Detection of Atmospheric Neutrinos and Relativistic Nuclear Structure Effects
Extracted source text ยท arXiv nucl-th/9602009
\documentstyle[preprint,aps]{revtex}
\tighten
%\documentstyle[aps]{revtex}
%\documentstyle[prl,aps,preprint,tighten,floats,epsf,rotate]{revtex}
\preprint{\vbox{Submitted to Physical Review {\bf C}
\hfill IU/NTC 96-02}}
\begin{document}
\title{Detection of Atmospheric Neutrinos and \\ Relativistic
Nuclear Structure Effects}
\author{Hungchong Kim~\footnote{E-mail:~hung@phya.yonsei.ac.kr}
$^{1,2}$, S. Schramm~\footnote{E-mail:~schramm@tpri6e.gsi.de}
$^3$, C. J. Horowitz
~\footnote{E-mail:~charlie@proteus.iucf.indiana.edu} $^1$}
\address{$^1$ Nuclear Theory Center, Indiana University,
Bloomington, Indiana 47408, USA \\
$^2$ Department of Physics,
Yonsei University, Seoul, 120-749, Korea \\
$^3$ GSI, D-64220 Darmstadt, Germany}
\date{\today}
\maketitle
\begin{abstract}
Neutrino-nucleus cross sections for the detection of atmospheric
neutrinos are calculated in relativistic impulse and random phase
approximations. Pion production is estimated via inclusive $\Delta$-
hole nuclear excitations. The number of pion events can confuse the
identification of the neutrino flavor.
As a further source of pions we
calculate coherent pion production (where the nucleus remains in the
ground state). We examine how these nuclear structure effects
influence atmospheric neutrino experiments and study possible
improvements in the present detector simulations.
\end{abstract}
\eject
\narrowtext
\section{Introduction}
\label{sec:cint}
Aside from the puzzling solar neutrino data there exist extensive
measurements of atmospheric neutrinos, which
are produced in cosmic ray interactions with the earth's atmosphere.
Cosmic rays, mostly protons or $\alpha$ particles, hitting
the atmosphere produce pions, which undergo the following
decays:
%\EQ
\begin{eqnarray}
\pi^{\pm}&\rightarrow &\mu^{\pm}+\nu_{\mu}\ (\bar{\nu}_{\mu})
\nonumber\\
& &\mu^{\pm}\rightarrow e^{\pm}+\nu_{e}\ (\bar{\nu}_{e})+
\bar{\nu}_{\mu}\ (\nu_{\mu})\nonumber\ .
\end{eqnarray}
%\EN
From this, one expects the neutrino flavor ratio,
\begin{eqnarray}
r= {\nu_\mu +{\bar \nu_\mu} \over \nu_e +{\bar \nu_e}}\ ,
\label{rat}
\end{eqnarray}
to be about 2. However, two existing experiments,
IMB~\cite{imb} and Kamiokande~\cite{kamiokande},
consistently measured a
ratio of about 1 by studying
charged-current neutrino-nucleus reactions [($\nu_e$, $e$) and
($\nu_\mu$, $\mu$)] in large underground
water detectors. This could mean either there is a depletion of
muon-type neutrinos or an enhancement of electron-type neutrinos.
While the separate $\nu_\mu$ and
$\nu_e$ neutrino fluxes differ in various calculations,
their ratio, however,
seems to be largely model-independent~\cite{eng}, and
the theoretical results are in clear contradiction with data.
This discrepancy is known as the ``atmospheric neutrino anomaly''.
One explanation of this result could be the existence
of neutrino oscillations of the muon neutrino into
some other neutrino
species~\cite{solar} which would reduce the $\nu_\mu$ flux.
However, before
drawing this conclusion, one should also investigate more
conventional sources
for uncertainties in the measured flux ratio, which originate from
nuclear physics and detector specifics.
One crucial point is the experimental separation of
electron and muon events.
Most data are measured using water detectors. In a charged-current
weak interaction with an $^{16}$O nucleus, the incoming neutrino
generates an outgoing lepton which is detected through its
\v{C}erenkov radiation. Electrons and muons are distinguished by the
characteristics of their tracks. Electrons produce a showering
``fuzzy'' track, whereas muons are identified by their nonshowering
track.
Pions constitute a major background in this way of determining
the neutrino flavors. In the experimental analysis, two- and
more ring events, which signal the production of several charged
particles, are rejected. However, a charged pion
can be confused with a muon event if the lepton energy is below
\v{C}erenkov radiation threshold, and a neutral pion, with
one decay photon missing,
can be counted as an electron event. Therefore,
a reasonable determination of the flavor ratio can only be
achieved when the pion events are properly taken into account.
One major source of pions is the decay of a $\Delta$ produced
in the neutrino scattering. The delta can
subsequently decay into a pion and
a nucleon. An estimate of the number of pion events is only
possible when the $\Delta$-h response is properly calculated.
Indeed, Monte Carlo simulation of the Kamiokande
experiment~\cite{monte} estimated the number of pion events based
on a rather complicated
nonrelativistic model~\cite{fogli}.
However, atmospheric neutrinos have a broad energy spectrum
ranging
from a few hundred MeV to several GeV.
It is therefore crucial to have a relativistic formalism that allows
the calculation of the cross section for arbitrary kinematics.
As another source of pions, we consider coherent pion
production and its role in the detection of atmospheric neutrinos.
Coherent pions are produced in both charged- and neutral-current
neutrino-nucleus interactions which leave the nucleus in its
ground state.
Traditionally, the coherent $\pi^0$'s are distinguished from the
resonantly produced pions due to their strongly
forward peaked angular
distribution. However, in atmospheric neutrino experiments, due
to a lack of
directional information of the incoming neutrinos, it is not
possible to
distinguish the coherent $\pi^0$ events from incoherent pions.
In the case of coherent charged pions the forward angle dominance
of the cross section leads to a small opening angle between two
charged particles ($\mu$ or $e$ and charged pion) in the final state,
and these two particles could be detected as one isolated
electromagnetic shower. In contrast to
incoherent charged pions whose \v{C}erenkov tracks are
usually identified as muon events, the special characteristics
of coherent pions
could lead to an increase of identified $\nu_e$ events
and thus to underestimating the $\nu_\mu$ to $\nu_e$ ratio.
This paper is organized as follows. In Sec.~\ref{sec:incl}
we present
the formalism for neutrino-nucleus cross section including p-h and
$\Delta$-h
nuclear excitations within the framework of
a relativistic mean-field model of the nucleus.
Sec.~\ref{sec:coh_for} contains
the formalism for coherent charged and neutral pion
production.
The resulting cross sections and production rates are shown in
Sec.~\ref{sec:result}. We close with a discussion and outlook in
Sec.~\ref{sec:con}.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{ Inclusive Cross Section}
\label{sec:incl}
We outline the formalism for inclusive neutrino-nucleus scattering
cross section
including p-h (nucleon particle-hole) and $\Delta$-h (delta-hole)
nuclear excitations. We consider a
neutrino with four-momentum $k=(E_\nu, {\bf k})$ which scatters from
a nucleus
via $W^{\pm}$ boson exchange producing a charged lepton with
four-momentum
$k'=(E_{\bf k'}, {\bf k'})$. Using impulse approximation,
the double differential scattering cross section from a target
nucleus with
mass number $A$ is
given by (we assume a symmetric $N=Z$ nucleus):
\begin{eqnarray}
{d^2\sigma\over {d\Omega _{\bf k'} d E_{\bf k'}}}=- {A G_F^2\,
{\rm cos^2\theta _c}
%\,|{\bf k}'|\over
\,|{\bf k}'|\over
{32 \pi^3 \rho E_\nu}}\,{{\rm Im}\,[L_{\mu\nu}\ \Pi_A^{\mu\nu}]}\ .
\label{dcross}
\end{eqnarray}
Here $\rho=2k_F^3/3\pi^2$ is the baryon density with Fermi
momentum $k_F$,
$\theta_c$ the Cabbibo angle (${\rm cos}^2\theta_c=0.95$), $G_F$
is the Fermi
constant.
The leptonic tensor $L_{\mu\nu}$ is defined as
\begin{equation}
L_{\mu\nu} = 8 \left( k_\mu k'_\nu + k_\nu k'_\mu -
k \cdot k' g_{\mu\nu} \mp i \epsilon_{\alpha\beta\mu\nu}
k^{\alpha} k'^{\beta} \right)\ ,
\label{lep}
\end{equation}
with the minus (plus) sign denoting neutrino (anti-neutrino)
scattering.
$\Pi^{\mu\nu}_A$ is the polarization tensor of the target nucleus
for the
charged weak current. Here we consider p-h, $\Pi^{\mu\nu}_{ph}$,
and
$\Delta$-h, $\Pi^{\mu\nu}_{\Delta h}$, contributions to the
polarization:
\begin{equation}
\Pi^{\mu\nu}_A=\Pi^{\mu\nu}_{ph}+\Pi^{\mu\nu}_{\Delta h}\ .
\end{equation}
In the impulse approximation, the p-h polarization takes a simple
form,
\begin{equation}
i\Pi^{\mu \nu}_{ph} = \int{d^4p\over (2\pi)^4}\,
{\rm Tr}[G(p+q)\,\Gamma^\mu \,G(p)\,\Gamma^\nu] \;,
\label{tun}
\end{equation}
with the weak-interaction vertex given in terms of single-nucleon
form factors parameterized from on-shell data,
\begin{eqnarray}
\Gamma^\mu(q^2)&=& F_1(q^2)\gamma^\mu +
iF_2(q^2) \sigma^{\mu\nu} {q_\nu \over 2M} -
G_A(q^2)\gamma^\mu \gamma^5 +
F_p(q^2)q^\mu\gamma^5\nonumber\ ,\\
&&\quad (q^2 \equiv q_{0}^{2}-{\bf q}^{2})
\label{cur}
\end{eqnarray}
The form factors $F_1, F_2, G_A$ and $F_p$ are given in the
appendix of
Ref.~\cite{hung}. The
pseudoscalar form factor $F_p$ is constructed from PCAC, and its
contribution
is suppressed by the small lepton mass.
In an
approximation where the nucleus is in a mean-field ground state of
the Walecka model,
the nucleon propagator $G(p)$
is given by~\cite{brian}
\begin{eqnarray}
G^*(p)=(\not\!{p}^{*} + M^*)
\biggl[{1\over p^{*2}-M^{*2} +i\epsilon} +
{i\pi \over E^*_{\bf p}} \delta({p}_{0}^{*}-E^{*}_{\bf p})
{\rm \theta}(k_F-|{\bf p}|)\biggr] \label{gmf} \;,
%&&\equiv G_F(p)+G_D(p) \;.
\end{eqnarray}
where the effective mass $M^*$ and energy $E_p^*$ are shifted from
their
free-space value by the scalar ($S$) and timelike component
($V$) of the mean fields,
\begin{equation}
M^*=M+S \;; \quad E^{*}_{\bf p} =\sqrt{{\bf p^2}+M^{*2}} \;; \quad
{p}^{*\mu}=(p^0-V,{\bf p}) \;.
\end{equation}
In the Fermi gas approximation of non-interacting particles,
$V$ and $S$
are set to zero.
Analytic expressions for the imaginary part of
the polarization [Eq.~(\ref{tun})] can be found in Ref.~\cite{hung}.
The impulse approximation can be improved
by including RPA effects. An RPA
calculation uses the same expression for the cross section as in
Eq.~(\ref{dcross}) with the replacement:
\begin{equation}
\Pi^{\mu\nu} \rightarrow \Pi^{\mu\nu}_{RPA}=
\Pi ^{\mu\nu}+\Delta\Pi_{RPA}^{\mu\nu} \;.
\end{equation}
$\Delta\Pi_{RPA}^{\mu\nu}$ represents many-body correlations mediated
by isovector particles, $\pi$ and $\rho$ incorporated with the
Landau-Migdal parameter $g'$.
The cross section involving $\Delta$-h excitations is calculated
similarly
in Hartree approximation~\cite{delta} using the $\Delta$-h
polarization
$\Pi_{\Delta h}^{\mu\nu}$.
The weak interaction contains vector current ($v$) and
axial-vector current ($a$) contributions,
therefore, we split $\Pi_{\Delta h}^{\mu\nu}$ into
\begin{equation}
\Pi^{\mu\nu}_{\Delta h} = (\Pi_{\Delta h}^{vv})^{\mu\nu}
+ (\Pi_{\Delta h}^{aa})^{\mu\nu}
+ (\Pi_{\Delta h}^{va})^{\mu\nu}
+ (\Pi_{\Delta h}^{av})^{\mu\nu}\ .
\end{equation}
Here $(\Pi_{\Delta h}^{va})^{\mu\nu}$ and
$(\Pi_{\Delta h}^{av})^{\mu\nu}$
are interference terms of the vector and axial-vector
currents. These polarization tensors are written in terms of the
spin $3/2$
delta propagator and appropriate weak vertices as,
\begin{eqnarray}
(\Pi_{\Delta h }^{ij})_{\mu\nu} =&-& i \int{d^4p\over (2\pi)^4}\,
{\rm Tr}[\Gamma^i_{\beta\mu}(-q,-p)\ S^{\beta\alpha}(p)\
\Gamma^j_{\alpha\nu}(q,p)\ G(p-q)\ ]\nonumber \\
&+& (q_\mu \rightarrow -q_\mu)\ ~~~~~~~~~~(i,j) = (a,v)\ .
\label{eq:pol}
\end{eqnarray}
The trace is taken for the Dirac matrices as well as the isospin
matrices.
For $S^{\mu\nu}(t)$ we take the Rarita-Schwinger
form of the free spin 3/2 propagator with momentum
$t$ :
\begin{equation}
S^{\mu\nu}(t)=-{\not\!t + M_\Delta \over t^2 - M_\Delta^2+i\epsilon}
\biggr
[ g^{\mu\nu}
-{1 \over 3} \gamma^\mu \gamma^\nu-{2 \over 3}
{t^\mu t^\nu \over M_\Delta^2} +
{t^\mu \gamma^\nu - t^\nu \gamma^\mu \over 3 M_\Delta } \biggl ]\ .
\end{equation}
The vector part of the nucleon-delta vertex has been studied in the
case of the
$\gamma N \Delta$ transition~\cite{wehr89},
\begin{eqnarray}
\Gamma^{V}_{\mu\nu} (q,p) &&= \sqrt{2} F^\Delta(Q^2) T^{\pm} \biggr
[ (-q_\mu \gamma_\nu+g_{\mu\nu}\not\!q) M_\Delta \gamma_5 +
(q_\mu p_\nu -q \cdot p g_{\mu\nu}) \gamma_5 \biggl]\nonumber \\
&&\equiv \sqrt{2} F^\Delta T^{\pm} V_{\mu\nu}\
\end{eqnarray}
with the isospin raising or lowering operator $T^{\pm}$ \cite{delta}.
The vertex for the axial $N\Delta$ transition is given
by~\cite{tjon,ann,towner}
\begin{eqnarray}
\Gamma^A_{\mu\nu}&=&-{r_{N \Delta} \over \sqrt{2} }
G_A T^{\pm}
(g_{\mu\nu}-{\gamma_\mu \gamma_\nu \over 4})\
\end{eqnarray}
with $r_{N\Delta} = f_{\pi N\Delta}/f_{\pi NN} \sim 2$.
The form factors $F^\Delta$ and $G_A$ are defined in the appendix of
Ref.~\cite{delta}.
In mean-field approximation, the propagation of a $\Delta$ is modified
by background scalar and vector fields similarly as the nucleon
propagator.
This can be understood from a chiral soliton model where the delta
is a rotational excitation of the nucleon. That is, in the medium,
the delta is expected to be influenced by the same magnitude of the
vector and
scalar potentials as the nucleon's because $\sigma$ and $\omega$ are
isoscalar.
The calculation of the cross section
proceeds in the same way as in the case of the free $\Delta$.
To include the delta decay width in our calculation, we average the
nuclear
response over the delta mass with a Breit-Wigner
distribution~\cite{wehr89}.
The averaged cross section follows as
\begin{eqnarray}
<{d^2\sigma\over {d\Omega _{k'} d E_{\bf k'}}}> &\sim&
\int d\mu^2 {d^2\sigma\over {d\Omega _{k'} d E_{\bf k'}}} (\mu)
f(\mu^2) ~/~ \int d\mu^2 f(\mu^2)~~~~, \\
f(\mu^2) &=&
{M_\Delta \Gamma_\Delta \over (M_\Delta^2-\mu^2)^2 +
M_\Delta^2 \Gamma_\Delta^2 }\
\end{eqnarray}
integrating from threshold to infinity.
However,
the decay width of a delta in the nuclear medium is not well
determined
both theoretically and experimentally. The $\pi$N decay
channel of the delta is partially blocked by Pauli blocking but
there is
an additional spreading width. Thus, determining the in-medium
decay width
is a nontrivial problem.
Here, we assume the in-medium width
to be the same as the free width $\Gamma_{\Delta}=115$ MeV.
More theoretical and experimental work has to be done regarding this
question, however.
\section{ Coherent Pion Production}
\label{sec:coh_for}
Coherent pions are produced as decay products of virtual p-h or
$\Delta$-h excitations of a finite nucleus. More specifically,
the momentum transfer from the incoming neutrino virtually excites
the nucleus through p-h or $\Delta$-h and the nucleus decays back to
its ground state by emitting a pion. Here we derive the formalism for
the $\Delta$-h excitation as it dominates
the p-h excitations of the nucleus. The corresponding scattering
diagram is
shown in Fig.~\ref{cohfig}.
Note that the coherent pion is on-shell even though the momentum
transfer in the initial scattering is space-like.
The missing momentum is provided
by the recoil of the nucleus.
In a local density approximation, the finite-nucleus polarization
$\Pi_{\rm FN}$$(q,q';\omega)$ with incoming and outgoing momenta
$q$ and $q'$, respectively, is approximated in terms of the nuclear
matter
polarization $\Pi_{\rm NM}$$(q,q';\omega)$,
\begin{equation}
\Pi^{\mu}_{\rm FN}(q,q';\omega) =
{\rm V}\ {\rm F}(|{\bf q} -{\bf q'}|)\
\Pi^{\mu}_{\rm NM}(q,q';\omega)~~ ,
\end{equation}
where ${\rm F}(|{\bf q} -{\bf q'}|)$ is the elastic
form factor of the target nucleus. We obtain
${\rm F}(|{\bf q} -{\bf q'}|)$
by Fourier transforming the nuclear ground state density
obtained from a self-consistent relativistic mean
field calculation for finite nuclei~\cite{horo81}.
The volume for a target nucleus with mass number A is given by
$${\rm V}={3 \pi^2 {\rm A} \over 2 k_F^3}\ .$$
As shown in Fig.~\ref{cohfig}, the incoming neutrino interacts
weakly
with the nucleus exciting an intermediate $\Delta$-h state.
The vertex of the weak interaction contains
vector and axial-vector parts while
the $\pi N\Delta$ vertex, on the other hand,
involves only the
axial-vector part contracted with a pion four-momentum $q'$.
Since the coherent-pion production cross section is dominated by
pion momenta ${\bf q'}$
parallel to the momentum transfer ${\bf q}$,
there is negligible contribution from the part of the polarization
that mixes the axial-vector and vector current.
We ignore this small contribution throughout this paper.
Consequently, the cross section for coherent pion production is
identical for neutrinos and antineutrinos.
The double differential cross section for coherent pion production is
\begin{eqnarray}
{d^2\sigma\over d E_\pi d\Omega_{\bf k'}}&=&
{|{\bf k'}| |{\bf q'}|\ G^2_F\ {\rm V}^2 \over
256 \pi^4 E_\nu}
\int^1_{-1} d {\rm cos} \theta_{q'}\ {\rm F}^2({\rm cos} \theta_{q'})
\biggr[ L_{00} |\Pi_{\rm NM}^{0}|^2\nonumber \\
&& + 2 L_{01}\ {\rm Re}[\Pi_{\rm NM}^{0} (\Pi^{1}_{\rm NM})^*]
+L_{11}|\Pi_{\rm NM}^{1}|^2 \biggl]\label{double}\ ,
\end{eqnarray}
where $\theta_{q'}$ is the angle between ${\bf q}$ and ${\bf q'}$.
Here we observed that the integral in Eq.~(\ref{double})
is dominated by small angles, which implies that
contributions from transverse
components of the polarization, $\Pi_{\rm NM}^{2}$ and
$\Pi_{\rm NM}^{3}$,
are suppressed. Energy conservation requires
the pion energy $E_\pi$ to be equal to the energy transfer to the
nucleus $q_0$.
The pion three-momentum $\bf{q'}$ is obtained from the mass-shell
condition
\begin{equation}
|{\bf q'}|=\sqrt{q_0^2-m_\pi^2}\ .
\end{equation}
The antisymmetric part of the leptonic tensor $L_{\mu\nu}$
does not contribute to
Eq.~(\ref{double}), as it is contracted with the vector axial-vector
mixing term of the polarization. The nuclear matter polarization,
$\Pi_{\rm NM}^\mu$, can be
evaluated for coherent charged or neutral pion production by taking
appropriate vertices.
The calculation of the real and imaginary parts of
the polarizations $\Pi_{\rm NM}^0$ and $\Pi_{\rm NM}^1$ is summarized
in Appendix.
In a relativistic mean field approximation,
the properties of the nucleon and $\Delta$ are modified
by strong scalar and vector fields in the
nuclear medium. These effects can be incorporated by
modifying
the nucleon and delta propagators as we described in the
inclusive calculation.
Since the axial-vector vertex does not contain the energy
of the $\Delta$ or nucleon, the contribution
from the constant vector mean-fields is eliminated in the calculation
of the polarization by a simple
change of integration variable.
As the $\Delta$ is unstable, we include a width $\Gamma_\Delta$
by using a complex mass $M^c_\Delta$
\begin{equation}
M_\Delta \rightarrow M^c_\Delta \equiv
M_\Delta-i\,\Gamma_\Delta /2\ ,
\end{equation}
in the denominator of the delta propagator~\cite{tjon}. This
prescription
yields similar results as the usual Breit-Wigner folding. Note, we
have not included pion distortions in Eq.~(\ref{double}). These may
somewhat reduce the cross section. The effect
of distortions and a possible $\Delta$-h spreading potential
remain to be
investigated.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Results}
\label{sec:result}
In this section, we present numerical results of our calculation.
We start with the inclusive cross sections and their role in
atmospheric
neutrinos followed by a discussion of the coherent pions.
In the case of the inclusive cross section, we
discuss the case of muon neutrinos,
but the general features of the results hold for electron neutrinos as
well. In addition to the relativistic Fermi gas description of
the nucleus, the effects of
mesonic mean fields are also considered in the mean-field
approximation (MFA). The
target nucleus is assumed to be $^{16}$O with a
Fermi momentum $k_F=225$ MeV.
In MFA, we use the same scalar and vector couplings for nucleon
and delta
(``universal coupling''). For $k_F=225$ MeV, the effective masses
are
\begin{equation}
M^*_\Delta = 931 {\rm MeV}\ ; \quad M^*_N = 638 {\rm MeV}\ ,
\end{equation}
and the vector self-energy is $V = 239$ MeV.
In atmospheric neutrino experiments such as Kamiokande and
IMB~\cite{imb,kamiokande}, the cross section of interest
is $d\sigma /dE_{\bf k'}$, which is
obtained from Eq.~(\ref{dcross}) by another numerical integration
over the scattering angle of the outgoing lepton.
Fig.~\ref{dsde} shows the cross section
$d\sigma /d E_{\bf k'}$ for $E_\nu=1$ GeV.
Note that the cross sections from the p-h excitation are peaked at a
large
lepton energy, which corresponds to a small nuclear excitation energy.
This is
the region where nucleon correlations are important. The nucleon
correlations
are modeled using RPA which includes a ($\rho + \pi +g'$)
residual isovector interaction. The RPA calculation\footnote{Note
that
our RPA calculations do not include
$\Delta$-h or mixtures of p-h and $\Delta$-h.} in the mean-field
ground state is also shown in Fig.~\ref{dsde}.
The cross section for stable deltas
vanishes around $E_{\bf k'}=0.7$ MeV, which gives the minimum energy
transfer for delta production.
The curves including the delta width, however,
spread over to higher lepton energies because of the averaging
process.
The mean fields reduce the p-h and the $\Delta$-h responses by 30\%
at the peaks. However, RPA effects reduce the peak by 50\% at high
muon
energies. The reduction of the $\Delta$-h response by mean fields
is interesting in the
Monte Carlo simulation of atmospheric neutrino experiments.
The strength of the $\Delta$-h cross section provides
the number of pions produced from the $\Delta$ decays.
These pions may cause an uncertainty in determining the lepton flavor
in the experiment.
Indeed, the Monte Carlo simulation of the
Kamiokande detector~\cite{monte} takes into account the number
of pion events based on a nonrelativistic model by Fogli and
Nardulli~\cite{fogli}. Their nonrelativistic results are
qualitatively
reproduced in our free delta calculations. Thus
mean-field effects are not included in the existing simulations
and could
provide significant systematic errors.
To get a rough estimate of the reduction in the number of events,
we fold
$d\sigma /d E_{\bf k'}$ with a simple model of the atmospheric
neutrino flux. We approximately fit the atmospheric muon neutrino
flux
with the formula
~\cite{kamiokande}
\begin{equation}
\phi_{\nu_\mu}(E_\nu) = 220\ E_\nu^{-2.5}
{1 \over {\rm m}^2\ {\rm sr}\ {\rm GeV}\ {\rm sec}}\ ,
\end{equation}
and assume
\begin{equation}
\phi_{{\bar \nu}_\mu}=\phi_{\nu_\mu}=2 \phi_{\nu_e}=2
\phi_{{\bar\nu}_e}\ .
\end{equation}
Using this flux, we calculate the yield Y in units of events per
Kton year of
detector exposure and per 100 MeV energy bin,
\begin{equation}
{\rm Y} =
1.1938 \times 10^{40}
\int \phi_{\nu_\mu}(E_{\nu_\mu})
{d\sigma \over d E_{\bf k'}} dE_{\nu_\mu}\label{events}\ ,
\end{equation}
and plot the result in Fig.~\ref{yield} (a) with respect to the
muon energy
for the p-h and $\Delta$-h excitations.
Note, the $\Delta$-h calculations include the decay width of the
delta of 115 MeV.
However, in the case of the inclusive process
the quantitative results are insensitive to the decay width.
At low muon energies,
the mean fields in the impulse approximation reduce the total p-h
events from
33 to 28 while including RPA yields 22 events (in a 100 MeV bin after
one Kton Year of exposure assuming 100 percent efficiency).
Thus, nuclear structure effects reduce the quasielastic
charged-current events substantially.
Since the electron events are also reduced by a similar amount by
RPA correlations, this reduction does not directly affect the
ratio $r$ of
Eq.~(\ref{rat}). However, it can change the
{\it energy dependence} of the cross section.
Also, as one can see from Fig.~\ref{yield}~(b), the $\Delta$-h
contribution
is about 40\% of the RPA calculation at its minimum and, as the
muon energy
increases, the $\Delta$-h response becomes more important.
The $\Delta$-h
response is even larger than the RPA response at high muon energies
(the
ratio is 1.06 at $E_{\bf k'} = 2$ GeV). Note that the $\Delta$-h
response
is reduced by 10\% to 25\% due to mean-fields. Incoherent pion
events
should be decreased correspondingly.
Therefore, the reduction due to the $\Delta$ mean-fields combined
with
the event reduction from RPA
should be properly taken into account in the simulation.
Not all deltas decay to $\pi$N in the nuclear medium.
In fact, in the medium the $\pi$N decay is partially
suppressed because of Pauli blocking~\cite{wehr90}. However, the
$\Delta$
in the medium has additional channels of pionless decay.
The coincidence experiments of ($p$, $n$) reactions at
the $\Delta$ resonance~\cite{had} showed that a large portion
of deltas in the medium decay without emitting pions.
Furthermore, in theoretical calculations, a value of about
70 MeV has been used as the 2p-2h partial width~\cite{moniz,jain}.
Thus, experimental and theoretical results indicate a
substantial amount of pionless delta decay in the medium.
This 2p-2h
decay mode of the delta has not been directly included in present
detector simulations. At best, its effects have been partially
incorporated using a very crude model of pion absorption without
any explicit reference to $\Delta$ production.
The pionless decay of the delta and the reduction due to the
mean-fields, suggest that the simulations may substantially
overcount the total number of pions.
At this moment, we do not know how these effects change the
result for the atmospheric neutrino anomaly.
But the results indicate that the Monte Carlo simulations in current
atmospheric neutrino experiments can be improved to treat pion events
more accurately.
Now we present results for coherent pion production in neutrino
scattering. Coherent pions may be important in atmospheric
neutrino experiments
because they could increase the electron-like events and
therefore reduce the ratio $r$ of Eq.~(\ref{rat}).
A coherent neutral pion produced in neutral-current scattering
($\nu$, $\nu'$)
immediately decays into two photons and may be counted as
an electron-like event. Likewise, a coherent charged pion
can also be confused with an electron-like event when the
two charged particles in the final state, muon (or
electron) and charged pion, move with small opening angle and make
an isolated electromagnetic shower. Indeed small-angle scattering
dominates the coherent cross section.
The single differential cross section
$d\sigma/dE_\pi$ is obtained by numerically integrating
Eq.~(\ref{double})
over the scattering solid angle. Again, we use the free
width $\Gamma_\Delta = 115~$MeV in our calculation.
Pauli-blocking and additional decay channels for the $\Delta$ in the
medium
will change this value (and generate a non-trivial
momentum-dependence).
However, as theoretical in-medium values are still very uncertain,
we will not explore modifications of the width at this point.
More work in this direction has to be done in the future.
In the atmospheric neutrino experiments, both $\mu$- and
$e$-neutrinos
participate in producing coherent pions. Furthermore, in contrast
to the inclusive
scattering process where the cross sections of antineutrinos are
strongly suppressed, neutrino and antineutrino contributions
to coherent pion production are of the same size.
Combining the events from $\mu$-type antineutrinos as well as $e$-type
neutrinos and antineutrinos could enhance coherent pion production
substantially,
and it is of interest to quantitatively compare the coherent
pion events
directly with the inclusive charged-current events through
p-h nuclear excitation for a realistic spectrum of neutrinos.
Using the $e$-neutrino flux ($\phi_{\nu_e}=\phi_{{\bar\nu}_e}$),
we calculate the coherent pion events and compare them with
electron production in
$\nu_e$ scattering. The neutral pion events per kiloton-year are
\begin{equation}
{\rm Y}(\pi^0) =
6 \times1.194 \times 10^{40} \int \phi (E_{\nu_e})
{d\sigma \over d E_\pi} (E_{\nu_e}) dE_{\nu_e}\label{po-events}\ .
\end{equation}
The factor 6 includes the contribution from antineutrinos
as well as a factor of 2 from $\mu$-neutrinos. Note, for the
purposes of illustration, let us assume all coherent
pion events result in electron-like tracks. The actual
identification
of these events depends on the details of a detector.
Fig.~\ref{ncohfig}~(a) shows the coherent $\pi^0$
events with and without mean fields and $\Delta$ decay width.
The electron events from are obtained from
\begin{equation}
{\rm Y}(e) =
1.194 \times 10^{40} \int \phi (E_{\nu_e})
{d\sigma \over d E_{\bf k'}} (E_{\nu_e})
dE_{\nu_e}\label{ve-events}\ .
\end{equation}
The differential cross section
$d\sigma/d E_{\bf k'}$ is calculated
in relativistic impulse and
in random phase approximation~\cite{hung}.
Note that the horizontal
axis in Fig.~\ref{ncohfig}~(a) shows either
the coherent pion or the electron energy.
Figure~\ref{ncohfig}~(a) also shows that at roughly $E=300$ MeV the
coherent pion events, neglecting decay width and
mean fields, are comparable in size to electron events calculated
in RPA.
This is certainly interesting because the electron-like events
can be increased by these pions and therefore might help to
explain the small ratio $r$
of atmospheric neutrinos at $E=300$ MeV. However,
inclusion of mean fields reduces the pion events by a
factor of 2. Furthermore,
the $\Delta$ decay width exacerbates the situation:
the number of events including mean fields and
$\Delta$ decay width (thin dashed line) are only 10\% of
the electron events calculated in RPA.
Next, we consider the coherent charged pion events.
In this case, the
\v{C}erenkov light detector
records the total energy of pion and muon (or electron)
assuming they are moving with small opening angle and make an
isolated
single track.
Overall energy conservation implies
\begin{equation}
E_{\nu_\mu}=E_\pi+E_\mu~~~ {\rm or}~~~~ E_{\nu_e}=E_\pi+E_e
\end{equation}
depending on the incoming neutrino type.
The coherent charged pion events are calculated from
\begin{eqnarray}
{\rm Y}(\pi^{\pm}) &=&
2 \times\ 1.194 \times 10^{40} \phi (E_{\nu_e})
\int^{E_\nu-m_e}_{m_\pi}
{d\sigma \over d E_\pi} dE_\pi\nonumber\ \\
&+&
4 \times\ 1.194 \times 10^{40} \phi (E_{\nu_e})
\int^{E_\nu-m_\mu}_{m_\pi}
{d\sigma \over d E_\pi} dE_\pi\
\label{c-events}\ .
\end{eqnarray}
The first term results from $e$-neutrinos (the factor of 2
includes contribution from antineutrinos) and the second term
originates from
muon-type neutrinos (with a total flux factor of 4).
Figure~\ref{ncohfig}~(b) shows the coherent charged pion
events and the electron events calculated in RPA.
As in the case of $\pi^0$s neglecting mean field and decay width
of the
$\Delta$ results in cross sections for the coherent pions comparable
or even larger than the quasielastic ones.
However, again mean fields and the
decay width substantially reduce the pion events: At $E=1$ GeV, the
full calculation
amounts to only 23\% of
the quasielastic electron events, but at $E=2$ GeV,
it increases to 30\%.
Since the pion events become important as $E$ increases, our
findings provide an important systematic error to
the interpretation of the recent Kamiokande experiment
in the multi-GeV energy range~\cite{mult-gev}.
The effect of neutral and charged coherent pions can be substantial,
the actual numbers are very sensitive to the delta decay width,
though.
Coherent $\pi^0$ production could be
large at $E=300$ MeV while coherent charged pions may become
important for high visible energies.
To solidify our results, however, a more elaborate study of the
$\Delta$ properties in the nucleus has to be performed.
\section{Summary}
\label{sec:con}
In this work, we have calculated neutrino-nucleus cross sections
using
a relativistic formalism to investigate possible improvements for
the Monte Carlo simulation of Kamiokande experiment.
We have found that RPA corrections reduce the charged-current neutrino
events by up to 37\% for particle-hole excitations.
Because of this reduction, the $\Delta$-h response could be
important and is found to give significant corrections to
quasi-elastic
nucleon knock-out processes. We have found that the relativistic
mean-fields reduce the $\Delta$-h responses by 10\% to 24\%, which
correspondingly reduce the pion events in atmospheric neutrino
experiments.
Combined with the additional channel of the non-pionic decay of the
delta in the medium, our findings suggest possible uncertainties
in the present Monte Carlo simulations of detectors and the
interpretation of the measured data.
We have also discussed coherent pion production in
neutrino-nucleus scattering.
We have shown that coherent pions can produce a sizeable background
in the atmospheric neutrino measurements. As neutral pions and
collinearly outgoing charged leptons and pions may mimic
electron-type events in the detector, this effect can
produce uncertainties in determining the flavor ratio of the
incoming neutrinos.
Coherent pions may provide important systematic errors for the
recent Kamiokande experiment in the multi-GeV energy
region~\cite{mult-gev}.
In order to get a clearer quantitative estimate of the influence
of coherent pions, improved calculations should be done and
these should be included in detector simulations
of atmospheric neutrino experiments
\section*{Acknowledgments}
This research was supported in part by the DOE under Grant
No. DE-FG02-87ER-40365
and the NSF under Grant No. NSF-PHY91-08036.
The work of Hunchong Kim was also
supported in part by the Basic Science Research Institute program
of the Korean Ministry of Education through grant no. BSRI-95-2425.
\eject
\appendix
\section*{}
Here we evaluate the nuclear matter polarization,
$(\Pi_{\rm NM})_{\mu}$,
for coherent pion production.
The polarization is defined as
\begin{eqnarray}
(\Pi^j_{\rm NM})_{\mu} &=& - i \int{d^4p\over (2\pi)^4}\,
{\rm Tr} \biggr[\Gamma^j_{\mu\beta}\ S^{\beta\alpha}(p')\
\Gamma^j_\alpha (q') \ G^o(p)\biggl]
+ (q_\alpha \rightarrow -q_\alpha)\nonumber \\
&\equiv& \Pi^j_{\mu\nu} (q) q'^\nu\;;~~~ j=(\pi^{\pm}, \pi^0) ,
\end{eqnarray}
where $p'=p+q$.
The superscript $j$ indicates the polarization for the charged
or neutral pion production.
The axial-vector ($\Gamma^j_{\mu\nu}$) and pion
$(\Gamma^j_\nu)$ vertices can be found from Ref.~\cite{tjon,ann,towner}.
The noninteracting nucleon propagator in Hartree approximation reads
\begin{equation}
G^o(p)=(\not\!p + M) {i\pi\over
E_{\bf p}}\delta(p_0-E_{\bf p}){\rm \theta}(k_F-|{\bf p}|)\ .
\end{equation}
The polarization $\Pi^j_{\mu\nu} (q)$ involves only axial-vector
vertex and
is given by
\begin{eqnarray}
\Pi^j_{\mu\nu} (q)&=&-i C^j(q^2)
\int{d^4p\over (2\pi)^4}{\rm Tr}
\Biggr[\Gamma_{\mu\beta} S^{\beta\alpha}(p')
\Gamma_{\alpha\nu}\ G^o(p)\Biggl] +
(q_\alpha \rightarrow -q_\alpha)\label{axial}\ ,
\end{eqnarray}
where
$\Gamma_{\alpha\nu}=g_{\alpha\nu}-\gamma_\alpha \gamma_\nu / 4$.
The off-shell term ($\gamma_\alpha \gamma_\nu / 4$) insures
that the axial
vector vertex satisfies the relation $\gamma^\mu \Gamma_{\mu\nu}=0$.
The coefficient functions $C^j(q^2)$ are given as
\begin{equation}
\cases{{4 f_{\pi N \Delta} \over 3 \sqrt{2} m_\pi}\
{\rm cos} \theta_c\
G_A(q^2)\ r_{N\Delta}~~~~{\rm for}~~~~j=\pi^{\pm}; \cr
\cr
{2 f_{\pi N \Delta} \over 3 m_\pi}
G_A(q^2)\ r_{N\Delta}~~~~~~~~~~~~~{\rm for}~~~~j=\pi^0\ . \cr}
\end{equation}
Here ${\rm \theta}_c$ is the Cabbibo angle and $G_A (q^2)$ is the
axial-vector form factor for the p-h excitation. The value
of the axial vector $N\Delta$ transition strength $r_{N\Delta}$
is somewhat
model dependent and we choose an intermediate value 2
in our calculation.
After some algebra, Eq.~(\ref{axial}) becomes
\begin{equation}
\Pi^j_{\mu\nu}(q) =- C^j (q^2) \int_M^{E_F}
dE_{\bf p}
\int_{-1}^{1} d \chi \,{ |{\bf p}| \over 8 \pi^2} {T_{\mu\nu}
\over (p+q)^2 - M_\Delta^2 + i\epsilon } +
(q_\alpha \rightarrow -q_\alpha)\ .
\label{co_pol}
\end{equation}
Here $T_{\mu\nu}$ is the result of the Dirac trace,
\begin{eqnarray}
T_{\mu\nu}&=&{1 \over 6 M_\Delta^2}\biggr[3p_\mu p'_\nu p'^2
+ 3 p_\nu p'_\mu p'^2 + p \cdot p'\ g_{\mu\nu} p'^2
- 16 p \cdot p' p'_\mu p'_\nu\nonumber \\
&&-3 M_\Delta^2 (p_\mu p'_\nu + p_\nu p'_\mu - 5 p\cdot p'g_{\mu\nu}
)\nonumber\\
&&+ 2 M M_\Delta(9 g_{\mu\nu}\,M^2_\Delta - g_{\mu\nu}p'^2
- 8 p'_\mu p'_\nu) \biggl]\ ,
\end{eqnarray}
and $E_F$ is the Fermi energy
$E_F=\sqrt{k_F^2+M^2}$,
and the $\chi={\rm cos}\theta$ is the angle between
{\bf q} and {\bf p}.
As long as $\Delta$ is stable, the angular
integration can be done analytically while the
remaining integration over
$E_{\bf p}$ is performed numerically. A similar calculation
in a mean
field approximation can be done by replacing $M$ and $M_\Delta$
by the effective masses. The contribution from vector mean field
is eliminated
by a change of variable in the energy integration.
We include the $\Delta$ decay width by replacing the $\Delta$
mass in the denominator of Eq.~(\ref{co_pol}) as
\begin{eqnarray}
M_\Delta \rightarrow M^c_\Delta \equiv
M_\Delta-i\Gamma/2\ ,
\end{eqnarray}
and the double integrations are done numerically.
\begin{references}
\bibitem{imb} {R. Becker-Szendy {\it et al.},
Phys.\ Rev.\ D {\bf 46}, 3720 (1992).}
\bibitem{kamiokande} {K. S. Hirata {\it et al.},
Phys.\ Lett. B {\bf 280}, 146 (1992);
K. S. Hirata {\it et al.},
Phys.\ Lett. B {\bf 205}, 416 (1988);
E. W. Beier {\it et al.},
Phys.\ Lett. B {\bf 283}, 446 (1992).}
\bibitem{eng} {J. Engel, E. Kolbe, K. Langanke, and P. Vogel,
Phys. \ Rev. \ D {\bf 48}, 3048 (1993).}
\bibitem{solar} {GALLEX Collaboration,
Phys. \ Lett. B {\bf 285}, 376 (1992) ;
A. I. Abazov {\it et al.},
Phys. \ Rev. \ Lett. {\bf 67}, 3332 (1991);
K. S. Hirata {\it et al.},
Phys.\ Rev. \ D {\bf 44}, 2241 (1991).}
\bibitem{monte} {Masayuki Nakahata {\it et al.},
J. \ Phys.\ Soc.\ of Japan {\bf 55} 3786 (1986).}
\bibitem{fogli} {G. L. Fogli and G. Nardulli,
Nucl.\ Phys. {\bf B160}, 116 (1979).}
\bibitem{hung} {Hungchong Kim, J. Piekarewicz and C.J. Horowitz,
Phys. \ Rev. \ C {\bf 51}, 2739 (1995).}
\bibitem{brian} {B. D. Serot and J. D. Walecka, in
{\it Advances in Nuclear Physics},
edited by J.W. Negele and E. Vogt
(Plenum, New York, 1986), Vol. 16.}
\bibitem{delta} {Hungchong Kim, S. Schramm and C. J. Horowitz,
To be published in Phys. Rev. C.}
\bibitem{wehr89} {K. Wehrberger, C. Bedau and F.Beck,
Nucl.\ Phys. {\bf A504}, 797 (1991).}
\bibitem{tjon} {M. J. Dekker, P. J. Brussaard, and J. A. Tjon,
Phys.\ Rev.\ C {\bf 49}, 2650 (1994).}
\bibitem{ann} {P. Carruthers and Michael Martin Nieto,
Ann.\ Phys. (N. Y.) {\bf 51}, 359 (1969).}
\bibitem{towner} {I. S. Towner,
Phys.\ Rep. {\bf 155}, 263 (1987).}
\bibitem{horo81} {C. J. Horowitz and B. D. Serot,
Nucl.\ Phys.\ {\bf A368}, 503 (1981).}
\bibitem{wehr90} {K. Wehrberger and R. Wittman,
Nucl.\ Phys.\ {\bf A513}, 603 (1990).}
\bibitem{had} {T. Hennino {\it et al.},
Phys.\ Lett. B {\bf 283}, 42 (1992);
J. Chiba {\it et al.},
Phys.\ Rev.\ Lett. {\bf 67}, 1982 (1991).}
\bibitem{jain} {B. K. Jain, J. T. Londergan and G. E. Walker,
Phys.\ Rev.\ C {\bf 37}, 1564 (1988).}
\bibitem{moniz} {E. J. Moniz and Alpar Sevgen,
Phys.\ Rev.\ C {\bf 24}, 224 (1981).}
\bibitem{mult-gev} {Y. Fukuda {\it et al.},
Phys.\ Lett. {\bf B335}, 237 (1994).}
\end{references}
\eject
\begin{figure}
\caption {Feynmam diagram for coherent pion production
through $\Delta$-h excitations.
In the charged-current reaction, $k'$ is the four momentum
of the muon or electron, and
$q$ is the momentum of the $W^{\pm}$ boson.
In the neutral-current
reaction, $k'$ is the momentum of the scattered neutrino
and $q$
the momentum of the exchanged $Z^0$.}\label{cohfig}
\end{figure}
\begin{figure}
\caption{Differential cross section
$d\sigma/dE_{\bf k'}$ as a function of
muon energy
$E_{\bf k'}$. The thick solid and dashed curves are for the
free $\Delta$-h calculations, without and with decay width,
respectively. The thick dot-dashed curve is the free p-h calculation.
The corresponding thin curves are for MFA.
The short dashed curve is the RPA
p-h calculation.}\label{dsde}
\end{figure}
\begin{figure}
\caption{(a) shows events per kiloton-year versus muon energy.
The bold
solid line represents
the free $\Delta$-h calculation with decay width 115 MeV, and the
bold dashed line is for $\Delta$-h in MFA. The thin lines
are the corresponding p-h results. Events calculated in RPA are
given by
the dot-dashed curve.
In figure (b), the solid curve is the ratio of
mean field to free Fermi gas calculations for p-h excitations
and the dashed curve is
the similar ratio of $\Delta$-h events. The dots gives the ratio of
RPA to free p-h
and the dot-dashed curve is the ratio of free $\Delta$-h over an RPA
calculation of p-h yields. }\label{yield}
\end{figure}
\begin{figure}
\caption{(a) shows electron events calculated in the Fermi gas model
(short-dashed curve)
and RPA (dot-dashed) involving only p-h excitations
versus electron energy. Also shown are the
coherent $\pi^0$ events
calculated for $\Delta$'s with (bold long-dashed)
and without decay
width (bold solid). The corresponding
thin curves are the cases including mean fields.
(b) shows the
similar curves for coherent charged pions. In this case,
the events are
plotted versus visible energy as discussed in the text.}
\label{ncohfig}
\end{figure}
\end{document}