1. Introduction and Statement of the Main Result
The classical Lipschitz spaces on
,
and
play an important role in function theory, harmonic analysis and partial differential equations. When
, they are defined as the set of functions
such that
. We can say that these classes are between the continuous functions
and derivable functions with continuous derivative,
However, Zygmund [
1] argued that for applications in harmonic analysis, when
, the space corresponds with a space bigger than
, the set of continuous functions
such that
. This space
is also known as the Zygmund space. For
,
is defined as the class of smooth functions such that their first-order derivatives belong to
To find some characterizations of these spaces, which are more suitable for some applications, is a recurrent object of research, such as expressions with finite differences, approximation properties, semigroup language, etc.; see [
2,
3,
4,
5,
6] for instance. Stein and Taibleson [
3,
4,
5,
6] gave the characterizations of bounded Lipschitz functions via the Poisson semigroup,
and the Gauss semigroup,
, and deserve a special mention. The semigroup language allows us to obtain regularity results in these spaces. In particular, we can prove the boundedness of some fractional operators, such as fractional Laplacians, fractional integrals, Riesz transforms and Bessel potentials, in a much more simple way than using the classical definition of the spaces.
Taibleson and Stein raised the question of analyzing some Lipschitz spaces adapted to different “Laplacians”, and to find pointwise and semigroup estimate characterizations. In the case of the Ornstein–Ulhenbeck operator
, in [
7], some Lipschitz classes were defined by means of its Poisson semigroup,
, and in [
8] a pointwise characterization was obtained for
. Sometimes, “Lipschitz classes” are also known as “Hölder classes”. For the case of Hermite operator
, adapted Hölder classes were defined pointwisely in [
9]. These last classes were characterized in [
10], also in the parabolic case, by using semigroups. The classical parabolic case was treated in [
11]. In [
12], they proved the characterization of Lipschitz spaces associated to the Schrödinger operators
, where
V is a non-negative potential satisfying a reverse Hölder inequality.
In this paper, we consider the high-order Schrödinger operators in
with
, that is,
, where
V is a non-negative potential satisfying the reverse Hölder inequality:
with an exponent
for every ball
B. In [
13], the author proved the fundamental solution estimates and
-estimates for some Schrödinger-type operators. In [
14], the authors proved that the Hardy spaces related with
is equivalent to the Hardy spaces related with
L, which is defined in [
15,
16]. Based on the corrigendum of [
14], throughout this paper, we always assume that the heat kernel,
, satisfies the following inequality:
for any
and
see [
14,
17]. Inspiring the results in [
12], we can consider the Lipschitz spaces associated with
as in the Schrödinger setting.
In [
18], the authors introduced the space
,
, as the set of functions which satisfy
where
is the critical radius associated to the potential
V, see (
7). Moreover, in [
12] (Definition 1.1) and for
, the authors consider the class of functions
, as the set of measurable functions satisfying
This space is endowed with the norm
They also proved that this space
can be described either as the set of functions such that
or as the set of functions such that
with the obvious norms; see Theorem 1.3 in [
12].
One of the main results of this paper is to show that, the above coincidence of smooth class of functions defined either in (
2), in (
3), or in (
4) also holds when considering the heat semigroup associated to the operator
. In order to state the main theorem, we shall present our definitions.
Denote
as the integral kernel of the heat semigroup of
generated by
. That means that, for suitable function
f,
Definition 1. Let We shall denote by the set of functions f such thatand The norm of this space is endowed withwhere is the infimum of the constants appearing in (5). Then, we obtain the following theorem.
Theorem 1. Let Thenwith equivalence of norms. The proof of this theorem will follow some ideas developed in [
12] partly. In particular, we shall need some new Lipschitz spaces associated to the operator
. These classes are introduced and discussed in
Section 2.
As an application of Theorem 1, we shall prove the regularity properties of some operators associated to
. We list here the operators which we are going to deal with. Their definitions are motivated by the Gamma formulas (see [
19]).
The Bessel potential of order
,
The fractional integral of order
,
The fractional “Laplacian” of order
,
The first-order Riesz transforms defined by
Then we have the following results.
Theorem 2. Let and denote the Bessel potential or the fractional integral of order β. Then,
- (i)
- (ii)
Theorem 3 (Hölder estimates)
. If and then, Theorem 4. - (1)
For , then , .
- (2)
For , then , .
Theorem 5. Let g be a measurable bounded function on and consider Then, for every , the multiplier operator of the Laplace transform type is bounded from into itself.
We organize the paper as follows. In
Section 2, we prove the characterization results but related to
In
Section 3, we give the proof of our characterization theorem listed above in a high-order Schrödinger setting. In the proof, we use the comparison of Lipschitz spaces associated to
and
. In
Section 4, we prove the boundedness of the Bessel potentials, fractional integrals, fractional “Laplacian” and the first-order Riesz transforms related with high-order Schrödinger operators
on the adapted Lipschitz spaces.
Along this paper, C denotes a constant that may not be the same in each appearance. We will write the constant with subindexes if we need to emphasize the dependence on some parameters.
2. Lipschitz Spaces Related to the Biharmonic Operator
We shall define a new Lipschitz space associated to the biharmonic operator , which will be called “the adapted Lipschitz space” later, in this section. In fact, we have double motivations. Firstly, we can use the definition to prove some regularity estimates for operators defined through . Secondly, as we have said, we shall need them as a tool to prove results for high-order Schrödinger operators of the type
Consider the following Cauchy problem for the biharmonic heat equation
The solution of this partial differential equation is given by
where
with
and
being the inverse Fourier transform.
denotes that the
v-th Bessel function and
is a normalization constant such that
For more details, see [
20,
21]. The following several results can be obtained by classical analysis (for details, see [
22]):
- (1)
If
,
then
and
- (2)
For the heat kernel of the semigroup , the following estimates are known.
Lemma 1 (See [
23] and ([
20], Lemma 2.4))
. For and , the following estimates hold:- (i)
- (ii)
- (iii)
- (iv)
There exist such that for
2.1. Lipschitz Spaces Associated to
We defined the spaces
in [
24].
Definition 2 (See ([
24], Definition 1.1))
. Let The spaces are defined asThe norm is endowed aswhere denotes the infimum of the constants appearing above. Remark 1. Because of the estimate(iii)in Lemma 1, in Definition 2, we can assume .
The proposition in the following means that we can use any bigger integer than in Definition 2.
Proposition 1 (See ([
24], Proposition 2.3))
. Let . A function if, and only if, for , we have and In [
24], the following theorem was proved.
Theorem 6 (See ([
24], Theorem 1.2)).
Let . Then the following three statements are equivalent:- (1)
.
- (2)
, where is the classical Poisson kernel.
- (3)
.
Additionally,where denotes the infimum of the constants appearing in . Remark 2. In [3], E. M. Stein proved that and are equivalent. In the following, we prove a parallel result to that of [
3] (Proposition 9 in p. 147), which gives the relationship between the Lipschitz function and its derivatives. First, we give a lemma and a theorem which will be used later.
Lemma 2 (See ([
24], Lemma 3.2))
. Let and . If , then for every such that , there exists a such that Theorem 7 (See ([
24], Theorem 1.6))
. Suppose that and Then,In this case, the following equivalence holds 2.2. Adapted Lipschitz Spaces Associated to
In this subsection, we should define a new Lipschitz class to prove our main results. For our results about the spaces adapted to the high order Schrödinger operator , it will be an auxiliary class. The crucial difference is that the functions do not need to be bounded.
Definition 3. Let We define the spaces as We endow this class with the normwhere . When , by the -boundedness of the heat semigroup , we know that, for every , is well defined. For functions , the following lemma says that the definition above has sense.
Lemma 3. Let f be a function with . Then, for every , is well defined. In addition, Moreover, belongs to .
Proof. If
, then the above integral is less than
If , the last integral is convergent and bounded by . The same arguments can be used for the derivatives , .
Moreover, if
, then
, for every
Indeed, observe that
If , the last integral is less than . In the case , the integral is less than
The proof of the second part of this lemma is similar to the proof of Lemma 2.15 in [
12], but with some notation differences. We sketch it here. □
Now we will prove a pointwise description of the adapted Lipschitz spaces associated to . First, we can obtain the proposition as follows.
Proposition 2. Let . A function if, and only if, for all , we have and
Proof. Let
. Using Lemma 1,
Then,
In addition, is trivial.
For the converse, since we can integrate on t as many times as we need to obtain □
We also need the following lemma in later proofs.
Lemma 4. Let and . If , then for every such that , there exists a such that Proof. If , by proceeding as before, we obtain that , , and we obtain the result by integrating the previous estimate times, since as as far as . □
Theorem 8. Let The following two statements are equivalent.
- (1)
.
- (2)
.
Proof. For any
and
, and, for every
,
, we have
In a parallel way, we can handle the two first summands. For the last summand, by using the chain rule, Lemmas 1 and 4, we have
Thus, by choosing , we obtain what we wanted.
For the converse, we assume that
and
. Since
and
we have
□
By the theorem above and Proposition 3.28 in [
12], we know that, when
, if
, then
4. Proof of Regularity Properties
In this section, we give the proof of Theorems 2–5. First, we prove a lemma in the following.
Lemma 12. Let and be either the operator or the operator . If f is a function such that for some , then is well defined and satisfies Moreover, if , then is well defined and Proof. If
for some
, then by Lemma 9, we obtain
The same estimate works for The proof in the second case runs parallel since Lemma 9 has an obvious version for bounded functions. □
Now, we are in a position to prove Theorem 2.
Proof of Theorem 2.
We prove only (i), and estimate (ii) can be proved similarly. Let
. Lemma 9 with
together with Fubini’s theorem allows us to obtain
By Lemma 10 with
and
such that
, we have
We can bound the function in the last integral by a uniform (in a neighborhood of t) integrable function (of s). This allows us to interchange the derivative with respect to t and the integral with respect to s in the above expression.
Let
. By using the above arguments again and using the hypothesis, we have
When , we apply Lemma 6 and we obtain for that . Then we can proceed as before.
Together with Lemma 12, we end the proof of the theorem. □
In order to prove Theorem 3, we need a lemma as follows.
Lemma 13. Let and f be a function in the space . Then is well defined and Proof. As
, by Lemma 9 we have
Now we shall estimate
. Let
. We have
If
, then by Lemma 10 (with
,
and
) we have
If
, then
and by Lemma 10 we obtain, for
,
If
is not divisible by 4, then we have, for
,
If
is divisible by 4, then
and, for
,
In order to solve the last integral, by changing variables
. Then we can proceed as in the proof of Proposition 5. We obtain, in this case,
Combining the estimates of I and , we can finish the proof of Lemma 13. □
Proof of Theorem 3. If
and
, then
Since , we have
By the same argument in the proof of Lemma 13, we have
Then, we will estimate
A and
B separately. By change of variables, we obtain
The last inequalities can be obtained by observing that inside the integrals together with the discussion about the sign of . Together with Lemma 13, we complete the proof of this theorem. □
Now, we will prove Theorem 4.
Proof of Theorem 4. Let and . We have by Theorem 2. By Theorem 11, this means that and , where denotes the adapted Lipschitz space associated to the biharmonic operator (see Definition 3). Therefore, by Proposition 6, we obtain and . Thus, Theorem 11 gives the first statement of the theorem.
Suppose and . By Theorem 11, this means that and . By Proposition 6, we obtain and . Again, by Theorem 11, this means that and by Theorem 2, we obtain . □
At last, we complete the proof of Theorem 5.
Proof of Theorem 5. Lemmas 9 and 10 guarantee the integrability of
as a function of
Hence, we can write
Now we estimate
I. Let
. If
is not divisible by 4, by Lemma 10, we obtain
If
is divisible by 4, by Lemma 10 we have
Up to now, we have shown that
Now we want to see that
. Fubini’s theorem together with Lemmas 9 and 10 allow us to interchange the integral with derivatives and kernels. Then,
We complete the proof of the theorem. □