1: In this section, I discuss chiral symmetry restoration at nonzero
2: temperature, and in particular the dependence of the symmetry
3: restoration temperature on the pion mass.
4: For simplicity, I restrict myselfe here to the two-flavour case.
5: My emphasis is not on the order of the transition as the
6: strange quark mass is varied but rather on how the temperature at
7: which the transition occurs (be it either a true phase transition or
8: just a crossover) depends on the pion mass.
9: Such dependence arises from two effects. First, of course, due to
10: explicit symmetry breaking occurring when $m_\pi>0$. Second, due to
11: the ``indirect'' dependence of spontaneous symmetry breaking, i.e., of
12: the condensate $\langle\bar{q}q\rangle$ resp.\ $f_\pi$, on the pion
13: mass (through pion loops, see previous section).
14:
15: In the following I use the counter-term renormalization scheme as discussed in
16: section V.B in \cite{Lenaghan:1999si}. In this scheme
17: a mass renormalization scale $\mu$ is
18: introduced and the couplings then depend on that scale. However, choosing
19: \be
20: \mu_{ren}^2
21: =\exp\left[\frac{m_\sigma^2(\ln\,m_\sigma^2-1)-m_\pi^2(\ln\,m_\pi^2-1)}
22: {m_\sigma^2-m_\pi^2}\right]~,
23: \ee
24: the four-point coupling $\lambda(\mu)=\lambda_{\rm tree}$ retains its
25: tree-level (classical) value~\cite{Lenaghan:1999si}. In other words, this
26: renormalization prescription evolves the renormalization scale $\mu_{ren}$
27: in such a way as to keep $\lambda$ constant.
28:
29: Explicitly, this leads to the following expressions for the
30: couplings~\cite{Lenaghan:1999si}:
31: \be
32: \lambda = \frac{1}{2} \frac{m_\sigma^2-m_\pi^2}{f_\pi^2},\quad
33: H=f_\pi\left(m_\sigma^2-2\lambda f_\pi^2 \right),\quad
34: m^2 = -\frac{1}{2} \left( m_\sigma^2-3 m_\pi^2\right)
35: -6\lambda Q_\mu(m_\pi),
36: \ee
37: where
38: \be
39: Q_\mu(M)\equiv \frac{1}{(4\pi)^2}\left[M^2\ln\frac{M^2}{\mu_{ren}^2}
40: -M^2+\mu_{ren}^2\right].
41: \ee
42: These equations determine the couplings in vacuum in terms of $m_\pi$,
43: $f_\pi$ and $m_\sigma$. The dependence of $f_\pi$ and $m_\pi$ on
44: the quark mass is
45: taken from the data of ref.~\cite{Chiu:2003iw}.
46: Roughly, for $m_\pi:0.4~{\rm GeV}\to 1$~GeV, $f_\pi$
47: increases by about 50~\%, leading to an increase of the explicit
48: symmetry breaking term $H$ by a factor of 10.
49: I also require the dependence of $m_\sigma$
50: on $m_\pi$, which I take from a computation with standard
51: Wilson fermions~\cite{Kunihiro:2003yj}. Those authors
52: find that $m_\sigma$ is essentially a linear function of
53: $m_\pi^2$. I checked how my results in Fig.~\ref{Fig_sigma} depend
54: on this assumption by using, alternatively, a linear dependence
55: $m_\sigma= m_\pi+{\rm const.}$, with
56: $m_\sigma=0.6$~GeV for $m_\pi=0.14$~GeV. I found essentially the same
57: dependence of $T_c$ on $m_\pi$.
58:
59: At nonzero temperature, one uses the
60: effective potential for composite operators, as discussed
61: in Sec.~\ref{seccjt}, to determine
62: the masses and the scalar condensate in the Hartree-Fock approximation.
63: This approximation is defined by only taking into account the
64: double-bubble diagrams shown in Figs.~\ref{paper1} d and e in the effective
65: potential $V$. The gap equations for
66: the condensate and the masses are given by
67: minimisation of this potential.
68: The expectation values of the one- and two-point functions in
69: the absence of external sources, $\sigma$ and
70: ${\cal S}$, ${\cal P}$,
71: are determined from the stationary points of $V$,
72: \bea
73: \left.\frac{\delta V}{\delta \bar\sigma}\right
74: |_{\bar\sigma=\sigma,\bar{S}={\cal S},\bar{P}={\cal P}}=0,
75: \left.\frac{\delta V}{\delta \bar{S}}\right
76: |_{\bar\sigma=\sigma,\bar{S}={\cal S},\bar{P}={\cal P}}=0,
77: \left.\frac{\delta V}{\delta \bar{P}}\right
78: |_{\bar\sigma=\sigma,\bar{S}={\cal S},\bar{P}={\cal P}}=0,
79: \eea
80: leading to the gap equations for the chiral condensate $\sigma$, and the
81: effective masses of the $\sigma$-meson and the pions,
82: $M_\sigma$ and $M_\pi$,
83: \bse\bea
84: H &=& \sigma \left[M_\sigma^2-2\lambda \sigma^2 \right], \label{eq9}\\
85: M_\sigma^2 &=& m^2+3 \lambda \left\{\sigma^2 +
86: [Q_T(M_\sigma)+Q_\mu(M_\sigma)]+
87: [Q_T(M_\pi)+Q_\mu(M_\pi)]\right\}, \hspace{.5cm} \label{eq10}\\
88: M_\pi^2 &=& m^2+ \lambda \left\{\sigma^2 +
89: [Q_T(M_\sigma)+Q_\mu(M_\sigma)]
90: +5[Q_T(M_\pi)+Q_\mu(M_\pi)]\right\}, \label{eq11}
91: \eea\ese
92: where the nonzero-temperature contribution of the tadpole diagram
93: is given by
94: \be
95: Q_T(m) \equiv
96: \frac{1}{2\pi^2}\int_0^\infty d q[\omega (q)]^{-1}
97: f[\omega(q)],
98: \ee
99: where $f(\omega)\equiv 1/[\exp(\omega/T)-1]$ is the Bose-Einstein
100: distribution function and $\omega(q)\equiv\sqrt{q^2+m^2}$ the
101: quasiparticle energy.
102: The self-consistent solution of the above
103: gap equations for a given vacuum pion mass determines
104: the temperature dependence of the scalar condensate as the order
105: parameter of chiral symmetry restoration. For explicitly broken chiral
106: symmetry, $H>0$, the transition in this approach is a crossover. I
107: define the crossover temperature $T_c$ by the peak of
108: $\partial \sigma / \partial T$.
109: \begin{figure}
110: \begin{center}
111: \includegraphics[height=6.3cm]{o4Tc}
112: \includegraphics[height=6.3cm]{o4sig}
113: \caption[The crossover temperature and the scalar condensate.]
114: {Left: The crossover temperature $T_c$ as a function of the
115: (vacuum) pion mass as obtained from the linear sigma model
116: with $O(4)$ symmetry in comparison to lattice
117: data~\protect\cite{Karsch:2000kv} for two and three flavours.
118: The scale for both $T_c$ and $m_\pi$ is set by the zero-temperature
119: string tension in the pure gauge theory, $\surd\sigma\simeq0.425$~GeV.
120: Right: The scalar condensate,
121: $\langle\sigma\rangle$, as a function of temperature for
122: various pion masses.}
123: \label{Fig_sigma}
124: \end{center}
125: \end{figure}
126: The dependence of $T_c$ on $m_\pi$ is depicted in Fig.~\ref{Fig_sigma}
127: (left), where I have also shown lattice results obtained with two and
128: three degenerate quark flavours, respectively~\cite{Karsch:2000kv} (the
129: $N_f=2$ data with standard action, the $N_f=3$ data with improved
130: p4-action). Driven by the increase of both $f_\pi$ and $H$ with
131: $m_\pi$, the linear sigma model predicts a rather rapid rise of $T_c$
132: with the pion mass, as compared to the data which is nearly flat on the
133: scale of the figure. While lattice data indicate a rather weak
134: dependence of $T_c$ on the quark mass (see also ref.~\cite{Bernard:1996cs}),
135: models with spontaneous symmetry breaking in the vacuum
136: naturally predict a rather steep rise of $T_c$
137: with the VEV $\sigma_{\rm vac}=f_\pi$, which itself
138: increases with the quark (or pion) mass. My findings here are in
139: qualitative agreement with those from ref.~\cite{Berges:1997eu} who employed
140: nonperturbative flow equations to compute the effective potential
141: for two-flavour QCD within the linear sigma model. They also find a
142: steeper slope of $T_c(m_\pi)$ than indicated by the lattice, even though their
143: analysis appears to predict a somewhat weaker increase of $f_\pi$
144: with $m_\pi$ than the data of~\cite{Chiu:2003iw}, which I employ here.
145:
146: Fig.~\ref{Fig_sigma} also shows the temperature dependence of the
147: $\sigma$ condensate (right). With $m_\sigma$ a linear function of
148: $m_\pi^2$~\cite{Kunihiro:2003yj}, the width of the crossover is
149: approximately independent of
150: the pion mass for 0.4~GeV$\lton m_\pi\lton1$~GeV, while I found
151: considerable broadening when $m_\sigma$ is linear in $m_\pi$ (not shown).
152: The chiral susceptibility $\partial \sigma/\partial T$ at
153: its maximum is $\approx 0.25$, i.e., the crossover is in fact
154: quite broad for the range of $m_\pi$ considered. Since this is
155: at variance with lattice data on QCD thermodynamics (pressure
156: and energy density as functions of temperature, see e.g.\ the review
157: in~\cite{Laermann:2003cv}), one might argue that the crossover is in
158: fact not driven by the order parameter field but by heavier degrees of
159: freedom~\cite{Karsch:2003zq,Karsch:2000kv,Gerber:1988tt}. Such degrees
160: of freedom could reduce the pion-mass dependence of
161: the transition substantially: using
162: three-loop chiral perturbation theory (i.e., the non-linear model),
163: Gerber and Leutwyler find~\cite{Gerber:1988tt} that
164: $T_c$ increases rapidly from
165: $\approx 190$~MeV in the chiral limit
166: (using their set of couplings) to $\approx 240$~MeV for
167: physical pion mass. However, when heavy states are included (in the
168: dilute gas approximation), then $T_c$ increases less rapidly, from
169: $\approx 170$~MeV in the chiral limit to $\approx 190$~MeV for
170: physical pion mass.
171:
172: \vfill