Non-perturbative thermodynamics of SU (N) gauge theories
J. Engels, J. Fingberg, F. Karsch, D.
Miller and M. Weber Fakultiit J~r Physik, Universitdt Bielefeld, W-4800 Bielefeld 1, FRG Received 20 September 1990The pressure near the deconfinement transition as determined up to now in lattice gauge theories shows unphysical behaviour:
it can become negative and may in SU (3) even have a gap at the transition. This has been attributed to the use of only perturba- tively known derivatives of coupling constants. We propose a method to evaluate the pressure, which works without these deriv- atives, and is valid on large lattices. In SU(2) we study the finite-volume effects and show that for lattices with spatial extent N, ~> 15 these effects are negligible. In SU ( 3 ) we then obtain a positive and continuous pressure. The influence of non-perturbative corrections to the fl-function on the energy density are investigated and found to be important, in particular for the latent heat.
1. Introduction
M o n t e Carlo s i m u l a t i o n s o f lattice gauge theories have p r o v e n to be a powerful tool to analyze the non- p e r t u r b a t i v e aspects o f these theories. This a p p r o a c h also o p e n e d for the first t i m e the p o s s i b i l i t y to study the f i n i t e - t e m p e r a t u r e Q C D phase t r a n s i t i o n f r o m first principles a n d to o b t a i n q u a n t i t a t i v e results for the t r a n s i t i o n t e m p e r a t u r e . The analysis o f o t h e r im- p o r t a n t t h e r m o d y n a m i c q u a n t i t i e s like energy den- sity, e n t r o p y d e n s i t y o r pressure h a v e also b e e n stud- ied on the lattice. T h e o p e r a t o r s used to extract these quantities have, however, an essential drawback: they involve in a d d i t i o n to certain m a t r i x e l e m e n t s o f the field strength t e n s o r also d e r i v a t i v e s o f the b a r e cou- plings with respect to t e m p e r a t u r e a n d / o r v o l u m e [ 1 ]. The relevant d e r i v a t i v e s o f the space-like (g~) a n d time-like (g~) couplings in the s t a n d a r d W i l s o n a c t i o n have been calculated p e r t u r b a t i v e l y [ 2 ]. O n e finds to O(g 2) **l
(0g;2
u g / ~ ' ¢ = L = c ' + O ( g 2); (1)~ The derivative with respect to the temperature Tcan be writ- ten in terms of the lattice anisotropy ~= a/a,, where a and a~
are the lattice spacings in the space-and time-like directions, respectively. Details about this and our notation can be found in ref. [ l ].
( 0 g 7 2
- - - ~ J ¢ = 1 = C'~ + O ( g 2 ) . ( 1 c o n t ' d ) In m o s t n u m e r i c a l calculations o f t h e r m o d y n a m i c a l q u a n t i t i e s p e r f o r m e d in the past the l e a d i n g - o r d e r w e a k - c o u p l i n g expressions for these d e r i v a t i v e s (c~,, c'~) have been used. Results for t h e r m o d y n a m i - cal q u a n t i t i e s o b t a i n e d in this way are thus not en- tirely n o n - p e r t u r b a t i v e .
N o n - p e r t u r b a t i v e results for the relevant deriva- tives o f the couplings have b e e n calculated at some selected p o i n t s [ 3 ]. These calculations indicate that at least for the SU ( 3 ) gauge t h e o r y at i n t e r m e d i a t e couplings ( g 2 ~ l ) d e v i a t i o n s from the p e r t u r b a - tively calculated values can be large. This does not c o m e as a surprise since it is k n o w n t h a t the deriva- tives o f g~ a n d g~ are related to the Q C D fl-function t h r o u g h
a~_~g a = g 3 ( ~ + Og~-2~
.
(
2)
As there are large d e v i a t i o n s from the p e r t u r b a t i v e fl-function o f the S U ( 3 ) gauge t h e o r y for g~> 1, it is to be expected that this is also true for the d e r i v a t i v e s ofg~ a n d g~.
A n o t h e r i n d i c a t i o n o f the i n a d e q u a c y o f the per- t u r b a t i v e relations for the d e r i v a t i v e s Og;c~) / O~ at in- t e r m e d i a t e couplings comes f r o m recent high-statis- tics calculations o f t h e r m o d y n a m i c a l quantities for
the SU ( 3 ) gauge theory on t h e r m a l lattices o f size N 3 × 4 [4,5]. Close to the critical coupling for the first-order d e c o n f i n e m e n t transition, which on lat- tices o f this size occurs close to 6 / g 2~ - 5.69, one ob- serves that the pressure, P, b e c o m e s negative. More- over, P is d i s c o n t i n u o u s at the critical point. This a p p r o a c h also leads to i n c o m p a t i b l e results for the la- tent heat o f the transition, if it is extracted either from the i n t e r a c t i o n measure d = ( ~ _ - 3 P ) / T 4 [ 1 ], which reflects the d e v i a t i o n s o f the t h e r m o d y n a m i c a l sys- tem from a massless ideal gas, or the enthalpy den- sity, e + P [4,5]. M o n t e Carlo s i m u l a t i o n s for these quantities have thus reached an accuracy where in- consistencies resulting from the usage o f operators for t h e r m o d y n a m i c a l observables, which are not entirely n o n - p e r t u r b a t i v e , b e c o m e visible.
2. Thermodynamics
One way o f arriving at c o m p l e t e l y n o n - p e r t u r b a - rive expressions for t h e r m o d y n a m i c a l observables is o u t l i n e d in ref. [ 3 ]: one first calculates the couplings given in eq. ( 1 ) n o n - p e r t u r b a t i v e l y at zero t e m p e r a - ture. These results can then be inserted in the stan- d a r d expressions for the t h e r m o d y n a m i c a l observables.
An alternative a p p r o a c h , which we w a n t to discuss here, is based on a calculation o f the free-energy den- s i t y , f and the i n t e r a c t i o n measure A. The free energy is related to the p a r t i t i o n function via
f = - T V -1 In Z . ( 3 )
On the lattice the l o g a r i t h m o f the p a r t i t i o n function m a y be calculated f r o m the e x p e c t a t i o n value o f the s t a n d a r d W i l s o n action, S, since the d e r i v a t i v e with respect to the bare coupling B = 2 N / g 2 is [ 1 ]
- 0 In Z / O f f = ( S ) = 3 N 3 N~( P~ + P~) , ( 4 ) where P~(~) is the space ( t i m e ) p l a q u e t t e with P = N -1 ( T r ( 1 - UUU*U t ) ) .
The physical free-energy density is then o b t a i n e d up to an integration constant, i.e. the value at Bo, from
= - 3 N , 4 d B ' [ 2 P o - ( P , + P d ] . ( 5 ) Bo
Here we have n o r m a l i z e d f r e m o v i n g the v a c u u m contribution at a p p r o x i m a t e l y T = 0 by subtracting the plaquette value Po on a s y m m e t r i c lattice ( N ~ = N o ) .
The i n t e r a c t i o n measure A can be o b t a i n e d as [ 1 ]
dg 4
A = ~ - 3 P -2
T 4 - a ~ - a 6 N N ~ [ 2 P o - ( P ~ + P O ] . ( 6 ) This relation involves the Q C D B-function, adg/da.
In the past it has been a p p r o x i m a t e d in the leading o r d e r by the p e r t u r b a t i v e weak-coupling expression, eq. ( 2 ) . However, as discussed a b o v e we should take into account d e v i a t i o n s from a s y m p t o t i c scaling in the ff-function in o r d e r to achieve a truly n o n - p e r t u r - b a t i v e calculation o f the t h e r m o d y n a m i c a l quan- tities. Starting with a calculation o f f a n d A has the a d v a n t a g e that we need to know only the n o n - p e r t u r - b a t i v e form o f t h e / ? - f u n c t i o n and the average pla- quette Po+ P,, i.e. the expectation value o f the action on a lattice with N~<N~ and on the c o r r e s p o n d i n g s y m m e t r i c lattice with N~ = N~.
Other t h e r m o d y n a m i c a l quantities can then be ob- t a i n e d from s t a n d a r d t h e r m o d y n a m i c a l relations, if we assume in a d d i t i o n h o m o g e n e i t y o f the system ~2 This general p r o p e r t y can usually be expected to hold in a given phase o f a very large system o f a single par- ticle type when only isotropic interactions are acting.
An i m p o r t a n t consequence o f the h o m o g e n e i t y o f the system is then
0 In Z In Z
0-I7 T-- V ' ( 7 )
which in fact relates the pressure, P, and the free-en- ergy density, f through the identity
P = - f . ( 8 )
G i v e n eq. ( 8 ) a n d the t h e r m o d y n a m i c relation
f = ¢ - T s , ( 9 )
we can then d e t e r m i n e the e n t r o p y density, s, using eqs. ( 5 ) a n d ( 6 ) ,3. In general we can expect eq. ( 8 ) to h o l d in the t h e r m o d y n a m i c limit. In a finite vol-
~2 An extensive quantity is said to be homogeneous of order one when an increase of the size of the system by a factor 2 leads to an increase of the quantity by the same factor.
~3 We note that the quantity (e+P)/T calculated on finite lat- tices is usually taken to be the entropy density, though this is true only for homogeneous systems.
ume, however, this relation is v i o l a t e d due to surface effects. F o r a d e r i v a t i o n o f n o n - p e r t u r b a t i v e ther- m o d y n a m i c s o f S U ( N ) gauge theories on finite lat- tices we therefore have to look s o m e w h a t closer at the p r o p e r t i e s o f the t h e r m o d y n a m i c a l functions:
( 1 ) The relation between P a n d f g i v e n in eq. ( 8 ) has to be investigated for finite volumes.
( 2 ) We need a n o n - p e r t u r b a t i v e fl-function.
We shall discuss these p o i n t s in the following sections.
3. The free-energy density and pressure
As p o i n t e d out a b o v e we w a n t to d e t e r m i n e the pressure f r o m the free-energy d e n s i t y by s i m p l y using eq. ( 8 ) . A first test o f eq. ( 8 ) consists in c o m p a r i n g the w e a k - c o u p l i n g e x p a n s i o n s o f P / T 4 a n d - - f i t 4.
T h e e x p a n s i o n s are o b t a i n e d from those o f the pla- quettes, which were calculated up to o r d e r g4 in ref.
[ 6 ]. O f course, the i n t e g r a t i o n c o n s t a n t to be used in eq. ( 5 ) is the value at g2 = 0, i.e. we can only test the a p p r o a c h to this point. The i n t e g r a n d c o n t a i n s first a g2-term, which m u s t d i s a p p e a r to leave a finite inte- gral. Indeed, with the values f r o m ref. [ 6 ] , we f i n d for N~= 16, N ~ = 4 a c a n c e l l a t i o n o f the g2-factor with an accuracy o f 10 - 6 . T h e next t e r m f r o m the integral a n d the c o r r e s p o n d i n g one f r o m the pressure weak- coupling e x p a n s i o n (see, e.g., ref. [7 ] ) c o i n c i d e within 5%.
We have s t u d i e d the p r o b l e m o f finite-size correc- tions to eq. ( 8 ) at i n t e r m e d i a t e fl-values in the case o f SU ( 2 ) . T h e r e we have a large set o f d a t a on lat- tices with N ~ = 8 , 12, 18, 26 a n d N ~ = 4 f r o m a finite- size analysis [ 8 ] a n d an i n v e s t i g a t i o n o f the interac- t i o n m e a s u r e A [ 7 ]. In a d d i t i o n we have calculated s o m e new p o i n t s on the 83)<4 lattice below f l = 2 . 2 4 . In fig. 1 we show the pressure d a t a d i v i d e d by T 4 re- sulting from the usual formula, ref. [ 1 ], involving the p e r t u r b a t i v e coupling derivatives. I n s i d e the fluctua- tion o f the d a t a p o i n t s (the error b a r s have been o m i t t e d for clarity, their size is o f the o r d e r o f the fluctuation for the d a t a f r o m each lattice size) the P / T 4 values show no significant finite-size effects. In- d e e d this is e x p e c t e d f r o m finite-size scaling t h e o r y [ 8 ] . The c o r r e s p o n d i n g d a t a for A = ( E - 3 P ) / T 4 are gathered in fig. 2. T h e y show a c o n s i d e r a b l e finite- size effect in the n e i g h b o u r h o o d o f the critical cou-
0 . 3
0 . 2
0 . 1
i
2 . 2 0
. . . . I . . . . I . . . . I
p / T 4 o x
8 u (2)
X 0
O 2 6 n
0 IB
X 't2 0 X,,,
Ix
A
~ o ~
~ ~ ~*X~×~ x ~ ~
~.~...a...~.~...,, ... ~ ... ~ ...
A
4 / g e
I I I I , , , i I , i i i I i i
2 . 2 5 2 . 3 0 2 . 3 5
Fig. 1. Pressure divided by T 4 for the SU(2) gauge theory plot- ted against B on N~=8, 12, 18, 26 and N,=4 lattices calculated with the conventional formula involving the perturbative cou- pling derivatives.
~.0
0 . 5
I I I I I I
SU (2)
AX X O 2 6
~ o 18
~ × ~
A a 8
A o
~ 0 v ~1
o ......................................................... ....
I I I I I I
2 . 2 2 . 3 2 . 4 2 . 5 2 . 6 4 / g 2
Fig. 2. The interaction measure A = ( ~ - 3 P ) / T 4 for SU(2) gauge theory plotted against fl, eq. (6), with adg-2/da calculated from the weak-coupling scaling relation. The lines are fits through the small-l/points, used for the integral in eq. (5) for No=8 (long dashes), 12 (short dashes) and 18 (solid line).
piing, tic = 2.3, up to about fl= 2.35. Since A is propor- tional to the integrand in eq. (5) we expect a similar finite-size effect for f / T 4.
We have e v a l u a t e d f f r o m eq. (5) in two steps: in the regions where our data populations were dense enough, we have used the histogram technique [9]
to find the integral contributions from the N 3 × 4 lat- tices, the contributions o f the symmetric lattice were calculated from spline fits to high-statistics 164 lat- tice plaquette values in the same fl-ranges; in the small-fl region we have supplemented the A data by fits through the remaining few points (the fits are also shown in fig. 2) and then integrated these fits to ob- tain the lower integral parts. The errors from the his- togram method are difficult to determine. Most probably they are much smaller than the fluctuation o f data points, since information from m a n y overlap- ping histograms (in the case o f the 83X 4 lattice we had 38, for 183X 4 still 20) was used simultaneously.
Errors from the fits to A and their integrals could shift the free-energy density curves slightly in the y-direc- tion but would not change their shapes.
The resulting curves for - f i T 4 on N~=8, 12, 18 lattices (for N~= 26 too few points were available) are compared in fig. 3 to the direct data for P / T 4 on the 183X4 lattice from fig. 1. We observe a strong volume dependence o f the free-energy density indi- cating that F = V/is not an extensive quantity on small lattices. On the other hand, the N o = 18 curve is in perfect agreement with the direct data above and al- ready at the transition point, i.e. for SU (2) we see in this region essentially no violation of the perturba- tive relations for the coupling derivatives. Below the phase transition the direct data should, however, de- viate more and more from the true non-perturbative result.
We conclude that for SU (2) on lattices with N~> 15 (and N~= 4) the finite-size dependence o f the free- energy density becomes negligible and therefore that eq. (8) is applicable. For SU (3) gauge theory we ex- pect similar finite-size effects; in contrast to SU (2), however, a noticeable difference to the perturbative calculation is anticipated.
To calculate the pressure via eq. (8) we have taken the SU (3) data on 163 × 4 and 16 4 lattices from ref.
[4]. In fig. 4a we show N 4 ( 2 P o - P , - P ~ ) as ob- tained from the A-data o f ref. [4], and the known factors in front o f this quantity in eq. (6). The solid
'i . . . .
-f/T 4 anc/
I . . . . P / T 4 I . . . . I 0S U
(a) /
/ o- - 18 0 /
O.t
. . . . 12
// ,~
8 / / / , , , /
0 . . .
,4/g 2
l t i i t i i i i I r t i l i i i
2.20 2.25 2.30 2.35
Fig. 3. Finite-size dependence of the free-energy density for the SU(2) gauge theory plotted against fl, Compared a r e - f i T 4 from eq. (5) for N,,=8 (long dashes), 12 (short dashes) and 18 (solid line) with P/T 4 (circles) from the direct formula with perturba- tive coupling derivatives for the 183× 4 lattice (same as in fig. 1).
curve in the plot is an interpolation o f these data points, with a gap assumed at f l = 5.6925. F r o m this interpolation we evaluated the integral in eq. (5). The resulting non-perturbative P I T 4 is shown in fig. 4b together with the Monte Carlo data ofref. [ 4 ], which were calculated with perturbative coupling deriva- tives. We note that by construction this pressure is always continuous across the deconfinement transi- tion. Evidently, our approach solved the unsatisfac- tory situation of a negative, discontinuous pressure.
4. Non-perturbative//-function and energy density Already the early Monte Carlo renormalization group ( M C R G ) studies [ 10,11 ] of the Q C D fl-func- tion have shown that there are considerable devia- tions from the weak-coupling scaling relation. In par- ticular in the case of SU (3) large deviations have been observed for fl> 5.7, which, however, seem to disap- pear rapidly above fl> 6.1. The numerical results for the discrete fl-function, Afl(fl), obtained from a start-
1 . 5
l . O
0 . 5
0 . 0
- 0 . 5 5 . 6
, I , i
N4 (2p o- (Pa+Pr))
p/T4
[](b)
a SU (3)
................. ......
5 . 8 6 , o
6/g 2
Fig. 4. (a) The difference between plaquette expectation values on asymmetric and symmetric lattices times N 4 from Monte Carlo S U ( 3 ) data [4] on 1 6 a x 4 and 16 4 lattices, respectively. The solid line is an interpolation through these data, used for the integral in eq. (5). (b) The non-perturbative result for the pressure in SU ( 3 ) gauge theory as a function offl along with the Monte Carlo data of ref. [4], based on perturbative coupling derivatives.
d a r d M C R G analysis, have been fitted with Pad6 ap- p r o x i m a n t s , in o r d e r to extract a n o n - p e r t u r b a t i v e r - function [ 12,13 ]. The a p p r o x i m a n t s are chosen such that they are consistent with the p e r t u r b a t i v e f o r m at large fl a n d r e p r o d u c e the o b s e r v e d scaling v i o l a t i o n s d o w n to f l ~ 5.7. To be specific we use the following p a r a m e t r i z a t i o n given in ref. [ 13 ]:
dg (1-alg2)2+a~g 4
a ~ a a = b o g 3 [l_(al+bl/2bo)gZ]2+a3g4, ( 1 0 ) where bo= 11N/48x 2 a n d bt = 34N2/768/~ 4 are the first two coefficients in the p e r t u r b a t i v e Q C D B-function for SU(N) gauge theories. In the case o f SU(3) we use the set o f coefficients { a ~ = 0 . 8 5 3 5 7 2 , a 2 = 0.0000093, a3 = 0 . 0 1 5 7 9 9 3 } , given in ref. [13].
We have used this n o n - p e r t u r b a t i v e r - f u n c t i o n a n d o u r i n t e r p o l a t i o n o f N 4 ( 2 P o - P , - P , ) , shown in fig.
4a, to extract the i n t e r a c t i o n m e a s u r e A from eq. ( 6 ) . This is p l o t t e d in fig. 5a as a solid curve. F o r c o m p a r - ison we show the u n c h a n g e d d a t a o f ref. [ 4 ] , which are b a s e d on the a s y m p t o t i c weak-coupling scaling relation. N e a r the transition point, a r o u n d r = 5.7, we observe a d r o p in A by about a factor two, which arises f r o m the n o n - p e r t u r b a t i v e r - f u n c t i o n . By construc- t i o n the two results a p p r o a c h each o t h e r at higher r - values. The non-perturbative energy density was then found by adding three times the pressure from eq. ( 8 ) ( s o l i d curve in fig. 4 b ) . In fig. 5b we c o m p a r e the result ( s o l i d c u r v e ) again with the u n c h a n g e d d a t a f r o m ref. [4 ]. A s i m i l a r d r o p as in A occurs; in par- ticular the slight p e a k in ~ / T 4 n e a r r = 5.83 has dis-
6
u ( e - 3 P ) / T 4
1
4 [] (a)
• [ ] ~
0 - - -
p I ~ I r
[] m [] o
e/T 4
m°
:
3
4 ~
( b )s u (3)
2
1 ~ 16ax4
0 I . . .
5,6
5 , 8 6 . 06/g 2
Fig. 5 T h e i n t e r a c t i o n m e a s u r e A = ( e - 3 P ) / T 4, ( a ) , a n d the en- ergy density divided by T 4, ( b ), plotted against fl for SU ( 3 ) gauge theory. The data are from ref. [4] and were calculated using per- turbative coupling derivatives. The solid lines show the corre- sponding results when the pressure is computed from eq. (8) and the non-perturbative r-function, eq. (10), is taken into account.
a p p e a r e d - the n o n - p e r t u r b a t i v e energy d e n s i t y is a very m o n o t o n o u s l y rising function a n d the gap A e /
T 4 is r e d u c e d considerably.
5. Summary and discussion
In s u m m a r i z i n g , we have seen that, u n d e r the as- s u m p t i o n o f the h o m o g e n e i t y o f the system, it is pos- sible to calculate the pressure n o n - p e r t u r b a t i v e l y on the lattice. By construction it is also c o n t i n u o u s at the first-order t r a n s i t i o n p o i n t s and it turns out to be positive everywhere. The i n f o r m a t i o n n e e d e d on an N 3 × N~ lattice is j u s t the e x p e c t a t i o n value o f the av- erage p l a q u e t t e (Po+P~) and a c o r r e s p o n d i n g value Po on a s y m m e t r i c N4~ lattice, i.e. essentially the actions.
F r o m the same lattice data, the average plaquettes, one then d e t e r m i n e s the i n t e r a c t i o n m e a s u r e e - 3 P a n d after that by the simple a d d i t i o n o f 3P the energy density. In this second step, however, one needs the //-function. The n o n - p e r t u r b a t i v e fl-function m a y change the shape o f the energy density a n d the size o f the latent heat density in SU ( 3 ) considerably. In fact it m a y resolve the p r o b l e m o f the N~-dependence o f A ~ / T 4 f o u n d in ref. [ 14]. Because o f the c o n t i n u i t y o f the n o n - p e r t u r b a t i v e pressure there is no further a m b i g u i t y in the d e t e r m i n a t i o n o f the latent heat density, as is the case in the usual gap calculations from the enthalpy density, e + P , or ~ - 3 P involving the d e r i v a t i v e s c~, c'~ and the a s y m p t o t i c fl-function.
This a m b i g u i t y a n d its origin, the i n c o m p l e t e knowledge o f the d e r i v a t i v e s o f the couplings with re- spect to the a n i s o t r o p y ~, was a l r e a d y n o t i c e d in one o f the first d e t e r m i n a t i o n s o f the latent heat d e n s i t y in SU ( 3 ) , ref. [ 15 ]. The cure in b o t h refs. [ 14 ] a n d [ 15 ] was to i m p o s e the c o n d i t i o n o f c o n t i n u i t y o f the pressure at the t r a n s i t i o n point. The two calculations still differ in so far as in ref. [ 14] the q u a n t i t y e + P , a n d in ref. [ 15 ] the q u a n t i t y E - 3P, was used to ex- tract the gap in the energy density. The e n t h a l p y den- sity e + P d e p e n d s on the difference o f the partial de- rivatives given in eq. ( 1 ) , while the i n t e r a c t i o n measure ~ - 3P d e p e n d s on the sum o f these deriva- tives via eq. ( 2 ) . I f one insists on using p e r t u r b a t i v e coupling derivatives, the second q u a n t i t y is to be pre- ferred, because, owing to eq. ( 2 ) , and the weak-cou- pling fl-function the sum o f the d e r i v a t i v e s is known up to o r d e r g2, the difference only to o r d e r gO.
In full Q C D theory, the same difficulties in the de- t e r m i n a t i o n o f the pressure appear, see, for example, ref. [ 16 ]. T h e r e too our p r o c e d u r e to o b t a i n a phys- ical pressure is applicable a n d certainly s u p e r i o r to the c o n v e n t i o n a l a p p r o a c h as long as M o n t e Carlo s i m u l a t i o n s for t h e r m o d y n a m i c quantities are re- stricted to rather small t e m p o r a l lattices (N~ < 10).
Acknowledgement
We are i n d e b t e d to H L R Z Jiilich, the B o c h u m U n i v e r s i t y c o m p u t e r centre a n d the J. von N e u m a n n c o m p u t e r centre, P r i n c e t o n for p r o v i d i n g the neces- sary c o m p u t e r time. We t h a n k the Deutsche F o r - schungsgemeinschaft for partial s u p p o r t u n d e r grant En 164/2.
References
[ 1 ] J. Engels, F. Karsch, I. Montvay and H. Satz, Nucl. Phys. B 205 [FS5] (1982) 545.
[2] F. Karsch, Nucl. Phys. B 205 [FS5] (1982) 285.
[3] G. Burgers, F. Karsch, A. Nakamura and I.O. Stamatescu, Nucl. Phys. B 304 (1988) 587.
[4] Y. Deng, Nucl. Phys. B (Proc. Suppl. ) 9 (1989) 334;
F.R. Brown, N.H. Christ, Y. Deng, M. Gao and T.J. Woch, Phys. Rev. Lett. 61 (1988) 2058.
[5] M. Fukugita, A. Ukawa and M. Okawa, Nucl. Phys. B 337 (1990) 181.
[6] U. Heller and F. Karsch, Nucl. Phys. B 251 [FSI3] (1985) 254.
[ 7 ] J. Engels, J. Fingberg, K. Redlich, H. Satz and M. Weber, Z.
Phys. C42 (1989) 341.
[8] J. Engels, J. Fingberg and M. Weber, Nucl. Phys. B 332 (1990) 737.
[9] A.M. Ferrenberg and R.H. Swendsen, Phys. Rev. Lett. 63 (1989) 1195.
[ 10 ] A. Hasenfratz, P. Hasenfratz, U. Heller and F. Karsch, Phys.
Lett. 140B (1984) 76; B 143 (1984) 193.
[ 11 ] R. Gupta, G. Guralnik, A. Patel, T. Warnock and C. Zemach, Phys. Rev. Lett. 53 (1984) 1721; Phys. Lett. B 161 (1985) 352.
[12] D. Petcher, Nucl. Phys. B 275 [FS17] (1986) 24l.
[ 13 ] J. Hoek, Nucl. Phys. B 339 (1990) 732.
[14] N.H. Christ and H.Q. Ding, Phys. Rev. Lett. 60 (1988) 1367.
[ 15 ] B. Svetitsky and F. Fucito, Phys. Lett. 131 B ( 1983 ) 165.
[ 16 ] J.B. Kogut and D.K. Sinclair, Pressures, energy densities and chiral condensates in SU(3) lattice gauge theory with fermions, Illinois preprint ILL-(TH)-90-7 dynamical
(1990).