hep-lat0510101/LS.tex
1: % ****** Start of file apssamp.tex ******
2: %
3: %   This file is part of the APS files in the REVTeX 4 distribution.
4: %   Version 4.0 of REVTeX, August 2001
5: %
6: %   Copyright (c) 2001 The American Physical Society.
7: %
8: %   See the REVTeX 4 README file for restrictions and more information.
9: %
10: % TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
11: % as well as the rest of the prerequisites for REVTeX 4.0
12: %
13: % See the REVTeX 4 README file
14: % It also requires running BibTeX. The commands are as follows:
15: %
16: %  1)  latex apssamp.tex
17: %  2)  bibtex apssamp
18: %  3)  latex apssamp.tex
19: %  4)  latex apssamp.tex
20: %
21: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
22: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
23: 
24: % Some other (several out of many) possibilities
25: %\documentclass[preprint,aps]{revtex4}
26: %\documentclass[preprint,aps,draft]{revtex4}
27: %\documentclass[prb]{revtex4}% Physical Review B
28: 
29: \usepackage{graphicx}% Include figure files
30: \usepackage{dcolumn}% Align table columns on decimal point
31: \def\bea{\begin{equation} }  
32: \def\eea{\end{equation} }  
33: \usepackage{bm}% bold math
34: 
35: %\nofiles
36: 
37: \begin{document}
38: 
39: \title{Loop States in Lattice Gauge Theories}% Manuscript Title:\\with Forced Linebreak}% Force line breaks with \\
40: 
41: \author{Manu Mathur}
42: %\altaffiliation[TIFR, Mumbai]{%S.N. Bose National Centre for Basic Sciences\\
43: %JD Block Physics Department, XYZ University.}
44: %Lines break automatically or can be forced with \\
45: %\author{Second Author}%
46: %\email{Second.Author@institution.edu}
47: \affiliation{% S. N. Bose National Centre for Basic Sciences} 
48: %Authors' institution and/or address\\
49: Tata Institute of Fundamental Research \\
50: Homi Bhabha Road, Mumbai 400 005, India \\ \\  
51: S. N. Bose National Centre for Basic 
52: Sciences\footnote{Permanent address, E.Mail: manu@bose.res.in}  \\ JD Block, Sector III, 
53: Salt Lake, Calcutta, India  
54: %% \textbackslash\textbackslash
55: }%
56: %%\author{Charlie Author}
57: %%%% \homepage{http://www.Second.institution.edu/~Charlie.Author}
58: %%\affiliation{
59: %%Second institution and/or address\\
60: %%This line break forced% with \\
61: %%}%
62: 
63: \date{\today}% It is always \today, today,
64:              %  but any date may be explicitly specified
65: 
66: \begin{abstract}
67: We solve the Gauss law as well as the corresponding Mandelstam
68: constraints of $(d+1)$ dimensional SU(2) lattice gauge theory
69: in terms of harmonic oscillator prepotentials. This enables us
70: to explicitly construct a complete orthonormal and manifestly
71: gauge invariant basis in the physical Hilbert space. Further, we show
72: that this gauge invariant description represents networks
73: of unoriented loops carrying certain non-negative abelian fluxes 
74: created by the harmonic oscillator prepotentials.
75: The loop network is characterized by $3(d-1)$ gauge invariant
76: integers at every lattice site which is the number of physical degrees
77: of freedom. Time evolution involves local fluctuations of these loops. 
78: The loop Hamiltonian is derived. The generalization to SU(N) gauge group 
79: is discussed.
80: \end{abstract}
81: 
82: \pacs{11.15. Ha }
83: 
84: \maketitle
85: 
86: \section{Introduction} 
87: 
88: The idea that gauge theories should be formulated completely in terms of 
89: loops in space carrying electric fluxes is quite old, appealing and
90: has long history \cite{mans}. It is widely believed that 
91: QCD  written in terms of such gauge invariant colorless loops, instead of 
92: colored quarks and gluons, is better suited to study non-perturbative long 
93: distance physics like color confinement. Further, since the introduction 
94: of $SL(2,C)$ Yang Mills connections as the basic variables for 
95: gravity \cite{ash}, the loop formalism goes beyond the color invariant 
96: description of gauge theories and has a much wider reach. To this end, the 
97: Wilson loop approach, though geometrical and manifestly gauge invariant, 
98: suffers from the serious problem of over completeness due to Mandelstam 
99: constraints \cite{mans,rst}. Therefore, a most economical as well as 
100: complete description of gauge theories in d dimension in terms of gauge 
101: invariant loop 
102: states is an important issue and the subject of the present work. 
103: One would like to solve the Mandelstam constraints in the loop basis  
104: and study the loop dynamics without making any approximations or 
105: taking any particular limits \cite{rst}. We use the recently proposed 
106: prepotential 
107: formulation of lattice gauge theories \cite{manu} to explicitly construct an 
108: orthonormal and manifestly gauge invariant basis in the physical 
109: Hilbert space and thus solve the SU(2) Gauss law as well as the associated 
110: Mandelstam constraints. The ideas can be generalized to SU(N) group 
111: and discussed in the last section. In the SU(2) case, we show that the basis 
112: vectors describe networks of loops which carry positive integer abelian fluxes
113: created by the prepotential operators. Further, the action of the 
114: Hamiltonian on the loop basis too has simple interpretation of counting, 
115: creating and destroying the above abelian flux lines on the links. 
116: The loop network is characterized by $3(d-1)$ angular momentum quantum numbers 
117: at every lattice site which is the number of the physical degrees of freedom 
118: of the SU(2) theory. Therefore, this loop state description is  also a duality 
119: transformation \cite{sharat1,rob}  where the effect of compactness of the 
120: gauge group is contained in the discrete angular momentum quantum numbers 
121: labeling the loop states. In the simpler context of compact (2+1) and (3+1) 
122: U(1) gauge theories such duality transformations are known to isolate the 
123: topological magnetic monopole degrees of freedom leading to confinement 
124: \cite{pol}. The plan of the paper is as follows. After a brief introduction, 
125: we first construct the 
126: loop states in d=2 and study their dynamics. This keeps the discussion simple 
127: and also illustrates 
128: all the ideas involved. The corresponding analysis and results in arbitrary 
129: d dimension is then obvious and done next.  The generalization to SU(N) case 
130: is discussed at the end. 
131: 
132: We start with SU(2) lattice gauge theory in (d+1) dimension. The Hamiltonian 
133: is \cite{kogut}:  
134: \bea 
135: H=\sum_{n,i}tr E(n,i)^{2} +K\sum_{plaquettes}tr \Big(U_{plaquette} + h.c\Big).     
136: \label{ham}   
137: \eea 
138: where K is the coupling constant. The index n labels the site of a d-dimensional
139: spatial lattice and $i,j (=1,2,...d)$ denote the unit vectors along the links.
140: Each link $(n, i)$ is associated with a symmetric top whose configuration,
141: i.e the rotation matrix from space fixed to body fixed frame, is given by the
142: operator valued SU(2) matrix $U(n,i)$. The angular momenta with respect 
143: to space fixed and the body fixed  frames are denoted by 
144: $E_{L}^{a}(n,i)$ and $E_{R}^{a}(n+i,i)$. More explicitly, 
145: $E_{L}(n,i)$ and $E_{R}(n+i,i)$ generate the gauge transformations 
146: at the lattice sites  n  and  n+i  respectively. 
147: They commute with each other and satisfy: 
148: $E_{L}(n,i).E_{L}(n,i) = E_{R}(n+i,i).E_{R}(n+i,i) \equiv E(n,i).E(n,i)$ 
149: as the total angular momentum is same in both the frames. 
150: Therefore, a complete basis at every 
151: link (n,i) is given by $|j(n,i),m(n,i),\tilde{m}(n,i)>$ where $j(n,i),m(n,i),
152: \tilde{m}(n,i)$  are the eigenvalues of $E(n,i).E(n,i), E_{L}^{3}(n,i)$ 
153: and $E_{R}^{3}(n+i,i)$ respectively. We now exploit the Schwinger boson 
154: representation of the angular momentum algebra \cite{schwinger} to 
155: define harmonic oscillator prepotentials on the links: 
156: \begin{eqnarray}
157: \label{sb}
158: E_{L}^{a}(n,i) &  \equiv &  a^{\dagger}(n,i)\frac{\sigma^{a}i}{2} a(n,i); \\ 
159: E_{R}^{a}(n+i,i) & \equiv & b^{\dagger}(n+i,i)\frac{\sigma^{a}}{2}b(n+i,i). 
160: \nonumber 
161: \end{eqnarray}
162: The gauge transformation properties of the angular momenta, 
163: $E_{L}(n, i) \rightarrow \Lambda(n) E_{L}(n, i) \Lambda^{\dagger}(n)$ and  
164: $E_{R}(n+i, i) \rightarrow \Lambda(n+i) E_{R}(n+i, i) \Lambda^{\dagger}(n+i)$, 
165: imply that the Schwinger bosons belong to the fundamental 
166: representations of the the gauge group, i.e: 
167: \bea
168: a^{\dagger}_{\alpha}(n,i)\rightarrow\Lambda(n)_{\alpha\beta} 
169: a^{\dagger}_{\beta}(n,i); b^{\dagger}_{\alpha}(n,i)\rightarrow\Lambda(n)_{\alpha\beta} b^{\dagger}_{\beta}(n,i). 
170: \label{gt3}
171: \eea
172: Therefore, $a^{\dagger}_{\alpha}(n,i)$ (left oscillator) and $b^{\dagger}_{\alpha}(n+i,i)$ 
173: (right oscillator) create spin half charges  
174: at left and right ends of the link (n,i) respectively. The total angular momentum 
175: being same in both the frames implies: 
176: \bea
177: a^{\dagger}(n,i).a(n,i) = b^{\dagger}(n+i,i).b(n+i,i)  \equiv N(n,i).  
178: \label{consho} 
179: \eea 
180: Thus,  besides SU(2) gauge invariance (\ref{gt3}) at every lattice site, we get 
181: an addition abelian gauge invariance on every lattice link:  
182: \begin{eqnarray} 
183: \label{u1}
184: a^{\dagger}_{\alpha}(n,i) & \rightarrow & \left(expi\theta(n,i)\right) ~ 
185: a^{\dagger}_{\alpha}(n,i);   \\ 
186: b^{\dagger}_{\alpha}(n+i,i) & \rightarrow & \left(exp-i\theta(n,i)\right) ~
187: b^{\dagger}_{\alpha}(n+i,i) \nonumber.   
188: \end{eqnarray} 
189: So we have gone from the electric field (or angular momentum), link operator 
190: description of  Kogut-Susskind Hamiltonian (\ref{ham}) to an equivalent description 
191: of SU(2) lattice gauge theory which is in terms of harmonic oscillator prepotentials
192: with $SU(2) \otimes U(1)$ gauge invariance. With the simple $SU(2) \otimes U(1)$  
193: gauge transformations (\ref{gt3}) and (\ref{u1}), we are well equipped to construct 
194: explicitly a manifestly gauge invariant and orthonormal loop basis. 
195:  
196: \section{The Loop States in d=2} 
197: 
198: \begin{figure}[t]
199: \begin{center}
200: \includegraphics[width=0.3\textwidth,height=0.3\textwidth]
201: {FIG1.eps}
202: \end{center}
203: \vspace{-5mm}
204: \caption{A graphical representation of the  $SU(2) \otimes U(1)$  
205: gauge transformations in d=2 at site n with $j_{1}=j_{2}=j_{3}=j_{4}=1$. 
206: The abelian flux lines graphically solves the abelian Gauss law 
207: (\ref{consho}). The (Young tableau) boxes at the ends of abelian flux 
208: lines represent the fundamental (spin 1/2) representation of SU(2) 
209: which acts at the lattice site n.}
210: \label{fig:figure1}
211: \end{figure}
212: The U(1) gauge invariance (\ref{u1}) and it's Gauss law (\ref{consho}) 
213: simply state that the number of the left ($a^{\dagger}(ni)$) and the right 
214: ($b^{\dagger}(n+i,i)$) oscillators is the same on any link. We denote this 
215: integer number by $2j(n,i)$. The abelian Gauss law constraints are solved 
216: easily by drawing  $2j(n,i)$ lines on every link (n,i). 
217: Each of these $2j(n,i)$ lines represents the U(1) charge (= $+ 1$) of 
218: $a^{\dagger}(n,i)$ (see (\ref{u1})) and henceforth will be called abelian charge 
219: or abelian flux line. To illustrate, a simple example with 
220: $j(n,1)=j(n,2)=j(n-1,1)=j(n-2,2) =1$ is shown in Figure 1.  With U(1) Gauss 
221: law satisfied, we now deal with SU(2) gauge invariance. 
222: Under SU(2) gauge transformations (\ref{gt3}), we note that 
223: $a^{\dagger}(n,i)$ [located at the starting point of the link 
224: (n,i)] and $b^{\dagger}(n,i)$ [located at the end point of the 
225: link (n-i,i)] transform together by $\Lambda(n)$. Therefore, it is convenient 
226: to group them together and define $a^{\dagger}[n,i]$ with i=1,2,3,4. 
227: More explicitly, $a^{\dagger}[n,1] \equiv a^{\dagger}(n,1), 
228: a^{\dagger}[n,2] \equiv a^{\dagger}(n,2), a^{\dagger}[n,3] 
229: \equiv b^{\dagger}(n,1), a^{\dagger}[n,4] \equiv b^{\dagger}(n,2)$. 
230: Therefore, each of the four $a^{\dagger}[n,i]$ can be represented by 
231: a Young tableau (YT) box belonging to the SU(2) group which acts 
232: at the site n. 
233: Thus, to get SU(2) gauge invariance,  we have to construct all possible 
234: spin singlets out of $\left(\sum_{i=1}^{2d=4} 2j[n,i]\right)$  YT boxes. 
235: This is a simple problem: all possible spin zero  operators are of the form 
236: $a^{\dagger}[n,i].\tilde{a}^{\dagger}[n,j] \equiv 
237: \epsilon_{\alpha\beta}a^{\dagger}_{\alpha}[n,i]
238: a^{\dagger}_{\beta}[n,j]$, where $\epsilon_{\alpha\beta}$ is the completely 
239: antisymmetric tensor and corresponds to putting two boxes of type i and j 
240: in a vertical column.  In Figure 1, we represent this by linking 
241: a line of type i with a line of type j ($i \neq j$) (see Figure 2).  
242: Thus for SU(2) gauge invariance all abelian 
243: flux lines must be mutually linked and no self linking is allowed. 
244: Therefore, the necessary condition on the number of abelian flux lines 
245: on the links [n,i] to give SU(2) gauge invariant state(s) at site n is: 
246: \begin{eqnarray} 
247: 2j(n,i) = \sum_{j\neq i} l_{ij},~~ l_{ij}=l_{ji},~~ l_{ij} \in {\cal Z_{+}},  
248: \label{part} 
249: \end{eqnarray}
250: where ${\cal Z_{+}}$ denotes the set of all non-negative integers and 
251: $l_{ij}$ are the linking numbers amongst $i$ and $j$ types of abelian 
252: flux lines. The partition (\ref{part}) represents the manifestly SU(2) 
253: gauge invariant state: 
254: \begin{eqnarray}
255: |\vec{l}> = \prod_{{}^{{i},{j}=1}_{~{j} > {i}}}^{2d = 4} 
256: \left(a^{\dagger}[i].\tilde{a}^{\dagger}[j]\right)^{l_{{i}j}}
257: |0> . 
258: \label{giv2} 
259: \end{eqnarray} 
260: The pattern $|l_{12}(n),l_{13}(n),l_{14}(n),l_{23}(n),l_{24}(n),l_{34}(n)>$ 
261: will be used to characterize the states in (\ref{giv2}).  
262: Thus given any network of loops on the lattice we can construct a 
263: manifestly $SU(2) \otimes U(1)$ gauge invariant basis (\ref{giv2}) 
264: characterized by $d(2d-1)$ integer quantum numbers at every lattice 
265: site. The disadvantage of the basis (\ref{giv2}) is that (like Wilson 
266: Loop basis) it is not orthonormal and it is over complete. 
267: To show this, we consider three distinct basis vectors contained 
268: in the set (\ref{giv2}): $|\vec{l}_1> = |100001>,  
269: |\vec{l}_2> = |010010>$ and $|\vec{l}_3> =|001100>$. We find 
270: that they are linearly related: $|\vec{l}_1> = |\vec{l}_2> - 
271: |\vec{l}_3>$ due to the identity: 
272: \begin{eqnarray}   
273: \left(a^{\dagger}[1].\tilde{a}^{\dagger}[2]\right) 
274: \left(a^{\dagger}[3].\tilde{a}^{\dagger}[4]\right)  
275: \equiv  \left(a^{\dagger}[1].\tilde{a}^{\dagger}[3]\right)
276: \left(a^{\dagger}[2].\tilde{a}^{\dagger}[4]\right) \nonumber \\ 
277: - \left(a^{\dagger}[1].\tilde{a}^{\dagger}[4]\right)
278: \left(a^{\dagger}[2].\tilde{a}^{\dagger}[3]\right).  ~~~ 
279: \label{mandi}
280: \end{eqnarray} 
281: Infact, the identity (\ref{mandi}) is the basic SU(2) Mandelstam 
282: identity written in terms of prepotentials. To solve the problem 
283: of over-completeness, we notice that the states (\ref{giv2}) obtained 
284: by different possible contractions of the abelian flux lines are all 
285: characterized by $|J[n,i].J[n,i] = j[n,i](j[n,i]+1), J 
286: \equiv J^{a}_{total} =0>$. However, the intermediate angular momentum 
287: quantum number labels are missing \cite{sharat1}. 
288: Therefore, we choose the following angular momentum addition scheme: 
289: $J[1]+J[2] \rightarrow J[12]+J[3] \rightarrow J[(12)3]+ J[4]=J  \equiv 0$ 
290: and label the the common eigenvectors by the corresponding eigenvalues: 
291: $|j_{1},j_{2},j_{12},j_{3},j_{123}=j_{4},j=j_{total}=0> \equiv 
292: |j_{1},j_{2},j_{12},j_{3},j_{4}>$. Thus, we get the (missing) operator 
293: $(J[n,1]+J[n,2])^{2}$ in this scheme which is  yet to be diagonalized in the basis 
294: (\ref{giv2}) with eigenvalue $j_{12}$. This diagonalization problem is again 
295: simple: after linking $l_{12}$ boxes from $2j_{1}$ YT boxes on the 
296: link (n,1) with $l_{12}$ boxes from $2j_{2}$ YT boxes on the link (n,2), 
297: we should be left with $2j_{12}$ boxes which are not linked (and symmetrized). 
298: Therefore, $l_{12}=j_{1}+j_{2}-j _{12}$. As total angular 
299: momentum is zero this also fixes $l_{34}=j_{3}+j_{4}-j _{12}$. 
300: This, in the example of Figure 1, is illustrated in Figure 2.  
301: The final orthonormal and manifestly SU(2) gauge invariant states are: 
302: \begin{eqnarray}
303: \label{fr}
304: |j_{1},j_{2},j_{12},(j_3),j_{123}=(j_{4})>  \equiv |j_1,j_2,j_{12} \rangle 
305: ~~~~ ~~~~~~\\
306: =  N(j) \sum_{{}^{l_{13},l_{14}}_{l_{23},l_{24}}}\hspace{-0.05cm}
307: {}^{{}^\prime} 
308: \prod_{{}^{i,j}_{i < j}} \big((l_{ij})!\big)^{-1} \big(a^{\dagger}[n,i].
309: \tilde{a}^{\dagger}[n,j]\big)^{l_{ij}} | 0 \rangle ~~~ \nonumber 
310: \end{eqnarray}
311: In (\ref{fr}), the prime over the summation means that the 
312: linking numbers $l_{13},l_{14},l_{23},l_{24}$ are summed over all possible values which 
313: are consistent with (\ref{part}). More simply, in the loop network language, the 
314: summations are over all possible contractions of the abelian flux lines keeping $l_{12}$ 
315: contractions fixed \footnote{Our derivation of (\ref{fr}) in arbitrary d dimension 
316: (see section III) is by exploiting the properties of SU(2) coherent states and will 
317: be published elsewhere \cite{manu2}.}. The normalization constant 
318: $N(j) = N\big(j_1,j_2,j_{12}\big)N\big(j_{12},j_{3},j_{123}\big)N
319: \big(j_{123}(=j_4)),j_4,0\big)$, where  
320: $N(j_1,j_2,j_3) = \big((j_1-j_2+j_3)!(-j_1+j_2+j_3)!(j_1+j_2-j_3)!\big)^{\frac{1}{2}} 
321: \big(\frac{(2j_3+1)}{(j_1+j_2+j_3+1)!}\big)^{\frac{1}{2}}$.   
322: In (\ref{fr}), on the left hand site $j_3$ and $j_4$ are within 
323: brackets as they are associated with the previous sites due to 
324: the U(1) Gauss law (\ref{u1}). 
325: \begin{figure}[t]
326: \begin{center}
327: \includegraphics[width=0.39\textwidth,height=0.13\textwidth]
328: {FIG2.eps}
329: \end{center}
330: \vspace{-5mm}
331: \caption{The loop state at site n $|j_{1}=j_{2}=j_{3}=j_{4}=j_{12} =1>$ 
332: from the orthonormal set (\ref{fr}) corresponding to Figure 1. The other  
333: state $|j_{1}=j_{2}=j_{3}=j_{4}=1, j_{12} =0>$ constructed from Figure 1 
334: and orthogonal to $|j_{1}=j_{2}=j_{3}=j_{4}=j_{12} =1>$ is 
335: represented by choosing the linking numbers $l_{12}=l_{34}=2$.}
336: \label{fig:figure2}
337: \end{figure}
338: Thus in d=2 the $SU(2)\otimes U(1)$ the gauge invariant 
339: loop network basis is characterized  by three physical 
340: (gauge invariant) quantum numbers per 
341: lattice site. We summarize the 
342: results obtained so far: In d = 2, the complete set of orthonormal gauge invariant 
343: states is isomorphic to the set of all possible loops which are labeled by 
344: the number of forward loop lines $j(n,1), j(n,2)$ and the linking number $l_{12}(n)$
345: at every lattice site.  The explicit construction is given by (\ref{fr}). 
346: 
347: \subsection{The Loop Space Dynamics} 
348: 
349: The Hamiltonian (\ref{ham}) has a very simple interpretation 
350: in the dual loop basis (\ref{fr}). The electric field term 
351: is now like potential energy term which simply counts the number 
352: of abelian flux lines. It's contribution to the energy is: 
353: $\sum_{links(l)} j(l)\left(j(l) +1\right)$. The plaquette term 
354: in (\ref{ham}) too has a simple meaning: it creates or 
355: annihilates the abelian flux lines on the plaquette. 
356: This can be seen by writing the link operator in terms of 
357: prepotentials: 
358: \begin{eqnarray}
359: U({\it l})_{\alpha\beta}  &=& F({\it l})  \big(a^{\dagger}_{\alpha}(l)
360: \tilde{b}^{\dagger}_{\beta}(l) +\tilde{a}_{\alpha}(l) 
361: {b}_{\beta}(l)\big) F(l) \nonumber \\
362: & \equiv &  U^{+}_{\alpha\beta}({\it l}) + U^{-}_{\alpha\beta}({\it l}).
363: \label{dhh}
364: \end{eqnarray}
365: Above, $F({\it l}) = 
366: \frac{1}{\sqrt{[a^{\dagger}({\it l}).a({\it l})+1]}}$.  
367: The transformation to prepotentials (\ref{dhh}) is obvious from the 
368: $SU(2) \otimes U(1)$ gauge transformations (\ref{gt3}) and (\ref{u1})
369: respectively. Looking at (\ref{dhh}) we realize that the operator 
370: $U^{+}({\it l})$ $(U^{-}({\it l}))$ creates (destroys) an abelian flux 
371: line on the link $({\it l})$ like in the case of compact QED 
372: (see also \cite{sharat2}).  We now compute the matrix elements of the 
373: Hamiltonian in the loop basis 
374: (\ref{fr}) for d=2. For convenience, we denote the four corners: 
375: n,n+1,n+1+2,n+2 of the plaquette 
376: located at n by a,b,c,d respectively and the associated loop 
377: basis vector as 
378: $|j_{abcd}> \equiv 
379: \prod_{x=a,b,c,d} \otimes |j_{1}^{x},j_{2}^{x},j_{12}^{x}>$. 
380: We consider the following $SU(2) \otimes U(1)$ gauge invariant part of 
381: the plaquette term in (\ref{ham}):   
382: $U_{plaq} = \left(a^{\dagger}[1].\tilde{a}^{\dagger}[2]\right)_{a}   
383: \left(a^{\dagger}[2].\tilde{a}^{\dagger}[3]\right)_{b}   
384: \left(a^{\dagger}[3].\tilde{a}^{\dagger}[4]\right)_{c}  
385: \left(a^{\dagger}[1].\tilde{a}^{\dagger}[4]\right)_{d}$ + h.c.   
386: The matrix elements are \cite{manu2}: 
387: \begin{eqnarray} 
388: &&<\bar{j}_{abcd}|U_{plaq}|j_{abcd}> =  
389: \Big(N_{+}~\delta_{\bar{j}_{1}^{a},{j}_{1}^{a}+\frac{1}{2}}  
390: \delta_{\bar{j}_{2}^{b},{j}_{2}^{b}+\frac{1}{2}}  
391: \delta_{\bar{j}_{1}^{d},{j}_{1}^{d}+\frac{1}{2}}  
392: \nonumber \\  
393: && \delta_{\bar{j}_{2}^{a},{j}_{2}^{a}+\frac{1}{2}} +    
394: N_{-}~\delta_{{j}_{1}^{a},{j}_{1}^{a}-\frac{1}{2}}  
395: \delta_{\bar{j}_{2}^{b},{j}_{2}^{b}-\frac{1}{2}}  
396: \delta_{\bar{j}_{1}^{d},{j}_{1}^{d}-\frac{1}{2}}  
397: \delta_{\bar{j}_{2}^{a},{j}_{2}^{a}-\frac{1}{2}}\Big) \nonumber \\ 
398: &&\left\{ \begin{array}{cccc}
399: j_{12}^{b} & \bar{j}_{12}^{b} & \frac{1}{2}  \\
400: \bar{j}_{1}^{a} & j_{1}^{a}  & j_{4}^{b}\\
401: \end{array} \right \} 
402: \left\{ \begin{array}{cccc}
403: j_{12}^{b} & \bar{j}_{12}^{b} & \frac{1}{2}  \\
404: \bar{j}_{2}^{b} & j_{2}^{b}  & j_{1}^{b}\\
405: \end{array} \right \} 
406: \left\{ \begin{array}{cccc}
407: j_{12}^{d} & \bar{j}_{12}^{d} & \frac{1}{2}  \\
408: \bar{j}_{1}^{d} & j_{1}^{d}  & j_{2}^{d}\\
409: \end{array} \right \}  \nonumber \\
410: && \left\{ \begin{array}{cccc}
411: j_{12}^{d} & \bar{j}_{12}^{d} & \frac{1}{2}  \\
412: \bar{j}_{2}^{a} & j_{2}^{a}  & j_{3}^{d}\\
413: \end{array} \right \}. ~~~
414: \label{me} 
415: \end{eqnarray} 
416: Above $N_{\pm}$ are the constants depending on the 
417: angular momentum quantum numbers on the plaquette 
418: $(abcd)$. The trivial $\delta$ functions over the 
419: quantum numbers which do not change are not shown.  
420: The 6-j symbols simply reflect the spin half 
421: nature of the prepotentials. The details will be given 
422: elsewhere \cite{manu2} (also see \cite{sharat1}).   
423: 
424: \section{The Loop States in d Dimension} 
425: 
426: It is easy to generalize d=2 construction of the previous section. 
427: We extend the angular momentum ladder and choose: $J[1]+J[2] \rightarrow J[12]+J[3] 
428: \rightarrow J[123] + ...\rightarrow J[12..2d-1]+J[2d] = J = 0$.  The states are now 
429: characterized as: $|j_1,j_2,..,j_d,(j_{d+1}),..,(j_{2d}),j_{12},j_{123},
430: ..j_{12...2d-1}=(j_{2d})> \equiv |j_1,j_2,..,j_d,j_{12},
431: j_{123},..,j_{12...2d-2}>$. Thus the loop network  
432: is labeled by $3(d-1)$ gauge invariant 
433: angular momentum quantum numbers at every lattice 
434: site which is the number of transverse physical 
435: degrees of the freedom of the gluons \cite{sharat1}. 
436: The states are again given by (\ref{fr}) with  
437: the constraints on the linking numbers which have to be 
438: generalized. 
439: The Young tableau arguments like in d=2 case, lead to:  
440: $l_{12}=j_{1}+j_{2}-j_{12},l_{13}+l_{23}=j_{12}+j_{3}-j_{123}, 
441: ......,l_{1,2d}+l_{2,2d}+.....l_{2d-1,2d} = 2j_{2d-1} 
442: \equiv 2j_{2d}$. 
443: Note that the last equation is an identity. 
444: As in the d=2 case, all possible contractions consistent with 
445: the above constraints and with the number of flux lines on the links 
446: (\ref{part}) are required to get the manifestly gauge invariant 
447: orthonormal loop basis. The dynamical matter fields  
448: are easy to incorporate, they will provide SU(2) charge sources and 
449: sinks to the abelian flux lines at their end points (see Figure 1) 
450: leading to additional color singlets. 
451: 
452: \section{SU(N) Lattice Gauge Theory} 
453: 
454: The SU(N) group has $(N-1)$ fundamental representations. Therefore, 
455: the defining equations for the SU(N) harmonic oscillator prepotentails 
456: \cite{manu1} on the links are: 
457: \begin{eqnarray}
458: E_{L}^{a}  \equiv  \sum_{r=1}^{(N-1)} a^{\dagger}[r]\frac{\lambda^{a}[r]}{2} a[r],  
459: E_{R}^{a}  \equiv   \sum_{r=1}^{(N-1)} b^{\dagger}[r]\frac{\lambda^{a}[r]}{2} b[r]  
460: \nonumber 
461: \end{eqnarray}
462: where the prepotential oscillators a and b are defined at the initial and the 
463: end point of the link $l$ respectively, the index r varies over the rank of the SU(N) 
464: group. Thus, the SU(N) lattice gauge theories in terms of prepotentials will 
465: have $SU(N) \otimes U(1)^{(N-1)}$ gauge invariance. This will lead to $(N-1)$ 
466: types of abelian flux lines in the SU(N) loop space. The role of $(N-1)$ 
467: abelian gauge groups in the confinement mechanism of SU(N) gauge theories 
468: has been emphasized by `t Hooft through the idea of abelian projection \cite{hooft}. 
469: It is interesting to imagine the background SU(2) quark anti-quark pair 
470: located at two different lattice sites in the prepotential formulation. 
471: The SU(2) fluxes will be neutralized {\it locally} but the abelian 
472: gauge invariance will demand the formation of an abelian flux line (string)
473: between quark anti-quark pair leading to color confinement in the strong 
474: coupling limit.  The  construction of SU(N) loop basis, the issue 
475: of color confinement and especially $N \rightarrow \infty$ limit are under 
476: investigation. 
477: 
478: I would like to thank Rajiv Gavai and Sourendu Gupta for the 
479: hospitality at TIFR, Mumbai where part of the paper was written.  
480: 
481: \begin{thebibliography}{99}
482: 
483: \bibitem{mans} S. Mandelstam, Phys. Rev. 175 (1968) 1580, A. M. Polyakov, 
484: Phys. Lett. {\bf B 82} (1979) 247, 
485: A. A. Migdal, Phys. Rep. {\bf 102} (1983) 199, 
486: S. Mandelstam, Phys. Rev. {\bf D 19} (1979) 2391,  
487: A.  M. Polyakov, Gauge Fields and Strings (Harwood, New York, 1987). 
488: \bibitem{ash} A. Ashtekar, Phys. Rev. Letts. {\bf 57} (1986) 2244, C. Rovelli, L. Smolin, 
489: Phys Rev. {\bf D 52} (1995) 5743.  
490: \bibitem{rst} Br\"ugmann B  1991 Phys. Rev. D {\bf 43} 566, Loll R 1992 Nucl. 
491: Phys B 368 121, Watson N J 1994 Phys. Letts. B 323 385, 
492: R. Gambini, Lorenzo Leal, Antoni Trias, Phys. Rev. {\bf D 39} (1989) 3127, 
493: Bartolo C,  Gambini R, Leal L 1989 Phys. Rev. {\bf D 39} 1756. 
494: \bibitem{manu} M. Mathur, to be published in Journal of Physics A.   
495: \bibitem{sharat1} R. Anishetty, H. S. Sharatchandra, Phys. Rev. Letts. {\bf 65} (1990) 81. 
496: H. S. Sharatchandra, Nuclear Physics  {\bf B 196} (1982) 62. 
497: \bibitem{sharat2} B. Gnanapragasam, H. S. Sharatchandra, Phys. Rev. {\bf D 45} 
498: (1992) R1010. 
499: \bibitem{pol} A. M. Polyakov, Phys. Letts. {\bf B 59}, (1975) 82, T. Banks, R. Myerson, 
500: J. Kogut, Nucl. Phys. {\bf B129} (1977) 473. 
501: \bibitem{kogut} J. Kogut, L Susskind, Phys Rev. {\bf D 10} (1974) 3468.  
502: \bibitem{schwinger} J. Schwinger 1952 D. Mattis, {\it The Theory of Magnetism} (Harper and Row, 1982). 
503: \bibitem{manu2} Manu Mathur, under preparation. 
504: \bibitem{rob} D. Robson, D. M. Weber, Z. Phys. {\bf C15} (1982), 199. \\ 
505: W. Furmanski, A. Kolawa, Nucl. Phys., {\bf B 291}, (1987) 594  
506: G. Burgio, R. De. Pietri, H. A. Morales-Tecotl, L. F. Urrutia, 
507: J. D. Vergara, Nucl. Phys. {\bf B 566}, (2000), 547.  
508: \bibitem{manu1} M. Mathur and D. Sen, J. Math. Phys. {\bf 42} (2001) 4181, 
509: M. Mathur, H. S. Mani, J. Math. Phys. {\bf 43} (2002) 5351.
510: \bibitem{hooft} G. `t Hooft,  Nucl.  Phys. {\bf B 190} [FS 3] (1981) 455.  
511: \end{thebibliography}
512: 
513: \end{document}
514: