An Integral formula on submanifolds of domains of Cn
Abstract
A Bochner-Martinelli-Koppelman type integral formula on submanifolds of pseudoconvex domains in Cn is derived; the result gives, in particular, integral formulas on Stein manifolds.
Full text
Publicacions
Ma emá iques,
Vol
35
(1991),
559-569
.
AN
INTEGRAL
FORMULA
ON
SUBMANIFOLDS
OF
DOMAINS
OF
Cn
Abs ac
A
Bochne -Ma inelli-Koppelman
ype
in eg al
o mula
on
submani olds
o
pseudocon ex
domains
in
C'
is
de i ed
;
he
esul
gi es,
in
pa icula ,
in eg al
o mulas
on
S ein
mani olds
.
The
me hod
o
in eg al
ep esen a ions
in se e al
complex
a iables
has
been
p o ed
o
be
qui e
e icien
in
cons uc ing
holomo ph_ic
unc ions
and
mo e
gene al analy ic objec s
(di e en ial
o ms
sol ing
he
á-equa ion,
sec ions o
holomo phic
ec o
bundles
e c)
;
see,
o
example,
Henkin and
Lei e e
[4],
[5],
Henkin and
Polyako
[6]
and
Range
[10]
.
This
me hod
e ol es
abou
Bochne -
Ma inelli-Koppelman's
in eg al
o mula
:
i
D
C
C''
ís
a
bounded
domain
in
C'
wi h
smoo h
bounda y
óD
hen
(1
.1)
.=
uAK,-
á
.AK
q
+8(J
UAK
y
-
1)
D
D
D
o
e e y
(0,
q)- o m
u
wi h
C
1
-coe icien s
on D,
whe e
K
9
a e app op ia ely
cons uc ed
ke nels
(see
O elid
[9])
.
The
in eg al
o mula
(1
.1)
canbe
modi ied
o
p oduce
a
a ie y o
o he
o mulas
wi h
which
a ious
p oblems
(such
as
he
á-equa ion
and
in e pola ion
p oblems)
can
be
sol ed
;
hus
(1
.1) is
he
i s
main
s ep
in
se e al
cons uc ions
.
The
pu pose
o his
pape
is
o
cons uc
an
analogue
o
(1
.1)
i
D
is
eplaced
by
a
_submani old
M
o
D
.
Mo e
p ecisely
le
D
_be
a
bounded
domain
in
C'
and
D
a
pseudocon ex
one
wi h
D
C
_D
;
le
M
be
a
(closed
and
comp_lex)
submani old
o
D
and
le
M
=
:
D
l
M
.
Assume
ha
áM
=
(0D)
n
M
_
is
smoo h
.
Then
we
will
cons uc
ke nels
K,«,
z)
de ined
o
z)
E
M
x
M
wi h
(
~
z so
ha
(1
.2)
u
(z)
=
U«)
A
Kq«,
z)-
SEaM
TELEMACHOS
HATZIAFRATIS
1
.
In oduc ion
CEM
S
-bu«)
A
Kq
(S,
z)
+
az
( EM
U(S)
A
K9-1
(
S,
z))
56
0
T
.
HATZIAFRATIS
o
u
E
Cho
a)
(M)
and z
E
M
.
In
[3]
we
de i ed
an
in eg al
o mula
like (1
.2)
in
he
case
M
is
a
comple e
in e sec ion
(i .e
.,
i
he e
exis
unc ions
hl,
..
.,
h
p
,
holomo phic
on
D
_
,
so
ha
M
=
{hl
=
. . .
=
h p
=
0}
and
dhl
n
.
.
.
n
dh
p
:~
0
on
M)
;
he cons uc ion
in
he
gene al
case
(Le
.,
when
M
is
no
necessa ily a
comple e
in e sec ion)
will
be
based
on
he
esul
o
[3]
.
A
simila
cons uc ion
was
ca ied
ou
by
Be nd sson
[1]
.
As
a
ma e
o
ac
ou
ke nels
a e
equi alen
o
hose
o
Be nd sson
bu
w i en
in
a
di e en
way
.
We
hink
howe e
ha
ou
e sion
o
he
cons uc ion
as well as
he
di e en
p oo
o
he
in eg al
o mula
a e o
u he
in e es
.
He e
is
an
ou line
o
he
cons uc ion
.
Fi _s
we
co e
M
by
su icien ly
small
open
se s
{U,}
(open
in
C
-
)
so
ha
each
M
n
U
o
is
a
comple e
in e sec ion
.
Then
he
esul
in
[3]
gi es ke nels
K9
o
which
(1
.2)
holds
in
M
n
U,
.
Such
ke nels
a e
no
unique
in
he
sense
ha
he e
a e
ce ain
choices
ha
can
be
made,
in
pa icula ,
K9
depends
on
He e
decomposi ions
o
he
holomo phic
unc ions
which
de ine
M
n
U
o
(as
he
se
o
he
common
ze os)
.
Bu
using
a
esul
o
Be nd sson
[1]
we
show
ha such
He e
decomposi ions
can
be
chosen
app op ia ely
so
ha
K9
=K9
onmnu,nu,
;
hus
we
can
de ine
a
global ke nel
K
9
(by
se ing
K
9
=
:
K9
on
M
n
U,)
.
Then
we
ha e
o
show
ha
(1
.2)
holds
;
and
his
is
done
along
he
same
lines
as in
[3]]
.
He e
is
an
ou line
o
his
p oo
in
he
case
u
is
a
holomo phic
unc ion
on
M
.
In
his
case
we
ha e
o
show
ha
(1
.3)
u(z)
=
~
u«)Ko«,
z),
z
E
M
CEa~u
Fix
a z
E
M
and
pick
a
U
o
wi h
z
E
U,
.
Bu ,
as
a
compu a ion
shows,
dCK0
=
0
( o
each
T
and
(
7~
z)
and
hence
dC[u«)Kó
«,
z)]
=
0
;
hus
(1
.3) is
equi alen
( ia
S okes'
heo em)
o
u(?)
=
u«)Ko«,
z)
CEaU,
which
holds
by
he
esul
o
[3]
(o
[2]
o
ha
case)
since
Ko
=
Kó
in
U
n
M
.
Ou
esul
gi es,
in
pa icula ,
in eg al
o mulas
o
domains
D
CC
X,
i
X
is
a
S ein
mani old,
ia
he
heo em
ha
S ein
mani olds
admi
embedding
in
some
ON
.
In eg al
o mulas
on
S ein
mani olds
ha e been
cons uc ed
p e iously,
using
di e en
echniques,
by
Henkin
and
Lei e e
[4]
.
Rela ed
is
also
he
wo k
o
Be nd sson
[1],
Ho mann
[7],
Palm
[9]
and
S ou
[11]
.
AN
INTEGRAL
FORMULA
ON
SUBMANIFOLDS
561
2
.
No a ion
Le
h
=
(h1,
. . . .
h
p
)
whe e
he
his a e
holomo phic
unc ions
de ined
in
some
open
se
o
Cn
and
suppose
ha
{hij
(~, z),
j
=
1,
. . . ,
n}
is
a He e
decomposi-
ion o
h
i
,
Le
.,
h?j
«,
z)
is
holomo phic
in
bo h
(
and
z
and
n
hi(S)
-
hi(z)
=
E
hij(S,z)(Sj
-
zj),i
=
1,
. .
.,p
.
j-1
We
associa e
o
hese da a he
ollowing
di e en ial
o ms
.
Fi s
n-P-1
5D(2
.1)
1
ah«,
z
)
=
cde [hlj,
. . .
,
h
p
j,
yj,
(C
+
z)yj]
whe e
y
=
(yl,
. . . ,
yn
)
is
some
smoo h
unc ion
(we
will
be mo e
speci ic
abou
y
la e )
and
c
is
no malizing
cons an
:
c
=
(-1)
-(-2-
1)
1
1
whe e
m
=
:
(2- i
^~
'
T c!
n-
p
.
In
he
de e minan
o
(2
.1),
j uns
om
j
=
1
up
o
j
=
n
o ming
he
(gis
+
az)yl
n
ows
o
i
;
also
he
column
o
di e en ial
o ms
in
he
de e minan
(n
-p
-
1)- imes
(as
i is
de e minan s
see
[4]
.
Also
de ine
(2
.2)
whe e
(a~
+
"
C7z)yn
n
-P
1
8h
1
h
ah
¡oh(C)12
de
á~P
,
d(j
Ioh(S)1
2
=
y
:
1
a(h1
. . . .
.
,
hP)
(~)
2
.
1<j
l
<
. .
.<jp<n
a(Sjl,
. . . ,
(j )
indica ed)
;
o
p ope ies
o
such
o
cou se
Q'«)
is
de ined
o
(
wi h
Joh(«
:yÉ
0
.
Wi h
his
no a ion
se
K
h
(~,
z)
=
ah«~
z)
n
~3h(~)
.
Le
ce,
be
he
pa
o
a
h
which
is
o
ype
(0,
q)
in
z
and
(0,
n
-p-
q
-
1)
in
~,
i
.
e
.,
Also
le
Kq
(~,
z)
=
a9
«,
z)
nQ
h
(~)
.
-
)
_~-
_
c q
=
c
C
n
q
-
1
de [hlj,
. . . .
h
p
j,
yj,
azyj,
aCyj
is
epea ed
O
cou se
Kq
depends
on
he
choice
o
he
He e
decomposi ions
{hij}
o
hi,
al hough
we
do no
indíca e
his
in
he
no a ion
;
bu
i
will
be
clea
om
he
con ex
which
choice
we
will
be
using
in
each
case
.
Also
we
ha e
been
agueabou
he
domains
in
which
he
abo e
unc ions
and
di e en ial
o ms
a e
de ined,
because
he e,
we
simply in oduced
no a ion
and
we
will
be
e y
speci ic
abou
his
la e
.
56
2
T
.
HATZIAFRATIS
3
.
Cons uc ion
o
he
ke nels
_
Fi s
we
desc ibe
he
se ing
.
Le_
D,
D
be
domains
in
C"
wi h
D
bound_ed,
D
pseudocon ex
and
D
C
D
.
Le
M
be
a
closed
complex
su_bmani old
o
D
o
(complex)
dimension
m
._
Se
M
=
DnM
and
aM
=
(OD)nM
and assume
ha
aD
is
smoo h
and
ha
M
mee s
aD
ans e sally
so
ha
aM
is
also
smoo h
.
Le
,y
=
(-y,,
. .
.,
-yn)
:
D
x
D
-
{~
=
z}
-
C'
be
a
smoo h
map
sa is ying
(3
.1)
((j
zj)
=
1
and
(-
z'
o
0
<
j(
-
zi
<
6
o
some
small
6
>
0
j=1
zl2
n
(j
7j
^y
.,
I-
(an
examPle
o
ls
g
.
l en
bY
-z'
all
(,
z
wi h
(
z)
.
7
hj
=
Z
In
his
se ing
we
will
now
con
_
s uc
he
ke nels
.
Le
{U
a
}
be a
se
o
small
con ex
open
se s
o
en
so
ha
M
C
U
a
U
aand
mo eo e
le
ha
=
(he,',
...
.
ha)
:
D
~
CP,
p=
:
n-
m,
be
holomo phic
maps
so
ha
MnU,
=
{z
E
U
a
:
ha
(z)
=
0} and
_
17h
a
l
~
0on
M
_
nU
a
;
ha
such
unc ions
há
exis
ollows
om
Ca an's
Theo em
A
since
D
is
assumed
o
be
pseudocon ex
.
Fu he mo e
he e
exis
p x p
ma ices
A,T
=
[(A ,)ik]1<i,k<p
o
unc ions
hi
~
í
hi
(A,T)ik,
holomo phic
in
U
a
n
UT,
so
ha
A,
T
Le
.,
(3
.2)
The
exis ence
o
such
ma ices
A
.T
ollows
om
Ca an's
Theo em
B
since
U
a
n
U
T
is
con ex
.
Lemma
1
.
ah
;
ah,
;
_
A
uT
~eu
a
nuTnM
.
ahp ahp
ac
;
aS
;
P oo
::
Di e en ia ing
(3
.2)
we
ob ain
ahá
a5j
ahk
(A,T)ik
-
+
k=1
a(j
k=1
p
há
=
E(A,T)ikhk,
i
=
1,
. .
.
,
p
.
k=1
hT a(AoT)ik
.
k
a(j
on
u,
nu
T
,
Since
hk
=
0
o
C
E
U
Q
n
U
T
n
M,
he
o mula
o
he
lemma
ollows
immedi-
a ely
.
The
ollowing
lemma
is
p o ed
by
Be nd sson
[1,
p
.414]
;
i_ s
P oo
is
based
on
Ca an's
Theo em
B,
using
again
he
pseudocon exi y
o
D
.
Lemma
2
.
The e
exis
unc ions
h
~
«,
z),
i
=
1,
. . . .
p,
j
=
1,
. . . ,
n,
holo-
mo phic
in
(C,
z)
E
(U,
n
M)
x
M
so
ha
hij
i
ihlj
(3
.3)
1
h-
j
j
h-
pi
Fu he mo e
h
~
«,
z)
Na e
holomo phic
ex ensions
in
C
in
a
neighbou hood
(in
en)
o
U
a
n
M,
sa is ying
Then
T
P oo
.
De ining
AN
INTEGRAL
FORMULA
ON
SUBMANIFOLDS
563
n
o j=1,
_,
n,CEu,nu,nM,zEM
(3
.4)
1
:
h
íj
(C,
z)
(SJ
-
zj)
=
h°
«)
o
zE
M
.
j=1
Lemma
3
.
Le
A
E
CPXP
and
B,
C
E
CnxP
so
ha
A
-
BT
=
C
T
(T
deno es
anspose)
.
and,
hence,
a
hh
'
=
de (A,)
de [C,
* *
*]
=
de (A)
de [B,
* *
*]
whe e
*
*
*
deno e
app op ia e
di e en ial
o ms
( he
same
on
bo h
sides
o
he
equa ion)
so
ha
he
ma ices
[B,
* *
*]
and
[C,
* *
*]
a e
n
x
n
.
P oo
:
This
is
a
s aigh o wa d
compu a ion
based
on
he
mul ilinea i y
o
he
de e minan s
(as
unc ions
o
hei
columns)
and
he
de ini ion
o
he de-
e minan s
.
(This
gene alizes
he
ac
ha
he
de e minan
o
he
p oduc
o
squa e
ma ices
is
equal
o
he
p oduc
o
he
de e minan s
o
hese
ma i-
ces)
.
Lemma
4
.
I
aho
is
he
o m
(2
.1)
associa ed
o
he
He e
decomposi ions
{h9'j
ande¿
h
is
de ned
simila ly,
hen
(3
.5)
ah'
=
de (A,
T
)
.
a
h
'
o (
E
U
o
n
U
T
n
M,
z
E
M
and
simila ly
H
T
,
we
ob ain
om
Lemma
2
ha
A
oT
-
(HT)
T
=
(Ha
)
T
o
(E
U,
n
U,
n
ú,
z
E
M
;
hus
Lemma
3
applies
and
gi es
(3
.5)
.
564
T
.
HATZIAFRATIS
Lemma
5
.
We
ha e
(3
.6)
ph
=
P oo
..
By
Lemma
1
we
ob ain
The e o e
1
de (
ph
°
E
U
Q
n
U
Tn
M
.
ah°
ahi
=A
aT
]
~N9k
1<i,k<p
N
.9k
1<i,k<p
.
a(hi,
. .
.,hP)
=
de (AQT)
a(hi
...
.,hP)
,
~
;
p
)
a«il
,
.. . ,
~
;p)
and
(3
.7)
j7h
o
l
2
=
1
de (A,)1
2
jVhT1
2
( his
also
shows
ha
de (A,
T
) ,-~
0
on
U
Qn
U
T
n
M)
.
Also
se ing
and
simila ly
o
BT
we
see
ha
Lemma
1
gi es
Hence,
by
Lemma
3,
8h1
aC,
. . .
acá
&h,'
ah,
aC
. . .
aCn
1T
.
(
B
T
)
T
=
(Bu
)T
(3
.8)
de
ah1
,
. . . ,
ah
P
C
;
a(
;
,
d
=
de (A,)
-
de
Now
(3
.6)
ollows
om
(2
.2),
(3
.7)
and
(3
.8)
.
We
a e
now
eady
o de ine
he
ke nels
:
n-p
ahi
ahT
~
,
a(,
. . .
,
,
dj
a(j
K«,
z)
=
:
K
h
'
(~,
z)
=
ah «,
z)
A
Qho
(~)
o
(
E
U
Qn
M,
z
E
M,
~
,~
z
.
By
Lemmms
4and
5,
i
ollows
ha
K(~,
z)
is
well-de ined
o
E
M,
z
E
M,
z
.
Simila ly
we
de ine
K
9
=
aq
,Q
h
o
q
>
0 and
K_1
=
0
.
AN
INTEGRAL
FORMULA
ON
SUBMANIFOLDS
565
4
.
The
in eg al
o mula
Wi h
he
ke nels
jus
cons uc ed
we
will
p o e
he
ollowing
heo em
.
Theo em
1
.
I
u
E
Cho
e)
(M)
(0
<
q
<
m)
and
z
E
M
hen
u(z)
=
~
u«)
n
Kq(C,
z)-
(EaM
Fi s
a
lemma
:
Lemma
6
.
We
ha e
(
,9
c
+
c
9z)K(S,
z)
=
0
.
P oo
.
:
I
su ices
o
show
(4
.1)
(8S
+
á
z
)
K
ho
(C,
z)
=
0
o
cE
U
o
l
M,
zE
M,
c
=~
z
.
Now
obse e
ha
(4
.2)
«I
-
zj)a
ho
(S,
z)
=
de
P oo
o
Theo em
1
:
I
su ices
o
show
ha
o
W
E
(Có
(M))(m,
m
-
q)
.
CEM
-
bu«)
A
K9
(S,
z)
+
áz
IEM
u«)
A
K
9
-1
(C,
z)]
.
(4
.3)
J
U(Z)
A
W(z)
=
J
EBM
u
n
K
q
A
W-
zEM
zEM
~
n,-
-1
hay
.
.
.
ha
y
7j
(ac
+
az)yi
J
2<j<n
(in
he
abo e
de e minan
j
uns
om
j
=
2 o
j
=
n
o ming
he
2nd
up o
he
n h
ow
o
i )
.
We
ob ained
(4
.2)
in
he
ollowing
way
:
(SI
-
z1) mul iplied
he
i s
ow
o
he
de e minan
which
de ines
aho
«,
z)
and
hen
we
added
o
ha
i s
ow
he
j h- ow
mul iplied
by
(Si
-
zj)
(j
=
2,
.
.
. ,
n)
.
Then
(4
.2)
ollows in
iew
o
he
i s
o
he
assump ions
in (3
.1)
and
(3
.4)
.
Now
(4
.2)
easily
implies
ha
(á
C
+
Ó
z
)
a
ho
=
0
;
since,
mo eo e ,
5CQ«)
=
0
(by
[2,
Co olla y
1,
p
.
76]),
we
ob ain
(4
.1)
which
comple es
he
p oo
o
he
lemma
.
áuAK
y
Acp+
J
(az(]
uAK
9
-1))A9
zemCEM
zEM
(EM
56
6
T
.
HATZIAFRATIS
Le
us
poin
ou
he
way
in
which
he
a ious
o ms
in
he
igh -hand
side
o (4
.3)
dependon
he
a iables
~
and
z
:
u_
«5u«)
u«),
K
9
=
K,«,
z),
K
9
-1
=
Kq-1
«,
z),
cp
=
;o(z)
and
8u
=
.
By
deg ee
easons,
(4
.3)
is
equi alen
. o
(4
.4)
L
u(z)
n
w(z)
=
u
n
K
A
cp-
EM
ó(MxM)
(in
ob aining
(4
.4)
we
used
also
he
ac
ha
J
lS~,z)
uAKAW=
uAKAW
E(óM)xM
8
(M
x
M)
which
holds
since
u
has
compac
suppo
in
M)
.
By
S okes'
o mula
(4
.5)
a(mxm)uAKAW=Ld[UAKAW]+L
uAKA<p
whe e
C
E
=
{((,
z)
E
M
x
M
:
j(
-
zi
=
E}
.
By
Lemma
6
and
deg ee
easons
L
buAKA(p+~
(áz(J
uAK))AW
MxM
M M
1
~~
d[uAKAW]
=
auAKAW-(-1)9J
uAKn77cp
;
MxM-{~~-z~<E}
Mxm
MxM
hence
(4
.4)
will
ollow
om
(4
.5)
as
soon
as
we
es ablish
he
ollowing
(4
.6)
lim
u
A
K
A
;o
=
u(z)
n
W(z)
.
E-0
£,z)EQ
.
EM
Bu
we
may
assume
wi hou
loss
o
gene ali y
ha
supp(u)
C
Ua
o
some
u
.
Then
o
E
small
enough
uAKAcw=
uAK
h
'n
;o
C
E
C
E
and
(4
.6)
ollows
om
(4
.7)
lim
uA
Kho
A
cp
=
u
(z)
1
w(z)
E-0
c
E
IzEM
Bu
(4
.7)
is
exac ly
wha
is
p o ed
in
[3,
p
.
339
341]
(in
ha pa
o he
p oo ,
(3
.4)
plays
an
impo an
ole)
.
This comple es he p oo
o
he
heo em
.
his
is
an
analogue
o
Cauchy's
in eg al
o mula
on
M
.
Con inuing
o
assume
D
o
be
s ic ly
pseudocon ex
and
he
gj's
as
abo e,
le
us
se
and
AN
INTEGRAL
FORMULA
ON
SUBMANIFOLDS
567
As
we
men ioned
in
he
in oduc ion
he
in eg al
o mula
ha
we
p o ed
can be
used
o
de i e
a ious
o he
in eg al
o mulas
.
He e
we
discuss
wo
examples
.
Wi h
no a ion
as
be o e
assume
u he mo e
ha
D
is
a
s ic ly
pseudocon ex
domain
.
Then,
by a
classical
cons uc ion
o
Henkin and Rami ez
(see
[51),
he e
exis
unc ions
g
j
(~, z),
j
=
1,
...
,
n
de ined
o
«,
z)
E
(aD)
x
D,
which
a e
smoo h
in (
and
holomo phic
in
z,
so
ha
G(~,
z)
_
:
~~
i
(~
j
-z
j
)gj
«,
z)
0
.
Thus
i
we
se
n-p-1
C«,
z)
_
[C(~,
z)1
-p
de [hi ,
. .
. ;
h'i,
g~,
DZgj
1
A
aho
(C)
o
(
E
U
naM
and
z
E
M
hen
C«,
z)
is
w
ell-de ined
o
E
aM
and
z
E
M,
holomo phic
in z
and
(as
i
ollows' om
heo em
1)
i
ep oduces
holomo phic
unc ions
on
M,
Le,
o
E
O(M)
l
C(M)we
Na e
?7,(S,z,~)=(1-
.1)
i'-
zI2+aG(S'z)
o (EaD,zeD,
.1E[0,11
L9(~,z,, )=CCn-p-1/
4
5
.
Applica ions
(z)
=
CEaM
(0C«,
z),
9
n-P-9-1
de
[hi ,
. . . ,
hpi,
l
a
;,
(aS~+
da
n
ah~
E
U
o
n
aM,
z
E
M
;
hen
L
.«,
z,
.
) is
well-de ined
o
-C
E
8M,
z
E
M
.
Also
se
B,«,
z)
=
:
L«,
z,
)1>,-o
;
zeM
;
hen
i is
clea
ha
B,(~,
z)
is
de ined
o
~,
z
E
M,
(
:,A
z
.
Using
he
abo e
ke nels
we
de ine
he
ollowing
ope a o s
:
T
9
u(z)
=
u(S)
A
Le
-
1( S,
z, )
+
u
(S)
n
BQ-1(S,z),z
E
M
(C,a)E(aM)X[0,1]
CEM