\documentclass[aps,prb,amssymb]{revtex4}
%\documentclass[preprint,showpacs,prl,aps,amssymb]{revtex4}
%\documentclass[twocolumn,prb,aps,amssymb,floatfix]{revtex4}
%\documentclass[twocolumn,showpacs,prl,aps,amsmath]{revtex4}
%\documentclass[aps,prl,twocolumn,amsmath,amssymb,
%superscriptaddress,floatfix,showpacs]{revtex4}

\usepackage{graphicx}% Include figure files
\usepackage{bm}% bold math

%\documentclass[aps,prb]{revtex4}


\def\ep {\epsilon}
\def\ek {\epsilon_k}
\def\e2 {\epsilon-\epsilon_k}
\def\be {\begin{equation}}
\def\ee {\end{equation}}
\def\bea {\begin{eqnarray}}
\def\eea {\end{eqnarray}}
\def\om {\omega}
\def\rs {\tau_{sp}^{-1}}

\begin{document}

\title{Scattering rate, transport and specific heat in a metal close
to a quantum critical point : emergence of a robust Fermi liquid picture ? }

\author{ George Kastrinakis}

\affiliation{Institute of Electronic Structure and Laser (IESL), 
Foundation for Research and Technology - Hellas (FORTH), 
P.O. Box 1527, Iraklio, Crete 71110, Greece$^*$}

\date{June 2005}

\begin{abstract} 

We calculate the low temperature one-particle scattering rate and the 
specific heat
in a weakly disordered metal close to a quantum critical point. To lowest
order in the fluctuation potential, we obtain typical Fermi-liquid
results proportional to $T^2$ and $T$ respectively, with prefactors 
which diverge as a power law of the control parameter upon approaching 
the critical point. The Kadowaki-Woods ratio 
is shown to be independent of the control parameter only for the
case of 3-D FM fluctuations. 
Our work is relevant for experiments on CeCoIn$_5$ and Sr$_3$Ru$_2$O$_7$.

\end{abstract}

\maketitle

\vspace{.3cm}

We calculate the one-particle scattering rate as a function of temperature
$T \rightarrow 0$.
We use the relation
\be
\text{Im} \Sigma^R(k,\ep) = \sum_{q} \int_{-\infty}^{\infty}d\om \;
\text{Im} G^R(q,\ep-\om) \text{Im} V^R(k-q,\om) \;
\{\coth(\om/2T) \; + \; \tanh((\ep-\om)/2T) \}\;\;,
\ee
to calculate the scattering rate, which equals twice Im$\Sigma^R$
- c.f. fig 1(a). This is
to lowest order in the magnetic fluctuation potential. 
Here 
\be
G(k,\ep)=\frac{1}{\ep-\ep_k + i \; \text{sgn}(\ep)/2\tau} \;\;
\ee
with $\tau$ being the impurity momentum relaxation time.

The dominant electron-electron interaction is assumed to be the
fluctuation propagator \cite{her,kim} (finite $q_0$ was considered in 
\cite{chu})
\be
V(q,\om)=\frac{g}{-i \om /(D q^2)+  (q-q_0)^2 \xi^2 + a} \;\;,
\ee
with $a$ measuring the distance from 
the quantum critical point (QCP). The control parameter $a$ 
depends on e.g. the magnetic field $H$, as in the systems of interest 
mentioned below like $a=h^s$ - typically $s={2/3}$ - $h=(H-H_c)/H_c$.
$q_0$=0 corresponds to ferromagnetic (FM) fluctuations.


Carrying-out this calculation yields in 3-D
\be
\text{Im} \Sigma^R(k_F,\ep_F,T) = const. T^2 f_3(a,q_0) \;\;,
\ee
with 
\bea
f_3(a,q_0)= \frac{1}{2 a \xi^2 q_0}\left\{ \frac{1}{q_{max}-q_0}
+\frac{1}{q_0} \right\} \;\;,\;\; q_0 = \text{finite} \;\;, \\
= \frac{1}{ a^2}   \;\;, \;\;  q_0 = 0 \;\;.
\eea


In 2-D we have 
\be
\text{Im} \Sigma^R(k_F,\ep_F,T) = const. T^2 f_2(a,q_0) \left\{
\ln\left(\frac{T_o}{T}\right)-1+\frac{3T}{T_o}
+O(T^2)\right\}\;\;,\;\; T_o = \frac{a^2 D'}{\xi^2} \;\;.
\ee
It turns out that $f_2(a,q_0) = f_3(a,q_0)$, so the prefactor dependence 
on $a$ is the same as in 3-D.


Taking the resistivity proportional to the scattering rate,
these results compare very favorably with data on CeCoIn$_5$ \cite{pag}
and Sr$_3$Ru$_2$O$_7$ 
\cite{gri}, where a typical Fermi-liquid (FL) 
$T$-dependence is seen in the resistivity $\rho(T)=\rho_0 + A(H) T^2$, with
$A(H)$ diverging around the transition, i.e. around $a=0$, as in our
result. In CeCoIn$_5$ it was found that $A(H) \propto a^{-2}$, as in the case
$q_0=$ above. The same 
result was obtained for the thermal conductivity \cite{pag}, 
which is easily understood
in the frame of our calculation: conduction electrons carry both charge
and heat, while interacting via $V(q,\om)$ (only {\em small} energy
transfer is involved with this $V(q,\om)$).

It should be possible to probe the scattering rate through angle-resolved
photoemission (ARPES) experiments.
 
We calculate the specific heat 
$C(T)= - T \partial^2 F(T)/\partial T^2 = \gamma T$ through 
the free energy $F(T)$ - c.f. fig 1(b). In 3-D we obtain for all $q_0$
\be
\gamma = const./(a \; \xi^2)\;\;.
\ee
In 2-D we have for all $q_0$
\be
\gamma = const./\xi^2 \;\;,
\ee
i.e. there is no diverging prefactor. 

The Kadowaki-Woods ratio is 
\bea
\frac{A(a)}{\gamma^2(a)} &=& c_1 \xi^4\;\;, \;\; 3-D\;,\; q_0 =0\;, \\
&=&  c_2\;  a \; \xi^2\;\;, \;\; 3-D\;,\; q_0 =\text{finite}\;, \\
&=& c_3 \xi^4/a^2 \;\;, \;\; 2-D\;,\; q_0 =0\;, \\
&=&  c_4 \xi^2/ a \;\;, \;\; 2-D\;,\; q_0 =\text{finite}\;, 
\eea
$c_{1-4}=$const.
It is independent of $a$ only in 3-D for $q_0=0$, while some simple
power law arises in the other cases considered. An $a$-independent ratio
was observed in CeCoIn$_5$ \cite{bia}, though in a more restricted range
of $H$ than the scaling of the coefficient $A(H)$.

We also consider an interaction with peaks 
at specific wavevectors $\vec{q}_{0i}$ ("AF" case)
\be
V_*(q,\om)=\sum_{i=1}^{n_d} 
\frac{g}{-i \om /(D q^2)+  (\vec{q}-\vec{q}_{0i})^2 \xi^2 + a} \;\;.
\ee
In 2-D for tetragonal symmetry $n_d=4$ and in 3-D for cubic symmetry $n_d=6$.
For small a, in 2-D and 3-D the potential $V_*(q,\om)$ 
gives the same scaling of the prefactors as for the case 3-D $q_0=$finite
above.


In all, a consistent Fermi liquid (FL) description seems to emerge from these
calculations. The renormalization of the fermions due to $V(q,\om)$ leads
back to the FL fixed point in a low-$T$ part of the phase diagram
- e.g. c.f. fig 1 of ref. \cite{pag}(case $n=2$). 
The effective mass of the fermions is not 
renormalized to first order in $V(q,\om)$ - c.f. the match between the one
particle scattering rate and the experimental resistivity, connected through
$\rho = m_*  \tau^{-1}/(n e^2)$.


\vspace{.3cm}
$^*$ e-mail : kast@iesl.forth.gr

\begin{thebibliography}{99}

\bibitem{her}
J. Hertz, Phys. Rev. B {\bf 14}, 1165 (1976).

\bibitem{kim}
Y.B. Kim and A.J. Millis, Phys. Rev. B {\bf 67}, 085102 (2003).

\bibitem{chu}
A.V. Chubukov, V.M. Galitski and V.M. Yakovenko, Phys. Rev. Lett. {\bf 94},
046404 (2005).

\bibitem{pag}
J. Paglione et al., cond-mat/0405157.

\bibitem{gri}
S.A. Grigera et al., Science {\bf 294}, 329 (2001).

\bibitem{bia}
A. Bianchi et al., Phys. Rev. Lett. {\bf 91}, 257001 (2003).

\end{thebibliography}


\begin{figure}[tb]
\begin{center}
  \centering

  \includegraphics[width=8truecm]{kw1.eps}
\vspace{.1cm}
\centerline{Fig. 1}
\caption{
Diagrams (a) for the self-energy and (b) for the free energy. The continuous
lines are the fermions and the dashed
line is the fluctuation mediated interaction $V(q,\om)$.}
%\label{fig:instability}
\end{center}
\end{figure}

\end{document}

