On the boundedness of the Weyl transform for homogeneous distributions on R 2

Starting from earlier work by M.W. Wong, we give necessary and sufﬁcient criteria, respectively, for the boundedness of the Weyl transform W σ as an operator on L 2 ( R ) if σ is a homogeneous distribution on R 2 .


Introduction and notation
In our first paper on Weyl transforms, see [8], we endeavored to generalize the theorems and examples in [11] concerning the boundedness and compactness, respectively, of Weyl transforms of radially symmetric functions to the case of radially symmetric distributions. (Incidentally, let me indicate three printing errors in [8]: On p. 769, the space D L 1 (R 2 ) should be replaced by D L 1 (R 1 ) and on p. 775, the expression e i exp(2x) was twice contorted to e i exp(2x); finally, in line 13 on p. 785, the factor s z 0 got lost in the expression (2π) −1 s z 0 e ih(s 0 ) 2π i/h (s 0 ).) In [8], we did not touch on Theorem 27.3 on tensor products in polar coordinates in [11, p. 128] for a variety of reasons: First, this theorem concerns a much more general situation (of which radially symmetric functions are just a tiny special case), second, the setting is much more intricate and arduous than for radially symmetric functions, third, we did not immediately succeed in finding a generalization to distributions with meaningful and manageable examples, and fourth, there is no hope (as far as we can see) for a single necessary and sufficient criterion for boundedness, in contrast to the radial case. D L p ,μ := D L p · (1 + |x| 2 ) −μ/2 , D L p ,μ := D L p · (1 + |x| 2 ) −μ/2 , 1 ≤ p ≤ ∞, μ ∈ R, see [6].
N, N 0 and Z denote the sets of positive, of non-negative, and of all integers, respectively. For sequences with the index sets N 0 and Z, respectively, we use the Banach spaces l p , 1 ≤ p ≤ ∞, and the Fréchet space s of fast decreasing sequences with its dual s of slowly increasing sequences.
S 1 denotes the unit circle in R 2 , which we identify with C, and we write ω for the generic variable on S 1 ⊂ C and dω for the measure of length on S 1 . With these notations, (2π) −1/2 ω k , k ∈ Z, is an orthonormal basis in L 2 (S 1 ) and f (ω) = k∈Z c k ω k is the usual Fourier series expansion for f ∈ L 2 (S 1 ) if c k = 1 2π The Heaviside function is denoted by Y , see [9, p. 36], and we write δ τ (t) ∈ D (R 1 t ), τ ∈ R, for the delta distribution with support in τ, which is the derivative of Y (t − τ ), i.e., φ, δ τ = φ(τ ) for φ ∈ D(R 1 ). In contrast, δ ∈ D (R n ) without any subscript stands for the delta distribution at the origin. For z ∈ C with Re z > −1, we write t z + for the locally integrable function Y (t)t z and we obtain, by analytic continuation, the analytic distribution-valued function We use the Fourier transform F in the form this being extended to § by continuity. For the partial Fourier transforms of a distribution T ∈ S (R m+n xy ) with respect to x ∈ R m or y ∈ R n , respectively, we use the notation F x T and F y T , respectively.
The Hermite polynomials H j and Hermite functions h j are defined as usually: see [5,Section 8.95], [11,Sections 18,19]. Similarly, the Laguerre polynomials L α j (x) of order α are defined by

Review of definitions and classical results
The Weyl transformation W can be described as a partial Fourier transform composed with a linear transformation in R 2n xy , to wit The Weyl transform W can be expressed by the "Wigner transform," which is a continuous sesquilinear mapping The mappings w and W are connected by Note that we use for w a normalization constant different from that in [11,Equ. (3.12), p. 15], and that we also changed, for clarity, the notation for the Wigner transform from W (φ, ψ) to w(φ, ψ).
In this study, we shall focus on symbols σ which are homogeneous distributions in two dimensions, and we shall derive necessary and sufficient criteria, respectively, for the boundedness of the corresponding pseudo-differential W σ on L 2 (R 1 ), i.e., criteria on whether W σ can be extended to a bounded mapping L 2 (R) → L 2 (R), or, in other words, whether the norm W σ φ 2 can be bounded by a constant multiple of φ 2 for φ ∈ S(R 1 ).
Let us now recapitulate some basic facts about homogeneous distributions σ on R 2 , see [6,7] for more details. If σ ∈ D (R 2 \{0}) is homogeneous of degree λ ∈ C, then the structure theorem (see [6,Sect. 2.5]) implies that σ can be extended to R 2 and can be represented in the form σ = F · r λ where the so-called "characteristic" F belongs to D (S 1 ) and If we identify S 1 ⊂ R 2 C, then 2π ω −k , F ∈ C for k ∈ Z and k∈Z c k ω k converges to F(ω) in D (S 1 ). The Hilbert spaces are the usual Sobolev spaces on S 1 .
The distribution-valued function z → t z + is meromorphic with simple poles at z = −k, k ∈ N. We have Res z=−k t z For λ ∈ C\(−2 − N 0 ), this implies that F · r λ is homogeneous not only in R 2 \{0}, but also on R 2 , whereas for λ = −2 − j, j ∈ N 0 , F · r λ is homogeneous on R 2 if and only if the Fourier coefficients Furthermore, if S λ (R 2 ), λ ∈ C, denotes the closed subspace of S (R 2 ) of distributions which are homogeneous of degree λ in R 2 , then F : is an isomorphism of locally convex spaces. More explicitly, if λ ∈ C\Z, then We finally observe that the factor in front of c k in formula (2.2) does not depend on the sign of k on account of the complement formula for the gamma function.
As we shall see in Proposition 3 below, a homogeneous symbol σ ∈ S (R 2 ) can have a bounded Weyl transform W σ only under the condition that the degree λ of homogeneity fulfills −2 ≤ Re λ ≤ 0. For homogeneous and radially symmetric symbols σ = r λ ∈ S (R 2 ), we have shown in [8,Sect. 5 Let us describe next what can be inferred from classical results. According to [11, Thms. 11.1, 11.3, pp. 55, 57], compare also [8, Prop. 1, p. 773], W σ is bounded and even compact as an operator on L 2 (R n ) if σ or Fσ belong to ∪ p∈ [1,2] , then we employ formula (2.2) and conclude once again that W σ ∈ Com(L 2 (R 1 )) from the asymptotic behaviour . We shall deduce more general results in Sect. 6 below.

The matrix representation of W
The normalized Hermite functions are an orthonormal basis in L 2 (R), see, e.g., [11,Thm. 19.3,p. 94]. This yields an isomorphism of Hilbert spaces: As was observed already by Schwartz, cf. [9, Ch. VII, Sect. 7, Ex. 7, pp. 260, 262], the sequence e j , j ∈ N 0 , also serves to identify the spaces S(R) and S (R) with the sequence spaces s(N 0 ) and s (N 0 ), respectively. Hence we obtain a commutative diagram: Here the vertical isomorphisms in the diagram are all given by T → ( e j , T ) j∈N 0 .
By the same token, using S ( on account of formula (2.1). Therefore, the operator W σ is bounded on L 2 (R) if and only if holds.
Let us specialize now to distributions σ ∈ S (R 2 ) that are tensor products in polar coordinates. For the radial dependency of σ, we take R ∈ S ([0, ∞)), which is the As in [8, p. 776], we observe that is an isomorphism of locally convex spaces.
If, furthermore, F ∈ D (S 1 ), then we define F · R ∈ S (R 2 ) by the following: Note that this is well-defined since the function is not uniquely determined up to a factor by the distribution σ = F · R since, e.g., F · δ = 0. In the following Proposition 1, we state formulas for the matrix coefficients of W σ with respect to the normalized Hermite functions e j , e k if σ = F · R ∈ S (R 2 ) is a tensor product in polar coordinates. This proposition is but a slight extension of Theorem 27.3 in [11, p. 128] from temperate functions to temperate distributions. Proposition 1 Let e j be the normalized Hermite functions as in formula (3.1) and L α k be the Laguerre polynomials as in Sect.
In the second part, we will closely follow the proof of Theorem 27.3 in [11, p. 129]. Assuming condition (3.5) and (1+|k|) μ c k k∈Z ∈ l 1 (Z), we set d k = C(1+|k|) μ |c k |, k ∈ Z. Then the sequence (d k ) k∈Z belongs to l 1 (Z) and i.e., B σ and thus also W σ are bounded. This completes the proof.
If we set F = 1 in Proposition 1, we obtain c k = δ k0 and hence the matrix B σ is diagonal with the eigenvalues  .2), p. 779], a representation of the eigenvalues λ j = a j j , j ∈ N 0 , by integrals with rational kernels was derived for σ ∈ S (R 2 ) radially symmetric and such that σ (x, y) = (F T )(x 2 + y 2 ) with T ∈ D L 1 ,−1 (R). This formula, i.e., was useful in some examples in calculating the asymptotic behavior of λ j for j → ∞, see [8,Sect. 5].
Since we will make no use here of the corresponding "rational" representation of a jk , let us just state the formula without proof: If j, k ∈ N 0 , j ≥ k, T ∈ D L 1 ,−1 (R) and R(r ) = (F T )(r 2 ) · r j−k holds in S ([0, ∞)), then (3.9) As in the diagonal case, formula (3.9) can be derived from formula (3.4) by Fourier transformation, compare the procedure in [11,Sects. 25,26].

The case of homogeneous symbols of degree 0
We specialize now our investigation to homogeneous symbols σ = F · r λ , λ ∈ C, F(ω) = k∈Z c k ω k ∈ D (S 1 ). We assume that σ is homogeneous on R 2 and hence that c k = 0 for |k| ≤ j if λ = −2 − j, j ∈ N 0 , cf. Sect. 2.
In order to calculate a jk (λ) for R(r ) = r λ + ∈ S ([0, ∞)) according to formula (3.4), we first assume that λ ∈ C\Z with Re λ > −2. Then [1, Equ. 11.44, p. 110] and [5, Equ. 9.131.1] yield the following for j, k ∈ N 0 with j ≥ k : By analytic continuation, formulas (4.1), (4.2) hold for λ ∈ C\Z and, due to the symmetry of (4.2) in j, k, for all j, k ∈ N 0 . But note that the hypergeometric function Because of that and because of the gamma factor in (4.1), (4.2), we have to exercise some care if λ ∈ Z.
Let us now specialize further and consider the case λ = 0, i.e., the case of homogeneous symbols σ = F(ω) of degree 0.
(b) Let us next estimate a jk (0). Using Euler's psi function ψ(x) = (log (x)) for x > 0 we obtain Equation (4.4) implies, incidentally, that α j is monotonically increasing since the same holds for the function ψ due to the logarithmic convexity of the gamma function, see, e.g., [5,Equ. 8.363.3].
Since C 1 + log x ≤ ψ(x) ≤ C 2 + log x holds for all x ≥ 1 2 and appropriate constants C 1 , C 2 ∈ R, we conclude that This implies Therefore, condition (3.5) in Proposition 1 holds for μ = 1 4 and this implies that W σ is bounded on L 2 (R) if the condition ( 4 √ |k| c k ) k∈Z ∈ l 1 (Z) is assumed. The proof is complete.

Necessary conditions
We shall prove here that σ ∈ S (R 2 ) homogeneous of degree λ can have a bounded Weyl transform only if the real part of λ lies in the interval [−2, 0] and if the characteristic F of σ belongs to the Sobolev space H (1/4+μ/2) (S 1 ). In deriving this result, we need yet another representation of a jk (λ) by the hypergeometric function.
If we use the second equation in [1,Equ. 11.44,p. 110] or, alternatively, apply (employing a limit procedure) [5,Equ. 9.131.2] to formula (4.1), then we obtain the following for λ ∈ C\Z and j, k ∈ N 0 with j ≥ k : Let us now digress a little bit and derive the "complement formula" for the matrix coefficients a jk (λ) from formula (5.1). Upon applying [5,Equ. 9.131.1] to formula (5.1) we obtain, still for j ≥ k, wherefrom we conclude that Note that the factor in (5.2) in front of a jk (−2 − λ) is symmetric in j and k. We observe that the special case of j = k in (5.2) was already given in [8, p. 782] and that formula (5.2) implies a jk (−1) = 0 if the minimum of j and k is odd, compare the remark following Proposition 2. Also note that formula (5.2) is a consequence of formula (2.2) and of the fact that W σ and W F σ are closely connected.
Proof (a) Let us first settle the case of homogeneous σ with support contained in {0}. Then λ = −2 − j, j ∈ N 0 , and σ is a linear combination of ∂ k x ∂ l y δ for k, l ∈ N 0 with k + l = j. This implies that W σ is a linear combination of the operators A kl , and the boundedness of A kl implies k = 0, and and the boundedness of A 0l implies l = 0 and hence σ is a multiple of δ.
(b) Let now σ = F · r λ with F(ω) = k∈Z c k ω k ∈ D (S 1 ) and assume that c m = 0 for some m ∈ Z. In particular, in the case of λ = −2 − j, j ∈ N 0 , the homogeneity of σ on R 2 implies that |m| > j.
Since W σ ∈ L(L 2 (R)), its matrix coefficients b jk = e j , W σ e k are bounded in modulus by W σ L(L 2 (R)) . Hence also the complex sequence a k,k+m (λ) = b k,k+m /c m , k ∈ N 0 , k ≥ −m, must be bounded.
Let us next determine the asymptotic behavior of a k+m,k (λ) for k → ∞ and m ∈ N 0 fixed, and for λ ∈ C\(−2 − N 0 ) or such that λ = −2 − j, j ∈ N 0 , and m > j.
Upon inserting the standard integral representation for the hypergeometric function (see [5,Equ. 9.111]) into formula (5.1) and using analytic continuation, we obtain Note that t is an entire distribution-valued function of λ which can be multiplied with (1−t) (m+λ)/2 + since the singular supports of the two distributions are disjoint. Also note that (1 − t) (m+λ)/2 + is analytic at the value of λ which we consider. We then obtain the asymptotic behavior of a k+m,k (λ) for k → ∞ by approximating (1 − 2t) k in formula (5.3) by (1 − t) 2k near zero and by (−t 2 ) k near one, respectively. This implies Hence, eventually, We remark that the special case of m = 0 was shown already in [8, p. 783].
. Altogether we conclude that F belongs to the Sobolev space H (1/4+μ/2) (S 1 ) and this completes the proof.

Remark
Let us yet comment on the situation for symbols σ that are homogeneous on R 2 \{0}, but only associated homogeneous on R 2 , i.e., σ = F · r −2− j , j ∈ N 0 , and F(ω) = k∈Z c k ω k with c k = 0 for some k ∈ Z with |k| ≤ j. In this case, W σ cannot be bounded on L 2 (R). We have omitted this case in Proposition 3 since the asymptotic expansion in formula (5.4) becomes slightly more complicated. As to be expected, logarithms appear. Just to get an idea, let us analyze the case of homogeneity λ = −2.
According to the complement formula (5.2), we obtain In particular, this implies On the other hand, a j j (−2) grows logarithmically for j → ∞ and thus W σ is not bounded if σ = r −2 = Pf λ=−2 r λ . In fact, setting m = 0 and k = j ∈ N in formula (5.3) yields .

A sufficient condition ensuring the boundedness of W
When comparing the three examples λ = 0 (in Proposition 2), λ = −1 (in the remark following Proposition 2) and λ = −2 (in the remark following Proposition 3), we see that we have to stipulate an l 1 condition on the Fourier coefficients c k of the characteristic F of the symbol σ = F · r λ in order to guarantee that W σ is bounded.
The following proposition incorporates these special cases into a general statement.
At the end of Sect. 2, we have seen that in the case of −2 < μ = Re λ < −1, assumption (A) implies the boundedness of W σ on L 2 (R) for σ = F · r λ . Instead, assumption (B) implies that W σ is bounded whenever μ ∈ [−2, 0] according to Proposition 4. If −2 < μ < − 3 2 , then (1 + |k|) 1/4+μ/2 k∈Z ∈ l 2 (Z) and hence (A) implies (B). So condition (B) is the less restrictive and more general one. In contrast, when (1 + 2 j ) −1/2 2 j/2 j −2 ≤ ∞ j=1 j −2 < ∞ but (c k ) k∈Z / ∈ l 2 (Z). Hence (B) holds and (A) fails. As was pointed out already in the introduction, in contrast to the case of radially symmetric distributions, we cannot hope for a condition that is both necessary and sufficient for the boundedness of W σ if σ is a tensor product in polar coordinates. This is a consequence of the fact that the Hermite functions in this case do not yield eigenfunctions of W σ . So the approach in Proposition 1 has its limitations from the word go.
in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.