Abstract
We relate the exponential integrability of the conjugate function \({\tilde{f}}\) to the size of the gap in the essential range of f. Our main result complements a related theorem of Zygmund.
Similar content being viewed by others
1 Introduction
We denote by \(L^p\) the usual Lebesgue spaces of functions on the unit circle \(\mathbb {T}\) with norm \(\Vert \cdot \Vert _p\). Given \(f\in L^1\), let u be the Poisson integral of f and denote by \({\tilde{u}}\) the harmonic conjugate function of u, normalized so that \({\tilde{u}}(0)=0\). Then \({\tilde{u}}(z)\) has nontangential limit \({\tilde{f}}(\theta )\) almost everywhere on \(\mathbb {T}\) and we call \({\tilde{f}}\) the conjugate function of f. Alternatively, the conjugate function \({\tilde{f}}\) can be defined as the principal value integral
for almost every \(\theta \). For further details and references, see Sect. 2.1 below.
The linear mapping \(f\mapsto {\tilde{f}}\) is referred to as the conjugation operator. If f is a trigonometric polynomial \(\sum _{n=-N}^N a_n e^{in\theta }\), then \({\tilde{f}}\) is a trigonometric polynomial of the same degree
where \(-i{{\,\mathrm{sgn}\,}}\) is the Fourier multiplier associated with the conjugation operator.
When \(1<p<\infty \), according to a famous theorem of M. Riesz, there is a constant \(C_p\) such that
for all \(f\in L^p\). In addition, although \(f\in L^\infty \) does not imply that \({\tilde{f}}\in L^\infty \) (see, e.g. [4]), the Hilbert transform still has very strong boundedness properties as can be seen in the following theorem, due to Zygmund [19].
Theorem A
(Zygmund) For \(f\in L^\infty \) with \(\Vert f\Vert _\infty \le \pi /2\) and \(\lambda <1\), there is a constant \(C_\lambda \) such that
and if f is continuous on \(\mathbb {T}\), then
for all \(\lambda <\infty \).
For the proof, see Corollary III.2.6 of [4]. It follows that
where \(C(\mathbb {T})\) stands for the space of continuous functions on \(\mathbb {T}\).
Let E be a measurable subset of \(\mathbb {T}\) and define
The Lebesgue measure of E is denoted by |E|. We note that the condition \(\Vert f_1\Vert _\infty <\pi /2\) above is optimal as seen by considering an interval \(E = [a,b] \subset (0,2\pi )\) and showing that
which is not integrable in any neighborhood of b. More generally, it follows from (1.6) and Theorem III.2.7 of [4] that \(\exp (\tfrac{\pi }{2}{\tilde{\rho }}_E) \notin L^1\) whenever E is a measurable subset of \(\mathbb {T}\) with \(0<|E|<2\pi \).
Notice that the conditions of (1.4) imply that if the function f is real valued and has jumps, then the size of each jump is strictly less than \(\pi \), while the size of each jump of \(\frac{\pi }{2}\rho _E\) is exactly \(\pi \) and \(\exp (\frac{\pi }{2}{\tilde{\rho }}_E)\notin L^1\) with E as above. Motivated by the study of the Fredholm properties of Toeplitz operators, Shargorodsky [11] proved that if g is real valued and \(\inf \mathcal {R}(g) > \pi /2\), where \(\mathcal {R}(g)\) stands for the essential range of g, then \(\exp (\widetilde{g\rho _E})\) is not integrable. These observations lead to the question of whether
Our main result answers this in the affirmative. Indeed, we give an elementary proof of the following result in Sect. 3.
Theorem 1
Suppose that \(f\in L^\infty \) with \(f\ge \pi /2\) almost everywhere and \(0<|E|<2\pi \). Then there is a positive constant C such that
for all \(\lambda \ge 0\). In particular,
Remark 1
In the preceding theorem, the conditions on E and that \(f\ge \pi /2\) almost everywhere are optimal—see Remark 3.
Remark 2
Notice that [11] contains the following local non-integrability result. Suppose that \(0<|E|<2\pi \) and \(\gamma \subset \mathbb {T}\) is an arc intersecting both E and \(\mathbb {T}\setminus E\) in sets of positive measure and at least one of these sets is not an arc (modulo sets of measure zero). If \(g\in L^\infty \) is real valued and \(\inf \mathcal {R}(g) > \pi /2\), then \(\exp (\widetilde{g\rho _E})\) is not integrable on \(\gamma \).
Notice that Theorem 1 generalizes the preceding result for \(\gamma = \mathbb {T}\). However, it remains an open problem whether our conclusion can be strengthened to that of [11] for the other arcs \(\gamma \ne \mathbb {T}\) if we only assume that \(g\ge \pi /2\) almost everywhere.
Previously in [9, 15], sufficient conditions for exponential integrability of \({\tilde{f}}\) were obtained in terms of the modulus of continuity of f in \(L^p\). In addition to these results and other intrinsic interest [4, 7, 18], the integrability of the exponential of conjugate functions plays an important role in the spectral theory of Toeplitz and related operators [10,11,12], scalar Riemann–Hilbert problems [13, 14, 17] and their applications.
2 Preliminaries
As indicated in the introduction, our approach is elementary and based on classical results of complex analysis which are briefly discussed in this section.
2.1 Poisson Integrals
For \(f\in L^1\), denote by P[f] the Poisson integral of f, that is,
where the Poisson kernel \(P_z\) is defined by
Recall that P[f] is harmonic in \(\mathbb {D}\) and if the function f is continuous at \(e^{i\theta }\), then
(see Theorem I.1.3 of [5]). To deal with discontinuities at \(e^{i\theta }\), define a cone \(\varGamma _\alpha \) by
for each \(\alpha <1\), and recall that a function \(\varphi : \mathbb {D}\rightarrow \mathbb {C}\) is said to have nontangential limit \(\varphi ^*(e^{i\theta })\) at \(e^{i\theta }\) if
for every \(\alpha <1\). Now, for any \(f\in L^1\), if \(u = P[f]\), then \(u^* = f\) almost everywhere by Fatou’s theorem.
Define
Then
for \(z\in \mathbb {D}\), where \(u = P[f]\) and \({\tilde{u}}\) is the harmonic conjugate function of u normalized so that \({\tilde{u}}(0)=0\). By Fatou’s theorem and Lemma III.1.1 of [4],
for almost every \(\theta \). For the integral representation of \({\tilde{f}}\) given in (1.1), see Lemma III.1.2 of [4]. In particular, it follows that the principal value in (1.1) exists almost everywhere. If \(f\in L^\infty \), then
2.2 Harmonic and Subharmonic Functions
In one of the key steps of the proof of the main theorem, we consider the Dirichlet problem of finding a unique bounded harmonic function on a simply connected domain \(\varOmega \) with prescribed boundary values.
If g is a continuous real-valued function on \(\partial \varOmega \), the Dirichlet problem of finding the bounded harmonic function \(u : \varOmega \rightarrow \mathbb {R}\) such that \(u = g\) on \(\partial \varOmega \) can be solved using the Poisson integral and the Riemann mapping theorem, which reduces the problem to the well-known case of the unit disk (see, e.g., [5]).
However, in our case, since the boundary functions are discontinuous (see Lemma 1 below), the following more general result is needed.
Theorem B
Let \(\varOmega \) be a simply connected domain and let g be a piecewise continuous function on \(\partial \varOmega \) with a finite number of discontinuities of the first kind at \(\xi _1,\ldots \xi _k\). Then there is at most one bounded harmonic function h on \(\varOmega \) such that \(h=g\) on \(\partial \varOmega \setminus \{\xi _1,\ldots , \xi _k\}\).
If such a bounded harmonic function h exists, then
for all \(z\in \varOmega \).
For the proof of the preceding result, see Theorems 5 and 6 of Sect. 42 of [8].
Regarding the values of a harmonic function in a domain \(\varOmega \) (a nonempty open connected set), we recall the maximum principle (Theorem 1.8 of [2]):
Theorem C
Suppose \(\varOmega \) is a domain, u is real-valued and harmonic on \(\varOmega \), and u has a maximum or a minimum in \(\varOmega \). Then u is constant.
Lemmas 1 and 2 below are utilized in key steps of the proof of the main result.
Lemma 1
Suppose that \(f \in L^\infty \) with \(f \ge \frac{\pi }{2}\) almost everywhere. Let \(\lambda \ge 0\) and
be the domain in Fig. 1 and put \(L = \partial G_\lambda \cap \{\mathrm {Im\,}z>\lambda +1\}\). Then there exists a unique bounded harmonic function \(v_\lambda \) on \(G_\lambda \) with the boundary values
Proof
Let \(\tau \) be a conformal map of \(G_\lambda \) onto the unit disk \(\mathbb {D}\). Then each straight-line piece of the boundary \(\partial G_\lambda \) is mapped to an arc on the circle \(\mathbb {T}\) (see Theorem II.3.4’ of [6]). Now consider the Poisson integral of the function that equals 1 on the arcs to which \(\tau \) maps L and 0 on the complementary arcs. The composition of the Poisson integral with \(\tau \) gives the desired harmonic function. Uniqueness and the bounds
follow from Theorems B and C. \(\square \)
The following characterization of subharmonic functions is also needed:
Theorem D
Let u be a function on \(\mathbb {D}\) and suppose it satisfies the following conditions:
-
(i)
\(\infty \le u < \infty \), \(u\not \equiv -\infty \),
-
(ii)
u is upper semi-continuous in \(\mathbb {D}\),
-
(iii)
for each \(z_0\in \mathbb {D}\), there is an \(r_0\) such that \(D(z_0,r_0)\subset \mathbb {D}\) and
$$\begin{aligned} u(z_0) \le \frac{1}{2\pi } \int _0^{2\pi } u(z_0 + re^{i\theta })\,d\theta \end{aligned}$$(2.9)for all \(0\le r\le r_0\).
Then u is subharmonic in \(\mathbb {D}\).
For the proof of the preceding result, see Theorem II.13 of [16].
Lemma 2
Let \(\lambda , f, G_\lambda \), and \(v_\lambda \) be as in Lemma 1 and \(E \subset \mathbb {T}\) be a measurable set with \(0<|E|<2\pi \). Define
where \(\rho _E\) is defined in (1.5). Let \(\varOmega _\lambda = X^{-1}(G_\lambda )\) and define \(H_\lambda : \mathbb {D}\rightarrow \mathbb {C}\) by
Then \(H_\lambda \) is continuous and subharmonic with \(0\le H_\lambda < 1\).
Proof
First, if \(z\in \partial \varOmega _\lambda \cap \mathbb {D}\), then \(X(z)\notin G_\lambda \) but \(X(z) \in \overline{G_\lambda \cap X(\mathbb {D})}\), and hence \(X(z) \in \partial G_\lambda \). Further, since \(|\mathrm {Re\,}X(z)|<\Vert f\Vert _\infty \) and \(z\in \mathbb {D}\),
and so \(v_\lambda (X(z)) = 0\) according to (2.7)). If \(z_k\rightarrow z\) in \(\varOmega _\lambda \), then, using the continuity of \(v_\lambda \) up to the point X(z), \(v_\lambda (X(z_k))\rightarrow 0\), which implies that \(H_\lambda \) is continuous on \(\mathbb {D}\). That \(0\le H_\lambda < 1\) follows from \(0< v_\lambda < 1\) (see (2.8)) and the definition of the function \(H_\lambda \).
The local mean value inequality in (2.9) holds for each point of \(\varOmega _\lambda \) because \(H_\lambda =v_\lambda \circ X\) is harmonic in \(\varOmega _\lambda \), and it holds for each point of \(\mathbb {D}\setminus \varOmega _\lambda \) because \(H_\lambda \ge 0\) equals zero there. Thus, by Theorem D, the function \(H_\lambda \) is subharmonic in \(\mathbb {D}\). \(\square \)
3 Proof of Theorem 1
Suppose that \(f\in L^\infty \) with \(f\ge \pi /2\) almost everywhere and \(0<|E|<2\pi \). Let
where \(\rho _E\) is defined in (1.5). Then
almost everywhere according to (2.5).
The proof of (1.8) consists of several steps.
Step 1
Since \(\mathrm {Re\,}X^* = f\rho _E \ge \pi /2\) almost everywhere on E and \(\mathrm {Re\,}X^* \le -\pi /2\) almost everywhere on \(\mathbb {T}\setminus E\), there is a point \(w\in \mathbb {D}\) such that \(|\mathrm {Re\,}X(w)|<\pi /2\).
Step 2
Let the open set \(G_\lambda \) (see Fig. 1) and \(L = \partial G_\lambda \cap \{\mathrm {Im\,}z>\lambda +1\}\) be defined as in Lemma 1.
Then Lemma 1 asserts that there exists a unique bounded harmonic function \(v_\lambda \) on \(G_\lambda \) with the boundary values
Step 3
Let \(\varOmega _\lambda = X^{-1}(G_\lambda )\). Then, according to Lemma 2, the function \(H_\lambda : \mathbb {D}\rightarrow \mathbb {C}\) defined by
is continuous and subharmonic with \(0\le H_\lambda < 1\).
Step 4
There is a \(C>0\), independent of \(\lambda \), such that
where \(E_\lambda \) is defined by (1.8) and w is defined in Step 1. Since \(H_\lambda \) is bounded, it trivially has a harmonic majorant, and hence, by Theorem I.6.7 of [4],
for \(z\in \mathbb {D}\). It follows from (2.5) that, for almost every \(\theta \in [0,2\pi ]\setminus E_\lambda \),
Therefore, using the definition of \(H_\lambda \) and the properties of \(v_\lambda ,\)
Now, by (3.2) and Lebesgue’s dominated convergence theorem,
where the constant is independent of \(\lambda \).
Step 5
It is difficult to obtain the desired estimate \((v_\lambda \circ X)(w) \ge Ce^{-\lambda }\) directly. Instead we estimate \(v_\lambda \) from below by another harmonic function \(g_\lambda \), defined on a vertical strip, which allows us to compute \(g_\lambda \) explicitly in the next step.
Let \(g_\lambda \) be the bounded harmonic function in the strip \(S=\{|\mathrm {Re\,}z|<\pi /2\}\) with the boundary values
Notice that \(\lim _{y\rightarrow \infty } v_0(\pm \pi /2 + i y) = 1\) (which can be verified using a conformal map of \(\varOmega _0\) onto \(\mathbb {D}\)). Therefore, since \(v_0\) is continuous, Theorem C implies that there is a positive constant C such that
for all \(y>2\). Clearly \(v_0(\pm \pi /2+iy) \ge 0 = Cg_0(\pm \pi /2+iy)\) for all \(y<2\). Thus, by Theorem B, \(v_0(z) - Cg_0(z) \ge 0\) for all \(z\in S\), and so for \(\lambda >0\) and \(z\in S\),
Consequently, by (3.1),
where the constant C is independent of \(\lambda \).
Step 6
Let \(\tau =x+iy = X(w)\), so \(|x|<\pi /2\) and \(y\in \mathbb {R}\). We show that there is a constant \(C>0\), independent of \(\lambda \), such that \(g_\lambda (\tau )\ge Ce^{-\lambda }\), which completes the proof.
defines a conformal mapping from S onto \(\mathbb {D}\) with \(F(ib)= i \tanh (\frac{1}{2} b) \rightarrow \pm i,\) as \(b \rightarrow \pm \infty \). Notice that
for \(z\in \mathbb {D}\). Thus, using the mapping \(z\mapsto -iz\), we see that
maps S conformally onto itself with \(\pm \frac{\pi }{2} \mapsto (0,\mp \infty )\) and \((0,\infty )\mapsto \frac{\pi }{2}\). Therefore,
To evaluate \(g_\lambda \) at \(\tau =x+iy\), write \(\tau _\lambda = \frac{\tau -i(\lambda +2)}{2}\) and notice that
using the formula \(\arg \tan (a+ib) = \arctan \frac{\sinh 2b}{\sin 2a}\) and the fact that \(\arctan \) is odd. Observe also that the expression for \(g_\lambda (\tau )\) is valid in S, but not necessarily on \(\partial S\) and the first expression for \(g_\lambda (z)\) is valid in \(\bar{S}\). We first established the expression for \(g_\lambda (z)\) with the desired boundary behavior on \(\partial S\) and then considered only the behavior inside S, where the expressions for \(g_\lambda (z)\) and \(g_\lambda (\tau )\) coincide. Since \(\arctan a = \frac{\pi }{2}-\arctan a^{-1}\) for \(a>0\), we have, for \(\lambda +2>y\),
which gives the estimate. This completes the proof of (1.8).
It remains to prove (1.9). As in Sect. 4 of Chapter I in [4], consider the distribution function
for \(\lambda >0\). Notice that for \(\lambda \ge 1\),
(see (1.8) for the definition of \(E_{\log \lambda }\)), and so by Lemma I.4.1 in [4],
4 Further Remarks
In this section we provide remarks and examples related to Theorem 1. We show first that the conditions in the theorem are optimal.
Remark 3
In Theorem 1, (i) the condition on E is optimal, and (ii) the condition that \(f\ge \pi /2\) almost everywhere is optimal.
Proof
(i) Let \(E=[0,2\pi ]\) and \(f=\pi /2\) on E. Then \(f\rho _E = f\) and so trivially \(e^{\widetilde{f\rho _E}}\in L^1\) by (1.3).
Further, by Corollary III.1.8 of [7],
which implies that there is no constant C for which (1.8) holds in this case.
(ii) Suppose that \(\pi /4<f<\pi /2\) on an interval \(I=(a,b)\subset (0,2\pi )\) and \(f=\pi /2\) on \([0,2\pi ]\setminus I\). Let \(E = [0,a+\delta ] \cup [b-\delta ,2\pi ]\) for some small \(\delta >0\). Then \(\exp (\widetilde{f\rho _E})\) is integrable by Zygmund’s Theorem A because the gap in the essential range of \(f\rho _E\) is strictly less than \(\pi \). Consequently, the condition is optimal for (1.9), and it must also be optimal for (1.8) because it was used to prove (1.9). \(\square \)
Remark 4
In addition to the example in the previous proof, there are functions f which are not constant and still satisfy \(|E_\lambda | \lesssim e^{-\lambda }\) as in (4.1). Indeed, let \(0<|E|<2\pi \) and let f be Hölder with \(f(\theta ) = \pi /2\) for all \(\theta \in \partial E\). If \(g=(f-\pi /2)\rho _E\), then it is not difficult to see that g is Hölder and hence \({\tilde{g}}\) is Hölder. Also,
Thus, by Exercise VI.18 of [4],
which implies that \(|E_\lambda | = \left| \{ t : |\widetilde{f\rho _E}(t)|>\lambda \}\right| \lesssim e^{-\lambda +\Vert {\tilde{g}}\Vert _\infty } \le Ce^{-\lambda }\) for some constant C.
In fact, (0.2) of [18], when \(\psi = \widetilde{f\rho _E} = {\tilde{g}} + \frac{\pi }{2}{\tilde{\rho }}_E\), implies the stronger result
Notice that the converse is not true, however, that is, (4.2) does not imply that
(see Wolff’s counterexample on page 52 of [18]).
Open Problem 1
In the preceding remark, when \(\psi = \frac{1}{2}\log |H|\) with H univalent and zero free, Baernstein [3, 18] posed a question of whether (4.2) implies (4.3). This seems to be still open.
Our next example concerns outer functions from the theory of Hardy spaces. Recall that an outer function is a function G on the unit disk which can be written in the form
where \(|\alpha | = 1\) and \(\varphi \) is a positive measurable function on \(\mathbb {T}\) such that \(\log \varphi \in L^1\). Similarly to (2.4),
where u is the Poisson integral of \(\log \varphi \) and v is the harmonic conjugate function of u so that \(e^{iv(0)}=\alpha \).
Let \(0<p<\infty \) and let f be analytic in \(\mathbb {D}\). Then the function f is in the Hardy space \(H^p\) if
Notice that \(G\in H^p\) if and only if \(\varphi \in L^p\) (see, e.g., Sect. II.4 of [4]). We can now use Theorem 1 to determine when certain outer functions are not in \(H^p\) as shown in the following example.
Example 1
Given a real-valued function f in \(L^\infty \), define
Using (4.5), it is easy to see that \(\varPhi _f\) and its inverse \(\varPhi _f^{-1}\) are both outer functions. Denote by \(\mathcal {R}(f)\) the essential range of f as before. Let
be an interval such that \(I\cap \mathcal {R}(f) = \emptyset \) and \(|I| \ge \frac{2\pi }{p}\). Then \(\varPhi _f, \varPhi _f^{-1}\notin H^p\). In particular, \(\varPhi _f^{\pm 1} \notin H^2\) if \(\mathcal {R}(f)\) has a gap of length \(\ge \pi \).
To see this, notice first that (similarly to (2.5))
so \(|(\varPhi _f^{\pm 1})^*(t)| = e^{\mp \frac{1}{2}{\tilde{f}}(t)}\). Since \(\varPhi _f\) is an outer function, it follows from Theorem 1 that \(\varPhi _f, \varPhi _f^{-1}\notin H^p\) if \(\frac{p}{2}|I|\ge \pi \).
To illustrate the effect of jumps in relation to exponential integrability (see Example 2), the following lemma will be needed. Its proof is included for completeness because we have not found it in the literature.
Lemma 3
For \(0<x<2\pi \), let
Then
as \(x\rightarrow 0+\), where B is a positive constant.
Proof
Define \(B_x(y) = \sum _{2\le n \le y} \cos nx\). Then, for \(0<x<\pi \) and \(y>2\), \(|B_x(y)| \le \frac{\pi }{x} + 1\). Therefore, integrating by parts,
and so
Suppose now that \(x<\tfrac{\pi }{2N}\). Then
Since
where B is a constant (see Exercise 8.20 of [1]), we get for \(x<\frac{\pi }{2N}\),
as \(N \rightarrow \infty .\)
Choosing \(N = \frac{1}{x}(\log \frac{1}{x} \log \log \frac{1}{x})^{-1/3}\), we obtain (4.6) as \(x\rightarrow 0+\). \(\square \)
Remark 5
In the following example, for small values of |x|, we only need the following consequence of the preceding lemma:
which can also be obtained using Theorem V.1.5 of [19].
Example 2
Let \(t_0\in (0,2\pi )\). For \(0<\delta <1\) define
and suppose that g is Hölder continuous elsewhere. By Theorem V.1.3 of [20], the series
converges uniformly and defines a continuous function on \([0,2\pi ]\). It is well known that
which is continuous on \([0,2\pi ]\setminus \{t_0\}\) (see Theorem I.2.6 of [20]). Define \(f=g+h\). Then \(\lim _{t\rightarrow t_0\pm } f(t) = \mp \pi /2\), so f is piecewise continuous with only one jump, which is of size \(\pi \). We want to determine whether \(\exp ({\tilde{f}})\) is integrable. Obviously we cannot use Zygmund’s Theorem A. Notice also that we cannot apply Theorem 1 because we do not know without further inspection whether \(f\ge \pi /2\) a.e. on \((0,t_0)\) and \(f\le -\pi /2\) a.e. on \((t_0,2\pi )\).
Now, similarly to (1.6), and using the fact that g is Hölder on \([0,2\pi ]\setminus \{t_0\}\), there is a constant \(C>0\) such that
for \(0\le t \le 2\pi \). To estimate \({\tilde{h}}\) near \(t_0\), notice first that Lemma 3 implies that
for t sufficiently close to \(t_0\). Therefore, for some constants C, we have
which is integrable in a neighborhood of \(t_0\), and hence \(e^{{\tilde{f}}}\in L^1\).
Now, of course, by Theorem 1, the integrability of \(e^{{\tilde{f}}}\) means that f has values in \((-\pi /2, \pi /2)\) on a set of positive measure. In fact, by V.2.13 of [19], the function h is positive on \((t_0, t_0+\epsilon )\) for some \(\epsilon >0\) and hence negative on \((t_0-\epsilon ,t_0)\) (as an odd function), and so indeed \(f<\pi /2\) on \((t_0-\epsilon , t_0)\) and \(f>-\pi /2\) on \((t_0, t_0+\epsilon )\).
Remark 6
We can use the previous example to construct a function \(f\in L^\infty \) such that \(\Vert f\Vert _\infty = \pi /2\), \(|f|<\pi /2\) and \(e^{{\tilde{f}}}\in L^1\). This should be compared with Zygmund’s result in (1.2).
Open Problem 2
In Theorem 1, it is assumed that \(f\in L^\infty \) as in Zygmund’s Theorem A. It is natural to ask whether the conclusion of Theorem 1 remains true for unbounded functions.
We finish this section with a connection between the estimate in (1.8) and the distance in \({{\,\mathrm{BMO}\,}}\) to \(L^\infty \) (see Sect. VI.6 of [4]).
Suppose that the conditions of Theorem 1 are satisfied, that is, \(f\in L^\infty \) with \(f\ge \pi /2\) almost everywhere and \(0<|E|<2\pi \). If \(\Vert g\Vert _\infty <\pi /2\), then by Theorems A and 1, \(\widetilde{f\rho _E} - {\tilde{g}}\notin L^\infty \), and so \({{\,\mathrm{dist}\,}}(\widetilde{f\rho _E}, L^\infty ) \ge \pi /2\), where
which is equivalent to \(\inf _{g\in L^\infty } \Vert \varphi -g\Vert _{{{\,\mathrm{BMO}\,}}}\) (see, e.g., page 250 in [4]). By Corollary VI.6.6, there is no \(\epsilon \in (0,1)\) such that
for all \(\lambda \ge 0\), where the supremum is taken over all arcs \(I\subset \mathbb {T}\) and the average \(\varphi _I\) is defined by \(\varphi _I = \frac{1}{|I|}\int _I \varphi \) for \(\varphi \in L^1\).
The same conclusion also follows directly from (1.8) if \(f\ge \pi /2\), which is no surprise because the requirement that f has a gap in its essential range is a stronger assumption than \({{\,\mathrm{dist}\,}}(\widetilde{f\rho _E}, L^\infty ) \ge \pi /2\). Indeed, assume that (1.8) holds and (4.7) does not hold, so that there exists \(\epsilon \in (0,1)\) satisfying the estimate (4.7). Denote \(h = f\rho _E\) so that \({\tilde{h}}_0 = \frac{1}{2\pi }\int _0^{2\pi } \tilde{h}\), and choose \(s \ge 0\) so large that \(\lambda = {\tilde{h}}_0 +s \ge 0.\) Then there exists a constant \(C > 0\) so that
so we have \(C e^{-\lambda } = C e^{-(\tilde{h}_0+s)} \le e^{-\frac{s}{\epsilon }}.\) Therefore \(Ce^{-\tilde{h}_0} \le e^{(1-\frac{1}{\epsilon })s} \rightarrow 0\) as \(s \rightarrow \infty ,\) which is a contradiction.
5 Complex-Valued Functions
While real-valued functions are of particular importance in the study of exponential integrability of their conjugate functions, especially in connection with applications, such as Riemann–Hilbert problems and spectral theory of Toeplitz operators above, it would also be of interest to consider the case of complex-valued functions. Indeed, as in Zygmund’s Theorem A, we may consider a complex-valued \(f\in L^\infty \) and ask under what conditions is \(\exp (|\widetilde{f\rho _E}|)\) not integrable. As in the proof of Theorem 1, we can define \(m(\lambda ) = |\{ t : \exp (|\widetilde{f\rho _E}(t)|)>\lambda \}|\) and show that if there is a constant \(|\{t : |\widetilde{f\rho _E}(t)|>\lambda \}| \ge Ce^{-\lambda }\) for all \(\lambda \ge 0\), then \(\exp (|\widetilde{f\rho _E}|)\) is not integrable.
However, in the complex case, the function \(f\rho _E\) no longer has a similar (geometric) meaning as in the real case where it can be related to a gap in the essential range. For this reason, we say that a set \(A\subset \mathbb {C}\) has a gap of size \(g>0\) if \(A = B \cup C\) for some sets B and C of positive measure with \({{\,\mathrm{dist}\,}}(A,B)\ge g\). With this, we can state (1.9) in Theorem 1 as follows: If \(f\in L^\infty \) is real and \(\mathcal {R}(f)\) has a gap of size at least \(\pi \), then \(\exp ({\tilde{f}})\) is not integrable.
Open Problem 3
Given a complex-valued function f in \(L^\infty \), find a converse to Zygmund’s Theorem A.
It may be useful to try to relate the exponential integrability of \({\tilde{f}}\) to the size of the gap in the essential range of f as in the real case.
References
Apostol, T.M.: Mathematical Analysis, 2nd edn. Addison-Wesley Publishing Co., Reading (1974)
Axler, S., Bourdon, P., Ramey, W.: Harmonic Function Theory, vol. 137, 2nd edn. Springer, New York (2001)
Baernstein, A.: Analytic functions of bounded mean oscillation. In: Aspects of Contemporary Complex Analysis (Proc. NATO Adv. Study Inst., Univ. Durham, Durham, 1979), pp. 3–36. Academic Press, London-New York (1980)
Garnett, J.B.: Bounded Analytic Functions, Revised 1, p. 236. Springer, New York (2007)
Garnett, J.B., Marshall, D.E.: Harmonic Measure. Reprint of the 2005 original. New Mathematical Monographs, p. 2. Cambridge University Press, Cambridge (2008)
Goluzin, G.M.: Geometric Theory of Functions of a Complex Variable. Translations of Mathematical Monographs, vol. 26. American Mathematical Society, Providence, RI (1969)
Katznelson, Y.: An Introduction to Harmonic Analysis, 3rd edn. Cambridge University Press, Cambridge (2004)
Lavrent’ev, M. A., Shabat, B. V.: Methods of the Theory of Functions in a Complex Variable. (Russian; translated into French, German and Spanish) Fifth edition. Nauka, Moscow (1987).
Lekishvili, M.M.: Exponential integrability of a conjugate function. Math. Zametki (Russian) 48(2), 155–157 (1990)
Perälä, A., Virtanen, J.A., Wolf, L.: A Riemann-Hilbert problem with a vanishing coefficient and applications to Toeplitz operators. Concr. Oper. 1, 28–36 (2013)
Shargorodsky, E.: On some geometric conditions of Fredholmity of one-dimensional singular integral operators. Integral Equ. Oper. Theory 20(1), 119–123 (1994)
Shargorodsky, E.: A remark on the essential spectra of Toeplitz operators with bounded measurable coefficients. Integral Equ. Oper. Theory 57(1), 127–132 (2007)
Shargorodsky, E., Virtanen, J.A.: Uniqueness results for the Riemann–Hilbert problem with a vanishing coefficient. Integral Equ. Oper. Theory 56(1), 115–127 (2006)
Shargorodsky, E., Toland, J.F.: Bernoulli free-boundary problems. Mem. Am. Math. Soc. 196(914), 70 (2008)
Storoženko, È.A.: Embedding in the class \(e^{L}\). Math. Zametki (Russian) 10, 17–24 (1971)
Tsuji, M.: Potential Theory in Modern Function Theory. Reprinting of the 1959 Original. Chelsea Publishing Co., New York (1975)
Virtanen, J.A.: A remark on the Riemann-Hilbert problem with a vanishing coefficient. Math. Nachr. 266, 85–91 (2004)
Wolff, T.H.: Counterexamples to two variants of the Helson-Szegő theorem. Dedicated to the memory of Tom Wolff. J. Anal. Math. 88, 41–62 (2002)
Zygmund, A.: Sur les fonctions conjugées. Fund. Math. 13, 284–303 (1929)
Zygmund, A.: Trigonometric Series, vol. I, 3rd edn. Cambridge University Press, Cambridge (2002)
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Alexey Karapetyants.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Gissy was supported by a University of Reading doctoral grant. Miihkinen was supported by Emil Aaltonen Foundation. Virtanen was supported in part by Engineering and Physical Sciences Research Council grant EP/T008636/1.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included 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/.
About this article
Cite this article
Gissy, H., Miihkinen, S. & Virtanen, J.A. On the Exponential Integrability of Conjugate Functions. J Fourier Anal Appl 27, 87 (2021). https://doi.org/10.1007/s00041-021-09885-4
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00041-021-09885-4