On Density of State of Quantized Willmore Surface-A Way to Quantized Extrinsic String in R^3
Extracted source text ยท arXiv solv-int/9707007
\mag=\magstephalf
%\mag=\magstep1
\pageno=1
\input amstex
%\baselineskip = 1.0 true cm
\documentstyle{amsppt}
\TagsOnRight
\interlinepenalty=1000
\hsize=6.4truein
\voffset=23pt
\vsize =9.8truein
\advance\vsize by -\voffset
%\baselineskip = 1.0 true cm
\nologo
\NoBlackBoxes
\font\twobf=cmbx12
\define \ii{\roman i}
\define \ee{\roman e}
\define \dd{\roman d}
\define \cn{\roman {cn}}
\define \dn{\roman {dn}}
\define \sn{\roman {sn}}
\define \am{\roman {am}}
\define \XMKdV{X_{\roman{MKdV}}}
\define \Xeigen{X_{\roman{eigen}}}
\define \Xelas{X_{\roman{elas}}}
\define \sc{{s_c}}
\define \bsc{{\bar s_c}}
\define \ts{{\tilde s}}
\define \Tr{{\roman{Tr}}}
\define \diag{{\roman{diag}}}
\define \cls{{\roman{cls}}}
%\define \qcl{{\roman{qcl}}}
\define \qcl{{\roman{}}}
\define \reg{{\roman{reg}}}
\define \KP{{\roman{KP}}}
\define \Em{{\roman{Em}}}
\define \Ker{{\roman{Ker}}}
\define \period{{\roman{period}}}
\define \traj{{\roman{traj}}}
\define \GL{{\roman{GL}}}
\define \PSL{{\roman{PSL}}}
\define \SSp{{\roman{Sp}}}
\define \SS{{\Cal S}}
\define \CC{{\Cal C}}
\define \tp{{t^+}}
\define \tm{{t^-}}
\define \tpm{{t^\pm}}
\define \DD{/\!\!\!\!\Cal{D}}
\define \tvskip{\vskip 0.5 cm}
{\centerline{\bf{On Density of State of Quantized Willmore
Surface}}}
{\centerline{\bf{---A Way to Quantized Extrinsic String in
$\Bbb R^3$---}}}
\author
\endauthor
\affil
Shigeki MATSUTANI\\
2-4-11 Sairenji, Niihama, Ehime, 792 Japan \\
\endaffil
\endtopmatter
%\baselineskip= 0.8 true cm
\subheading{Abstract}
Recently I quantized an elastica with Bernoulli-Euler functional
in two-dimensional space using the modified KdV hierarchy.
In this article, I will quantize a Willmore surface,
or equivalently
a surface with the Polyakov extrinsic curvature
action, using the modified Novikov-Veselov (MNV) equation.
In other words, I show that the density of state of the
partition function for the quantized Willmore surface
is expressed by volume of
a subspace of the moduli of the MNV equation.
%\subheading{PACS numbers}:
%\baselineskip= 0.8 true cm
\document
%\newpage
%\baselineskip= 0.8 true cm
\tvskip
\centerline{\twobf \S 1. Introduction }
\tvskip
In series of works [1-7], I have considered the correspondence
between an immersed object
and the Dirac operator confined there. The Dirac operator confined
in an immersed object is uniquely determined by
the procedure which I proposed [1-4] and can be regarded as
the representation matrix of the symmetry of the immersed object
[1-7].
I had been studying it mainly on an elastica in a plane [1-6]
and showed that the Dirac operator confined in an elastica is
identified with the Lax operator of the
modified KdV equation while the mathematical deformation of the
elastica
obeys the modified KdV hierarchy [6].
By investigating other quantum equations [8-9],
I conjectured that such correspondence between the Dirac operator
and geometry
can be extended to higher dimensional immersed objects [2,3,4].
Couple years ago Konopelchenko [10,11] discovered that a conformal
surface $\SS$ immersed in three dimensional flat space $\Bbb R^3$
obeys the Dirac
equation, which I will call
Konopelchenko-Kenmotsu-Weierstrass-Enneper (KKWE) [10-15]
equation here,
$$
\partial f_1 = V f_2, \quad \bar \partial f_2= -V f_1,
\tag 1-1
$$
where
$$
V:= \frac{1}{2}\sqrt{\rho} H, \tag 1-2
$$
$H$ is the mean curvature of the surface $\SS$ parameterized
by complex $z$ and $\rho$ is the conformal metric induced from
$\Bbb R^3$.
The KKWE equation completely exhibits the immersed geometry as
the old Weierstrass-Enneper equation expresses the minimal
surface [10-15]. In ref.[16] I showed that it is identified with
the Dirac operator confined in the surface
$\SS$ and by quantizing the Dirac field I found that the quantized
symmetry of the Dirac operator is also in agreement with the
symmetry of the surface itself [17].
In other words, this KKWE equation is the equation which
I conjectured before [2,3] and I had been searching for.
Even though for more general surface, which is not
conformal, the KKWE type equation was discovered by Burgress
and Jensen [18] following my prescriptions [1],
their equation is not easy to deal with and I could not find
meaningful results.
However the KKWE equation is very useful to investigate the
immersed object and in terms of (1-1), Konopelchenko,
Taimanov and other Russian group found
non-trivial results related to the immersed surface [10-15,19,20].
By physical investigating the KKWE equation and its quantized
version, the Willmore functional [21,22] and the modified
Novikov-Veselov (MNV) equation
naturally appears [10-15,17,19,20]. The Willmore functional
is given as [21,22],
$$
W = \int_{\Cal S} \dd \text{vol}\ H^2 , \tag 1-3
$$
where "$\dd \text{vol}$" is a volume form of the surface $\SS$.
The harmonic map associated with this functional has been studied in
the differential geometry [21,22].
On the other hands, Polyakov introduced an extrinsic curvature
action in the string theory
and the theory of two-dimensional gravity from renomalizability [23].
However his action is just the Willmore functional (1-3).
Thus his program recently was investigated by Carroll and
Konopelchenko [19]
and Grinevich and Schmidt [20] using KKWE equation (1-1).
The main theme of this article is to quantize the Willmore surface
but I emphasis that it means the study on the quantization of
the Polyakov extrinsic curvature action.
It should be noted that the elastica problem of $\Bbb R^2$
has very similar structure
of the Willmore surface problem of $\Bbb R^3$ [10-14].
Corresponding to the Willmore functional (2-20), there is
Bernoulli-Euler functional for an elastica [24],
$$
E=\int \dd q^1\ k^2, \tag 1-4
$$
where $k$ is a curvature of the elastica [24].
While the Willmore surface is related to
the modified Novikov-Veselov (MNV) equation,
the elastica is related to the modified KdV equation
[1-7,25-27].
Recently I exactly quantized the elastica of the Bernoulli-Euler
functional (1-4) preserving its local length [25].
Then I found that its moduli is completely represented
by the MKdV equation and closely related to the two-dimensional
quantum gravity [28-30].
The quantized elastica obeys the MKdV equation and at a
critical point, the Painlev\'e equation of the first kind appears
[25] while in the quantized two-dimensional gravity which is
defined at a critical point
of the discrete tiling model, there appears the Painlev\'e equation
of the first kind
with the KdV hierarchy [28-30].
In this article instead of the local length preserving,
I will impose that the surface
preserves its complex structure and
will quantize the Willmore functional.
Then I will show that the MNV hierarchy appears as
the quantized motion of a Willmore surface in the path integral.
The organization of this article is as follows.
Section 2 reviews the argument of
the quantized elastica following to
that in ref.[25]. In \S 3, I will quantize the Willmore surface
and then the density of states of the Willmore functional
is given as the volume of the MNV equation.
Section 4 gives the discussion for the results.
\tvskip
\centerline{\twobf \S 2. Quantization of Elastica}
\tvskip
I will denote by $\CC$ a shape of the elastica embedded
in a complex plane $\Bbb C $
and by $X(s)$ its affine vector [6]:
$$
S^1 \ni s \mapsto X(s) \in \CC \subset \Bbb C ,
\quad X(s+L)=X(s),
\tag 2-1
$$
where $L$ is the length of the elastica.
I will fix the metric of the curve $\CC$ induced from
the natural metric of $\Bbb C $; $\dd s = \sqrt{\dd X \dd \bar X}$.
The Frenet-Serret relations are expressed as [6,25-27]
$$
\psi_0:=\exp(\ii \phi/2)=\sqrt{\partial_s X}, \tag 2-2
$$
$$
\pmatrix \partial_s & v \\
v & -\partial_s \endpmatrix
\pmatrix \psi_0 \\ \ii \psi_0\endpmatrix
=0 ,
\quad v:= \frac{1}{2} k :=\frac{1}{2}\partial_s \phi,
\tag 2-3
$$
where $\phi$ is a real valued function of $s$ and $k$ is
the curvature of the curve $\CC$,
$\phi(s+L)=\phi(s)$ and $k(s+L)=k(s)$.
The energy functional of the elastica, which I will call
Bernoulli-Euler functional here [24], is given as
$$
E=\int_0^L \dd s\ k^2=4 \int_0^L \dd s\ v^2 ,\tag 2-4
$$
and shape of a static elastica is realized as its stationary point
satisfied with the boundary conditions.
I assume that the elastica does not stretch and preserves
its local infinitesimal length.
I will consider quantization of
a closed elastica whose local length preserves in the
quantization process.
The partition function of the elastica is given as [25],
$$
\Cal Z=\int D X \exp
\left(-\beta \int^L_0 \dd s\left[ k^2\right] \right) .
\tag 2-5
$$
Since there is trivial affine symmetry of the centroid of
the elastica and the partition function diverges,
I will regularize it,
$$
\Cal Z_{\reg}=\frac{\Cal Z}{\text{Vol}(\text{Aff})},
\tag 2-6
$$
where $\text{Vol}(\text{Aff})$ is the volume of the affine
transformation.
Next I will consider the condition of local length preserving.
In the path integral, I must pay attentions upon the
higher perturbations of $\epsilon$
to gain an exact result.
Hence I will assume that $X$ is parameterized by a parameter $t$
and the difference between perturbed affine vector $X_\epsilon$
and unperturbed one $X$
can be expressed by [6,25-27],
$$
X_\epsilon(s,t):=\ee^{\epsilon \partial_{t}}X_{\qcl}(s,t),
\quad \epsilon \partial_{t} X =X_\epsilon - X
+ \Cal O(\epsilon^2). \tag 2-7
$$
with the relation
$$
\partial_{t} X_{\qcl} = (u_1 + \ii u_2 )\exp(\ii \phi_{\qcl}) ,
\quad u_1(L)=u_1(0), \quad u_2(L)=u_2(0),
\tag 2-8
$$
where $u$'s are real function of $s$ and $t$.
This is virtual dynamics of the curve [6].
As well as the argument in refs.[6,25-27], due to the isometry
condition, I require $ [\partial_{t},\partial_s]=0$ for $X$.
Then the isometry condition exactly
preserves, $\dd s\equiv \dd s_\epsilon$ for
$\dd s_\epsilon :=\sqrt{ \partial_s
\bar X_\epsilon\partial_s X_\epsilon}\dd s$.
Even though $\epsilon$ is constant,
dependence of the variation upon the position $s$ comes from
the "equation of motion" (2-8) and $u_a(s)$, $a=1,2$.
Hence the deformation (2-7) contains non-trivial ones.
From $ [\partial_{t},\partial_s]=0$, I have the relation [26,27],
$$
- \partial_{t} \exp(\ii \phi_{\qcl})
= \left((u_{1s}-u_2 k_{\qcl})
+\ii(u_{2s}+u_1 k) \right) \exp(\ii \phi_{\qcl}) .
\tag 2-9
$$
Noting that $\phi$ and $u$'s are real valued,
(3-9) is reduced to two differential equations
and by partially solving one of them,
I obtain the "equation of motion" of the deformation,
$$
\partial_s u_1 = k_{\qcl} u_2,
\quad u_1 =\int^s \dd s \ u_2 k_{\qcl}
=: \partial_s^{-1} u_2 k_{\qcl} ,
$$
$$
\partial_t k = \Omega u_2.
\tag 2-10
$$
Here $\partial^{-1}_s$ is the pseudo-differential operator
with a parameter $c\in\Bbb R$ as an integral constant and
$$
\Omega:= \partial_s^2 +\partial_s
k_{\qcl}\partial_s^{-1} k_{\qcl} . \tag 2-11
$$
is the Gel'fand-Dikii operator for the MKdV equation [26,27].
In ref.[6,25], instead of the single deformation parameter,
I used the infinite dimensional parameters
$\bold t=(t_1,t_3,\cdots)$ and investigated the moduli space
of the partition function.
Then the minimal set of the virtual equations of motion,
which satisfies the physical requirements, is given as
$$
\partial_{t_{2n+1}} k_{\qcl}
=- \Omega^n \partial_s k_{\qcl} , \quad
\partial_{t_{2n+3}} k_{\qcl}
=\Omega \partial_{t_{2n+1}} k_{\qcl}, \quad
(n=1,2,\cdots).
\tag 2-12
$$
They are the MKdV hierarchy [27,28]. As in ref.[6],
I stated that these relations (2-12)
should be regarded as the Neither currents for the
immersed object and
$t$'s should be considered as the Schwinger proper times,
in ref.[25] I showed that (2-12) means
the quantum fluctuations and a kind of currents
of the quantized Neither
theorem or the Ward-Takahashi identities.
However by the studying the moduli of the quantize elastica,
the nontrivial deformation obeys the MKdV equation
$$
\partial_t v + 6 v^2 \partial_s v + \partial_s^3 v=0,
\tag 2-13
$$
because the solutions of the higher order equations
belonging to the MKdV hierarchy
are also satisfied with the MKdV equation.
Here it is a very remarkable fact that for the variation
of $t$ obeying the MKdV equation,
the Bernoulli-Euler functional is invariant,
$$
\partial_t \int \dd s\ v(s,t)^2
=\frac{1}{4}\partial_t E =0, \tag 2-14
$$
because
$$
\partial_t\int \dd s v^2=- \int \dd s \partial_s
(\frac{3}{2} v^4
+\frac{1}{2}(v \partial_s^2 v-(\partial_s v)^2)=0. \tag 2-15
$$
Since the MKdV problem is an initial value problem,
for any regular shape of elastica
satisfied with the boundary conditions, the "time" $t$
development of the curvature can be expressed.
In other words, for given any regular curve,
there exists family of the solutions of the MKdV equation
(2-13) which contains the given curve.
Due to the integrability and (2-15),
during the motion of $t$, the energy functional does not
change its value.
Hence the trajectory of the deformation parameter $t$ draws the
functional space which has the same of value of the
energy functional.
This remind me of the fact that in the group theory
the character of a group is invariant among the elements belonging
the same conjugate class. In fact in the Sato theory, the solutions
space of the MKdV equation is acted by the affine Lie algebra
$A^{(1)}$ [31].
Thus I can estimate the functional space for each functional value.
In other words by investigating the moduli of the MKdV equation
which is satisfied with
the boundary conditions,
$$
v(0)=v(L), \quad X(0)=X(L), \tag 2-16
$$
the measure of the functional integral $\dd \mu$ can be decomposed,
$$
\dd \mu= \sum_E \dd \mu_E. \tag 2-17
$$
So I let the set of these trajectories which occupy
the same energy $E$ be denoted as $\Xi_E$.
Hence the partition function can be represented as
$$
\Cal Z_\reg = \int \dd \mu \exp(-\beta E )
=\sum_E \exp(-\beta E) \int_{\Xi_E}
\dd \mu_E =\sum_E \exp(-\beta E) \text{Vol}(\Xi_E)
\tag 2-18
$$
where
$$
\text{Vol}(\Xi_E)= \int_{\Xi_E} \dd \mu_E \tag 2-19
$$
is the volume of the trajectories $\Xi_E$.
In ref.[25], I explicitly expressed $\dd \mu$ in terms of
the moduli of the MKdV equation. According to the arguments
in ref.[25], for a case of the solution represented by
the hyperelliptic function of
genus $g$, $\dd \mu_E$ is roughly expressed as
$\dd t_3 \wedge \dd t_5 \wedge$
$\cdots \wedge \dd t_{2g-1}$ where
$\bold t_g:=(t_1,t_3,\cdots,t_{2g-1})$
is a subset of the infinite dimensional deformed parameters
such as (2-12).
Even though I introduced the infinite dimensional coordinates
$\bold t$ in ref.[25], they are often reduced to finite
dimensional space, as the Jacobi variety with
finite dimension is embedded in the universal grassmannian manifold
in the Sato theory [25,31,32]. I showed that
$\Xi_E$ is given as the
real subspace of the Jacobi variety corresponding to the
hyperelliptic curve, which is the trajectory space of the
solution [25].
I will note that the volume of $\Xi_E$ is estimated by the unit of
the elastica length $L$. Due to the
complex structure of the moduli of the MKdV equation,
which is expressed as the
Jacobi variety of the hyperelliptic curve and can be performed
the coordinate transformation such as rotation, the volume
can be evaluated in terms of the elastica length $L$ [25,32].
However since the dimension of the trajectory space $\Xi_E$
differs depending upon
the energy $E$, the sum of terms with different dimensional
volume appear. It seems to be fancy
but noting the facts that the dimension of the energy functional
$E$ is the inverse of the length and
that $\beta /$[length] is order unit, the multiple of the length
can be interpreted as
the multiple of the quantizing parameter $\beta^{-1}$.
Hence such summation has physical meanings.
(2-18) means that the density of state of the Bernoulli-Euler
functional system is completely represented by the moduli and
solutions spaces of the MKdV equation.
These space is acted by the infinite dimensional Lie group [31].
Then the measure $\dd \mu$ can be regarded as the Haar measure for
the subalgebra of
infinite dimensional Lie algebra $A^{(1)}$ [25].
\tvskip
\centerline{\twobf \S 3. Quantization of Willmore surface}
\tvskip
I will denote by $\SS$ a shape of a compact surface immersed
in the three dimensional space
$\Bbb R^3 \approx \Bbb C \times \Bbb R$,
and by $(Z(z,\bar z):=X^1+\ii X^2,X^3(z,\bar z))$
its affine vector :
$$
\Sigma \ni z \mapsto (Z,X^3) \in \SS \subset
\Bbb C \times \Bbb R. \tag 3-1
$$
Here $\Sigma$ can be expressed as $\Sigma = \Bbb C/\Gamma$
where $\Gamma$ is a Fuchian group and then $\Sigma$ is a
complex analytic object [33].
The volume element and the infinitesimal length the surface
$\SS$ are given by
$$
\dd \text{vol}=\frac{\ii}{2}\rho\dd z\wedge \dd \bar z
=:\frac{\ii}{2}\rho \dd^2 z, \quad
\dd s^2= \rho \dd z \dd \bar z.
\tag 3-2
$$
The Konopelchenko-Kenmotsu-Weierstrass-Enneper (KKWE) relation
will be expressed as
$$
\psi_+=\ii \sqrt{\bar \partial_z\bar Z},
\quad\psi_-=-\ii \sqrt{ \partial_z\bar Z},
\quad \partial X^3 = - \bar \psi_+ \bar \psi_- ,\tag 3-3
$$
$$
\DD \psi =\pmatrix \partial & -V \\
V & \bar \partial \endpmatrix \pmatrix \psi_+\\
\psi_-\endpmatrix =0 , \quad
V=\frac{1}{2}H \rho,
\tag 3-4
$$
where $V$ is a real valued function of $z$ and $\bar z$
and $H$ is the mean curvature of $\SS$.
The general form might be expressed by the Dirac operator
which was calculated by Burgess and Jensen following
my prescription of the confinement Dirac operator.
However due to the conformal structure of the surface $\SS$,
the equation becomes simpler and given by (3-4) [16]
while due to the isometry condition, the Frenet-Serret relation
becomes simple (2-2).
As the energy integral of the Willmore surface or the Polyakov
extrinsic curvature action is given as
$$
E=\int \rho \dd^2 z \ H^2=4\int \dd^2 z\ V^2 .
\tag 3-5
$$
The Willmore surface is realized as its stationary point.
As I quantized the elastica using the MKdV equation [25],
I will consider quantization of such a surface.
The partition function of the surface is also given as [25],
$$
\tilde {\Cal Z}_\reg =\frac{\int D X \exp\left(-\beta
\int \rho \dd^2 z H^2\right)}
{\text{Vol}(\text{Aff})} ,
\tag 3-6
$$
where $\text{Vol}(\text{Aff})$ means the volume of the affine
transformation in $\Bbb R^3$.
I will search for the deformation flow of the surface which
preserves the
Willmore function or the Polyakov extrinsic curvature action and
complex structure. My question is what equation the deformation
flow obeys.
By Taimanov and Konopelchenko, the answer of the question was
founded that the modified Novikov-Veselov (MNV) equation preserves
the complex structure and the functional (3-5) [10-14].
The MNV equation is given as
$$
V_t=V_\tp+V_\tm,
$$ $$
V_\tp=\partial^3 V+3 \partial V U+\frac{3}{2}V \partial U, \quad
V_\tm=\bar \partial^3 V+3 \bar \partial
V \bar U+\frac{3}{2}V \bar \partial \bar U,
$$ $$
\bar \partial U=\partial V^2 ,\quad
\partial \bar U=\bar\partial V^2.
\tag 3-7
$$
Along the line $z = \bar z$, the MNV equation (3-7)
is reduced to the MKdV equation (2-13).
As the Frenet-Serret relation can be regarded as the inverse
scattering system of the MKdV equation,
the KKWE equation can be also regarded as the inverse scattering
system of the MNV equation.
$$
(\partial_\tpm -B^\pm)\DD +[\DD,A^\pm]=0, \tag 3-8
$$
recovers (3-7) for
$$
A^+= \pmatrix \partial^3 & -3(\partial V) \partial +3 VU\\
0 & \partial^3 +3U \partial +3(\partial U)/2 \endpmatrix,
$$
$$
B^+=3\pmatrix 0 & (\partial V)\partial - VU \\
-(\partial V)\partial-(\partial^2 V)-UV & 0 \endpmatrix,
$$
$$
A^-= \pmatrix \bar \partial^3 + \bar U \bar \partial
+3\bar \partial \bar U/2 & 0\\
3\bar \partial V \bar \partial -3V \bar U & \bar \partial^3
\endpmatrix,
$$
$$
B^-=3\pmatrix 0 & (\bar \partial V)\bar
\partial +(\bar \partial^2 V)-V\bar U \\
-(\bar \partial V)\bar \partial+ V\bar U & 0 \endpmatrix.
\tag 3-9
$$
The variation of the Dirac field is given as
$$
\partial_t\psi=\partial_\tp\psi+ \partial_\tm\psi,\quad
\partial_\tpm\psi = A^\pm \psi. \tag 3-10
$$
For the variation of $t$ obeying the MNV equation, the Willmore
functional is invariant,
$$
4 \partial_t \int \dd^2 z\ V^2=\partial_t E =0. \tag 3-11
$$
because the integrand can be expressed by the
boundary quantities [13],
$$
V^2_t = \partial (V \partial^2 V-\frac{1}{2}
(\partial V)^2 +\frac{3}{2}V^2 U)+
\bar \partial (V \bar\partial^2 V-\frac{1}{2}
(\bar\partial V)^2 +\frac{3}{2}V^2\bar U). \tag 3-12
$$
Next I will check the preserving the complex structure of
surface for the MNV flows
following the argument of Taimanov.
First I will remark that the metric is represented by the
Dirac field as
$$
\rho = (|\psi_1|^2+|\psi_2|^2)^2, \tag 3-13
$$
owing to the relation (3-3). Thus if the relation (3-3) is covariant
or preserves for a point of the MNV flows,
the conformal structure (3-13) maintains.
Thus I will evaluate $\partial_t Z=\partial_\tp Z+\partial_\tm Z$
and $\partial_t X^3$.
By straight forward computations, these values calculated as [13]
$$
\partial_\tpm Z =2 \ii \int^{z(\bar z) }
\dd (f_\pm+g_\pm), \tag 3-14
$$
and
$$
\partial_t X^3 =- \int^{z(\bar z) }
\dd (h_1+h_2) \tag 3-15
$$
where
$$
\split
f_+ &:=\frac{3}{2} U \psi_-^2, \quad
g_+ := \psi_- \partial^2 \psi_-
-\frac{1}{2} (\partial \psi_-)^2\\
f_- &:=\frac{3}{2} \bar U \psi_+^2, \quad
g_- := \psi_+ \bar \partial^2 \psi_+
-\frac{1}{2} (\bar \partial \psi_+)^2
,\endsplit
\tag 3-16
$$
$$
\split
h_1&= \bar \psi_+ \partial^2 \psi_- + \psi_-
\partial^2 \bar \psi_+ -\partial \psi_-\partial
\bar \psi_+ + 3 U\bar \psi_+ \psi_- ,\\
h_2&= \psi_+ \bar\partial^2 \bar\psi_- +
\bar\psi_- \bar\partial^2 \psi_+ -
\bar\partial \bar\psi_- \bar\partial \psi_+ +
3 \bar U \psi_+\bar \psi_- .\endsplit \tag 3-17
$$
%$\partial_+:=\partial$, $\partial_-:=\bar \partial$,
$U^{(+)}=U$ and $U^{(-)}=\bar U$.
Here $\dd f= \partial f \dd z+ \bar \partial f \dd \bar z$.
Let the infinitesimal flows obeying the MNV equation (3-7)
module $\epsilon^2$ be denoted as
$$
(Z_\epsilon,X_\epsilon^3):=(Z,X^3)+
\epsilon \partial_t(Z,X^3)+ \Cal O(\epsilon^2). \tag 3-18
$$
(3-14)-(3-17) means the infinitesimal variation is
given as the integral of the closed form defined over
$\Sigma$ [34]
and can be regarded as a single function of $\Sigma$.
Since $(Z,X^3)$ is also a periodic function of $\Sigma$,
$(Z_\epsilon,X_\epsilon^3)$
is globally defined over $\Sigma$ as a function of $\Sigma$.
On the other hand, (3-14)-(3-17) guarantee
that $[\partial,\bar \partial]X^i_\epsilon=0$, which
means that I can locally define the independent
coordinates $z$ and $\bar z$ for $X^i_\epsilon$ surface;
I can locally find a conformal coordinate system of
an open set of $\Sigma$. Furthermore due to the global
properties, their coordinate system can be extended to the
global coordinate and the connection of each open set are
trivial due to $[\partial,\bar \partial]X^i_\epsilon=0$
for any point of $\Sigma$.
Hence the MNV flows preserves the complex structure of the surface
$\SS$.
I will emphasis that
the MNV problem is also an initial value problem, for any shape of
compact conformal surface, the "time" $t$ development of
the surface can be expressed
and these conserves the energy functional and complex structure.
Hence the trajectory of the deformation parameter $t$
means the states of the same energy and its volume is the
density of states of each energy $E$.
As I did for the quantization elastica,
the measure of the functional integral $\dd \mu$ can be decomposed,
$$
\dd \tilde \mu= \sum_E \dd \tilde \mu_E \tag 3-19
$$
and moduli of $\tilde {\Cal Z}_\reg$ restricted by $E$ is
denoted as $\tilde \Xi_E$
$$
\tilde {\Cal Z}_\reg= \int \dd \tilde
\mu \exp(-\beta E )=\sum_E \exp(-\beta E)
\int_{\tilde \Xi_E} \dd\tilde \mu_E
=\sum_E \exp(-\beta E) \text{Vol}(\tilde \Xi_E) \tag 3-20
$$
As I did for the Bernoulli-Euler functional, the density of state
of the Willmore functional system
might be completely represented by the moduli of the MNV equation.
However the Willmore surface has no natural length because for
a global scale transformation $z \to \lambda z$
($\lambda >0$), the mean curvature changes as $H \to H/\lambda$
and the Willmore surface is invariant.
Hence for given energy $E$, there are infinite degenerate states
regarding to the global scaling parameter $\lambda \in (0,\infty)$
and the regularized partition function
$\tilde {\Cal Z}_\reg$ also diverges.
However along the line of $z=\bar z$, the deformation of MNV flows
obeys the MKdV equation
which conserves local length of the line. In other words,
on the MNV flows,
the length of the line is a conserved quantity and is well-defined.
Hence in terms of this length, I can redefine the partition
function by fixing the length of the line,
$$
{\Cal Z}_\reg:=\tilde {\Cal Z}_\reg |_{(\text{the length of }z
=\bar z) =L}. \tag 3-21
$$
Due to the compactness of the surface $\SS$, $L$ must be finite.
By fixing the scale of the surface, I will define decomposed
measure and the space of the trajectories,
$\dd \mu_{E,L}:=\dd \tilde \mu|_L$, $\Xi_{E,L}:={\tilde \Xi_E}|_L$,
$$
{\Cal Z}_\reg =\sum_E \int_{ \Xi_{E,L}} \dd\tilde
\mu_{E,L}\exp(-\beta E)
=\sum_E \exp(-\beta E) \text{Vol}( \Xi_{E,L}) \tag 3-22
$$
The physical meaning of the summation in (3-22) is justified
similar to (2-19).
\tvskip
\centerline{\twobf \S 4. Discussion}
\tvskip
Carroll and Konopelchenko also proved that the MNV flows conserves
the extrinsic string action for the case that $\rho H$
is constant; here the extrinsic string action consists of the
Nambu-Goto action, the Wess-Zumino-Witten type geometrical action,
and the Polyakov extrinsic curvature action (3-3).
Hence my result (3-2) can be extended to such a case and then
it means that the algorithm of the
calculation of the partition function of the extrinsic string
in $\Bbb R^3$ is essentially
the same as above arguments.
In other words, the quantization of the string immersed
in $\Bbb R^3$ can be partially performed even though
only the string in $\Bbb R^n$ $n<3$ had been studied as
the two dimensional gravity [28-30].
In ref. [2] and [3], I stated that as the self-dual Yang-Mills
equation can be expressed by
the integrable equation and be represented by the Dirac operator
and as the MKdV equation governs the "virtual" motion of
the elastica and can be
written by the Dirac operator, the higher dimensional
soliton surface might be expressed by the Dirac operator
and has a physical meaning.
This conjecture was proved by the discoveries and studies of
Konopelchenko and Taimanov.
Using the Dirac operator, they investigated the surface itself
and derived non-trivial results [10-14].
Categories of the linear analytic system of Dirac operator,
the geometry and integrable system are closely connected with
each other.
Their connections should be interpreted as functors among them
and the exact sequences and the associated cohomorogy in individual
categories should be expressed by common language.
Thus the expression of these systems should be unified and then
using the relations their common hidden symmetries might be
revealed. I believe that this quantization of surface and
quantization of elastica [25]
contributes such studies.
Finally I will comment upon open problems related to this system.
In ref.[25], I found that
at the critical point of the quantized elastica, a certain
expectation value obeys the Painlev\'e
equation of the first kinds.
I have a question what equation appears at a critical point
in the quantized Willmore surface system. If exists, it might
be related to the higher dimensional analogue of the Painlev\'e
equation.
Furthermore since the MNV equation is an initial value problem,
more general Riemannian surfaces can be allowed, at least,
an initial condition even though the energy
manifold of the inverse scattering system is given as only
hyperelliptic curves [6].
Hence I have another question whether there is an analytical
connection between the general Riemannian surface or
the general Fuchian group and the hyperelliptic function
of the MKdV equation.
If there is, this system should be algebraically studied.
%\newpage
\tvskip
\centerline{\twobf Acknowledgment}
\tvskip
I would like to thank Y. \^Onishi
for helpful discussions and continuous encouragement
and Prof. B.~G.~Konopelchenko for sending me his interesting works.
I also acknowledge that the seminars on
differential geometry, topology, knot theory and group theory with
Prof. K.~Tamano influenced this work.
%\enddocument
%\newpage
\Refs
\ref \no 1 \by S.~Matsutani and H.~Tsuru \jour Phys. Rev A
\vol 46 \yr1992\page1144-7 \endref
\ref \no 2 \by S.~Matsutani \jour Prog. Theor. Phys. \vol 91
\yr1994\page1005-37 \endref
\ref \no 3 \by S.~Matsutani
\jour Phys. Lett. A \vol 189 \yr 1994\pages 27-31\endref
\ref \no 4 \by S.~Matsutani \jour J. Phys. A: Math. \& Gen.
\vol 28 \yr 1995 \pages1399-1412 \endref
\ref \no 5 \by S.~Matsutani \jour Int. J. Mod. Phys. A \vol 10
\yr 1995 \pages3091-3107 \endref
\ref \no 6 \by S.~Matsutani \jour Int. J. Mod. Phys. A \vol 10
\yr 1995 \page 3109-3130 \endref
\ref \no 7 \by S.~Matsutani \jour Thesis in Tokyo Metropolitan Univ.
\yr1996\endref
\ref \no 8 \by S.~Matsutani \jour Phys. Rev. A \vol 47 \yr 1993
\pages 686-689\endref
\ref \no 9 \by S.~Matsutani \jour J. Phys A \vol 26 \yr 1993
\pages 5133-5143\endref
\ref \no 10 \by B.~G.~Konopelchenko \jour Studies in Appl.~Math.
\vol 96 \yr1996 \pages 9-51 \endref
\ref \no 11 \by B.~G.~Konopelchenko and I.~A.~Taimanov
\jour J.~Phys.~A: Math.~\& Gen.
\vol 29 \yr1996 \pages 1261-65 \endref
\ref \no 12 \by B.~G.~Konopelchenko and I.~A.~Taimanov \paper
Generalized Weierstarass formulae,
soliton equations and Willmore surfaces I.
Tori of revolution and the mKDV equation \jour dg-ga/9506011
\endref
\ref \no 13 \by I.~A.~Taimanov
\paper Modified Novikov-Veselov equation and
differential geometry of surface \jour dg-ga/9511005 \endref
\ref \no 14 \by I.~A.~Taimanov
\paper Surface revolution in terms of soliton \jour dg-ga/9610013
\endref
\ref \no 15 \by K.~Kenmotsu \jour Math. Ann. \year 1979
\pages89-99 \vol 245 \endref
\ref \no 16\by S.~Matsutani \jour J. Phys. A: Math.
\& Gen. \vol 30 \yr 1997 \pages4019-4029 \endref
\ref \no 17 \by S.~Matsutani \paper Immersion Anomaly of
Dirac Operator on Surface in $\Bbb R^3$
\jour submitted to J. Phys. A: Math. \& Gen. \endref
\ref \no 18 \by M.~Burgess and B.~Jensen \jour Phys. Rev. A
\vol48 \yr1993\page1861-6 \endref
\ref \no 19 \by R.~Carroll and B.~Konopelchenko
\jour Int. J. Mod. Phys. \vol A11 \yr 1996 \pages1183-1216
\endref
\ref \no 20 \by P.~G.~Grinevich and M.~U.~Schmidt \paper
Conformal invariant functionals of immersioons of tori
into $\Bbb R^3$ \jour dg-ga/9702015 \endref
\ref \no 21 \by T.~J.~Willmore \jour J. Lond. Math. Soc. \vol 2
\yr 1971\pages307-310\endref
\ref \no 22 \by T.~J.~Willmore \book Riemannian Geometry
\publ Oxford \yr 1993 \publaddr Oxford \endref
\ref \no 23 \by A.~M.~Polyakov \book Gauge Fields and Strings
\publ Harwood Academic Publishers \yr 1987 \publaddr London \endref
\ref \no 24 \by A.~E.~H.~Love
\book A Treatise on the Mathematical Theory of Elasticity
\publ Cambridge Univ. Press \yr 1927 \publaddr Cambridge \endref
\ref \no 25 \by S.~Matsutani
\paper Statistical Mechanics of Elastica on Plane as a Model of
Supercoiled DNA-Origin of the MKdV hierarchy-
\jour submitted to J. Phys. A: Math. \& Gen.,
solv-int/9707003 \endref
\ref \no 26 \by R.~E.~Goldstein and D.~M.~Petrich
\jour Phys. Rev. Lett.\vol 67 \yr 1991, \page 3203-3206 \endref
\ref \no 27 \by R.~E.~Goldstein and D.~M.~Petrich
\jour Phys. Rev. Lett.\vol 67 \yr 1992, \page 555-558 \endref
\ref \no 28 \by M.~R.~Douglas and S.~H.~Shenker
\jour Nucl.~Phys.~B \vol 335 \yr 1990 \page 635-654 \endref
\ref \no 29 \by D.~J.~Gross and A.~A.~Migdal
\jour Phys.~Rev.~Lett. \vol 64 \yr 1990
\page 717-720 \endref
\ref \no 30\by K.~Sogo \jour J.~Phys.~Soc.~Jpn \vol 62
\yr 1993 \page 1887-1894 \endref
\ref \no 31 \by M. Jimbo and T. Miwa\jour Publ. RIMS, Kyoto Univ.
\vol 19 \yr 1983 \page 943-1001 \endref
\ref \no 32 \by I.~M.~Krichever \jour Funct. Anal. Appl. \vol 11
\yr 1977 \page 12-26 \endref
\ref \no 33 \by H. Hochstadt
\book The Functions of Mathematical Physics \publ
Wiely\publaddr New York\yr1971\endref
\ref \no 34 \by M.~Nakahara\book Geometry, Topology and Physics
\publ Adam Hilger\publaddr New York \yr 1990\endref
\endRefs
\enddocument
/E