1: \section{RG-flows for compact groups: CFT approach}
2: \label{sec:cft}
3:
4: In order to make contact with genuine \textsc{cft} techniques, we must
5: identify the relevant operators which are responsible for the $(\h,
6: \hb)$ deformation of the $G \times H$ original \textsc{wzw} model ($H
7: = U(1)^{\rank G}$). At lowest approximation, all we need is their
8: conformal dimensions in the unperturbed theory.
9:
10: Following~\cite{Zamolodchikov:1986gt},
11: let ${\cal L}_0$ be the unperturbed (conformal) action and $ \mathcal{O}_i$
12: the operators of conformal dimension $\Delta_i$. Consider the perturbed model,
13: with Lagrangian
14: \begin{equation}
15: {\cal L} = {\cal L}_0 + g^i \mathcal{O}_i\ .
16: \end{equation}
17: \begin{comment}where $S_0$ is the initial -- unperturbed -- action, $g^i$ are
18: coupling constants and $\mathcal{O}_i$ the added operators. It is then
19: well known that the tree-level\footnote{In this framework the
20: deformations with respect to the \textsc{wzw} model are treated as
21: classical effects.}
22: \end{comment}
23: The tree-level beta-functions read:
24: \begin{equation}
25: \label{eq:cft-beta}
26: \beta^i (g) = (\Delta_i-1)g^i\ ,
27: \end{equation}
28: where $g^i$ is supposed to be small, for the perturbative
29: expansion of $\beta^i$ to hold\footnote{One should be very careful in the
30: choice of signs in these formulae. In~\cite{Zamolodchikov:1986gt} the time
31: variable, in fact, describes the evolution of the system towards the infrared
32: and as such it is opposite with respect to the $t = \log \mu $ convention
33: that we used in the previous section (as in \cite{Witten:1983ar}).}.
34:
35: The $G \times H$ primary operator we need can be written as follows:
36: \begin{equation}
37: \mathcal{O} = \sum_{\ss{a,b}} \braket{ t^{\ss{a}} g t^{\ss{b}} g^{-1}}
38: \braket{ t^{\ss{a}} \del g g^{-1} } \braket{ t^{\ss{b}} g^{-1} \db g}
39: = \sum_{\ss{a,b}} \Phi^{\ss{ab}} J^{\ss{a}} \bar J^{\ss{b}}\ ,
40: \end{equation}
41: where $\Phi^{\ss{ab}}$ is a primary field transforming in the adjoint
42: representation of the left and right groups $G$. As such, the total conformal
43: dimensions are \cite{Knizhnik:1984nr}
44: \begin{equation}
45: \Delta = \bar \Delta = 1 + \frac{g^\ast}{g^\ast + k}\ ,
46: \end{equation}
47: where $g^\ast$ is the dual Coxeter number and as such the operator is
48: irrelevant (in the infrared).
49:
50: Specializing this general construction to our case we find that the
51: action for the fields in Eq.~(\ref{eq:4d-2par-su2}) is:
52: \begin{equation}
53: \label{eq:cft-action}
54: {\cal L} = \frac{k}{4 \pi} \left\{ {\cal L}_0 + \left( \frac{\h}{\hb}- 1\right)
55: \sum_{\ss{a,b}} \Phi^{\ss{ab}} J^{\ss{a}} \bar J^{\ss{b}}
56: + \frac{\h}{\hb} \left( \h + \hb \right) \sum_i J^{a_i} \bar J^i
57: + \frac{\h}{\hb} \left( \h - \hb \right) \sum_{i,\ss{a}} J^i \Phi^{a_i \ss{a}}
58: \bar J^{\ss{a}} \right\} \ ,
59: \end{equation}
60: where $\ss{a}$ runs over all currents, $i$ over the internal currents
61: in $H$ and $J^{a_i}$ is the \textsc{wzw} current of the Cartan
62: subalgebra of $G$ coupled to the internal $\bar J^i$. The extra terms
63: can be interpreted as combinations of operators in the $G \times
64: H$ model. The beta-functions are thus computed following
65: Eq.~(\ref{eq:cft-beta}) with the coupling $g=h/\hb-1$. We obtain
66: \begin{equation} \label{eq:hhb-beta}
67: %CFT:\hskip 1 cm
68: \Deriv{}{t} \left( \frac{\h}{\hb} -1 \right)
69: \Bigg|_{\hb = \h} = \frac{ g^\ast}{g^\ast + k }\left({h\over\hb}-1\right) + \cdots =
70: \left(\frac{g^\ast}{k } - \frac{g^{\ast 2}}{k^2} \right)
71: \left({h\over\hb}-1\right) + \cdots\ ,
72: \end{equation}
73: where the dots after the first equality denote higher order terms
74: in the $(h/\hb-1)$-expansion and after the second equality, in
75: addition to that, higher order terms in the $1/k$-expansion.
76: This result is the same as the one in \cite{Knizhnik:1984nr} since, as we have mentioned,
77: there is the a local field redefinition that maps this model at the conformal point to
78: the $G\times H$ \textsc{wzw} model.
79: The above result
80: is to be compared with the results following from
81: Eq.~(\ref{eq:hhb-beta}) when they are expanded around
82: $\h=\hb$. We obtain:
83: \begin{equation} \label{eq:hhb-beta1}
84: %\sigma-model :\hskip .5 cm
85: \Deriv{}{t} \left( \frac{\h}{\hb} -1 \right)
86: \Bigg|_{\hb = \h} = \left(\frac{g^\ast}{k } -
87: \frac{g^{\ast 2}}{k^2}\right)\left({h\over \hb}-1\right)+
88: {1\over 2}\left(- {g^\ast\over k} + 7 {{g^\ast}^2\over k^2}\right)\left({h\over \hb}-1\right)^2
89: +\cdots\ .
90: \end{equation}
91: We see that these results agree to first order in the coupling $h/\hb-1$.
92: % But there's more: as pointed out before the conformal
93: %model ($\hb = \h$) is exact because it coincides with a $G \times H$
94: %\textsc{wzw} model after a suitable field redefinition for any value of $\h$.
95: %As a consequence the equations remain valid for any finite $\h$. This is
96: %reassuring both for the validity of the geometrical approach\footnote{There is
97: % no doubt on the method itself. It could simply fail to describe the desired
98: % phenomenon due to an inappropriate ansatz for the off-criticality excursion
99: % in parameter space.} and for the conclusions on the stability picture of the
100: %models under consideration.
101:
102: The extra information that we obtain from this calculation is about
103: the interpretation for the two-loop beta-function we described
104: in the previous section. The one-loop
105: corrections to~(\ref{eq:cft-beta}) are of the form $ C_{ijk} \,
106: g^i \, g^j$, where $ C_{ijk}$ are related to the three-point function
107: of the unperturbed theory \cite{Zamolodchikov:1986gt}.
108: This coefficient is a measure of the
109: dimension of the operator $\mathcal{O}_i$ in the theory perturbed by
110: the set of all operators. Such a computation goes beyond
111: the scope of the present note. Nevertheless,
112: (\ref{eq:hhb-beta1}) predicts the coefficient of the term $(h/\hb-1)^2$ to second
113: order in the $1/k$-expansion and it seems that such a computation is feasible from
114: the \textsc{cft} viewpoint at least as a series expansion in $1/k$.
115:
116: \begin{comment}
117: Note that the perturbation at hand is described from the one side as
118: a one-loop (quantum) effect in terms of target space
119: renormalization and from the other side as a tree-level
120: (classical) effect from the \textsc{cft} point of view. This brings
121: us to the natural question: what is the one-loop correction to the
122: beta-functions in the present \textsc{cft} framework, and how does
123: this compare to the equivalent order in the geometrical approach
124: (two-loop, see Appendix \ref{sec:two-loop-equations} and Sec.
125: \ref{sec:renormalization} for comments)? This must be crucial for
126: clarifying the issue of new fixed points.
127:
128:
129: An important point must be made here. The dimensions for the operators
130: in Eq.~(\ref{eq:cft-action}) are calculated in the $SU(2) \times U(1)
131: $ \textsc{wzw} model framework. This is clearly not enough if we want
132: a result extending the $\h \to 0$ validity range for
133: Eq.~(\ref{eq:hhb-beta}), although -- as we've already remarked -- the
134: results are $\h$ independent. Nevertheless a complete \textsc{cft}
135: description for the squashed \textsc{wzw} models would be required for
136: a complete analysis: a first step in this direction will be made in
137: Sec.~\ref{sec:poisson-brackets} where we discuss the Poisson brackets among the currents
138: generating the symmetries of the model.
139:
140:
141:
142:
143: \texttt{[Add more comments on the finte-$\h$ validity.]}
144:
145:
146: As already mentioned, our results of Secs.~\ref{sec:compact-groups}
147: and \ref{sec:cft} can be generalized straightforwardly to any compact
148: group (and some actually are already general enough). We will not
149: expand more in this direction but move instead to the less familiar
150: and richer situation of non-compact groups.
151:
152:
153: \end{comment}
154:
155:
156:
157: %%% Local Variables:
158: %%% mode: latex
159: %%% TeX-master: "Transmutation"
160: %%% End:
161: