Ab-initio calculations of charge symmetry breaking in the A=4 hypernuclei
Extracted source text ยท arXiv 1512.01049
%%%%%%%%%%%%%%%%%%%%%%%%%gg15prl.tex%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%\documentclass[aps,tightenlines,showpacs,nofootinbib,twocolumn,floatfix]
\documentclass[aps,prl,showpacs,twocolumn,groupedaddress,superscriptaddress]
{revtex4}
%\usepackage{epsfig}
\usepackage{graphicx}
\usepackage{dcolumn}
\usepackage{longtable}
\usepackage{amsmath}
%\tightenlines
\begin{document}
\title{{\it Ab initio} calculations of charge symmetry breaking in the $A=4$
hypernuclei}
\author{Daniel Gazda}\thanks{gazda@ujf.cas.cz}
\affiliation{Nuclear Physics Institute, 25068 \v{R}e\v{z}, Czech Republic}
\affiliation{ECT$^{\ast}$, Villa Tambosi, 38123 Villazzano (Trento), Italy}
\affiliation{Department of Fundamental Physics, Chalmers University of
Technology, SE-412 96 G\"{o}teborg, Sweden}\author{Avraham Gal}
\thanks{avragal@savion.huji.ac.il}\affiliation{Racah Institute of Physics,
The Hebrew University, Jerusalem 91904, Israel}
\date{\today}
\begin{abstract}
%\rule{0ex}{3ex}
We report on {\it ab initio} no-core shell model calculations of the mirror
$\Lambda$ hypernuclei $_{\Lambda}^4$H and $_{\Lambda}^4$He, using the
Bonn-J\"{u}lich leading-order chiral effective field theory hyperon-nucleon
potentials plus a charge symmetry breaking $\Lambda$-$\Sigma^0$ mixing vertex.
In addition to reproducing rather well the $0^{+}_{\rm g.s.}$ and $1^{+}_{
\rm exc.}$ binding energies, these four-body calculations demonstrate for the
first time that the observed charge symmetry breaking splitting of mirror
levels, reaching hundreds of keV for $0^{+}_{\rm g.s.}$, can be reproduced
using realistic theoretical interaction models, although with a non-negligible
momentum cutoff dependence. Our results are discussed in relation to recent
measurements of the $_{\Lambda}^4$H($0^{+}_{\rm g.s.}$) binding energy
[MAMI A1 Collaboration, Phys. Rev. Lett. {\bf 114}, 232501 (2015)] and the
$_{\Lambda}^4{\rm He}(1^{+}_{\rm exc.})$ excitation energy [J-PARC E13
Collaboration, Phys. Rev. Lett. {\bf 115}, 222501 (2015)].
\end{abstract}
\pacs{21.80.+a, 13.75.Ev, 11.30.-j, 21.60.De}
%\keywords{Hypernuclei, Hyperon-nucleon interactions, Symmetry and
%conservation laws, Ab initio methods.}
\maketitle
\noindent
{\bf Introduction.}~~Charge symmetry in hadronic physics is broken in QCD
by the up-down light quark mass difference and by the up and down quark
QED interactions. Recent lattice QCD+QED simulations of octet baryon mass
differences within isospin multiplets, such as the neutron-proton mass
difference $\Delta_{np}$ which vanishes in the limit of charge symmetry,
account nicely for the observed charge symmetry breaking (CSB) pattern in
the lowest-mass nonstrange as well as strange baryon spectrum \cite{LQCD15}.
A comparable level of precision in reproducing theoretically CSB
effects in the baryon-baryon interaction is lacking \cite{LQCD13}.
In practice, introducing two charge-dependent contact interaction terms
in chiral effective field theory (EFT) applications, one is able at
next-to-next-to-next-to-leading order (N3LO) to account quantitatively
for the charge dependence of the low-energy nucleon-nucleon ($NN$)
scattering parameters~\cite{entem03}. For strangeness $S=-1$, however,
given that low-energy $\Lambda p$ cross sections are poorly known and
$\Lambda n$ scattering data do not exist, the available chiral EFT
hyperon-nucleon ($YN$) interactions \cite{polinder06,haidenbauer13}
do not include charge-dependent interaction terms. Potentially unique
information on CSB in the $\Lambda N$ interaction and in $\Lambda$
hypernuclei is provided by the large $\Lambda$ separation-energy difference
$\Delta B^{J=0}_{\Lambda}$=350$\pm$60~keV~\cite{davis05} in the $A=4$
mirror hypernuclei $0^+$ ground states (g.s.) and the apparently negligible
difference $\Delta B^{J=1}_{\Lambda}$ in the $1^+$ excited states \cite{E13},
see Fig.~\ref{fig:exp}. Here, $\Delta B^{J}_{\Lambda}\equiv B^{J}_{\Lambda}
(_{\Lambda}^4{\rm He})-B^{J}_{\Lambda}(_{\Lambda}^4{\rm H})$. The recent
precise measurement of the $_{\Lambda}^4{\rm H}_{\rm g.s.}\to{^4{\rm He}}+
\pi^-$ decay at the Mainz Microtron (MAMI) \cite{MAMI15} reaffirms
a substantial CSB g.s. splitting $\Delta B^{J=0}_{\Lambda}$=270$\pm$95~keV,
which is consistent with the emulsion value cited above. Note that
$\Delta B^{J=0}_{\Lambda}$ is considerably larger than the $\approx$70~keV
assigned to CSB splitting in the mirror core nuclei $^3$H and
$^3$He \cite{miller06}.
\begin{figure}[htbp]
\begin{center}
\includegraphics[width=4.8cm]{gg15prlrev_fig1.eps}
\caption{(${_{\Lambda}^4{\rm H}},{_{\Lambda}^4{\rm He}}$) mirror hypernuclei
level diagram. The $0^+_{\rm g.s.}$ $\Lambda$ separation energies
$B_{\Lambda}$, loosely termed $\Lambda$ binding energies, are from emulsion
work \cite{davis05}, and the $1^{+}_{\rm exc.}$ $B_{\Lambda}$ values follow
from $\gamma$-ray measurements of the excitation energies $E_{\gamma}$
\cite{E13}.}
\label{fig:exp}
\end{center}
\end{figure}
Dalitz and von Hippel \cite{DvH64} suggested that the SU(3) octet
$\Lambda_{I=0}$ and $\Sigma^0_{I=1}$ hyperons are admixed in the physical
$\Lambda$ hyperon, thereby generating a CSB direct $\Lambda N$ potential
$V_{\rm CSB}$ that consists of isovector meson exchanges, notably a long-range
one-pion exchange (OPE) component. Although these exchanges are forbidden in
the $\Lambda N$ channel by the strong interactions (SI), they do contribute
strongly to the $\Lambda N\leftrightarrow\Sigma N$ coupling potential.
Quite generally, the matrix element of $V_{\rm CSB}$ arising from
$\Lambda-\Sigma^0$ mixing is related to the SI $I_{NY}=1/2$ matrix
element $\langle N\Sigma|V_{\rm SI}|N\Lambda\rangle$ by \cite{gal15}
\begin{equation}
\langle N\Lambda|V_{\rm CSB}|N\Lambda\rangle = -0.0297\,\tau_{Nz}\,
\frac{1}{\sqrt{3}}\,\langle N\Sigma|V_{\rm SI}|N\Lambda\rangle ,
\label{eq:OME}
\end{equation}
where
%the z component of the isospin Pauli matrix ${\vec\tau}_N$ assumes the values
$\tau_{Nz}=\pm 1$ for protons and neutrons, respectively,
%the isospin Clebsch-Gordan coefficient $1/\sqrt{3}$ accounts for the
%$N\Sigma^0$ amplitude in the $I_{NY}=1/2$ $N\Sigma$ state,
and the space-spin structure of this $N\Sigma$ state is taken identical
with that of the $N\Lambda$ state embracing $V_{\rm CSB}$. The CSB scale
coefficient 0.0297 in (\ref{eq:OME}) follows from the $\Lambda-\Sigma^0$
mass-mixing matrix element \cite{gal15a}
\begin{equation}
\langle\Sigma^0|\delta M|\Lambda\rangle=\frac{1}{\sqrt 3}
(\Delta_{\Sigma^0\Sigma^+}-\Delta_{np})=1.14(5)~{\rm MeV}
\label{eq:deltaM}
\end{equation}
and has been used in all previous CSB works listed below. A visualization
of Eq.~(\ref{eq:OME}) is provided by the CSB $\Lambda N$ interaction diagram
of Fig.~\ref{fig:diag_csbls_h}, where the $V_{\Lambda N - \Sigma N}$ blob
represents {\it any} SI isovector meson exchange or contact term such as
introduced in chiral EFT models \cite{polinder06}.
Precise four-body calculations using the Nijmegen soft-core realistic meson
exchange $YN$ interaction models NSC97$_{\rm e,f}$~\cite{NSC97}, which include
charge-dependent interactions induced by $\Lambda-\Sigma^0$ mixing and meson
mixings, produced at most 30\% of the observed CSB g.s. splitting $\Delta
B^{J=0}_{\Lambda}$ \cite{nogga01,nogga02,haidenbauer07,nogga13,nogga14}.
Below we comment on this insufficiency. More recent Nijmegen \cite{ESC10} or
quark-cluster \cite{kyoto07} models have not been used in four-body studies.
With SI $\Lambda N\leftrightarrow\Sigma N$ potential energy contributions
of order 10~MeV \cite{nogga01}, and with a CSB scale of order 3\%,
Eq.~(\ref{eq:OME}) could yield CSB contributions of order 300~keV.
Reproducing the observed CSB splitting poses a challenge for microscopic
$YN$ interaction models.
In this Letter we report on detailed {\it ab initio} no-core shell model
(NCSM) calculations of the $A=4$ $\Lambda$ hypernuclei that employ the SI
Bonn-J\"{u}lich LO chiral EFT $YN$ interaction potentials \cite{polinder06},
plus a CSB $\Lambda-\Sigma^0$ mixing interaction potential $V_{\rm CSB}$
generated by applying Eq.~(\ref{eq:OME}) to each one of the $\Lambda N
\leftrightarrow\Sigma N$ $V_{\rm SI}$ components in this LO version. CSB meson
mixings, with negligible contributions in the $A$=4 hypernuclei \cite{coon99},
are disregarded here. In addition to reproducing reasonably well the
$0^{+}_{\rm g.s.}$ and $1^{+}_{\rm exc.}$ binding energies, these four-body
calculations establish for the first time as large CSB splittings $\Delta
B^{J=0}_{\Lambda}$ as suggested by experiment, see Fig.~\ref{fig:exp},
although with a non-negligible cutoff dependence. We also discuss possible
implications to the recent Bonn-J\"{u}lich-Munich NLO chiral EFT $YN$
interaction model \cite{haidenbauer13}.
\begin{figure}[t!]
\begin{center}
\includegraphics[width=60mm]{gg15prlrev_fig2.eps}
\caption{CSB $\Lambda N$ interaction diagram describing a SI $V_{\Lambda N -
\Sigma N}$ interaction followed by a CSB $\Lambda-\Sigma^0$ mass-mixing
vertex.}
\label{fig:diag_csbls_h}
\end{center}
\end{figure}
%\newline
\noindent
{\bf Methodology.}~~The nuclear NCSM technique used in the present four-body
calculations employs realistic two-body and three-body model interactions
and is formulated in a translationally invariant Jacobi-coordinate
harmonic-oscillator (HO) basis \cite{navratil00}. Antisymmetrization with
respect to nucleons is exercised in order to satisfy the Pauli principle. The
resulting Hamiltonian is diagonalized in finite four-body HO bases, admitting
all HO excitation energies $N\hbar\omega$, $N\leq N_{\rm max}$, up to
$N_{\rm max}$ HO quanta. Extrapolated energy values $E(\omega)$, $N_{\rm max}
\to\infty$, are obtained by fitting an exponential function to $E(N_{\rm max},
\omega \text{~fixed})$ sequences in the vicinity of the variational minima
with respect to the HO basis frequency $\omega$. The reliability of such
extrapolations is then reflected in the independence of $E(\omega)$ of the
frequency $\omega$.
\begin{table}[t!]
\caption{Cutoff dependence of $\Lambda$ separation energies $B^{J}_{\Lambda}$
in $_{\Lambda}^4{\rm H}$ and $_{\Lambda}^4{\rm He}$ (all in MeV) from {\it ab
initio} NCSM calculations at $\hbar\omega$=30(32)~MeV for $J$=0(1), using N3LO
(LO) chiral $NN$ ($YN$) interactions \cite{entem03} (\cite{polinder06}) plus
Coulomb interactions, and $V_{\rm CSB}$ generated by Eq.~(\ref{eq:OME}) from
the LO SI $YN$ potentials. Experimental values are from Fig.~\ref{fig:exp}.}
\begin{ruledtabular}
\begin{tabular}{lccccc}
Cutoff & 550 & 600 & 650 & 700 & Experiment \\
\hline
$B^{J=0}_{\Lambda}(_{\Lambda}^4{\rm H})$ & 2.556 & 2.308 & 2.154 & 2.196 &
2.04$\pm$0.04 \\
$B^{J=0}_{\Lambda}(_{\Lambda}^4{\rm He})$ & 2.586 & 2.444 & 2.398 & 2.490 &
2.39$\pm$0.03 \\
$B^{J=1}_{\Lambda}(_{\Lambda}^4{\rm H})$ & 1.744 & 1.359 & 1.067 & 0.877 &
0.95$\pm$0.04 \\
$B^{J=1}_{\Lambda}(_{\Lambda}^4{\rm He})$ & 1.572 & 1.166 & 0.839 & 0.654 &
0.98$\pm$0.03 \\
\end{tabular}
\end{ruledtabular}
\label{tab:B_L(A=4)}
\end{table}
This NCSM technique, extended recently to light hypernuclei
\cite{gazda14,wirth14}, is applied here to the $A$=4 mirror hypernuclei
using chiral N3LO $NN$ and N2LO $NNN$ interactions \cite{entem03,navratil07},
respectively, both with a momentum cutoff of 500 MeV. These together with
the Coulomb interaction reproduce the binding energies of the $A$=3 core
nuclei. For the SI $YN$ coupled-channel potentials $V_{\rm SI}$, we use the
Bonn-J\"{u}lich LO chiral EFT SU(3)-based model with cutoff momenta $\Lambda$
from 550 to 700 MeV \cite{polinder06} plus $V_{\rm CSB}$ evaluated from
$V_{\rm SI}$ by using Eq.~(\ref{eq:OME}). Baryon mass differences within
isomultiplets are incorporated. The reported calculations consist of
fully converged $^3$H and $^3$He binding energies, and ($_{\Lambda}^4$H,
$_{\Lambda}^4$He) $0^+_{\rm g.s.}$ and $1^{+}_{\rm exc.}$ binding energies
extrapolated to infinite model spaces from $N_{\rm max}=18(14)$ for $J=0(1)$.
The $NNN$ interaction is excluded from the calculations reported here,
in order to save computing time, after verifying that its inclusion makes
a difference of only a few keV in the calculation of the CSB splittings
$\Delta B^{J}_{\Lambda}$ for both $J=0,1$.
\newline
\noindent
{\bf Results.}~~The cutoff dependence of $\Lambda$ separation energies in both
$A$=4 mirror hypernuclei, obtained from NCSM calculations with LO chiral EFT
coupled-channel $YN$ potentials \cite{polinder06} and $V_{\rm CSB}$ from
Eq.~(\ref{eq:OME}), is shown in Table~\ref{tab:B_L(A=4)}. We used
$N_{\rm max}\to\infty$ extrapolated binding-energy values for the
$_{\Lambda}^4{\rm He}$ and $_{\Lambda}^4{\rm H}$ $J$=0(1) levels at fixed
$\hbar\omega$=30(32)~MeV, which is where the absolute variational minima
occur for $\Lambda$=550 and 600~MeV. For higher values of $\Lambda$ the
four-body absolute variational minima occur at slightly higher $\hbar\omega$
values. Although the spread of $B^{J}_{\Lambda}(\hbar\omega)$ values for
a given cutoff momentum is of the order of 100~keV, it is considerably smaller
and in fact marginal for the CSB splittings $\Delta B^{J}_{\Lambda}$ on which
we focus here, as demonstrated by Fig.~4 below.
%The spread generated thereby away from the listed $B_{\Lambda}$ values
%is maximal at $\Lambda$=700~MeV, amounting to 60 (170)~keV for $J$=0(1).
The $\Lambda$ separation energies listed in Table~\ref{tab:B_L(A=4)}
show a moderate cutoff dependence for the $0^+_{\rm g.s.}$ mirror levels
and a stronger dependence for the $1^{+}_{\rm exc.}$ mirror levels,
with mean values for their charge-symmetric (CS) averages given by
${\overline B}^{\rm CS}_{\Lambda}(0^+_{\rm g.s.})$=2.39$^{+0.18}_{-0.12}$~MeV
and ${\overline B}^{\rm CS}_{\Lambda}(1^{+}_{\rm exc.})$=1.16$^{+0.50}_{-0.39}
$~MeV which compare well with the CS-averaged experimental values derived from
the last column in Table~\ref{tab:B_L(A=4)}. Furthermore, considering NCSM
$N_{\rm max}\to\infty$ extrapolation uncertainties, our CS-averaged
$B_{\Lambda}$ values are in fair agreement with those reported
in other four-body calculations using CS LO $YN$ chiral EFT
interactions \cite{haidenbauer07,nogga13,nogga14,gazda14,wirth14}.
A detailed analysis of calculational uncertainties will be given elsewhere.
\begin{figure}[t!]
\begin{center}
\includegraphics[width=80mm]{gg15prlrev_fig3.eps}
\caption{(color online). Cutoff momentum dependence of excitation energies
$E_{\rm x}$(0$^+_{\rm g.s.}$$\to$1$^+_{\rm exc.}$) in $_{\Lambda}^4{\rm H}$
(squares) and $_{\Lambda}^4{\rm He}$ (circles) in {\it ab initio} NCSM
calculations, at $\hbar\omega$=30(32)~MeV for $J$=0(1), for LO chiral EFT
coupled-channel $YN$ potentials \cite{polinder06} with (solid lines) and
without (dashed lines) $V_{\rm CSB}$ derived from these SI potentials using
Eq.~(\ref{eq:OME}). The dotted horizontal lines denote $E_{\rm x}$ values
from $\gamma$-ray measurements \cite{E13}.}
\label{fig:ex-l_d}
\end{center}
\end{figure}
Shown in Fig.~\ref{fig:ex-l_d} by solid lines is the cutoff momentum
dependence of the $0^+_{\rm g.s.}\to 1^{+}_{\rm exc.}$ excitation
energies $E_{\rm x}$ formed from the $B_{\Lambda}$ values listed
in Table~\ref{tab:B_L(A=4)} for both $A$=4 mirror hypernuclei.
As observed in several few-body calculations of $s$-shell
hypernuclei \cite{GL88,akaishi00,hiyama02,akaishi02}, $E_{\rm x}$ is
strongly correlated with the $\Lambda N\leftrightarrow\Sigma N$ coupling
potential which in the present context, through $\Lambda-\Sigma^0$ mixing,
gives rise to CSB splittings of the $A$=4 mirror levels.
Figure \ref{fig:ex-l_d} demonstrates a steady rise of both
$E_{\rm x}(_{\Lambda}^4{\rm He})$ and $E_{\rm x}(_{\Lambda}^4{\rm H})$
as a function of the cutoff momentum $\Lambda$, with a CS-averaged value
${\overline E}_{\rm x}^{\rm CS}$=1.23$^{+0.35}_{-0.32}$~MeV compared to
1.25$\pm$0.02~MeV deduced from the two $\gamma$-ray energies shown in
Fig.~\ref{fig:exp}. A steady rise is also observed in the difference
$\Delta E_{\rm x}^{\rm CSB}$ with a mean value 380$^{+140}_{-180}$~keV
compared to 320$\pm$20~keV, again from Fig.~\ref{fig:exp}. In agreement with
previous calculations \cite{nogga01,nogga02,haidenbauer07,nogga13,nogga14},
residual CSB contributions of up to 30~keV from electromagnetic mass
differences, mostly of $\Sigma$ hyperons, and from the increased Coulomb
repulsion in the $^3$He core of $_{\Lambda}^4{\rm He}$, survive upon
switching off $V_{\rm CSB}$, as demonstrated by the slight difference
between the two middle dashed lines in the figure.
%\vspace{0.5cm}
\begin{figure}[t!]
\begin{center}
\includegraphics[width=80mm]{gg15prlrev_fig4.eps}
\caption{(color online). Dependence of the separation-energy differences
$\Delta B_{\Lambda}$ between $_{\Lambda}^4{\rm He}$ and $_{\Lambda}^4{\rm H}$,
for $0^+_{\rm g.s.}$ (upper curve) and for $1^+_{\rm exc.}$ (lower curve)
on the HO $\hbar\omega$ in {\it ab initio} NCSM calculations using LO chiral
EFT coupled-channel $YN$ potentials with cutoff momentum $\Lambda$=600~MeV
\cite{polinder06} plus $V_{\rm CSB}$ derived from these SI potentials using
Eq.~(\ref{eq:OME}). Results for other values of $\Lambda$ are shown at
$\hbar\omega$=30(32)~MeV for $J$=0(1).}
\label{fig:csb-om}
\end{center}
\end{figure}
In Fig.~\ref{fig:csb-om} we show the $\hbar\omega$ dependence of
separation-energy differences $\Delta B^{J}_{\Lambda}$ between $_{\Lambda}^4{
\rm He}$ and $_{\Lambda}^4{\rm H}$ levels of a given spin $J$, for $0^+_{
\rm g.s.}$ and $1^{+}_{\rm exc.}$, using $N_{\rm max}\to\infty$ extrapolated
values for the four possible binding energies which are calculated for a
cutoff $\Lambda$=600~MeV and including $V_{\rm CSB}$ from Eq.~(\ref{eq:OME}).
Extrapolation uncertainties for $\Delta B^{J}_{\Lambda}$ are about 20~keV.
The variation of $\Delta B^{J=0}_{\Lambda}$ in the spanned $\hbar\omega$
range amounts to a few keV, whereas that of $\Delta B^{J=1}_{\Lambda}$ is
larger, amounting to $\sim$30~keV. It is worth noting that the difference
$\Delta B^{J=0}_{\Lambda}-\Delta B^{J=1}_{\Lambda}$ between the upper and
lower curves assumes at $\Lambda$=600~MeV the value 0.33$\pm$0.04~MeV,
in perfect agreement with the difference $E_{\gamma}(_{\Lambda}^{4}{\rm He})
-E_{\gamma}(_{\Lambda}^{4}{\rm H})=0.32\pm 0.02$~MeV between the two $\gamma$
ray energies shown in Fig.~\ref{fig:exp}. The figure also demonstrates
a strong cutoff dependence of $\Delta B^{J=0}_{\Lambda}$, varying between
30 and 300~keV upon increasing $\Lambda$, together with a considerably weaker
cutoff dependence of $\Delta B^{J=1}_{\Lambda}$, varying between $-$170
and $-$230~keV. Note that $\Delta B^{J=0}_{\Lambda}$ comes out invariably
positive, whereas $\Delta B^{J=1}_{\Lambda}$ is robustly negative. With mean
values ${\overline{\Delta B}}^{J=0}_{\Lambda}$=176$^{+118}_{-146}$~keV and
${\overline{\Delta B}}^{J=1}_{\Lambda}=-204^{+32}_{-24}$~keV, the mean
values ${\overline{\Delta B}}^J_{\Lambda}$ satisfy
\begin{equation}
{\overline{\Delta B}}^{J=1}_{\Lambda} \approx
-\,{\overline{\Delta B}}^{J=0}_{\Lambda} < 0.
\label{eq:minus}
\end{equation}
\noindent
{\bf Discussion.}~~To understand the CSB pattern Eq.~(\ref{eq:minus}) for the
$A$=4 hypernuclei, we note that the SI $\Lambda N\leftrightarrow\Sigma N$
coupling potential in the LO chiral EFT $YN$ model of Ref.~\cite{polinder06}
consists of a pseudoscalar (PS) meson exchange, dominated by OPE, plus
two $s$-wave interaction contact terms (CT) of which the $^3S_1$ CT is
negligible and the $^1S_0$ CT is large. In a zeroth-order single-particle
description of the $A$=4 hypernuclei, and using Eq.~(\ref{eq:OME}), these
$\Lambda N\leftrightarrow\Sigma N$ coupling-potential components contribute
to the CSB separation-energy differences as follows:
\begin{equation}
\Delta B^{J=0}_{\Lambda} = \frac{3}{2}C_1-\frac{1}{2}C_0,\;\;\;\;
\Delta B^{J=1}_{\Lambda} = \frac{1}{2}C_1+\frac{1}{2}C_0,
\label{eq:CTa}
\end{equation}
with $C_S=C_S^{\rm CT}+C_S^{\pi}$ the sum of contributions to the
triplet ($S$=1) and singlet ($S$=0) matrix elements from CT and from OPE.
The ${\vec\sigma}_Y\cdot{\vec\sigma}_N$ spin dependence of $C_S^{\pi}$
leads in this approximation to $C_1^{\pi}=-\frac{1}{3}C_0^{\pi}$.
Recalling that these matrix elements already incorporate isospin, both
$C_0^{\pi}$ and $C_0^{\rm CT}$ are negative. Hence,
\begin{equation}
\Delta B^{J=0}_{\Lambda} \approx -\,(C_0^{\pi}+\frac{1}{2}C_0^{\rm CT}) > 0,
\label{eq:CTb1}
\end{equation}
\begin{equation}
\Delta B^{J=1}_{\Lambda} \approx +\,(\frac{1}{3}C_0^{\pi} +
\frac{1}{2}C_0^{\rm CT}) < 0,
\label{eq:CTb2}
\end{equation}
in agreement with the signs of the calculated CSB splittings. In the
limit that $C_0^{\pi}$ is negligible with respect to $C_0^{\rm CT}$,
Eq.~(\ref{eq:minus}) is recovered. We conclude that it is the sizable $^1S_0$
$\Lambda N\leftrightarrow\Sigma N$ coupling-potential CT in the LO chiral EFT
$YN$ interaction model \cite{polinder06} that makes it possible to generate
sufficiently large values of $\Delta B^{J=0}_{\Lambda}$ to explain the
observed CSB splitting of the $0^+_{\rm g.s.}$ mirror levels. However,
the opposite-sign values of $\Delta B^{J=1}_{\Lambda}$ appear too large
with respect to the near degeneracy observed for the $1^{+}_{\rm exc.}$
mirror levels, even when updated values from the latest MAMI measurement
are considered \cite{MAMI16}.
In contrast to the ability of the LO chiral EFT $YN$ interaction model to
generate sizable CSB g.s. splittings $\Delta B^{J=0}_{\Lambda}$ owing to
a dominant $^1S_0$ $\Lambda N\leftrightarrow\Sigma N$ coupling-potential CT,
the $\Lambda N\leftrightarrow\Sigma N$ coupling potential in NSC97 models
is dominated by a $^3S_1 - {^3D_1}$ tensor component which is ineffective
in generating a large CSB contribution when used in the right-hand side
of Eq.~(\ref{eq:OME}). The reason is that the SI $\Lambda N$ states on
the left-hand side, in the case of NSC97, are dominated by purely $s$-wave
channels \cite{nogga01}. The NSC97 $^1S_0$ $\Lambda N\leftrightarrow\Sigma N$
coupling-potential contribution that replaces $C_0^{\pi}$ in
Eq.~(\ref{eq:CTb1}) is too weak to generate on its own a sizable $\Delta
B^{J=0}_{\Lambda}$. A detailed account of this item will be given elsewhere.
It is tempting to speculate on the $A$=4 CSB separation-energy differences
$\Delta B^J_{\Lambda}$ anticipated from applying Eq.~(\ref{eq:OME}) to the
recently published NLO chiral EFT $YN$ interaction \cite{haidenbauer13}.
The $\Lambda N\leftrightarrow\Sigma N$ coupling-potential contact terms differ
considerably in NLO from those in LO, with a very large $C_1^{\rm CT}$ that
dominates in NLO over $C_0^{\rm CT}$, and with a new $^3S_1$--$^3D_1$ CT.
It is fair to assume that PS one- and two-meson exchange contributions in
NLO are still dominated by OPE. Dominance of $C_1^{\rm CT}$ over all other
allowed contributions would result in {\it negative} values of $\Delta
B^J_{\Lambda}$, with $\Delta B^{J=0}_{\Lambda}$ 3 times as large as
$\Delta B^{J=1}_{\Lambda}$; this would disagree with the observed positive
value for $\Delta B^{J=0}_{\Lambda}$, see Fig.~\ref{eq:OME}, confirmed
also by the new MAMI measurement \cite{MAMI15}. We note, furthermore,
that the NLO version underestimates the $A$=4 hypernuclear g.s. separation
energy, with $B_{\Lambda}^{\rm CS}\approx$1.5--1.6~MeV \cite{nogga13},
compared to $\approx$2.2~MeV from Fig.~\ref{fig:exp}. Three-body $YNN$
interaction terms introduced in higher-order versions in order to recover the
missing g.s. attraction might provide additional source of CSB in $\Lambda$
hypernuclei. However, expecting that the dominant $YNN$ terms correspond
to $\Sigma^{\ast}(1385)NN$ intermediate states \cite{petschauer15} and
realizing that, unlike $\Sigma^0$, $\Sigma^{\ast 0}({\frac{3}{2}}^+)$
cannot mix with $\Lambda^0({\frac{1}{2}}^+)$ to generate CSB, these $YNN$
interaction terms will not produce as strong CSB as evaluated here using
Eq.~(\ref{eq:OME}), which is based on the Dalitz--von Hippel $\Lambda^0 -
\Sigma^0$ mixing mechanism \cite{DvH64}. It is therefore questionable whether
the NLO version \cite{haidenbauer13} offers an advantage over the LO version
\cite{polinder06} for $\Lambda$ hypernuclei, given also that both provide
comparably reasonable fits to the low-energy $YN$ scattering data.
\newline
\noindent
{\bf Summary and outlook.}~~In conclusion, we have presented the first CSB
{\it ab initio} calculation in hypernuclei with chiral EFT coupled-channel
$YN$ interactions, showing that the LO version \cite{polinder06} is capable
of producing a {\it large} CSB $0^+_{\rm g.s.}$ splitting ${\overline{
\Delta B}}^{J=0}_{\Lambda}\sim 180\pm 130$~keV. This is consistent with a g.s.
splitting of 270$\pm$95~keV reported by the MAMI experiment \cite{MAMI15}.
Our NCSM calculation reproduces quantitatively and with weak cutoff dependence
the $0^+_{\rm g.s.}$ binding energies of the $A$=4 mirror hypernuclei,
whereas the $1^{+}_{\rm exc.}$ binding-energy calculation, which is known
to be numerically more challenging \cite{nogga01}, displays a strong
cutoff dependence. The calculated CSB $1^{+}_{\rm exc.}$ splitting is of
opposite sign to that of the $0^+_{\rm g.s.}$ splitting and fairly large:
${\overline{\Delta B}}^{J=1}_{\Lambda}\approx -200\pm 30$~keV, with a weak
cutoff dependence. While the latest results from MAMI suggest a smaller
negative CSB splitting of $-83\pm 94$~keV for the $1^{+}_{\rm exc.}$
mirror levels \cite{MAMI16}, the measurement systematic uncertainty is still
too large to rule out the prediction of the LO version.
In future work it would be of great interest to apply the CSB generating
equation (\ref{eq:OME}) in {\it ab initio} calculations of the $A$=4 mirror
hypernuclei using the recent NLO EFT version \cite{haidenbauer13}, and also
to readjust the $\Lambda N\leftrightarrow\Sigma N$ contact terms in NLO by
imposing the accurate CSB datum $E_{\gamma}(_{\Lambda}^4{\rm He})-E_{\gamma}
(_{\Lambda}^4{\rm H})=0.32\pm 0.02$~MeV, so it is reproduced in four-body
calculations with as weak cutoff dependence as possible. Another natural
follow-up would be to extend these CSB calculations in LO and NLO to $p$-shell
hypernuclei. Recent shell model calculations \cite{gal15}, using a schematic
$\Lambda N\leftrightarrow\Sigma N$ coupling-potential model, suggest that CSB
splittings of g.s. mirror levels in $p$-shell hypernuclei decrease in size
with respect to $A=4$, and perhaps even reverse sign, in rough agreement with
old emulsion data \cite{davis05}. Such extensions of the present work pose
a valuable theoretical challenge to the microscopic understanding of strange
nuclear systems.
\newline
\noindent
{\bf Acknowledgments}.~~We are grateful to Petr Navr\'{a}til for providing
valuable advice and help on extensions of nuclear-physics NCSM codes, to
Johann Haidenbauer and Andreas Nogga for providing us with the input LO EFT
$YN$ potentials used in the present work, to Nir Barnea for stimulating
discussions on EFT few-body applications, and to Ji\v{r}\'{i} Mare\v{s} for
a critical reading of this manuscript. The research of D.G. was supported by
the Grant Agency of the Czech Republic (GACR) Grant No. P203/15/04301S.
\begin{thebibliography}{99}
\bibitem{LQCD15} S.~Borsanyi {\it et al.}, Science {\bf 347}, 1452 (2015).
\bibitem{LQCD13} S.R.~Beane {\it et al.} (NPLQCD Collaboration), Phys. Rev.
C {\bf 88}, 024003 (2013).
\bibitem{entem03} D.R.~Entem and R.~Machleidt, Phys. Rev. C {\bf 68},
041001(R) (2003).
\bibitem{polinder06} H.~Polinder, J.~Haidenbauer, and U.-G.~Mei{\ss}ner,
Nucl. Phys. A {\bf 779}, 244 (2006).
\bibitem{haidenbauer13} J.~Haidenbauer, S.~Petschauer, N.~Kaiser,
U.-G.~Mei{\ss}ner, A.~Nogga, and W.~Weise, Nucl. Phys. A {\bf 915}, 24 (2013).
\bibitem{davis05} D.H.~Davis, Nucl. Phys. A {\bf 754}, 3c (2005).
\bibitem{E13} T.O.~Yamamoto {\it et al.} (J-PARC E13 Collaboration),
Phys. Rev. Lett. {\bf 115}, 222501 (2015).
\bibitem{MAMI15} A.~Esser {\it et al.} (A1 Collaboration), Phys. Rev. Lett.
{\bf 114}, 232501 (2015).
\bibitem{miller06} G.A.~Miller, A.K.~Opper, and E.J.~Stephenson, Annu. Rev.
Nucl. Part. Sci. {\bf 56}, 253 (2006).
\bibitem{DvH64} R.H.~Dalitz and F.~von Hippel, Phys. Lett. {\bf 10}, 153
(1964).
\bibitem{gal15} A.~Gal, Phys. Lett. B {\bf 744}, 352 (2015).
\bibitem{gal15a} see Comment by A.~Gal, Phys. Rev. D {\bf 92}, 018501 (2015)
and Reply by R.~Horsley {\it et al.} (QCDSF-UKQCD Collaboration), Phys. Rev.
D {\bf 92}, 018502 (2015), where the incompleteness of existing LQCD
calculations of $\langle\Sigma^0|\delta M|\Lambda\rangle$ is discussed.
\bibitem{NSC97} Th.A.~Rijken, V.G.J.~Stoks, and Y.~Yamamoto, Phys. Rev. C
{\bf 59}, 21 (1999).
\bibitem{nogga01} A.~Nogga, Ph.D. Dissertation submitted to the Ruhr-
\newline
\noindent
Universit\"{a}t, Bochum (2001), http://www-brs.ub.ruhr-
\newline
\noindent
uni-bochum.de/netahtml/HSS/Diss/NoggaAndreas/.
\bibitem{nogga02} A.~Nogga, H.~Kamada, and W.~Gl\"{o}ckle, Phys. Rev. Lett.
{\bf 88}, 172501 (2002).
\bibitem{haidenbauer07} J.~Haidenbauer, U.-G.~Mei{\ss}ner, A.~Nogga, and
H.~Polinder, in {\it Topics in Strangeness Nuclear Physics}, Lecture Notes
in Physics {\bf 724}, edited by P.~Byd\v{z}ovsk\'{y}, J.~Mare\v{s}, and
A. Gal (Springer, New York, 2007), pp.~113-140.
\bibitem{nogga13} A.~Nogga, Nucl. Phys. A {\bf 914}, 140 (2013) and
references to earlier works cited therein.
\bibitem{nogga14} A.~Nogga, Few-Body Syst. {\bf 55}, 757 (2014).
\bibitem{ESC10} Th.A.~Rijken, M.M.~Nagels, and Y.~Yamamoto,
Prog. Theor. Phys. Suppl. {\bf 185}, 14 (2010).
\bibitem{kyoto07} Y.~Fujiwara, Y.~Suzuki, and C.~Nakamoto,
Prog. Part. Nucl. Phys. {\bf 58}, 439 (2007).
\bibitem{coon99} S.A.~Coon, H.K.~Han, J.~Carlson, and B.F.~Gibson,
in {\it Meson and Light Nuclei '98}, edited by J.~Adam, P.~Byd\v{z}ovsk\'{y},
J.~Dobe\v{s}, R.~Mach, and J.~Mare\v{s} (WS, Singapore, 1999), pp.~407-413,
arXiv:nucl-th/9903034.
\bibitem{navratil00} P.~Navr\'{a}til, G.P.~Kamuntavi\v{c}ius,
and B.R.~Barrett, Phys. Rev. C {\bf 61}, 044001 (2000).
\bibitem{gazda14} D.~Gazda, J.~Mare\v{s}, P.~Navr\'{a}til, R.~Roth, and
R.~Wirth, Few-Body Syst. {\bf 55}, 857 (2014).
\bibitem{wirth14} R.~Wirth, D.~Gazda, P.~Navr\'{a}til, A.~Calci,
J.~Langhammer, and R.~Roth, Phys. Rev. Lett. {\bf 113}, 192502 (2014).
Note that with the same Jacobi-coordinate NCSM methodology and the same
$NN$ and $YN$ interactions used in the present paper, these authors found
$_{\Lambda}^3$H to be particle-stable, with $\Lambda$ separation energy
$B_{\Lambda}$=110$\pm$10~keV for cutoff 600 MeV, consistent with experiment
\cite{davis05} and with Faddeev calculations reported by Haidenbauer
{\it et al.} \cite{haidenbauer07}. CSB from $\Lambda - \Sigma^0$ mixing has
no effect in $_{\Lambda}^3$H calculations.
\bibitem{navratil07} P.~Navr\'{a}til, Few-Body Syst. {\bf 41}, 117 (2007).
\bibitem{GL88} B.F.~Gibson and D.R.~Lehman, Phys. Rev. C {\bf 37}, 679 (1988).
\bibitem{akaishi00} Y.~Akaishi, T.~Harada, S.~Shinmura, and K.S.~Myint,
Phys. Rev. Lett. {\bf 84}, 3539 (2000).
\bibitem{hiyama02} E.~Hiyama, M.~Kamimura, T.~Motoba, T.~Yamada,
and Y.~Yamamoto, Phys. Rev. C {\bf 65}, 011301(R) (2002).
\bibitem{akaishi02} H.~Nemura, Y.~Akaishi, and Y.~Suzuki, Phys. Rev. Lett.
{\bf 89}, 142504 (2002).
\bibitem{MAMI16} F.~Schulz {\it et al.} (A1 Collaboration), Nucl. Phys. A
{\bf 954}, 149 (2016). Adopting their value of $B_{\Lambda}(_{\Lambda}^4{
\rm H})=2.157\pm 0.006({\rm stat.})\pm 0.077({\rm syst.})$~MeV, the observed
CSB splittings would change to $233\pm 92$~keV for $0^+_{\rm g.s.}$ and to
$-83\pm 94$~keV for $1^{+}_{\rm exc.}$. We thank Dr. Patrick Achenbach for
briefing us of these values prior to publication.
\bibitem{petschauer15} S.~Petschauer, N.~Kaiser, J.~Haidenbauer,
U.-G.~Mei{\ss}ner, and W.~Weise, Phys. Rev. C {\bf 93}, 014001 (2016),
and references cited therein to earlier work.
\end{thebibliography}
\end{document}