arXiv:1512.07532v1 [math.CV] 23 Dec 2015
Some Hilbert Spaces related with the the Dirichlet space Nicola Arcozzi; Karl-Mikael Perfekt; Pavel Mozolyako, Stefan Richter, Giulia Sarfatti December 24, 2015 Abstract We do a preliminary study of the reproducing kernel Hilbert space having as kernel kd , where d is a positive integer and k is the reproducing kernel of the analytic Dirichlet space.
Consider the Dirichlet space D on the unit disc {z ∈ C : |z| < 1} of the complex plane. It might be defined as the Reproducing Kernel Hilbert Space (RKHS) having kernel kz (w) = k(w, z) =
∞ X (zw)n 1 1 log = . zw 1 − zw n=0 n + 1
˜ d , with d in N, having kernel k d . Our interest in these spaces We are here interested in the spaces D was motivated by the study of Hankel operators on D. ˜ d and for their dual spaces in Theorem We will provide several equivalent norms for the spaces D 1. Then, we discuss some of their properties. More precisely, we will investigate: ˜ d and its dual space HSd in connection with Hankel operators of Hilbert-Schmidt class on - D the Dirichlet space D; ˜d; - the complete Nevanlinna-Pick property for D - the Carleson measures for these spaces. ˜d live infinitely close to D in the scale of the weighted It an interesting fact that the spaces D Dirichlet spaces defined by the norms Z +π ZZ dA(z) 2 ϕ(eit ) 2 dt + kϕk2Ds = |f ′ (z)| (1 − |z|2 )s , 0 ≤ s < 1, 2π π −π |z| 0 such that for all k ≥ 0 we have cad (k) ≤ Consequently, if t ∈ [0, 1), then c
1 1 log t 1−t
d
≤
logd−1 (k + 2) ≤ Cad (k). k+1
∞ X logd−1 (k + 2) k=0
k+1
tk ≤ c
1 1 log t 1−t
d
.
Proof of Lemma 1. We will prove the Lemma by induction on d ∈ N. It is obvious for d = 1. Thus let d ≥ 2 and suppose the Lemma is true for d − 1. Also we observe that there is a constant c > 0 such that for all k ≥ 0 and 0 ≤ n ≤ k we have c logd−2 (k + 2) ≤ logd−2 (n + 2) + logd−2 (k − n + 2) ≤ 2 logd−2 (k + 2). Then for k ≥ 0 ad (k) =
X
n1 +···+nd =k k X
=
1 n+1 n=0
≈
X
=
1 (n1 + 1) . . . (nd + 1)
n2 +···+nd =k−n
≈
1 (n2 + 1) . . . (nd + 1)
1 logd−2 (k − n + 2) n+1 k−n+1 n=0
by the inductive assumption
k 1 X logd−2 (n + 2) + logd−2 (k − n + 2) 2 n=0 (n + 1)(k − n + 1)
≈ logd−2 (k + 2) =
X
k X
1 (n + 1)(k − n + 1) n=0
by the earlier observation
k logd−2 (k + 2) X 1 1 + k+2 n+1 k−n+1 n=0
logd−1 (k + 2) . k+1
Next, we prove the equivalence [ϕ]HSd ≈ |[ϕ]|HSd in (5). Lemma 2. Let d ∈ N. Then d−1 Z 1 logd−1 (k + 2) 1 1 k dt ≈ log , t t 1−t k+1 0
3
k ≥ d.
Given the Lemma, we expand |[ψ]|2HSd
2 b = |ψ(0)| +
2 b = |ψ(0)| +
2 b ≈ |ψ(0)| +
≈ [ψ]2HSd ,
ZZ
|
∞ X
D k=1
∞ X
k=1 ∞ X k=1
k−1 2 b ψ(k)kz | logd−1
Z b 2 k ψ(k) 2
1
logd−1
0
1 dA(z) 2 1 − |z| π
1 k−1 t dt 1−t
d−1 b 2 log (k + 2) k 2 ψ(k) k+1
as wished. Proof of Lemma 2. The case d = 1 is obvious, thus we will assume that d ≥ 2. We will also assume that k ≥ 2. Then by Lemma 1 we have Z
0
1
t
k
1 1 log t 1−t
d−1
dt
≈ =
Z
1
0
∞ X
tk
∞ X logd−2 (n + 2) n t dt n+1 n=0
logd−2 (n + 2) = S1 + S2 , (n + 1)(n + k + 1) n=0
where S1
=
≈ ≈ =
k−1 X
logd−2 (n + 2) (n + 1)(n + k + 1) n=0 k−1 1 X logd−2 (n + 2) k + 1 n=0 n+1 Z k+2 logd−2 (t) 1 dt k+1 1 t
1 logd−1 (k + 2) d−1 k+1
4
and S2
∞ X
=
logd−2 (n + 2) (n + 1)(n + k + 1)
n=k ∞ X
≤
n=k+1
logd−2 (n + 1) n2
j+1
∞ kX −1 X logd−2 (n + 1) ≤ n2 j j=1 n=k
∞ X
d−2
d−2
log
≤ logd−2 (k + 2)
∞ X
≤
(j + 1)
j=1
= ≤
n=kj
log
d−2
log
d−2
(k + 2) k+1
(j + 1)d−2
j=1 ∞ X
log
d−2
∞ kj −1
1 dx x2
k+1 kj − 1
(j + 1)d−2
k+1 (k − 1)k j−1
(j + 1)d−2
3 2j−1
j=1
logd−1 (k + 2) k+1
1 n2
(j + 1)d−2
j=1
∞ (k + 2) X
k+1
= o
Z
j=1
∞ (k + 2) X
k+1
≤
k
kj+1 X−1
!
since k ≥ 2
.
˜ d and HSd under the duality pairing given by the inner product in Now, the duality between D D is easily seen comparing the norms [·]d and [·]HSd : they are weighted ℓ2 norms and duality is established by means of Cauchy-Schwarz inequality. We still have to prove that [ϕ]d ≈ |[ϕ]|d . This is equivalent to proving that the dual space of HSd with respect to the Dirichlet inner product < ·, · >D is the Hilbert space with the norm |[·]|d . d 1 Let d ∈ N and set, for 0 ≤ t < 1, wd (t) = 1t log 1−t and, for |z| < 1, Wd (z) = wd (|z|2 ).
1 d ) . Then Lemma 3. Let d ∈ N and for t ∈ [0, 1) define wd (t) = ( 1t log 1−t
Z
1
1−ε
wd (t)dt ·
Z
1
1−ε
1 dt ≈ ε2 wd (t)
as ε → 0.
1 d Proof. Write w(t) ˜ = (log 1−t ) , and note that it suffices to establish the lemma for w ˜ in place of
5
w. Let ε > 0. Then w ˜ is increasing in [0, 1) and w(1 ˜ − εk+1 ) = (k + 1)d (log 1ε )d , hence Z
1
w(t)dt ˜ = 1−ε
≤ =
∞ Z X k=1 ∞ X
k=1
∞ X
k=1
1−εk+1
w(t)dt ˜
1−εk
w(1 ˜ − εk+1 )(εk − εk+1 ) 1 (k + 1)d (log )d εk (1 − ε) ε
1 1 ≈ ε(log )d ε (1 − ε)d For 1/w ˜ we just notice that it is decreasing and hence Z
1 1−ε
Thus as ε → 0 we have ε2 ≤
1 ε 1 dt ≤ ε= w(t) ˜ w(1 ˜ − ε) (log 1ε )d
Z
1
w(t)dt ˜
1−ε
Z
1
1 dt = O(ε2 ). w(t) ˜
1−ε
For 0 < h < 1 and s ∈ [−π, π) let Sh (eis ) be the Carleson square at eis , i.e. Sh (eis ) = {reit : 1 − h < r < 1, |t − s| < h}. A positive function W on the unit disc is said to satisfy the B´ekoll´e-Bonami condition (B2) if Z Z 1 W dA · dA ≤ ch4 W is is Sh (e ) Sh (e ) for every Carleson square Sh (eis ). If d ∈ N and if Wd (z) = wd (|z|2 ), then Z
Sh
(eis )
Wd dA ·
Z
Sh
(eis )
1 dA = h2 Wd
Z
1
1−h
wd (t)dt ·
Z
1
1−h
1 dt ≈ h4 wd (t)
by Lemma 3. Thus Wd satisfies the condition (B2). Furthermore, note that if f (z) = is analytic in the open unit disc, then Z
|z|