1: In this section I consider the Polyakov-loop model as introduced in
2: Sec. 1.4. As discussed, the effective potential for three
3: colours with explicit breaking of the $Z(3)$ symmetry is given by,
4: \begin{equation}
5: V(\ell) = - b_1 \frac{\ell + \ell^*}{2}
6: - \frac{b_2}{2} |\ell|^2
7: - \frac{b_3}{3}\frac{\ell^3 + \ell^*\!^3}{2}
8: + \frac{1}{4} \left( |\ell|^2 \right)^2,
9: \; \quad\quad(N_c=3,~m_\pi<\infty)~.
10: \label{e3_Q}
11: \end{equation}
12:
13: First, I consider the case where $b_1$ is very small, and take
14: the term linear in $\ell$ as a perturbation; then
15: the weakly first-order phase transition of the
16: pure gauge theory persists (in what follows, the critical temperature
17: in the pure gauge theory with $b_1=0$ will be denoted $T_c^*$).
18: The critical temperature is determined from
19: \be \label{b2Tc}
20: b_2(T_c) = -\frac{2}{9} b_3^2 \left(1+\frac{27}{2}\frac{b_1}{b_3^3}
21: \right) + {\cal O}(b_1^2)~.
22: \ee
23: The order parameter jumps at $T_c$,
24: from
25: \be \label{ellTc-}
26: \ell_0(T_c^-) = \frac{9}{2} \frac{b_1}{b_3^2} + {\cal O}(b_1^2)~,
27: \ee
28: to
29: \be \label{ellTc+}
30: \ell_0(T_c^+) = \frac{2}{3}b_3 - \frac{9}{2} \frac{b_1}{b_3^2}
31: + {\cal O}(b_1^2)~.
32: \ee
33: Note that numerically $\ell_0(T_c^-)$ could be much larger than $b_1$
34: if the phase transition in the pure gauge theory is weak and so the
35: correlation length $\xi=1/m_\ell$ near $T_c$ is large (i.e., if
36: $b_3$ is small), as indeed appears to be the case for $N_c=3$
37: colours~\cite{Kaczmarek:1999mm}. In other words, it could be that on the lattice
38: $\ell$ quickly develops a non-vanishing expectation value at $T_c^-$
39: already for rather large quark (or pion) masses, but this does not
40: automatically imply a large explicit symmetry breaking (see also
41: Fig.~\ref{Fig_ell} below).
42:
43: From Eq.~(\ref{b2Tc}) one can estimate the shift of $T_c$ induced by
44: letting $m_\pi<\infty$.
45: Writing the argument of $b_2$ in that equation as
46: $T_c^* + \Delta T_c$ and expanding to first-order in $\Delta T_c$ one
47: obtains
48: \bea \label{DelTc}
49: \frac{\Delta T_c}{T_c^*} &=& -3\; \frac{b_1}{b_3}~\left(
50: T\frac{\partial b_2}{\partial T}\right)^{-1}_{T=T_c^*} + {\cal O}(b_1^2)\nn\\
51: &=&-\frac{2}{3}~\ell_0(T_c^-)~b_3\left(
52: T\frac{\partial b_2}{\partial T}\right)^{-1}_{T=T_c^*}+ {\cal O}(b_1^2)~.
53: \eea
54: The shift in $T_c$ with decreasing pion mass is proportional to
55: the expectation value of the Polyakov-loop just below $T_c$; all other
56: factors on the right-hand side of Eq.~(\ref{DelTc}) do not depend on
57: $b_1$ or $m_\pi$.
58: Numerical values for $b_3$ and for $b_2(T)$ were obtained
59: in~\cite{Dumitru:2000in,Dumitru:2001bf,Scavenius:2001pa,Scavenius:2002ru}
60: by fitting the effective potential~(\ref{e3_N=3})
61: to the pressure and energy density of the pure gauge theory with three
62: colours; those are $b_3\approx0.9$ and $b_3^2/[T_c^* \partial
63: b_2(T_c^*)/\partial T]\approx 1$, to within 10\%. I therefore expect
64: that numerically $\Delta T_c/T_c^*$ is roughly equal to $\ell_0(T_c^-)$.
65:
66: The Eqs.~(\ref{ellTc-},\ref{ellTc+}) seem to indicate that
67: the discontinuity of $\ell_0$ at $T_c$ vanishes, i.e., that the phase
68: transition turns into a crossover, at a pion mass such that
69: $b_1(m_\pi)=2b_3^3/27$. However, one cannot really extend the ${\cal
70: O}(b_1)$ estimates to the endpoint of the line of first-order
71: transitions because it applies, near $T_c$, only if $-4b_2(T_c) \ll
72: b_3^2$, which translates into $b_1 \ll b_3^3/108$, see Eq.~(\ref{b2Tc}).
73: To find the endpoint of the line of first-order transitions I
74: solve numerically for the global minimum of~(\ref{e3_Q}) as a function
75: of $b_2$, for given $b_1$, see Fig.~\ref{Fig_ell} (left). The
76: numerical solution is ``exact'' and does not assume
77: that $b_1$ is small. I employ $b_3=0.9$ to properly account for the
78: small latent heat of the pure gauge
79: theory~\cite{Dumitru:2000in,Dumitru:2001bf,Dumitru:2001xa,Scavenius:2001pa,Dumitru:2001vc,Dumitru:2002cf,Scavenius:2002ru}.
80: Also, for $b_1=0$, this $b_3$ corresponds to $\ell_c=0.6$, which is close
81: to the expectation value of the renormalised (fundamental) loop for
82: the $N_c=3$ pure gauge theory~\cite{Kaczmarek:2002mc,Petreczky:2004pz,Zantow:2003uh,Dumitru:2003hp}.
83:
84: Clearly, for very small $b_1$ the order parameter $\ell_0$ jumps at
85: some $b_2^c\equiv b_2(T_c)$, i.e., the first-order phase
86: transition persists. (The abscissa is normalised by
87: $|b_2(T_c^*)| = 2b_3^2/9$.)
88: I find that the discontinuity vanishes at $b_1^c=0.026(1)$, so
89: there is no true phase transition for $b_1 > b_1^c$. Nevertheless,
90: I define $b_2^c$ even in the crossover regime via
91: the peak of $\partial\ell_0(b_2)/\partial b_2$.
92: The shift of $b_2^c$ with increasing $b_1$ can now be converted
93: into the shift of $T_c$ itself by expanding about $T_c^*$:
94: \be \label{DelTc2}
95: \frac{\Delta T_c}{T_c^*} = {\Delta b_2^c}~\left(
96: T\frac{\partial b_2}{\partial T}\right)^{-1}_{T=T_c^*}~,
97: \ee
98: as already discussed above. I also note that from Fig.~\ref{Fig_ell}
99: (left) the susceptibility for the Polyakov-loop at its maximum is
100: $\partial \ell_0/\partial b_2\simeq 3.5$, 2, 1.5 for $b_1=0.06$, 0.1,
101: and 0.126, respectively. That is, the crossover is rather sharp for
102: the values of $b_1$ shown in the figure.
103:
104: Explicit breaking of the Z(3) symmetry of the gauge theory
105: has previously been studied
106: in~\cite{Green:1983sd,Meisinger:1995qr,Alexandrou:1998wv}, and has been
107: identified as the essential factor in determining the endpoint of
108: deconfining phase transitions. Moreover, while the term $\sim b_1$
109: quickly washes out the transition, those studies showed that along the
110: line of first-order transitions the shift of $T_c$ (or, alternatively, of
111: the critical coupling $\beta_c$) is moderate, which agrees with
112: my findings. However, the numerical values for the critical ``external field"
113: at the endpoint obtained in~\cite{Meisinger:1995qr,Alexandrou:1998wv}
114: from actual
115: Monte-Carlo simulations can not be compared directly to my estimate for
116: $b_1^c$ because I work here with the renormalised (continuum-limit) loop,
117: not the bare loop.
118: Ref.~\cite{Karsch:2000kv} studied finite-temperature
119: QCD with $N_f=3$ flavours and various quark masses on the lattice (with
120: improved p4-action), and determined the critical (or crossover)
121: temperature as a
122: function of the pion mass. Using Eq.~(\ref{DelTc2}) one can
123: match $\Delta T_c/T_c^*$ to the data from~\cite{Karsch:2000kv} to
124: determine $b_1(m_\pi)$. In other words, I extract the
125: function $b_1(m_\pi)$ required to match the effective
126: Lagrangian~(\ref{e3_Q}) to $T_c(m_\pi)$ found on the lattice.
127: The result is shown in Fig.~\ref{Fig_ell} on the right. (Again, the pion mass
128: is normalised to the zero-temperature string tension in the pure gauge
129: theory, $\surd\sigma\simeq0.425$~GeV.)
130: \begin{figure}
131: \begin{center}
132: \includegraphics[height=6.2cm]{ell_b2}
133: \includegraphics[height=6.2cm]{b1_mpi}
134: \caption[The expectation value for the Polyakov-loop
135: and the explicit symmetry breaking term.]
136: {Left: The expectation value for the Polyakov-loop,
137: $\ell_0(b_2(T))$, for various values of the explicit
138: symmetry breaking coupling, $b_1$. All curves terminate at
139: $\ell_0=1\Leftrightarrow T=\infty$.
140: Right: $b_1$ as a function of $m_\pi$, obtained by matching to three-flavour
141: lattice data for $T_c(m_\pi)$. The solid line corresponds to an exponential
142: increase of $b_1$ with decreasing $m_\pi$, see text.
143: The broken horizontal line displays the
144: endpoint of the line of first-order phase transitions in terms of
145: $b_1$; the intersection with the $b_1(m_\pi)$ curve then gives the
146: corresponding pion mass.} \label{Fig_ell}
147: \end{center}
148: \end{figure}
149: Evidently, the $\approx33\%$ reduction of $T_c$ from $m_\pi=\infty$
150: (pure gauge theory) to $m_\pi/\surd\sigma\approx 1$ requires only
151: small explicit breaking of the $Z(3)$ symmetry for the Polyakov-loop
152: $\ell$: I find that $b_1<0.15$ even for the smallest pion
153: masses available on the lattice. This is due to the rather weak first-order
154: phase transition of the pure gauge theory with $N_c=3$ colours,
155: reflected by the strong dip of the string tension
156: in the confined phase near $T_c^-$ and of the Polyakov-loop screening
157: mass $m_\ell$ in the deconfined phase near $T_c^+$~\cite{Kaczmarek:1999mm};
158: cf.\
159: also the discussion in~\cite{Pisarski:2000eq,Pisarski:2002ji,
160: Dumitru:2000in,Dumitru:2001bf,Dumitru:2001xa,Dumitru:2001vc,Dumitru:2002cf}.
161:
162: Moreover, $b_1(m_\pi)$ appears to follow the
163: expected behaviour $\sim\exp(-m_\pi)$. The exponential fit shown by the
164: solid line corresponds to $b_1(m_\pi) = a \exp(-b \; m_\pi/\surd\sigma)$,
165: with $a=0.19$ and $b=0.47$~. Surprisingly, by naive extrapolation
166: one obtains a pretty small
167: explicit symmetry breaking even in the chiral limit,
168: $b_1\approx 0.2$.
169:
170: Finally, the endpoint of the line of
171: first-order transitions at $b_1^c=0.026$ (indicated by the dashed
172: horizontal line) intersects the curve $b_1(m_\pi)$ at
173: $m_\pi/\surd\sigma\approx 4.2$. For heavier pions the theory exhibits a
174: first-order deconfining phase transition, which then turns into a
175: crossover for $m_\pi\lton4.2\;\surd\sigma\approx 1.8$~GeV
176: ($\hat{=}\,m_q\approx 0.9$~GeV). According to my estimate, the
177: endpoint of the line of first-order transitions occurs at $\Delta
178: T_c/T_c^*\approx 12\% $, which is slightly less than a previous
179: (qualitative) estimate of 26\% from ref.~\cite{Ogilvie:1999if}.
180:
181: Of course, so far my analysis is restricted to pion masses
182: $m_\pi/\surd\sigma\gsim1$. On the other hand,
183: one might cross a chiral critical point for some pion
184: mass $m_\pi/\surd\sigma<1$~\cite{Gavin:1993yk}. Attempting a fit
185: with the model~(\ref{e3_Q}) beyond that point would then lead to
186: deviations from $b_1\sim\exp(-m_\pi)$.
187:
188:
189: