On the normally torsion-freeness of square-free monomial ideals
Extracted source text · arXiv 2311.11850
\documentclass[12pt]{amsart}
\usepackage{amsmath, amsthm, amscd, amsfonts, amssymb, enumerate, verbatim, newlfont, calc, graphicx, color}
%\usepackage[bookmarksnumbered, colorlinks, plainpages]{hyperref}
\usepackage{tikz}
\usepackage{doi}
\newtheorem{theorem}{Theorem}[section]
\newtheorem{question}[theorem]{Question}
\newtheorem{questions}[theorem]{Questions}
\newtheorem{lemma}[theorem]{Lemma}
\newtheorem{proposition}[theorem]{Proposition}
\newtheorem{conjecture}[theorem]{Conjecture}
\newtheorem{corollary}[theorem]{Corollary}
\newtheorem{definition}[theorem]{Definition}
\newtheorem{condition}[theorem]{Condition}
\newtheorem{notation}[theorem]{Notation}
\newtheorem{remark}[theorem]{Remark}
\newtheorem{remarks}[theorem]{Remarks}
\newtheorem{example}[theorem]{Example}
\newtheorem{acknowledgement}{Acknowledgement}
\usepackage{mathtools}
\usepackage{pstricks}
\usepackage{epsfig}
\usepackage{pst-grad}
\usepackage{pst-plot}
\usepackage{subfig}
\textwidth=13cm
\begin{document}
\author[M. Nasernejad, V. Crispin Qui{\~{n}}onez, and W. Hochst\"attler]{Mehrdad ~Nasernejad$^{1,*}$, Veronica Crispin Qui\~{n}onez$^{2}$, Winfried Hochst\"attler$^{3}$}
\title[Normally torsion-free square-free monomial ideals] {On the normally torsion-freeness of square-free monomial ideals}
\subjclass{Primary 13B25, 13F20; Secondary 05C25, 05E40}
\keywords{Normally torsion-free monomial ideals, Embedded associated primes, Dominating ideals, Trees}
\thanks{$^*$Corresponding author}
\thanks{E-mail addresses: m$\_$nasernejad@yahoo.com, veronica.crispin@math.uu.se, and winfried.hochstaettler@fernuni-hagen.de}
\maketitle
\begin{center}
{\it
$^{1}$Univ. Artois, UR 2462, Laboratoire de Math\'{e}matique de Lens (LML), \\ F-62300 Lens, France \\
$^{2}$Department of Mathematics, Uppsala University, S-751 06, Uppsala, Sweden\\
$^{3}$FernUniversit\"{a}t in Hagen, Fakult\"{a}t f\"{u}r Mathematik und Informatik, 58084 Hagen, Germany
}
\end{center}
\vspace{0.4cm}
%%% ----------------------------------------------------------------------
\begin{abstract}
Let $I\subset R=K[x_1, \ldots, x_n]$ be a square-free monomial ideal, $\mathfrak{q}$ be a prime monomial ideal in $R$, $h$ be a square-free monomial in $R$ with $\mathrm{supp}(h) \cap (\mathrm{supp}(\mathfrak{q}) \cup \mathrm{supp}(I))=\emptyset$, and $L:=I\cap (\mathfrak{q}, h)$. In this paper,
we first focus on the associated primes of powers of $L$ and explore the normally torsion-freeness of $L$. We also give an application on a combinatorial result. Next, we study when a square-free monomial ideal is minimally not normally torsion-free. Particularly, we introduce a class of square-free monomial ideals, which are minimally not normally torsion-free.
\end{abstract}
%%% ----------------------------------------------------------------------
\vspace{0.4cm}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Preliminaries}
Let $R$ be a commutative Noetherian ring and $I$ be an ideal of $R$. A prime ideal $\mathfrak{p}\subset R$ is an {\it associated prime}
of $I$ if there exists an element $v$ in $R$ such that $\mathfrak{p}=(I:_R v)$. The {\it set of associated primes} of $I$, denoted by $\mathrm{Ass}_R(R/I)$, is the set of all prime ideals associated to $I$. A well-known result of Brodmann \cite{BR} showed that the sequence $\{\mathrm{Ass}_R(R/I^k)\}_{k \geq 1}$ of associated prime ideals is stationary for large $k$. That is, there exists a positive integer $k_0$ such that $\mathrm{Ass}_R(R/I^k)=\mathrm{Ass}_R(R/I^{k_0})$ for all $k\geq k_0$. The minimal such $k_0$ is called the
{\it index of stability} of $I$ and $\mathrm{Ass}_R(R/I^{k_0})$ is called the {\it stable set }
of associated prime ideals of $I$, which is denoted by $\mathrm{Ass}^{\infty }(I).$
An ideal $I\subset R$ is said to be {\it normally torsion-free} if, for all $k\geq 1$, we have $\mathrm{Ass}_R(R/I^k)\subseteq \mathrm{Ass}_R(R/I)$,
see \cite[Definition 4.3.28]{V1}. It is well-known that a finite simple undirected graph is bipartite if and only if its edge ideal is normally torsion-free, if and only if its cover ideal is normally torsion-free, consult \cite{GRV, SVV} for more information. In addition, in \cite[Corollary 4.6]{HRV}, it has been shown that every transversal polymatroidal ideal is normally torsion-free.
It has been proved in \cite[Theorem 3.2]{KHN1} that the Alexander dual of path ideals generated by all paths of length 2 in rooted trees are normally
torsion-free. Also, in \cite[Theorem 3.3]{N3}, the authors showed that the Alexander dual of the monomial
ideal generated by the paths of maximal lengths in a rooted starlike tree is normally torsion-free.
Furthermore, in \cite[Lemma 5.15]{NQKR}, it has been presented a characterization of all normally torsion-free $t$-spread principal Borel ideals.
More recently, in \cite[Theorem 4.6]{MNQ}, the authors proved that the edge ideal of any $\mathbf{t}$-spread $d$-partite hypergraph
$K^{\textbf{t}}_{V}$ is normally torsion-free.
However, normally torsion-freeness has been explored in the literature, but little is known for the normally torsion-free monomial ideals which are not square-free, for instance refer to \cite{OL}.
In this work, which has three main results, our aim is to continue studying the concept of normally torsion-freeness of square-free monomial ideals.
In order to reach this purpose, we first prove the following lemma:
\bigskip
\textbf{Lemma \ref{NTF1}.} Let $I\subset R=K[x_1, \ldots, x_n]$ be a normally torsion-free square-free monomial ideal, $\mathfrak{q}$ be a prime monomial ideal in $R$, and $h$ be a square-free monomial in $R$ with $\mathrm{supp}(h) \cap (\mathrm{supp}(\mathfrak{q}) \cup \mathrm{supp}(I))=\emptyset$
such that $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \bigcap_{x_r\in \mathrm{supp}(h)}(\mathfrak{q}, x_r)$ (respectively, $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \mathfrak{q}$) is a minimal primary decomposition of $I \cap (\mathfrak{q}, h)$ (respectively, $I \cap \mathfrak{q}$). Let $I$ and $I \cap \mathfrak{q}$ be normally torsion-free. Then
$L:=I \cap (\mathfrak{q}, h)$ is normally torsion-free.
\bigskip
As an application of Lemma \ref{NTF1}, we re-prove that dominating ideals of trees are normally torsion-free, see Theorem \ref{DI-TREES}, which has been shown in \cite[Theorem 2.3]{NQ} by combinatorial tools.
Next, we establish the second main result of this text in the following lemma:
\bigskip
\textbf{Lemma \ref{Ass-maximal1}.} Let $I \subset R=K[x_1, \ldots, x_n]$ be a square-free monomial ideal and $\mathfrak{q}$ be a prime monomial ideal in $R$. Let $L:=I \cap (\mathfrak{q}, x_r)$ such that $x_r\notin \mathrm{supp}(\mathfrak{q}) \cup \mathrm{supp}(I)$ and $\mathrm{Ass}(L)=\mathrm{Ass}(I) \cup \{(\mathfrak{q}, x_r)\}$. If $\mathfrak{p}=(L^s:x_r^{\theta}v)$ for some $s\geq 2$, $\theta \geq 0$, and monomial $v$ with
$x_r\nmid v$, then $\mathfrak{p}\setminus x_r \in \mathrm{Ass}(I^{\theta}(I\cap \mathfrak{q})^{s-\theta})$.
\bigskip
In broadly speaking, normally torsion-free square-free monomial ideals are closely related to Mengerian hypergraphs. A hypergraph is called
\textit{Mengerian} if it satisfies a certain min-max equation, which is known as the Mengerian property in hypergraph theory or has the max-flow min-cut property in integer programming. In fact, it is well-known that if $\mathcal{C}$ is a clutter and $I=I(\mathcal{C})$ its edge ideal, then $I$ is normally torsion-free if and only if $I^{k}=I^{(k)}$ for all $k\geq 1$, if and only if $\mathcal{C}$ is Mengerian, if and only if $\mathcal{C}$ has the max-flow min-cut (MFMC for short) property, see \cite[Theorem 14.3.6]{V1}. On the other hand, one of the famous and long-standing questions in this subject is
\textit{Conforti-Cornu\'e{j}ols conjecture}, which says that a hypergraph has the packing property if and only if it is Mengerian, see
\cite[Conjecture 14.3.19]{V1}.
Furthermore, finding a counterexample for this conjecture is too hard, and a well-known result in this theme is due to H\`a and Morey, which says that
a minimal counterexample to the Conforti-Cornu\'e{j}ols conjecture cannot be unmixed, see \cite[Corollary 3.10]{HM}. Three definitions arise along this conjecture are as follows:
\begin{enumerate}
\item A clutter $\mathcal{C}$ is {\it minimally non-MFMC} if it does not have the MFMC property
but all its proper minors do. Equivalently, $I(\mathcal{C})$ is {\it minimally not normally torsion-free} if $I(\mathcal{C})$ is not normally torsion-free but all its proper minors are normally torsion-free.
\item A clutter is {\it minimally nonideal} if it is not ideal but all its proper minors are.
\item A clutter is {\it minimally nonpacking} if it does not pack but all its proper minors do.
\end{enumerate}
As the third main result, we concentrate on the first definition above, and show the following proposition:
\bigskip
\textbf{Proposition \ref{MNNT}.} Let $I\subset R=K[x_1, \ldots, x_n]$ be a square-free monomial ideal and $\mathfrak{m}=(x_1, \ldots, x_n)$
such that there exists some $ m\geq 1$ with $\mathrm{Ass}(I^s)=\mathrm{Min}(I)$ for all $1\leq s \leq m$, and $\mathrm{Ass}(I^s)= \mathrm{Min}(I) {\cup} \{\mathfrak{m}\}$ for all $s\geq m+1$ such that $\mathfrak{m}\setminus x_j \notin I^s\setminus x_j$ for all $j=1, \ldots, n$.
Then $I$ is minimally not normally torsion-free. In particular, we have $m\geq \beta_1(I)$, where $\beta_1(I)$ denotes the maximum cardinality of an independent set in $I$.
Finally, we close this paper with introducing a class of square-free monomial ideals originating from graph theory, which are minimally not normally torsion-free,
refer to Example \ref{Exam.MNNT}.
\bigskip
Throughout this paper, $R=K[x_1, \ldots, x_n]$ stands for a polynomial ring over a field $K$ in the $n$ indeterminates $x_1, \ldots, x_n$.
Moreover, $\mathcal{G}(I)$ denotes the unique minimal set of monomial generators of a monomial ideal $I\subset R$.
The {\em support} of a monomial $u\in R$, denoted by $\mathrm{supp}(u)$, is the set of variables that divide $u$. Moreover, for a monomial ideal $I$, we set $\mathrm{supp}(I)=\bigcup_{u \in \mathcal{G}(I)}\mathrm{supp}(u)$. Furthermore, when there is no confusion about the underlying ring, we will denote the set of associated primes of $I$ simply by $\mathrm{Ass}(R/I)$ or $\mathrm{Ass}(I)$. Also, any necessary definitions related to graph theory can be found in \cite{West}.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Some results on normally torsion-free square-free monomial ideals }
In this section, we present some results on normally torsion-free square-free monomial ideals.
To show Proposition \ref{PRODUCT}, we employ Lemma \ref{PRODUCT1}. We start with proving Lemma \ref{PRODUCT1} in the following result.
\begin{lemma}\label{PRODUCT1}
Let $I\subset R=K[x_1, \ldots, x_r]$ be a square-free monomial ideal and $\mathfrak{q}$ be a prime monomial ideal in $R$ such that $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \mathfrak{q}$ is a minimal primary decomposition of
$I \cap \mathfrak{q}$. Let $I$ and $I \cap \mathfrak{q}$ be normally torsion-free.
Then, for all $m,n \geq 1$, we have $I^n(I\cap \mathfrak{q})^m=I^{n+m} \cap \mathfrak{q}^m$.
\end{lemma}
\begin{proof}
Fix $m,n \geq 1$. It is straightforward to verify that $I^n(I\cap \mathfrak{q})^m\subseteq I^{n+m} \cap \mathfrak{q}^m$.
It thus remains to prove the reverse inclusion. To accomplish this, take an arbitrary monomial $u$ in $I^{n+m} \cap \mathfrak{q}^m$.
It follows from $u\in I^{n+m}$ that there exist monomials $g_1, \ldots, g_n, h_1, \ldots, h_m\in \mathcal{G}(I)$ and some monomial $M$ such that
$u=g_1 \cdots g_n h_1 \cdots h_m M$. We also deduce from $u\in \mathfrak{q}^m$ that $u=x_{i_1} \cdots x_{i_m}w$, where
$x_{i_1}, \ldots, x_{i_m}\in \mathfrak{q}$ and some monomial $w$. Hence, we get $x_{i_1} \cdots x_{i_m} \mid g_1 \cdots g_n h_1 \cdots h_m M$.
Without loss of generality, we partite the set $\{x_{i_1}, \ldots, x_{i_m}\}$ into $S_1, \ldots, S_a$, where $a\leq n$,
$S_{a+1}, \ldots, S_b$, where $b-a\leq m$, $S_{b+1}$ sets such that $\bigcup_{i=1}^{b+1}S_i=\{x_{i_1}, \ldots, x_{i_m}\}$ and
for all $x_j \in S_k$ with $1\leq k \leq a$ (respectively, $x_j \in S_z$ with $a+1 \leq z \leq b$, and $x_j \in S_{b+1}$), we have
$x_j \mid g_k$ (respectively, $x_j \mid h_z$, and $x_j \mid M$). This implies that $g_k \in \mathfrak{q}^{|S_k|}$ (respectively,
$h_z \in \mathfrak{q}^{|S_z|}$, and $M\in \mathfrak{q}^{|S_{b+1}|}$). Now, we can distribute the rest $n-a+m-b+a=m+n-b$ monomials of types $g_{i}$'s and $h_j$'s such that, for $k=1, \ldots, a$, $g_kH_k \in \mathfrak{q}^{|S_k|}\cap I^{|S_k|}$, where $H_k$ is the product of $|S_k|-1$ monomials of $m+n-b$ monomials of types $g_{i}$'s and $h_j$'s (respectively, for $z=a+1, \ldots, b$,
$h_z T_z\in \mathfrak{q}^{|S_z|}\cap I^{|S_z|}$, where $T_z$ is the product of $|S_z|-1$ monomials of $m+n-b$ monomials of types $g_{i}$'s and $h_j$'s, and $ML\in \mathfrak{q}^{|S_{b+1}|}\cap I^{|S_{b+1}|}$, where $L$ is the product of $|S_{b+1}|$ monomials of $m+n-b$ monomials of types $g_{i}$'s and $h_j$'s).
Since $I$ and $I \cap \mathfrak{q}$ are normally torsion-free, one can deduce from \cite[Proposition 4.3.29 and Exercise 6.1.25]{V1} that, for all $\ell \geq 1$, we obtain $I^\ell=I^{(\ell)}=\bigcap_{\mathfrak{p}\in \mathrm{Min}(I)} \mathfrak{p}^\ell$ and
$(I\cap \mathfrak{q})^\ell=(I\cap \mathfrak{q})^{(\ell)}=\bigcap_{\mathfrak{p}\in \mathrm{Min}(I\cap \mathfrak{q})} \mathfrak{p}^\ell.$
By virtue of $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \mathfrak{q}$ is a minimal primary decomposition, we get $ \mathrm{Min}(I\cap \mathfrak{q})=\mathrm{Min}(I) \cup \{\mathfrak{q}\}$, and so $(I\cap \mathfrak{q})^\ell=I^\ell \cap \mathfrak{q}^\ell$.
In particular, we have $g_kH_k \in (\mathfrak{q}\cap I)^{|S_k|}$ for $k=1, \ldots, a$ (respectively,
$h_zT_z \in (\mathfrak{q}\cap I)^{|S_z|}$ for $z=a+1, \ldots, b$, and $ML\in (\mathfrak{q}\cap I)^{|S_{b+1}|}$).
It follows from $\bigcup_{i=1}^{b+1}S_i=\{x_{i_1}, \ldots, x_{i_m}\}$ that $\sum_{i=1}^{b+1}|S_i|=m$. This gives rise to
$u\in I^n(I\cap \mathfrak{q})^m$.
Consequently, for all $m,n \geq 1$, we obtain $I^n(I\cap \mathfrak{q})^m=I^{n+m} \cap \mathfrak{q}^m$, as claimed.
\end{proof}
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
\begin{proposition}\label{PRODUCT}
Let $I\subset R=K[x_1, \ldots, x_r]$ be a square-free monomial ideal and $\mathfrak{q}$ be a prime monomial ideal in $R$ such that $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \mathfrak{q}$ is a minimal primary decomposition of
$I \cap \mathfrak{q}$. Let $I$ and $I \cap \mathfrak{q}$ be normally torsion-free. Then, for all $m,n \geq 1$, the following statements hold:
\begin{itemize}
\item[(i)] $\mathrm{Ass}(I^n(I\cap \mathfrak{q})^m)=\mathrm{Min}(I) \cup \{\mathfrak{q}\}$.
\item[(ii)] The monomial ideal $I^n(I\cap \mathfrak{q})^m$ is normally torsion-free.
\end{itemize}
\end{proposition}
\begin{proof}
(i) Fix $m,n \geq 1$. It follows from Lemma \ref{PRODUCT1} that $I^n(I\cap \mathfrak{q})^m=I^{n+m} \cap \mathfrak{q}^m$. Thus, we must show that $\mathrm{Ass}(I^{n+m} \cap \mathfrak{q}^m)=\mathrm{Min}(I) \cup \{\mathfrak{q}\}$. To do this, let $\mathrm{Ass}(I)=\{\mathfrak{p}_1, \mathfrak{p}_2,
\ldots, \mathfrak{p}_t\}$. Since $I$ is normally torsion-free, one can conclude from \cite[Proposition 4.3.29 and Exercise 6.1.25]{V1} that
$I^{n+m}= \mathfrak{p}^{n+m}_1 \cap \mathfrak{p}^{n+m}_2 \cap \cdots \cap \mathfrak{p}^{n+m}_t$ is a minimal primary decomposition of $I^{n+m}$. We establish $\mathfrak{p}^{n+m}_1 \cap \mathfrak{p}^{n+m}_2 \cap \cdots \cap \mathfrak{p}^{n+m}_t \cap \mathfrak{q}^m$
is a minimal primary decomposition of $I^{n+m} \cap \mathfrak{q}^m$. For this, we prove
$\mathfrak{p}^{n+m}_1 \cap \mathfrak{p}^{n+m}_2 \cap \cdots \cap \mathfrak{p}^{n+m}_t \nsubseteq \mathfrak{q}^m$ and also
$\bigcap_{i=1, i\neq j}^{t} \mathfrak{p}^{n+m}_i \cap \mathfrak{q}^m \nsubseteq \mathfrak{p}^{n+m}_j,$
where $j=1, \ldots, t$. If $\mathfrak{p}^{n+m}_1 \cap \mathfrak{p}^{n+m}_2 \cap \cdots \cap \mathfrak{p}^{n+m}_t \subseteq \mathfrak{q}^m$, then $\sqrt{I^{n+m}} \subseteq \sqrt{\mathfrak{q}^m}$, and so $\sqrt{I}\subseteq \mathfrak{q}$. Since $I=\sqrt{I}$, we get $I\subseteq \mathfrak{q}$, which contradicts the fact that $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \mathfrak{q}$ is a minimal primary decomposition. Suppose, on the contrary, that $\bigcap_{i=1, i\neq j}^{t} \mathfrak{p}^{n+m}_i \cap \mathfrak{q}^m \subseteq \mathfrak{p}^{n+m}_j$ for some $1\leq j \leq t$. As $\mathfrak{p}^{n+m}_1 \cap \mathfrak{p}^{n+m}_2 \cap \cdots \cap \mathfrak{p}^{n+m}_t$ is a minimal primary decomposition of $I^{n+m}$, this implies that there exists an element $u \in \bigcap_{i=1, i\neq j}^{t} \mathfrak{p}^{n+m}_i \setminus \mathfrak{p}^{n+m}_j$.
On account of $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \mathfrak{q}$ is a minimal primary decomposition, we obtain $\mathfrak{q}\nsubseteq \mathfrak{p}_j$. Hence, there exists some element $v\in \mathfrak{q}\setminus \mathfrak{p}_j$. This yields that $uv^m \in \bigcap_{i=1, i\neq j}^{t} \mathfrak{p}^{n+m}_i \cap \mathfrak{q}^m$, and so $uv^m \in \mathfrak{p}^{n+m}_j$. In view of \cite[Corollary 6.1.8]{V1}, we know that $ \mathfrak{p}^{n+m}_j$ is $\mathfrak{p}_j$-primary. We can now derive from $uv^m \in \mathfrak{p}^{n+m}_j$ that $u\in \mathfrak{p}^{n+m}_j$ or $v^m\in \sqrt{ \mathfrak{p}^{n+m}_j}$. We therefore get, $u\in \mathfrak{p}^{n+m}_j$ or $v\in \mathfrak{p}_j$. This leads to a contradiction. Hence, $\bigcap_{i=1}^{t}\mathfrak{p}^{n+m}_i \cap \mathfrak{q}^m$ is a minimal primary decomposition of $I^{n+m} \cap \mathfrak{q}^m$. Consequently, $\mathrm{Ass}(I^n(I\cap \mathfrak{q})^m)=\mathrm{Min}(I) \cup \{\mathfrak{q}\}.$
(ii) Set $L:=I^n(I\cap \mathfrak{q})^m$ and fix $t\geq 1$. It follows from part (i) that
$$\mathrm{Ass}(L^t)=\mathrm{Ass}(I^{nt}(I \cap \mathfrak{q})^{mt})=\mathrm{Min}(I) \cup \{\mathfrak{q}\}=\mathrm{Ass}(L).$$
This means that $L$ is normally torsion-free, as claimed.
\end{proof}
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
To establish Lemma \ref{NTF1}, we need to use Lemma \ref{Ass-contraction}. For this purpose, we state and prove it in the next lemma.
\begin{lemma} \label{Ass-contraction}
Let $I\subset R=K[x_1, \ldots, x_n]$ be a normally torsion-free square-free monomial ideal, $\mathfrak{q}$ be a prime monomial ideal in $R$, and $h$ be a square-free monomial in $R$ with $\mathrm{supp}(h) \cap (\mathrm{supp}(\mathfrak{q}) \cup \mathrm{supp}(I))=\emptyset$ such that $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \bigcap_{x_r\in \mathrm{supp}(h)}(\mathfrak{q}, x_r)$ is a minimal primary decomposition of $L:=I \cap (\mathfrak{q}, h)$. If $\mathfrak{p}\in\mathrm{Ass}(L^s)\setminus \mathrm{Min}(L)$ for some $s\geq 2$,
then $(\mathfrak{q}, x_r) \subsetneq \mathfrak{p}$ for some $x_r \in \mathrm{supp}(h)$.
\end{lemma}
\begin{proof}
Let $\mathfrak{p}\in\mathrm{Ass}(L^s)\setminus \mathrm{Min}(L)$ for some $s\geq 2$. Since $\mathfrak{p}\notin \mathrm{Min}(L)$, this implies that $(\mathfrak{q}, x_r) \neq \mathfrak{p}$ for any $x_r \in \mathrm{supp}(h)$. We claim that there exists some $x_r \in \mathrm{supp}(h)$ such that
$x_r \in \mathfrak{p}$. Suppose, on the contrary, that $x_r \notin \mathfrak{p}$ for any $x_r \in \mathrm{supp}(h)$.
Due to $\mathfrak{p}\in\mathrm{Ass}(L^s)$, it follows readily from \cite[Theorem 4.6(viii)]{RNA} that $\mathfrak{p} \in \mathrm{Ass}(L(\mathfrak{p})^s)$.
Thanks to $x_r \notin \mathfrak{p}$ for any $x_r \in \mathrm{supp}(h)$, we get $(\mathfrak{q}, h)(\mathfrak{p})=(1)$, and by virtue of
$L(\mathfrak{p})=I(\mathfrak{p}) \cap (\mathfrak{q}, h)(\mathfrak{p})$, this leads to $L(\mathfrak{p})=I(\mathfrak{p})$. In addition, the assumption $\mathrm{supp}(h) \cap (\mathrm{supp}(\mathfrak{q}) \cup \mathrm{supp}(I))=\emptyset$ yields that $L=I\cap \mathfrak{q} +hI$. It follows immediately from $L \subseteq \mathfrak{p}$ and $x_r \notin \mathfrak{p}$ for any $x_r \in \mathrm{supp}(h)$ that $I \subseteq \mathfrak{p}$.
One can now deduce from \cite[Theorem 3.15]{SN} that $I(\mathfrak{p})$ is normally torsion-free, and so $L(\mathfrak{p})$ is normally torsion-free as well. We therefore have $\mathfrak{p} \in \mathrm{Ass}(L(\mathfrak{p}))$, and \cite[Theorem 4.6(viii)]{RNA} gives that $\mathfrak{p} \in \mathrm{Ass}(L)$. This implies that $\mathfrak{p}\in \mathrm{Min}(L)$, which is a contradiction. Accordingly, there exists some $x_r \in \mathrm{supp}(h)$ such that
$x_r \in \mathfrak{p}$, say $x_{r_0}$. As $L^s \subseteq \mathfrak{p}$ and $x_{r_0} \in \mathfrak{p}$, this forces us to obtain $(\mathfrak{q}, x_{r_0}) \subsetneq \mathfrak{p}$, as claimed.
\end{proof}
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
We are in a position to provide the first main result of this paper in the following lemma.
\begin{lemma} \label{NTF1}
Let $I\subset R=K[x_1, \ldots, x_n]$ be a normally torsion-free square-free monomial ideal, $\mathfrak{q}$ be a prime monomial ideal in $R$, and $h$ be a square-free monomial in $R$ with $\mathrm{supp}(h) \cap (\mathrm{supp}(\mathfrak{q}) \cup \mathrm{supp}(I))=\emptyset$
such that $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \bigcap_{x_r\in \mathrm{supp}(h)}(\mathfrak{q}, x_r)$ (respectively, $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \mathfrak{q}$) is a minimal primary decomposition of $I \cap (\mathfrak{q}, h)$ (respectively, $I \cap \mathfrak{q}$). Let $I$ and $I \cap \mathfrak{q}$ be normally torsion-free. Then
$L:=I \cap (\mathfrak{q}, h)$ is normally torsion-free.
\end{lemma}
\begin{proof}
Let $\mathfrak{p}\in\mathrm{Ass}(L^s)\setminus \mathrm{Min}(L)$ for some $s\geq 2$. It follows at once from Lemma \ref{Ass-contraction} that
$(\mathfrak{q}, x_r) \subsetneq \mathfrak{p}$ for some $x_r \in \mathrm{supp}(h)$. Hence, we can view $\mathfrak{p}$ as
$\mathfrak{p}=(\mathfrak{q}', \mathfrak{q}'')$ with $\mathfrak{q} \subsetneq \mathfrak{q}'\subseteq \mathrm{supp}(\mathfrak{q}) \cup
\mathrm{supp}(I)$ and $\mathfrak{q}'' \subseteq \mathrm{supp}(h)$.
Since $L=I \cap (\mathfrak{q}, h)$ and $\mathrm{supp}(h) \cap (\mathrm{supp}(\mathfrak{q}) \cup \mathrm{supp}(I))=\emptyset$, we get
$L=I \cap \mathfrak{q} + hI$. We first claim that, for all $1 \leq \ell \leq s$, we have
$(L^s:h^{\ell})=I^{\ell-1}(I\cap \mathfrak{q})^{s-\ell+1}+I^{\ell}L^{s-\ell}.$
To achieve this, we proceed by induction on $\ell$. Assume that $\ell=1$. Since
$L^s=(I\cap \mathfrak{q})^s + \sum_{i=1}^{s} (hI)^{i}(I\cap \mathfrak{q})^{s-i}$,
we can conclude the following equalities
\begin{align*}
(L^s: h)&=((I\cap \mathfrak{q})^s:h) + \sum_{i=1}^{s} ((hI)^{i}(I\cap \mathfrak{q})^{s-i}: h) \\
&=(I\cap \mathfrak{q})^s + I\sum_{i=0}^{s-1} (hI)^{i}(I\cap \mathfrak{q})^{s-1-i} \\
&=(I\cap \mathfrak{q})^s + IL^{s-1}.
\end{align*}
Hence, the assertion is true for the case in which $\ell=1$. Now, suppose, inductively, that $\ell >1$ and that the result has been shown for all values less than $\ell$. In particular, we deduce from the inductive hypothesis that
\begin{equation}\label{eq33}
(L^s:h^{\ell-1})=I^{\ell-2}(I\cap \mathfrak{q})^{s-\ell+2}+I^{\ell-1}L^{s-\ell+1}.
\end{equation}
Note that a similar argument yields the following equality
\begin{equation}\label{eq44}
(L^{s-\ell+1}:h)=(I\cap \mathfrak{q})^{s-\ell+1} + IL^{s-\ell}.
\end{equation}
By using (\ref{eq33}) and (\ref{eq44}), we get the following equalities
\begin{align*}
(L^s:h^{\ell}) &=(I^{\ell-2}(I\cap \mathfrak{q})^{s-\ell+2}+I^{\ell-1}L^{s-\ell+1}:h)\\
&=I^{\ell-2}(I\cap \mathfrak{q})^{s-\ell+2} + I^{\ell-1}((I\cap \mathfrak{q})^{s-\ell+1} + IL^{s-\ell})\\
&=I^{\ell-2}(I\cap \mathfrak{q})^{s-\ell+2} + I^{\ell-1}(I\cap \mathfrak{q})^{s-\ell+1} + I^{\ell}L^{s-\ell} \\
&=I^{\ell-1}(I\cap \mathfrak{q})^{s-\ell+1}+I^{\ell}L^{s-\ell}.
\end{align*}
This completes the inductive step, and so the claim has been shown by induction. In particular, $(L^s:h^s)=I^s$. Moreover, we have the following equalities
$$((L^s:h^{\ell}),h)=I^{\ell-1}(I\cap \mathfrak{q})^{s-\ell+1}+I^{\ell}L^{s-\ell} + hR=I^{\ell}(I\cap \mathfrak{q})^{s-\ell} + hR,$$
for all $1 \leq \ell \leq s$. In what follows, our aim is to verify this claim that, for all $1 \leq \ell \leq s$, we have
$\mathfrak{p}\in \mathrm{Ass}(L^s:h^{\ell}).$ To establish the assertion for $\ell=1$, we consider the following short exact sequence, where
$\psi(\overline{t})= h\overline{t}$ and $\phi(\overline{t}) =\overline{t}$,
\begin{equation}\label{eq55}
0\longrightarrow R/(L^s:h)\stackrel{\psi}{\longrightarrow} R/L^s \stackrel{\phi}{\longrightarrow} R/(L^s,h) \longrightarrow 0.
\end{equation}
We thus deduce from (\ref{eq55}) that $\mathrm{Ass}(L^s) \subseteq \mathrm{Ass}(L^s:h) \cup \mathrm{Ass}(L^s,h).$
This gives rise to $\mathfrak{p} \in \mathrm{Ass}(L^s:h)$ or $\mathfrak{p} \in \mathrm{Ass}(L^s,h).$
Assume that $\mathfrak{p} \in \mathrm{Ass}(L^s,h).$
Since $L^s +hR= (I \cap \mathfrak{q})^s + hR$, It follows immediately from \cite[Theorem 3.3]{SNQ} that
$\mathfrak{q}'\in \mathrm{Ass}(I \cap \mathfrak{q})^s.$ Due to $I \cap \mathfrak{q}$ is normally torsion-free, this gives that
$\mathfrak{q}'\in \mathrm{Min}(I \cap \mathfrak{q}).$ Because $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \mathfrak{q}$ is a minimal primary decomposition and $\mathfrak{q} \subsetneq \mathfrak{q}'$, we get a contradiction. Hence, we obtain $\mathfrak{p} \notin \mathrm{Ass}(L^s,h)$, and so
$\mathfrak{p} \in \mathrm{Ass}(L^s:h)$. Consequently, the claim holds for $\ell=1$.
Now, suppose, inductively, that $\ell >1$ and that the claim has been proven for all values less than $\ell$.
We now consider the following short exact sequence,
\begin{equation}\label{eq66}
0\longrightarrow R/(L^s:h^{\ell})\stackrel{\psi}{\longrightarrow} R/(L^s:h^{\ell-1}) \stackrel{\phi}{\longrightarrow}
R/((L^s:h^{\ell-1}),h) \longrightarrow 0.
\end{equation}
It follows rapidly from the inductive hypothesis that $\mathfrak{p}\in \mathrm{Ass}(L^s:h^{\ell-1}).$
By (\ref{eq66}), one has $\mathrm{Ass}(L^s:h^{\ell-1}) \subseteq \mathrm{Ass}(L^s:h^{\ell}) \cup
\mathrm{Ass}((L^s:h^{\ell-1}),h)$, and so $\mathfrak{p}\in \mathrm{Ass}(L^s:h^{\ell})$ or $\mathfrak{p}\in \mathrm{Ass}((L^s:h^{\ell-1}),h)$. Let $\mathfrak{p}\in \mathrm{Ass}((L^s:h^{\ell-1}),h)$. Thanks to $((L^s:h^{\ell-1}),h)=I^{\ell-1}(I \cap \mathfrak{q})^{s-\ell+1}+hR$, \cite[Theorem 3.3]{SNQ} implies that $\mathfrak{q}'\in \mathrm{Ass}(I^{\ell-1}(I \cap \mathfrak{q})^{s-\ell+1}).$
According to Proposition \ref{PRODUCT}, we can derive that $\mathfrak{q}'\in \mathrm{Min}(I) \cup \{\mathfrak{q}\}$.
Since $\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p} \cap \mathfrak{q}$ is a minimal primary decomposition and $\mathfrak{q} \subsetneq \mathfrak{q}'$, this leads to a contradiction. We thus get $\mathfrak{p}\in \mathrm{Ass}(L^s:h^{\ell})$. This completes the inductive step, and hence the claim has been proven by induction. In particular, we have $\mathfrak{p}\in \mathrm{Ass}(L^s:h^{s})$, and so $\mathfrak{p}\in \mathrm{Ass}(I^s)$. Thanks to $I$ is normally torsion-free, this implies that $\mathfrak{p}\in \mathrm{Min}(I)$, a contradiction.
Consequently, $L$ is normally torsion-free, as desired.
\end{proof}
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
The theorem below has been shown in \cite[Theorem 2.3]{NQ} by combinatorial tools, but, as an application of Lemma \ref{NTF1}, we re-prove it.
To formulate Theorem \ref{DI-TREES}, we recall here the required definitions. Let $G$ be a simple finite undirected graph.
For each vertex $v \in V(G)$, the {\em closed neighborhood} of $v$ in $G$ is defined as follows:
\[
N_G[v]=\{u \in V(G) : \{u,v\} \in E(G)\} \cup \{v\}.
\]
A subset $S\subseteq V(G)$ is called a {\em dominating set} of $G$ if $S \cap N_G[v] \neq \emptyset$ for all $v\in V(G)$. A dominating set is called {\it minimal} if it does not properly contain any other dominating set of $G$. A minimum dominating set of $G$ is a minimal dominating set with the smallest size. The {\em dominating number} of $G$, denoted by $\gamma(G)$, is the size of its minimum dominating set, that is,
\[
\gamma(G)=\mathrm{min} \{|S|: S \text{ is a minimal dominating set
of } G\}.
\]
The {\em closed neighborhood ideal} of $G$ has been introduced as
\[
NI(G)= (\prod_{x_j \in N[x_i]} x_j: x_i \in V(G)),
\]
and the {\em dominating ideal} of $G$ is defined as
\[
DI(G)=(\prod_{x_i \in S}x_i: S \text{ is a minimal dominating set of } G).
\]
It is shown in \cite[Lemma 2.2]{SM} that $DI(G)$ is the Alexander dual of $NI(G)$, that is, $DI(G)=NI(G)^{\vee}$.
In order to find more properties of closed neighborhood ideals and dominating ideals, the reader may refer to
\cite{NBR, NQBM, SM}.
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
\begin{theorem}\label{DI-TREES}
The dominating ideals of trees are normally torsion-free.
\end{theorem}
\begin{proof}
Suppose that $T=(V(T), E(T))$ is a tree. We use induction on $n:=|E(T)|$. By \cite[Lemma 3.12]{SN} and the fact that every prime monomial ideal is normally torsion-free, one can establish the claim for any $n\leq 3$. Now, suppose, inductively, that $n>3$ and that the result has been shown for all trees whose edge sets have cardinalities less than $n$. Assume that $T=(V(T), E(T))$ is a tree with $n=|E(T)|$. Using the fact that $T$ is a tree and $n>3$, there exists at least one leaf such as $w$ such that $w$ has a neighbor such as $v$ with either $\mathrm{deg}_T(v)=2$ or $v$ has another leaf in its neighbor set, see for example \cite[Proposition 3.9]{KHN3}. Since $w\in N_T(v)$, we can view the tree $T$ as the union of a tree such as $T_1$ and the edge $\{v,w\}$, where $v$ is the cut-point of this union. Note that, based on the inductive hypothesis, $DI(T_1)$ is normally torsion-free.
Here, one may consider the following three cases:
\vspace{1mm}
\textbf{Case 1.} $v$ has another leaf in its neighbor set. Hence, $\prod_{j\in N_{T_1}[v]}x_j \notin \mathcal{G}(NI(T_1))$.
On account of \cite[Lemma 2.2]{SM} and the fact that $v\in N_{T_1}[v]$, we obtain the following equalities
$$DI(T)=DI(T_1) \cap (x_w,x_j~:~ j\in N_{T_1}[v])R \cap (x_w, x_v)R=DI(T_1) \cap (x_w, x_v)R,$$
where $R=K[x_{\alpha}~:~ \alpha \in V(T)]$. Since $x_w \notin \mathrm{supp}(DI(T_1))$, we can derive from \cite[Lemma 2.17]{NQBM}
that $DI(T)$ is normally torsion-free.
\vspace{1mm}
\textbf{Case 2.} $\mathrm{deg}_T(v)=2$ and $\prod_{j\in N_{T}[c]}x_j \notin \mathcal{G}(NI(T))$, where $N_{T}(v)=\{c\}$.
This means that there is a leaf in the neighbor set of $c$. Let $L$ be a square-free monomial ideal such that $\mathcal{G}(L)= \mathcal{G}(NI(T_1))\setminus \{x_cx_v \}$. We thus get
$$DI(T)=L^\vee \cap (x_w,x_c,x_v)R \cap (x_w,x_v)R=L^\vee \cap (x_w,x_v)R,$$ where $R=K[x_{\alpha}~:~ \alpha \in V(T)]$ and
$L^{\vee}$ denotes the Alexander dual of $L$. Thanks to $L^{\vee}=DI(T_1)/x_v$ and $DI(T_1)$ is normally torsion-free, it follows at once from
\cite[Theorem 3.19]{SN} that $L^{\vee}$ is normally torsion-free. As $x_v, x_w\notin \mathrm{supp}(L^{\vee})$, \cite[Proposition 3.3]{NQ}
gives that $DI(T)$ is normally torsion-free.
\vspace{1mm}
\textbf{Case 3.} $\mathrm{deg}_T(v)=2$ and $\prod_{j\in N_{T}[c]}x_j \in \mathcal{G}(NI(T))$, where $N_{T_1}(v)=\{c\}$.
Notice that it follows readily from $N_{T_1}(v)=\{c\}$ that one can view the tree $T_1$ as the union of a tree such as $T_2$ and the edge $\{c,v\}$, where $c$ is the cut-point of this union.
Assume that $L_1$ and $L_2$ are two square-free monomial ideals such that
$\mathcal{G}(L_1)=\mathcal{G}(NI(T_1))\setminus \{x_cx_v \}\; \; \text{and}\; \; \mathcal{G}(L_2)=\mathcal{G}(L_1)\cup \{\prod_{j\in N_{T_1}[c]}x_j\}.$
In view of \cite[Lemma 2.2]{SM}, one gains the following equalities
$$DI(T_1)=L^{\vee}_1\cap (x_c, x_v)\; \; \text{and}\;\; L^{\vee}_2=L^{\vee}_1 \cap (x_j~:~ j\in N_{T_1}[c]),$$
where $L^{\vee}_1$ (respectively, $L^{\vee}_2$) denotes the Alexander dual of $L_1$ (respectively, $L_2$).
It is not hard to check that $L^{\vee}_1=DI(T_1)/x_v$. Due to $DI(T_1)$ is normally torsion-free, one can deduce from \cite[Theorem 3.19]{SN} that
$L^{\vee}_1$ is normally torsion-free. Put $\mathfrak{q}:=(x_j~:~ j\in N_{T_2}[c])$. Then we have
$L^{\vee}_2=L^{\vee}_1 \cap (\mathfrak{q}, x_v).$
On account of $DI(T_2)$ is normally torsion-free and $DI(T_2)=L^{\vee}_1\cap \mathfrak{q}$, this implies that $L^{\vee}_1\cap \mathfrak{q}$
is normally torsion-free. Notice that $x_v \notin \mathrm{supp}(\mathfrak{q}) \cup \mathrm{supp}(L^{\vee}_1)$.
By virtue of Lemma \ref{NTF1}, one can immediately deduce that $L^{\vee}_2$ is normally torsion-free. Once again, \cite[Lemma 2.2]{SM} yields that
$$DI(T)=L^{\vee}_2\cap (x_w, x_j~:~ j\in N_{T_1}[v])R \cap (x_w,x_v)R=L^{\vee}_2 \cap (x_w,x_v)R,$$ where $R=K[x_{\alpha}~:~ \alpha \in V(T)]$.
Due to $x_w \notin \mathrm{supp}(L^{\vee}_2)$ and $L^{\vee}_2$ is normally torsion-free, one can immediately conclude from
\cite[Lemma 2.17]{NQBM} that $DI(T)$ is normally torsion-free.
This completes the inductive step, and hence the claim has been proven by induction.
\end{proof}
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
\section{Some results on the embedded associated primes}
This section has two folds. We first give some results on the embedded associated prime ideals of powers of square-free monomial ideals. Next, we turn our attention to study when a square-free monomial ideal is minimally not normally torsion-free. Particularly, we introduce a class of square-free monomial ideals, which are minimally not normally torsion-free.
To understand Lemma \ref{Ass-maximal1} and Proposition \ref{MNNT}, one needs to recollect some definitions. Assume that $I$ is a square-free monomial ideal and $\Gamma \subseteq \mathcal{G}(I)$, where $\mathcal{G}(I)$ denotes the unique minimal set of monomial generators of the monomial ideal $I$. We say that $\Gamma$ is an {\it independent} set in $I$ if $\mathrm{gcd}(f,g)=1$ for each $f,g\in \Gamma$ with $f\neq g$. We denote the maximum cardinality of an independent set in $I$ by $\beta_1(I)$.
Furthermore, for $1\leq j \leq n$, the { \it contraction} of $I$ at $x_j$, denoted by $I/x_j$, is obtained by setting $x_j=1$ in each $u\in \mathcal{G}(I)$. In addition, the { \it deletion} of $I$ at $x_j$, denoted by $I\setminus x_j$, is formed by putting $x_j=0$ in each $u\in \mathcal{G}(I)$, refer to \cite[page 303]{HM}.\par
We are ready to give the second main result of this paper in the following lemma, which is an affirmative answer to \cite[Question 3.11]{NQ1}.
\begin{lemma} \label{Ass-maximal1}
Let $I \subset R=K[x_1, \ldots, x_n]$ be a square-free monomial ideal and $\mathfrak{q}$ be a prime monomial ideal in $R$. Let $L:=I \cap (\mathfrak{q}, x_r)$ such that $x_r\notin \mathrm{supp}(\mathfrak{q}) \cup \mathrm{supp}(I)$ and $\mathrm{Ass}(L)=\mathrm{Ass}(I) \cup \{(\mathfrak{q}, x_r)\}$. If $\mathfrak{p}=(L^s:x_r^{\theta}v)$ for some $s\geq 2$, $\theta \geq 0$, and monomial $v$ with
$x_r\nmid v$, then $\mathfrak{p}\setminus x_r \in \mathrm{Ass}(I^{\theta}(I\cap \mathfrak{q})^{s-\theta})$.
\end{lemma}
\begin{proof}
Let $\mathfrak{p}=(L^s:x_r^{\theta}v)$ for some $s\geq 2$, $\theta \geq 0$, and monomial $v$ with
$x_r\nmid v$.
Since $x_r^{\theta}v\notin L^s$ and $L^s=\sum_{i=0}^{s} (x_rI)^{i}(I\cap \mathfrak{q})^{s-i}$, this gives that $x_r^{\theta}v \notin (x_rI)^{i} (I\cap \mathfrak{q})^{s-i}$ for all $i=0, \ldots, s$. Thanks to $L$ is a square-free monomial ideal, one can deduce from \cite[Proposition 4.8]{SNQ} that $\theta \leq s-1$. In particular, we have $x_r^{\theta}v \notin (x_rI)^{\theta}(I\cap \mathfrak{q})^{s-\theta}$, and so
$v \notin I^{\theta}(I\cap \mathfrak{q})^{s-\theta}$. Now, let $x_t$ be an arbitrary element in $\mathfrak{p}\setminus x_r$. It follows from
$\mathfrak{p}=(L^s: x_r^{\theta}v)$ that $x_tx_r^{\theta}v\in L^s$, and so $x_tv\in (L^s:x_r^{\theta})$. Because $x_r \nmid x_tv$, this yields that
$x_tv\in (L^s:x_r^{\theta})\setminus x_r$. By remembering the following chain
\begin{equation}\label{eq77}
(I\cap \mathfrak{q})^s \subseteq I(I\cap \mathfrak{q})^{s-1} \subseteq I^2(I\cap \mathfrak{q})^{s-2} \subseteq \cdots \subseteq I^{s-1}(I\cap \mathfrak{q}) \subseteq I^s,
\end{equation}
we get $x_tv\in I^{\theta}(I\cap \mathfrak{q})^{s-\theta}$. We can conclude from $v\notin I^{\theta}(I\cap \mathfrak{q})^{s-\theta}$ and
$x_tv \in I^{\theta}(I\cap \mathfrak{q})^{s-\theta}$ for all $x_t \in \mathfrak{p}\setminus x_r$ that
$(I^{\theta}(I\cap \mathfrak{q})^{s-\theta}:v)=\mathfrak{p}\setminus x_r$. This gives rise to $\mathfrak{p}\setminus x_r \in \mathrm{Ass}(I^{\theta}(I\cap \mathfrak{q})^{s-\theta})$, and the proof is complete.
\end{proof}
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
As an immediate consequence of Lemma \ref{Ass-maximal1}, we get the corollary below.
\begin{corollary} \label{Cor.Ass-maximal1}
Let $I \subset R=K[x_1, \ldots, x_n]$ be a square-free monomial ideal and $\mathfrak{q}$ be a prime monomial ideal in $R$. Let $L:=I \cap (\mathfrak{q}, x_r)$ such that $x_r\notin \mathrm{supp}(\mathfrak{q}) \cup \mathrm{supp}(I)$ and $\mathrm{Ass}(L)=\mathrm{Ass}(I) \cup \{(\mathfrak{q}, x_r)\}$. If $\mathfrak{p}=(x_1, \ldots, x_n)=(L^s:v)$ for some $s\geq 2$ with
$x_r\nmid v$, then $\mathfrak{p}\setminus x_r \in \mathrm{Ass}((I\cap \mathfrak{q})^s)$.
\end{corollary}
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
\begin{example}\label{Exam.1}
Let $R=K[x_1, \ldots, x_8]$, $\mathfrak{q}:=(x_1,x_3,x_7)$, $x_r:=x_8$, and
\begin{align*}
I =&(x_1,x_4,x_7)\cap (x_1,x_4,x_2)\cap (x_2,x_4,x_6)\cap (x_2,x_6,x_5)\cap (x_3,x_5,x_6)\\
&\cap (x_5,x_7,x_3)\\
=&(x_1x_2x_3,x_2x_3x_4,x_1x_2x_5,x_4x_5,x_1x_3x_6,x_3x_4x_6,x_1x_5x_6,x_2x_3x_7,\\
& x_2x_5x_7,x_1x_6x_7, x_2x_6x_7,x_4x_6x_7).
\end{align*}
Let $L:=I\cap (\mathfrak{q}, x_8)$. Using {\it Macaulay2} \cite{GS}, we obtain $(L^2: v)=(x_1, \ldots, x_8)$, where
$v=x_2x_3x_4x_5x_7$. Since $x_8\nmid v$, it follows immediately from Corollary \ref{Cor.Ass-maximal1} that
$(x_1, \ldots, x_7)\in \mathrm{Ass}((I\cap \mathfrak{q})^2)$.
\end{example}
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
In the following result we are going to introduce a class of square-free monomial ideals which are minimally not normally torsion-free.
An ideal $I'$ is a {\it minor} of a square-free monomial ideal $I$ if $I'$ can be obtained from $I$ by a sequence of deletions and contractions in any order, see \cite[Definition 6.5.3]{V1}. Recall that a square-free monomial ideal $I \subset R=K[x_1, \ldots, x_n]$ is said to be {\it minimally not normally torsion-free} when $I$ is not normally torsion-free but all its proper minors are normally torsion-free, refer to \cite{HM}.
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
\begin{proposition}\label{MNNT}
Let $I\subset R=K[x_1, \ldots, x_n]$ be a square-free monomial ideal and $\mathfrak{m}=(x_1, \ldots, x_n)$
such that there exists some $ m\geq 1$ with $\mathrm{Ass}(I^s)=\mathrm{Min}(I)$ for all $1\leq s \leq m$, and $\mathrm{Ass}(I^s)= \mathrm{Min}(I) {\cup} \{\mathfrak{m}\}$ for all $s\geq m+1$ such that $\mathfrak{m}\setminus x_j \notin I^s\setminus x_j$ for all $j=1, \ldots, n$.
Then $I$ is minimally not normally torsion-free. In particular, we have $m\geq \beta_1(I)$, where $\beta_1(I)$ denotes the maximum cardinality of an independent set in $I$.
\end{proposition}
\begin{proof}
Since $\mathfrak{m}\in \mathrm{Ass}(I^s)$ for all $s\geq m+1$, this implies that $I$ is not normally torsion-free. To complete the proof, we have to show that $I/x_j$ and $I\setminus x_j$ are normally torsion-free for all $1\leq j \leq n$. To do this, fix $1\leq j \leq n$. We first verify that $I/x_j$ is normally torsion-free. To see this, we need to show that $\mathrm{Ass}((I/x_j)^k) \subseteq \mathrm{Ass}(I/x_j)$ for all $k\geq 1$. Fix $k \geq 1$.
By \cite[Lemma 4.10(viii)]{RNA}, we have
$\mathrm{Ass}((I/x_j)^k)=\mathrm{Ass}(I^k/x_j)=\{\mathfrak{q} : \mathfrak{q} \in \mathrm{Ass}_{R}(R/I^k)~\mathrm{and}~ x_j\notin\mathfrak{q}\}.$ Choose an arbitrary element $\mathfrak{q}\in \mathrm{Ass}((I/x_j)^k)$. We thus get
$\mathfrak{q} \in \mathrm{Ass}_{R}(R/I^k)$ and $x_j\notin\mathfrak{q}$. Due to $x_j\notin\mathfrak{q}$, this yields that $\mathfrak{q}\neq \mathfrak{m}$, and so $\mathfrak{q}\in \mathrm{Min}(I)$. Accordingly, we have $\mathfrak{q}\in \mathrm{Ass}(I)$, and therefore
$\mathfrak{q}\in \mathrm{Ass}(I/x_j)$. This gives rise to $I/x_j$ is normally torsion-free.
To finish the proof, we demonstrate that $I\setminus x_j$ is normally torsion-free. To accomplish this, we must establish
$\mathrm{Ass}((I\setminus x_j)^k) \subseteq \mathrm{Ass}(I \setminus x_j)$ for all $k\geq 1$. Fix $k \geq 1$.
One can derive from \cite[Definition 4.3.22 and Exercise 6.1.25]{V1} that $I^{(k)}=\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p}^k$, and so
$I^k=\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p}^k$ when $1\leq k\leq m$, and
$I^k=\bigcap_{\mathfrak{p}\in \mathrm{Ass}(I)}\mathfrak{p}^k \cap Q$ with $\sqrt{Q}=\mathfrak{m}$ when $k\geq m+1$. Let
$\mathrm{Ass}(I)=\{\mathfrak{p}_1, \ldots, \mathfrak{p}_z\}$.
From \cite[Lemma 4.12(vi)]{RNA}, we have a primary decomposition of $I^k\setminus x_j$ as follows:
$$I^k\setminus x_j = \bigcap_{i=1}^z(\mathfrak{p}_i\setminus x_j)^k \; \; \text{or} \; \; I^k\setminus x_j = \bigcap_{i=1}^z(\mathfrak{p}_i\setminus x_j)^k \cap (Q\setminus x_j).$$
Let $\mathrm{Ass}(I\setminus x_j)=\mathrm{Min}(I\setminus x_j)=\{{\mathfrak{p}}_{i_1}\setminus x_j, \ldots, {\mathfrak{p}}_{i_h}\setminus x_j\},$ where $\{i_1, \ldots, i_h\}\subseteq \{1, \ldots, z\}$.
We note if $I^k\setminus x_j = \bigcap_{i=1}^z(\mathfrak{p}_i\setminus x_j)^k \cap (Q\setminus x_j),$
then we have $\bigcap_{t=1}^h(\mathfrak{p}_{i_t}\setminus x_j)^k \subseteq Q\setminus x_j$.
Indeed, if it were $\bigcap_{t=1}^h(\mathfrak{p}_{i_t}\setminus x_j)^k \nsubseteq Q\setminus x_j$, then we can derive
that $\mathfrak{m}\setminus x_j \in \mathrm{Ass}(I^k\setminus x_j)$.
This contradicts our assumption. We thus get $(I\setminus x_j)^k=({\mathfrak{p}}_{i_1}\setminus x_j)^k \cap \cdots \cap ({\mathfrak{p}}_{i_h}\setminus x_j)^k$ is a minimal primary decomposition of $(I\setminus x_j)^k$.
As $\mathrm{Min}(I\setminus x_j)=\{{\mathfrak{p}}_{i_1}\setminus x_j, \ldots, {\mathfrak{p}}_{i_h}\setminus x_j\},$ it follows from
\cite[Exercise 6.1.25]{V1} that $(I\setminus x_j)^{(k)}=({\mathfrak{p}}_{i_1}\setminus x_j)^k \cap \cdots \cap ({\mathfrak{p}}_{i_h}\setminus x_j)^k.$
This leads to $(I\setminus x_j)^{(k)}=(I\setminus x_j)^k$. We can finally deduce from \cite[Proposition 4.3.29]{V1} that
$I\setminus x_j$ is normally torsion-free. The last assertion can be deduced from \cite[Corollary 3.5]{SNQ}, and the proof is over.
\end{proof}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
We close this paper with the following example. In fact, by employing Proposition \ref{MNNT}, we are going to introduce a class of square-free monomial ideals, which are minimally not normally torsion-free.
\begin{example} \label{Exam.MNNT}
Let $C_{2n+1}=(V(C_{2n+1}), E(C_{2n+1}))$ be the odd cycle graph with $V(C_{2n+1})=\{x_1, \ldots, x_{2n+1}\}$ and
$$E(C_{2n+1})=\{\{x_i, x_{i+1}\} : i=1, \ldots, 2n\} \cup \{\{x_1, x_{2n+1}\}\}.$$ Also, let $J(C_{2n+1})$ denote the cover ideal of $C_{2n+1}$ in
the polynomial ring $R=K[x_1, \ldots, x_{2n+1}]$ over a field $K$ and $\mathfrak{m}=(x_1, \ldots, x_{2n+1})$ be the maximal ideal. By \cite[Proposition 3.6]{NKA}, we know that
$\mathrm{Ass}(J(C_{2n+1}))=\mathrm{Min}(J(C_{2n+1}))$ and $\mathrm{Ass}(J(C_{2n+1})^s)=\mathrm{Min}(J(C_{2n+1}))\cup \{\mathfrak{m}\}$ for all $s\geq 2$. It is easy to see that for each $1\leq j \leq 2n+1$ with $x_0=x_{2n+1}$ and $x_{2n+2}=x_1$, we have
$J(C_{2n+1})\setminus x_j=(x_{j-1}) \cap (x_{j+1}) \cap J(P)$, where $P$ is a path with $x_{j-1}, x_{j+1}\notin V(P)$. From
\cite[Corollary 2.6]{GRV}, $J(P)$ is normally torsion-free, and in view of \cite[Proposition 3.3]{NQ}, we dedude that $J(C_{2n+1})\setminus x_j$ is normally torsion-free, and so $\mathfrak{m}\setminus x_j \notin \mathrm{Ass}(J(C_{2n+1})^s\setminus x_j)$ for all $s$. It follows now from Proposition \ref{MNNT} that $J(C_{2n+1})$ is minimally not normally torsion-free. In particular, we have $\beta_1(J(C_{2n+1}))=1$.
\end{example}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section*{ORCID}
Mehrdad ~Nasernejad: \textsc{https://orcid.org/0000-0003-1073-1934}\\
Veronica Crispin Qui$\mathrm{\tilde{n}}$onez: \textsc{https://orcid.org/0000-0002-2146-7051}\\
Winfried Hochstättler: \textsc{https://orcid.org/0000-0001-7344-7143} \\
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section*{The conflict of interest and data availability statement}
We hereby declare that this manuscript has no associated data and also there is no conflict of interest in this manuscript.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section*{Acknowledgment.}
%First, the authors are deeply grateful to the anonymous referee for careful reading of the manuscript, and for his/her valuable suggestions which led to %several improvements in the quality of this paper.
This paper was prepared when Mehrdad Nasernejad visited the Department of Mathematics of
Uppsala University in 2023. In particular, Mehrdad Nasernejad would like to thank Uppsala
University for its hospitality.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{thebibliography}{999}
\bibitem{BR} M. Brodmann, \textit{Asymptotic stability of $\mathrm{Ass}(M/I^{n}M$).} Proc. Amer. Math. Soc.
\textbf{74} (1979), 16--18.
\bibitem{GRV} I. Gitler, E. Reyes, and R. H. Villarreal, {\it Blowup algebras of ideals of vertex covers of bipartite graphs,} Contemp. Math. {\bf 376} (2005), 273-279.
\bibitem{GS} D. R. Grayson and M. E. Stillman, Macaulay2, a software system for research in algebraic geometry,
\url{http://www.math.uiuc.edu/ Macaulay2/}.
\bibitem{HM} H. T. H\`a and S. Morey, {\it Embedded associated primes of powers of square-free monomial ideals,} J. Pure Appl. Algebra {\bf 214} (2010), 301--308.
\bibitem{HRV} J. Herzog, A. Rauf, and M. Vladoiu, {\it The stable set of associated prime ideals of a polymatroidal ideal,} J. Algebraic Combin. {\bf 37} (2013), 289--312.
\bibitem{KHN1} K. Khashyarmanesh and M. Nasernejad, \textit{ A note on the Alexander dual of path ideals of rooted trees,} Comm. Algebra {\bf 46} (2018), 283--289.
\bibitem{KHN3} K. Khashyarmanesh and M. Nasernejad, {\it On the minimal path cover sets of union of certain graphs}, Quaest. Math. {\bf 43(7)} (2020), 881--901.
\bibitem{MNQ} A. Musapa\c{s}ao\u{g}lu, M. Nasernejad, and A. A. Qureshi, {\it The edge ideals of $\textbf{t}$-spread $d$-partite hypergraphs}, Collectanea Mathematica, 2023, \doi{10.1007/s13348-023-00410-y}.
\bibitem{NBR} M. Nasernejad, S. Bandari, and L. G. Roberts, \textit{Normality and associated primes of closed neighborhood ideals and dominating ideals},
Journal of Algebra and Its Applications, 2023, \doi{10.1142/S0219498825500094}.
\bibitem{N3} M. Nasernejad and K. Khashyarmanesh, \textit{ On the Alexander dual of the path ideals of rooted and unrooted trees,} Comm. Algebra
{\bf 45} (2017), 1853--1864.
\bibitem{NKA} M. Nasernejad, K. Khashyarmanesh, and I. Al-Ayyoub, \textit{ Associated primes of powers of cover ideals under graph operations,} Comm. Algebra, {\bf 47} (2019), No. 5, 1985--1996.
\bibitem{NQ1} M. Nasernejad, V. C. Qui$\mathrm{\tilde{n}}$onez, and J. Toledo, \textit{ Normally torsion-freeness and normality criteria for monomial ideals}, Submitted.
\bibitem{NQ} M. Nasernejad and A. A. Qureshi, \textit{Algebraic implications of neighborhood hypergraphs and their transversal hypergraphs}, 2023, To appear in Communications in Algebra, \url{ArXiv:2208.14040v1}.
\bibitem{NQBM} M. Nasernejad, A. A. Qureshi, S. Bandari, and A. Musapa\c{s}ao\u{g}lu, \textit{Dominating ideals and closed neighborhood ideals of graphs}, Mediterr. J. Math. {\bf 19} (2022), no. 4, Paper No. 152, 18 pp.
\bibitem{NQKR} M. Nasernejad, A. A. Qureshi, K. Khashyarmanesh, and L. G. Roberts, \textit{Classes of normally and nearly normally torsion-free monomial ideals}, Comm. Algebra, {\bf 50(9)} (2022), 3715--3733, \doi{10.1080/00927872.2022.2042546}
\bibitem{OL} A. Olteanu, {\it Normally torsion-free lexsegment ideals,} Algebra Colloq. {\bf 22} (2015), 23--34
\bibitem{RNA} S. Rajaee, M. Nasernejad, and I. Al-Ayyoub, \textit{Superficial ideals for monomial ideals,} J. Algebra Appl. {\bf 16}(2) (2018) 1850102 (28 pages).
\bibitem{SN} M. Sayedsadeghi and M. Nasernejad, \textit{Normally torsion-freeness of monomial ideals under monomial operators,} Comm. Algebra {\bf 46}(12) (2018), 5447--5459.
\bibitem{SNQ} M. Sayedsadeghi, M. Nasernejad, and A. A. Qureshi, \textit{On the embedded associated primes of monomial ideals,} Rocky Mountain J. Math. {\bf 52(1)} (2022), 275-287. \doi{10.1216/rmj.2022.52.275}
\bibitem{SM} L. Sharifan and S. Moradi, \textit{Closed neighborhood ideal of a graph}, Rocky Mountain J. Math. {\bf 50}(3) (2020), 1097--1107.
\bibitem{SVV} A. Simis, W. Vasconcelos, and R. Villarreal, {\it On the ideal theory of graphs,} J. Algebra {\bf 167} (1994), 389--416.
\bibitem{V1} R. H. Villarreal, \textit{Monomial Algebras.} 2nd. Edition, Monographs and Research Notes in Mathematics, CRC Press, Boca Raton, FL, 2015.
\bibitem{West} D. West, \textit{Introduction to Graph Theory.} 2nd. Edition, Prentice-Hall, Upper Saddle River, 2001.
\end{thebibliography}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% ------------------------------------------------------------------------
\end{document}
% ------------------------------------------------------------------------