hep-lat0612001/3f.tex
1: \documentclass[aps,showpacs,preprintnumbers,amsmath,nofootinbib,amssymb]{revtex4}
2: \usepackage[dvips]{epsfig}
3: 
4: \newcommand{\nc}[1]{\newcommand{#1}}
5: 
6: %*** text ***
7: 
8: \nc{\its}[1]{\itshape #1 \upshape}
9: \nc{\mc}[3]{\multicolumn{#1}{#2}{#3}}
10: 
11: %*** graphics ***
12: 
13: \nc{\bc}{\begin{center}}
14: \nc{\ec}{\end{center}}
15: \nc{\ig}[1]{\bc \includegraphics{#1} \ec}
16: 
17: %*** mathematical mode ***
18: 
19: \nc{\bo}[1]{\mbox{\boldmath \( #1 \! \! \)  \unboldmath}}
20: \nc{\be}{\begin{eqnarray}}
21: \nc{\ee}{\end{eqnarray}}
22: \nc{\bew}{\begin{eqnarray*}}
23: \nc{\eew}{\end{eqnarray*}}
24: \nc{\bs}{\begin{subeqnarray}}   %*** requires subeqnarray.sty
25: \nc{\es}{\end{subeqnarray}}     %*** requires subeqnarray.sty
26: \nc{\nnn}{\nonumber \\}
27: \nc{\f}[2]{\frac{#1}{#2}}
28: \nc{\td}[2]{\f{d #1}{d #2}}
29: \nc{\pd}[2]{\f{\partial #1}{\partial #2}}
30: \nc{\suli}{\sum\limits}
31: \nc{\proli}{\prod\limits}
32: \nc{\ili}{\int\limits}
33: \nc{\sr}[2]{\stackrel{#1}{#2}}
34: \nc{\dps}{\displaystyle}
35: \nc{\ket}[1]{\left| #1 \right>}
36: \nc{\bra}[1]{\left< #1 \right|}
37: \nc{\bracket}[2]{\left< #1 \right| \left. \! #2 \right>}
38: \nc{\norm}[1]{\left\| #1 \right\|}
39: \nc{\lndm}[1]{\pd{^{#1} \ln{\det{M}}}{\mu^{#1}}}
40: \nc{\pdmm}[1]{M^{-1} \pd{^{#1} M}{\mu^{#1}}}
41: \nc{\pdm}{M^{-1}\pd{M}{\mu}}
42: \nc{\trac}[1]{\mbox{Tr}\left(#1\right)}
43: \nc{\hm}{\hat{m}}
44: 
45: \def\lsim{\raise0.3ex\hbox{$<$\kern-0.75em\raise-1.1ex\hbox{$\sim$}}}
46: \def\gsim{\raise0.3ex\hbox{$>$\kern-0.75em\raise-1.1ex\hbox{$\sim$}}}
47: 
48: \epsfxsize=5.8cm
49: 
50: \begin{document}
51: 
52: \title{Study of the finite temperature transition in 3-flavor QCD using the R and RHMC algorithms}
53: 
54: \author{M. Cheng$^{\rm a}$, N. H. Christ $^{\rm a}$, M.A. Clark$^{\rm b}$, 
55: J. van der Heide$^{\rm d}$,
56: C. Jung$^{\rm c}$, F. Karsch$^{\rm c,d}$, O. Kaczmarek$^{\rm d}$,\\ 
57: E. Laermann$^{\rm d}$, R. D. Mawhinney$^{\rm a}$, C. Miao$^{\rm d}$,
58: P. Petreczky$^{\rm c,e}$, K. Petrov$^{\rm f}$,
59: C. Schmidt$^{\rm c}$, W. Soeldner$^{\rm c,d}$, and T. Umeda$^{\rm c}$
60: }
61: 
62: %\address{
63: \affiliation{
64: $^{\rm a}$ Physics Department, Columbia University, New York, NY 10027, USA\\
65: $^{\rm b}$ Center for Computational Science, Boston University, Boston, MA 02215\\
66: $^{\rm c}$Physics Department, Brookhaven National Laboratory, 
67: Upton, NY 11973, USA \\
68: $^{\rm d}$Fakult\"at f\"ur Physik, Universit\"at Bielefeld, D-33615 Bielefeld,
69: Germany\\
70: $^{\rm e}$ RIKEN-BNL Research Center, Brookhaven National Laboratory, 
71: Upton, NY 11973, USA \\
72: $^{\rm f}$Niels Bohr Institute, University of Copenhagen, 
73: Blegdamsvej 17, DK-2100 Copenhagen, Denmark\\
74: }
75: 
76: \date{\today}
77: \preprint{BNL-NT-06/45}
78: \preprint{BI-TP 2006/40}
79: \preprint{CU-TP-1163}
80: 
81: \begin{abstract}
82: We study the finite temperature transition in  
83: QCD with three flavors of equal masses using the R and RHMC algorithm on lattices
84: with temporal extent $N_{\tau}=4$ and $6$.
85: For the transition temperature in the continuum limit we find $r_0 T_c=0.429(8)$ for the 
86: light pseudo-scalar mass
87: corresponding to the end point of the 1st order transition region.
88: When comparing the results obtained with the R and RHMC algorithms for
89: p4fat3 action we see no significant step-size errors down to a lightest
90: pseudo-scalar mass of $m_{ps} r_0=0.4$. 
91: \end{abstract}
92: 
93: \pacs{11.15.Ha, 11.10.Wx, 12.38Gc, 12.38.Mh}
94: 
95: \maketitle
96: 
97: \section{Introduction}
98: \label{intro}
99: Lattice QCD has established the existence of a transition from hadron gas to a 
100: new state of strongly interacting matter where quarks and gluons are no longer confined
101: inside hadrons and which is usually called the quark gluon plasma \cite{cargese,reviews}. 
102: The nature of this 
103: transition depends on the quark content and quark masses. For infinite or very large quark
104: masses the transition is a 1st order deconfining transition. In the opposite case of zero
105: quark masses one may have a 2nd order chiral phase transition for 2 flavors or
106: a 1st order chiral phase transition for 3 flavors. For intermediate masses the transition is just
107: a rapid crossover, meaning that thermodynamic quantities change very rapidly in a narrow
108: temperature interval. The boundary of the 1st order transition region of 3 flavor QCD as a function of mass 
109: has been studied using improved (p4) \cite{karschlat03} and standard staggered actions
110: \cite{christian,norman}. There is a significant discrepancy regarding the value of the quark masses, or
111: equivalently the value of the pseudo-scalar meson masses where the transition changes
112: from crossover to 1st order. With an improved action it was found that the 1st order transition ends
113: for a pseudo-scalar meson mass of about $70$ MeV, while with the standard action it ends for
114: pseudo-scalar meson masses of about $190$ MeV \cite{christian,phil2} or larger \cite{norman}.
115: 
116: It has been observed that the pressure and energy density normalized by its ideal gas
117: value shows almost the same behavior as a function of $T/T_c$ for SU(3) pure gauge theory,
118: 2 flavor, 2+1 flavor and 3 flavor QCD \cite{cargese}. Thus  flavor and quark mass dependence
119: of these quantities in the first approximation, is determined by flavor and quark mass dependence of 
120: transition temperature $T_c$. Therefore it is very instructive to
121: study the flavor dependence of the transition temperature. Such a study has been 
122: performed in Ref. \cite{karsch01} on lattices with temporal extent $N_{\tau}=4$ using the
123: improved staggered p4 action. 
124: 
125: 
126: In the past most simulations with 2 and 3 flavors of staggered fermions have been done using the
127: hybrid molecular dynamics R (HMDR) algorithm \cite{hybrid}, often called simply the R-algorithm. It has finite step-size errors
128: of ${\cal O}(dt^2)$ where a step-size $dt$ is used in  the molecular dynamics evolution. Recently the rational hybrid Monte Carlo (RHMC) algorithm has been
129: invented which allows simulations of  theories with fractional powers of the fermion determinant, for example
130: 2 and 3 flavors of staggered fermions, without finite step-size errors \cite{rhmc}. Therefore the most recent
131: thermodynamics studies use the RHMC algorithm \cite{aoki06,philipsen06,SK06,us}.
132: It has been observed in Ref. \cite{philipsen06} that the use of the exact RHMC algorithm
133: reduces the value of the critical quark mass where the transition turns to 1st order by $25 \%$ in the
134: case of the standard staggered action.
135: 
136: 
137: 
138: The purpose of this paper is twofold. First we would like to study the transition in 3 flavor QCD,
139: extending the previous studies to smaller quark masses and smaller lattice spacings using the 
140: improved p4 staggered fermion action. Second we would like to compare the R-algorithm with the
141: new RHMC algorithm. 
142: The rest of the paper is organized as follows. In section II
143: we discuss the calculational set-up. In section III we show our results for finite temperature calculations.
144: Section IV discusses the zero temperature simulations needed to set the scale and the transition
145: temperature in physical units. In section V we present a comparison of the R and RHMC algorithms.
146: Finally, section VI contains our conclusions. A technical discussion of different fat link actions is
147: given in  Appendix A. In Appendix B we discuss the eigenvalues and eigenvectors of the p4 action in the free field limit.
148: 
149: 
150: 
151: 
152: \section{Lattice Formulation and setup}
153: \label{setup}
154: 
155: Most of the simulations discussed in this paper were done using the 
156: p4fat3 action, which is also simply called the p4 action \cite{karsch01}.
157: To improve the rotational symmetry, which is violated on the lattice, bent 3-link terms are
158: added to the 1-link term of the standard staggered action \cite{Heller}. 
159: Properly chosen coefficients of the 1-link and the 3-link terms  can eliminate,
160: at tree level, the  ${\cal O}(a^2)$ errors in the dispersion relation for staggered fermions \cite{Heller}.
161: The violation of flavor symmetry which is present in the staggered fermion formulation 
162: can be significantly reduced by replacing the normal link in the 1-link term by a fat link
163: which is a sum a of normal link and 3-link staples \cite{tb97}. This 3-link fattening is the origin of the name p4fat3.
164: Though this type of fat link action is a big improvement over the standard staggered
165: fermion action, further improvement of the flavor symmetry can be obtained by adding 5-
166: and 7-link staples \cite{kostas}. In fact one can eliminate the effect of flavor symmetry breaking at
167: order ${\cal O}(g^2 a^2)$ using a suitable combination of 3-, 5- and 7- link staples leading to
168: what is called the fat7 action \cite{kostas}.  We have also done calculations with the p4
169: action with fat7 fat links. Unfortunately it turns out that on the  coarse lattices used in our study of  3 flavor QCD
170: thermodynamics this action has some undesirable features. It leads to the occurrence of a bulk transition, which
171: we will discuss in more detail in Appendix A.
172: 
173: To study staggered fermions with less than four  flavors,  we use the rooting
174: procedure, i.e. each fermion flavor is represented by  $({\rm det} \mathcal{M})^{1/4}$, where $\mathcal{M}$ is the staggered fermion Dirac operator. 
175: For a recent discussion of this procedure see Ref. \cite{sharpe}.
176: Most of our simulations have been done using the standard R-algorithm \cite{hybrid}. As in Ref. \cite{karsch01}
177: the step-size of the molecular dynamics evolution was set to $dt=m/2.5$ for the staggered quark mass $m$.
178: 
179: 
180: We also performed calculations with the RHMC algorithm. In this algorithm an optimal rational
181: approximation is used to evaluate the fractional power of the determinant \cite{rhmc}.  
182: More precisely one finds the optimal approximation for $(\mathcal{M}^{\dagger} \mathcal{M})^{\nu}$ where $\mathcal{M}=2 m +D$ is
183: the usual staggered fermion matrix and for the three flavors $\nu=3/8$. Using 
184: sufficiently high order polynomials   the rational approximation can be made arbitrarily precise for
185: the given spectral range of the fermion operator. 
186: For the standard staggered fermion action the spectral range of $\mathcal{M}^{\dagger} \mathcal{M}$ is well 
187: known, the smallest eigenvalue of this matrix is $4 m^2$. The largest eigenvalue can be estimated in the free field limit to
188: be  $\lambda^2_{max}=16+4 m^2$.
189: For the case of the p4 staggered action the smallest eigenvalue is the same, while the
190: largest eigenvalue in the free case is $\lambda_{max}^2=50/9+4 m^2$. In appendix B we give
191: the derivation of this result.
192: It turns out that
193: in all our simulations the largest eigenvalue was smaller than 5.0, so we choose 
194: $\lambda_{max}^2=5.0$
195: as the upper limit on it. 
196: For the range of the quark masses studied by us, which include quark masses
197: as light as 1/20th of the strange quark mass, it is sufficient to use polynomials 
198: of degree 12 to achieve
199: machine precision with the rational approximation. 
200: Therefore in the Metropolis accept/reject step we used
201: polynomials of degree 12. In the molecular dynamics evolution we used a less 
202: stringent approximation of
203: the determinant since any errors in the evolution, including the $dt^2$ 
204: step-size errors, are eliminated by the
205: accept/reject step. It has been found that in most cases it is sufficient to use 
206: polynomials of degree 5-6
207: without compromising the acceptance rate. 
208: Furthermore, the stopping criteria for the conjugate gradient inversions can be relaxed
209: to $10^{-5}$ in the molecular dynamics evolution without significant 
210: effect on the acceptance rate. In the Monte-Carlo accept/reject step we typically use  $10^{-12}$ for
211: the conjugate gradient stopping criteria. The length of the trajectory was $\tau_{MD}=0.5$
212: in units of molecular dynamics time.
213: 
214: 
215: The gauge fields and quarks contribute different amounts to the force in 
216: the molecular dynamics evolution.
217: The contribution of the gauge fields is larger than that of quarks. On the other hand   
218: the cost of the evaluation of the force coming from the gauge fields is small. 
219: The opposite is true
220: for the part of the force coming from the fermions. Therefore it is reasonable to 
221: integrate the gauge force on finer molecular dynamics time scale than the fermion force \cite{sextwein}.
222: In our simulations we typically used 10 gauge field updates per fermion update. The 3 flavors
223: of staggered fermions are simulated as 1+1+1, i.e. we used a factor $({\rm det} \mathcal{M}^{\dagger} \mathcal{M})^{1/8}$  
224: for each fermion flavor. Although this increases the number of inversions, the reduced force allows
225: for a larger step-size $d t$ for the same acceptance rate. The step-size $dt$ was chosen such that 
226: the acceptance is about $70 \%$. To achieve this, the stepsize was typically of the order
227: of the strange quark mass used in our 2+1 flavor study \cite{us}.
228: 
229: 
230: 
231: \begin{table}[t]
232: %\vspace*{-0.5cm}
233: \begin{center}
234: \vspace{0.3cm}
235: \begin{tabular}{|c|r|c|c|c|}
236: \hline
237: $N_\tau$ &   $m$     & $N_\sigma~~$     &\# $\beta$ values & max. no. of traj. \\
238: \hline
239: 4             & 0.100      & 16                     & 4                         &  42000 \\
240: ~             & 0.050      & 8,~16                & ~8,~14                &  4130,~ 2650\\
241: ~             & 0.025      & 8,~16                & ~8,~~9                &  6250, ~8650 \\
242: ~             & 0.010      & 8,~16                & ~9,~~6                &  2460, ~4660 \\
243: ~             & 0.005      & 12,~16              & ~11,~8                &  1360,~3000  \\
244: \hline
245: 6            & 0.100       & 16                       &  16                    & 6000             \\
246: ~            & 0.050       & 16                       &  14                    & 7900             \\
247: ~            & 0.020       & 16                       &  10                    & 16000           \\
248: ~            & 0.010       & 16                       &   9                     & 10900           \\
249: \hline
250: \end{tabular}
251: \end{center}
252: \caption{
253: Parameters of the numerical simulations
254: }
255: \label{tab:parameter}
256: \end{table}
257: 
258: 
259: \section{Finite temperature simulations}
260: 
261: Our studies of the QCD transition at finite temperature have been
262: performed on lattices of size $N_\sigma^3\times N_\tau$. The 
263: lattice spacing, $a$, relates the spatial ($N_\sigma$) and temporal 
264: ($N_\tau$) size of the lattice to the physical volume 
265: $V= (N_\sigma a)^3$ and temperature $T=1/N_\tau a$, respectively.
266: The lattice spacing, and thus the temperature, is controlled by the
267: gauge coupling, $\beta = 6/g^2$, as well as the bare quark masses.
268: The parameters of our finite temperature simulations are given in Table \ref{tab:parameter}.
269: We extended the results of Ref. \cite{karsch01} in two respects.
270: Compared to Ref. \cite{karsch01} we have added a smaller mass value $m=0.005$ for $N_{\tau}=4$ and
271: extended the runs at larger quark masses to achieve a better statistical accuracy.
272: In addition we have studied the finite temperature transition on $N_{\tau}=6$ 
273: lattices for two values of the quark mass $m$. All the results presented in this section have been
274: obtained with the R algorithm.
275: 
276: 
277: As mentioned in the introduction, in 3 flavor QCD for small quark mass we have a 1st order phase transition
278: which turns into rapid crossover at the quark mass corresponding to the light pseudo-scalar mass
279: of about $70$ MeV \cite{christian}.
280: The transition is signaled by a 
281: rapid change in bulk thermodynamic observables (energy density, pressure)
282: as well as in the chiral condensates and the Polyakov loop expectation value,
283: \begin{eqnarray}
284: \frac{\langle \bar{\psi}\psi \rangle}{T^3} &=& \frac{1}{3} \frac{1}{VT^2}
285: \frac{\partial \ln Z}{\partial m}
286: = \frac{N_\tau^2}{4 N_\sigma^{3}} 
287: \left\langle {\rm Tr}\; M^{-1}(m)\right\rangle
288: \label{orderparameter1}  
289: , \\
290: \langle L\rangle &=& \left\langle \frac{1}{3N_\sigma^{3}} 
291: {\rm Tr}\sum_{\bf x} 
292: \prod_{x_0 = 1}^{N_\tau} U_{(x_0, {\bf x}),\hat{0}} \right\rangle \; , 
293: \label{orderparameter}
294: \end{eqnarray}
295: which are order parameters for a true phase transition in the zero 
296: and infinite quark mass limit, respectively. 
297: Note that we have defined the chiral condensate per flavor
298: degree of freedom, hence the factor 1/3 in Eq.  (\ref{orderparameter1}). 
299: 
300: We use  the Polyakov loop susceptibility as well as the disconnected part 
301: of the chiral susceptibility to locate the transition temperature.
302: \begin{eqnarray}
303: \chi_L &\equiv& N_\sigma^{3}  \left( 
304: \langle  L^2 \rangle - \langle L \rangle^2 \right) \; ,
305: \label{sus_L}\\
306: \frac{\chi_q}{T^2} &\equiv& \frac{N_\tau}{16N_\sigma^{3}} \left( 
307: \left\langle \left( {\rm Tr}\; M^{-1}(m)\right)^2 \right\rangle -
308: \left\langle {\rm Tr\;} M^{-1}(m)\right\rangle^2\right) 
309: . \label{sus_chi}
310: \label{sus_m}
311: \end{eqnarray}
312: We calculate the value of the Polyakov loop at the end of every trajectory.
313: For each tenth trajectory we calculate the value of $\bar \psi \psi$ and $\chi_q$ using ten Gaussian
314: random vectors.
315: In Fig. \ref{fig:suscnt4} we show the disconnected part of the chiral susceptibility
316: calculated on our $N_{\tau}=4$ lattice with different spatial volumes at different
317: quark masses.
318: In Fig. \ref{fig:suscnt6} we show the Polyakov loop $\chi_L$ and chiral $\chi_q$ susceptibilities
319: calculated on $16^3 \times 6$ lattice. 
320: The location of peaks in the susceptibilities
321: has been determined using Ferrenberg-Swedsen re-weighting for several values $\beta$
322: in the vicinity of the transition.  Errors on the peak location have been obtained from
323: a jackknife analysis where Ferrenberg-Swedsen re-weighting has been performed on different
324: sub-samples. The resulting pseudo-critical couplings are shown in Table \ref{tab:betac}.
325: In finite volume the pseudo-critical couplings $\beta_c$ determined from the Polyakov loop
326: correlator and chiral susceptibility are generally different. In the case of the crossover
327: this difference can persist even in the infinite volume limit. From  Table \ref{tab:betac} we see
328: that in most cases the two pseudo-critical couplings are identical within statistical errors even for
329: small volume $8^3 \times 4$. 
330: The cases where this difference is the largest are the cases where $\beta_{c,q}$ has large
331: statistical errors. For example for the $16^3\times 6$ lattice and $m=0.05$ we find
332: $\beta_{c,L}-\beta_{c,q}=.0113(328)$.
333: Therefore we have also calculated the weighted average of  $\beta_{c,L}$ and $\beta_{c,q}$
334: which is shown in the last column of Table \ref{tab:betac} together with the corresponding
335: error. This error is calculated from the statistical errors and the difference between the central
336: values added quadratically. 
337: The difference in pseudo-critical couplings determined on $8^3 \times 4$ and
338: $16^3 \times 4$ lattices is typically small, indicating small finite volume effects.
339: As a general tendency the pseudo-critical coupling $\beta_c$ shifts toward smaller values
340: with increasing volume.  
341: \begin{figure}
342: \includegraphics[width=7cm]{V_CS_p4fat3_4_m0.005.eps}
343: \includegraphics[width=7cm]{V_CS_p4fat3_4_m0.010.eps}
344: \includegraphics[width=7cm]{V_CS_p4fat3_4_m0.025.eps}
345: \includegraphics[width=7cm]{V_CS_p4fat3_4_m0.050.eps}
346: \caption{The disconnected part of the chiral susceptibility calculated 
347: on $N_{\tau}=4$ lattices at different quark masses.}
348: \label{fig:suscnt4}
349: \end{figure}
350: In agreement with earlier calculations we find that the position of
351: peaks in $\chi_q$ and $\chi_L$ show only little volume dependence and that
352: the peak height changes only little, although the maxima become somewhat more
353: pronounced on the larger lattices. This is consistent with the transition being
354: a crossover rather than a true phase transition in the infinite volume limit
355: for the range of quark masses explored by us.
356: \begin{figure}
357: \includegraphics[width=8cm]{M_CS_p4fat3_06_3f.eps}
358: \includegraphics[width=8cm]{M_LS_p4fat3_06_3f.eps}
359: \caption{The disconnected part of the chiral susceptibility (left) and the
360: Polyakov loop susceptibility (right) calculated on $16^3 \times 6$ lattices.}
361: \label{fig:suscnt6}
362: \end{figure}
363: \begin{table}
364: \begin{tabular}{|c|c|c|c|c|c|}
365: \hline
366: $N_{\tau}$   &   $m$     & $N_{\sigma}$  &   $\beta_{c,L}$ [from $\chi_L$]   &    $\beta_{c,q}$ [from $\chi_q$]   & $\beta_c$ [averaged] \\
367: \hline
368:   4                &   0.100   & 16                    &   3.4800(27)                               &    3.4804(24)                                &  3.4802(18)   \\
369:                     &   0.050   & 16                    &   3.3884(32)                               &    3.3862(47)                                &  3.3877(34)   \\
370:                     &               & 8                      &   3.4018(35)                               &    3.3930(201)                              &  3.4015(94) \\
371:                     &   0.025   & 8                      &   3.3294(27)                               &    3.3270(28)                                &   3.3283(31)   \\
372:                     &   0.010   & 16                    &   3.2781(7)                                 &    3.2781(4)                                  &   3.2781(3)     \\
373:                     &               & 8                      &   3.2858(71)                               &    3.2820(61)                                &   3.2836(60) \\
374:                     &   0.005   & 16                    &   3.2656(13)                               &    3.2678(12)                                &   3.2667(24)  \\
375:                     &               & 12                    &   3.2659(13)                               &    3.2653(12)                                &   3.2656(10)  \\
376: \hline
377: 6                  &   0.200   & 16                     &  3.8495(11)                               &    3.9015(279)                              &    3.8495(520) \\
378:                     &   0.100   & 16                     &  3.6632(55)                               &    3.6855(105)                              &    3.6680(228) \\
379:                     &   0.050   & 16                     &  3.6076(24)                               &    3.6189(328)                              &    3.6077(115) \\
380:                     &   0.020   & 16                     &  3.4800(110)                             &    3.4800(80)                                &    3.4800(65)  \\
381:                     &   0.010   & 16                     &  3.4518(50)                               &    3.4510(83)                                &    3.4516(44) \\
382: \hline
383: \end{tabular}
384: \caption{Critical couplings determined from the location of peaks in the Polyakov loop susceptibility as well
385: as in the disconnected parts of the chiral susceptibilities. The last column gives the average of $\beta_{c,L}$
386: and $\beta_{c,q}$ with combined statistical and systematic errors.}
387: \label{tab:betac}
388: \end{table}
389: 
390: 
391: 
392: 
393: \section{Zero temperature calculations and the transition temperature}
394: 
395: In order to determine the transition temperature in units of some physical 
396: quantity we performed
397: zero temperature calculations on $16^3 \times 32$ lattices in the vicinity of 
398: the pseudo-critical 
399: coupling $\beta_c$. 
400: The parameters of these calculations together with the accumulated statistics 
401: are summarized in Table \ref{tab:T=0}.
402: We have calculated the static quark potential
403: and meson correlators on each 10th trajectory generated.
404: 
405: The static potential has been calculated using  the ratios of the 
406: Wilson loops at two neighboring time-slices and extrapolating them to infinite time
407: separation with the help of constant plus exponential form.
408: The spatial transporters in the Wilson loop have been constructed 
409: from spatially
410: smeared links with APE smearing. The weight of the 3 link staple was $\gamma=0.4$
411: and we used ten steps of APE smearing. From the 
412: static potential we have determined the string tension and the Sommer parameter $r_0$ defined
413: as \cite{Sommer}  
414: \begin{equation}
415: r^2\frac{{\rm d} V_{\bar{q}q}(r)}{{\rm d}r}\biggl|_{r=r_0} = 1.65.
416: \label{r0}
417: \end{equation}
418: When extracting $r_0$ and the string tension  on coarse lattices, such as the ones 
419: used in thermodynamics studies, the
420: violation of rotational symmetry has to be taken into account. We do this using the procedure described
421: in detail in our recent paper \cite{us}.
422: The value of Sommer scale and the string tension are given in Table \ref{tab:T=0} for different quark masses.
423: Having determined $r_0$ for different gauge couplings and quark masses allows us to perform interpolations
424: of $r_0/a$ in these parameters. As in Ref. \cite{us} we use the following renormalization group inspired
425: interpolation ansatz \cite{allton}
426: \begin{equation}
427: (r_0/a)^{-1} =
428: R(\beta) (1 +B \hat{a}^2(\beta) + C \hat{a}^4(\beta))
429: {\rm e}^{A (2 m_l + m_s)+D}
430: \; .
431: \label{interpolate}
432: \end{equation}
433: Here $R(\beta)$ is 2-loop beta function of 3 flavor QCD.
434: In the interpolation we also used the values of $r_0/a$ determined in our 2+1 flavor study \cite{us}
435: in addition to those shown in Table \ref{tab:T=0}, giving $A=1.45(5)$, $B=1.20(17)$,  $C=0.21(6)$
436: and $D=2.41(5)$ with $\chi^2/dof=0.9$. This was the reason for  using the notation
437: $m_l$ and $m_s$  for the light and the strange quark mass in Eq. (\ref{interpolate}) . For 3 degenerate flavors of course $m_s=m_l=m$.
438: In  Table \ref{tab:T=0}  we also show the values of $r_0/a$ obtained from this ansatz for each of
439: the parameter sets.  
440:    
441: 
442: Meson masses have been calculated using the four local staggered meson operators. We used 
443: point-wall meson correlation functions with a $Z_2$ wall source. To extract the meson masses
444: from
445: the correlation functions we used a double exponential ansatz which 
446: takes into account the two lowest
447: states with opposite parity.  The two lowest pseudo-scalar meson masses 
448: as well as the lightest vector meson mass
449: are shown in  Table \ref{tab:T=0}. The breaking of the flavor symmetry can be 
450: quantified by the quadratic
451: splitting of the pseudo-scalar masses: 
452: $\Delta_{ps}=(m_{ps2}^2-m_{ps}^2) r_0^2$, where $m_{ps2}$ is the mass of the 
453: lightest pseudo-scalar non-Goldstone meson that is present with staggered fermions. 
454: This quantity should
455: be quark mass independent for sufficiently small  quark masses and should vanish 
456: as ${\cal O}(a^2)$ when the 
457: continuum limit ( $a \rightarrow 0$) is approached. 
458: In the last column of Table \ref{tab:T=0} we show the value 
459: of  $\Delta_{ps}$ from our scale setting run for $N_{\tau}=4$ and $N_{\tau}=6$.  
460: As we see from the table this quantity does
461: not decrease quite as fast as $a^2$. This is an indication that on the coarse 
462: lattice, corrections to asymptotic scaling are
463: still important.
464: 
465: \begin{table}
466: \begin{tabular}{|c|c|c|c|c|c|c|c|c|c|}
467: \hline
468: $\beta$ &  $m$   & \# traj    &  $m_{ps}$  & $m_{ps2}$ & $m_V$     & $r_0$      & $(r_0)_{smooth}$ &  $\sqrt{\sigma}$  &  $\Delta_{ps}$ \\
469: \hline     
470: 3.3877  &  0.050 &  7800    &  0.7084(1)  & 1.094(7)     & 1.310(20)   & 2.066(7)[7]       &  2.061            &   0.552(12)[12]         &   2.97(7)   \\
471: 3.3270  &  0.025 &  12000  &  0.5118(3)  & 0.998(24)   & 1.222(32)   & 1.982(14)[13]   &  1.989            &   0.564(11)[11]         &   2.90(20) \\
472: 3.2680  &  0.005 &  1500    &  0.2341(9)  & 0.860(90)   & 1.250(50)   & 1.888(15)[9]     &  1.888            &   0.587(17)[17]         &   2.44(55)  \\
473: \hline
474: 3.46345&   0.020 &  4420   &  0.4413(8)  &   0.665(5)   &  0.908(11)  &  2.797(20)[20]  &  2.813            &   0.404(6)[6]             &   1.94(9)  \\
475: 3.4400  &   0.010 &  4290   &  0.3210(7)  &   0.594(7)   &  0.882(20)  &  2.770(13)[13]  &  2.779            &   0.405(6)[6]             &   1.92(7)  \\
476: \hline
477: \end{tabular}
478: \caption{Parameters of the zero temperature simulations, meson masses, the Sommer scale $r_0$ and the string tension.
479: Also shown is the value of $r_0$ obtained from the interpolation formula (\ref{interpolate}). The upper part of the table refers to
480: scale setting runs for our $N_{\tau}=4$ lattices while the lower part 
481: to our $N_{\tau}=6$ calculations. In the last column the splitting between the lightest non-Goldstone and the Goldstone pseudo-scalar meson masses squared is shown.
482: All dimensionfull quantities are given in units of the lattice spacings.
483: }
484: \label{tab:T=0}
485: \end{table}
486: 
487: With the help of the interpolation formula we can calculate the lattice spacing in units of $r_0$ and thus 
488: $r_0 T_c$ for different pseudo-scalar meson masses, which is shown in Fig. \ref{fig:tc}. The error in $a(\beta_c)$ results in an error in the value of
489: $r_0 T_c$ which is shown in Fig.   \ref{fig:tc} as thin error-bars. 
490: The uncertainty in $\beta_c$ itself also contribute to the uncertainty in $r_0 T_c$, which is shown
491: as a thick error-bar in the figure. 
492: 
493: If there is a critical point in the $(m,T)$-plane, 
494: then universality dictates that
495: $T_c(m)-T_c(m^e) \sim (m-m^e)^{1/(\delta \beta)}$ with $\beta$ and $\delta$ 
496: being critical exponents. 
497: In the case of three degenerate flavors the line of the 1st order transition in the 
498: $(m,T)$ plane ends in a critical end-point $m_e, T_c(m^e)$ 
499: belonging to the Z(2) universality class. For this universality class we have $\delta \beta=1.5654$. 
500: Therefore
501: we attempted a combined continuum and chiral extrapolation using the following extrapolation
502: ansatz 
503: \begin{equation}
504: r_0 T_c(m_{ps},N_{\tau})=r_0 T_c|_{cont}(m_{ps}^e)+ A \left ( 
505: \left (r_0 m_{ps} \right )^2 - \left( r_0 m_{ps,c} \right )^2 \right)^{1/(\delta \beta)} 
506: + B/N_{\tau}^2.
507: \label{ansatz}
508: \end{equation}
509: The value of the quark mass where the transition changes from 1st order to crossover, i.e.
510: the mass corresponding to the end-point, has been estimated in \cite{karschlat03} using
511: $N_{\tau}=4$ lattices to be $m^e=0.0007(4)$. This translates into the value of the 
512: pseudo-scalar mass
513: \begin{equation}
514: r_0 m_{ps}^e =0.16^{+3}_{-5}.
515: \end{equation}
516: It turns out that this large uncertainty in the value of $m_{ps}^e$ produces
517: an uncertainty in the extrapolated value of $T_c(m_{ps}^2)$ which is much smaller
518: than the statistical errors. The extrapolation according to Eq. (\ref{ansatz})
519: yields
520: \begin{equation}
521: r_0 T_c (m_{ps}^e)  = 0.429(8),~~\frac{T_c(m_{ps}^e)}{\sqrt{\sigma}}=0.391(9).
522: \end{equation}
523: For the fit to $r_0 T_c$ data we get $\chi^2/dof=0.7$, while for the fit to
524: $T_c(m_{ps}^e)/\sqrt{\sigma}$ data we have $\chi^2/dof=0.4$. The quark mass
525: dependence of the transition temperature is described by Eq. (\ref{ansatz})
526: only for $m>m^e$. For smaller quark masses the transition is first order and
527: $T_c$ depends linearly on the quark mass, i.e. we expect 
528: $T_c(m_{ps},N_{\tau})=T_c|_{cont}^{chiral} + A m_{ps}^2 + B/N_{\tau}^2$.
529: If we would insist on the linear dependence of the transition temperature
530: on the quark mass  in the entire mass range, the combined chiral and
531: continuum extrapolation would give
532: \begin{equation}
533: r_0 T_c  = 0.419(9),~~\frac{T_c}{\sqrt{\sigma}}=0.383(10).
534: \end{equation}
535: The $\chi^2/dof$ we have found for this fit is almost the same as for
536: the one above. The value of $T_c/\sqrt{\sigma}$ is slightly smaller than
537: the estimate of Ref. \cite{karsch01} based on $N_{\tau}=4$ lattice and larger
538: quark masses.
539: This is due to the continuum extrapolation performed in the present work.
540: Note, however, that the value $T_c(0)^{N_t=4}/\sqrt{\sigma}=0.417(9)$ is 
541: entirely consistent with Ref. \cite{karsch01}.
542: The value of $T_c$ could be compared with  the corresponding $2+1$ flavor value
543: $T_c r_0=0.444(6)[+12][-6]$ in the limit of vanishing $u$ and $d$ quark masses but fixed physical value of $m_s$  \cite{us}. Thus the flavor
544: dependence of $r_0 T_c$ is about or smaller than $5\%$. One should also note that
545: the difference between the transition temperature calculated on $N_{\tau}=4$ and
546: $N_{\tau}=6$ lattices is very similar to that found in 2+1 flavor case \cite{us}.
547: \begin{figure}
548: \includegraphics[width=7cm]{tc_r0_pir0_nf3_exp.eps}
549: \includegraphics[width=7cm]{tc_sqrtsigma_pir0_nf3_exp.eps}
550: \caption{The value of the transition temperature in units of $r_0$ (left) and in units of $\sqrt{\sigma}$ (right) calculated on 
551: $N_{\tau}=4$ and $N_{\tau}=6$ lattices as function of $m_{ps}$ together with continuum extrapolated value.
552: The vertical line and band indicates the value of $m_{ps}$ where the transition becomes 1st order.}
553: \label{fig:tc}
554: \end{figure}
555: In Ref. \cite{karsch01} the transition temperature in the chiral limit has
556: also been estimated in units of the vector mass. Our estimate for
557: $T_c/m_V|_{m=0}$ is consistent with that result.
558: 
559: 
560: 
561: \section{Comparison of R and RHMC algorithm}
562: 
563: We investigated the effect of the finite step-size errors of the R algorithm on the properties of
564: the finite temperature transition. We performed calculations on $8^3 \times 4$ lattices
565: using the p4fat3 action as well as the p4fat7 action and the later will be described in  
566: Appendix A in more detail.  
567: In the calculations with the p4fat3 action we used a quark mass of $m=0.01$, while
568: in case of p4fat7 action we used two quark masses $m=0.1$ and $0.035$.
569: In our calculations with the R algorithm the step-size of the molecular dynamics
570: evolution was set to be $dt=m/2.5$. Some additional calculations have been
571: done at twice smaller step-size $dt=m/5$. We have calculated the chiral condensate
572: and the Polyakov loop and determined the pseudo-critical couplings which are
573: summarized in Table \ref{tab:rcomp}.
574: In Fig. \ref{fig:pbpsusc_alg_comp} we compare the chiral 
575: condensate susceptibility  calculated using the R-algorithm and RHMC algorithm for 
576: the p4fat3 action.  We find that for the p4fat3 action the results obtained with R and RHMC algorithms
577: are identical within statistical errors.
578: 
579: The situation is different for
580: p4fat7. 
581: In Fig. \ref{fig:fat7_alg_comp} the expectation value of the Polyakov loop and the chiral
582: condensate calculated with the two algorithms are shown for two values of the quark mass.
583: Here we see significant differences
584: in the value of the chiral condensate and Polyakov loops calculated with the R algorithm and step-size 
585: $dt=m/2.5$ and the corresponding result obtained with RHMC algorithm. We see also
586: a small but statistically significant difference in the value of the pseudo-critical coupling calculated
587: with the two algorithms, c.f. Table \ref{tab:rcomp}. 
588: The difference becomes much less visible when the step-size is decreased
589: to $dt=m/5$. Figure \ref{fig:fat7_alg_comp} also suggests that the difference between the results of the
590: two algorithms becomes larger for the smaller quark mass.
591: \begin{figure}
592: \includegraphics[width=7cm]{R_CS_p4fat3_8884_m0.01.eps}
593: \caption{The comparison of the disconnected part of chiral condensate susceptibility calculated with the R and RHMC algorithms
594: for p4fat3 with $m=0.01$.}
595: \label{fig:pbpsusc_alg_comp}
596: \end{figure}
597: \begin{figure}
598: \includegraphics[width=7cm]{R_L_p4fat7_8884_m0.1.eps}
599: \includegraphics[width=7cm]{R_L_p4fat7_8884_m0.035.eps}
600: \includegraphics[width=7cm]{R_C_p4fat7_8884_m0.1.eps}
601: \includegraphics[width=7cm]{R_C_p4fat7_8884_m0.035.eps}
602: \caption{The Polyakov loops (upper part) and the chiral condensate (lower part) calculated
603: with RHMC and R algorithms for the p4fat7 action with $m=0.1$ and $m=0.035$. }
604: \label{fig:fat7_alg_comp}
605: \end{figure}
606: 
607: For the p4fat7 action we also performed a zero temperature calculation on $16^3 \times 32$ lattices using the RHMC algorithm to determine the scale.
608: We have calculated meson masses as well the static quark potential. From the later we have extracted $r_0$. 
609: The results of these calculations  are also given in Table \ref{tab:rcomp}. Now we can estimate the transition temperature in units of $r_0$ for $m=0.035$
610: calculated with the two algorithms. For the R-algorithm we get $r_0 T_c=0.542(3)$ while for the RHMC algorithm  we have 
611: $r_0 T_c=0.552(3)$. Thus in the case of the p4fat7 action the R algorithm underestimates the transition temperature roughly by $2 \%$.
612: \begin{table}
613: \begin{tabular}{|c|c|c|c|c|c|c|}
614: \hline
615: Action  &  $m$    & Algorithm  & $\beta_{c, L}$   & $\beta_{c,q}$  &  $m_{ps}$  &   $r_0/a$   \\
616: \hline
617: p4fat3   &  0.010  & HMDR       & 3.2858(71)        &  3.2820(61)     &                   &                  \\
618:              &             & RHMC       & 3.2820(11)        &  3.2820(11)     &                   &                  \\
619: \hline
620: p4fat7   & 0.100  & HMDR      &  2.9850(25)       & 2.9753(53)       &                   &                  \\
621:              &             & RHMC      & 2.9939(33)        & 2.9831(20)       &                   &                  \\
622: \hline
623: p4fat7   & 0.035   & HMDR      &  2.7514(6)         & 2.7485(7)         &  0.7884(5)  &  2.1661(123) \\
624:              &             & RHMC      &  2.7540(6)         & 2.7515(7)         &  0.7897(7)  & 2.2063(108) \\
625: \hline
626: \end{tabular}
627: \caption{Comparison of the R and the RHMC algorithms for the pseudo-critical couplings and
628: the scale at the transition temperature.} 
629: \label{tab:rcomp}
630: \end{table}
631: 
632: \section{Conclusions}
633: 
634: In this paper we have studied the phase transition in 3 flavor QCD at finite
635: temperature using $N_{\tau}=4$ and $N_{\tau}=6$ lattices. For the quark mass
636: corresponding to the second order end-point we find the critical 
637: temperature to be $r_0 T_c=0.429(8)$. The transition temperature in the
638: chiral limit is about $2\%$ smaller than the above value. For a given 
639: pseudo-scalar meson mass the difference between the transition temperature in 
640: 3 flavor and 2+1 flavor case is less than $5\%$. We also find that the cut-off
641: dependence of the transition temperature in 3 and 2+1 flavor QCD is very similar.
642: Furthermore, we find that finite step-size errors present in the R algorithm are
643: negligible, at least for the p4fat3 action at the quark masses studied.
644: 
645: 
646: \section*{Acknowledgments}
647: \label{ackn}
648: This work has been supported in part by contracts DE-AC02-98CH1-886 
649: and DE-FG02-92ER40699 with the U.S. Department of Energy, 
650: the Helmholtz Gesellschaft under grant
651: VI-VH-041 and the Deutsche Forschungsgemeinschaft under grant GRK 881.  
652: The work of C.S. has been supported through LDRD funds from Brookhaven
653: National Laboratory.
654: The majority of the calculations reported here were carried out using
655: the QCDOC supercomputers of the RIKEN-BNL Research Center and the U.S.
656: DOE.  In addition some of the work was done using the APE1000 supercomputer
657: at Bielefeld University
658: 
659: 
660: 
661: 
662: \section*{Appendix A} 
663: In this appendix we are going to discuss the
664: properties of the finite temperature transition in the case of the p4fat7
665: action and compare it to the p4fat3 case.  In general the gauge transporter
666: in the 1-link term of the p4 action can be replaced by combination of
667: the link variable and different staples, called the fat link, without
668: changing the naive continuum limit. This is true provided the coefficient
669: of different terms in the fat link satisfy appropriate normalization
670: conditions.  For example in the case of the fat link with the three link
671: staple only, this condition reads $c_1 +6 c_3=1$.  Introducing five
672: and seven link staples  in addition to the three link staples give
673: the so-called fat7 link \cite{kostas}.  In this case the normalization
674: condition reads: $c_1+6c_3+24 c_5+48 c_7=1$. It is possible to
675: eliminate the leading order coupling to the high momentum gluons with
676: momenta $(0,\pi,0,0)$, $(0,\pi,\pi,0)$ and $(0,\pi,\pi,\pi)$, i.e. to
677: suppress the flavor changing interaction at order $g^2 a^2$ if the
678: coefficients are chosen as \cite{kostas}
679: \begin{equation}
680: \frac{c_3}{c_1}=\frac{1}{2},~\frac{c_5}{c_1}=\frac{1}{8},~\frac{c_7}{c_1}=\frac{1}{48}.
681: \end{equation}
682: This gives then the value $c_1=-1/8$ for the coefficient of the 1- link term to get
683: the naive continuum limit.
684: 
685: In Fig. \ref{fig:pbpfta7} we show the chiral condensate calculated on
686: the $N_{\tau}=4$ and $N_{\tau}=6$ lattices.  We can see that for small
687: quark masses the transition becomes strongly first order. The value of
688: the chiral condensate in the low temperature phase is also much larger
689: than for the p4fat3 action discussed in the main text. For the
690: smallest quark mass the discontinuity in the chiral condensate is
691: about the same for $N_{\tau}=4$ and $N_{\tau}=6$ lattice, although we would
692: expect it to decrease by roughly a factor of $(6/4)^3 \simeq 3$ when
693: going from $N_{\tau}=4$ to $N_{\tau}=6$.  This could mean that we are dealing
694: with a bulk transition.  In Fig. \ref{fig:betac} we also compare the
695: pseudo-critical couplings for the p4fat3 and p4fat7 actions. We see
696: that for large quark masses the $N_{\tau}$ dependence of the
697: pseudo-critical coupling is similar, though their values are
698: significantly different.  For small quark masses the pseudo-critical
699: couplings calculated for $N_{\tau}=4$ and $6$ come very close
700: together, again suggesting that the transition may be a bulk
701: transition. We did calculations also on $8^4$ lattice at $m=0.01$.
702: In Fig. \ref{fig:betac} we compare the chiral condensate calculated on $8^3 \times 4$, $16^3 \times 6$ and
703: $8^4$ lattices. We see a sharp drop in the value of the chiral condensate, which occurs at the same $\beta_c$ for
704: $N_{\tau}=6$ and $8$. This again indicates a bulk transition.
705: \begin{figure}
706: \includegraphics[width=7cm]{M_C_p4fat7_04_3f.eps}
707: \includegraphics[width=7cm]{M_C_p4fat7_06_3f.eps}
708: \caption{The chiral condensate calculated on $N_{\tau}=4$ (left) and
709: $N_{\tau}=6$ (right) lattices for p4fat7 action.}
710: \label{fig:pbpfta7}
711: \end{figure}
712: 
713: One may wonder which feature of the p4fat7 action is responsible for
714: the bulk transition.  The main difference of the p4fat7 action
715: compared to p4fat3 action as well as to other fat link action (e.g.
716: ASQTAD) is the negative sign of the one link term. 
717: Close to the continuum limit the normalization condition
718: $c_1 + 6 c_3 + 24 c_5 + 48 c_7 = 1$ should insure that the
719: combination of 1-, 3-, 5- and 7-link terms will describe a
720: conventionally normalized, positive Dirac kinetic energy.
721: However, at stronger coupling where the gauge field are
722: more disordered, the staples with many links are expected
723: to give a significantly smaller contribution and the 1-link
724: term may dominate, resulting in an effective kinetic energy
725: term with a possibly negative sign.  While this would
726: simply correspond to non-standard sign and normalization
727: conventions for the Dirac kinetic energy, it raises the
728: possibility that this effective kinetic energy term will
729: change sign as one passes from strong to weak coupling.
730: Such a sign change could induce a bulk transition.  In addition,
731: the change in magnitude of the coefficient in the effective
732: kinetic energy (small for strong coupling and large for
733: weak coupling) would appear reversed in the chiral condensate
734: (large for strong coupling and small for weak coupling)
735: consistent with the observed behavior.
736: To verify this we did calculations with p4fat7 action but with
737: different coefficients which we call the p4fat7' action.  The
738: coefficients were chosen to be
739: \begin{equation}
740:   c_1=\frac{3}{4} \cdot  \frac{1}{8},~\frac{c_3}{c_1}=\frac{1}{2},~\frac{c_5}{c_1}=\frac{1}{8},~\frac{c_7}{c_1}=\frac{1}{48}.
741: \end{equation}         
742: For this action we found no evidence for a strong first order
743: transition but only a crossover. This can be seen for example in the
744: behavior of the chiral condensate shown in Fig. \ref{fig:pbpfat7p}.
745: Both the value of the chiral condensate and the location of the
746: transition point is very similar to that of the p4fat3 action.
747: \begin{figure}
748: \includegraphics[width=7cm]{betac.eps}
749: \includegraphics[width=7cm]{C_p4fat7_nt.eps}
750: \caption{The pseudo-critical couplings for the p4fat3 and p4fat7 actions (left) and the
751: chiral condensate calculated with the p4fat7 action on $8^3 \times 4$, $16^3 \times 6$ and
752: $8^4$ lattices at $m=0.01$ (right). 
753: }
754: \label{fig:betac} 
755: \end{figure}
756: 
757: \begin{figure}
758:   \includegraphics[width=7cm]{C_p4fat7_prime++_8884_m0.01.eps}
759:   \includegraphics[width=7cm]{CS_p4fat7_prime_8884_m0.01.eps}
760:   \caption{The chiral condensate (left) and its susceptibility (right)
761:     for the p4fat7' action calculated on $8^3 \times 4$ lattice at $m=0.01$. We also show the chiral condensate
762: calculated with p4fat7 and p4fat3 actions at the same quark mass. The data points for p4fat3 action have been shifted horizontally by $\Delta\beta=-0.437$ 
763: for better visibility. The band in the right figure corresponds to Ferrenberg Swendsen re-weighting.}
764: \label{fig:pbpfat7p}
765: \end{figure}
766: 
767: We also calculated the eigenvalues $\lambda=\imath \lambda' +2 m$ of the p4fat7
768: Dirac operator. The normalized distribution of the lowest 50
769: eigenvalues $\lambda'$ is shown in Figs.  \ref{fig:eig_m01} and
770: \ref{fig:eig_m001} for the quark masses $m=0.1$ and $m=0.01$,
771: respectively. We have used 100 configurations for $m=0.01$, and 200
772: configurations for $m=0.1$.  
773: Given the above definition of $\lambda'$ the breaking of the chiral
774: symmetry manifests itself in a non-zero density at $\lambda' \simeq 0$.
775: In Fig. \ref{fig:eig_m01} we show the eigenvalue distribution for the
776: larger quark mass $m=0.1$.  It shows the expected features: large
777: density of eigenvalues near $\lambda \simeq 0$ in the low temperature
778: phase ($\beta=2.96$), significant drop of eigenvalue density around
779: zero at the transition ($\beta=3.0$) and zero density of eigenvalues
780: at the origin for the deconfined phase ($\beta=3.04$).  The situation
781: is different for the smallest quark mass $m=0.01$, where we see
782: non-zero density of eigenvalues at $\lambda \simeq 0$ even in the
783: deconfined phase ($\beta=2.66$). The large decrease in the density of
784: eigenvalues at the origin when going from the confined phase
785: ($\beta=2.635$) to the deconfined explains the large drop in the value of
786: the chiral condensate.
787: \begin{figure}
788: \includegraphics[width=5.5cm]{p4fat7_83x4_b2.96_m0.10_hist.eps}
789: \includegraphics[width=5.5cm]{p4fat7_83x4_b3.00_m0.10_hist.eps}
790: \includegraphics[width=5.5cm]{p4fat7_83x4_b3.04_m0.10_hist.eps}
791: \caption{The eigenvalue distributions of the p4fat7 Dirac operator for
792:   $m=0.1$ calculated on $8^3 \times 4$ lattice in the confined phase
793:   ($\beta=2.96$), at the transition ($\beta=3.00$) and in the
794:   deconfined phase ($\beta=3.04$).  }
795: \label{fig:eig_m01}
796: \end{figure}
797: 
798: \begin{figure}  
799: \includegraphics[width=5.5cm]{p4fat7_83x4_b2.635_m0.01_hist.eps}              
800: \includegraphics[width=5.5cm]{p4fat7_83x4_b2.64_m0.01_hist.eps}
801: \includegraphics[width=5.5cm]{p4fat7_83x4_b2.66_m0.01_hist.eps}
802: \caption{The eigenvalue distributions of the p4fat7 Dirac operator for
803:   $m=0.01$ calculated on $8^3 \times 4$ lattice in the confined phase
804:   ($\beta=2.635$), at the transition ($\beta=2.64$) and in the
805:   deconfined phase ($\beta=2.66$).  }
806: \label{fig:eig_m001}
807: \end{figure}
808: 
809: 
810: 
811: 
812: 
813: 
814: \section*{Appendix B}
815: In this appendix we discuss the calculations of the largest and the
816: smallest eigenvalue of the staggered Dirac operator in the free field
817: limit. Let us start our discussion with case of the standard staggered
818: fermions.  The free-field staggered Dirac operator acting on a
819: single-component fermion field is given by
820: \begin{eqnarray}
821:   \mathcal{M} \chi(x) = \sum_{\mu}\eta_{\mu}(x)\left(\chi(x+\mu) - 
822:     \chi(x-\mu) \right) + 2m\chi(x),
823: \end{eqnarray}
824: where $\eta_{\mu}(x) = (-1)^{x_0 + x_1 + \ldots x_{\mu-1}}$ are the
825: normal staggered phases. Consider $\mathcal{M}$ acting on a momentum
826: eigenstate
827: \begin{eqnarray}
828:   \chi(x) & = & \chi(p) \, \mathrm{e}^{\imath p \cdot x} \label{chifield}, \\
829:   \mathcal{M} \chi(x) & = & \sum_{\mu}\left[\chi(p) \,
830:     \mathrm{e}^{\imath p \cdot x}\,
831:     (\mathrm{e}^{\imath p_{\mu}} - \mathrm{e}^{-\imath p_{\mu}}) \, 
832:     \mathrm{e}^{\pi_{\mu} \cdot x} \right]
833:   + 2m \,\chi(p) \,\mathrm{e}^{\imath p \cdot x},
834: \end{eqnarray}
835: where $\pi_{0} = (0,0,0,0), \pi_{1} = (\pi, 0, 0, 0), \pi_{2} = (\pi,
836: \pi, 0, 0), \ldots$ Thus, the staggered Dirac operator has
837: non-diagonal terms that couple together states at different corners of
838: the Brillouin zone.
839: 
840: In momentum space, we can write the fermion matrix as
841: \begin{eqnarray}
842:   \mathcal{M}_{p',p} & = & \sum_{\mu} 2\imath\sin(p_{\mu}) \,
843:   \delta_{p',p+\pi_{\mu}} + 2m \,\delta_{p',p} \\
844:   \mathcal{M^{\dagger}}_{p'',p'} & = & \sum_{\mu} 
845:   -2\imath\sin(p''_{\mu})\, \delta_{p''+\pi_{\mu}, p'} +
846:   2m \,\delta_{p'',p'}.
847: \end{eqnarray}
848: Then, we have for $\mathcal{M}^{\dagger}\mathcal{M}$
849: \begin{eqnarray}
850:   \mathcal{M^{\dagger}}_{p'',p'}\mathcal{M}_{p',p}  & = 
851:   & 4 \sum_{\mu}\sum_{\nu}\sin(p_{\mu}) \sin(p''_{\nu}) \,
852:   \delta_{p''+\pi_{\nu}, p+\pi_{\mu}} + 4m^2 \delta_{p,p''} \\
853:   & = & 4 \sum_{\mu}\sum_{\nu}\sin(p_{\mu})
854:   \sin(p_{\nu}+(\pi_{\mu})_{\nu} - (\pi_{\nu})_{\nu}) \, \delta_{p''+
855:     \pi_{\nu}, p+\pi_{\mu}} +  4m^2 \delta_{p,p''} \\
856:   & = & 4\sum_{\mu}\sum_{\nu<\mu}\sin(p_{\mu})
857:   \sin(p_{\nu}+\pi) \,\delta_{p'',p+\pi_{\mu}-\pi_{\nu}} +  \nonumber\\
858:   & & 4\sum_{\mu}\sum_{\nu>\mu}\sin(p_{\mu})\sin(p_{\nu}) \,\delta_{p'',p+
859:     \pi_{\mu}-\pi_{\nu}} +  \nonumber \\
860:   &  & 4\sum_{\mu} \sin^2(p_{\mu}) \, \delta_{p,p''} +
861:   4m^2  \delta_{p,p''}. \label{eq:intmunu}
862: \end{eqnarray}
863: Noticing that $p+\pi_{\mu}-\pi_{\nu} = p + \pi_{\nu}-\pi_{\mu}$ (mod
864: 2$\pi$), and interchanging $\mu$ and $\nu$ labels in the second piece
865: of Eq.~(\ref{eq:intmunu}), we see that all the off-diagonal pieces of
866: $\mathcal{M}^{\dagger}\mathcal{M}$ cancel, and we are left only with
867: the diagonal piece,
868: \begin{eqnarray}
869:   (\mathcal{M}^{\dagger}\mathcal{M})_{p'',p} = 
870:   4\sum_{\mu}\sin^2(p_{\mu}) \, \delta_{p,p''} +
871:   4m^2 \delta_{p,p''}.
872: \end{eqnarray}
873: As expected, we have a hard lower bound on the eigenvalue spectrum of
874: $\lambda^2_{min}=4m^2$.  The upper bound $\lambda^2_{max} = 16 + 4m^2$ is
875: realized when $p =
876: (\frac{\pi}{2},\frac{\pi}{2},\frac{\pi}{2},\frac{\pi}{2})$.
877:  
878: The situation for free-field p4 fermions is similar, but a little bit
879: more complicated.  Here, the p4 Dirac operator is given by
880: \begin{eqnarray}
881:   \mathcal{M} \chi(x) & = & 2m\chi(x) + c_{1,0}\sum_{\mu}\eta_{\mu}(x) 
882:   \left( \chi(x+\mu) - \chi(x-\mu) \right)+ {}  \\
883:   & & + c_{1,2}\sum_{\mu}\sum_{\nu \neq \mu}\eta_{\mu}(x)
884:   \left(\chi(x+\mu+2\nu)+\chi(x+\mu-2\nu)-
885:     \chi(x-\mu+2\nu)-\chi(x-\mu-2\nu)\right),\nonumber
886: \end{eqnarray}
887: where $c_{1,0}=\frac{3}{4}$ and $c_{1,2}=\frac{1}{24}$. Again, we can
888: examine how $\mathcal{M}$ acts on the momentum eigenstate in
889: Eq.~(\ref{chifield}).  As before, we see that we have off-diagonal
890: pieces that come about as a direct result of the presence of staggered
891: phases.
892: \begin{eqnarray}
893:   \mathcal{M} \chi(x) & = & \sum_{\mu}\chi(p)\,
894:   \mathrm{e}^{\imath (p+\pi_{\mu}) \cdot x} \,\imath \, h_{\mu}(p) + 
895:   2m \,\chi(p) \, \mathrm{e}^{\imath p \cdot x}, \\
896:   h_{\mu}(p) & = & 2c_{1,0}\sin(p_{\mu}) + 
897:   4c_{1,2}\sum_{\nu \neq \mu}\sin(p_{\mu})\cos(2p_{\nu}).
898: \end{eqnarray}
899: Or, in matrix form,
900: \begin{eqnarray}
901:   \mathcal{M}_{p',p} & = & \sum_{\mu} \imath h_{\mu}(p) \,
902:   \delta_{p',p+\pi_{\mu}} + 2m \, \delta_{p',p}.
903: \end{eqnarray}
904: Calculating $\mathcal{M}^{\dagger}\mathcal{M}$, we see that the
905: off-diagonal pieces are eliminated in the same way as in the naive
906: staggered case. Thus, we are left with only diagonal terms,
907: \begin{eqnarray}
908:   (\mathcal{M}^{\dagger}\mathcal{M})_{p'',p} =  \sum_{\mu}h_{\mu}^2(p) \,
909:   \delta_{p,p''} + 4m^2 \delta_{p,p''}.
910: \end{eqnarray}
911: We see that $\lambda^2_{min} = 4m^2$.  Finding $\lambda^2_{max}$ requires
912: us to maximize the function $\sum h^2_{\mu}(p)$.  Doing this, we find
913: the maximum when two of the components of $p$ are equal to $\pi/2$ and
914: the other two components are equal to $0$.  For example, $p_{max} =
915: (\pi/2,0,\pi/2,0)$.  This yields $\lambda^2_{max} = 50/9+4m^2$ for the
916: p4 action.  A similar calculation for the Naik action yields
917: $\lambda^2_{max} = 196/9+4m^2$.
918: 
919: For completeness, we also quote the eigenvectors of the free p4 Dirac
920: operator. Using a slightly different method we find,
921: \begin{equation}
922: \label{eq:ansatzup4}
923: \chi(x) \equiv \chi_\rho(X) = \left[ \imath \sum_\mu
924:   \Gamma_{\rho\rho'}^\mu(p) \, \sin p_\mu \left(
925:   2 c_{1,0}+ 4 c_{1,2} \sum_{\nu \ne \mu}\cos 2 p_\nu 
926:   \right) + (\lambda -2 m) \, \delta_{\rho\rho'} \right] u^0_{\rho'} \; 
927: \mathrm{e}^{\imath 2 p \cdot X},
928: \end{equation}
929: where we implicitly sum over $\rho'$. Here we have used hypercube
930: coordinates $X$ and $\rho$, where $X$ labels the hypercube and $\rho$
931: is the offset within the hypercube, $x = 2 X + \rho$ with
932: $\rho_\mu=0,1$.  Furthermore,
933: $\Gamma_{\rho\rho'}^\mu(p)$ is defined by
934: $\Gamma_{\rho\rho'}^\mu(p) :=\eta_\mu(\rho)\; [
935: \delta_{(\rho+\hat{\mu}),\rho'} + \delta_{(\rho-\hat{\mu}),\rho'}]
936: \mathrm{e}^{\imath p \: (\rho-\rho')}$ and $u_{\rho}^0$ is a constant
937: vector depending only on $\rho$.  Finally,  the eigenvalue of
938: the free p4 Dirac operator is
939: \begin{equation}
940:   \label{eq:eigp4full}
941:   \lambda = 2 m \pm \imath \sqrt{\sum_{\mu} \sin^2 p_\mu
942:   \left( 2 c_{1,0} + 4 c_{1,2}\sum_{\nu \ne \mu}
943:     \cos 2 p_\nu \right)^2 }.
944: \end{equation}
945: 
946: 
947: 
948: 
949: \begin{thebibliography}{99}
950: \bibitem{cargese} 
951: F.~Karsch,
952: %``Lattice QCD at high temperature and density,''
953: Lect.\ Notes Phys.\  {\bf 583}, 209 (2002)
954: 
955: \bibitem{reviews} 
956: E.~Laermann and O.~Philipsen,
957: %``Status of lattice QCD at finite temperature,''
958: Ann.\ Rev.\ Nucl.\ Part.\ Sci.\  {\bf 53}, 163 (2003)
959: %\cite{Petreczky:2004xs}
960: %\bibitem{Petreczky:2004xs}
961:   P.~Petreczky,
962:   %``QCD thermodynamics on lattice,''
963:   Nucl.\ Phys.\ Proc.\ Suppl.\  {\bf 140}, 78 (2005)
964: %  [arXiv:hep-lat/0409139].
965: %%CITATION = HEP-LAT 0409139;%%
966: 
967: 
968: \bibitem{karschlat03}
969: %\cite{Karsch:2003va}
970: %\bibitem{Karsch:2003va}
971:   F.~Karsch, et al.,
972:   %``Where is the chiral critical point in 3-flavor QCD?,''
973:   Nucl.\ Phys.\ Proc.\ Suppl.\  {\bf 129}, 614 (2004)
974: %  [arXiv:hep-lat/0309116].
975:   %%CITATION = HEP-LAT 0309116;%%
976: 
977: \bibitem{christian}
978: %\bibitem{Schmidt:2002uk}
979:   C.~Schmidt, C.~R.~Allton, S.~Ejiri, S.~J.~Hands, O.~Kaczmarek, F.~Karsch and E.~Laermann,
980:   %``The quark mass and mu dependence of the QCD chiral critical point,''
981:   Nucl.\ Phys.\ Proc.\ Suppl.\  {\bf 119}, 517 (2003)
982:   [arXiv:hep-lat/0209009].
983:   %%CITATION = HEP-LAT 0209009;%%
984: %\cite{Karsch:2001nf}
985: %\bibitem{Karsch:2001nf}
986:   F.~Karsch, E.~Laermann and C.~Schmidt,
987:   %``The chiral critical point in 3-flavor QCD,''
988:   Phys.\ Lett.\ B {\bf 520}, 41 (2001)
989: %  [arXiv:hep-lat/0107020].
990:   %%CITATION = HEP-LAT 0107020;%%
991: 
992: 
993: \bibitem{phil2}
994: %\cite{deForcrand:2003ut}
995: %\bibitem{deForcrand:2003ut}
996:   P.~de Forcrand and O.~Philipsen,
997:   %``The phase diagram of N(f) = 3 QCD for small baryon densities,''
998:   Nucl.\ Phys.\ Proc.\ Suppl.\  {\bf 129}, 521 (2004)
999: %  [arXiv:hep-lat/0309109].
1000:   %%CITATION = HEP-LAT 0309109;%%
1001: 
1002: \bibitem{norman}
1003: N.H. Christ and X. Liao, Nucl. Phys.
1004:  B (Proc. Suppl.) {\bf 119}, 514 (2003)
1005: 
1006: \bibitem{karsch01}
1007: F.~Karsch, E.~Laermann, A.~Peikert, Nucl.\ Phys.\ B {\bf 605}, 579 (2001)
1008: 
1009: \bibitem{hybrid} 
1010: S.~A.~Gottlieb, W.~Liu, D.~Toussaint, R.~L.~Renken and R.~L.~Sugar,
1011: %``HYBRID MOLECULAR DYNAMICS ALGORITHMS FOR THE NUMERICAL SIMULATION OF
1012: %QUANTUM CHROMODYNAMICS,''
1013: Phys.\ Rev.\ D {\bf 35}, 2531 (1987)
1014: 
1015: \bibitem{rhmc} I. Horv\'ath, A. D. Kennedy and S. Sint,
1016: Nucl. Phys. B {\bf 73}, 834 (1999); 
1017: M. A. Clark, A. D. Kennedy and Z. Sroczynski,
1018: Nucl. Phys. Proc. Suppl. {\bf 140}, 835 (2005)
1019: 
1020: 
1021: \bibitem{aoki06}
1022: Y.~Aoki, Z.~Fodor, S.~D.~Katz and K.~K.~Szabo,
1023: %``The equation of state in lattice QCD: With physical quark masses towards
1024: %the continuum limit,''
1025: \emph{JHEP} {\bf 0601}, 089 (2006)
1026: 
1027: %\cite{deForcrand:2006pv}
1028: \bibitem{philipsen06}
1029:   P.~de Forcrand and O.~Philipsen,
1030:   % ``The chiral critical line of N(f) = 2+1 QCD at zero and non-zero baryon
1031:   %density,''
1032:   arXiv:hep-lat/0607017
1033:   %%CITATION = HEP-LAT 0607017;%%
1034: 
1035: 
1036: %\cite{Kogut:2006jg}
1037: \bibitem{SK06}
1038:   J.~B.~Kogut and D.~K.~Sinclair,
1039:   %``The RHMC algorithm for theories with unknown spectral bounds,''
1040:   arXiv:hep-lat/0608017
1041:   %%CITATION = HEP-LAT 0608017;%%
1042: 
1043: \bibitem{us}
1044: %\cite{Cheng:2006qk}
1045: %\bibitem{Cheng:2006qk}
1046:   M.~Cheng {\it et al.},
1047:   %``The transition temperature in QCD,''
1048:   Phys.\ Rev.\ D {\bf 74}, 054507 (2006)
1049: %  [arXiv:hep-lat/0608013].
1050: 
1051: 
1052: \bibitem{peikert_pressure}
1053: F.~Karsch, E.~Laermann and A.~Peikert,
1054: %``The pressure in 2, 2+1 and 3 flavour QCD,''
1055: Phys.\ Lett.\ B {\bf 478}, 447 (2000)
1056: 
1057: %\bibitem{Bernard04}
1058: %C.~Bernard {\it et al.}  [MILC Collaboration],
1059: %``QCD thermodynamics with three flavors of improved staggered quarks,''
1060: %Phys.\ Rev.\ D {\bf 71}, 034504 (2005).
1061: 
1062: 
1063: 
1064: \bibitem{Heller}
1065: U.~M.~Heller, F.~Karsch and B.~Sturm,
1066: %``Improved staggered fermion actions for QCD thermodynamics,''
1067: Phys.\ Rev.\ D {\bf 60}, 114502 (1999)
1068: 
1069: \bibitem{tb97}
1070: T. Blum et al., Phys. Rev. {\bf D55}, 1133 (1997)
1071: 
1072: \bibitem{kostas}
1073: %\cite{Orginos:1999cr}
1074: %\bibitem{Orginos:1999cr}
1075:   K.~Orginos, D.~Toussaint and R.~L.~Sugar  [MILC Collaboration],
1076:   %``Variants of fattening and flavor symmetry restoration,''
1077:   Phys.\ Rev.\ D {\bf 60}, 054503 (1999)
1078:  % [arXiv:hep-lat/9903032].
1079:   %%CITATION = HEP-LAT 9903032;%%
1080: 
1081: %\cite{Sharpe:2006re}
1082: \bibitem{sharpe}
1083:   S.~R.~Sharpe,
1084:   %``Rooted staggered fermions: Good, bad or ugly?,''
1085:   arXiv:hep-lat/0610094
1086: 
1087: 
1088: \bibitem{sextwein}
1089: J. C. Sexton, D.H. Weingarten, Nucl. Phys. {\bf B380}, 665 (1992)
1090: 
1091: 
1092: %\bibitem{cheng}
1093: %A.~Peikert, B.~Beinlich, A.~Bicker, F.~Karsch and E.~Laermann,
1094: %  ``Staggered fermion actions with improved rotational invariance,''
1095: %Nucl.\ Phys.\ Proc.\ Suppl.\  {\bf 63}, 895 (1998);\\
1096: %M.~Cheng  [RBC-Bielefeld Collaboration],
1097: %  ``Scaling test of the P4-improved staggered fermion action,''
1098: %PoS {\bf LAT2005}, 045 (2006)
1099: 
1100: 
1101: \bibitem{Sommer}
1102: R. Sommer, Nucl. Phys. {\bf B411} (1994) 839
1103: 
1104: %\bibitem{gray}
1105: %A. Gray et al., Phys. Rev. {\bf D72} (2005) 094507. 
1106: 
1107: 
1108: %\bibitem{MILCpotential}
1109: %C.~Bernard {\it et al.}  [MILC Collaboration], Phys. Rev. D70 (2004) 094505.
1110: 
1111: \bibitem{allton} C. Allton, Nucl. Phys. B [Proc. Suppl.] {\bf 53}, 867 (1997)
1112: %\bibitem{MILCmasses}
1113: %C.~W.~Bernard {\it et al.} [MILC Collaboration],
1114: %``The QCD spectrum with three quark flavors,''
1115: %Phys.\ Rev.\ D {\bf 64} (2001) 054506
1116: %\bibitem{kogut}J. Kogut and D.K. Sinclair, O(N)
1117: %\bibitem{cond_scaling} chiral condensate scaling
1118: 
1119: %\bibitem{zantow}
1120: %O.~Kaczmarek, F.~Karsch, F.~Zantow and P.~Petreczky,
1121: %``Static quark anti-quark free energy and the running coupling at finite
1122: %temperature,''
1123: %Phys.\ Rev.\ D {\bf 70}, 074505 (2004)
1124: %[Erratum-ibid.\ D {\bf 72}, 059903 (2005)]
1125: 
1126: %\bibitem{Gottlieb}
1127: %S.~Gottlieb {\it et al.},
1128: %``Onium masses with three flavors of dynamical quarks,''
1129: %PoS {\bf LAT2005}, 203 (2006).
1130: %\bibitem{Davies}
1131: %C.T.H. Davies et al., Phys. Rev. Lett. 92, 022001 (2004).
1132: 
1133: 
1134: %\bibitem{takaishi}
1135: %P. de Forcrand, Phys. Rev. {\bf E 73}, 036706 (2006)
1136: %\bibitem{omelyan}
1137: %I.P. Omelyan, I.M. Mryglod, R. Folk, Phys. Rev. {\bf E66}, 026701 (2002)
1138: 
1139: 
1140: %\bibitem{hasenbusch}
1141: %M. Hasenbusch, Phys. Lett. {\bf 519}, 177 (2001)
1142: 
1143: \end{thebibliography}
1144: \end{document}
1145: