleon from this spin-isospin component:

∣[3]S=12MST=12MT⟩=∑[f2]S2T2M2MT2⟨[3]1212∣[f2]S2T2⟩⟨[f2]S2M2T2MT2∣[f2]S2M2T2MT2⟩×(S2M212σ∣12MS)(T2MT212τ∣12MT)χ12σspin(m)χ12τisospin(m),(36)\left|[3]S=\frac{1}{2}M_ST=\frac{1}{2}M_T\right\rangle =\sum_{[f_2]S_2T_2M_2M_{T_2}}\langle[3]\frac{1}{2}\frac{1}{2}\mid[f_2]S_2T_2\rangle \langle[f_2]S_2M_2T_2M_{T_2}\mid[f_2]S_2M_2T_2M_{T_2}\rangle \times(S_2M_2\frac{1}{2}\sigma\mid\frac{1}{2}M_S)(T_2M_{T_2}\frac{1}{2}\tau\mid\frac{1}{2}M_T)\chi^{\mathrm{spin}}_{\frac{1}{2}\sigma}(m)\chi^{\mathrm{isospin}}_{\frac{1}{2}\tau}(m), \tag*{(36)}

where ⟨[3]1212∣[f2]S2T2⟩\langle[3]\frac{1}{2}\frac{1}{2}\mid[f_2]S_2T_2\rangle is the fractional parentage expansion coefficient and ⟨[f2]S2M2T2MT2⟩\langle[f_2]S_2M_2T_2M_{T_2}\rangle is the two-body spin-isospin function, and then performing all summations over spin and isospin projections, one obtains

⟨[3]12MS′12MT′∣∑i=1,2,3Sμ,it3i∣[3]12MS12MT⟩=3⟨[3]12MS′12MT′∣Sμ,mt3m∣[3]12MS12MT⟩=MTδMT,MT′⟨ψt∣Sμ∣ψt⟩.(37)\left\langle[3]\frac{1}{2}M'_S\frac{1}{2}M'_T\left|\sum_{i=1,2,3}S_{\mu,i}t_{3i}\right|[3]\frac{1}{2}M_S\frac{1}{2}M_T\right\rangle =3\left\langle[3]\frac{1}{2}M'_S\frac{1}{2}M'_T\right|S_{\mu,m}t_{3m}\left|[3]\frac{1}{2}M_S\frac{1}{2}M_T\right\rangle =M_T\delta_{M_T,M'_T}\langle\psi_t\mid S_{\mu}\mid\psi_t\rangle. \tag*{(37)}

Therefore, for the isospin projection MT=1/2M_T = 1/2, associated with triton,

∑i=1,2,3⟨ψt(1,2,3)∣Sμ,igsi∣ψt(1,2,3)⟩=gn⟨ψt∣Sμ∣ψt⟩=12gn(12Mt′1μ|12Mt)⟨12∥S∥12⟩.(38)\sum_{i=1,2,3} \langle\psi_t(1,2,3) \mid S_{\mu,i}g_{si} \mid\psi_t(1,2,3) \rangle = g_n \langle\psi_t \mid S_\mu\mid\psi_t \rangle = \frac{1}{\sqrt{2}}g_n \left(\frac{1}{2}M_t'1\mu\middle| \frac{1}{2}M_t\right)\left\langle\frac{1}{2} \middle\| S \middle\| \frac{1}{2}\right\rangle. \tag*{(38)}

Then, supplementing the triton wave function with spin functions of the remaining neutron and proton, one has

\begin{aligned} &\langle[[\psi_t \otimes\chi^{\mathrm{spin}}_{\frac{1}{2}}(p)]_0 \otimes\chi^{\mathrm{spin}}_{\frac{1}{2}}(n)]_{\frac{1}{2}\sigma_n} \sum_{i=1,2,3} S_{\mu,i}g_{si}\\ &\qquad\times|\chi^{\mathrm{spin}}_{\frac{1}{2}}[\psi_t \otimes\chi^{\mathrm{spin}}_1(p,n)]_{S_iM_{S_i}}\rangle = g_n\frac{1}{\sqrt{2}}(1\mu S_iM_{S_i}\middle|\frac{1}{2}\sigma_n). \tag*{(39)} \end{aligned}

For i=4i = 4, which is a proton in deuteron, the spin part is simpler,

\begin{aligned} &\langle[[\psi_t \otimes\chi^{\mathrm{spin}}_{\frac{1}{2}}(p)]_0 \otimes\chi^{\mathrm{spin}}_{\frac{1}{2}}(n)]_{\frac{1}{2}\sigma_n}\middle|S_\mu(p)g_{sp}\\ &\qquad\times|[\psi_t \otimes\chi^{\mathrm{spin}}_1(p,n)]_{S_iM_{S_i}}\rangle = g_p\sqrt{\frac{1}{2}}(1\mu S_iM_{S_i}\middle|\frac{1}{2}\sigma_n), \tag*{(40)} \end{aligned}

which is exactly the same as in the case of i=1,2,3i = 1, 2, 3 except for the neutron gyromagnetic factor gng_n being replaced by the proton gyromagnetic factor gpg_p.

For i=5i = 5, which is a neutron in deuteron, the spin matrix element is

\begin{aligned} &\langle[[\psi_t \otimes\chi^{\mathrm{spin}}_{\frac{1}{2}}(p)]_0 \otimes\chi^{\mathrm{spin}}_{\frac{1}{2}}(n)]_{\frac{1}{2}\sigma_n}\middle|S_\mu(n)g_{sn}\\ &\qquad\times|[\psi_t \otimes\chi^{\mathrm{spin}}_1(p,n)]_{S_iM_{S_i}}\rangle = W(\frac{1}{2}\frac{1}{2}S_i;01)\langle\frac{1}{2}\middle\|S\middle\|\frac{1}{2}\rangle(\frac{1}{2}M_{S_i}1\mu\middle|\frac{1}{2}\sigma_n), \tag*{(41)} \end{aligned}

which is non-zero only for Si=1/2S_i = 1/2. Given that the M1 transitions considered in this paper come from the resonance component with spin Si=3/2S_i = 3/2, the contribution from nucleon 5 (which is the neutron in deuteron) is set to zero.

Finally, the matrix elements (39) and (40) should be further multiplied by the isospin part of the M1 matrix element, which is

\begin{aligned} &\langle\chi^{\mathrm{isospin}}_{\frac{1}{2}\frac{1}{2}}(1,2,3)\chi^{\mathrm{isospin}}_{\frac{1}{2}-\frac{1}{2}}(4)\chi^{\mathrm{isospin}}_{\frac{1}{2}\frac{1}{2}}(n)\\ &\qquad\times|\chi^{\mathrm{isospin}}_{\frac{1}{2}\frac{1}{2}}(1,2,3)\chi^{\mathrm{isospin}}_{00}(p,n)\rangle =(\frac{1}{2}\ -\frac{1}{2}\middle|00)=-\frac{1}{\sqrt{2}}. \tag*{(42)} \end{aligned}

Spatial matrix element for M1 magnetization amplitude

The calculate the M1 amplitude the integration should be performed either over dr⃗αndr⃗ptd\vec r_{\alpha n}d\vec r_{pt} or over dr⃗npdr⃗dtd\vec r_{np}d\vec r_{dt}. This task is simplified if integration is performed over dr⃗αndr⃗dtd\vec r_{\alpha n}d\vec r_{dt} - the variables of the continuum state wave functions. Because r⃗tp=85r⃗dt−45r⃗αn\vec r_{tp} = \frac{8}{5}\vec r_{dt} - \frac{4}{5}\vec r_{\alpha n} and r⃗np=85r⃗αn−65r⃗dt\vec r_{np} = \frac{8}{5}\vec r_{\alpha n} - \frac{6}{5}\vec r_{dt}, one has dr⃗αndr⃗pt=(35)3dr⃗αndr⃗dtd\vec r_{\alpha n}d\vec r_{pt} = \left(\frac{3}{5}\right)^3d\vec r_{\alpha n}d\vec r_{dt}. Then the product of the radial parts of the overlap IptI_{pt} and the deuteron wave function ψd\psi_d could be represented as

Ipt(rpt)φd(rnp)=4π∑lmFl(rdt,rαn)Ylm∗(r^dt)Ylm(r^αn),I_{pt}(r_{pt})\varphi_d(r_{np}) = 4\pi\sum_{lm}F_l(r_{dt},r_{\alpha n})Y^*_{lm}(\hat r_{dt})Y_{lm}(\hat r_{\alpha n}),
Fl(rdt,rαn)=12∫−11dμ Pl(μ)φd(∣65rdt−85rαn∣)×Ipt(∣85rdt−45rαn∣).(43)F_l(r_{dt},r_{\alpha n}) = \frac{1}{2}\int_{-1}^{1}d\mu\,P_l(\mu)\varphi_d\left(\left|\frac{6}{5}r_{dt}-\frac{8}{5}r_{\alpha n}\right|\right) \times I_{pt}\left(\left|\frac{8}{5}r_{dt}-\frac{4}{5}r_{\alpha n}\right|\right). \tag*{(43)}

Using the results for both spin and isospin matrix elements derived in the subsection above, one can obtain that the M1 amplitude contains the matrix element over the angular variables

F=−i∑lmFl(rdt,rαn)⟨YLf′Mf′(r^αn)Ylm(r^dt)|gn[∇⃗t×L⃗tϕλμ(rt))ν+gp[∇⃗p×L⃗pϕλμ(rp))ν|YLdt′Mdt′(r^dt)Ylm(r^αn)⟩.(44)F=-i\sum_{lm}F_l(r_{dt},r_{\alpha n})\left\langle Y_{L'_fM'_f}(\hat r_{\alpha n})Y_{lm}(\hat r_{dt})\middle|g_n\left[\vec\nabla_t\times\vec L_t\phi_{\lambda\mu}(r_t)\right)_\nu+g_p\left[\vec\nabla_p\times\vec L_p\phi_{\lambda\mu}(r_p)\right)_\nu\middle|Y_{L'_{dt}M'_{dt}}(\hat r_{dt})Y_{lm}(\hat r_{\alpha n})\right\rangle. \tag*{(44)}

The first term is further developed using

⟨Ylm(r^dt)|[∇⃗t×L⃗tϕλμ(rt))ν|YLdt′Mdt′(r^dt)⟩=k∑λ′μ′(−)λ′cλλ′(λμ1ν∣λ′μ′)jλ′(25kγrdt)×(Ldt′0λ′0∣Lf′0)L^dt′λ^′4πL^f′,(45)\begin{aligned} \left\langle Y_{lm}(\hat r_{dt})\middle|\left[\vec\nabla_t\times\vec L_t\phi_{\lambda\mu}(r_t)\right)_\nu\middle|Y_{L'_{dt}M'_{dt}}(\hat r_{dt})\right\rangle ={}&k\sum_{\lambda'\mu'}(-)^{\lambda'}c_{\lambda\lambda'}(\lambda\mu1\nu|\lambda'\mu')j_{\lambda'}\left(\frac{2}{5}k_\gamma r_{dt}\right)\\ &\times(L'_{dt}0\lambda'0|L'_f0)\frac{\hat L'_{dt}\hat\lambda'}{\sqrt{4\pi}\hat L'_f}, \tag*{(45)} \end{aligned}

with

cλλ′={−(λ+1)λ2λ−1,λ′=λ−1,0,λ′=λ,λλ+12λ+3,λ′=λ+1.(46)c_{\lambda\lambda'}= \begin{cases} -(\lambda+1)\sqrt{\dfrac{\lambda}{2\lambda-1}}, & \lambda'=\lambda-1,\\ 0, & \lambda'=\lambda,\\ \lambda\sqrt{\dfrac{\lambda+1}{2\lambda+3}}, & \lambda'=\lambda+1. \end{cases} \tag*{(46)}

To develop the second term of (44), one has to use a general expansion for r=α1r1+α2r2\mathbf{r}=\alpha_{1}\mathbf{r}_{1}+\alpha_{2}\mathbf{r}_{2}

−i[∇×L~ϕμ(r)]ν=4πk∑λ′μ′λ1λ2μ1μ2i−λ1+λ2−λ′cλλ′(λ10λ20∣λ′0)jλ1(α1kr1)jλ2(α2kr2)×(λ1μ1λ′μ′∣λ2μ2)(λμ1ν∣λ′μ)Yλ1μ1∗(r^1)Yλ2μ2(r^2)(47)-\mathrm{i}\left[\nabla\times\widetilde{\mathbf{L}}_{\phi\mu}(\mathbf{r})\right]_{\nu}=\sqrt{4\pi k}\sum_{\lambda'\mu'\lambda_{1}\lambda_{2}\mu_{1}\mu_{2}}\mathrm{i}^{-\lambda_{1}+\lambda_{2}-\lambda'}c_{\lambda\lambda'}\left(\lambda_{1}0\lambda_{2}0\mid\lambda'0\right)j_{\lambda_{1}}(\alpha_{1}kr_{1})j_{\lambda_{2}}(\alpha_{2}kr_{2})\times\left(\lambda_{1}\mu_{1}\lambda'\mu'\mid\lambda_{2}\mu_{2}\right)\left(\lambda\mu_{1}\nu\mid\lambda'\mu\right)Y_{\lambda_{1}\mu_{1}}^{*}(\widehat{\mathbf{r}}_{1})Y_{\lambda_{2}\mu_{2}}(\widehat{\mathbf{r}}_{2}) \tag*{(47)}

In the case of rp=65rdt−45rαn\mathbf{r}_{p}=\frac{6}{5}\mathbf{r}_{dt}-\frac{4}{5}\mathbf{r}_{\alpha n} the final result for FF becomes

F=kL^dt′4π∑λ′μ′cλλ′FLdt′λ′Lf′(rdt,rαn)(λμ1ν∣λ′μ′)(λ′μ′Ldt′Mdt′∣Lf′Mf′),(48)F=k\frac{\widehat{L}'_{dt}}{\sqrt{4\pi}}\sum_{\lambda'\mu'}c_{\lambda\lambda'}F_{L'_{dt}\lambda'L'_{f}}(\mathbf{r}_{dt},\mathbf{r}_{\alpha n})\left(\lambda\mu_{1}\nu\mid\lambda'\mu'\right)\left(\lambda'\mu'L'_{dt}M'_{dt}\mid L'_{f}M'_{f}\right), \tag*{(48)}

where

FLdt′λ′Lf′(rdt,rαn)=gpGLdt′λ′Lf′(rdt,rαn)+gn(Ldt′0λ′0∣Lf′0)jλ′(65krdt)FLf′(rdt,rαn)(49)F_{L'_{dt}\lambda'L'_{f}}(\mathbf{r}_{dt},\mathbf{r}_{\alpha n})=g_{p}G_{L'_{dt}\lambda'L'_{f}}(\mathbf{r}_{dt},\mathbf{r}_{\alpha n})+g_{n}\left(L'_{dt}0\lambda'0\mid L'_{f}0\right)j_{\lambda'}\left(\frac{6}{5}kr_{dt}\right)F_{L'_{f}}(\mathbf{r}_{dt},\mathbf{r}_{\alpha n}) \tag*{(49)}

with

GLdt′λ′Lf′(rdt,rαn)=∑λ1λ2lCλ1λ2lFl(rdt,rαn)×jλ1(65krdt)jλ2(45krαn)(50)G_{L'_{dt}\lambda'L'_{f}}(\mathbf{r}_{dt},\mathbf{r}_{\alpha n})=\sum_{\lambda_{1}\lambda_{2}l}C_{\lambda_{1}\lambda_{2}l}F_{l}(\mathbf{r}_{dt},\mathbf{r}_{\alpha n})\times j_{\lambda_{1}}\left(\frac{6}{5}kr_{dt}\right)j_{\lambda_{2}}\left(\frac{4}{5}kr_{\alpha n}\right) \tag*{(50)}
Cλ1λ2l=∑λ1λ2 iλ1+λ2−λ′λ^12λ^22(λ10Ldt′0∣l0)(λ20Lf′0∣l0)(λ10λ20∣λ′0)W(λ′λ1Lf′l;λ2Ldt′).(51)C_{\lambda_{1}\lambda_{2}l}=\sum_{\lambda_{1}\lambda_{2}\,} \mathrm{i}^{\lambda_{1}+\lambda_{2}-\lambda'}\widehat{\lambda}_{1}^{2}\widehat{\lambda}_{2}^{2}\left(\lambda_{1}0L'_{dt}0\mid l0\right)\left(\lambda_{2}0L'_{f}0\mid l0\right)\left(\lambda_{1}0\lambda_{2}0\mid\lambda'0\right)W(\lambda'\lambda_{1}L'_{f}l;\lambda_{2}L'_{dt}). \tag*{(51)}

Then standard algebra manipulations lead to (22) and (23).

Exchange contribution to magnetization

The simplest way to calculate the exchange contribution

⟨Φαn,Mn(41)(1,2,3,4;5)|Mdt,MdMtλ(3,4,5;1,2)⟩(52)\left\langle\Phi_{\alpha_{n},M_{n}}^{(41)}(1,2,3,4;5)\middle|M_{dt,M_{d}M_{t}}^{\lambda}(3,4,5;1,2)\right\rangle \tag*{(52)}

to the M1 amplitude is to use the long-wave approximation, justified by small values of the kγrk_{\gamma}r contributing to the process. In this approximation, one has to deal with matrix elements of the operator ∑igsiSμ,i=(gn+gp)∑iSμ,i+(gn−gp)∑iSμ,it3i\sum_{i}g_{s_{i}}S_{\mu,i}=(g_{n}+g_{p})\sum_{i}S_{\mu,i}+(g_{n}-g_{p})\sum_{i}S_{\mu,i}t_{3i}. This could be achieved by making fractional parentage decomposition of the α\alpha-particle wave function in the exit channel into the product of two deuteron wave functions keeping only the terms corresponding to the deuteron ground state wave function,

ψα(1,2,3,4)≈[ψd(1,2)×ψd(3,4)]00Idd(r~dd),(53)\psi_{\alpha}(1,2,3,4)\approx[\psi_{d}(1,2)\times\psi_{d}(3,4)]_{00}I_{dd}(\widetilde{\mathbf{r}}_{dd}), \tag*{(53)}

with the isospin Clebsch-Gordon coefficient embedded in the overlap function Idd(r~dd)I_{dd}(\widetilde{\mathbf{r}}_{dd}). The triton wave function in the initial state is also expanded as

ψt(3,4,5)=Ind(r~nd)φd(r~34)[χ1/2spin(3,4)χ1/2spin(5)]JtMt(54)\psi_{t}(3,4,5)=I_{nd}(\widetilde{\mathbf{r}}_{nd})\varphi_{d}(\widetilde{\mathbf{r}}_{34})[\chi_{1/2}^{\mathrm{spin}}(3,4)\chi_{1/2}^{\mathrm{spin}}(5)]_{J_{t}M_{t}} \tag*{(54)}

Then considering only the ss-wave components in the deuteron wave functions and overlaps one has to calculate the matrix element

⟨[χ1/2spin(1,2)×χ1/2spin(3,4)]00χ1/2spin(5)X0isospin|∑i=15Sμ,igsi|[[χ1/2spin(3,4)×χ1/2spin(5)]Jt×χ1/2spin(1,2)]SiMSiXisospin⟩,(55)\left\langle[\chi_{1/2}^{\mathrm{spin}}(1,2)\times\chi_{1/2}^{\mathrm{spin}}(3,4)]_{00}\chi_{1/2}^{\mathrm{spin}}(5)X_{0}^{\mathrm{isospin}}\middle|\sum_{i=1}^{5}S_{\mu,i}g_{s_{i}}\middle|[[\chi_{1/2}^{\mathrm{spin}}(3,4)\times\chi_{1/2}^{\mathrm{spin}}(5)]_{J_{t}}\times\chi_{1/2}^{\mathrm{spin}}(1,2)]_{S_{i}M_{S_{i}}}X^{\mathrm{isospin}}\right\rangle, \tag*{(55)}

where XisospinX^{\mathrm{isospin}} is the isospin part of the five-body system. It is easy to show that

∑i=1,2⟨χ1Md′spin(1,2)χ00isospin(1,2)|Sμ,it3i|χ1Mdspin(1,2)χ00isospin(1,2)⟩=0(56)\sum_{i=1,2}\left\langle\chi_{1M'_{d}}^{\mathrm{spin}}(1,2)\chi_{00}^{\mathrm{isospin}}(1,2)\middle|S_{\mu,i}t_{3i}\middle|\chi_{1M_{d}}^{\mathrm{spin}}(1,2)\chi_{00}^{\mathrm{isospin}}(1,2)\right\rangle=0 \tag*{(56)}

and

∑i=1,2⟨χ1Md′spin(1,2)χ00isospin(1,2)|Sμ,i|χ1Mdspin(1,2)χ00isospin(1,2)⟩=−2(1μ1Md∣1Md′).(57)\sum_{i=1,2}\left\langle\chi_{1M'_{d}}^{\mathrm{spin}}(1,2)\chi_{00}^{\mathrm{isospin}}(1,2)\middle|S_{\mu,i}\middle|\chi_{1M_{d}}^{\mathrm{spin}}(1,2)\chi_{00}^{\mathrm{isospin}}(1,2)\right\rangle=-\sqrt{2}\left(1\mu1M_{d}\mid1M'_{d}\right). \tag*{(57)}

The same is true for nucleons 3 and 4. Also,

⟨χ12σ5′spinχ1212isospin(5)|Sμgs5|χ12σ5spinχ1212isospin(5)⟩=gn(12σ51μ|12σ5′)12⟨12|S|12⟩(58)\left\langle\chi_{\frac{1}{2}\sigma'_{5}}^{\mathrm{spin}}\chi_{\frac{1}{2}\frac{1}{2}}^{\mathrm{isospin}}(5)\middle|S_{\mu}g_{s_{5}}\middle|\chi_{\frac{1}{2}\sigma_{5}}^{\mathrm{spin}}\chi_{\frac{1}{2}\frac{1}{2}}^{\mathrm{isospin}}(5)\right\rangle=g_{n}\left(\frac{1}{2}\sigma_{5}1\mu\middle|\frac{1}{2}\sigma'_{5}\right)\frac{1}{\sqrt{2}}\left\langle\frac{1}{2}\middle|S\middle|\frac{1}{2}\right\rangle \tag*{(58)}

Then, after further algebraic manipulations, one obtains that the spin-isospin latrix element (55) is equal to

12(gn+gp)(12MSi1μ|12σnf)∑S(−)S+SiS^iS^W(1112S;1Si)[(−)Jt+1−SiδSJt+(−)3/2−SS^J^tW(112Si1;SJt)]+(−)Jt+1−SignW(11Si12;0Jt)J^t(12MSi1μ|12σnf)(59)\frac{1}{2}(g_n+g_p)\left(\frac{1}{2}M S_i 1\mu\middle|\frac{1}{2}\sigma n_f\right)\sum_S(-)^{S+S_i}\hat{S}_i\hat{S}W\left(11\frac{1}{2}S;1S_i\right)\left[(-)^{J_t+1-S_i}\delta_{SJ_t}+(-)^{3/2-S}\hat{S}\hat{J}_tW\left(1\frac{1}{2}S_i1;SJ_t\right)\right] +(-)^{J_t+1-S_i}g_nW\left(11S_i\frac{1}{2};0J_t\right)\hat{J}_t\left(\frac{1}{2}M S_i 1\mu\middle|\frac{1}{2}\sigma n_f\right) \tag*{(59)}

For Si=3/2S_i = 3/2 and Jt=1/2J_t = 1/2 it is equal to zero.

References

  1. [1]M.J. Moran and B. Chang, MOLE: A new high-energy gamma-ray diagnostic. United States: N. p., 1992. Web. doi:10.2172/5656147DOI
  2. [2]W. Buss, H. Wäffler, and B. Ziegler, Radiative capture of deuterons by 3H, Physics Letters 4, 198 (1963).
  3. [3]A. Kosiara and H. Willard, Phys. Lett. 15, B 32, 99 (1970).
  4. [4]V. M. Bezotosnyi, V. A. Zhm ailo, L. M. Surov, and M. S.Shvetsov, Sov. J. Nucl. Phys. 10, 127 (1970).
  5. [5]F. E. Cecil and F. J. Wilkinson, Phys. Rev. Lett. 53, 767 (1984).
  6. [6]G. L. Morgan, P. W. Lisowski, S. A. Wender, R. E. Brown, N. Jarmie, J. F. Wilkerson, and D. M. Drake, Phys. Rev. C 33, 1224 (1986).
  7. [7]J. E. Kammeraad, J. Hall, K. E. Sale, C. A. Barnes, S. E. Kellogg, and T. R. Wang, Phys. Rev. C 47, 29 (1993).
  8. [8]Y. Kim et al, Phys. Rev. C 85, 061601 (2012).
  9. [9]J. Jeet, A. B. Zylstra, M. Rubery, Y. Kim, K. D. Meaney, C. Forrest, V. Glebov, C. J. Horsfield, A. M. McEvoy, and H. W. Herrmann, Phys. Rev. C 104, 054611 (2021).
  10. [10]K. D. Meaney, Y. Kim, H. Geppert-Kleinrath, H. W. Herrmann, A. S. Moore, E. P. Hartouni, D. J. Schlossberg, E. Mariscal, J. Carrera, and J. A. Church, Phys. Plasmas 28, 102702 (2021).
  11. [11]Z. Mohamed, Y. Kim, and J. Knauer, Front. Phys. 10, 10.3389/fphy.2022.944339 (2022).
  12. [12]Z. L. Mohamed, Y. Kim, J. P. Knauer, and M. S. Rubery, Phys. Rev. C 107, 014606 (2023).
  13. [13]K. D. Meaney, M. W. Paris, G. M. Hale, Y. Kim, J. Kuczek, and A. C. Hayes, Phys. Rev. C 109, 034003 (2024)
  14. [14]A. Dal Molin, G. Marcer, M. Nocente, M. Rebai, D. Rigamonti, M. Angelone, A. Bracco, F. Camera, C. Cazzaniga, T. Craciunescu, G. Croci, M. Dalla Rosa, L. Giacomelli, G. Gorini, Y. Kazakov, E.M. Khilkevitch, A. Murao, E. Panontin, E. Perelli Cippo, M. Pillon, O. Putignano, J. Scionti, A. E. Shevelev, A. Žohar, and M. Tardocchi, Phys. Rev. Lett. 133, 055102 (2024)
  15. [15]M. Rebai, D. Rigamonti, A. Dal Molin, G. Marcer, A. Bracco, F. Camera, D. Farina, G. Gorini, E. Khilkevitch, M. Nocente, E. P. Cippo, O.r Putignano, J. Scionti, A. Shevelev, A. Zohar, and M. Tardocchi, Phys. Rev. C 110, 014625 (2024)
  16. [16]C. J. Horsfield, M. S. Rubery, J. M. Mack, H. W. Herrmann, Y. Kim, C. S. Young, S. E. Caldwell, S. C. Evans, T. S. Sedillo, A. M. McEvoy, N. M. Hoffman, M. A. Huff, J. R. Langenbrunner, G. M. Hale, D. C. Wilson, W. Stoeffl, J. A. Church, E. M. Grafil, E. K. Miller, and V. Yu Glebov Phys. Rev. C 104, 024610 (2021)
  17. [17]N.K. Timofeyuk, G.W. Bailey and M.R. Gilbert, Phys. Rev. C 110, 014612 (2024)
  18. [18]D.Baye, Phys. Rev. C 86, 034306 (2012)
  19. [19]D. Baye and P. Descouvemont, Nucl. Phys. A 443, 302 (1985)
  20. [20]J. Dohet-Eraly and D. Baye, Phys. Rev. C 88, 024602 (2013)
  21. [21]G.R. Satchler, L.W. Owen, A.J. Elwyn, G.L. Morgan and R.L. Walter, Nucl. Phys. A 112, 1 (1968)
  22. [22]H. Fiedeldey, Nucl. Phys. A 77, 149 (1966)
  23. [23]P. Navrátil and S. Quaglioni, Phys. Rev. Lett. 108, 042503 (2012).
  24. [24]S.B. Dubovichenko and N.A. Burkova and A.V. Dzhazairov-Kakhramanov and R.Ya. Kezerashvili and Ch.T. Omarov and A.S. Tkachenko and D.M. Zazulin, Nucl. Phys. A 987, 46 (2019)
  25. [25]B.Wu, H.Duan, J.Liu Nucl. Phys. A 1017, 122340 (2022)
  26. [26]Y. Yamaguchi, Phys Rev. 95, 1628 (1954)
  27. [27]H. De Vries, C. W. De Jager, and C. De Vries, At. Data and Nucl. Data Tables, 36 (1987) 495 (1987)
  28. [28]M. J. G. Borge, L. Johannsen, B. Jonson, T. Nilsson, G. Nyman, K. Riisager, O. Tengblad, and K. Wilhelmsen Rolander, Nucl. Phys. A 560, 664 (1993).
  29. [29]E.M. Tursunov, D. Baye and P. Descouvemont, Phys.Rev. C 73, 014303 (2006); Erratum Phys.Rev. C 74, 069904 (2006)
  30. [30]E.M. Tursunov, D. Baye and P. Descouvemont, Phys.Rev. C 97, 014302 (2018)

Paper details

vlnc:W_24p4uunu
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