Constructions of mutually unbiased entangled bases
Extracted source text · arXiv 1911.08761
\documentclass[pra,aps,floatfix,amsmath,superscriptaddress,onecolumn]{revtex4}
\usepackage{amssymb}
\usepackage{graphicx}
\usepackage{graphics}
\usepackage{amsmath}
\usepackage{amsthm}
\usepackage{color}
\usepackage{dsfont}
\usepackage[colorlinks,citecolor=blue,linkcolor=blue]{hyperref}
\usepackage{enumerate}
\usepackage{mathtools}
\usepackage[justification=centering]{caption}
\def\ra{\rangle}
\def\la{\langle}
\def\bb{\mathbb}
\newtheorem{theorem}{Theorem}
\newtheorem{conjecture}{Conjecture}
\newtheorem{lemma}[theorem]{Lemma}
\newtheorem{cor}[theorem]{Corollary}
\newtheorem{eg}[theorem]{Example}
\newtheorem{pro}[theorem]{Proposition}
\newcommand{\bea}{\begin{eqnarray}}
\newcommand{\eea}{\end{eqnarray}}
\newcommand{\be}{\begin{equation}}
\newcommand{\ee}{\end{equation}}
\newcommand{\ba}{\begin{equation}\begin{aligned}}
\newcommand{\ea}{\end{aligned}\end{equation}}
\newcommand{\rank}{\text{Rank}}
\newtheorem{definition}{Definition}
\theoremstyle{remark}
\newtheorem{remark}{Remark}
%\newcommand{\tripartite}{\text{tripartite}}
\newtheorem{example}{Example}
\def\be{\begin{equation}}
\def\ee{\end{equation}}
%\def\ba{\begin{align}}
%\def\ea{\end{align}}
\newcommand{\mU}{\mathcal{U}}
\newcommand{\mA}{\mathcal{A}}
\newcommand{\mF}{\mathcal{F}}
\newcommand{\C}{{\mathbb{C}}}
\newcommand{\A}{{\mathbb{A}}}
\newcommand{\mI}{\mathcal{I}}
\newcommand{\mH}{\mathcal{H}}
\newcommand{\mrH}{\mathrm{H}}
\newcommand{\mM}{\mathcal{M}}
\newcommand{\mN}{\mathcal{N}}
\newcommand{\mT}{\mathcal{T}}
\newcommand{\mW}{\mathcal{W}}
\newcommand{\mK}{\mathcal{K}}
\newcommand{\mB}{\mathcal{B}}
\newcommand{\bv}{\mathbf{v}}
\newcommand{\bc}{\mathbf{c}}
\newcommand{\bu}{\mathbf{u}}
\newcommand{\spa}{\text{span}}
\newcommand{\mS}{\mathcal{S}}
\newcommand{\lr}{\rangle\langle}
\newcommand{\tr}{{\rm Tr}}
\newcommand{\mc}[1]{\mathcal{#1}}
\newcommand{\mbf}[1]{\mathbf{#1}}
\newcommand{\mbb}[1]{\mathbb{#1}}
\newcommand{\mrm}[1]{\mathrm{#1}}
\newcommand{\cH}{\mathcal{H}}
\newcommand{\bra}[1]{\langle #1|}
\newcommand{\ket}[1]{|#1\rangle}
\newcommand{\braket}[2]{\langle{#1}|{#2}\rangle}
%inner product
\newcommand{\ip}[2]{\langle #1|#2\rangle}
%outer productt
\newcommand{\mbN}{\mathbb{N}}
\newcommand{\mbR}{\mathbb{R}}
%%%%%%%%%%% Begin DJ latex definitions %%%%%%%%
\def\>{\rangle}
\def\<{\langle}
\def\diag{ \mathrm{diag}}
\def\supp{ \mathrm{supp}}
\begin{document}
%\preprint{APS/123-QED}
\title{Constructions of mutually unbiased entangled bases}% Force line breaks with \\
\author{Fei Shi}
\email{shifei@mail.ustc.edu.cn}
\affiliation{School of Cyber Security,
University of Science and Technology of China, Hefei, 230026, People's Republic of China.}
\author{Xiande Zhang}
\email{drzhangx@ustc.edu.cn}
\affiliation{School of Mathematical Sciences,
University of Science and Technology of China, Hefei, 230026, People's Republic of China.}
\author{Lin Chen}
\email{linchen@buaa.edu.cn}
\affiliation{School of Mathematical Sciences, Beihang University, Beijing 100191, China}
\affiliation{International Research Institute for Multidisciplinary Science, Beihang University, Beijing 100191, China}
\begin{abstract}
We construct two mutually unbiased bases by maximally entangled states (MUMEB$s$) in $\C^{2}\otimes \C^{3}$. This is the first example of MUMEB$s$ in $\C^{d}\otimes \C^{d'}$ when $d\nmid d'$, namely $d'$ is not divisible by $d$. We show that they cannot be extended to four MUBs in $\C^6$. We propose a recursive construction of mutually unbiased bases formed by special entangled states with a fixed Schmidt number $k$ (MUSEB$k$s). It shows that
$\min \{t_{1},t_{2}\}$ MUSEB$k_{1}k_{2}$s in $\C^{pd}\otimes \C^{qd'}$ can be constructed from $t_{1}$ MUSEB$k_{1}$s in $\C^{d}\otimes \C^{d'}$ and $t_{2}$ MUSEB$k_{2}$s in $\C^{p}\otimes \C^{q}$ for any $d,d',p,q$. Further, we show that three MUMEB$s$ exist in $\C^{d}\otimes \C^{d'}$ for any $d,d'$ with $d\mid d'$, and two MUMEB$s$ exist in $\C^{d}\otimes \C^{d'}$ for infinitely many $d,d'$ with $d\nmid d'$.
% t gives a construction of MUSEB$k$s in some systems. We also give a lower bound of $M_{d}(d,qd)$ and $M_{d}(pd,qd)$ for general $d\geq 2$ and $p,q\geq 1$. In addition, we construct two MUMEB$s$ in $\C^{2}\otimes \C^{3}$, which also can be used to construct MUMEBs in some systems $\C^{d}\otimes \C^{d'}$ where $d\nmid d'$. Thus our results promote and cover all the previous related work.
\end{abstract}
%\pacs{03.67.Mn, 03.65.Ud.}% PACS, the Physics and Astronomy
% Classification Scheme.
%\keywords{mutually unbiased bases, maximally entangled states, special entangled states}%Use showkeys class option if keyword
%display desired
\maketitle
\vspace{-0.5cm}
~~~~~~~~~~\indent{\it Keywords}: mutually unbiased bases, maximally entangled states, special entangled states
%\end{CJK*}
%\tableofcontents
\section{Introduction}\label{intro}
The notion of mutually unbiased bases (MUBs) became an essential feature of quantum
mechanics in 1960 in the works of Schwinger \cite{Schwinger}. In his work, Schwinger realized that no
information can be retrieved when a quantum state which is prepared in a basis
state is measured with respect to the basis mutually unbiased with the prepared one.
This observation has a striking application in the well-known BB84 quantum key distribution (QKD) protocol \cite{Bennett}.
Estimating an unknown physical transformation of a state, which is described by a unitary operator, was studied in \cite{Acin}. It is a more challenging problem compared to state estimation because it requires optimal measurements of states both post and prior to transformation. Such estimation has an immediate application in the alignment of reference frames using quantum spins \cite{Chiribella2004}, and in the field of quantum cryptography but with a lower security threshold \cite{Lucamarini,Chiribella2008}. To improve the security, a notion of ``mutually unbiased unitary-operator bases" (MUUBs) was put forward in \cite{Scott} for unitary operators acting on a $d$-dimensional Hilbert space as a set of pairwise \emph{mutually unbiased} unitary operator bases. The concept MUUBs was shown to be equivalent to MUBs formed by maximally entangled states (MUMEBs) \cite{Shaari}. In the same paper, Shaari et al. \cite{Shaari} generalized the original MUUBs for unitary operators acting on \emph{subspaces} of a Hilbert space, and considered the distinguishability of unitaries selected from a set of MUUBs and its use in a QKD setup.
In this paper, we consider another generalization of MUUBs. Motivated by the equivalence between MUUBs in $\C^{d}$ and MUBs formed by maximally entangled states (MUMEBs) in $\C^{d}\otimes \C^{d}$, we study mutually unbiased bases formed by special entangled states with a fixed Schmidt number $k$ (MUSEB$k$s) in a general bipartite state space $\C^{d}\otimes \C^{d'}$ $(1\leq k\leq d,d')$,
Let $M_{k}(d,d')$ denote the maximum size of MUSEB$k$s in $\C^{d}\otimes \C^{d'}$, i.e., the maximum number of mutually unbiased bases formed by special entangled states with Schmidt number $k$. The central problem here is to determine $M_{k}(d,d')$ in a given bipartite state space and provide bounds for this value.
There have been results in \cite{Tao,Zhang,Liu,Xu,Xu1,Cheng,Han,Scott,Shaari} in which MUMEBs were known only when $d\mid d'$, namely $d'$ is divisible by $d$. We give the first example of MUMEBs when $d\nmid d'$, and propose a simple construction of MUSEB$k$s for general systems via matrix spaces. Consequently, we improve all previous results on MUSEB$k$s (MUMEBs). See Table~\ref{all} for a summary and comparison with known results.
The rest of this paper is organized as follows. In Sec.~\ref{Def}, we introduce some related notations and terminologies, and give a one-to-one correspondence between MUSEB$k$s in $\C^{d}\otimes \C^{d'}$ and MUSEB$k$s in $\mathcal{M}_{d\times d'}$. Sec.~\ref{sec:first} provides a construction of two MUMEBs in $\C^{2}\otimes \C^{3}$, which is the first MUMEBs when $d\nmid d'$. In Sec.~\ref{sec:appli}, we show that the two MUMEBs in $\C^{2}\otimes \C^{3}$ constructed in Sec.~\ref{sec:first} cannot be extended to 4 MUBs in $\C^6$. In Sec.~\ref{Mks}, we establish a construction of
$\min \{t_{1},t_{2}\}$ MUSEB$k_{1}k_{2}$s in $\C^{pd}\otimes \C^{qd'}$ from $t_{1}$ MUSEB$k_{1}$s in $\C^{d}\otimes \C^{d'}$ and $t_{2}$ MUSEB$k_{2}$s in $\C^{p}\otimes \C^{q}$, which yields several new lower bounds in Sec.~\ref{Mbs}. We give the main conclusion and some open problems in Sec.~\ref{conpen}.
\begin{widetext}
\begin{table}[htbp]
\caption{Results about MUSEB$k$s in $\C^{d}\otimes \C^{d'}$}($d,q$ have a factorization of prime power, $p_1^{a_1}\leq\cdots\leq p_s^{a_s}$,$(p_1')^{a_1'}\leq\cdots\leq (p_s')^{a_s'}$)\label{all}
\centering
\renewcommand\tabcolsep{20.0pt}
\begin{tabular}{ccc}
\hline\noalign{\smallskip}
Conditions &Bounds & References\\
\noalign{\smallskip}\hline\noalign{\smallskip}
$d\geq2$ &$M_d(d,d)\leq d^2-1$ &\cite{Scott}\\
$d=2,3,5,7,11$ &$M_d(d,d)=d^2-1$ &\cite{Scott} \\
$d=p_{1}^{a_{1}}\cdots p_{s}^{a_{s}}$, $p_1^{a_1}\geq 3$ &$M_d(d,d)\geq p_1^{a_1}-1$ &\cite{Liu}\\
$d\geq 3$, prime power &$M_d(d,d)\geq 2(d-1)$ &\cite{Xu}\\
$d=p_{1}^{a_{1}}\cdots p_{s}^{a_{s}}$, odd &$ M_d(d,d)\geq 2(p_1^{a_1}-1)$ &\cite{Cheng}\\
$d$, odd prime power &$M_d(d,d)\geq (d^2-1)/2$ &\cite{Xu1}\\
$d=2^{t}p_{1}^{a_{1}}\cdots p_{s}^{a_{s}}$, $t=0$ &$M_d(d,d)\geq(p_1^{2a_1}-1)/2$ &This paper\\
$d=2^{t}p_{1}^{a_{1}}\cdots p_{s}^{a_{s}}$, $t=1$ &$M_d(d,d)\geq3$ &This paper\\
$d=2^{t}p_{1}^{a_{1}}\cdots p_{s}^{a_{s}}$, $t\geq 2$ &$M_d(d,d)\geq \min\{2(2^t-1),(p_{1}^{2a_{1}}-1)/2\}$ &This paper\\\hline
$M_d(d,d')\geq 2, \ell\geq 1$ &$M_d(d,2^ld')\geq 2$ &\cite{Zhang}\\
$q$ and $ d$, both prime powers &$M_{d}(d,qd)\geq \min\{q,M_{d}(d,d)\}$ &\cite{Xu}\\
$q=(p_{1}')^{a_{1}'}\cdots (p_{s}')^{a_{t}'}$, $d$ odd &$M_{d}(d,qd)\geq \min\{(p_{1}')^{a_{1}'}+1,M_{d}(d,d)\}$ &\cite{Cheng} \\
$d, q\geq 2$ &$M_{d}(d,qd)\geq \min\{N(q),M_{d}(d,d)\}\geq 3$ &This paper\\
$d,q\geq 1$ &$M_{2d}(2d,3qd)\geq2$. &This paper\\\hline
$k=2^l$ &$M_2(3,4k)\geq 2$ &\cite{Han}\\
$k\geq 1$ &$M_2(3,2k)\geq 2$ &This paper \\
$d, p\geq 2$, $q\geq 1$ &$M_{d}(pd,qd)\geq \min\{N(p),M_{d}(d,qd)\}\geq 3$ &This paper \\
\noalign{\smallskip}\hline
\end{tabular}\label{Res}
\end{table}
\end{widetext}
% \begin{enumerate}
% \item if $M_{d}(d,d')\geq 2$, then $M_{d}(d,qd')\geq 2$ for any $q\geq 2$.
%This improves the result in \cite{Zhang}, which only shows that if $M_{d}(d,d')\geq 2$, then $M_{d}(d,2^{l}d')\geq 2$.
% \item We also give a lower bound of $M_{d}(d,qd)$ and $M_{d}(pd,qd)$ for general $d\geq 2, p,q\geq 1$, which impoves and covers all the lower bounds of $M_{d}(d,qd)$ in \cite{Xu,Xu1,Cheng,Liu}.
% \item In addition, We construct two MUMEBs in $\C^{2}\otimes \C^{3}$, which give a construction of MUMEBs in $\C^{2d}\otimes \C^{3kd}$. It means that we can construct MUMEBs in some systems $\C^{d}\otimes \C^{d'}$ where $d\nmid d'$.
% \item We also show that $M_{2}(3,2k)\geq 2$ which is better than the result in \cite{Han}, while it only shows that $M_{2}(3,4k)\geq 2$ where $k=2^{l}$.
% \end{enumerate}
%It gives a method to construct MUSEB$d$s in $\C^{pd}\otimes \C^{d'}$ from MUBs in $\C^{p}$ and MUMEBs in $\C^{d}\otimes \C^{d'}$. Specially, we show that Thus our results promote and cover all the previous work about MUSEB$k$s (MUMEBs).
\section{Definition and preliminary}\label{Def}
%A set of orthonormal bases $\mathcal{B}_{1}, \mathcal{B}_{2}, \cdots, \mathcal{B}_{m}$ in $\C^{d}$ is called mutually unbaised bases (MUBs) if every pair of $\mathcal{B}_{i}$ and $\mathcal{B}_{j}$ is mutually unbiased. Let $N(d)$ denote the maximum size $m$ of MUBs $\{\mathcal{B}_{1}, \mathcal{B}_{2}, \cdots, \mathcal{B}_{m}\}$ in $\C^{d}$. It is shown that $N(d)\leq d+1$ and $N(d)=d+1$ when $d$ is a prime power \cite{Andreas}. An open problem is to determine $N(d)$ when $d$ is not a prime power. $N(d)$ has a elementary lower bound. Let $d=p_{1}^{a_{1}}p_{1}^{a_{2}}\cdots p_{s}^{a_{s}}$ be a factorization of $d$ into distinct primes $p_{i}$. Then
%\begin{equation}\label{nd}
%N(d)\geq \min\{N(p_{1}^{a_{1}}),N(p_{2}^{a_{2}}),\cdots ,N(p_{s}^{a_{s}})\}.
%\end{equation}
%Actually, this bound can be improved in some special cases \cite{Pawe}.
Assume that $1\leq k \leq \min\{d,d'\}$. A (pure) state $\ket{\psi}\in \C^{d}\otimes \C^{d'}$ is called
a \emph{special entangled state with Schmidt number $k$} if it has Schmidt decomposition:
\begin{equation*}
\psi=\frac{1}{\sqrt{k}}\sum_{i=0}^{k-1}\ket{i}\ket{i'},
\end{equation*}
where $\{\ket{i}\}$ and $\{\ket{i'}\}$ are orthonormal sets of $\C^{d}$ and $\C^{d'}$, respectively \cite{Shi}.
In particular, a special entangled state with Schmidt number $k$ is a product state when $k=1$, and a maximally entangled state when $k=\min\{d,d'\}$. A basis $\mathcal{B}$ is called a special entangled basis with Schmidt number $k$ (SEB$k$) in $\C^{d}\otimes \C^{d'}$ if it contains $dd'$ pairwise orthogonal special entangled states with Schmidt number $k$ \cite{Shi,Guo15}. In particular, $\mathcal{B}$ is a maximally entangled state basis (MEB) when $k=\min\{d,d'\}$.
Two orthonormal bases $\mathcal{B}_{1}=\{\ket{\phi_{i}}\}_{i=1}^{dd'}$ and $\mathcal{B}_{2}=\{\ket{\psi_{i}}\}_{i=1}^{dd'}$ of $\C^{d}\otimes \C^{d'}$ are said to be \emph{mutually unbiased} if
\begin{equation*}
|\braket{\phi_{i}}{\psi_{j}}|=\frac{1}{\sqrt{dd'}}, \ (1\leq i, j\leq dd').
\end{equation*}
A set of SEB$k$s $\mathcal{A}=\{\mathcal{B}_{1}, \mathcal{B}_{2}, \ldots, \mathcal{B}_{m}\}$ in $\C^{d}\otimes \C^{d'}$ is called mutually unbaised (MUSEB$k$s) if every pair $\mathcal{B}_{i}$ and $\mathcal{B}_{j}$ are mutually unbiased. MUSEB$k$s are MUUBs when $k=d=d'$, and MUMEBs when $k=\min\{d,d'\}$.
When $k=d'=1$, MUSEB$k$s are just MUBs \cite{Andreas}, for which the maximum size is denoted by $N(d)$ conventionally. It is shown that $N(d)\leq d+1$ and $N(d)=d+1$ when $d$ is a prime power \cite{Andreas}. An open problem is to determine $N(d)$ when $d$ is not a prime power, which has an elementary lower bound ($d=p_1^{a_1}\dots p_s^{a_s}$, $p_1^{a_1}\leq\cdots\leq p_s^{a_s}$):
\begin{equation}\label{nd}
N(d)\geq \min\{N(p_{1}^{a_{1}}),\ldots ,N(p_{s}^{a_{s}})\}=p_{1}^{a_{1}}+1.
\end{equation}
Actually, this bound can be improved in some special cases \cite{Pawe}. It is shown that a complete set of MUBs of a bipartite system contains a fixed amount of entanglement \cite{wpz11}.
%
%Mutually unbiased bases (MUBs) have many applications in quantum information and quantum computation, such as quantum key distribution \cite{Cerf}, cryptographic protocols \cite{Cerf,Brie}, mean king's problem \cite{Engl}, quantum teleportation and super dense coding \cite{Durt,Klim, Revz}.
Now we define similar concepts in matrix spaces. Let $\mathcal{M}_{d\times d'}$ be the Hilbert space of all $d\times d'$ complex matrices equipped with inner product
\begin{equation*}\label{inn}
(A,B)=\text{Tr}(A^{\dag}B).
\end{equation*}
There is a one-to-one relation between $\C^{d}\otimes \C^{d'}$ and $ \mathcal{M}_{d\times d'}$ \cite{Guo15}:
\begin{equation*}
\begin{aligned}
|\psi_{i}\rangle=&\sum_{k,l}a_{k,l}^{(i)}|k\rangle|l'\rangle\in \C^{d}\otimes \C^{d'} \Longleftrightarrow A_{i}=[a_{k,l}^{(i)}]\in \mathcal{M}_{d\times d'},\\
&Sr(|\psi_{i}\rangle)=\text{rank}(A_{i}), \ \langle\psi_{i}|\psi_{j}\rangle=\text{Tr}(A_{i}^{\dag}A_{j}),
\end{aligned}
\end{equation*}
where $\{k\}$ and $\{l'\}$ are the computational bases of $\C^{d}$ and $\C^{d'}$, respectively, and $Sr(|\psi_{i}\rangle)$ denotes the Schmidt number of $|\psi_{i}\rangle$.
Let $\{\frac{1}{\sqrt{k}},\frac{1}{\sqrt{k}},\ldots,\frac{1}{\sqrt{k}}\}_{k}$ denote $k$ values of $\frac{1}{\sqrt{k}}$.
A $d\times d'$ matrix is called a \emph{$k$-singular-value-$\frac{1}{\sqrt{k}}$ matrix} if its nonzero singular values are $\{\frac{1}{\sqrt{k}},\frac{1}{\sqrt{k}},\ldots,\frac{1}{\sqrt{k}}\}_{k}$. Then $|\psi_{i}\rangle$ is a special entangled state with Schmidt number $k$ if and only if $A_{i}$ is a $k$-singular-value-$\frac{1}{\sqrt{k}}$ matrix. Especially, when $k=d=d'$, $|\psi_{i}\rangle$ is a maximally entangled state if and only if $\sqrt{d}A_{i}$ is a unitary matrix. We give the definitions of SEB$k$s and MUSEB$k$s in $\mathcal{M}_{d\times d'}$.
\begin{definition}
A set of $k$-singular-value-$\frac{1}{\sqrt{k}}$ matrices $\{A_{i}\}_{i=1}^{dd'}$ of $\mathcal{M}_{d\times d'}$ is called an SEB$k$ if $\text{Tr}(A_{i}^{\dag}A_{j})=\delta_{ij}$.
\end{definition}
\begin{definition}\label{defmu}
Two SEB$k$s $\mathcal{B}_{1}=\{A_{i}\}_{i=1}^{dd'}$ and $\mathcal{B}_{2}=\{B_{i}\}_{i=1}^{dd'}$ in $\mathcal{M}_{d\times d'}$ are said to be \emph{mutually unbiased} if
\begin{equation*}
|\text{Tr}(A_{i}^{\dag}B_{j})|=\frac{1}{\sqrt{dd'}}, \ (1\leq i, j\leq dd').
\end{equation*}
A set of SEB$k$s $\mathcal{A}=\{\mathcal{B}_{1}, \mathcal{B}_{2}, \ldots, \mathcal{B}_{m}\}$ in $\mathcal{M}_{d\times d'}$ is called MUSEB$k$s if every pair $\mathcal{B}_{i}$ and $\mathcal{B}_{j}$ are mutually unbiased.
\end{definition}
Due to the one-to-one relation, an SEB$k$ of $\mathcal{M}_{d\times d'}$ is an SEB$k$ of $\C^{d}\otimes \C^{d'}$;
MUSEB$k$s in $\mathcal{M}_{d\times d'}$ are MUSEB$k$s in $\C^{d}\otimes \C^{d'}$. When $k=d=d'$, Definition~\ref{defmu} gives MUMEBs in $\mathcal{M}_{d\times d}$, which corresponds to MUMEBs in $\C^{d}\otimes \C^{d}$. In this case, multiplying each matrix by a scaler $\sqrt{d}$ gives a set of mutually unbiased unitary bases (MUUBs) in $\C^{d}$ defined in \cite{Scott}. So Definition~\ref{defmu} can be viewed as a generation of MUUBs in \cite{Scott}, from which we know $M_{d}(d,d)\leq d^{2}-1$ and $M_{d}(d,d)=d^{2}-1$ when $d=2,3,5,7,11$.
\begin{remark}
In \cite{Shaari}, the authors gave another definition of MUUBs: Consider two distinct orthogonal bases, $\mA_0$ and $\mA_1$,
composed of unitary transformations for some subspace
of the vector space $\C^{d}$. $\mA_0$ and $\mA_1$ are MUUBs
provided that \[|\text{Tr}(A_0^{(i)\dag}A_1^{(j)})|^2=c, ~\forall~A_0^{(i)}\in \mA_0, ~A_1^{(j)}\in \mA_1,\]
for $i,j = 1, . . . ,n$ and some constant $c\neq 0$.
When $\mA_0$ and $\mA_1$ are composed of unitary transformations for the whole vector space $\C^{d}$, and $c=1$, then this definition reduces to Definition~\ref{defmu} for the case $k=d=d'$ by multiplying each matrix by a scaler $1/\sqrt{d}$.
\end{remark}
\section{The first MUMEBs with $d\nmid d'$}\label{sec:first}
In previous works of \cite{Tao,Zhang,Liu,Xu,Xu1,Cheng}, all constructions about MUMEBs are for $\C^{d}\otimes \C^{d'}$ with $d\mid d'$. Are there MUMEBs in $\C^{d}\otimes \C^{d'}$ when $d\nmid d'$? We next give an example in this case.
\begin{theorem}
\label{C23}
There are two MUMEBs in $\C^{2}\otimes \C^{3}$. Namely, $M_{2}(2,3)\geq 2$.
\end{theorem}
\begin{proof}
Let
$$\ket{\phi_{m,n}}=\frac{1}{\sqrt{2}}\sum_{p=0}^{1}(-1)^{np}\ket{p}\ket{(p\oplus m)'},$$ where $ n=0,1$, $m=0,1,2,$ and
$p\oplus m:=p+m \pmod 3$. It is easy to check that $\{\ket{\phi_{m,n}}\}$ is an MEB in $\C^{2}\otimes \C^{3}$.
Let $A$ be a $2\times2$ unitary matrix
$$
A=\left(
\begin{matrix}
A_{00} & A_{01}\\
A_{10} & A_{11}
\end{matrix}\right),
$$
and let
$$
\left(
\begin{matrix}
\ket{a_{0}}\\
\ket{a_{1}}
\end{matrix}\right)=\left(
\begin{matrix}
A_{00} & A_{01}\\
A_{10} & A_{11}
\end{matrix}\right)\left(
\begin{matrix}
\ket{0}\\
\ket{1}
\end{matrix}\right).
$$
Denote
$$\ket{\psi_{m,n}}=\frac{1}{\sqrt{2}}\sum_{p=0}^{1}(-1)^{np}\ket{a_{p}}\ket{(p\oplus m)'},$$ where $ n=0,1$ and $m=0,1,2.$
Then $\{\ket{\psi_{m,n}}\}$ is also an MEB in $\C^{2}\otimes \C^{3}$.
By definition, $\{\ket{\phi_{m,n}}\}$ and $\{\ket{\psi_{m,n}}\}$ are two MUMEBs if and only if
$$\left|\sum_{p_{2}=0}^{1}\sum_{p_{1}=0}^{1}(-1)^{\lambda}\braket{(p_{1}\oplus m_{1})'}{(p_{2}\oplus m_{2})'}A_{p_{2},p_{1}}\right|=\frac{2}{\sqrt{6}},$$for all $ \lambda=0,1$ and $m_{1},m_{2}=0,1,2.$
When $m_{1}=m_{2}$, we have $|A_{0,0}\pm A_{1,1}|=\frac{2}{\sqrt{6}}$, which implies $A_{0,0}\bot A_{1,1}$ and $|A_{0,0}|=|A_{1,1}|=\frac{1}{\sqrt{3}}$; when $m_{2}\equiv m_{1}+1 \pmod 3$, we have $|A_{0,1}|=\frac{2}{\sqrt{6}}$; when $m_{1}\equiv m_{2}+1 \pmod 3$, we have $|A_{1,0}|=\frac{2}{\sqrt{6}}$.
So we can assume that
\begin{equation}\label{aia}
A=\frac{1}{\sqrt{3}}\left(
\begin{matrix}
e^{i\theta_{1}} &\sqrt{2}e^{i\theta_{2}}\\
\sqrt{2}e^{i\theta_{3}} &e^{i(\theta_{1}+\frac{\pi}{2})}
\end{matrix}\right),
\end{equation} for some $\theta_i$, $i=1,2,3$.
Since $A$ is a unitary matrix, $\theta_{i}$ must satisfy the following condition:
\begin{equation}\label{iai}
\theta_{2}+\theta_{3}-2\theta_{1}=2k\pi+\frac{3\pi}{2}, \text{ for some } k\in Z.
\end{equation}
Taking
$$\theta_{1}=0,\ \theta_{2}=\frac{3\pi}{2},\ \theta_{3}=0,$$
we have
$$
A=\frac{1}{\sqrt{3}}\left(
\begin{matrix}
1 &-\sqrt{2}i\\
\sqrt{2} &i
\end{matrix}\right) \text{ and}
$$
$$
\ket{a_{0}}=\frac{1}{\sqrt{3}}(\ket{0}-\sqrt{2}i\ket{1}),\ \ket{a_{1}}=\frac{1}{\sqrt{3}}(\sqrt{2}\ket{0}+i\ket{1}).
$$
Then $\{\ket{\phi_{m,n}}\}$ and $\{\ket{\psi_{m,n}}\}$ are as follows:
\begin{equation}\label{16}
\left\{
\begin{aligned}
&\ket{\phi_{0,0}}=\frac{1}{\sqrt{2}}(\ket{00'}+\ket{11'})\\
&\ket{\phi_{0,1}}=\frac{1}{\sqrt{2}}(\ket{00'}-\ket{11'})\\
&\ket{\phi_{1,0}}=\frac{1}{\sqrt{2}}(\ket{01'}+\ket{12'})\\
&\ket{\phi_{1,1}}=\frac{1}{\sqrt{2}}(\ket{01'}-\ket{12'})\\
&\ket{\phi_{2,0}}=\frac{1}{\sqrt{2}}(\ket{02'}+\ket{10'})\\
&\ket{\phi_{2,1}}=\frac{1}{\sqrt{2}}(\ket{02'}-\ket{10'})
\end{aligned}
\right. ,
\end{equation}
\begin{equation}\label{17}
\left\{
\begin{aligned}
&\ket{\psi_{0,0}}=\frac{1}{\sqrt{6}}(\ket{00'}+\sqrt{2}\ket{01'}-\sqrt{2}i\ket{10'}+i\ket{11'})\\
&\ket{\psi_{0,1}}=\frac{1}{\sqrt{6}}(\ket{00'}-\sqrt{2}\ket{01'}-\sqrt{2}i\ket{10'}-i\ket{11'})\\
&\ket{\psi_{1,0}}=\frac{1}{\sqrt{6}}(\ket{01'}+\sqrt{2}\ket{02'}-\sqrt{2}i\ket{11'}+i\ket{12'})\\
&\ket{\psi_{1,1}}=\frac{1}{\sqrt{6}}(\ket{01'}-\sqrt{2}\ket{02'}-\sqrt{2}i\ket{11'}-i\ket{12'})\\
&\ket{\psi_{2,0}}=\frac{1}{\sqrt{6}}(\ket{02'}+\sqrt{2}\ket{00'}-\sqrt{2}i\ket{12'}+i\ket{10'})\\
&\ket{\psi_{2,1}}=\frac{1}{\sqrt{6}}(\ket{02'}-\sqrt{2}\ket{00'}-\sqrt{2}i\ket{12'}-i\ket{10'})
\end{aligned}
\right. .
\end{equation}
Thus the above $\{\ket{\phi_{m,n}}\}$ and $\{\ket{\psi_{m,n}}\}$ are two MUMEBs in $\C^{2}\otimes \C^{3}$.
\end{proof}
%\begin{remark}
Note that we can also take
$\theta_{1}=\frac{\pi}{4},\ \theta_{2}=\frac{3\pi}{4},\ \theta_{3}=\frac{5\pi}{4}$ in Eq. (\ref{aia}), then
$$
A=\frac{1}{\sqrt{3}}\left(
\begin{matrix}
\frac{\sqrt{2}}{2}(1+i) &-1+i\\
-1-i&\frac{\sqrt{2}}{2}(-1+i)
\end{matrix}\right) \text{ and}
$$
\begin{equation}\label{othermu}
\ket{a_{0}} =\frac{1}{\sqrt{3}}(\frac{\sqrt{2}}{2}(1+i)\ket{0}+(-1+i)\ket{1}),
\ket{a_{1}} =\frac{1}{\sqrt{3}}((-1-i)\ket{0}+\frac{\sqrt{2}}{2}(-1+i)\ket{1}).
\end{equation}
The above $\{\ket{\phi_{m,n}}\}$ and $\{\ket{\psi_{m,n}}\}$ are also two MUMEBs in $\C^{2}\otimes \C^{3}$. Unfortunately, if $A_{1}$ and $A_{2}$ are two matrices that satisfy the conditions of Eq.~(\ref{aia}) and Eq.~(\ref{iai}), then $A_{1}A_{2}$ cannot satisfy these conditions. It means we cannot construct three MUMEB$s$ in $\C^{2}\otimes \C^{3}$ by this way.
%\end{remark}
%\begin{remark}
We should mention that our method of constructing two MUMEBs in $\C^{2}\otimes \C^{3}$ in Theorem~\ref{C23} is similar to that of constructing MUMEBs in $\C^{d}\otimes \C^{kd}$ from \cite{Tao}, where a unitary matrix $A$ of size $kd\times kd$ is applied on the latter subspace $\C^{kd}$. However, using their method, we find that such a $3\times 3$ unitary matrix $A$ does not exist. So we cannot construct MUMEBs in $\C^{2}\otimes \C^{3}$ by the method in \cite{Tao}.
%\end{remark}
%\begin{remark}\label{matr}
Before closing this section, we convert $\{\ket{\phi_{m,n}}\}$ and $\{\ket{\psi_{m,n}}\}$ of Eq.~(\ref{16}) and Eq.~(\ref{17}) to MUMEBs $\mathcal{R}_{1}$ and $\mathcal{R}_{2}$ in $\mathcal{M}_{2\times 3}$, respectively.
\begin{align*}
\mathcal{R}_{1}=\{&\frac{1}{\sqrt{2}}\left(
\begin{matrix}
1 & 0 & 0\\
0 & 1 & 0
\end{matrix}\right), \ \frac{1}{\sqrt{2}}\left(
\begin{matrix}
1 & 0 & 0\\
0 & -1 & 0
\end{matrix}\right), \ \frac{1}{\sqrt{2}}\left(
\begin{matrix}
0 & 1 & 0\\
0 & 0 & 1
\end{matrix}\right), \\
&\frac{1}{\sqrt{2}}\left(
\begin{matrix}
0 & 1 & 0\\
0 & 0 & -1
\end{matrix}\right), \ \frac{1}{\sqrt{2}}\left(
\begin{matrix}
0 & 0 & 1\\
1 & 0 & 0
\end{matrix}
\right), \ \frac{1}{\sqrt{2}}\left(
\begin{matrix}
0 & 0 & 1\\
-1 & 0 &0
\end{matrix}
\right)\},\\
\mathcal{R}_{2}=\{&\frac{1}{\sqrt{6}}\left(
\begin{matrix}
1 & \sqrt{2} & 0\\
-\sqrt{2}i & i & 0
\end{matrix}\right), \ \frac{1}{\sqrt{6}}\left(
\begin{matrix}
1 & -\sqrt{2} & 0\\
-\sqrt{2}i & -i & 0
\end{matrix}
\right), \\
&\frac{1}{\sqrt{6}}\left(
\begin{matrix}
0 & 1 & \sqrt{2}\\
0 & -\sqrt{2}i & i
\end{matrix}
\right),
\frac{1}{\sqrt{6}}\left(
\begin{matrix}
0 & 1 & -\sqrt{2}\\
0 & -\sqrt{2}i & -i
\end{matrix}\right), \\
&\frac{1}{\sqrt{6}}\left(
\begin{matrix}
\sqrt{2}& 0 & 1\\
i & 0 &-\sqrt{2}i
\end{matrix}
\right), \ \frac{1}{\sqrt{6}}\left(
\begin{matrix}
-\sqrt{2}& 0 & 1\\
-i & 0 &-\sqrt{2}i
\end{matrix}
\right)\}.
\end{align*}
%\end{remark}
\section{Application}\label{sec:appli}
In this section, we propose an application of the two MUMEBs in $\C^2\otimes\C^3$ constructed in Eqs. (\ref{16}) and (\ref{17}) in Theorem \ref{thm:mumeb}. This is to tackle the special case of a long-standing open problem, namely whether four six-dimensional MUBs exist. The study of MUBs has extensive physical applications in quantum cryptography, tomography, and the construction of Wigner functions. It has been shown that there exist three six-dimensional MUBs, and widely believed that four six-dimensional MUBs may not exist in spite of great efforts from mathematical and physical community
\cite{Boykin05,bw08,bw09,jmm09,mub09,bw10,deb10,wpz11,mw12ijqi,mw12jpa135307,rle11,Sz12,Goyeneche13,mw12jpa102001,mb15,mpw16,
cy17,chen2018mutually,liang2019h_2}.
To explain the application, we refer to the vectors in an orthonormal basis in $\C^n$ as the column vectors of a unitary matrix. In particular we refer to the computational basis as the identity matrix. Hence, the existence of four six-dimensional MUBs is equivalent to the existence of the identity matrix, and three complex Hadamard matrices (CHMs) which are MU. If they exist then we refer to the three CHMs as the so-called MUB trio \cite{cy17}. From Lemma 11 (ii) and matrix $Y_6$ of \cite{cy17}, we have found the following criterion of determining when a CHM belongs to an MUB trio.
\begin{lemma}
\label{le:mub=trio}
\begin{enumerate}[(i)]
\item A CHM belongs to an MUB trio if and only if so does its transpose.
\item If a CHM has a $2\times3$ real submatrix then it does not belong to any MUB trio.
\end{enumerate}
\end{lemma}
Now we are in a position to present the application of the two MUMEBs in Eqs. (\ref{16}) and (\ref{17}).
\begin{theorem}
\label{thm:mumeb}
The two MUMEBs in Eqs. (\ref{16}) and (\ref{17}) cannot be extended to four MUBs.
\end{theorem}
\begin{proof}
In the following, we express the two MUMEBs as two unitary matrices $U$ and $V$, respectively.
$$
U= \left(
\begin{matrix}
\frac{1}{\sqrt{2}} & \frac{1}{\sqrt{2}} & 0 & 0 & 0 & 0 \\
0&0 & \frac{1}{\sqrt{2}} & \frac{1}{\sqrt{2}} &0 &0 \\
0&0 &0 &0 & \frac{1}{\sqrt{2}} & \frac{1}{\sqrt{2}} \\
0&0 &0 &0 & \frac{1}{\sqrt{2}} & -\frac{1}{\sqrt{2}}\\
\frac{1}{\sqrt{2}} & -\frac{1}{\sqrt{2}} & 0&0 &0 &0 \\
0&0 & \frac{1}{\sqrt{2}} & -\frac{1}{\sqrt{2}} & 0& 0
\end{matrix}
\right),
$$
and
$$
V=\left(
\begin{matrix}
\frac{1}{\sqrt{6}} & \frac{1}{\sqrt{6}} & 0 & 0 & \frac{\sqrt{2}}{\sqrt{6}} & -\frac{\sqrt{2}}{\sqrt{6}} \\
\frac{\sqrt{2}}{\sqrt{6}} & -\frac{\sqrt{2}}{\sqrt{6}} &
\frac{1}{\sqrt{6}} & \frac{1}{\sqrt{6}} & 0 & 0 \\
0 & 0 & \frac{\sqrt{2}}{\sqrt{6}} & -\frac{\sqrt{2}}{\sqrt{6}} & \frac{1}{\sqrt{6}} & \frac{1}{\sqrt{6}} \\
-\frac{\sqrt{2}i}{\sqrt{6}} & -\frac{\sqrt{2}i}{\sqrt{6}} & 0 & 0 & \frac{i}{\sqrt{6}} & -\frac{i}{\sqrt{6}} \\
\frac{i}{\sqrt{6}} & -\frac{i}{\sqrt{6}} & -\frac{\sqrt{2}i}{\sqrt{6}} &-\frac{\sqrt{2}i}{\sqrt{6}} & 0 & 0 \\
0& 0& \frac{i}{\sqrt{6}} & -\frac{i}{\sqrt{6}} &
-\frac{\sqrt{2}i}{\sqrt{6}} & -\frac{\sqrt{2}i}{\sqrt{6}}
\end{matrix}
\right).
$$
One can verify that the lower left $2\times3$ submatrix of $U^{\dagger} VQ$ is
$\left(
\begin{matrix}
-\frac{1}{\sqrt{6}} & -\frac{1}{\sqrt{6}} & \frac{1}{\sqrt{6}} \\
\frac{1}{\sqrt{6}} &\frac{1}{\sqrt{6}}&\frac{1}{\sqrt{6}}
\end{matrix}
\right)$, where the diagonal unitary $Q=\diag(-i,-i,1,1,1,1)$. Hence $U^{\dagger} V Q$ does not belong to any MUB trio in terms of Lemma \ref{le:mub=trio}. We have proven the assertion.
\end{proof}
Our results restrict the form of MUB trio, if it really exists. Following the proof of Theorem \ref{thm:mumeb}, one can similarly show that the other pair of MUMEBs constructed at Eq.~(\ref{othermu}) cannot be extended to four MUBs too. We may conjecture that, any two MUMEBs cannot be extended to four MUBs.
\section{A Recursive Construction of MUSEBks}\label{Mks}
In this section, we give a construction of MUSEB$k_{1}k_{2}$s in $\C^{pd}\otimes \C^{qd'}$ from MUSEB$k_{1}$s in $\C^{d}\otimes \C^{d'}$ and MUSEB$k_{2}$s in $\C^{p}\otimes \C^{q}$ for any positive integers $d$, $d'$, $p$ and $q$.
For any two sets $\mathcal{S}=\{A_{1},A_{2},\ldots,A_{n}\}$ and $\mathcal{T}=\{B_{1},B_{2},\ldots,B_{m}\}$, define $\mathcal{S}^{\mathrm{T}}:=\{A_{1}^{\mathrm{T}},A_{2}^{\mathrm{T}},\cdots,A_{n}^{\mathrm{T}}\},$ and
$$\mathcal{S}\otimes\mathcal{T}:=\{A_{i}\otimes B_{j}\mid i=1,\ldots, n, j=1,\ldots,m\}.$$
%Then $\mathcal{S}\otimes\mathcal{T}$ is a set of $mn$ matrices.
\begin{theorem}
\label{mum}
If there are $t_{1}$ MUSEB$k_{1}$s in $\C^{d}\otimes \C^{d'}$ and $t_{2}$ MUSEB$k_{2}$s in $\C^{p}\otimes \C^{q}$, then there are $ \min \{t_{1},t_{2}\}$ MUSEB$k_{1}k_{2}$s in $\C^{pd}\otimes \C^{qd'}$. Namely,
\begin{equation}
M_{k_{1}k_{2}}(pd,qd')\geq \min\{M_{k_{1}}(d,d'),M_{k_{2}}(p,q)\}.
\end{equation}
\end{theorem}
\begin{proof}We prove it by using the language of MUSEB$k$s in $\mathcal{M}_{d\times d'}$. Let $\mathcal{S}_{1}=\{A_{1},A_{2},\ldots,A_{dd'}\}$ and $\mathcal{S}_{2}=\{B_{1},B_{2},\ldots,B_{dd'}\}$ be two MUSEB$k_{1}$s in $\mathcal{M}_{d\times d'}$, that is, $A_{i},B_{j}$ are $k_1$-singular-value-$\frac{1}{\sqrt{k_1}}$ matrices, $Tr(A_{i}^{\dag}A_{j})=\delta_{i,j}$, $Tr(B_{i}^{\dag}B_{j})=\delta_{i,j}$ and $|\text{Tr}(A_{i}^{\dag}B_{j})|=\frac{1}{\sqrt{dd'}}$.
Similarly, let $\mathcal{T}_{1}=\{C_{1},C_{2},\ldots,C_{pq}\}$ and $\mathcal{T}_{2}=\{D_{1},D_{2},\ldots,D_{pq}\}$ be two MUSEB$k_{2}$s in $\mathcal{M}_{p\times q}$, then $C_{i},D_{j}$ each have $k_{2}$ singular values $\frac{1}{\sqrt{k_{2}}}$, $Tr(C_{i}^{\dag}C_{j})=\delta_{i,j}$, $Tr(D_{i}^{\dag}D_{j})=\delta_{i,j}$ and $|\text{Tr}(C_{i}^{\dag}D_{j})|=\frac{1}{\sqrt{pq}}$.
Then $\mathcal{S}_{1}\otimes \mathcal{T}_{1}$ and $\mathcal{S}_{2}\otimes \mathcal{T}_{2}$ are two sets of $dd'pq$ matrices in $\mathcal{M}_{pd\times qd'}$. We prove that they are two MUSEB$k_{1}k_{2}$s in $\mathcal{M}_{pd\times qd'}$.
First, we show that $\mathcal{S}_{1}\otimes \mathcal{T}_{1}$ and $\mathcal{S}_{2}\otimes \mathcal{T}_{2}$ are two SEB$k_{1}k_{2}$s. We only prove it for $\mathcal{S}_{1}\otimes \mathcal{T}_{1}$, the other one is similar. Since each $A_{i} \otimes C_{j}\in \mathcal{S}_{1}\otimes \mathcal{T}_{1}$ has nonzero singular values $\{\frac{1}{\sqrt{k_{1}k_{2}}},\frac{1}{\sqrt{k_{1}k_{2}}},\ldots,\frac{1}{\sqrt{k_{1}k_{2}}}\}_{k_{1}k_{2}}$, it is a $k_1k_2$-singular-value-$\frac{1}{\sqrt{k_1k_2}}$ matrix. Next, $\text{Tr}[(A_{i}\otimes C_{j})^{\dag}(A_{i'}\otimes C_{j'})]=\text{Tr}(A_{i}^{\dag}A_{i'})\cdot\text{Tr}(C_{j}^{\dag}C_{j'})=\delta_{i,i'}\delta_{j,j'}$, so they are orthogonal if $(i,j)\neq (i',j')$.
Finally, the mutually unbiased property follows from the fact that $|\text{Tr}[(A_{i}\otimes C_{j})^{\dag}(B_{i'}\otimes D_{j'})]|=|\text{Tr}(A_{i}^{\dag}B_{i'})|\cdot|\text{Tr}(C_{j}^{\dag}D_{j'})|=\frac{1}{\sqrt{pdqd'}}$.
Let $t=\min\{t_{1},t_{2}\}$. Assume that $\mathcal{S}_{1},\mathcal{S}_{2},\ldots,\mathcal{S}_{t}$ are $t$ MUSEB$k_{1}$s in $\mathcal{M}_{d\times d'}$, and $\mathcal{T}_{1},\mathcal{T}_{2},\ldots,\mathcal{T}_{t}$ are $t$ MUSEB$k_{2}$s in $\mathcal{M}_{p\times q}$.
It is easy to see that $\{\mathcal{S}_{1}\otimes \mathcal{T}_{1}, \mathcal{S}_{2}\otimes \mathcal{T}_{2},\ldots \mathcal{S}_{t}\otimes \mathcal{T}_{t}\}$ is a set of $t$ MUSEB$k_{1}k_{2}$s in $\mathcal{M}_{pd\times qd'}$ from the above analysis.
\end{proof}
Theorem~\ref{mum} gives a useful generic construction of MUSEB$k$s (MUMEBs). We illustrate its importance in several cases.
When $k_1=d\leq d'$ and $k_2=p\leq q$, we obtain the case for MUMEBs,
\begin{equation}\label{meb}
M_{pd}(pd,qd')\geq \min\{M_{d}(d,d'),M_{p}(p,q)\}.
\end{equation}
\begin{example}\label{rem1} There are five MUMEBs in $\C^{2}\otimes \C^{4}$ and three MUMEBs in $\C^{2}\otimes \C^{6}$ \cite{Tao}, then we can construct at least three MUMEBs in $\C^{4}\otimes \C^{24}$, that is, $M_{4}(4,24)\geq \min\{M_{2}(2,4),M_{2}(2,6)\}\geq 3$.
\end{example}
%
%
% Note that a special entangled state with Schmidt number $k$ in $\C^{d}\otimes \C^{d'}$ is also a special entangled state with Schmidt number $k$ in $\C^{d'}\otimes \C^{d}$. Then an SEB$k$ in $\C^{d}\otimes \C^{d'}$ is an SEB$k$ in $\C^{d'}\otimes \C^{d}$ and similarly MUSEB$k$s in $\C^{d'}\otimes \C^{d}$ are MUSEB$k$s in $\C^{d'}\otimes \C^{d}$. Namely, $M_{k}(d,d')=M_{k}(d',d)$.
Observing that Theorem~\ref{mum} does not require $d\leq d'$ and $p\leq q$, and the fact that $M_{k}(d,d')=M_{k}(d',d)$, we can get the following different example.
\begin{example}
From five MUMEBs in $\C^{2}\otimes \C^{4}$ and three MUMEBs in $\C^{2}\otimes \C^{6}$, we can also construct three MUSEB$4$s in $\C^{8}\otimes \C^{12}$, that is, $M_{4}(8,12)\geq \min\{M_{2}(4,2),M_{2}(2,6)\}\geq 3$.
\end{example}
Now we consider Theorem~\ref{mum} when some parameters of $d,d',p,q$ equal one. Observing that $\mathcal{M}_{1\times d'}$ is exactly the same space as $\C^{d'}$, an SEB$1$ in $\mathcal{M}_{1\times d'}$ is an orthonormal basis in $\C^{d'}$, and MUSEB$1$s in $\mathcal{M}_{1\times d'}$ are indeed MUBs in $\C^{d'}$. Namely, $M_{1}(1,d')=N(d')$. Hence, Eq~(\ref{nd}) is a special case of Theorem~\ref{mum} when $k_1=k_2=d=p=1$. When $p=1$ (or $q=1$), we have the following corollary.
\begin{cor}
\label{seb}
If there are $t_{1}$ MUSEBks in $\C^{d}\otimes \C^{d'}$, $t_{2}$ MUBs in $\C^{q}$, then there are at least $\min\{t_{1},t_{2}\}$ MUSEBks in $\C^{d}\otimes \C^{qd'}$. Namely,
\begin{equation}\label{k1k}
M_{k}(d,qd')\geq \min\{N(q),M_{k}(d,d')\}.
\end{equation}
If $k=d\leq d'$, then
\begin{equation}\label{mdq}
M_{d}(d,qd')\geq \min\{N(q),M_{d}(d,d')\},
\end{equation}
and
\begin{equation}\label{mpd}
M_{d}(pd,d')\geq \min\{N(p),M_{d}(d,d')\}.
\end{equation}
\end{cor}
\begin{remark}
As $N(q)\geq 3$ for general $q\geq 2$ \cite{Andreas}, by Eq.~(\ref{mdq}), \emph{if $M_{d}(d,d')\geq 2$, then $M_{d}(d,qd')\geq 2$ for any $q\geq 2$.}
This improves the result in \cite{Zhang}, which only showed that if $M_{d}(d,d')\geq 2$, then $M_{d}(d,2^{l}d')\geq 2$.
\end{remark}
%In \cite{Han}, the authors proposed an open problem of constructing MUSEB$k$s.
From Eq.~(\ref{mpd}), we give a method to construct MUSEB$k$s from MUBs and MUMEBs. In fact, we can construct MUSEB$d$s in $\C^{pd}\otimes \C^{d'}$ from MUBs in $\C^{p}$ and MUMEBs in $\C^{d}\otimes \C^{d'}$ for any $p\geq 2$.
\begin{example}\label{33}
We can construct three MUSEB$3$s in $\C^{6}\otimes \C^{6}$ from three MUMEBs in $\C^{3}\otimes \C^{3}$ and three MUBs in $\C^{2}$. Namely, $M_{3}(6,6)\geq 3$. For three MUMEBs in $\C^{3}\otimes \C^{3}$, we construct the following three bases in $\mathcal{M}_{3\times 3}$ with $w=e^{\frac{2\pi i}{3}}$ and $i=\sqrt{-1}$.
\begin{widetext}
\begin{align*}
\mathcal{S}_{1}=\{&\frac{1}{3\sqrt{3}}\left(
\begin{matrix}
w+2 &w+2 & w+2\\
2w^{2}+1 &w+2 &w^{2}+2w\\
w^{2}+2w &w+2 &2w^{2}+1
\end{matrix}\right), \ \frac{1}{3\sqrt{3}}\left(
\begin{matrix}
w+2 &w^{2}+2w &2w^{2}+1\\
2w^{2}+1 &w^{2}+2w &w+2\\
w^{2}+2w &w^{2}+2w &w^{2}+2w
\end{matrix}
\right),
\frac{1}{3\sqrt{3}}\left(
\begin{matrix}
w+2 &2w^{2}+1 &w^{2}+2w\\
2w^{2}+1 &2w^{2}+1 &2w^{2}+1 \\
w^{2}+2w &2w^{2}+1 &w+2
\end{matrix}
\right),\\
&\frac{1}{3\sqrt{3}}\left(
\begin{matrix}
w+2 & w+2 &w+2\\
w+2 & w^{2}+2w &2w^{2}+1\\
w+2 & 2w^{2}+1 &w^{2}+2w
\end{matrix}\right),
\frac{1}{3\sqrt{3}}\left(
\begin{matrix}
w+2 &w^{2}+2w &2w^{2}+1\\
w+2 &2w^{2}+1 &w^{2}+2w\\
w+2 &w+2 &w+2
\end{matrix}
\right), \ \frac{1}{3\sqrt{3}}\left(
\begin{matrix}
w+2 &2w^{2}+1 &w^{2}+2w\\
w+2 &w+2 &w+2 \\
w+2 &w^{2}+2w &2w^{2}+1
\end{matrix}
\right),\\
&\frac{1}{3\sqrt{3}}\left(
\begin{matrix}
w+2 &w+2 &w+2\\
w^{2}+2w &2w^{2}+1 &w+2\\
2w^{2}+1 &w^{2}+2w &w+2
\end{matrix}\right), \ \frac{1}{3\sqrt{3}}\left(
\begin{matrix}
w+2 &w^{2}+2w &2w^{2}+1 \\
w^{2}+2w &w+2 &2w^{2}+1\\
2w^{2}+1 &2w^{2}+1 &2w^{2}+1
\end{matrix}
\right),
\frac{1}{3\sqrt{3}}\left(
\begin{matrix}
w+2 &2w^{2}+1 &w^{2}+2w\\
w^{2}+2w &w^{2}+2w &w^{2}+2w\\
2w^{2}+1 &w+2 &w^{2}+2w
\end{matrix}
\right)\},\\
\mathcal{S}_{2}=\{&\frac{1}{3\sqrt{3}}\left(
\begin{matrix}
0 & 3w^{2} & 0\\
3w & 0 & 0\\
0 & 0 & 3
\end{matrix}\right), \ \frac{1}{3\sqrt{3}}\left(
\begin{matrix}
0 & 3 & 0\\
3w & 0 & 0\\
0 & 0 & 3w^{2}
\end{matrix}
\right), \ \frac{1}{3\sqrt{3}}\left(
\begin{matrix}
0 & 3w & 0\\
3w & 0 & 0\\
0 & 0 & 3w
\end{matrix}
\right),
\frac{1}{3\sqrt{3}}\left(
\begin{matrix}
3w^{2} & 0 & 0\\
0 & 0 & 3w\\
0 & 3 & 0
\end{matrix}\right), \ \frac{1}{3\sqrt{3}}\left(
\begin{matrix}
3w^{2}& 0 & 0\\
0 & 0 & 3\\
0 &3w & 0
\end{matrix}
\right),\\
&\frac{1}{3\sqrt{3}}\left(
\begin{matrix}
3w^{2} & 0 & 0\\
0 & 0 & 3w^{2}\\
0 & 3w^{2} & 0
\end{matrix}
\right),
\frac{1}{3\sqrt{3}}\left(
\begin{matrix}
0 & 0 & 3w^{2}\\
0 & 3w & 0\\
3 & 0 & 0
\end{matrix}\right), \ \frac{1}{3\sqrt{3}}\left(
\begin{matrix}
0 &0 &3w\\
0 &3w^{2} & 0\\
3 &0 & 0
\end{matrix}
\right), \ \frac{1}{3\sqrt{3}}\left(
\begin{matrix}
0 & 0 &3\\
0 &3 &0\\
3 &0 & 0
\end{matrix}
\right)\},\\
\mathcal{S}_{3}=\{&\frac{1}{\sqrt{3}}\left(
\begin{matrix}
1 & 0 & 0\\
0 & 1 & 0\\
0 & 0 & 1
\end{matrix}\right), \ \frac{1}{\sqrt{3}}\left(
\begin{matrix}
1 & 0 & 0\\
0 & w & 0\\
0 & 0 & w^{2}
\end{matrix}
\right), \ \frac{1}{\sqrt{3}}\left(
\begin{matrix}
1 & 0 & 0\\
0 & w^{2} & 0\\
0 & 0 & w
\end{matrix}
\right),
\frac{1}{\sqrt{3}}\left(
\begin{matrix}
0 & 0 & 1\\
1 & 0 & 0\\
0 & 1 & 0
\end{matrix}\right), \ \frac{1}{\sqrt{3}}\left(
\begin{matrix}
0 & 0 & w^{2}\\
1 & 0 & 0\\
0 & w & 0
\end{matrix}
\right),
\frac{1}{\sqrt{3}}\left(
\begin{matrix}
0 & 0 & w\\
1 & 0 &0\\
0 & w^{2} & 0
\end{matrix}
\right),\\
&\frac{1}{\sqrt{3}}\left(
\begin{matrix}
0 & 1 & 0\\
0 & 0 & 1\\
1 & 0 & 0
\end{matrix}\right), \ \frac{1}{\sqrt{3}}\left(
\begin{matrix}
0 &w &0\\
0 &0 & w^{2}\\
1 &0 & 0
\end{matrix}
\right), \ \frac{1}{\sqrt{3}}\left(
\begin{matrix}
0 &w^{2} &0\\
0 &0 &w\\
1 &0 & 0
\end{matrix}
\right)\}.
\end{align*}
For three MUBs in $\C^{2}$, choose
$$\mathcal{T}_{1}=\{(0,1),(1,0)\}, \ \mathcal{T}_{2}=\{\frac{1}{\sqrt{2}}(1,1),\frac{1}{\sqrt{2}}(1,-1)\}, \
\mathcal{T}_{3}=\{\frac{1}{\sqrt{2}}(1,i),\frac{1}{\sqrt{2}}(1,-i)\}.$$
\end{widetext}
Then $\{\mathcal{S}_{1}\otimes \mathcal{T}_{1}, \mathcal{S}_{2}\otimes \mathcal{T}_{2},\mathcal{S}_{3}\otimes \mathcal{T}_{3}\}$ is a set of three MUMEBs in $\C^{3}\otimes \C^{6}$ by Eq.~(\ref{mdq}), and $\{\mathcal{T}_{1}^{\mathrm{T}}\otimes \mathcal{S}_{1}\otimes \mathcal{T}_{1}, \mathcal{T}_{2}^{\mathrm{T}}\otimes\mathcal{S}_{2}\otimes \mathcal{T}_{2},\mathcal{T}_{3}^{\mathrm{T}}\otimes\mathcal{S}_{3}\otimes \mathcal{T}_{3}\}$ is a set of three MUSEB$3$s in
$\C^{6}\otimes \C^{6}$ by Eq.~(\ref{mpd}).
\end{example}
\section{Lower Bounds for MUMEBs}\label{Mbs}
%In this section, we give lower bounds of $M_{d}(d,d')$ for general $d\leq d'$. %We also construct two MUMEB$s$ in $\C^{2}\otimes \C^{3}$, which can be used to construct MUMEBs in some systems $\C^{d}\otimes \C^{d'}$ where $d\nmid d'$.
Through out this section, we always assume that $d=2^{t}p_{1}^{a_{1}}p_{2}^{a_{2}}\cdots p_{s}^{a_{s}}$, where $p_{1},p_{2},\ldots p_{s}$ are distinct odd primes such that $3\leq p_{1}^{a_{1}}\leq p_{2}^{a_{2}}\leq \cdots \leq p_{s}^{a_{s}}$. For convenience, we define $M_{1}(1,1):=\infty$.
\begin{theorem}\label{dddlower}
For any positive integer $d\geq 2$, we have $M_{d}(d,d)\geq 3$. In particular,
\begin{enumerate}[(i)]
\item $M_{d}(d,d)\geq \frac{p_{1}^{2a_{1}}-1}{2}$ when $t=0$;
\item $M_{d}(d,d)\geq 3$ when $t=1$;
\item $M_{d}(d,d)\geq \min\{2(2^t-1),\frac{p_{1}^{2a_{1}}-1}{2}\}$ when $t\geq 2$.
\end{enumerate}
\end{theorem}
\begin{proof} By Eq.~(\ref{meb}), we have
\begin{equation*}
M_{d}(d,d)\geq
\min\{M_{2^t}(2^t,2^t),M_{p_{1}^{a_{1}}}(p_{1}^{a_{1}},p_{1}^{a_{1}}),\ldots,M_{p_{s}^{a_{s}}}(p_{s}^{a_{s}},p_{s}^{a_{s}})\}.
\end{equation*}
Since $M_{p_{i}^{a_{i}}}(p_{i}^{a_{i}},p_{i}^{a_{i}})\geq\frac{p_{i}^{2a_{i}}-1}{2}$ by \cite{Xu1}, $M_{d}(d,d)\geq\min\{M_{2^t}(2^t,2^t), \frac{p_{1}^{2a_{1}}-1}{2}\}$. Then the three cases follow from the fact that $M_1(1,1)=\infty$, $M_{2}(2,2)=3$ and $M_{2^t}(2^t,2^t)\geq 2(2^t-1)$ by \cite{Xu}.
\end{proof}
Theorem~\ref{dddlower} gives a general lower bound $M_{d}(d,d)\geq 3$ for any $d\geq 2$, which is similar to the case of MUBs $N(q)\geq 3$ for any $q\geq 2$.
By Eq.~(\ref{mdq}) and Eq.~(\ref{mpd}), we have the following corollary.
% and Eq~(\ref{mpd})
\begin{cor}
\label{ddqdl} For all integers $d, p, q\geq 2$, we have
\begin{equation}\label{dkd1}
M_{d}(d,qd)\geq \min\{N(q),M_{d}(d,d)\}\geq 3,
\end{equation}
\begin{equation}
M_{d}(pd,qd)\geq \min\{N(p),M_{d}(d,qd)\}\geq 3. %\text{ and }
\end{equation}
%and
% \begin{equation}\label{dkd2}
% M_{d}(pd,qd)\geq \min\{N(p),M_{d}(d,qd)\}\geq 3.
% \end{equation}
\end{cor}
\begin{remark}
In \cite{Cheng}, it was shown that $M_{d}(d,qd)\geq \min\{(p_{1}')^{a_{1}'}+1,M_{d}(d,d)\}$ when $d$ is an \emph{odd} number.
Eq. (\ref{dkd1}) extends this result for any number $d\geq 2$.
\end{remark}
%, where $q\geq 2$ and $(p_{1}')^{a_{1}'}$ is the smallest prime power divisor of $q$
We have the following corollary by applying Eq.~(\ref{meb}) with Theorem~\ref{C23} and Eq.~ (\ref{dkd1}).
\begin{cor}\label{21k}
There are at least two MUMEBs in $\C^{2d}\otimes \C^{3qd}$ for any positive integers $q$ and $d$. Namely,
\begin{equation}\label{002}
M_{2d}(2d,3qd)\geq 2.
\end{equation}
There are at least two MUSEB$2$s in $\C^{3}\otimes \C^{2k}$ for any $k\geq 1$. Namely,
\begin{equation}\label{112}
M_{2}(3,2k)\geq 2.
\end{equation}
\end{cor}
\begin{remark}
Corollary~\ref{21k} shows that two MUMEBs exist in $\C^{d}\otimes \C^{d'}$ for infinitely many parameters $d,d'$ satisfying $d\nmid d'$. The result of Eq.~(\ref{112}) is better than that in \cite{Han}, which only showed that $M_{2}(3,4k)\geq 2$ where $k=2^{l}$.
\end{remark}
\begin{example}
There are two MUMEBs in $\C^{6}\otimes \C^{9}$, $\mathcal{R}_{1}\otimes \mathcal{S}_{1},\mathcal{R}_{2}\otimes \mathcal{S}_{2}$. Namely, $M_{6}(6,9)\geq 2$. Here, $\mathcal{R}_{1}, \mathcal{R}_{2}$ are from Sec.~\ref{sec:first}, and $\mathcal{S}_{1}, \mathcal{S}_{2}$ are constructed in Example~\ref{33}.
%
%We convert $\{\ket{\phi_{m,n}}\}$ and $\{\ket{\psi_{m,n}}\}$ of Eqs~(\ref{16}) and (\ref{17}) to MUMEBs $\mathcal{R}_{1}$ and $\mathcal{R}_{2}$ in $\mathcal{M}_{d\times d}$, respectively.
%
%Then $\{\mathcal{R}_{1}\otimes \mathcal{S}_{1},\mathcal{R}_{2}\otimes \mathcal{S}_{2}\}$ are two MUMEBs in $\C^{6}\otimes \C^{9}$ by Eq~(\ref{002}), where
\end{example}
\section{Conclusion}
\label{conpen}
In this paper we studied constructions of MUSEB$k$s (MUMEBs) and provided several better lower bounds on the maximum size of MUSEB$k$s. See Table~\ref{all} for a conclusion of all known results. By similar arguments as in \cite{Shaari}, one can consider the issue of distinguishability of SEB$k$s selected from a set of MUSEB$k$s and the use of MUSEB$k$s in a QKD setup. We can also consider the mutually unbiased measurements consisting of MUSEB$k$s, which plays a special role in the problem of state determination \cite{Woot}.
For the existence of MUSEB$k$s, there are still many open questions. For example, which $d$ can achieve the upper bound $M_{d}(d,d)\leq d^{2}-1$ besides $2,3,5,7,11$? Can we improve the lower bound of $M_{d}(d,qd)\geq 3$? Are there three MUMEBs in $\C^{2}\otimes \C^{3}$, while there are three MUBs in $\C^{6}$. Are there MUSEB$k$s in $\C^{d}\otimes \C^{d'}$
when $k\nmid d d'$? Are there MUMEB in $\C^{d}\otimes \C^{d'}$ for any $d\nmid d'$? The minimum unsolved case is $\C^{2}\otimes \C^{5}$.
\begin{acknowledgements}
FS and XZ were supported by NSFC under Grant No. 11771419, the Fundamental Research Funds for the Central Universities, and Anhui Initiative in Quantum Information Technologies under Grant No. AHY150200. LC was supported by the NNSF of China (Grant No. 11871089), and the Fundamental Research Funds for the Central Universities (Grant Nos. KG12080401 and ZG216S1902).
\end{acknowledgements}
%We gave a one-to-one correspondence between MUSEB$k$s in $\C^{d}\otimes \C^{d'}$ and MUSEB$k$s in $\mathcal{M}_{d\times d'}$, and we constructed $\min \{t_{1},t_{2}\}$ MUSEB$k_{1}k_{2}$s in $\C^{pd}\otimes \C^{qd'}$ from $t_{1}$ MUSEB$k_{1}$s in $\C^{d}\otimes \C^{d'}$ and $t_{2}$ MUSEB$k_{2}$s in $\C^{p}\otimes \C^{q}$. In addition, We first constructed MUMEBs in some systems $\C^{d}\otimes \C^{d'}$ where $d\nmid d'$.
\bibliographystyle{unsrt}
\bibliography{MUSEBk}
%\begin{thebibliography}{99}
% \bibitem{Schwinger}
% J. Schwinger, Unitary operator bases, Proc. Nat. Acad. Sci. U.S.A., \textbf{46}: 570–579, (1960).
% \bibitem{Bennett}
% C. H. Bennett and G. Brassard, Proceedings of the IEEE International Conference on Computers, Systems and Signal Processing, Bangalore, India (IEEE, New York, 1984), pp. 175-179.
% \bibitem{Acin}A. Ac\'in, E. Jan\'e and G. Vidal, Optimal estimation of quantum dynamics, Phys. Rev. A \textbf{64}, 050302(R) (2001).
% \bibitem{Chiribella2004}
% G. Chiribella, G. M. D'Ariano, P. Perinotti and M. F. Sacchi, Efficient use of quantum resources for the transmission of a reference frame, Phys. Rev. Lett. \textbf{93}, 180503 (2004).
% \bibitem{Lucamarini}
% M. Lucamarini and S. Mancini, Secure deterministic communication without entanglement, Phys. Rev. Lett. \textbf{94}, 140501 (2005).
% \bibitem{Chiribella2008}
% G. Chiribella, G. M. D'Ariano and P. Perinotti, Optimal cloning of unitary transformation, Phys. Rev. Lett. \textbf{101}, 180504 (2008).
% \bibitem{Cerf}
% N. J. Cerf, M. Bourennane, A. Karlsson and N. Gisin, Security of quantum key distribution using $d$-level
% systems, Phys. Rev. Lett. \textbf{88}, 127902 (2002).
% \bibitem{Brie}
% S. Brierley, Quantum key distribution highly sensitive to eavesdropping, arXiv:0910.2578v1 (2009).
% \bibitem{Engl}
% B. G. Englert and Y. Aharonov, The mean king's problem: prime degrees of freedom, Phys. Lett. A \textbf{284},
% 1-5 (2001).
% \bibitem{Durt}
% T. Durt, If $1 = 2\oplus3$, then $1=2\odot3$: Bell states, finite groups, and mutually unbiased bases, a unifying approach, arXiv:0401046v2 (2004).
% \bibitem{Klim}
% A. B. Klimov, D. Sych, L. L. Sanchez-Soto and G. Leuchs, Mutually unbiased bases and generalized
% Bell states, Phys Rev. A \textbf{79}, 052101 (2009).
% \bibitem{Revz}
% M. Revzen, Maximally entangled states via mutually unbiased collective bases,
% Phys. Rev. A \textbf{81}, 012113 (2010).
% \bibitem{Andreas}
% A. Klappenecker and M. R\"{o}tteler, Construction of mutually unbaised bases, International Conference on Finite Fields and Applications (Springer, Berlin, Heidelberg, 2003), pp. 137-144.
% \bibitem{Pawe}
% P. Wocjan and T. Beth, New construction of mutually unbiased bases
% in square dimensions, arXiv:0407081v1 (2004).
% \bibitem{Shi}
% F. Shi, X. D. Zhang and Y. Guo, Constructions of unextendible entangled bases, Quantum Inf. Process. \textbf{18}, 324 (2019).
% \bibitem{Guo15}
% Y. Guo, S. P. Du and X. L. Li, Entangled bases with fixed Schmidt number, J. Phys. A: Math Theor. \textbf{48}
% 245301 (2015).
% \bibitem{Tao}
% Y. H. Tao, H. Nan, J. Zhang and S. M. Fei, Mutually unbiased maximally entangled bases
% in $\C^{d}\otimes \C^{kd}$, Quantum Inf. Process. \textbf{14}, 2291-2300 (2015).
% J. Zhang, Y. H. Tao, H. Nan and S. M. Fei, Construction of mutually unbiased bases in $\C^{d}\otimes \C^{2^{l}d'}$, Quantum Inf. Process. \textbf{14}, 2635-2644 (2015).
% \bibitem{Liu}
% J. Y. Liu, M. H. Yang and K. Q. Feng, Mutually unbiased maximally entangled bases
% in $\C^{d}\otimes \C^{d}$, Quantum Inf. Process. \textbf{16}, 159 (2017).
% \bibitem{Xu}
% D. M. Xu, Construction of mutually unbiased maximally entangled
% bases through permutations of Hadamard matrices, Quantum Inf. Process. \textbf{16}, 65 (2017).
% \bibitem{Xu1}
% D. M. Xu, Trace-2 excluded subsets of special linear groups over finite
% fields and mutually unbiased maximally entangled bases, Quantum Inf. Process. \textbf{18}, 213 (2019).
% \bibitem{Cheng}
% X. Y. Cheng. and Y. Shang, New bounds of mutually unbiased maximally entangled bases in $\C^{d}\otimes \C^{kd}$, Quantum Inf. Comput. \textbf{18}(13–14), 1152-1164 (2018).
% \bibitem{Han}
% H. F. Han, G. J. Zhang, X. L. Yong, L. S. Xu and Y. H. Tao, Mutually unbiased special entangled bases
% with Schmidt number 2 in $\C^{3}\otimes \C^{4k}$, Quantum Inf. Process. \textbf{17}, 58 (2018).
% \bibitem{Scott}
% A. J. Scott, Optimizing quantum process tomography with
% unitary $2$-designs, J. Phys. A: Math. Theor. \textbf{41}, 055308 (2008).
% \bibitem{Shaari}
% J. S. Shaari, R. N. M. Nasir and S. Mancini, Mutually unbiased unitary bases, Phys, Rev. A \textbf{94}, 052328 (2016).
% \bibitem{Woot}
% W. K. Wootters and B. D. Fields, Optimal state-determination by mutually unbiased measurements, Ann.
% Phys. \textbf{191}, 363–381 (1989).
%\end{thebibliography}
\end{document}
%
% ****** End of file apssamp.tex ******
\begin{thebibliography}{10}
\bibitem{Schwinger}
Julian Schwinger.
\newblock Unitary operator bases.
\newblock {\em Proceedings of the national academy of sciences of the United
States Of America}, 46(4):570, 1960.
\bibitem{Bennett}
Charles~H Bennett and Gilles Brassard.
\newblock Proceedings of the ieee international conference on computers,
systems and signal processing.
\bibitem{Acin}
A~Acin, E~Jan{\'e}, and G~Vidal.
\newblock Optimal estimation of quantum dynamics.
\newblock {\em Physical Review A}, 64(5):050302, 2001.
\bibitem{Chiribella2004}
Giulio Chiribella, GM~D’Ariano, P~Perinotti, and MF~Sacchi.
\newblock Efficient use of quantum resources for the transmission of a
reference frame.
\newblock {\em Physical review letters}, 93(18):180503, 2004.
\bibitem{Lucamarini}
Marco Lucamarini and Stefano Mancini.
\newblock Secure deterministic communication without entanglement.
\newblock {\em Physical review letters}, 94(14):140501, 2005.
\bibitem{Chiribella2008}
Giulio Chiribella, Giacomo~Mauro D’Ariano, and Paolo Perinotti.
\newblock Optimal cloning of unitary transformation.
\newblock {\em Physical review letters}, 101(18):180504, 2008.
\bibitem{Scott}
Andrew~James Scott.
\newblock Optimizing quantum process tomography with unitary $2$-designs.
\newblock {\em Journal of Physics A: Mathematical and Theoretical},
41(5):055308, 2008.
\bibitem{Shaari}
Jesni~Shamsul Shaari, Rinie~NM Nasir, and Stefano Mancini.
\newblock Mutually unbiased unitary bases.
\newblock {\em Physical Review A}, 94(5):052328, 2016.
\bibitem{Tao}
Yuan-Hong Tao, Hua Nan, Jun Zhang, and Shao-Ming Fei.
\newblock Mutually unbiased maximally entangled bases in $\mathbb{C}^{d}\otimes
\mathbb{C}^{kd}$.
\newblock 2015.
\bibitem{Zhang}
Jun Zhang, Yuan-Hong Tao, Hua Nan, and Shao-Ming Fei.
\newblock Construction of mutually unbiased bases in $\mathbb{C}^{d}\otimes
\mathbb{C}^{2^ld'}$.
\newblock 2015.
\bibitem{Liu}
Junying Liu, Minghui Yang, and Keqin Feng.
\newblock Mutually unbiased maximally entangled bases in $\mathbb{C}^{d}\otimes
\mathbb{C}^{d}$.
\newblock {\em Quantum Information Processing}, 16(6):159, 2017.
\bibitem{Xu}
Dengming Xu.
\newblock Construction of mutually unbiased maximally entangled bases through
permutations of hadamard matrices.
\newblock {\em Quantum Information Processing}, 16:1--11, 2017.
\bibitem{Xu1}
Dengming Xu.
\newblock Trace-2 excluded subsets of special linear groups over finite fields
and mutually unbiased maximally entangled bases.
\newblock {\em Quantum Information Processing}, 18(7):213, 2019.
\bibitem{Cheng}
Xiaoya Cheng and Yun Shang.
\newblock New bounds of mutually unbiased maximally entangled bases in
$\mathbb{C}^{d}\otimes \mathbb{C}^{kd}$.
\newblock {\em Quantum Information \& Computation}, 18(13-14):1152--1164, 2018.
\bibitem{Han}
Yi-Fan Han, Gui-Jun Zhang, Xin-Lei Yong, Ling-Shan Xu, and Yuan-Hong Tao.
\newblock Mutually unbiased special entangled bases with schmidt number $2$ in
$\mathbb{C}^{3}\otimes \mathbb{C}^{4k}$.
\newblock {\em Quantum Information Processing}, 17(3):58, 2018.
\bibitem{Shi}
Fei Shi, Xiande Zhang, and Yu~Guo.
\newblock Constructions of unextendible entangled bases.
\newblock {\em Quantum Information Processing}, 18(10):324, 2019.
\bibitem{Guo15}
Yu~Guo, Shuanping Du, Xiulan Li, and Shengjun Wu.
\newblock Entangled bases with fixed schmidt number.
\newblock {\em Journal of Physics A: Mathematical and Theoretical},
48(24):245301, 2015.
\bibitem{Andreas}
Andreas Klappenecker and Martin R{\"o}tteler.
\newblock Constructions of mutually unbiased bases.
\newblock In {\em International Conference on Finite Fields and Applications},
pages 137--144. Springer, 2003.
\bibitem{Pawe}
Pawel Wocjan and Thomas Beth.
\newblock New construction of mutually unbiased bases in square dimensions.
\newblock {\em arXiv preprint quant-ph/0407081}, 2004.
\bibitem{wpz11}
M~Wiesniak, T~Paterek, and A~Zeilinger.
\newblock Entanglement in mutually unbiased bases.
\newblock {\em New Journal of Physics}, 13(5):053047, 2011.
\bibitem{Boykin05}
P.~O. {Boykin}, M.~{Sitharam}, P.~H. {Tiep}, and P.~{Wocjan}.
\newblock {Mutually Unbiased Bases and Orthogonal Decompositions of Lie
Algebras}, June 2005.
\bibitem{bw08}
Stephen Brierley and Stefan Weigert.
\newblock Maximal sets of mutually unbiased quantum states in dimension 6.
\newblock {\em Phys. Rev. A}, 78:042312, Oct 2008.
\bibitem{bw09}
Stephen Brierley and Stefan Weigert.
\newblock Constructing mutually unbiased bases in dimension six.
\newblock {\em Phys. Rev. A}, 79:052316, May 2009.
\bibitem{jmm09}
Philippe Jaming, M\'at\'e Matolcsi, P\'eter M\'ora, Ferenc Sz\"oll\"osi, and
Mih\'aly Weiner.
\newblock A generalized pauli problem and an infinite family of mub-triplets in
dimension 6.
\newblock {\em Journal of Physics A: Mathematical and Theoretical},
42(24):245305, 2009.
\bibitem{mub09}
Stephen Brierley.
\newblock {\em Mutually Unbiased Bases in Low Dimensions}.
\newblock PhD thesis, University of York, Department of Mathematics. 2009.
\bibitem{bw10}
Stephen Brierley and Stefan Weigert.
\newblock {Mutually Unbiased Bases and Semi-definite Programming}.
\newblock {\em Journal of Physics: Conference Series}, 254:012008, June 2010.
\bibitem{deb10}
T.~{Durt}, B.-G. {Englert}, I.~{Bengtsson}, and K.~{Zyczkowski}.
\newblock {On mutually unbiased bases}.
\newblock {\em Int. J. Quantum Information}, 8(4):535--640, 2010.
\bibitem{mw12ijqi}
Daniel McNulty and Stefan Weigert.
\newblock On the impossibility to extend triples of mutually unbiased product
bases in dimension six.
\newblock {\em International Journal of Quantum Information}, 10(05):1250056,
2012.
\bibitem{mw12jpa135307}
Daniel McNulty and Stefan Weigert.
\newblock All mutually unbiased product bases in dimension 6.
\newblock {\em Journal of Physics A: Mathematical and Theoretical},
45(13):135307, 2012.
\bibitem{rle11}
Philippe Raynal, Xin L\"u, and Berthold-Georg Englert.
\newblock Mutually unbiased bases in six dimensions: The four most distant
bases.
\newblock {\em Phys. Rev. A}, 83:062303, Jun 2011.
\bibitem{Sz12}
Ferenc Sz\"oll\"osi.
\newblock Complex hadamard matrices of order 6: a four-parameter family.
\newblock {\em Journal of the London Mathematical Society}, 85(3):616--632,
2012.
\bibitem{Goyeneche13}
D~Goyeneche.
\newblock Mutually unbiased triplets from non-affine families of complex
hadamard matrices in dimension 6.
\newblock {\em Journal of Physics A: Mathematical and Theoretical},
46(10):105301, 2013.
\bibitem{mw12jpa102001}
Daniel McNulty and Stefan Weigert.
\newblock The limited role of mutually unbiased product bases in dimension 6.
\newblock {\em Journal of Physics A: Mathematical and Theoretical},
45(10):102001, 2012.
\bibitem{mb15}
Andrew~S. Maxwell and Stephen Brierley.
\newblock {On properties of Karlsson Hadamards and sets of mutually unbiased
bases in dimension six}.
\newblock {\em Linear Algebra and its Applications}, 466(0):296 -- 306, 2015.
\bibitem{mpw16}
Daniel McNulty, Bogdan Pammer, and Stefan Weigert.
\newblock Mutually unbiased product bases for multiple qudits.
\newblock {\em Journal of Mathematical Physics}, 57(3), 2016.
\bibitem{cy17}
Lin Chen and Li~Yu.
\newblock Product states and schmidt rank of mutually unbiased bases in
dimension six.
\newblock {\em Journal of Physics A: Mathematical and Theoretical},
50(47):475304, 2017.
\bibitem{chen2018mutually}
Lin Chen and Li~Yu.
\newblock Mutually unbiased bases in dimension six containing a product-vector
basis.
\newblock {\em Quantum Information Processing}, 17(8):198, 2018.
\bibitem{liang2019h_2}
Mengfan Liang, Mengyao Hu, Lin Chen, and Xiaoyu Chen.
\newblock The $h_2$-reducible matrix in four six-dimensional mutually unbiased
bases.
\newblock {\em Quantum Information Processing}, 18(11):352, 2019.
\bibitem{Woot}
William~K Wootters and Brian~David Fields.
\newblock Optimal state-determination by mutually unbiased measurements.
\newblock {\em Annals of Physics}, 191(2):363--381, 1989.
\end{thebibliography}