1: \documentclass{JHEP3}
2:
3: \usepackage{epsfig}
4:
5: \newcommand{\be}{\begin{equation}}
6: \newcommand{\ee}{\end{equation}}
7:
8:
9:
10: \title{Phases of three dimensional
11: large N QCD on a continuum torus}
12: \author{R. Narayanan
13: \\Department of Physics, Florida International University, Miami,
14: FL 33199, USA\\E-mail: \email{rajamani.narayanan@fiu.edu}}
15: \author{ H. Neuberger
16: \\ Rutgers University, Department of Physics and Astronomy,
17: Piscataway, NJ 08855, USA\\E-mail: \email
18: {neuberg@physics.rutgers.edu} }
19: \author{F. Reynoso
20: \\Department of Physics, Florida International University, Miami,
21: FL 33199, USA\\E-mail: \email{freyn001@fiu.edu}}
22:
23: \abstract {It is established by numerical means that continuum
24: large N QCD defined on a three dimensional torus
25: can exist in four different phases.
26: They are (i) confined phase; (ii) deconfined phase;
27: (iii) small box at zero temperature and (iv) small box
28: at high temperatures.
29: }
30:
31:
32: \keywords{Large N QCD, Lattice Gauge Field Theories}
33:
34: \preprint{}
35:
36: \begin{document}
37:
38: \section{Introduction.}
39:
40: Yang Mills theory on an $l^d$ continuum torus
41: ($d>2$) exhibits a phenomenon referred to as
42: continuum reduction whereby the theory
43: for $l > l_1 > 0$
44: is independent of $l$
45: \footnote{We use $l_1$ to denote the critical size
46: as opposed to $l_c$ since we will have a sequence
47: of critical sizes.}~\cite{Narayanan:2003fc,Kiskis:2003rd}.
48: At $l=l_1$, the theory goes from the confined phase ($0$c: $l > l_1$)
49: to the deconfined phase ($1$c: $l < l_1$). The order parameter is
50: the Polyakov loop and rotational symmetry is spontaneously
51: broken. More phases were conjectured to exist~\cite{Kiskis:2003rd}
52: in the continuum theory and these
53: are refereed to as $X$c phase with $X=2,\cdots, d$.
54:
55: The aim of this paper is to numerically
56: establish the existence of the
57: $2$c and $3$c phase in addition to the
58: $0$c and $1$c phase for the continuum Yang-Mills theory
59: on a periodic torus. We will use the Polyakov
60: loop to define an order parameter
61: to be labeled, $P$~\cite{Bhanot:1982sh},
62: and it will take values in the range [0,0.5].
63: If the $U(1)$ symmetry\footnote{The
64: $U(1)$ symmetry is the limit of the $Z_N$ symmetry
65: as $N\rightarrow\infty$.} under which the Polyakov loop
66: transforms non-trivially is spontaneously broken,
67: then $\bar P < 0.5$.
68: Let $\bar P_{x,y,z}$, be
69: the order parameters
70: in the three directions.
71: Then, $\bar P_x=\bar P_y=\bar P_z=0.5$
72: in the $0$c phase.
73:
74: There are three possibilities for the $1$c phase
75: and one of them is characterized by
76: $\bar P_y=\bar P_z=0.5$
77: and $\bar P_x < 0.5$ with rotational symmetry still
78: present in the $(y,z)$ plane. It is difficult to
79: numerically establish the order of the transition from
80: the $0$c to $1$c phase and we will leave it unresolved
81: in this paper.
82:
83: The $2$c phase also has three possibilities
84: and one of them is characterized by $\bar P_x=\bar P_y < 0.5$
85: and $\bar P_z=0.5$ with rotational symmetry present in the $(x,y)$
86: plane. This transition occurs when $l=l_2 < l_1$.
87: One can argue that the $1$c to $2$c phase transition
88: is first order as follows. In the $1$c phase, $\bar P_x$ was
89: less than $0.5$ and $\bar P_y$ was $0.5$. Since $\bar P_x=\bar P_y$
90: in the
91: $2$c phase, it is necessary for at least $\bar P_x$ or $\bar P_y$
92: to change discontinuously at the $1$c to $2$c transition.
93: If one operator shows a discontinuity, all operators
94: will generically show discontinuities and this
95: signals a first order transition.
96:
97: Rotational symmetry is restored in the $3$c phase and
98: $\bar P_x=\bar P_y=\bar P_z < 0.5$. For the same reason as above,
99: we expect the $2$c to $3$c phase transition to be
100: of first order. This transition occurs when $l=l_3 < l_2$.
101:
102: The $2$c phase is characterized by
103: two short directions and one infinitely long direction
104: since the theory will not depend on the length of
105: the direction where the U(1) is not broken. Therefore,
106: this phase describes large N QCD in a small box at
107: zero temperature (or infinite time). Confinement
108: cannot be addressed in the $2$c phase since we do
109: not have large Wilson loops. The $3$c phase
110: describes large N QCD in a small box at high
111: temperatures. The $2$c to $3$c transition is
112: like the transition seen in perturbation theory
113: on $S^2\times S^1$\cite{Papadodimas:2006jd}
114: where $S^2$ replaces the two torus along which
115: the U(1) symmetry is broken in the $2$c phase.
116:
117: We extend our discussion to include $l_x \times l_y \times
118: l_z$ torus with $l_x < l_y < l_z$.
119: The transition from $0$c to $1$c will occur at $l_x=l_1$
120: and this is independent of $l_y$ and $l_z$.
121: The transition from $1$c to $2$c will occur at $l_y=l_2(l_x)$
122: with $0 \le l_x \le l_2$. Furthermore, $l_2(l_2)=l_2$
123: and our numerical results will show that $l_2(0) > 0$.
124: Finally, there is no dependence on $l_z$.
125: Continuing along the same lines, we can say that the
126: $2$c to $3$c transition occurs at $l_z=l_3(l_x,l_y)$
127: with $l_3=l_3(l_3,l_3)$. It is possible that one
128: can obtain this critical size for small
129: $l_x$ and $l_y$ by perturbation theory
130: but it is necessary to consider the zero momentum modes
131: of the gauge fields on all three directions. We do
132: not address this problem in the paper.
133:
134: The results in this paper complement the results in the
135: closely related paper by Bursa and Teper~\cite{Bursa:2005tk}.
136: We mainly focus on the continuum limit of the various phases
137: using Polyakov loops. The paper
138: is organized as follows. We define the relevant technical
139: details in section~\ref{latdef}. Our
140: numerical results showing the existence of the
141: various phases are presented in section~\ref{polydet}.
142:
143: \section{Technical details \label{latdef}}
144: \subsection{Lattice gauge action}
145: We used the single plaquette Wilson action given by
146: \begin{eqnarray}
147: S=\frac{\beta}{4N}\sum_{x,i\ne j} Tr[ U_{ij}(n)
148: +U_{ij}^\dagger (x) ] \\
149: U_{ij}(n)=U_i (n) U_j (n+\hat i) U_i^\dagger (n+\hat j)
150: U_j^\dagger (n)
151: \end{eqnarray}
152: $n$ is a three component integer
153: vector labeling the site, $i$ labels a direction and
154: $\hat i$ denotes
155: a unit vector in the $i$ direction. The link matrices $U_\mu(n)$
156: are in $SU(N)$.
157: We define
158: \be
159: b=\frac{\beta}{2N^2}
160: \ee
161: and take
162: the large $N$ limit with $b$ held fixed.
163:
164: All computations were done on a
165: $L_x \times L_y \times L_z$ periodic lattice with $L_x \le L_y = L_z$.
166: One gauge field update of the whole lattice~\cite{Kiskis:2003rd}
167: is one Cabibo-Marinari heat-bath
168: update of the whole lattice
169: followed by one SU(N) over-relaxation update of the whole lattice.
170: The code was run on two clusters, one with 48 nodes and another
171: with 31 nodes. The nodes in the cluster were simply used to
172: generate more statistics using a parallel random number generator
173: and generating independent configurations with the same set of
174: parameters on different nodes.
175:
176: \subsection{Determination of the critical sizes}
177: Given the lattice coupling $b$ and lattice sizes $L_x$ and $L_y$,
178: the dimensionless physical sizes are defined as
179: \be
180: l_{x,y}=\lim_{b\rightarrow\infty} L_{x,y}/b_{\rm tad}.
181: \ee
182: The tadpole improved coupling~\cite{Lepage:1996jw},
183: $b_{\rm tad}$ is defined as
184: \be
185: b_{\rm tad} = b e(b) = b \langle
186: \frac{1}{12NL_x L_y L_z}\sum_{n,i\ne j} Tr[ U_{ij}(n)
187: +U_{ij}^\dagger (n) ] \rangle \label{plaq}
188: \ee
189:
190: \subsection{Lattice bulk transition}
191: Since the computations in this paper use the Wilson gauge action
192: on the lattice, it is necessary to address the
193: unphysical transition which is the extension
194: of the Gross-Witten transition~\cite{Gross:1980he} in QCD${}_2$.
195: The order parameter for this transition is the plaquette
196: operator. The third order transition analytically computed
197: in QCD${}_2$ remains to be true based on a numerical
198: investigation in QCD${}_3$~\cite{Bursa:2005tk}
199: and the critical point is $b=0.43$.
200: This lattice transition does not survive the continuum
201: limit and we will work with $b>0.43$ throughout this paper
202: in order to describe continuum physics.
203:
204: \subsection{An order parameter}
205: An order parameter suitable for studying the phase transitions
206: we are interested in is~\cite{Bhanot:1982sh}
207: \begin{eqnarray}
208: \bar P_{x,y,z} &=& \left < P_{x,y,z} \right > \cr
209: P_{x,y,z} &=& \frac{1}{2 L_x L_y L_z} \sum_n 1 - \left | \frac{1}{N}
210: Tr {\cal P}_{x,y,z}(n)
211: \right |^2 \cr
212: {\cal P}_{x,y,z}(n) &=& \prod_{m=1}^{L_{x,y,z}} U_i(n+m\hat i).
213: \end{eqnarray}
214: The quantity $P_{x,y,z}$ takes values in the range $[0,0.5]$ on
215: any gauge field background and we choose the
216: $x$, $y$ and $z$
217: directions on each configuration
218: such that $P_x < P_y < P_z$.
219:
220: Although this observable needs to be renormalized, we found it
221: sufficient to work with the unrenormalized operator and
222: we also did not have to smear the link variables.
223: The eigenvalues,$e^{i\theta_k}$; $k=1,\cdots, N$,
224: of the Polyakov loop operator,
225: ${\cal P}_{x,y,z}(n)$, are gauge invariant.
226: $P_{x,y,z}=0.5$ implies a uniform distribution of the eigenvalues
227: of ${\cal P}_{x,y,z}(n)$.
228: A departure from $P_{x,y,z}=0.5$ implies the presence of
229: a peak in the distribution of the eigenvalues of ${\cal P}_{x,y,z}(n)$
230: and a breaking of $Z_N$ symmetry associated with this operator.
231: There is no gap in the distribution of the eigenvalues
232: of ${\cal P}_{x,y,z}$ when $Z_N$ is broken.
233:
234: \section{Transitions in Polyakov loops \label{polydet}}
235:
236: All computations were done using $N=47$. Having picked a lattice
237: size $L_x \times L_y \times L_y$, each run listed in Table~\ref{tab1}
238: was a closed loop in the lattice coupling $b$.
239: The third column in Table~\ref{tab1}
240: shows the range of $b$ and the step size in $b$. The fourth
241: column, $N_t$, shows the number of thermalization
242: sweeps at the two end points. Only one measurement was done per node
243: at each $b$ and
244: the fifth column, $N_b$, shows the
245: number of sweeps done at each intermediate $b$ before the measurement.
246: For example,
247: the run on $3^3$ lattice,
248: started at a $b=0.5$ and went up to a $b=2.5$.
249: A total of $2000$ sweeps were performed at $b=0.5$ and $b=2.5$ and
250: a total of $1000$ sweeps were performed for all $b$ in between $0.5$
251: and $2.5$. The step size in $b$ was $0.05$ and this code was run on
252: the 31 node cluster with one measurement at each $b$ per node. All
253: values of $b$ between $0.5$ and $2.5$ had two sets of measurements;
254: one on the way up in $b$ and one the way down in $b$.
255:
256: \TABLE{
257: \begin{tabular}{ccccccccc}
258: $L_x$ & $L_y$ & $b$ & $N_t$ & $N_b$ & $N_{\rm cfg}$
259: & $\frac{L_x}{b_{\rm tad}}(0{\rm c}-1{\rm c})$
260: & $\frac{L_y}{b_{\rm tad}}(1{\rm c}-2{\rm c})$
261: & $\frac{L_z}{b_{\rm tad}}(2{\rm c}-3{\rm c})$ \cr
262: \hline
263: 3 & 3 & [0.5,2.5;0.05] & 2000 & 1000 & 31 & 6.14(33) & 4.02(14) & 2.27(17)\\
264: 4 & 4 & [0.5,3.5;0.05] & 3000 & 600 & 48 & 5.76(21) & 3.83(28) & 2.14(17)\\
265: 5 & 5 & [0.5,2.5;0.05] & 2000 & 400 & 31 & 5.60(47) & 3.73(35) & 2.17(23)\\
266: 6 & 6 & [0.5,4.5;0.10] & 2000 & 400 & 31 & 5.37(48) & 3.82(70) & 1.99(38)\\
267: 5 & 6 & [1.5,3.5;0.10] & 3000 & 600 & 48 & & 2.70(25) & \\
268: 4 & 5 & [0.5,3.5;0.05] & 3000 & 600 & 48 & 6.00(46) & 2.47(23) & \\
269: 3 & 4 & [0.5,2.5;0.05] & 3000 & 600 & 48 & 6.56(76) & 2.07(20) & \\
270: 4 & 6 & [2.0,5.0;0.10] & 3000 & 600 & 48 & & 1.78(17) & \\
271: 3 & 5 & [2.0,4.0;0.10] & 3000 & 600 & 48 & & 1.53(15) & \\
272: 3 & 6 & [2.0,8.0;0.20] & 2000 & 400 & 31 & & 1.14(11) & \\
273: \hline
274: \end{tabular}
275: \caption{\label{tab1} The parameters of all the runs used
276: to study the transitions in Polyakov loops along with the
277: results for the critical sizes.}}
278:
279: \subsection{Details of the data analysis}
280:
281: The plaquette
282: as defined in (\ref{plaq}) was measured on all configurations and
283: this was used to obtain the tadpole improved coupling, $b_{\rm tad}$.
284: Figure~\ref{444} shows the results for all three Polyakov loop
285: observables as a function of $\frac{4}{b_{\rm tad}}$ for the data
286: obtained on the $4^3$ lattice (second row in Table~\ref{tab1}).
287: The hysteresis is clear in both
288: $\bar P_x$ and $\bar P_y$ for the $1$c-$2$c transition and
289: it is seen in all the $\bar P_{x,y,z}$
290: for the $2$c-$3$c transition.
291: The critical size for the various transitions along with the
292: error is obtained by locating the two points
293: (one for upward direction and the second for the
294: downward direction) where the error
295: is largest in the observable that is broken. The
296: vertical lines in Figure~\ref{444}
297: shows the critical sizes along with the errors
298: and these results are shown in the
299: last three columns of Table~\ref{tab1}.
300:
301: Within the $2$c phase one sees a difference in $\bar P_{x}$
302: and $\bar P_y$. But this is just a consequence of our
303: choice of observable. Note that we have picked $P_x < P_y$
304: on every configuration. If we assume two
305: independent Gaussian random variables, $\alpha$ and $\beta$,
306: that have the same mean and variance, then one can
307: show that the variables $P_x$ and $P_y$ defined as the
308: minimum and maximum of $\alpha$ and $\beta$ will
309: be distributed such that
310: \be
311: \frac{ \bar P_y - \bar P_x}
312: {\sqrt{ \left< P^2_{x,y} \right> - P^2_{x,y} }} =
313: \frac{2}{\sqrt{\pi -1}}.
314: \ee
315: Our data within the $2$c phase is consistent with the above
316: equation.
317:
318: We did not choose a range in $b$ such that all transitions
319: are seen on all $L_x$, $L_y$ pairs since some of them were
320: used only to investigate the $1$c-$2$c transition. But,
321: we always picked a range such that the end points are
322: in one of the four phases.
323:
324: \FIGURE[htp]{
325: \epsfig{file=poly_444_600.eps, height=3.5in }
326: \caption{Plot of $\bar P_{x,y,z}$ for the data
327: in the second row of Table~\ref{tab1}
328: showing all three transitions.}
329: \label{444}}
330:
331: \subsection{$0$c-$1$c transition}
332:
333: Let us focus on the seventh column in Table~\ref{tab1}
334: to study the confinement-deconfinement transition. The results on
335: $3^3$, $4^3$, $5^3$ and $6^3$ show that the $0$c-$1$c transition
336: is physical since the critical size, $\frac{L_x}{b_{\rm tad}}$,
337: is the same on all four lattices within errors. The results
338: here are consistent with the older results presented
339: in~\cite{Narayanan:2003fc}. We also studied the $0$c-$1$c transition
340: on $4\times 5^2$ and $3\times 4^2$ and found that the critical
341: size is independent of $L_y$ as expected. Figure~\ref{P1} shows
342: that the six results for the $0$c-$1$c transition do scale properly
343: and we estimate the continuum critical size to be
344: $l_1=5.90(47)$.
345: If we take the central value for the dimensionless string tension
346: from~\cite{Bringoltz:2006zg}, namely, $\sqrt{\sigma}=0.1975$,
347: then we get
348: \be
349: \frac{1}{l_1\sqrt{\sigma}}=0.86(7)
350: \ee
351: and this is consistent~\cite{Liddle:2005qb}
352: with saying that $\frac{1}{l_1}$
353: is the deconfinement temperature.
354:
355: \FIGURE[htp]{
356: \epsfig{file=P1.eps, height=3.5in }
357: \caption{Plot of $\bar P_x$ showing the $0$c-$1$c transition.}
358: \label{P1}}
359:
360: \subsection{$1$c-$2$c transition on $L^3$ lattices}\label{1to2}
361:
362: The physical size associated with the $1$c-$2$c transition, $l_2$.
363: is expected to depend on $l_x$, the
364: temperature in the deconfined phase.
365: We first estimate the critical
366: size on lattices with $L_x=L_y$. We use the data on
367: $3^3$, $4^3$, $5^3$ and $6^3$.
368: The four results show continuum scaling as can be seen
369: from Figure~\ref{P2a} and we conclude that
370: the $1$c-$2$c transition exists in the continuum limit.
371: We estimate $l_2(l_2)=3.85(43)$.
372: As mentioned before, $1$c phase is the deconfined phase.
373: The system is a small finite box at zero temperature
374: in the $2$c phase. This transition occurs on a
375: $l^3$ torus when the temperature is $1.53(21)$ times
376: the deconfinement temperature.
377:
378: \FIGURE[htp]{
379: \epsfig{file=P2a.eps, height=3.5in}
380: \caption{Plot of $\bar P_y$ showing the $1$c-$2$c transition
381: on lattices with $L_x=L_y$.}
382: \label{P2a}}
383:
384: \subsection{$2$c-$3$c transition on $L^3$ lattices}
385:
386: We also investigated the $2$c-$3$c transition on
387: $3^3$, $4^3$, $5^3$ and $6^3$.
388: Here again,
389: the four results show continuum scaling as can be seen
390: from Figure~\ref{P3} and we conclude that
391: the $2$c-$3$c transition also exists in the continuum limit.
392: The transition size will depend on $l_x$ and $l_y$ when
393: both of them are smaller than $l_z$. But, we only estimate
394: $l_3(l_3,l_3)=2.14(26)$ here.
395:
396: Large N QCD on a very small torus, $l^3$,
397: for $l< l_3$ feels
398: the size of the box and the temperature is high.
399: Large N QCD is in a small box of size $l$
400: at zero temperature if $l > l_3$ and
401: it
402: undergoes a phase transition into the deconfined phase
403: when the box
404: size is $1.80(30)$ times $l_3$.
405:
406: \FIGURE[htp]{
407: \epsfig{file=P3.eps, height=3.5in}
408: \caption{Plot of $\bar P_y$ showing the $2$c-$3$c transition
409: on lattices with $L_x=L_y$.}
410: \label{P3}}
411:
412: \subsection{Phase diagram for $l_x \le l_y \le l_z$}
413:
414: The single scale in the $1$c phase is $l_x$ which
415: can also be thought of as inverse temperature in the
416: deconfined phase. The $2$c phase has two scales,
417: namely the size of the two dimensional box $l_x$ and $l_y$
418: with $l_x \le l_y$.
419: If $l_y > l_2(l_x)$, then the theory does not depend on
420: $l_y$ and we are in the deconfined phase. We considered
421: the special case of $l_x=l_y$ in section~\ref{1to2}.
422: We extended this to the case when $l_x < l_y$.
423: For this purpose,
424: we considered the lattices listed in last six rows of Table~\ref{tab1}.
425: The phase transition in $\bar P_y$ is shown in Figure~\ref{P2b}.
426: There is an
427: obvious dependence of the critical size $l_2$ on $l_x$.
428:
429: Figure~\ref{xvsy} summarizes the various phases by focusing
430: on the $(l_x,l_y)$ plane at $l_z=l_y$.
431: The dependence of $l_2(l_x)$ is shown using the shaded
432: square points in Figure~\ref{xvsy}. The dashed line is
433: a quadratic fit to the seven points and we note that
434: $l_2(0) > 0$.
435: In order to get an overall picture, we have also shown
436: the $0$c-$1$c transition in Figure~\ref{xvsy}. The
437: dotted line indicates that the $0$c-$1$c transition
438: does not depend on $l_y$ for $l_y > l_x$. Figure~\ref{xvsy}
439: also shows the $0$c, $1$c and $2$c phases for $l_x \le l_y \le l_z$.
440: For completeness, we have also shown the $2$c-$3$c transition
441: as seen on this specific $(l_x,l_y)$ plane restricted to $l_x=l_y$.
442: Like the $1$c-$2$c transition, the $2$c-$3$c transition will also
443: show a dependence on $l_x$ for $l_x< l_y$
444: and the $2c$ phase will not reach the $l_x=l_y$ line
445: for $l_x=l_y < l_2$. Furthermore, the
446: $2$c-$3$c transition curve will change as one changes the
447: $l_z$ that defines the $(l_x,l_y)$ plane. We have not
448: investigated these details pertaining to the $2$c-$3$c
449: transition in this paper. But, we should remark that the
450: rest of the phase diagram does not depend on $l_z$ for
451: $l_z > l_y > l_x$.
452:
453: \FIGURE[htp]{
454: \epsfig{file=P2b.eps, height=3.5in}
455: \caption{Plot of $\bar P_y$ showing the $1$c-$2$c transition
456: on lattices with $L_x \le L_y$.}
457: \label{P2b}}
458:
459: \FIGURE[htp]{
460: \epsfig{file=xvsy.eps, height=3.5in}
461: \caption{Phase diagram for $l_x \le l_y \le l_z$}
462: \label{xvsy}}
463:
464:
465: \section{Conclusions}
466:
467: Large $N$ QCD in three dimensions on a $l^3$
468: continuum torus exists in four different phases. The theory
469: is in the confined phase ($0$c) for $l \ge 5.90(47)= l_1$
470: and physics does not depend on the box size. This critical
471: size is the inverse of the deconfinement temperature,
472: $T_c=\frac{1}{l_1}$,
473: and the
474: theory is in the deconfined phase ($1$c) for
475: $1 < \frac{T}{T_c} < 1.53(21)$.
476:
477: The system is in a finite box and feels the
478: effect of temperature ($3$c phase) when $l < 2.14(26)= l_3$.
479: The temperature has no effect if $1 < \frac{l}{l_3} < 1.80(30)$
480: ($2$c phase).
481: The system goes into the deconfined phase if
482: $\frac{l}{l_3} > 1.80(30)$.
483:
484: All phase transitions are most likely first order in nature.
485: We have provided arguments for this scenario when going
486: from $1$c-$2$c and $2$c-$3$c.
487: Previous results~\cite{Liddle:2005qb,firstorder}
488: indicate that the deconfinement phase transition is also first order.
489:
490: The $1$c to $2$c transition on a $l_x \times l_y \times l_z$ torus
491: with $l_x \le l_y \le l_z$ depends on $l_x$. The critical line is
492: given by $l_2(l_x)=0.56 + 1.08 l_x - 0.059 l_x^2$ and this is valid
493: for $0 \le l_x \le 3.85(23)$ and it is independent of $l_z$. This
494: transition has been analyzed for one point on the $(l_x,l_y)$ plane
495: in~\cite{Bursa:2005tk} and we are in agreement with the result in that
496: paper.
497:
498: \acknowledgments
499:
500: R. N. and F.R. acknowledge partial support by the NSF under grant number
501: PHY-055375
502: R.N. also acknowledges partial support from Jefferson Lab. The Thomas Jefferson National
503: Accelerator Facility (Jefferson Lab) is operated by the
504: Southeastern Universities Research Association (SURA) under DOE
505: contract DE-AC05-84ER40150. H. N. acknowledges partial
506: support by the DOE under grant number DE-FG02-01ER41165 at Rutgers,
507: an Alexander von Humboldt award and the hospitality of the Physics
508: department at Humboldt University, Berlin.
509:
510: \begin{thebibliography}{99}
511: %\cite{Narayanan:2003fc}
512: \bibitem{Narayanan:2003fc}
513: R.~Narayanan and H.~Neuberger,
514: %``Large N reduction in continuum,''
515: Phys.\ Rev.\ Lett.\ {\bf 91}, 081601 (2003)
516: [arXiv:hep-lat/0303023].
517: %%CITATION = PRLTA,91,081601;%%
518:
519: %\cite{Kiskis:2003rd}
520: \bibitem{Kiskis:2003rd}
521: J.~Kiskis, R.~Narayanan and H.~Neuberger,
522: %``Does the crossover from perturbative to nonperturbative physics in QCD
523: %become a phase transition at infinite N?,''
524: Phys.\ Lett.\ B {\bf 574}, 65 (2003)
525: [arXiv:hep-lat/0308033].
526: %%CITATION = PHLTA,B574,65;%%
527:
528: %\cite{Bhanot:1982sh}
529: \bibitem{Bhanot:1982sh}
530: G.~Bhanot, U.~M.~Heller and H.~Neuberger,
531: %``The Quenched Eguchi-Kawai Model,''
532: Phys.\ Lett.\ B {\bf 113}, 47 (1982).
533: %%CITATION = PHLTA,B113,47;%%
534:
535: %\cite{Papadodimas:2006jd}
536: \bibitem{Papadodimas:2006jd}
537: K.~Papadodimas, H.~H.~Shieh and M.~Van Raamsdonk,
538: %``A second order deconfinement transition for large N 2+1 dimensional
539: %Yang-Mills theory on a small S**2,''
540: arXiv:hep-th/0612066.
541: %%CITATION = HEP-TH/0612066;%%
542:
543: %\cite{Bursa:2005tk}
544: \bibitem{Bursa:2005tk}
545: F.~Bursa and M.~Teper,
546: %``Strong to weak coupling transitions of SU(N) gauge theories in 2+1
547: %dimensions,''
548: Phys.\ Rev.\ D {\bf 74}, 125010 (2006)
549: [arXiv:hep-th/0511081].
550: %%CITATION = PHRVA,D74,125010;%%
551:
552:
553: %\cite{Lepage:1996jw}
554: \bibitem{Lepage:1996jw}
555: G.~P.~Lepage,
556: %``Redesigning lattice QCD,''
557: arXiv:hep-lat/9607076.
558: %%CITATION = HEP-LAT/9607076;%%
559:
560: %\cite{Gross:1980he}
561: \bibitem{Gross:1980he}
562: D.~J.~Gross and E.~Witten,
563: %``Possible Third Order Phase Transition In The Large N Lattice Gauge
564: %Theory,''
565: Phys.\ Rev.\ D {\bf 21}, 446 (1980).
566: %%CITATION = PHRVA,D21,446;%%
567:
568: %\cite{Bringoltz:2006zg}
569: \bibitem{Bringoltz:2006zg}
570: B.~Bringoltz and M.~Teper,
571: %``A precise calculation of the fundamental string tension in SU(N) gauge
572: %theories in 2+1 dimensions,''
573: Phys.\ Lett.\ B {\bf 645}, 383 (2007)
574: [arXiv:hep-th/0611286].
575: %%CITATION = PHLTA,B645,383;%%
576:
577: %\cite{Liddle:2005qb}
578: \bibitem{Liddle:2005qb}
579: J.~Liddle and M.~Teper,
580: %``The deconfining phase transition for SU(N) theories in 2+1 dimensions,''
581: PoS {\bf LAT2005}, 188 (2006)
582: [arXiv:hep-lat/0509082].
583: %%CITATION = POSCI,LAT2005,188;%%
584:
585: %\cite{firstorder}
586: \bibitem{firstorder}
587: P.~de Forcrand and O.~Jahn,
588: %``Deconfinement transition in 2+1-dimensional SU(4) lattice gauge theory,''
589: Nucl.\ Phys.\ Proc.\ Suppl.\ {\bf 129}, 709 (2004)
590: [arXiv:hep-lat/0309153];
591: %%CITATION = NUPHZ,129,709;%%
592: K.~Holland,
593: %``Another weak first order deconfinement transition: Three-dimensional SU(5)
594: %gauge theory,''
595: JHEP {\bf 0601}, 023 (2006)
596: [arXiv:hep-lat/0509041].
597: %%CITATION = JHEPA,0601,023;%%
598:
599: \end{thebibliography}
600:
601:
602: \end{document}
603: