The Kastler-Kalau-Walze type

theorem for 6-dimensional manifolds with boundary

Jian Wang, Yong Wang∗

School of Mathematics and Statistics, Northeast Normal University, Changchun, 130024, P.R.China
arXiv:1211.6223v2 [math.DG] 27 Apr 2015

Abstract
In this paper, we define lower dimensional volumes of spin manifolds with boundary. We compute the lower
(1,3)
dimensional volume Vol6 for 6-dimensional spin manifolds with boundary and the gravity on boundary is
derived by the noncommutative residue associated with Dirac operators. For 6-dimensional manifolds with
boundary, we also get a Kastler-Kalau-Walze type theorem for a general fourth order operator .
Keywords: lower-dimensional volumes; noncommutative residue; gravitational action; perturbations of
Dirac operators.
2000 MSC: 53G20, 53A30, 46L87

1. Introduction
The noncommutative residue plays a prominent role in noncommutative geometry [1][2]. Connes [3]
used the noncommutative residue to derive a conformal 4-dimensional Polyakov action analogy. Connes [4]
proved that the noncommutative residue on a compact manifold M coincided with the Dixmier’s trace on
pseudodifferential operators of order −dimM . Several years ago, Connes made a challenging observation
that the noncommutative residue of the square of the inverse of the Dirac operator was proportional to the
Einstein-Hilbert action, which we call the Kastler-Kalau-Walze theorem. Kastler[5] gave a brute-force proof
of this theorem. Kalau and Walze [6] proved this theorem in the normal coordinates system simultaneously.
Ackermann [7] gave a note on a new proof of this theorem by means of the heat kernel expansion.
Recently, Ponge defined lower dimensional volumes of Riemannian manifolds by the Wodzicki residue
[8]. Fedosov et al defined a noncommutative residue on Boutet de Monvel’s algebra and proved that it
was a unique continuous trace [9]. Wang generalized the Connes’ results to the case of manifolds with
boundary in [10] [11] , and proved a Kastler-Kalau-Walze type theorem for the Dirac operator and the
signature operator for lower-dimensional manifolds with boundary[12] [13]. The purpose of papers [12] [13]
is to derive the gravitational action by the noncommutative residue associated with Dirac operators for spin
manifolds with boundary, but the boundary term vanished in Wang’s results. The motivation of this paper
is to derive the gravitational action on boundary by the noncommutative residue associated with Dirac
operators. In other words, we want to get a nonvanishing boundary term. In [12] [13], Wang computed
] + D−1 ◦ π + D−1 ] and Wres[π
Wres[π ] + D−2 ◦ π + D−2 ], where the two operators are symmetric. In this paper,
] + D−1 ◦π + D−3 ]. Since D−1 and D−3 are not
for 6-dimensional manifolds with boundary, we compute Wres[π
symmetric, the boundary term is the integral of the extrinsic scalar curvature and the gravitational action
on boundary emerges. We also get a generalized Kastler-Kalau-Walze theorem associated with a general
fourth order operators for 6-dimensional spin manifolds.
This paper is organized as follows: In Section 2, we define lower dimensional volumes of spin manifolds
with boundary. In Section 3, for 6-dimensional spin manifolds with boundary and the associated Dirac

∗ Corresponding author; Email address: [email protected] (Yong Wang)

Email address: [email protected] (Jian Wang)

Preprint submitted to Elsevier October 12, 2018

 (1,3)
operator D and D3 , we compute the lower dimensional volume Vol6 and get a Kastler-Kalau-Walze type
theorem in this case. In Section 4, we get a Kastler-Kalau-Walze type theorem associated with the fourth
order operators for 6-dimensional spin manifolds.

2. Lower dimensional volumes of spin manifolds with boundary

In order to define lower dimensional volumes of spin manifolds with boundary, we need some basic
facts and formulae about Boutet de Monvel’s calculus and the definition of the noncommutative residue for
manifolds with boundary. We can find them in Section 2,3 [11] and Section 2.1[12].
Let M be a n-dimensional compact oriented spin manifold with boundary ∂M . We assume that the
metric g M on M has the following form near the boundary,
1
gM = g ∂M + dx2n , (2.1)
h(xn )

where g ∂M is the metric on ∂M . Let D be the Dirac operator associated to g on the spinors bundle
S(T M )[12]. Let p1 , p2 be nonnegative integers and p1 + p2 ≤ n.
Definition 2.1. Lower dimensional volumes of spin manifolds with boundary are defined by

] + D−p1 ◦ π + D−p2 ]
Voln(p1 ,p2 ) M := Wres[π (2.2)

where the related definitions, see Section 2, 3[12].

Denote by σl (A) the l-order symbol of an operator A. By (2.1.4)-(2.1.8)[12], we get
Z Z Z
] + D−p1 ◦ π + D−p2 ] =
Wres[π traceS(T M) [σ−n (D−p1 −p2 )]σ(ξ)dx + Φ, (2.3)
M |ξ|=1 ∂M

and
Z Z +∞ ∞ X
X (−i)|α|+j+k+1
Φ = × traceS(T M) [∂xj n ∂ξα′ ∂ξkn σr+ (D−p1 )(x′ , 0, ξ ′ , ξn )
|ξ ′ |=1 −∞ j,k=0 α!(j + k + 1)!

×∂xα′ ∂ξj+1
n
∂xkn σl (D−p2 )(x′ , 0, ξ ′ , ξn )]dξn σ(ξ ′ )dx′ , (2.4)

where the sum is taken over r − k − |α| + l − j − 1 = −n, r ≤ −p1 , l ≤ −p2 .

Since [σ−n (D−p1 −p2 )]|M has the same expression as σ−n (D−p1 −p2 ) in the case of manifolds without
boundary, so locally we can use the computations [5], [6], [8], [12] to compute the first term. The following
proposition is the motivation of the definition of lower dimensional volumes of spin manifolds with boundary
[13].
Proposition 2.2. [13] Lower dimensional volumes of spin manifolds with boundary are given by
1) When p1 + p2 = n, then Voln(p1 ,p2 ) M = c0 VolM .
R
2) when p1 + p2 ≡ nmod1, Voln(p1 ,p2 ) M = ∂M Φ.
3) Z
(1,1) Ω4 (1,1)
Vol4 =− sdvolM ; Vol3 = c1 Vol∂M (2.5)
3 M
where c0 , c1 are constants and s is the scalar curvature.

2
3. A Kastler-Kalau-Walze type theorem for 6-dimensional spin manifolds with boundary
(1,3)
In this section, We compute the lower dimensional volume Vol6 for 6-dimensional spin manifolds with
boundary and get a Kastler-Kalau-Walze type theorem in this case.
Firstly, we compute the symbol σ(D−3 ) of D−3 . Recall the definition of the Dirac operator D [9][14].
Let ∇L denote the Levi-Civita connection about g M . In the local coordinates {xi ; 1 ≤ i ≤ n} and the fixed
orthonormal frame {ee1 , · · · , f
en }, the connection matrix (ωs,t ) is defined by

∇L (ee1 , · · · , f
en ) = (ee1 , · · · , f
en )(ωs,t ). (3.1)

The Dirac operator is defined by

n
X h 1X i
D= c(eei ) eei − ωs,t (eei )c(ees )c(eet ) . (3.2)
i=1
4 s,t

where c(eei ) denotes the Clifford action.

Recall the definition of the Dirac operator D2 in [5], [6] and [13], we have
X h i 1
D2 = − g i,j ∂i ∂j + 2σi ∂j + (∂i σj ) + σi σj − Γki,j ∂i − Γki,j σk + s. (3.3)
i,j
4
P
where σi := − 41 s,t ωs,t (∂i )es et .
Combining (3.2) and (3.3), we have
n
X  X X
D3 = c(eel )hel , dxl i (− g i,j ∂i ∂j )∂l − ∂l (g i,j )∂i ∂j
i=1 i,j i,j
h X i X
+ − g i,j (2σi ∂j − Γki,j ∂i ) ∂l − ∂l (g i,j )(2σi ∂j − Γki,j ∂i )
i,j i,j
X  
− g i,j 2(∂l σi )∂j − ∂l (Γki,j )∂i
i,j
h X   1 i
+ − g i,j (∂i σj ) + σi σj − Γki,j σk + s ∂l
i,j
4
X   1
− ∂l (g i,j ) (∂i σj ) + σi σj − Γki,j σk + ∂l s
i,j
4
X   1 
i,j k
− g ∂l (∂i σj ) + σi σj − Γi,j σk + ∂l s
i,j
4

1X X X
− ωs,t (eel )c(el )c(ees )c(eet ) − g i,j ∂i ∂j − g i,j (2σi ∂j − Γki,j ∂i )
4 s,t i,j i,j
X   1 
− g i,j (∂i σj ) + σi σj − Γki,j σk + s . (3.4)
i,j
4

3
Then, we obtain
√
σ3 (D3 ) = −1c(ξ)|ξ|2 , (3.5)
1 X
σ2 (D3 ) = c(ξ)(2σ k − Γk )ξk − |ξ|2 ωs,t (eel )c(el )c(ees )c(eet ), (3.6)
4 s,t
Xn  X
3
σ1 (D ) = c(eel )hel , dxl i − ∂l (g i,j )(2σi ∂j − Γki,j ∂i )
i=1 i,j
X X  
i,j
+ g ∂l (Γki,j )∂i − g i,j 2(∂l σi )∂j
i,j i,j
h X   1 i 
+ − g i,j (∂i σj ) + σi σj − Γki,j σk + s ∂l
i,j
4
1X h X i
− ωs,t (eel )c(el )c(ees )c(eet ) − g i,j (2σi ∂j − Γki,j ∂i ) , (3.7)
4 s,t i,j
Xn  X   1
σ0 (D3 ) = c(eel )hel , dxl i − ∂l (g i,j ) (∂i σj ) + σi σj − Γki,j σk + ∂l s
i=1 i,j
4
X   1 
− g i,j ∂l (∂i σj ) + σi σj − Γki,j σk + ∂l s
i,j
4
1 X h X   1 i
− ωs,t (eel )c(el )c(ees )c(eet ) − g i,j (∂i σj ) + σi σj − Γki,j σk + s . (3.8)
4 s,t i,j
4

Write
∞
X
√
Dxα = (− −1)|α| ∂xα ; σ(D3 ) = p3 + p2 + p1 + p0 ; σ(D−3 ) = q−j . (3.9)
j=3

By the composition formula of psudodifferential operators, we have

X 1
1 = σ(D3 ◦ D−3 ) = ∂ξα [σ(D)]Dxα [σ(D−1 )]
α
α!
= (p3 + p2 + p1 + p0 )(q−3 + q−4 + q−5 + · · · )
X
+ (∂ξj p3 + ∂ξj p2 + ∂ξj p1 + ∂ξj p0 )(Dxj q−3 + Dxj q−4 + Dxj q−5 + · · · )
j
X
= p3 q−3 + (p3 q−4 + p2 q−3 + ∂ξj p3 Dxj q−3 ) + · · · , (3.10)
j

Then we obtain X
q−3 = p−1 −1 −1
3 ; q−4 = −p3 [p2 p3 + ∂ξj p3 Dxj (p−1
3 )]. (3.11)
j

By Lemma 2.1 in [12] and (3.4)-(3.11), we obtain

4
Lemma 3.1.
√
−1c(ξ)
σ−1 (D−1 ) = ; (3.12)
|ξ|2
c(ξ)σ0 (D)c(ξ) c(ξ) X  
σ−2 (D−1 ) = + c(dx j ) ∂xj [c(ξ)]|ξ| 2
− c(ξ)∂ xj (|ξ| 2
) ; (3.13)
|ξ|4 |ξ|6 j
√
−1c(ξ)
σ−3 (D−3 ) = ; (3.14)
|ξ|4
√
c(ξ)σ2 (D3 )c(ξ) −1c(ξ)  4
σ−4 (D−3 ) = + |ξ| c(dxn )∂xn c(ξ ′ ) − 2h′ (0)c(dxn )c(ξ)
|ξ|8 |ξ|8

+2ξn c(ξ)∂xn c(ξ ′ ) + 4ξn h′ (0) , (3.15)
P
where σ0 (D) = − 14 s,t ωs,t (eei )c(eei )c(ees )c(eet ).
Since Φ is a global form on ∂M , so for any fixed point x0 ∈ ∂M , we can choose the normal coordinates
U of x0 in ∂M (not in M ) and compute Φ(x0 ) in the coordinates U e = U × [0, 1) ⊂ M and the metric
1 ∂M 2 M e
+ dxn . The dual metric of g on U is h(xn )g ∂M 2
+ dxn . Write gijM ∂
= g M ( ∂x ij
, ∂ ); gM =
h(xn ) g i ∂xj

g M (dxi , dxj ), then

 1   
[g ∂M ] 0 i,j
i,j
h(xn )[g∂M ] 0
M
[gi,j ] = h(xn ) i,j ; [gM ]= , (3.16)
0 1 0 1
and
∂M M
∂xs gij (x0 ) = 0, 1 ≤ i, j ≤ n − 1; gij (x0 ) = δij . (3.17)
∂M
Let n = 6 and {e1 , · · · , en−1 } be an orthonormal p frame field in U p about g which is parallel along
∂
geodesics and ei (x0 ) = ∂x i
(x 0 ), then { e
e 1 = h(x )e
n 1 , · · · , ]
e n−1 = h(xn n−1 f
)e , en = dxn } is the or-
e M ∼ e ∗ n
thonormal frame field in U about g . Locally S(T M )|Ue = U × ∧C ( 2 ). Let {f1 , · · · , f8 } be the orthonor-
mal basis of ∧∗C ( n2 ). Take a spin frame field σ : U e → Spin(M ) such that πσ = {ee1 , · · · , ef n }, where
π : Spin(M ) → O(M ) is a double covering, then {[(σ, fi )], 1 ≤ i ≤ 8} is an orthonormal frame of
S(T M )|Ue . In the following, since the global form Φ is independent of the choice of the local frame, so we
can compute trS(T M) in the frame {[(σ, fi )], 1 ≤ i ≤ 8}. Let {E1 , · · · , En } be the canonical basis of Rn and
c(Ei ) ∈ clC (n) ∼
= Hom(∧∗C ( n2 ), ∧∗C ( n2 )) be the Clifford action. By [Y], then

∂ ∂
c(eei ) = [(σ, c(Ei ))]; c(eei )[(σ, fi )] = [(σ, c(Ei )fi )]; = [(σ, )], (3.18)
∂xi ∂xi
∂
then we have ∂xi c(eei ) = 0 in the above frame. By Lemma 2.2 in [12], we have
Lemma 3.2. [12] For n-dimensional spin manifolds with boundary,

∂xj (|ξ|2gM )(x0 ) = 0, if j < n; = h′ (0)|ξ ′ |2g∂M , if j = n. (3.19)

′
∂xj [c(ξ)](x0 ) = 0, if j < n; = ∂xn [c(ξ )](x0 ), if j = n, (3.20)

where ξ = ξ ′ + ξn dxn .
Next we compute σ0 (D)(x0 ). By Lemma 2.3 in [12], we have
Lemma 3.3. [12] For 6-dimensional spin manifolds with boundary,
5
σ0 (D)(x0 ) = − h′ (0)c(dxn ). (3.21)
4
5
 Now we can compute Φ (see formula (2.4) for the definition of Φ), since the sum is taken over −r − l +
k + j + |α| = 5, r ≤ −1, l ≤ −4, then we have the following five cases:
case a) I) r = −1, l = −3 k = j = 0, |α| = 1
From (2.4) we have
Z Z +∞ X
case a) I) = − trace[∂ξα′ πξ+n σ−1 (D−1 ) × ∂xα′ ∂ξn σ−3 (D−3 )](x0 )dξn σ(ξ ′ )dx′ . (3.22)
|ξ ′ |=1 −∞ |α|=1

By Lemma 3.2, for i < n, then

√
−3 −1c(ξ)
∂xi σ−3 (D )(x0 ) = ∂xi [ ](x0 )
|ξ|4
√ √
−1∂xi [c(ξ)](x0 ) 2 −1c(ξ)∂xi [|ξ|2 ](x0 )
= − = 0. (3.23)
|ξ|4 |ξ|6
Then case a) I) vanishes.
case a) II) r = −1, l = −3 k = |α| = 0, j = 1
From (2.4) we have
Z Z +∞
1
case a) II) = − trace[∂xn πξ+n σ−1 (D−1 ) × ∂ξ2n σ−3 (D−3 )](x0 )dξn σ(ξ ′ )dx′ . (3.24)
2 |ξ′ |=1 −∞
By (2.2.23) in [12], we have
∂xn [c(ξ ′ )](x0 ) √ h ic(ξ ′ ) c(ξ ′ ) + ic(dxn ) i
πξ+n ∂xn σ−1 (D−1 )(x0 )||ξ′ |=1 = + −1h′ (0) + . (3.25)
2(ξn − i) 4(ξn − i) 4(ξn − i)2
By (3.14), we obtain
√  (20ξn2 − 4)c(ξ ′ ) + 12(ξn3 − ξn )c(dxn ) 
∂ξ2n σ−3 (D−3 ) = −1 ; (3.26)
(1 + ξn2 )4
Since n = 6, trS(T M) [id] = dim(∧∗ (3)) = 8. By the relation of the Clifford action and trAB = trBA, then
we have the equalities:
tr[c(ξ ′ )c(dxn )] = 0; tr[c(dxn )2 ] = −8; tr[c(ξ ′ )2 ](x0 )||ξ′ |=1 = −8;
tr[∂xn c(ξ ′ )c(dxn )] = 0; tr[∂xn c(ξ ′ )c(ξ ′ )](x0 )||ξ′ |=1 = −4h′ (0). (3.27)
By (3.25)-(3.27) and direct computations ,we have
nh ∂ [c(ξ ′ )](x ) √  ic(ξ ′ ) c(ξ ′ ) + ic(dxn ) i
xn 0
tr + −1h′ (0) +
2(ξn − i) 4(ξn − i) 4(ξn − i)2
h√  (20ξ 2 − 4)c(ξ ′ ) + 12(ξ 3 − ξ )c(dx ) io
n n n n
× −1 (x0 )||ξ′ |=1
(1 + ξn2 )4
−8 − 24iξn + 40ξn2 + 24iξn3
= h′ (0) . (3.28)
(ξn − i)6 (ξn + i)4
By (3.24) and (3.28), we obtain
Z Z +∞
4 + 12iξn − 20ξn2 − 12iξn3
case a) II) = h′ (0) dξn σ(ξ ′ )dx′
|ξ ′ |=1 −∞ (ξn − i)6 (ξn + i)4
Z
4 + 12iξn − 20ξn2 − 12iξn3
= h′ (0)Ω4 dξn dx′
Γ+ (ξn − i)6 (ξn + i)4
2πi h 4 + 12iξn − 20ξn2 − 12iξn3 i(5)
= h′ (0)Ω4 |ξn =i dx′
5! (ξn + i)4
15
= − πh′ (0)Ω4 dx′ , (3.29)
16
6
where Ω4 is the canonical volume of S 4 .
case a) III) r = −1, l = −3 j = |α| = 0, k = 1

From (2.4) we have

Z Z +∞
1
case a) III) = − trace[∂ξn πξ+n σ−1 (D−1 ) × ∂ξn ∂xn σ−3 (D−3 )](x0 )dξn σ(ξ ′ )dx′ . (3.30)
2 |ξ ′ |=1 −∞

By (2.2.29) in [12], we have

c(ξ ′ ) + ic(dxn )
∂ξn πξ+n σ−1 (D−1 )(x0 )||ξ′ |=1 = − . (3.31)
2(ξn − i)2

By (3.14), we obtain
h (1 − 5ξ 2 )c(dx ) √
√ n 6ξn c(ξ ′ ) i 4 −1ξn ∂xn c(ξ ′ )(x0 )
∂ξn ∂xn σ−3 (D−3 )(x0 )||ξ′ |=1 = −2 −1h′ (0) n
− − . (3.32)
(1 + ξn2 )4 (1 + ξn2 )4 (1 + ξn2 )3

By (3.27), (3.31) and (3.32), we obtain

nh c(ξ ′ ) + ic(dx ) i h √  (1 − 5ξ 2 )c(dx ) √
n ′ n n 6ξn c(ξ ′ )  4 −1ξn ∂xn c(ξ ′ )(x0 ) io
tr × 2 −1h (0) − + (x0 )||ξ′ |=1
2(ξn − i)2 (1 + ξn2 )4 (1 + ξn2 )4 (1 + ξn2 )3
8i − 32ξn − 8iξn2
= h′ (0) . (3.33)
(ξn − i)5 (ξn + i)4

Then
Z Z +∞
−4i + 16ξn + 4iξn2
case a) III) = h′ (0) 5 4
dξn σ(ξ ′ )dx′
|ξ ′ |=1 −∞ (ξn − i) (ξn + i)
2πi h −4i + 16ξn + 4iξn2 i(4)
= h′ (0) |ξn =i Ω4 dx′
4! (ξn + i)4
25 ′
= πh (0)Ω4 dx′ . (3.34)
16
case b) r = −1, l = −4, k = j = |α| = 0

From (2.4) and an integration by parts, we have

Z Z +∞
case b) = −i trace[πξ+n σ−1 (D−1 ) × ∂ξn σ−4 (D−3 )](x0 )dξn σ(ξ ′ )dx′
|ξ ′ |=1 −∞
Z Z +∞
= i trace[∂ξn πξ+n σ−1 (D−1 ) × σ−4 (D−3 )](x0 )dξn σ(ξ ′ )dx′ . (3.35)
|ξ ′ |=1 −∞

In the normal coordinate, g ij (x0 ) = δij and ∂xj (g αβ )(x0 ) = 0, if j < n; = h′ (0)δβα , if j = n. So by Lemma
A.2 in [12], we have Γn (x0 ) = 25 h′ (0) and Γk (x0 ) = 0 for k < n. By the definition of δ k and Lemma 2.3 in

7
[12], we have δ n (x0 ) = 0 and δ k = 41 h′ (0)c(eek )c(f
en ) for k < n. Then by (3.15), we obtain
√
c(ξ)σ2 (D3 )c(ξ) c(ξ) X −1c(ξ)
σ−4 (D−3 ) = − ∂ξj [c(ξ)|ξ| 2
]D xj [ ]
|ξ|8 |ξ|4 j |ξ|4
1 1 X 5 ′ 5 ′ 
′ 2
= c(ξ) h (0)c(ξ) ξk c( ee k )c(f
e n ) − h (0)ξ n c(ξ) − h (0)|ξ| c(dx n ) c(ξ)
|ξ|8 2 2 4
k<n
√
−1c(ξ)  4 ′ ′ ′ ′

+ |ξ| c(dx )∂
n xn c(ξ ) − 2h (0)c(dx n )c(ξ) + 2ξ n c(ξ)∂ xn c(ξ ) + 4ξ n h (0)
|ξ|8
1  1 5 5 
= 8
c(ξ) h′ (0)c(ξ)c(ξ ′ )c(dxn ) − h′ (0)ξn c(ξ) − h′ (0)|ξ|2 c(dxn ) c(ξ)
|ξ| 2 2 4
1  4 ′ ′
+ 8 i|ξ| c(ξ)c(dxn )∂xn c(ξ ) − 2ih (0)c(ξ)c(dxn )c(ξ)
|ξ|

+2iξn c(ξ)c(ξ)∂xn c(ξ ′ ) + 4iξn h′ (0)c(ξ)
1  11
= ( ξn (1 + ξn2 ) + 8iξn )h′ (0)c(ξ ′ )
(1 + ξn2 )4 2
7 15

+ − 2i + 6iξn2 − (1 + ξn2 ) + ξn2 (1 + ξn2 ) h′ (0)c(dxn )
4 4 
−3iξn (1 + ξn2 )∂xn c(ξ ′ ) + i(1 + ξn2 )c(ξ ′ )c(dxn )∂xn c(ξ ′ ) . (3.36)

By (3.27), (3.31) and (3.36), we obtain

tr[∂ξn πξ+n σ−1 (D−1 ) × σ−4 (D−3 )]

h c(ξ ′ ) + ic(dx ) 1  11

n
= tr − 2
× 2 4
ξn (1 + ξn2 ) + 8iξn h′ (0)c(ξ ′ )
2(ξn − i) (1 + ξn ) 2
7 15

+ − 2i + 6iξn2 − (1 + ξn2 ) + ξn2 (1 + ξn2 ) h′ (0)c(dxn )
4 4 i
−3iξn (1 + ξn2 )∂xn c(ξ ′ ) + i(1 + ξn2 )c(ξ ′ )c(dxn )∂xn c(ξ ′ )
7 + 6i − (20 − 15i)ξn + (7 − 6i)ξn2 + 15iξn3
= h′ (0) . (3.37)
(ξn − i)5 (ξn + i)4

Then
Z Z +∞
7 + 6i − (20 − 15i)ξn + (7 − 6i)ξn2 + 15iξn3
case b) = ih′ (0) dξn σ(ξ ′ )dx′
|ξ ′ |=1 −∞ (ξn − i)5 (ξn + i)4
2πi h 7 + 6i − (20 − 15i)ξn + (7 − 6i)ξn2 + 15iξn3 i(4)
= ih′ (0) |ξn =i Ω4 dx′
4! (ξn + i)4
25 35i
= (− − )πh′ (0)Ω4 dx′ . (3.38)
8 16
case c) r = −2, l = −3, k = j = |α| = 0

From (2.4) we have

Z Z +∞
case c) = −i trace[πξ+n σ−2 (D−1 ) × ∂ξn σ−3 (D−3 )](x0 )dξn σ(ξ ′ )dx′ . (3.39)
|ξ ′ |=1 −∞

8
By (2.2.34) , (2.2.37) and (2.2.40) in [13], we have
h c(ξ)p (x )c(ξ) + c(ξ)c(dx )∂ [c(ξ ′ )](x ) i
0 0 n xn 0
πξ+n σ−2 (D−1 )(x0 )||ξ′ |=1 = πξ+n
(1 + ξn2 )2
h c(ξ)c(dx )c(ξ) i
n
−h′ (0)πξ+n
(1 + ξn )3
:= B1 − B2 . (3.40)

where
−1
B1 = [(2 + iξn )c(ξ ′ )p0 c(ξ ′ ) + iξn c(dxn )p0 c(dxn )
4(ξn − i)2
+(2 + iξn )c(ξ ′ )c(dxn )∂xn c(ξ ′ ) + ic(dxn )p0 c(ξ ′ ) + ic(ξ ′ )p0 c(dxn ) − i∂xn c(ξ ′ )]
1 h5 5i ′ i
′ ′ ′ ′ ′
= h (0)c(dx n ) − h (0)c(ξ ) − (2 + iξ n )c(ξ )c(dxn )∂ξ c(ξ ) + i∂ξ c(ξ ) ; (3.41)
4(ξn − i)2 2 2 n n

h′ (0) h c(dxn ) c(dxn ) − ic(ξ ′ ) 3ξn − 7i i

B2 = + 2
+ 3
[ic(ξ ′ ) − c(dxn )] . (3.42)
2 4i(ξn − i) 8(ξn − i) 8(ξn − i)

By (3.14), we obtain
−4iξn c(ξ ′ ) + (i − 3iξn2 )c(dxn )
∂ξn σ−3 (D−3 ) = . (3.43)
(1 + ξn2 )3
By (3.27), (3.42) and (3.43), we have

tr[B2 × ∂ξn σ−3 (D−3 )(x0 )]||ξ′ |=1

n h′ (0) h c(dx ) c(dxn ) − ic(ξ ′ ) 3ξn − 7i i
n ′
= tr + + [ic(ξ ) − c(dxn )]
2 4i(ξn − i) 8(ξn − i)2 8(ξn − i)3
′ 2
−4iξn c(ξ ) + (i − 3iξn )c(dxn ) o
×
(1 + ξn2 )3
4i − 11ξn − 6iξn2 + 3ξn3
= h′ (0) . (3.44)
(ξn − i)5 (ξn + i)3

Similarly, we have

tr[B1 × ∂ξn σ−3 (D−3 )(x0 )]||ξ′ |=1

n 1 h5 5i i
= tr 2
h′ (0)c(dxn ) − h′ (0)c(ξ ′ ) − (2 + iξn )c(ξ ′ )c(dxn )∂ξn c(ξ ′ ) + i∂ξn c(ξ ′ )
4(ξn − i) 2 2
−4iξn c(ξ ) + (i − 3iξn )c(dxn ) o
′ 2
×
(1 + ξn2 )3
3 + 12iξn + 3ξn2
= h′ (0) . (3.45)
(ξn − i)4 (ξn + i)3

Combining (3.44) and (3.45), we obtain

Z Z +∞
′ −7i + 26ξn + 15iξn2
case c) = −ih (0) dξn σ(ξ ′ )dx′
|ξ ′ |=1 −∞ (ξn − i)5 (ξn + i)3
2πi h −7i + 26ξn + 15iξn2 i(5)
= −ih′ (0) |ξn =i Ω4 dx′
4! (ξn + i)3
55 ′
= πh (0)Ω4 dx′ . (3.46)
16

9
Since Φ is the sum of the cases a), b) and c), so
15 35i
Φ=( − )πh′ (0)Ω4 dx′ . (3.47)
16 16
Now recall the Einstein-Hilbert action for manifolds with boundary [12][13][14],
Z Z
1
IGr = sdvolM + 2 Kdvol∂M := IGr,i + IGr,b , (3.48)
16π M ∂M

where X i,j
K= Ki,j g∂M ; Ki,j = −Γni,j , (3.49)
1≤i,j≤n−1

and Ki,j is the second fundamental form, or extrinsic curvature. Taking the metric in Section 2, then by
Lemma A.2 [12], for n = 6, then
5
K(x0 ) = − h′ (0); IGr,b = −5h′ (0)Vol∂M . (3.50)
2
Then we obtain
Theorem 3.4. Let M be a 6-dimensional compact spin manifold with the boundary ∂M and the metric g M
c, then
as above and D be the Dirac operator on M
Z Z
(1,3)
Vol6 ] + D−1 ◦ π + D−3 ] = − 5Ω5
= Wres[π
7i 3
sdvolM + ( − )πΩ4 KdVol∂M . (3.51)
3 M 8 8 ∂M

] + D−1 ◦π + D−1 ] and Wres[π

Remark 3.5. In [12] [13], Wang computed Wres[π ] + D−2 ◦π + D−2 ]. In that cases,
the boundary terms vanished, where the two operators are symmetric. Theorem 3.4 states the boundary terms
] + D−1 ◦ π + D−3 ]. The reason is that D−1 and D−3 are not symmetric.
is non-zero when we compute Wres[π
Let
] + D−1 ◦ π + D−3 ] = Wres
Wres[π ] i [π + D−1 ◦ π + D−3 ] + Wres
] b [π + D−1 ◦ π + D−3 ], (3.52)
where Z Z
] i [π + D−1 ◦ π + D−3 ] =
Wres traceS(T M) [σ−6 (D−1−3 )]σ(ξ)dx (3.53)
M |ξ|=1

and
] b [π + D−1 ◦ π + D−3 ]
Wres
Z Z Z +∞ X ∞ X
(−i)|α|+j+k+1
= × traceS(T M) [∂xj n ∂ξα′ ∂ξkn σr+ (D−1 )(x′ , 0, ξ ′ , ξn )
∂M |ξ ′ |=1 −∞ j,k=0 α!(j + k + 1)!

×∂xα′ ∂ξj+1
n
∂xkn σl (D−3 )(x′ , 0, ξ ′ , ξn )]dξn σ(ξ ′ )dx′ (3.54)

] + D−1 ◦ π + D−3 ].
denote the interior term and boundary term of Wres[π
Combining (3.48), (3.51) and (3.52), we obtain
Corollary 3.6. Let M be a 6-dimensional compact spin manifold with the boundary ∂M and the metric
c, then
g M as above and D be the Dirac operator on M
−3 ] + −1
IGr,i = Wresi [π D ◦ π + D−3 ];
80πΩ5
16 ] b [π + D−1 ◦ π + D−3 ].
IGr,b = Wres (3.55)
(7i − 3)πΩ4

10
 (1,3)
Nextly, for 5-dimensional spin manifolds with boundary, we compute Vol5 . By Proposition 2.2 (2) in
[13], we have Z
] + D−1 ◦ π + D−3 ] =
Wres[π Φ. (3.56)
∂M
By (2.4), when n = 5, we have r − k − |α| + l − j − 1 = −5, r ≤ −1, l ≤ −3, so we get r = −1, l = −3, k =
|α| = j = 0, then
Z Z +∞
+
Φ= traceS(T M) [σ−1 (D−1 )(x′ , 0, ξ ′ , ξn ) × ∂ξn σ−3 (D−3 )(x′ , 0, ξ ′ , ξn )]dξn σ(ξ ′ )dx′ . (3.57)
|ξ ′ |=1 −∞

By (2.2.44) in [12], we have

c(ξ ′ ) + ic(dxn )
πξ+n σ−1 (D−1 )(x0 )||ξ′ |=1 = . (3.58)
2(ξn − i)
From (3.14) we obtain
−4iξn c(ξ ′ ) + (i − 3iξn2 )c(dxn )
∂ξn σ−3 (D−3 ) = . (3.59)
(1 + ξn2 )3
Since n = 5, tr(id) = dim(S(T M )) = 4. By the relation of the Clifford action and trAB = trBA, then we
have the equalities:
tr[c(ξ ′ )c(dxn )] = 0; tr[c(dxn )2 ] = −4; tr[c(ξ ′ )2 ](x0 )||ξ′ |=1 = −4. (3.60)
Hence from (3.58), (3.59), and (3.60), we have
h c(ξ ′ ) + ic(dx )   −4iξ c(ξ ′ ) + (i − 3iξ 2 )c(dx ) i
n n n n
tr ×
2(ξn − i) (1 + ξn2 )3
2i − 6ξn
= . (3.61)
(ξn − i)3 (ξn + i)3
Then
Z Z +∞
2i − 6ξn
Φ = dξn σ(ξ ′ )dx′
|ξ ′ |=1 −∞ (ξn − i)3 (ξn + i)3
2πi h 2i − 6ξn i(2)
= |ξn =i Ω3 dx′
2! (ξn + i)3
3πi
= Ω3 dx′ . (3.62)
4
(1,3) 3πi
By Vol5 = 4 Ω3 Vol∂M and IGr,b = −4h′ (0)Vol∂M , we have
Theorem 3.7. Let M be a 5-dimensional compact spin manifold with the boundary ∂M and the metric g M
c, then
as in Section 2 and D be the Dirac operator on M
(1,3) ] + D−1 ◦ π + D−3 ] = 3πi
Vol5 = Wres[π Ω3 Vol∂M , (3.63)
4
16ih′ (0) ] + −1
IGr,b = Wres[π D ◦ π + D−3 ], (3.64)
3πΩ3
where Vol∂M denotes the canonical volume of ∂M .

4. A Kastler-Kalau-Walze type theorem for perturbations of Dirac operators

In [15], Connes and Moscovici defined a twisted spectral triple and considered the perturbations of Dirac
operator eh Deh . In this section, for perturbations of Dirac operators, we compute the lower dimensional
(2,2)
volume Vol6 for 6-dimensional spin manifolds with boundary and get a Kastler-Kalau-Walze type theorem
in this case.
11
 ] + (f D−2 ) ◦ π + D−2 ] for the Dirac operators
4.1. Wres[π
Let M be a 6-dimensional compact spin manifold with the boundary ∂M and the metric g M as Section 2
] + (f D−2 ) ◦ π + D−2 ] for a smooth function f (x).
c, we will compute Wres[π
and D be the Dirac operator on M
By the Kastler-Kalau-Walze type theorem in [6], we get the lower dimensional volume for 6-dimensional
spin manifolds without boundary. An application of Theorem 1 in [13] shows that
Lemma 4.1. Let M be a 6-dimensional compact spin manifold without boundary, then
Z
5Ω5
W res[f D−4 ] = − f sdvolM . (4.1)
3 M
R
Therefore, we only need to compute ∂M Φ. By Lemma 1 in [13], we have
Lemma 4.2. For 6-dimensional compact spin manifold with the boundary ∂M and the metric g M as above,
then

σ−2 (D−2 ) = |ξ|−2 ; (4.2)

σ−2 (f D−2 ) = f |ξ|−2 ; (4.3)
√ √
σ−3 (D−2 ) = − −1|ξ|−4 ξk (Γk − 2δ k ) − −1|ξ|−6 2ξ j ξα ξβ ∂j g αβ . (4.4)

Since Φ is a global form on ∂M , so for any fixed point x0 ∈ ∂M , we can choose the normal coordinates
U of x0 in ∂M (not in M ) and compute Φ(x0 ) in the coordinates U e = U × [0, 1) ⊂ M and the metric
1 ∂M 2
h(xn ) g + dx n . For details, see Section 2.2.2[12]. Now we can compute Φ (see formula (2.4) for the
definition of Φ), since the sum is taken over −r − l + k + j + |α| = 5, r, l ≤ −2, then we have the following
five cases:
case a) I) r = −2, l = −2 k = j = 0, |α| = 1
From (2.4) we have
Z Z +∞ X
case a) I) = −f trace[∂ξα′ πξ+n σ−2 (D−2 ) × ∂xα′ ∂ξn σ−2 (D−2 )](x0 )dξn σ(ξ ′ )dx′ . (4.5)
|ξ ′ |=1 −∞ |α|=1

By Lemma 3.2, for i < n, then

∂xi (|ξ|2 )(x0 )

∂xi σ−2 (D−2 )(x0 ) = ∂xi (|ξ|−2 )(x0 ) = − = 0. (4.6)
|ξ|4
Then case a) I) vanishes.
case a) II) r = −2, l = −2 k = |α| = 0, j = 1
From (2.4) we have
Z Z +∞
1
case a) II) = − trace[∂xn πξ+n σ−2 (f D−2 ) × ∂ξ2n σ−2 (D−2 )](x0 )dξn σ(ξ ′ )dx′
2 |ξ′ |=1 −∞
Z Z +∞
1
= − f trace[∂xn πξ+n σ−2 (D−2 ) × ∂ξ2n σ−2 (D−2 )](x0 )dξn σ(ξ ′ )dx′
2 |ξ′ |=1 −∞
Z Z +∞
1
− ∂xn (f ) trace[πξ+n σ−2 (D−2 ) × ∂ξ2n σ−2 (D−2 )](x0 )dξn σ(ξ ′ )dx′ . (4.7)
2 |ξ ′ |=1 −∞

By case a) II) in [13], we have

Z Z +∞
1 5
− f trace[∂xn πξ+n σ−2 (D−2 ) × ∂ξ2n σ−2 (D−2 )](x0 )dξn σ(ξ ′ )dx′ = − πf h′ (0)Ω4 dx′ , (4.8)
2 |ξ′ |=1 −∞ 8

where Ω4 is the canonical volume of S 4 .

12
 On the other hand, by (14) in [13], we have
−2 + 6ξn2
∂ξ2n (σ−2 (D−2 ))(x0 ) = ∂ξ2n (|ξ|−2 )(x0 ) = . (4.9)
(1 + ξn2 )3
By (4.2) and the Cauchy integral formula , then
Z 1
1 (ηn +i)(ξn +iu−ηn )
πξ+n σ−2 (D−2 )(x0 )||ξ′ |=1 = lim− dηn
2πi u→0 Γ+ (ηn − i)
−i
= . (4.10)
2(ξn − i)
Combining (4.9) and (4.10), we have
h −i −2 + 6ξn2 i i − 3iξn2 8i − 24iξn2
tr × 2 3
= 4 3
tr[id] = , (4.11)
2(ξn − i) (1 + ξn ) (ξn − i) (ξn + i) (ξn − i)4 (ξn + i)3
where n = 6, trS(T M) [id] = dim(∧∗ (3)) = 8.
From (4.7) and (4.11), we have
Z Z +∞
1
− ∂xn f trace[πξ+n σ−2 (D−2 ) × ∂ξ2n σ−2 (D−2 )](x0 )dξn σ(ξ ′ )dx′
2 |ξ ′ |=1 −∞
Z Z +∞
1 8i − 24iξn2
= − ∂xn (f ) 4 3
(x0 )dξn σ(ξ ′ )dx′
2 |ξ ′ |=1 −∞ (ξn − i) (ξn + i)
1 2πi h 8i − 24iξn2 i(3)
= − ∂xn (f ) |ξn =i Ω3 dx′
2 3! (ξn + i)3
= 3πi∂xn (f )Ω4 dx′ . (4.12)
Hence in this case,
5
case a) II) = − πf h′ (0)Ω4 dx′ + 3πi∂xn (f )Ω4 dx′ . (4.13)
8
case a) III) r = −2, l = −2 j = |α| = 0, k = 1
From (2.4) we have
Z Z +∞
1
case a) III) = − trace[∂ξn πξ+n σ−2 (f D−2 ) × ∂ξn ∂xn σ−2 (D−2 )](x0 )dξn σ(ξ ′ )dx′
2 |ξ′ |=1 −∞
Z Z +∞
1
= − f trace[∂ξn πξ+n σ−2 (D−2 ) × ∂ξn ∂xn σ−2 (D−2 )](x0 )dξn σ(ξ ′ )dx′ . (4.14)
2 |ξ′ |=1 −∞
By case a) III) in [13], we have
Z Z +∞
1 5
− trace[∂ξn πξ+n σ−2 (D−2 ) × ∂ξn ∂xn σ−2 (D−2 )](x0 )dξn σ(ξ ′ )dx′ = πh′ (0)Ω4 dx′ . (4.15)
2 |ξ′ |=1 −∞ 8
Combining (4.14) and (4.15), we obtain
5
case a) III) = πf h′ (0)Ω4 dx′ . (4.16)
8
Since case b), case c) has the same expression with the case of case b), case c) in [13] multiplied by
a function f , so we can use the same way to compute the two terms. An application of (21) and (24) in
[13] shows that the sum of case b) and case c) is zero. Hence we conclude that, the sum of case a), case
b) and case c) is
Φ = 3πi∂xn (f )Ω4 dx′ . (4.17)
In summary, we have proved
13
Theorem 4.3. Let M be a 6-dimensional compact spin manifold with the boundary ∂M and the metric g M
c, then
as above and D be the Dirac operator on M
Z Z
] + (f D−2 ) ◦ π + D−2 ] = − 5Ω5
Wres[π f sdvolM + 3πiΩ4 ∂xn (f )|xn =0 dvol∂M . (4.18)
3 M ∂M

] + (D2 + f )−1 ◦ π + D−2 ] for the Dirac operators

4.2. Wres[π
In this subsection, let M be a 6-dimensional compact spin manifold with the boundary ∂M and the
metric g M as Section 2 and D be the Dirac operator on M ]
c, we will compute Wres[(D 2
+ f )−1 D−2 ] for
smooth function f (x).
Firstly, we compute the symbol σ((D2 + f )−1 ) of (D2 + f )−1 . By the definition of the Dirac operator D
in section 3 and Lemma 1 in [13], we have

σ−2 ((D2 + f )−1 ) = σ−2 (D−2 ); σ−3 ((D2 + f )−1 ) = σ−3 (D−2 ). (4.19)

Let D̃ = D4 + f D2 be the fourth order operators. Write

∞
X
√
Dxα = (− −1)|α| ∂xα ; σ(D̃) = p̃4 + p̃3 + p̃2 + p̃1 + p̃0 ; σ((D̃)−1 ) = q̃−j . (4.20)
j=4

Since the smooth function f (x) ∈ p̃0 , then we have

σ4 (D̃4 ) = σ4 (D4 ); σ3 (D̃4 ) = σ3 (D4 ); σ2 (D̃4 ) = σ2 (D4 ) + f |ξ|2 . (4.21)

By the composition formula of psudodifferential operators, we have

X 1
1 = σ(D̃ ◦ D̃−1 ) = ∂ξα [σ(D̃)]Dxα [σ((D̃)−1 )]
α
α!
= (p̃4 + p̃3 + p̃2 + p̃1 + p̃0 )(q̃−4 + q̃−5 + q̃−6 + · · · )
X
+ (∂ξj p̃4 + ∂ξj p̃3 + ∂ξj p̃2 + ∂ξj p̃1 + ∂ξj p̃0 )
j
 
× Dxj q̃−4 + Dxj q̃−5 + Dxj q̃−6 + · · ·
X 
+ ∂ξi ∂ξj (p̃4 + p̃3 + p̃2 + p̃1 + p̃0 )
i,j
 
× Dxi Dxj q̃−4 + Dxi Dxj q̃−5 + Dxi Dxj q̃−6 + · · ·
 X 
= p̃4 q̃−4 + p̃4 q̃−5 + p̃3 q̃−4 + ∂ξj p̃4 Dxj q̃−4 + · · · . (4.22)
j

Then we obtain

q̃−4 = p̃−1
4 ; (4.23)
X
q̃−5 = −p̃−1
4 [p̃3 q̃−4 + ∂ξj p̃4 Dxj (q̃−4 )]; (4.24)
j
X X
0 = p̃4 q̃−6 + p̃3 q̃−5 + p̃2 q̃−4 + ∂ξj p̃3 Dxj q̃−4 + ∂ξj p̃4 Dxj q̃−5
j j
X
+ ∂ξi ∂ξj p̃4 Dxi Dxj q̃−4 . (4.25)
i,j

Hence from (4.21)-(4.25) and the recursion formula, we have

14
Lemma 4.4. For a 6-dimensional compact spin manifold with the boundary ∂M , then

q−6 = −f |ξ|2 + σ−6 (D−4 ). (4.26)

By Lemma 4.4 and [5], [6], we obtain

Theorem 4.5. let M be a 6-dimensional compact spin manifold without boundary and fi (i = 1, 2, 3) be the
smooth function, then
Z
5
Wres[(D4 + f1 D2 + f2 D + f3 )−1 ] = −Ω5 ( s + 8f1 )dvolM . (4.27)
M 3

Remark 4.6. When fi = 0 (i = 1, 2, 3), we get the classical Kastler-Kalau-Walze type theorem.
On the other hand, it is straightforwardR to see from (4.19) we can directly get the same results with
Theorem 1 in [13]. Hence we conclude that ∂M Φ = 0. In summary, we have proved
Theorem 4.7. let M be a 6-dimensional compact spin manifold with the boundary ∂M and the metric g M
c, then
as Section 2 and D be the Dirac operator on M
Z
] + 2 −1 + −2 5
Wres[π (D + f ) ◦ π D ] = −Ω5 ( s + 8f )dvolM . (4.28)
M 3

Acknowledgements

This work was supported by Fok Ying Tong Education Foundation under Grant No. 121003 and NSFC.
11271062. The author also thank the referee for his (or her) careful reading and helpful comments.

References

References
[1] V. W. Guillemin.: A new proof of Weyl’s formula on the asymptotic distribution of eigenvalues. Adv. Math. 55, no. 2,
131-160, (1985).
[2] M.Wodzicki.: local invariants of spectral asymmetry. Invent. Math. 75(1), 143-178, (1995).
[3] A. Connes.: Quantized calculus and applications. XIth International Congress of Mathematical Physics(Paris,1994), In-
ternat Press, Cambridge, MA, 15-36, (1995).
[4] A. Connes.: The action functinal in Noncommutative geometry. Comm. Math. Phys. 117, 673-683, (1998).
[5] D. Kastler.: The Dirac Operator and Gravitation. Comm. Math. Phys. 166, 633-643, (1995).
[6] W. Kalau and M. Walze.: Gravity, Noncommutative geometry and the Wodzicki residue. J. Geom. Physics. 16, 327-
344,(1995).
[7] T. Ackermann.: A note on the Wodzicki residue. J. Geom.Phys. 20, 404-406, (1996).
[8] R. Ponge.: Noncommutative geometry and lower dimensional volumes in Riemannian geometry. Letters in Mathematical
Physics. 83, no.1, 19-32, (2008).
[9] B. V. Fedosov, F. Golse, E. Leichtnam, E. Schrohe.: The noncommutative residue for manifolds with boundary. J. Funct.
Anal. 142, 1-31, (1996).
[10] Y. Wang.: Diffential forms and the Wodzicki residue for Manifolds with Boundary. J. Geom. Physics. 56, 731-753, (2006).
[11] Y. Wang.: Diffential forms the Noncommutative Residue for Manifolds with Boundary in the non-product Case, Letters
in Mathematical Physics. 77, 41-51, (2006).
[12] Y. Wang.: Gravity and the Noncommutative Residue for Manifolds with Boundary. Letters in Mathematical Physics. 80,
37-56, (2007).
[13] Y. Wang.: Lower-Dimensional Volumes and Kastler-kalau-Walze Type Theorem for Manifolds with Boundary . Commun.
Theor. Phys. Vol 54, 38-42, (2010).
[14] Y. Yu.: The Index Theorem and The Heat Equation Method, Nankai Tracts in Mathematics-Vol.2, World Scientific
Publishing, (2001).
[15] A. Connes, H. Moscovici.: Type III and spectral. Traces in Geometry, Number Theory and Quantum Fields, Aspects of
Mathematics E38, Vieweg Verlag, 57-71, (2008).

15