On the abscissas of a Dirichlet series and its subseries supported on prime factorization
Extracted source text · arXiv 2503.23349
%------------------------------------------------------------
%------------------------------------------------------------
%\documentclass[oneside,11pt,reqno, psamsfonts]{amsart}
\documentclass[a4paper,oneside,15pt, reqno, psamsfonts]{amsart}
%------------------------------------------------------------
\usepackage{etoolbox}
%-----------------------end -------------------
\makeatletter
\patchcmd\maketitle
{\uppercasenonmath\shorttitle}
{}
{}{}
\patchcmd\maketitle
{\@nx\MakeUppercase{\the\toks@}}
{\the\toks@}
{}
{}{}
%----titre lowercase-----------
\patchcmd\@settitle{\uppercasenonmath\@title}{\Large}{}{}
%\patchcmd\@settitle{\uppercasenonmath\@title}{} {}{}
%-------------------------
\patchcmd\@setauthors
{\MakeUppercase{\authors}}
{\authors}
{}{}
\makeatother
%------------------------------------------------------------
%\usepackage[notref,notcite]{showkeys}
\usepackage{amsmath,amssymb,amsthm, amsfonts}
%\usepackage{relsize}
\usepackage{mathrsfs}
\usepackage{color}
\usepackage{float}
\usepackage{url}
\usepackage{tikz-cd}
\usepackage{tikz}
\usepackage{amsthm}
\usetikzlibrary{arrows}
\usetikzlibrary{calc}
\usepackage{enumitem}
\usepackage{mathtools}
%\usepackage{pxfonts}
%-----------------
%------------------------------------------------------------
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
%\usepackage{refcheck}
%------------------------------------------------------------
%\textheight 22.5truecm \textwidth 14.5truecm
%\setlength{\oddsidemargin}{0.35in}\setlength{\evensidemargin}{0.35in}
%\setlength{\topmargin}{-.5cm}
%\usepackage{geometry}
%\geometry{left=2cm,right=2cm,top=2cm,bottom=2cm}
% ------------------------------------------------------------
\newtheorem{theorem}{Theorem}[section]
\newtheorem{observation}[theorem]{Observation}
\newtheorem{question}[theorem]{Question}
%\newtheorem{question}{Question}[section]
\newtheorem{step}{Step}
\newtheorem{definition}{Definition}[section]
\newtheorem{definitions}{Definitions}[section]
\newtheorem{notation}{Notation}[section]
\newtheorem{proposition}{Proposition}[section]
\newtheorem{corollary}[proposition]{Corollary}
\newtheorem{lemma}{Lemma}[section]
\newtheorem{remark}[lemma]{Remark}
\newtheorem{example}[lemma]{Example}
\numberwithin{equation}{section}
\newcommand{\beq}{\begin{eqnarray}}
\newcommand{\eeq}{\end{eqnarray}}
\newcommand{\beqn}{\begin{eqnarray*}}
\newcommand{\eeqn}{\end{eqnarray*}}
\newcommand{\rar}{\rightarrow}
\newcommand*{\Ge}{\geqslant}
\newcommand*{\Le}{\leqslant}
\newcommand*{\inp}[2]{\langle{#1},\,{#2} \rangle}
\newcommand*{\mf}{\mathbf}
\newcommand*{\child}[1]{\mathsf{Chi}(#1)}
\pagestyle{plain}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\newcommand*{\uwam}[1]{\marginpar{\sf\tiny{{\textcolor{red}
{$\blacktriangleleft$} \textcolor{blue}{#1}
\textcolor{red}{$\blacktriangleright$}}}\normalsize}}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%% hyperref %%%%%%%%%%%%%%%%%%%%
%\usepackage[colorlinks=true,pagebackref=true]{hyperref}
\usepackage[colorlinks=true]{hyperref}
\hypersetup{urlcolor=blue, citecolor=red , linkcolor= blue}
%----------
\usepackage[capitalise,noabbrev,nameinlink]{cleveref} % \labelcref pour les equations
%\title{\noindent\rule{\textwidth}{0.03cm}\\\texttt{\normalsize A convergence problem of Dirichlet series, and results on multipliers of Hilbert spaces of Dirichlet series}\\\noindent\rule{\textwidth}{0.03cm}}
%%\noindent\rule{\textwidth}{0.05cm}
%\date{}
%\author{\normalsize \texttt{Chaman Kumar Sahu}\\ \normalsize \texttt{Department of Mathematics}\\
% \normalsize \texttt{Indian Institute of Technology Bombay}\\ \normalsize \texttt{Powai, Mumbai-400076, India}\\ \normalsize \texttt{Email: sahuchaman9@gmail.com}}
%----------
%%%%%%%%%%%% \begin{document} %%%%%%%%%%%%%%%%%%%%
\begin{document}
\title[On the abscissas of a Dirichlet series and its subseries supported on prime factorization]
{On the abscissas of Dirichlet series and its subseries supported on prime factorization}
\author[C. K. Sahu]{Chaman Kumar Sahu}
\address{Department of Mathematics, Indian Institute of Technology Bombay,
Mumbai 400076, India}
\email{sahuchaman9@gmail.com, chamanks@math.iitb.ac.in}
% \subjclass[2010]{Primary 11M99, 46E22; Secondary 47B35, 47G10.}
%
\subjclass[2020]{Primary 30B50, 46E22.}
\keywords{Dirichlet series, Absolute convergence}
\maketitle
% \title[\textcolor{blue}{\normalsize }]
% {\textcolor{blue}{\normalsize A convergence problem of Dirichlet series, and multiplication operators on Hilbert spaces of Dirichlet series}}
%
% \author{Chaman Kumar Sahu}
% \address{Department of Mathematics and Statistics\\
% Indian Institute of Technology Bombay, India}
% \email{sahuchaman9@gmail.com}
\begin{abstract}
For a sequence $\{a_n\}_{n \Ge 1} \subseteq (0, \infty)$ and a Dirichlet series $f(s) = \sum_{n=1}^\infty a_n n^{-s},$ let $\sigma_a(f)$ denote the abscissa of absolute convergence of $f$ and let
\beqn
\delta_a(f): = \inf\Bigg\{\Re(s) : \sum\limits_{\substack{j= 1 \\\tiny{\mbox{\bf gpf}}(j) \Le p_n }}^\infty a_j j^{-s} < \infty ~\text{for all}~ n \Ge 1\Bigg\},
\eeqn
where $\{p_j\}_{j \Ge 1}$ is an increasing enumeration of prime numbers and $\text{\bf gpf}(n)$ denotes the greatest prime factor of an integer $n \Ge 2.$
One significant aspect of these abscissas is their crucial role in analyzing the multiplier algebra of Hilbert spaces associated with diagonal Dirichlet series kernels.
The main result of this paper establishes that $\sigma_a(f)- \delta_a(f)$ can be made arbitrarily large, meaning that it can be equal to any non-negative real number. As an application, we determine the multiplier algebra in some cases and, in others, gain insights into the structure of the multiplier algebra of certain Hilbert spaces of Dirichlet series.
% We next derive that $\mathscr H_\mf d$ is closed under the Dirichlet product by a weighted $\ell^1$-space. However, a similar result appears in \cite[Proposition~4.3]{BCD} for certain Banach spaces of Dirichlet series under the additional assumption
% that constant functions belong to the space. But, in our result, constant functions need not be in $\mathscr H_\mf d.$
\end{abstract}
% \textcolor{red}{{\it Subject classifications (2020):} Primary 30B50, 46E22; Secondary 11Z05.}
\section{Introduction}
Let $\mathbb N,$~$\mathbb R$ and $\mathbb C$ denote the sets of positive integers, real and complex numbers, respectively.
For $s \in \mathbb C,$ let $\Re(s),$ $|s|,$ $\overline{s}$ and $\arg(s)$ denote the real part, the modulus, the complex conjugate and the argument of $s,$ respectively.
For $\rho \in \mathbb R,$ let $\mathbb H_\rho$ denote the right half-plane $\{s \in \mathbb C : \Re(s) > \rho\}.$
A {\it Dirichlet series} is a series of the form
\beqn
f(s) = \sum_{j=1}^{\infty} a_{j} j^{-s},
\eeqn
where $a_j \in \mathbb C,~j \Ge 1$ and $s$ is a complex variable.
If $a_j = 1$ for all $j \Ge 1,$ then we have the Riemann zeta function, denoted by $\zeta.$
If $f$ is convergent at $s=s_{0}$, then it converges
uniformly throughout the angular region $\{s \in \mathbb C : |\arg(s-s_0)| < \frac{\pi}{2} -\delta\}$ for every positive real number $\delta < \frac{\pi}{2}.$ Consequently, $f$ defines a holomorphic function on $\mathbb H_{\Re(s_0)}$ (refer to \cite[Chapter~IX]{Ti} for the basic theory of Dirichlet series).
The abscissa of {\it absolute convergence} $\sigma_{a}(f)$ of $f$ is the extended real number defined by
\beqn
% \label{sigma-form}
\sigma_{a}(f) = \inf \Big\{\Re(s) : \sum_{n=1}^{\infty} |a_{n}| n^{-\Re(s)} < \infty \Big\}.
\eeqn
For an integer $n \Ge 2,$ let $\text{\bf gpf}(n)$ denote the greatest prime factor of $n,$ and let $\text{\bf gpf}(1) = 1.$ Define
\beqn
% \label{delta-form}
\delta_a(f) := \inf\Bigg\{\Re(s) : \sum\limits_{\substack{j= 1 \\\tiny{\mbox{\bf gpf}}(j) \Le p_n }}^\infty |a_j| j^{- s} < \infty ~\text{for all}~ n \in \mathbb N\Bigg\},
\eeqn
where $\{p_j\}_{j=1}^\infty$ is the increasing enumeration of prime numbers.
Let $\mathcal D$ denote the set of all functions which are representable by a convergent Dirichlet series in some right half plane.
% \begin{remark}
% Let $f(s) = \sum\limits_{n=1}^\infty a_nn^{-s}$ and $g(s) = \sum\limits_{n=1}^\infty b_nn^{-s}$ be in $\mathcal D.$ If there exists $n_0 \in \mathbb N$ such that $a_n = b_n$ for all $n \Ge n_0,$ then $\sigma_a(f) = \sigma_a(g)$ and $\delta_a(f) = \delta_a(g).$
% %The latter one could possibly be $-\infty.$
% \uwam{do we need this remark: No!?}
% \end{remark}
%\subsection{A convergence problem of Dirichlet series}
If $f \in \mathcal D$ has all positive coefficients, then
$\delta_a(f) \Le \sigma_a(f).$ One of the importance of these abscissas lies in the fact that they appear in the study of Hilbert space associated to diagonal Dirichlet series kernels and its multiplier algebra (see, \cite{Sa}, \cite[Theorem~3.1]{St} and Proposition~\ref{stetler_result_repeat}).
Thus, the following natural question arises.
%Thus, the following natural question arises.
\begin{question}\label{Q-1}
For $r \Ge 0,$ does there exist $f \in \mathcal D$ with positive coefficients such that $\sigma_a(f) - \delta_a(f) = r?$
\end{question}
If $r =1,$ Question~\ref{Q-1} has an affirmative answer. Indeed, consider $\zeta(s) = \sum_{n=1}^\infty n^{-s},$ which has all coefficients equal to $1.$ Then, $\sigma_a(\zeta) = 1,$ and for every $n \in \mathbb N,$ since the subseries
\beq \label{zeta-n}
\zeta_n(s): = \sum\limits_{\substack{j= 1 \\\tiny{\mbox{\bf gpf}}(j) \Le p_n }}^\infty j^{- s} = \prod_{j=1}^n \frac{1}{1-p_j^{-s}}
\eeq
converges on $\mathbb H_0$ (see \cite[Lemma~3.2]{St}) and diverges at $0$, it follows that $\delta_a(\zeta) = 0.$ Therefore, $\sigma_a(\zeta) - \delta_a(\zeta) = 1.$
From the definitions of both abscissas, one might naturally expect that the difference between these abscissas can not be too large. However, in relation to Question~\ref{Q-1}, we establish a surprising result in this context, as presented below.
\begin{theorem}\label{arbitrary-large} The following statements are true.
\begin{itemize}
\item [$(i)$] For every real number $r \Ge 0$, there exists a Dirichlet series $f$ with positive coefficients such that $\sigma_a(f)-\delta_a(f) = r$.
\item [$(ii)$] If $r \Ge 1,$ there exists a Dirichlet series $f$ with multiplicative coefficients $($i.e., $a_{mn} = a_ma_n$ whenever $\gcd(m,n)=1)$ such that $\sigma_a(f)-\delta_a(f) = r$.
%\item [$(iii)$] There exists a Dirichlet series $f$ with positive coefficients such that $\sigma_a(f) = \infty$ and $\delta_a(f) = 0.$
\end{itemize}
\end{theorem}
%A proof of Theorem~\ref{arbitrary-large} is provided in Section~\ref{S2}. In relation to Theorem~\ref{arbitrary-large}, we provide two schemes to form Dirichlet series for the case $r = 0.$ The second methods seems very natural in the sense that many interesting examples falls into that category (See, Example~\ref{}).
A proof of Theorem~\ref{arbitrary-large} is given in Section~\ref{S2}. Regarding Theorem~\ref{arbitrary-large}(i), we present two approaches for constructing Dirichlet series in the case $r=0$. The second approach appears quite natural, as it encompasses many interesting examples (see, Example~\ref{example-inte}).
%For $r > 1,$ the Dirichlet series formed in the proof of Theorem~\ref{arbitrary-large} has multiplicative coefficients,
% \subsection{Multipliers of Hilbert spaces of Dirichlet series}
% For $\mf d = \{d_n\}_{n=1}^\infty \subseteq (0, \infty),$ let $D(s) = \sum_{n=1}^\infty d_n n^{-s}.$ The sequence $\mf d$ is called {\it admissible} if $\sigma_a(D) < \infty.$ With this sequence, we now associate a positive semi-definite Dirichlet series kernel $\kappa_\mf d$ given by
% \beqn
% % \label{kappa-d}
% \kappa_\mf d(s, u) = \sum_{n=1}^\infty d_n n^{-s -\overline{u}},\quad s, u \in \mathbb H_{\frac{\sigma_a(D)}{2}}.
% \eeqn
% Clearly, $\kappa_\mf d$ converges absolutely on $\mathbb H_{\frac{\sigma_a(D)}{2}} \times \mathbb H_{\frac{\sigma_a(D)}{2}},$ and does not extend on any larger domain.
% By the Moore's theorem \cite[Theorem~2.14]{P-R}, there exists an unique reproducing kernel Hilbert space $\mathscr H_\mf d$ associated with $\kappa_\mf d,$ which is given by
% \beqn
% \mathscr H_\mf d = \Big\{\sum_{n=1}^\infty a_n n^{-s} ~~:~~ \sum_{n=1}^\infty \frac{|a_n|^2}{d_n} < \infty\Big\}.
% \eeqn
% %We use the convention that $\mf 1 \notin \mathscr H_\mf d$ whenever $d_1 = 0.$ In this case, constant functions do not belong to $\mathscr H_\mf d.$
% By Cauchy-Schwarz inequality, every function in $\mathscr H_\mf d$ converges absolutely on $\mathbb H_{\frac{\sigma_a(D)}{2}},$ and hence $\Sigma(\mathscr H_\mf d) \Le \frac{\sigma_a(D)}{2}$ (see \eqref{abs-space}). Furthermore, $\kappa_\mf d(\cdot, \frac{\sigma_a(D)}{2}+\epsilon) \in \mathscr H_\mf d$ for any $\epsilon > 0,$ which admits a pole at $s = \frac{\sigma_a(D)}{2}-\epsilon.$ Thus,
% \beq
% \label{abs_H_d}
% %\sigma_{c}(\mathscr H_\mf d) =
% \Sigma(\mathscr H_\mf d) = \frac{\sigma_a(D)}{2}
% \eeq
% implying that the common domain of analyticity of $\mathscr H_\mf d$ is $\mathbb H_{\frac{\sigma_a(D)}{2}}.$
%
% Let $X$ be a non-empty domain in $\mathbb C$ and
% let $H^{\infty}(X)$ denote the Banach space of bounded holomorphic functions on $X$ endowed with supremum norm.
% Let $\kappa: X \times X \rar \mathbb C$ be a positive semi-definite kernel, and let $\mathcal H_\kappa$ be the reproducing kernel Hilbert space of complex-valued functions on $X$ associated to $\kappa.$ A function $\varphi :X \rar \mathbb C$ is called {\it multiplier of $\mathcal H_\kappa$} if $\varphi \cdot f \in \mathcal H_\kappa$ for every $f \in \mathcal H_\kappa.$ Denote by $\mathcal M(\mathcal H_\kappa),$ the set of all multipliers of $\mathcal H_\kappa.$
%
% \begin{remark}\label{abs-mult}
% For an admissible sequence $\mf d,$
% \beq \label{abs-mult-est}
%\Sigma(\mathcal M(\mathscr H_\mf d)) \Le \frac{\sigma_a(D)}{2}.
% \eeq
%Indeed, since constant functions are in $\mathscr H_\mf d,$ we get that $\mathcal M(\mathscr H_\mf d) \subseteq \mathscr H_\mf d.$ Hence,
%\beqn
%\Sigma(\mathcal M(\mathscr H_\mf d)) \Le \Sigma(\mathscr H_\mf d) \overset{\eqref{abs_H_d}}= \frac{\sigma_a(D)}{2}.
%\eeqn
%It is important to note that the conclusion remains valid even when
%$d_1 = 0.$ In this scenario, constant functions are not in $\mathscr H_\mf d$ and thus, the reasoning proceeds as follows:
%%for any distinct primes $p, q$ and $\varphi \in \mathcal M(\mathscr H_\mf d),$
%%\beq
%%\label{series_forms-1}
%%\varphi \cdot p^{-s} = \sum_{m=1}^\infty a_m m^{-s} ~~\text{and}~~ \varphi \cdot q^{-s} = \sum_{m=1}^\infty b_m m^{-s}, \quad s \in \mathbb H_{\frac{\sigma_a(D)}{2}}
%%\eeq
%%belong to $\mathscr H_\mf d.$ This immediately yields that
%%\beqn
%%\sum_{m=1}^\infty a_m (qm)^{-s} = \sum_{n=1}^\infty b_n (pn)^{-s}, \quad s \in \mathbb H_{\frac{\sigma_a(D)}{2}}.
%%\eeqn
%%By the uniqueness of Dirichlet series $($see \cite[Theorem~6]{HR}$),$ $a_m = 0$ for all integers $m \Ge 1$ with $p \nmid m,$ and hence by \eqref{series_forms-1},
%%\beqn
%%\varphi(s) = \sum_{m=1}^\infty a_m \Big(\frac{m}{p}\Big)^{-s} = \sum_{m = 1}^\infty a_{pm} m^{-s}, \quad s \in \mathbb H_{\frac{\sigma_a(D)}{2}}.
%%\eeqn
%%Since $\sum_{m=1}^\infty a_m m^{-s} \in \mathscr H_\mf d$ converges absolutely on $\mathbb H_{\frac{\sigma_a(D)}{2}}$ $($see \eqref{abs_H_d}$)$, $\sigma_a(\varphi) \Le \frac{\sigma_a(D)}{2},$ proving \eqref{abs-mult-est}.
%for any distinct primes $p, q$ and $\varphi \in \mathcal{M}(\mathscr{H}_\mf{d})$, we have
%\beq
%\label{series_forms-1}
%\varphi \cdot p^{-s} = \sum_{m=1}^\infty a_m m^{-s} \quad \text{and} \quad \varphi \cdot q^{-s} = \sum_{m=1}^\infty b_m m^{-s}, \quad s \in \mathbb{H}_{\frac{\sigma_a(D)}{2}}
%\eeq
%which both belong to $\mathscr{H}_\mf{d}$. This directly leads to the identity
%\beqn
%\sum_{m=1}^\infty a_m (qm)^{-s} = \sum_{n=1}^\infty b_n (pn)^{-s}, \quad s \in \mathbb{H}_{\frac{\sigma_a(D)}{2}}.
%\eeqn
%By the uniqueness of Dirichlet series (see \cite[Theorem~6]{HR}), it follows that $a_m = 0$ for all integers $m \geq 1$ with $p \nmid m$. Consequently, using \eqref{series_forms-1}, we obtain
%\beqn
%\varphi(s) = \sum_{m=1}^\infty a_m \Big(\frac{m}{p}\Big)^{-s} = \sum_{m=1}^\infty a_{pm} m^{-s}, \quad s \in \mathbb{H}_{\frac{\sigma_a(D)}{2}}.
%\eeqn
%Since the series $\sum_{m=1}^\infty a_m m^{-s} \in \mathscr{H}_\mf{d}$ converges absolutely in $\mathbb{H}_{\frac{\sigma_a(D)}{2}}$ (see \eqref{abs_H_d}), we conclude that $\sigma_a(\varphi) \leq \frac{\sigma_a(D)}{2}$, proving \eqref{abs-mult-est}.
% \end{remark}
%
%Recall that for two normed spaces $H_j (j=1,2),$~ $H_1$ is contractively included in $H_2$ if $H_1 \subseteq H_2$ (as a subset) and the inclusion map $H_1 \hookrightarrow H_2$ is a contraction. There exist various families of Hilbert spaces of Dirichlet series for which the multiplier algebra has been characterized (see, for example, \cite{BCD, Mc-1, Sa, St}). We establish conditions on $\mf{d}$ that guarantee a weighted $\ell^1$-space is contractively embedded in $\mathcal{M}(\mathscr{H}_\mf{d})$. In other words, we show that the Hilbert space $\mathscr{H}_\mf{d}$ is closed under the Dirichlet product with a weighted $\ell^1$-space of Dirichlet series (cf. \cite{Mash}).
%
% \begin{theorem}\label{l1-subsapce-cor}
% Let $\mf d=\{d_n\}_{n=1}^\infty$ be an admissible sequence. If, for each prime $p,$ the condition
% \beq \label{nece-suf-con}
% \frac{d_m}{d_{mp}} \Le p^{-\delta_a(D)}, \quad m \Ge 2
% \eeq
% holds,
% then the weighted $\ell^1$-space
% \beqn
% \Big\{\sum_{n=1}^\infty a_n n^{-s} : \sum_{n=1}^\infty |a_n| n^{-\frac{\delta_a(D)}{2}} < \infty \Big\}
% \eeqn
% is contractively contained in $\mathcal M(\mathscr H_\mf d).$
% \end{theorem}
%
% A similar result as above appears in \cite[Proposition~4.3]{BCD} for certain Banach spaces of Dirichlet series under the additional assumption
% that constant functions belong to the space.
%The Section~\ref{S3} presents a proof of Theorem~\ref{l1-subspace} along wiht some examples illustrating Theorem~\ref{l1-subsapce-cor}.
%
%
%\begin{proposition}
% \label{l1-subspace}
% For an admissible sequence $\mf d=\{d_n\}_{n=1}^\infty,$ let $\{\frac{d_m}{d_{mn}}\}_{m=2}^\infty$ be a bounded sequence for every integer $n \Ge 1$ and let $\mf w = \{w_n\}_{n=1}^\infty$ be defined by $w_n = \big(\sup\limits_{m \Ge 2}\frac{d_m}{d_{mn}}\big)^{1/2},~ n \Ge 1.$ If $\ell^1(\mf w)$ is given by
% \beqn
% %\label{wiener-contain}
% \ell^1(\mf w):= \Big\{\sum_{n=1}^\infty a_n n^{-s}: \sum_{n=1}^\infty |a_n|w_n < \infty \Big\},
% \eeqn
% then $\ell^1(\mf w)$ is contractively contained in $\mathcal M(\mathscr H_\mf d)$.
%\end{proposition}
% For later purpose, .
%
% which follows the arguments given as in the proof of \cite[Proposition~4.3]{BCD}. In our proof, we use Proposition~\ref{stetler_result_repeat} to conclude that a cauchy sequence in $\mathcal M(\mathscr H_\mf d)$ converges pointwise on $\mathbb H_{\frac{\delta_a(D)}{2}}$ as $\mathcal M(\mathscr H_\mf d) \subseteq \mathscr H_\mf d$ may or may not be true. This is where our proof diverts a little from that of \cite[Proposition~4.3]{BCD}.
%However, in Theorem~\ref{l1-subspace}, constant functions need not be in $\mathscr H_\mf d$.
\section{Proof of Theorem~\ref{arbitrary-large}}\label{S2}
We begin with the following lemma, which shall be used in the proof of Theorem~\ref{arbitrary-large}.
\begin{lemma}\label{eventually-equal-lemma}
Let $f(s) = \sum_{n=1}^\infty a_n n^{-s} \in \mathcal D$ with $a_n > 0$ for all $n \Ge 1$ and let
\beqn
f_n(s) = \sum\limits_{\substack{j= 1 \\\tiny{\mbox{\bf gpf}}(j) \Le p_n }}^\infty a_j j^{- s}, \quad n \Ge 1.
\eeqn
Then,
$
%\label{eventually-equal}
\delta_a(f) = \sup\{\sigma_a(f_n) : n \Ge m\}
$
for any integer $m \Ge 1.$
\end{lemma}
\begin{proof}
If there exists a real number $\sigma$ such that $f_n(\sigma) < \infty$ for all integers $n \Ge 1,$ then $\sigma_a(f_n) \Le \sigma$ for all $n \Ge 1,$ and hence $\sup\{\sigma_a(f_n) : n \Ge m\} \Le \sigma$ for any integer $m \Ge 1.$ Thus, it follows that $s_m:= \sup\{\sigma_a(f_n) : n \Ge m\} \Le \delta_a(f).$ To see the reverse inequality,
note first that $\sigma_a(f_n) < s_m + \epsilon$ for any $\epsilon > 0$ and all integers $n \Ge m.$ This implies that $f_n(s_m + \epsilon) < \infty$ for all integers $n \Ge m,$ and hence $f_n(s_m + \epsilon) < \infty$ for all integers $n \Ge 1$ as $\{f_n(\cdot)\}_{n=1}^\infty$ is an increasing sequence of functions. Therefore, $\delta_a(f) \Le s_m + \epsilon,$ and hence $\delta_a(f) \Le s_m$ as $\epsilon > 0$ is arbitrary. This completes the proof.
\end{proof}
\begin{proof}[\textbf{Proof of Theorem~\ref{arbitrary-large}}]
%We categorize the proof into three cases based on the range of the given positive number $r.$
(i) Let $r \Ge 0$ and consider the Dirichlet series $g$ defined as
\beqn
g(s) = \sum_{n=1}^\infty \frac{p_n^{r-1-s}}{1 - p_n^{-s}},
\eeqn
where $\{ p_n \}_{n=1}^{\infty}$ denotes the increasing enumeration of all prime numbers. Define $g_n$ as the subseries of $g$ consisting of terms supported on integers that can be factorized using only the first $n$ prime numbers. Then, $g_n$ takes the form
\beqn
g_n(s) = \sum_{j=1}^{n} \frac{p_j^{r-1-s}}{1 - p_j^{-s}}.
\eeqn
Since $\sigma_a(g_n) = 0$ for every integer $n \Ge 1$, Lemma~\ref{eventually-equal-lemma} implies that $\delta_a(g) = 0$.
Next, we determine $\sigma_a(g)$. To this end, observe that applying \cite[Eq.~(0.4)]{DGMS} yields
\beqn
g(\sigma) \Le \frac{1}{1 - 2^{-\sigma}} \sum_{n=1}^{\infty} p_n^{r-1-\sigma} < \infty, \quad \text{for } \sigma > r.
\eeqn
This establishes that $\sigma_a(g) \Le r$.
Furthermore,
\beqn
g(r) = \sum_{n=1}^{\infty} \frac{1}{p_n - p_n^{1-r}} \Ge \sum_{n=1}^{\infty} \frac{1}{p_n}.
\eeqn
By another application of \cite[Eq.~(0.4)]{DGMS}, it follows that $\sigma_a(g) = r$.
Now, consider a Dirichlet series $f$ given by
\beqn
f(s) &=& \zeta(s-r+1) (1 +g(s)).
\eeqn
Since both $\zeta(\cdot-r+1)$ and $g$ converge absolutely on $\mathbb H_r$, it follows from \cite[Theorem~4.3.1]{QQ} that $\sigma_a(f) \Le r.$ Let $(c_m)_{m=1}^\infty$ be the positive sequence determined uniquely by the relation
\beq
\label{coeff-i}
f(s) &=& \Big(\sum_{m=1}^\infty m^{r-1}m^{-s}\Big) \Big(1+ \frac{2^{r-1-s}}{1 - 2^{-s}} + \frac{3^{r-1-s}}{1 - 3^{-s}} + \frac{5^{r-1-s}}{1 - 5^{-s}} + \cdots \Big) \notag\\
&=& \sum_{m=1}^\infty c_m m^{-s}, \quad s \in \mathbb H_r.
\eeq
Then, $c_p = 2 p^{r-1}$ for any prime $p,$ so that
\beqn
f(r) \Ge \sum_{j=1}^\infty c_{p_j} p_j^{-r} = 2 \sum_{j=1}^\infty \frac{1}{p_j} = \infty,
\eeqn
which implies that $f$ diverges at $r$ and hence, $\sigma_a(f) = r.$
% It now remains to verify that $\delta_a(f) = 0$.
% To establish this, note that for any integer $n \Ge 1$ and $s \in \mathbb H_0,$
% \beqn
% \sum\limits_{\substack{j = 1 \\\tiny{\mbox{\bf gpf}}(j) \Le p_n }}^\infty c_j j^{-s} = \zeta_n(s-r+1) \Big(1+ \frac{2^{r-1-s}}{1 - 2^{-s}} + \frac{3^{r-1-s}}{1 - 3^{-s}} + \cdots + \frac{p_n^{r-1-s}}{1 - p_n^{-s}}\Big),
% \eeqn
% where $\zeta_n$ is given by \eqref{zeta-n}.
% This immediately gives that $\delta_a(f) \Le 0$. Furthermore, since $c_{2^m} \Ge 2^{r-1}$ for all integers $m \Ge 1$, the series
% \beqn
% \sum\limits_{\substack{j = 1 \\\tiny{\mbox{\bf gpf}}(j) \Le2}}^\infty c_j = \sum_{m =0}^\infty c_{2^m}~~\text{diverges}.
% \eeqn
% Therefore, we obtain that $\delta_a(f) = 0$, implying that $\sigma_a(f) - \delta_a(f) = r.$
%
It remains to verify that \(\delta_a(f) = 0\). To establish this, observe that for any integer $n \Ge 1$ and $s \in \mathbb H_0$, we have
\beqn
\sum\limits_{\substack{j = 1 \\ \tiny{\textbf{gpf}}(j) \Le p_n }}^\infty c_j j^{-s} = \zeta_n(s - r + 1) \left(1+ \frac{2^{r-1-s}}{1 - 2^{-s}} + \frac{3^{r-1-s}}{1 - 3^{-s}} + \cdots + \frac{p_n^{r-1-s}}{1 - p_n^{-s}}\right),
\eeqn
where \(\zeta_n\) is given by equation \eqref{zeta-n}.
From this, it follows immediately that $\delta_a(f) \Le 0$. Furthermore, since $c_{2^m} \Ge 2^{r-1}$ for all integers $m \Ge 1$, the series
\beqn
\sum\limits_{\substack{j = 1 \\ \tiny{\textbf{gpf}}(j) \Le 2}}^\infty c_j = \sum_{m =0}^\infty c_{2^m}
\eeqn
diverges.
Thus, we conclude that \(\delta_a(f) = 0\), which implies that $\sigma_a(f) - \delta_a(f) = r$.
\noindent \textbf{Another approach to form a Dirichlet series for $r=0$}:
To see a Dirichlet series satisfying $\sigma_a= \delta_a,$ let $g(s) = \sum\limits_{n=2}^\infty b_nn^{-s} \in \mathcal D$ be such that $b_n >0$ for all integers $n \Ge 2$ (this construction remains valid even if $\sigma_a(g)=-\infty$, meaning that $g$ is an entire Dirichlet series). Clearly, $g$ is a strictly decreasing function on $(\sigma_a(g), \infty)$ and $g(\sigma) \rar 0$ as $\sigma \rar \infty$. For $\alpha > 0,$ consider the function $f$ given by
\beqn
%\label{coeff-ii}
f(s): = \frac{1}{\alpha-g(s)}.
\eeqn
Assume that there exists $\rho \in (\sigma_a(g), \infty)$ such that $g(\rho) = \alpha.$
Then, we claim that
\beq \label{sigma=delta=rho}
\text{$f \in \mathcal D$ and $\sigma_a(f) = \delta_a(f) = \rho.$}
\eeq
Note first that such a real number $\rho$ is unique as $g$ is injective, and $\alpha > 0.$ Since $g(\rho)= \alpha$ and $g$ is decreasing, we have that $g(\rho + \epsilon) < \alpha$ for every $\epsilon > 0,$ and hence
\beqn
|\alpha-g(s)| \Ge \alpha-g(\rho + \epsilon) > 0, \quad s \in \mathbb H_{\rho+ \epsilon}.
\eeqn
Also, since $g$ is absolutely convergent at $\rho,$ by \cite[Theorem~4.4.3]{QQ}, $f \in \mathcal D$ with $\sigma_a(f) \Le \rho,$ so that write
\beq
\label{coeff-ii}
f(s) = \sum_{n=1}^\infty c_n n^{-s}.
\eeq
Furthermore, since $f$ blows up at $s = \rho,$ we get that $\sigma_a(f) = \rho.$ This also yields that $\delta_a(f) \Le \rho.$
It now remains to verify that $\delta_a(f) \Ge \rho.$
For every integer $n \Ge 1,$ let
\beqn
f_n(s): = \sum\limits_{\substack{j = 1 \\\tiny{\mbox{\bf gpf}}(j) \Le p_n }}^\infty c_j j^{-s} ~~\text{and}~~g_n(s): = \sum\limits_{\substack{j = 2 \\\tiny{\mbox{\bf gpf}}(j) \Le p_n }}^\infty b_j j^{-s}.
\eeqn
Then, for every $n \in \mathbb N,$ we have that $\sigma_a(f_n) \Le \sigma_a(f) = \rho$ and
\beq
\label{fn-case}
f_n(s) = \frac{1}{\alpha-g_n(s)}, \quad s \in \mathbb H_\rho.
\eeq
If $\sigma_a(f_m) = \rho$ for some $m \in \mathbb N,$ then it follows that $\sigma_a(f_n) \Ge \rho$ for all $n \Ge m,$ and hence by Lemma~\ref{eventually-equal-lemma}, $\delta_a(f) \Ge \rho,$ completing the claim. Otherwise, assume that $\sigma_a(f_n)<\rho$ for all $n \in \mathbb N.$ Let $x_0 < \sigma_a(g)$ be a real number very close to $\sigma_a(g).$ Since $g(x_0)$ is divergent, there exists a positive integer $n_0 \equiv n_0(x_0)$ such that
$g_n(x_0) > \alpha$ for all $n \Ge n_0.$ Note that $g \Ge g_n$ on $(\sigma_a(g), \infty)$, $g_n$ is decreasing and $g(\rho) = \alpha$. Thus, there exists a real number $\rho_n < \rho$ such that
$
g_n(\rho_n) = \alpha,~ n \Ge n_0.
$
This implies that $g_n(\rho_n + \epsilon) < \alpha$ for every $\epsilon > 0$ as $g_n$ is a decreasing function, and hence
\beqn
|\alpha-g_n(s)| \Ge \alpha-g_n(\rho_n +\epsilon) > 0, \quad s \in \mathbb H_{\rho_n+ \epsilon}.
\eeqn
Also, since $g_n$ is absolutely convergent at $\rho_n,$ we get that $\sigma_a(f_n) \Le \rho_n$ for all integers $n \Ge n_0.$
% \beqn
% \sigma_a(f_n) \Le \rho_n,\quad n \Ge n_0.
% \eeqn
Moreover, since $f_n$ blows up at $s = \rho_n$ (see \eqref{fn-case}),
\beqn
% \label{n-n0}
\text{$\sigma_a(f_n) = \rho_n$ for all integers $n \Ge n_0.$}
\eeqn
Further,
note that $\rho_n \Le \rho$ for all integers $n\Ge n_0$ and $\{\rho_n\}_{n=1}^\infty$ is an increasing sequence, which yields that $\rho_n \rar \rho_0 \Le \rho$ as $n \rar \infty.$ By using the fact that $g_n$ decreases, we get that
\beq\label{decreasing-prop}
g_n(\rho) \Le g_n(\rho_0) \Le g_n(\rho_n) =\alpha, \quad n \Ge n_0.
\eeq
Letting $n \rar \infty$ in \eqref{decreasing-prop} leads to $\lim_{n \rar \infty}g_n(\rho_0) = \alpha,$
which yields that $g$ also converges at $\rho_0$ with $g(\rho_0) = \alpha.$ Thus, since $g$ is injective over $(\sigma_a(g), \infty)$, it follows that $\rho_0= \rho,$ and hence $\rho_n$ converges to $\rho$ as $n \rar \infty.$ This combined with Lemma~\ref{eventually-equal-lemma} implies that $\delta_a(f) = \rho$ completing the proof of \eqref{sigma=delta=rho}. Therefore, we finally have that $\sigma_a(f) - \delta_a(f) = 0.$
(ii) Let $r \Ge 1.$ For any prime $p$ and $m \in \mathbb N,$ define $c_{p^m} = p^{r-1}$ and extend it multiplicatively for all integers $n \Ge 1.$
Consider $f(s) = \sum_{n=1}^\infty c_n n^{-s}.$ Since $c_n > 0$ for all $n \in \mathbb N,$
\beqn
\sum_{n=1}^\infty \frac{1}{p_n} = \sum_{n=1}^\infty c_{p_n} {p_n}^{-r} \Le f(r),
\eeqn
which together with the fact that $\sum_{n=1}^\infty \frac{1}{p_n}$ diverges (see, \cite[Eq.~(0.4)]{DGMS}) implies that $f$ also diverges at $s = r.$ Hence,
\beq \label{one-estimate}
\sigma_a(f) \Ge r.
\eeq
Moreover, since $r \Ge 1,$ we have that $c_{p^m} \Le p^{(r-1) m}$ for any prime $p$ and positive integer $m.$ This combined with multiplicativity of $\{c_n\}_{n \Ge 1}$ yields that $c_n \Le n^{r-1},~n \in \mathbb N,$ and hence for any $\sigma > r,$
\beqn
\sum_{n=1}^\infty c_n n^{-\sigma} \Le \sum_{n=1}^\infty n^{-(\sigma+1 - r)} < \infty
\eeqn
implying that $\sigma_a(f) \Le r.$ Thus, by \eqref{one-estimate}, $\sigma_a(f) = r.$
It now suffices to claim that $\delta_a(f) = 0.$ To see this, note first that for any integer $n \Ge 1,$
\beqn
% \label{delta-w}
\prod_{j=1}^n \Big(1+\frac{p_j^{r-1-s}}{1-p_j^{-s}}\Big)
\eeqn
is a finite product of Dirichlet series absolutely convergent on $\mathbb H_0.$
This combined with the multiplicativity of $\{c_n\}_{n \Ge 1}$ and the fundamental theorem of arithmetic yields that
\beqn
\sum\limits_{\substack{j = 1 \\\tiny{\mbox{\bf gpf}}(j) \Le p_n }}^\infty c_j j^{-s} = \prod_{j=1}^n \Big(1+\frac{c_{p_j} p_j^{-s}}{1-p_j^{-s}}\Big)
= \prod_{j=1}^n \Big(1+\frac{p_j^{r-1-s}}{1-p_j^{-s}}\Big)
\eeqn
converges on $\mathbb H_0$ implying that
$\delta_a(f) \Le 0.$ Also, since $\sum_{n=0}^\infty c_{2^n}$ is a divergent series, we obtain that $\delta_a(f) = 0.$ Finally, $\sigma_a(f) - \delta_a(f) = r.$
%Let $r \in (0, 1).$ Consider a Dirichlet series $g$ defined by
% \beqn
% g(s) = \sum_{n=1}^\infty \frac{p_n^{r-1-s}}{1-p_n^{-s}},
% \eeqn
% where $\{p_n\}_{n=1}^\infty$ be an increasing enumeration of all prime numbers. Let $g_n$ represent the subseries of $g$ consisting of terms supported on integers that can be factorized using only the first $n$ primes numbers. Then, $g_n$ will be of the form
% \beqn
% g_n(s) = \sum_{j=1}^n \frac{p_j^{r-1-s}}{1-p_j^{-s}},
% \eeqn
% so that $\sigma_a(g_n) = 0$ for every integer $n \Ge 1,$ and hence by Lemma~\ref{eventually-equal-lemma}, $\delta_a(g) = 0.$
% We now compute $\sigma_a(g).$ To see this, note that by an application of \cite[Eq.~(0.4)]{DGMS},
% \beqn
% g(\sigma) \Le \frac{1}{1-2^{-\sigma}}\sum_{n=1}^\infty p_n^{r-1-\sigma} < \infty, \quad \sigma > r,
% \eeqn
% and hence $\sigma_a(g) \Le r.$
% Moreover,
% \beqn
% g(r)= \sum_{n=1}^\infty \frac{1}{p_n - p_n^{1-r}} \Ge \sum_{n=1}^\infty \frac{1}{p_n}.
% \eeqn
% Again, by \cite[Eq.~(0.4)]{DGMS}, we finally obtain that $\sigma_a(g) = r.$
\end{proof}
\begin{remark}
% A careful analysis of the proof of Theorem~\ref{arbitrary-large}(ii) shows that even if $r \Ge 0$, then by using the same approach as in Theorem~\ref{arbitrary-large}(ii), $\delta_a(f) = 0$ and $\sigma_a(f) \Ge r$. But, to see the value of $\sigma_a(f)$, we crucially used the assumption that $r \Ge 1$. So, in the remaining case $r \in [0, 1)$, we do not know the exact value of $\sigma_a(f).$
A careful analysis of the proof of Theorem~\ref{arbitrary-large}(ii) reveals that even if $r \Ge 0$, applying the same approach as in Theorem~\ref{arbitrary-large}(ii) yields $\delta_a(f) = 0$ and $\sigma_a(f) \Ge r$. However, determining the exact value of $\sigma_a(f)$ relied crucially on the assumption that $r \Ge 1$. Consequently, for the remaining case where $r \in [0,1)$, we do not know the precise value of $\sigma_a(f)$.
\end{remark}
% We now present a family of examples that demonstrate the second approach in the proof~(ii) of Theorem~\ref{arbitrary-large}. A specific instance of this has previously appeared in \cite{Mc-1}.
We now present a family of examples illustrating the second approach employed in the proof of Theorem~\ref{arbitrary-large}(i). A particular case of this has previously been presented in \cite{Mc-1}.
\begin{example}\label{example-inte}
For any integer $m \Ge 1,$ let $\rho_m \in (1, \infty)$ be the unique real number satisfying $\zeta^m(\rho_m) = 2.$ By \cite[Equation~1.2.2]{Ti},
\beqn
f_m(s) = \frac{1}{2- \zeta^m(s)} = \sum_{n=1}^\infty d_m(n) n^{-s}, \quad s, u \in \mathbb H_{\rho_m},
\eeqn
where, for every integer $n \Ge 2,$ $d_m(n)$ denotes the number of ways of expressing $n$ as a product of $m$ factors $($the order also matters$).$ Therefore, by Theorem~\ref{arbitrary-large}(i), $\sigma_a(f_m) = \delta_a(f_m) = \rho_m.$
% For $n \Ge 2,$ let $H(n)$ denote the total number of ways to factor a natural number $n$ into an ordered product of integers greater than $1,$ and let $H(1) =1$ $($refer to \cite{CLM} for properties of $\{H(n)\}_{n=1}^\infty).$ By \cite[Eq.~1.2.15]{Ti},
% \beqn
% f(s) = \frac{1}{2 - \zeta(s)} = \sum_{n=1}^\infty H(n) n^{-s}, \quad s \in \mathbb H_\rho,
% \eeqn
% where $\rho \approx 1.72864724$ be the unique real number such that $\zeta(\rho) = 2.$
% Therefore, by Theorem~\ref{arbitrary-large}$,$ $\sigma_a(f) = \delta_a(f) = \rho.$
\end{example}
\section{An application of Theorem~\ref{arbitrary-large}}
For $\{a_n\}_{n=1}^\infty \subseteq (0, \infty),$ let $f(s) = \sum_{n=1}^\infty a_n n^{-s}$ with $\sigma_a(f) < \infty.$ With this sequence, we now associate a positive semi-definite Dirichlet series kernel $\kappa$ given by
\beq
\label{kappa-d}
\kappa(s, u) = \sum_{n=1}^\infty a_n n^{-s -\overline{u}},\quad s, u \in \mathbb H_{\frac{\sigma_a(f)}{2}}.
\eeq
Clearly, $\kappa$ converges absolutely on $\mathbb H_{\frac{\sigma_a(f)}{2}} \times \mathbb H_{\frac{\sigma_a(f)}{2}},$ and does not extend on any larger domain.
By the Moore's theorem \cite[Theorem~2.14]{P-R}, there exists an unique reproducing kernel Hilbert space $\mathscr H_\kappa$ associated with $\kappa,$ which is given by
\beqn
\mathscr H_\kappa = \Big\{\sum_{n=1}^\infty b_n n^{-s} ~~:~~ \sum_{n=1}^\infty \frac{|b_n|^2}{a_n} < \infty\Big\}.
\eeqn
%We use the convention that $\mf 1 \notin \mathscr H_\mf d$ whenever $d_1 = 0.$ In this case, constant functions do not belong to $\mathscr H_\mf d.$
By Cauchy-Schwarz inequality, every function in $\mathscr H_\kappa$ converges absolutely on $\mathbb H_{\frac{\sigma_a(f)}{2}},$ and hence \beqn
\sup\{\sigma_a(h) : h \in \mathscr H_\kappa\} \Le \frac{\sigma_a(f)}{2}.
\eeqn
% For $\mathcal S\subset \mathcal D$, let
% \beq \label{abs-space}
% \Sigma(\mathcal S): = \sup\{\sigma_a(f) : f \in \mathcal S\}
% \eeq
% so that the common domain of analyticity is $\mathbb H_{\Sigma(\mathcal S)}.$
Furthermore, $\kappa(\cdot, \frac{\sigma_a(f)}{2}+\epsilon) \in \mathscr H_\kappa$ for any $\epsilon > 0,$ which admits a pole at $s = \frac{\sigma_a(f)}{2}-\epsilon.$ Thus,
\beq
\label{abs_H_d}
%\sigma_{c}(\mathscr H_\mf d) =
\sup\{\sigma_a(h) : h \in \mathscr H_\kappa\} = \frac{\sigma_a(f)}{2}
\eeq
implying that the common domain of analyticity of $\mathscr H_\kappa$ (the largest right half-plane where all series from $\mathscr H_\kappa$ converges) is $\mathbb H_{\frac{\sigma_a(f)}{2}}.$
The following is a well known fact related to the multipliers of $\mathscr H_\kappa$ (see, \cite[Theorem~3.1]{St}).
\begin{proposition}\label{stetler_result_repeat}
Let $\kappa$ be the kernel defined by \eqref{kappa-d}. Then, the multiplier algebra
$\mathcal M(\mathscr H_\kappa)$ of $\mathscr H_\kappa$ is contractively included in $H^\infty(\mathbb H_{\frac{\delta_a(f)}{2}}) \cap \mathcal D.$
\end{proposition}
As a consequence of Theorem~\ref{arbitrary-large} and Proposition~\ref{stetler_result_repeat}, we gain some insights into the structure of the multiplier algebra.
\begin{corollary}
For any non-negative real number $r,$ let $f$ be the Dirichlet series formed in the proof of Theorem~\ref{arbitrary-large}. Let $\kappa: \mathbb H_r \times \mathbb H_r \rar \mathbb C$ be a kernel given by
\beqn
\kappa(s, u) = f(s+\overline{u}) = \sum_{m=1}^\infty c_m m^{-s-\overline{u}}, \quad s, u \in \mathbb H_{\frac{r}{2}}.
\eeqn
%where $f$ is the Dirichlet series formed in the proof of Theorem~\ref{arbitrary-large}.
\begin{itemize}
\item [$(i)$] If $\{c_m\}_{m \Ge 1}$ is determined either by \eqref{coeff-i} $($with $r \in [0,1))$ or \eqref{coeff-ii}, then $\mathcal M(\mathscr H_\kappa) \subsetneq H^\infty(\mathbb H_{0}) \cap \mathcal D.$
\item [$(ii)$] If $\{c_m\}_{m \Ge 1}$ is the multiplicative sequence formed as in proof (ii) of Theorem~\ref{arbitrary-large}, then
$\mathcal M(\mathscr H_\kappa) = H^\infty(\mathbb H_0) \cap \mathcal D$ (isometrically).
\end{itemize}
\end{corollary}
\begin{proof}
By Theorem~\ref{arbitrary-large}, we have that
\beq
\label{sigma-delta}
\text{ $\sigma_a(f) = r$ and $\delta_a(f) = 0$.}
\eeq
(i) By Proposition~\ref{stetler_result_repeat}, we already know that $\mathcal M(\mathscr H_\kappa) \subseteq H^\infty(\mathbb H_{0}) \cap \mathcal D.$ To see the strict containment, since $r = \sigma_a(f) \in [0,1)$ and $\mathcal M(\mathscr H_\kappa) \subseteq \mathscr H_\kappa$, it follows from \eqref{abs_H_d} and \eqref{sigma-delta} that
\beqn
\sup\{\sigma_a(h) : h \in \mathcal M(\mathscr H_\kappa)\}
&\Le&
\sup\{\sigma_a(h) : h \in \mathscr H_\kappa\}\\
&=& \frac{r}{{2}}\\
&<& \frac{1}{2} = \sup\{\sigma_a(h) : h \in H^\infty(\mathbb H_{0}) \cap \mathcal D\},
\eeqn
where the last equality holds by an application of \cite[Theorem~4.1]{DGMS}.
Therefore, $\mathcal M(\mathscr H_\kappa) \subsetneq H^\infty(\mathbb H_{0}) \cap \mathcal D.$
(ii) By assumption, $c_{p^m} = p^{r-1}$ which extends multiplicatively for all integers $n \Ge 1.$ Then, \cite[Corollary~4.2]{St} (also, see \cite[Theorem~4.1]{Sa}) combined with the condition \eqref{sigma-delta} yields the desired conclusion.
\end{proof}
%Here is a special case of the part~(ii) of the proof of Theorem~\ref{arbitrary-large}, which appeared also in \cite{Mc-1}.
%We now provide a family of examples illustrating part (ii) of the proof of Theorem~\ref{arbitrary-large}. A particular case of it has appeared in \cite{Mc-1}.
% \section{Proof of Theorem~\ref{l1-subspace} and an application}\label{S3}
% The following lemma characterizes all sequences $\mf d$ such that Dirichlet polynomials are included in $\mathcal M(\mathscr H_\mf d),$ which shall be required to prove Theorem~\ref{l1-subspace}.
%
%
% \begin{lemma} \label{monomials-characterization}
% Let $\mf d=\{d_n\}_{n=1}^\infty$ be an admissible sequence such that $d_1 = 0.$ Then, $n^{-s} \in \mathcal M(\mathscr H_\mf d)$
% if and only if $\{\frac{d_m}{d_{mn}}\}_{m=2}^\infty$ is a bounded
% sequence. In this case,
% \beq \label{iso-prime-leval}
% \|n^{-s}\|^2_{\mathcal M(\mathscr H_\mf d)} = \sup_{m \Ge 2} \frac{d_m}{d_{mn}}, \quad n \Ge 1.
% \eeq
% \end{lemma}
% \begin{proof}
% Assume that $n^{-s} \in \mathcal M(\mathscr H_\mf d).$ Then,
% \beq
% \label{lower-norm}
% \frac{d_{m}}{d_{mn}} = \frac{\| (mn)^{-s}\|^2}{\|m^{-s}\|^2} \Le \|n^{-s}\|^2_{\mathcal M(\mathscr H_\mf d)} < \infty, \quad m \Ge 2
% \eeq
% implying that $\{\frac{d_m}{d_{mn}}\}_{m=2}^\infty$ is bounded.
% Conversely, assume that $\{\frac{d_m}{d_{mn}}\}_{m=2}^\infty$ is bounded. Then, for any integer $n \Ge 1$ and $f(s) = \sum\limits_{m=2}^\infty a_m m^{-s} \in \mathscr H_\mf d,$ we have that
% $n^{-s}f \in \mathcal D$ with
% $n^{-s}f(s) = \sum\limits_{m=2}^\infty a_m (mn)^{-s}.$
% By the assumption, since
% \beqn
% \sum\limits_{m=2}^\infty \frac{|a_m|^2}{d_{mn}} \Le \Big(\sup_{m \Ge 2}\frac{d_m}{d_{mn}}\Big) \|f\|_{\mathscr H_\mf d}^2 < \infty,
% %\overset{\eqref{nece-suf-con}}\Le p^{-\delta_a(D)} \|f\|_{\mathscr H_\mf d}^2,
% \eeqn
% it follows that $n^{-s}f \in \mathscr H_\mf d,$ and hence $n^{-s} \in \mathcal M(\mathscr H_\mf d)$ with
% \beq
% \label{upper-norm}
% \|n^{-s}\|_{\mathcal M(\mathscr H_\mf d)} \Le \Big(\sup_{m \Ge 2}\frac{d_m}{d_{mn}}\Big)^{1/2},\quad n \Ge 1.
% \eeq
% completing the equivalence. Finally, the desired equality \eqref{iso-prime-leval} follows, by using \eqref{lower-norm} and \eqref{upper-norm}.
% \end{proof}
%If $d_1 > 0,$ then $\mathcal D[s] \subseteq \mathcal M(\mathscr H_\mf d)$
%if and only if $\{\frac{d_m}{d_{mn}}\}_{m=1}^\infty$ is bounded
%for every integer $n \Ge 1.$ This can also be proven using the same approach as in the proof of the previous lemma.
%% \medskip \noindent
%% We are now ready to prove Theorem~\ref{l1-subspace}.
%
% \begin{proof}[\textbf{Proof of Theorem}~\ref{l1-subspace}]
% Suppose that $d_1 = 0$ and $\{\frac{d_m}{d_{mn}}\}_{m=2}^\infty$ is a bounded sequence for every integer $n \Ge 1.$ Then, by Lemma~\ref{monomials-characterization}, the set $\mathcal D[s]$ of Dirichlet polynomials is contained in $\mathcal M(\mathscr H_\mf d).$
% Let $\varphi(s) = \sum_{n=1}^\infty a_n n^{-s} \in \ell^1(\mf w)$ and let $\varphi_n(s) = \sum_{j=1}^n a_j j^{-s},~ n \Ge 1.$ Then, $\{\varphi_n\}_{n=1}^\infty$ is a cauchy sequence in $\mathcal M(\mathscr H_\mf d)$ as $\|\varphi_n - \varphi_m\|_{\mathcal M(\mathscr H_\mf d)} \Le \sum_{j=m+1}^n |a_j|w_j$ for positive integers $m \Le n.$ Thus, by Proposition~\ref{stetler_result_repeat}, $\{\varphi_n\}_{n=1}^\infty$ is cauchy in $H^\infty(\mathbb H_{\frac{\delta_a(D)}{2}}) \cap \mathcal D,$ and hence
% \beq
% \label{uniform-conv}
% \text{$\varphi_n$ converges uniformly to $\varphi$ on $\mathbb H_{\frac{\delta_a(D)}{2}}.$}
% \eeq
% Furthermore, we also have that $\{M_{\varphi_n}\}_{n=1}^\infty$ is cauchy in $\mathcal B(\mathscr H_\mf d),$ the Banach space of bounded linear operators on $\mathscr H_\mf d.$ This implies that $M_{\varphi_n}$ converges to some $T \in \mathcal B(\mathscr H_\mf d)$ and for $g \in \mathscr H_\mf d,$
% \beqn
% T(g)(s) = \lim_{n \rar \infty} \varphi_n(s)g(s) \overset{\eqref{uniform-conv}}= \varphi(s)g(s) = M_\varphi(g)(s), \quad s \in \mathbb H_{\frac{\sigma_a(D)}{2}}.
% \eeqn
% Therefore, $M_\varphi = T \in \mathcal B(\mathscr H_\mf d),$ and hence, $\varphi \in \mathcal M(\mathscr H_\mf d),$ showing that $\ell^1(\mf w) \subseteq M(\mathscr H_\mf d).$ To see that the inclusion is a contraction, note that since $\varphi_n$ converges to $\varphi$ in $\mathcal M(\mathscr H_\mf d),$ we get
% \beqn
% \|\varphi \|_{\mathcal M(\mathscr H_\mf d)} = \lim_{n \rar \infty}\|\varphi_n\|_{\mathcal M(\mathscr H_\mf d)} \Le \sum_{n=1}^\infty |a_n| w_n.
% \eeqn
% This completes the proof.
%\end{proof}
%
% \medskip \noindent
% For $r \in \mathbb R,$ let $\mathcal A_r$ be defined by
% \beqn
% \mathcal A_r := \Big\{ \sum_{n=1}^\infty a_n n^{-s} : \sum_{n=1}^\infty |a_n| n^{-r} < \infty \Big\}.
% \eeqn
% Clearly, $\Sigma(\mathcal A_r) \Le r.$
% Moreover, for every $\epsilon > 0,$ let $f_\epsilon(s) = \zeta(s+\epsilon-r+1)$. Then, $f_\epsilon \in \mathcal
% A_r$ and $\sigma_a(f_\epsilon) = r-\epsilon,$ which implies that
% $
% %\label{abs_A_r}
% \Sigma(\mathcal A_r) = r.
% $
% %It follows that the common domain of analyticity of $\mathcal A_r$ is the right half-plane $\mathbb H_r.$
% The set $\mathcal A_r$ becomes a Banach space when endowed with the norm
% \beqn
% \Big\|\sum_{n=1}^\infty a_n n^{-s}\Big\|_{\mathcal A_r} : = \sum_{n=1}^\infty |a_n| n^{-r}.
% \eeqn
% The space $\mathcal A_0$ is known as Wiener-Dirichlet algebra (refer to \cite{BFLQ} for more details).
%
% %a translated space of Wiener-Dirichlet algebra is included in $\mathcal M(\mathscr H_\mf d).$
%
% \begin{corollary}
%% \hookrightarrow H^\infty(\mathbb H_{\frac{\delta_a(D)}{2}}) \cap \mathcal D.
%% $
% \end{corollary}
% \begin{proof}
% Since \eqref{nece-suf-con} holds, by Lemma~\ref{monomials-characterization}, $p^{-s} \in \mathcal M(\mathscr H_\mf d)$ for every prime $p,$ and hence the set $\mathcal D[s]$ of Dirichlet polynomials is contained in $\mathcal M(\mathscr H_\mf d)$ as $\mathcal M(\mathscr H_\mf d)$ is an algebra. Thus, by Lemma~\ref{monomials-characterization}, $\{\frac{d_m}{d_{mn}}\}_{m =2}^\infty$ is a bounded sequence for all integers $n \Ge 1,$ and set $x_n: = \big(\sup_{m \Ge 2} \frac{d_m}{d_{mn}}\big)^{1/2},~ n \Ge 1.$ Then,
% by Lemma~\ref{monomials-characterization}, for any positive integer $n = \prod_{j=1}^k p_{i_j}^{r_j},$
% \beqn
% \|n^{-s}\|_{\mathcal M(\mathscr H_\mf d)} = x_n \Le \prod_{j=1}^k x_{p_{i_j}^{r_j}} \Le \prod_{j=1}^k x_{p_{i_j}}^{r_j}
% \overset{\eqref{nece-suf-con}}\Le \prod\limits_{j=1}^k p_{i_j}^{-\frac{\delta_a(D) r_j}{2}} = \|n^{-s}\|_{\infty, \mathbb H_{\frac{\delta_a(D)}{2}}}.
% \eeqn
% This combined with Proposition~\ref{stetler_result_repeat} yields that
% $x_n = \|n^{-s} \|_{\mathcal M(\mathscr H_\mf d)} = n^{-\frac{\delta_a(D)}{2}}$ for all integers $n \Ge 1.$
% Therefore, by Theorem~\ref{l1-subspace},
% \beqn
% \mathcal A_{\frac{\delta_a(D)}{2}}= \Big\{\sum_{n=1}^\infty a_n n^{-s} : \sum_{n=1}^\infty |a_n| n^{-\frac{\delta_a(D)}{2}} < \infty \Big\} \subseteq \mathcal M(\mathscr H_\mf d),
% \eeqn
% completing the proof.
% %which together with by gives that $ \subseteq
% \end{proof}
%
%It is worth noting that, under the assumptions of corollary above, it follows from \cite[Corollary~9.2]{DGMS} that $H^\infty(\mathbb H_{\frac{\delta_a(D)-1}{2}}) \cap \mathcal D \subseteq \mathcal A_{\frac{\delta_a(D)}{2}},$ and hence by Proposition~\ref{stetler_result_repeat} and Corollary~\ref{l1-subsapce-cor},
%\beqn
%H^\infty(\mathbb H_{\frac{\delta_a(D)-1}{2}}) \cap \mathcal D \subseteq \mathcal M(\mathscr H_\mf d) \subseteq H^\infty(\mathbb H_{\frac{\delta_a(D)}{2}}) \cap \mathcal D,
%\eeqn
%so that $\mathcal M(\mathscr H_\mf d)$ lies between two $H^\infty$-spaces of Dirichlet series.
\vskip.4cm
\textit{Acknowledgment:}
The author gratefully acknowledges the support provided by the institute post-doctoral fellowship of IIT Bombay. The author expresses gratitude to O. F. Brevig and H. Ahmed for helpful discussions regarding Question~\ref{Q-1}.
%\vskip.3cm
\begin{thebibliography}{10}
% \bibitem{BCD} J. Bueno-Contreras, G. P. Curbera, O. Delgado,
% The Cesàro space of Dirichlet series and its multiplier algebra. {\it J. Math. Anal. Appl.}, {\bf 475} (2019), no. 2, 1448–1471.
%
% \bibitem{BFLQ} F. Bayart, C. Finet, D. Li, H. Queffélec, Composition operators on the Wiener-Dirichlet algebra, {\it J. Operator Theory} {\bf 60} (2008), 45–70.
%
%
% \bibitem{CLM} B. Chor, P. Lemke, Z. Mador, On the number of ordered factorizations of natural numbers, {\it Discrete Math.} {\bf 214} (2000), no. 1-3, 123–133.
%
\bibitem{DGMS} A. Defant, D. Garcia, M. Maestre, P. Sevilla-Peris, {\it Dirichlet series and holomorphic functions in high dimensions}, {\bf 37}, Cambridge University Press, Cambridge, 2019, xxvii+680 pp.
% \bibitem{HR} G. H. Hardy and M. Riesz, {\it The General Theory of Dirichlet Series}, Cambridge Tracts
% in Mathematics and Mathematical Physics, 18. Stechert-Hafner, Inc., New York, 1964,
% vii+78 pp.
%
% \bibitem{Mash} J. Mashreghi, An inequality for Cauchy products, {\it Proceedings of the American Mathematical Society}, {\bf 145}(2017), 3013-3016.
%
\bibitem{Mc-1} J. E. McCarthy, Hilbert spaces of Dirichlet series and their multipliers, {\it Trans. Amer. Math. Soc.} {\bf 356} (2004), 881-893.
\bibitem{P-R} V. Paulsen, M. Raghupathi, {\it An introduction to the theory of reproducing kernel Hilbert spaces}, Cambridge Studies in Advanced Mathematics, 152. Cambridge University Press, Cambridge, 2016. x+182 pp.
%
\bibitem{QQ} H. Queffelec, M. Queffelec, {\it Diophantine approximation and Dirichlet series}. Second edition. Texts and Readings in Mathematics, 80. Hindustan Book Agency, New Delhi; Springer, Singapore, 2020. 287 pp.
\bibitem{Sa} C. K. Sahu, Multipliers of the Hilbert spaces of Dirichlet series,
{\it New York J. Math.} {\bf 29}, (2023), 323–334.
%
\bibitem{St} E. Stetler, {\it Multiplication operators on Hilbert spaces of Dirichlet series}. Thesis (Ph.D.) - University of Florida, 2014. ProQuest LLC, Ann Arbor, MI, 102 pp.
%
\bibitem{Ti} E. C. Titchmarsh, {\it The theory of the Riemann zeta-function}, Second edition, Oxford University Press, New York, 1986. x+412 pp.
\end{thebibliography}
\end{document}