On self-adjointness of a Schroedinger operator
Extracted source text ยท arXiv funct-an/9607002
%----------- version 15.8.96 small changes in formula (26) ------
%--------------------- version 28.7.96 --------------------------
%--------------------- version 14.7.96 ---------------------------
%--------------------- version 10.7.96 ---------------------------
%-----------------------------------------------------------------
%------------------------- This is AMS-TeX ---------------------
%-----------------------------------------------------------------
\input amstex
\documentstyle{amsppt}
%\TagsOnRight
\nopagenumbers
\magnification=\magstep1
%Greek letters
\def\alp{\alpha} \def\Alp{\Alpha}
\def\bet{\beta}
\def\gam{\gamma} \def\Gam{\Gamma}
\def\del{\delta} \def\Del{\Delta}
\def\eps{\varepsilon}
\def\zet{\zeta}
\def\tet{\theta} \def\Tet{\Theta}
\def\iot{\iota}
\def\kap{\kappa}
\def\lam{\lambda} \def\Lam{\Lambda}
\def\sig{\sigma} \def\Sig{\Sigma}
\def\vphi{\varphi}
\def\ome{\omega} \def\Ome{\Omega}
%Caligraphic roman letters
\def\calA{{\Cal A}}
\def\calB{{\Cal B}}
\def\calC{{\Cal C}}
\def\calD{{\Cal D}}
\def\calE{{\Cal E}}
\def\calF{{\Cal F}}
\def\calG{{\Cal G}}
\def\calH{{\Cal H}}
\def\calI{{\Cal I}}
\def\calJ{{\Cal J}}
\def\calK{{\Cal K}}
\def\calL{{\Cal L}}
\def\calM{{\Cal M}}
\def\calN{{\Cal N}}
\def\calO{{\Cal O}}
\def\calP{{\Cal P}}
\def\calQ{{\Cal Q}}
\def\calR{{\Cal R}}
\def\calS{{\Cal S}}
\def\calT{{\Cal T}}
\def\calU{{\Cal U}}
\def\calV{{\Cal V}}
\def\calW{{\Cal W}}
\def\calX{{\Cal X}}
\def\calY{{\Cal Y}}
\def\calZ{{\Cal Z}}
%Euler Fraktur letters
\def\grG{{\frak G}} \def\grg{{\frak g}}
\def\grT{{\frak T}} \def\grt{{\frak t}}
%Capital roman double letters
\def\QQ{\Bbb Q}
\def\RR{\Bbb R}
\def\ZZ{\Bbb Z}
\def\CC{\Bbb C}
\def\NN{\Bbb N}
\def\otet{\overline\tet}
\def\oxi{\overline\xi}
\def\obet{\overline\bet}
\def\tilrho{{\widetilde{\rho}}}
\def\tilP{{\widetilde{P}}}
\def\nek{,\ldots,}
\def\sdp{\times \hskip -0.3em {\raise 0.3ex
\hbox{$\scriptscriptstyle |$}}} % semidirect product
%words in roman font
\def\End{\operatorname{End}}
\def\dist{\operatorname{dist}}
\def\olim{\operatornamewithlimits{\overline{lim}\,}}
\def\const{\operatorname{const}}
\def\RANK{\operatorname{rank}}
\def\Hom{\operatorname {Hom}}
\def\im{\operatorname {im}}
\def\ind{\operatorname{ind}}
%---------------------------------------------------------------
%---------------------------------------------------------------
\def\d{\partial}
\def\we{\bigwedge^\b(T^*M)}
%\def\wee[1]{\text{\( \bigwedge^{#1}(T^*M)\)}}
\def\wec{\Ome^\b_c(M)}
\def\lwe{L^2\Ome^\b(M)}
%\def\lwee[1]{L^2\Ome^{#1}(M)}
\def\n{\nabla}
\def\b{\bullet}
\def\p{\partial}
\def\>{\rangle}
\def\<{\langle}
\NoBlackBoxes
\document
%-------------------------------------------------------------------
%-------------------------------------------------------------------
\topmatter
\title On self-adjointness of a Schr\"odinger operator on differential
forms
\endtitle
\rightheadtext{On self-adjointness of a Schr\"odinger operator}
\author Maxim Braverman \endauthor
\address
School of Mathematical Sciences,
Tel-Aviv University,
Ramat-Aviv 69978, Israel
\endaddress
\email
maxim\@math.tau.ac.il
\endemail
\thanks{The research was supported by US - Israel Binational Science
Foundation grant No. 9400299}
\endthanks
\subjclass{Primary: 58G25 Secondary: 35P05} \endsubjclass
\abstract
Let $M$ be a complete Riemannian manifold and let $\Ome^\b(M)$ denote
the space of differential forms on $M$. Let $d:\Ome^\b(M)\to
\Ome^{\b+1}(M)$ be the exterior differential operator and let
$\Del=dd^*+d^*d$ be the Laplacian. We establish a sufficient condition
for the Schr\"odinger operator $H=\Del+V(x)$ (where the potential
$V(x):\Ome^\b(M)\to \Ome^\b(M)$ is a zero order differential operator)
to be self-adjoint. Our result generalizes a theorem by I.~Oleinik
about self-adjointness of a Schr\"odinger operator which acts on the
space of scalar valued functions.
\endabstract
\endtopmatter
%-------------------------------------------------------------------
%-------------------------------------------------------------------
\subheading{1. Introduction} Suppose $M$ is a complete Riemannian
non-compact manifold. We will assume that $M$ is oriented and
connected. Let $T^*M$ denote the cotangent bundle to $M$ and let $\we
= \bigoplus_i \bigwedge^i(T^*M)$ denote the exterior algebra of $T^*M$. We
denote by $\lwe$ the space of square integrable complex valued
differential forms on $M$, i.e. the space of sections of $\we\otimes
\CC$ which are square integrable with respect to the scalar product %
$$
\<\alp,\bet\> \ = \ \int_M \, \alp\wedge *\obet,
\qquad \alp,\bet\in \lwe.
\tag1 $$
Here $*$ denotes the Hodge operator associated to the Riemannian
metric on $M$. Note that $L^2\Ome^0(M)$ is just the space of square
integrable complex valued functions on $M$.
Let $d:\lwe\to L^2\Ome^{\b+1}(M)$ denote the exterior differential and
let $d^*$ be the operator formally adjoint to $d$ with respect to the
scalar product (1).
Let $\Del=dd^*+d^*d$ be the Laplacian and consider the Schr\"odinger
operator
$$
H \ = \ \Del \ + \ V(x): \ \lwe \ \to \ \lwe
\tag2 $$
where the {\it potential} $V(x)$ is a measurable section of the bundle
$\End\big(\we\big)$ of endomorphisms of $\we$ which belongs to the
class $L^\infty_{loc}$ (i.e. such that for any compact set $K\subset M$
there exists a constant $C_K>0$ such that $|V(x)|\le C_K$ for almost all
$x\in K$).
We denote by $H_0$ the restriction of $H$ on the space $\wec$ of
smooth differential forms with compact support. The purpose of this
paper is to introduce a sufficient condition on the potential $V(x)$
for operator $H_0$ to be self-adjoint.
%-------------------------------------------------------------------
\subheading{2. Statement of results} For $x,y\in M$ let $\dist(x,y)$
denote the Riemannian distance between $x$ and $y$. Fix a point $p\in
M$ and set $r(x)=\dist(x,p)$.
Fix $x\in M$. The Riemannian metric on $M$ defines a scalar product
$\<\cdot,\cdot\>_x$ on the fiber $\bigwedge^\b(T^*_xM)\otimes \CC$ of
the vector bundle $\we\otimes \CC$. As usual, we write $V(x)\ge C$ if
$$
\<V(x)\, \xi,\xi\>_x \ \ge \ C\, \<\xi,\xi\>_x
\tag3 $$
for any $\xi\in \bigwedge^\b(T^*_xM)\otimes \CC$. Note that it follows
from (3) that $V(x)$ is a self-adjoint endomorphism of
$\bigwedge^\b(T^*_xM)$.
\proclaim{Theorem A} Assume that for almost all $x\in M$ the potential
$V(x)$ of the operator (2) satisfies the estimate
$$
V(x) \ \ge \ -Q(x),
\tag4 $$
where $1\le Q(x)\le \infty$ and $Q^{-1/2}(x)$ is a Lipschitz function
on $M$ such that
$$
|\, Q^{-1/2}(x) \ - \ Q^{-1/2}(y)\, | \ \le \ K\dist(x,y)
\quad \text{for any} \quad x,y\in M.
\tag5 $$
If for any piecewise smooth curve $\gam:[0,\infty)\to M$ such that
$\lim_{t\to\infty} r(\gam(t))=\infty$ the integral
$$
\int_\gam \,
Q^{-1/2}(x) \, d\gam \ = \ \infty
\tag6 $$
then the operator $H_0$ is essentially self-adjoint.
\endproclaim
For the case of a Schr\"odinger operator acting on scalar valued
functions this theorem was established by I.~Oleinik \cite{O2}.
Note that $Q(x)$ may be equal to infinity on a set of positive
measure.
As a simple consequence of Theorem~A we obtain the following
\proclaim{Theorem B} Suppose that for almost all $x\in M$ the
potential $V(x)$ satisfies the estimate $V(x)\ge -q\big(r(x)\big)$,
where $1\le q\le \infty$ and $q^{-1/2}(t)$ is a Lipschitz function
on $\RR$ such that $\int_0^\infty \, q^{-1/2}(t) \, dt = \infty$.
Then the operator $H_0$ is essentially self-adjoint.
%
In particular, if $M=\RR^n$ and $V(x)\ge -C|x|^2$ then the operator
$H_0$ is essentially self-adjoint.
\endproclaim
\subheading{\it Remark} Theorem~A remains true if we replace
$\lwe$ by the space of square integrable forms on $M$ with values in a
flat Hermitian vector bundle $\calF$ over $M$, provided that the
Hermitian structure on $\calF$ is flat. In this case the differential
$d$ should be replaced by the covariant differential associated to the
flat structure on $\calF$. The proof is a verbatim repetition of the
proof for the scalar case, cf. bellow. However the notation in the
vector valued case is more complicated.
%-------------------------------------------------------------------
\subheading{3. Historical remarks} An analogue of Theorem~B for the
case \/ $M=\RR^1$ \/ was established by Sears \cite{Se}. B.~Levitan
\cite{Le} proved the Sears theorem for the Schr\"odinger operator
acting on scalar valued functions on $M=\RR^n$. F.~Rofe-Beketov
\cite{RB} extended these results to the case where the potential
$V(x)$ can not be estimate by a function depending only on
$\dist(x,p)$. Many results and references about the essential
self-adjointness of Schr\"odinger operators on $\RR^n$ may be found in
\cite{RS}.
I.~Oleinik \cite{O1,O2} established Theorem~A for the Schr\"odinger
operator acting on scalar valued functions on a complete Riemannian
manifold.
Essential self-adjointness of a pure Laplacian (without lower order
terms) on differential forms on a complete Riemannian manifold was
first stated and proved by M.~P.~Gaffney \cite{Ga1,Ga2}. A number of
related results may be found in \cite{Sh}.
In \cite{BFS}, Theorem~B is established for the case where $M$ is
a manifold with cylindrical ends and the potential $V(x)\ge 0$. The
result is used there to study Witten deformation of the Laplacian on a
non-compact manifold.
%-------------------------------------------------------------------
\subheading{4. Acknowledgment} I would like to thank M.~Shubin for
posing the problem and for essential help.
I am very grateful to I.~Oleinik for pointing out a gap in a
preliminary version of the paper and for drawing my attention to the
paper \cite{O2}.
I am also thankful to M.~Farber and V.~Matsaev for valuable
discussions.
%-------------------------------------------------------------------
\subheading{5. The domain of $\calD(H_0^*)$} Let $H_0^*$ denote the
operator adjoin to $H_0$. The domain $\calD(H_0^*)$ of $H_0^*$
consists of forms $\alp\in \lwe$ such that $H\alp$ understood in the
sense of distributions also belongs to $\lwe$.
The operator $H_0$ is symmetric. Hence, to show that its closure is
self-adjoint it is enough to show that the adjoint operator $H_0^*$ is
symmetric. In other words we have to prove that
$$
\int_M \,
(H\alp\wedge *\obet - \alp\wedge *\overline{H\bet})
\ = \ 0
\tag7 $$
for any $\alp, \bet\in \calD(H_0^*)$.
To prove (7) we need some information about the behavior of
differential forms from $\calD(H_0^*)$. The main result of this
section is the following lemma, which provides us with this
information.
%
%
\proclaim{Lemma 1} If $\alp\in \calD(H_0^*)$ then the forms
$Q^{-1/2}d\alp, Q^{-1/2}d^*\alp$ are square integrable.
\endproclaim
%
\subheading{\it Remark} 1. \ By the standard theory of elliptic
operators any $\alp\in \calD(H_0^*)$ belongs to the Sobolev space
$H_{loc}^2$. Hence, $d\alp, d^*\alp$ are locally square
integrable. Thus the lemma provides us with an information about the
behavior of the forms from $\calD(H_0^*)$ at infinity.
2. \ For the Schr\"odinger operator on scalar valued functions on
$\RR^n$ an analogous lemma was established in \cite{RB}. The proof was
adopted in \cite{O1,O2} to the case of a Riemannian manifold. In our
proof we follow rather closely the lines of \cite{O2}. However, the
fact that we deal with differential forms rather than with scalar
valued functions demands a more careful analysis.
%
\demo{Proof}
Recall that we fixed a point $p\in M$ and that for any $x\in M$ we
denoted by $r(x)$ the Riemannian distance between $x$ and $p$.
It is shown in \cite{O2, Proof of Lemma~1} that for any $R>0,\eps>0$
there exist smooth functions $r_{R,\eps}(x),F_{R,\eps}(x)$ on $M$
which approximate the Lipschitz functions $r(x),Q^{-1/2}(x)$ in the
sense that
$$
\aligned
|r_{R,\eps}(x)-r(x)| \ < \ \eps, \qquad
Q^{-1/2}(x)-\eps \ &< \
F_{R,\eps}(x) \ < \ (1+\eps) \, Q^{-1/2}(x),\\
\olim_{\eps\to 0} |d r_{R,\eps}(x)|\le 1, \qquad \ \
&\olim_{\eps\to 0} |dF_{R,\eps}(x)| \ \le \ K,
\endaligned
\tag8 $$
for any \/ $x\in r_{R,\eps}^{-1}([0,R+1])$. Here $K$ is the same
constant as in (5).
Let $\Psi:[0,+\infty)\to [0,1]$ be a smooth function which is equal
to one when $t\le 1/2$ and which is equal to zero when $t\ge 1$. Set
$$
\psi_{R,\eps}(x) \ = \
\cases
\Psi\big(\frac{r_{R,\eps}(x)}R\big)\, F_{R,\eps}(x)
\quad &\text{if} \quad r_{R,\eps}(x)\le R; \\
0 \quad &\text{outside of the set} \ r_{R,\eps}(x)\le R.
\endcases
\tag9 $$
For any $R>0$ the functions $\psi_{R,\eps}, \, \eps<1$ vanish outside
of the compact set $r^{-1}([0,R+1])$. Hence, it follows from (8) and
(4) that there exist a constant $K_1>0$ not depending on $R$ and a
number $\eps_R>0$ (which does depend on $R$) such that
$$
% \olim_{\eps\to 0}
% |\psi_{R,\eps}(x)| \ \le \ Q^{-1/2}(x) \ \le \ 1,
% \qquad
|d \psi_{R,\eps}(x)| \ \le \ K_1, \qquad
\psi_{R,\eps}^2(x) \ \le \ 2, \qquad
\left|\, \int_M\, \psi_{R,\eps}^2\, \alp\wedge *V\alp\, \right|
\ \le \ 2\, \|\alp\|^2,
\tag10 $$
for any $x\in M, \, R>1, \, 0<\eps<\eps_R, \, \alp\in \lwe$. Here
$\|\alp\|=\<\alp,\alp\>^{\frac12}$ denotes the $L^2$-norm of the form
$\alp$.
Functions \/ $\psi_{R,\eps}$ \/ have compact support. Hence, in view
of the remark~1 after the statement of the lemma, the forms \/
$\psi_{R,\eps}d\alp$ \/ and \/ $\psi_{R,\eps}d^*\alp$ \/ are square
integrable. Assume that \/ $\alp\in \calD(H_0^*)$ \/ is a real valued
form and set
$$
J_{R,\eps}^2 \ = \
\|\psi_{R,\eps}d\alp\|^2 + \|\psi_{R,\eps}d^*\alp\|^2 \ = \
\int_M \psi_{R,\eps}(x)^2\,
\big(d\alp\wedge *d\alp + d^*\alp\wedge *d^*\alp\big).
\tag11 $$
It follows from (8), (9) that to prove the lemma it is
enough to show that
$$
\olim_{R\to \infty} \, \olim_{\eps\to 0} J_{R,\eps}
\ < \ \infty.
\tag12 $$
Let us first rewrite the integrand in (11) in a more convenient
form. In the calculations bellow we use the equality (cf.
\cite{Wa, \S6.1}) \/ $d^*\alp= (-1)^{|\alp|}*^{-1}d*\alp$ \/ where
$|\alp|$ denotes the degree of the differential form $\alp$.
%
$$
\alignat 2
\ \psi_{R,\eps}^2\, d\alp\wedge &&*d\alp \ =& \
d\big(\psi_{R,\eps}^2\alp\wedge*d\alp\big) -
2\psi_{R,\eps}d\psi_{R,\eps}\wedge\alp\wedge *d\alp +
\psi_{R,\eps}^2\alp\wedge*d^*d\alp, \\
%
\ \psi_{R,\eps}^2\, d^*\alp\wedge &&*d^*\alp \ =& \
(-1)^{|\alp|} \psi_{R,\eps}^2\, d^*\alp\wedge d*\alp \tag13 \\
&&= -d&\big(\psi_{R,\eps}^2d^*\alp\wedge*\alp\big) +
2 \psi_{R,\eps}d\psi_{R,\eps}\wedge d^*\alp\wedge *\alp +
\psi_{R,\eps}^2dd^*\alp\wedge *\alp.
\endalignat
$$
It follows now from (13), (10) and from the Stokes theorem that, if
$R>1, \, \eps<\eps_R$, then
$$
\alignat 2
\|\psi_{R,\eps}d\alp\|^2
\ &=&& \
\int_M \, \psi_{R,\eps}^2\, d\alp\wedge *d\alp
\ = \
-2\<d\psi_{R,\eps}\wedge \alp,\psi_{R,\eps}d\alp\> +
\<\alp,\psi_{R,\eps}^2d^*d\alp\>
\\
% &\le&& \
% 2 \|d\psi_{R,\eps}\wedge \alp\|\, \|\psi_{R,\eps} d\alp\| +
% \<\alp,\psi_{R,\eps}^2d^*d\alp\>
% \\
&\le&& \
2K_1\|\alp\|\, \|\psi_{R,\eps}d\alp\| +
\<\alp,\psi_{R,\eps}^2d^*d\alp\>,
%
%
\\
%
%
\|\psi_{R,\eps}d^*\alp\|^2
\ &=&& \
\int_M \, \psi_{R,\eps}^2\, d^*\alp\wedge *d^*\alp
\ = \
2\<d\psi_{R,\eps}\wedge\psi_{R,\eps}d^*\alp,\alp\> +
\<\psi_{R,\eps}^2dd^*\alp,\alp\>
\\
% &\le&& \
% 2\|d\psi_{R,\eps}\wedge\psi_{R,\eps}d^*\alp\|\, \|\alp\| +
% \<\psi_{R,\eps}^2dd^*\alp,\alp\>
% \\
&\le&& \
2K_1 \|\psi_{R,\eps}d^*\alp\|\, \|\alp\| +
\<\alp,\psi_{R,\eps}^2dd^*\alp\>.
\endalignat
$$
Summing these two equations we obtain
$$
\multline
J_{R,\eps}^2 \ \le \
2K_1\|\alp\| \big(
\|\psi_{R,\eps}d\alp\| + \|\psi_{R,\eps}d^*\alp\|
\big) +
\<\alp,\psi_{R,\eps}^2\Del\alp\> \\
\le \
4K_1\|\alp\|\, J_{R,\eps} +
\int_M\,
\psi_{R,\eps}^2\big(\alp\wedge *H\alp - \alp\wedge *V\alp\big) \\
\le \
4K_1\|\alp\|\, J_{R,\eps} +
2 \|\alp\|\, \|H\alp\| +2 \|\alp\|^2.
\endmultline
\tag14 $$
Here the last inequality follows from (10).
It follows from (14) that the set $\{J_{R,\eps}: \, R>1,
\eps<\eps_R\}$ is bounded from above. Hence (12) holds. The proof of
the lemma is completed.
$\square$\enddemo
%-----------------------------------------------------------
\subheading{6. Proof of Theorem~A} We apply a modification of the
method used in \cite{RB} suggested by I.~Oleinik \cite{O2}.
The quantity
$$
\tilrho(x,y) = \inf_{\gam}\, \int_\gam Q^{-1/2}(x)\, d\gam,
\tag15 $$
where the infimum is taken over all piecewise smooth curves connecting
the points $x,y\in M$, is called {\it generalized distance} between $x$
and $y$. It is a symmetric function in $x,y$ which satisfies the
triangular inequality. The first metric axiom is not valid in
general. Note, however, that (6) implies, that the sets
$P^{-1}([0,R])$ are compact for any $R>0$.
Recall that in Section~2 we have fixed a point $p\in M$.
Set \/ $P(x)=\tilrho(x,p)$. Then (cf. \cite{O2, Lemma~2})
$$
|P(x)-P(y)| \ \le \
Q^{-1/2}(x)\dist(x,y) \ + \frac{K}2 (\dist(x,y))^2
\tag16 $$
for any $x,y\in M$. It follows (cf. \cite{O2}) that for any $R>0,
\eps>0$ there exists a smooth function $\tilP_{R,\eps}(x)$ which
approximates $P(x)$ in the sense that
$$
|\tilP_{R,\eps}(x)-P(x)|\le \eps, \qquad
\olim_{\eps\to 0} |d\tilP_{R,\eps}(x)|\le Q^{-1/2}(x),
\tag17 $$
for any $x\in P^{-1}([0,R+1])$.
Assume that $\eps<1$ so that $\tilP_{R,\eps}^{-1}([0,R])\subset
P^{-1}([0,R+1])$. Let us define a piecewise smooth function
$P_{R,\eps}(x)$ on $M$ by the formula
$$
P_{R,\eps}(x)=\cases
\tilP_{R,\eps}(x) \quad &\text{if} \quad
\tilP_{R,\eps}(x)\le R; \\
R \quad &\text{outside the set} \quad
\tilP_{R,\eps}(x)\le R.
\endcases
\tag18 $$
By (17), the inequality
$$
\olim_{\eps\to 0} |dP_{R,\eps}(x)| \ \le \ Q^{-1/2}(x)
\tag19 $$
holds almost everywhere on $M$.
Recall from Section~5, that the statement of Theorem~A is equivalent
to equality (7). Fix $\alp,\bet\in \calD(H_0^*)$ and consider the
following approximation of the integral (7)
$$\multline
I_{R,\eps} \ = \ \int_M \, \left(1-\frac{P_{R,\eps}}R\right)\,
\big(H\alp\wedge*\obet -
\alp\wedge*\overline{H\bet}\big) \\
\ = \
\int_M \, \left(1-\frac{P_{R,\eps}}R\right)\,
\big(\Del\alp\wedge*\obet -
\alp\wedge*\overline{\Del\bet}\big).
\endmultline
\tag20 $$
By the Fatou theorem (\cite{RS, Theorem I.17}), it is enough to show that
$$
\olim_{R\to \infty} \, \olim_{\eps\to 0} I_{R,\eps} \ = \ 0.
\tag21 $$
We will need the following ``integration by parts'' lemma
\footnote{I learned this lemma from M.~Shubin.}
%
\proclaim{Lemma 2} Let $\phi:M\to \RR$ be a smooth function with compact
support. Then
$$
\multline
\ \int_M\, \phi \, \Del\alp\wedge *\obet \\
\ = \
\int_M\,
\phi \, \big(d\alp\wedge*d\obet + d^*\alp\wedge*d^*\obet\big)
\ + \
\int_{M} \, d\phi\wedge \big(
\obet\wedge *d\alp - d^*\alp\wedge*\obet \big)
\endmultline \tag22
$$
for any $\alp,\bet\in \calD(H_0^*)$.
\endproclaim
Note that, by remark 1 after the statement of Lemma~1, all the
integrals in (22) have sense.
%
%
\demo{Proof}
Recall that \/ $d^*u= (-1)^{|u|}*^{-1}d*u$ \/ where $|u|$ denotes
the degree of the differential form $u$. Hence, if $|u|=|v|-1$, then
$$
\phi du\wedge*v \ = \ \phi u\wedge*d^*v \ - \
d\phi\wedge u\wedge *w \ + \ d\,(\phi u\wedge*v)
\tag23 $$
Substituting into (23) first \/ $u=d^*\alp,\, v=\obet$ \/ and
then \/ $u=\obet,\, v=d\alp$ \/ we obtain
$$
\align
\phi dd^*\alp\wedge*\obet \ &= \
-d\phi\wedge d^*\alp\wedge *\obet \ + \
\phi d^*\alp\wedge*d^*\obet \ + \
d\, (\phi d^*\alp\wedge*\obet), \\
\phi d^*d\alp\wedge *\obet \ &= \ \phi \obet\wedge*d^*d\alp \ = \
d\phi\wedge \obet\wedge d\alp \ + \
\phi d\alp\wedge *d\obet \ - \
d\, (\phi \obet\wedge*d\alp).
\endalign
$$
In the last equality we used that $u\wedge*v=v\wedge *u$ for any
differential forms $u,v$ of the same degree. Summing the above
equations, integrating over $M$ and using the Stokes theorem we get
(22).
$\square$\enddemo
Using definition (20) of $I_{R,\eps}$ and Lemma~2 we
obtain
$$
I_{R,\eps} \ = \ \frac1R\, \int_M \, dP_{R,\eps}\wedge \big(
\obet\wedge *d\alp - d^*\alp\wedge*\obet -
\alp\wedge *d\obet + d^*\obet\wedge*\alp \big).
\tag24 $$
Let \/ $d\mu(x)$ \/ denote the Riemannian density on \/ $M$. For any
\/ $\xi\in \bigwedge^k(T^*M)\otimes \CC$ \/ we denote by \/ $|\xi|$ \/
its norm with respect to the scalar product on \/
$\bigwedge^\b(T^*M)\otimes\CC$ \/ induced by the Riemannian structure
on $M$. Then
$$
|\<\alp,\bet\>| \ \le \ \int_M\, |\alp\wedge*\obet|\, d\mu(x)
\ \le \ \int_M\, |\alp|\, |\bet|\, d\mu(x) \ \le \
\|\alp\|\, \|\bet\|
\tag25 $$
for any $\alp,\bet\in \lwe$.
Let us estimate the behavior of the right hand side of (24)
as $\eps\to 0$. For the first term we obtain
$$
\multline
\olim_{\eps\to 0}
\left|\,
\frac1R \int_M\, dP_{R,\eps}\wedge \obet\wedge *d\alp
\right|
\ \le \
\frac1R\, \olim_{\eps\to 0} \int_M \,
|dP_{R,\eps}|\, |d^*\alp|\, |\obet|\, d\mu(x)
\\ \le \
\frac1R\, \int_M \,
|Q^{-1/2} d^*\alp|\, |\obet|\, d\mu(x)
\ \le \
\frac{\|Q^{-1/2}d^*\alp\|\, \|\bet\|}R.
\endmultline
\tag26 $$
In the second inequality in (26) we used the estimate
(19). The last inequality in (26) follows from
Lemma~1.
Analogously, one can estimate the other terms in the right hand side of
(24). That proves (21) and Theorem~A. \
$\square$
%--------------------------------------------------------------------
%--------------------------------------------------------------------
\Refs
\widestnumber\key{BFK3s}
\ref\key BS
\by F.~A.~Berezin, M.~A.~Shubin
\book The Schr\"odinger equation
\publaddr Kluwer, Dordrecht \yr1991
\endref
\ref\key BFS
\by M.~Braverman, M.~Farber, M.~Shubin
\paper The Novikov-Bott inequalities on a manifold with a cylindrical
ends
\jour In preparation
\endref
\ref\key Ga1
\by M.~P.~Gaffney
\paper The Harmonic operators for exterior differential forms
\jour Proc. Nat. Acad. Sci. USA \vol37 \yr1951 \pages48--50
\endref
\ref\key Ga2
\by M.~P.~Gaffney
\paper A special Stokes's theorem for complete Riemannian manifolds
\jour Ann. of Math. \vol60 \yr1954 \pages 140--145
\endref
\ref\key Le
\by B.~M.~Levitan
\paper On a theorem of Titchmarsh and Sears
\jour Usp. Math. Nauk \vol16 \yr1961 \pages 175--178
\endref
\ref\key O1
\by I.~M.~Oleinik
\paper On the essential self-adjointness of the Schr\"odinger operator
on a complete Riemannian manifold
\jour Mathematical Notes \vol 54 \yr1993 \pages934--939
\endref
\ref\key O2
\by I.~M.~Oleinik
\paper On the connection of the classical and quantum mechanical
completeness of a potential at infinity on complete Riemannian
manifolds
\jour Mathematical Notes \vol 55 \yr1994 \pages380--386
\endref
\ref\key RB
\by F.~S.~Rofe-Beketov
\paper Self-adjointness conditions for the Schr\"odinger operator
\jour Mat. Zametki \vol8 \pages741--751 \yr1970
\endref
\ref\key RS
\by M. Reed, B. Simon
\book Methods of modern mathematical physics, Vol. I, II
\publaddr Academic Press, London \yr1978
\endref
\ref\key Se
\by D.~B.~Sears
\paper Note on the uniqueness of Green's functions associated with
certain differential equations
\jour Canadian J. Math. \vol2 \yr1950
\pages314--325
\endref
\ref\key Sh
\by M.~A.~Shubin
\paper Spectral theory of elliptic operators on non-compact manifolds
\jour Ast\'erisque \vol207 \pages 37--108 \yr1992
\endref
\ref\key Wa
\by F.~W.~Warner
\book Foundations of differentiable manifolds and Lie groups
\publaddr Graduate Texts in Mathematics, Springer-Verlag, New-York, Berlin,
Heidelberg, Tokyo \yr1983
\endref
\endRefs
\enddocument