1 Introduction
Let
$a,b,$
and n be positive integers. For any integers x and y, we denote by
$(x, y)$
the greatest common divisor of integers x and y and by
$[x,y]$
their least common multiple. Let
$\mathbb Z$
denote the ring of integers and let
$|T|$
stand for the cardinality of a finite set T of integers. Let f be an arithmetic function and let
$S=\{x_1, \ldots , x_n \}$
. Let
$(f(S))$
and
$(f[S])$
denote the
$n\times n$
matrices whose
$(i,j)$
-entries are
$f((x_i, x_j))$
and
$f([x_i, x_j])$
, respectively. Define
$\langle n\rangle :=\{1,\ldots , n\}$
. For any real number e, let
$\xi _e$
be the arithmetic function defined for any positive integer x by
$\xi _e(x)=x^e$
. The
$n\times n$
matrices
$(\xi _a(x_i, x_j)) =((x_i, x_j)^a)$
(abbreviated by
$(S^a)$
) and
$(\xi _a[x_i, x_j])=([x_i, x_j]^a)$
(abbreviated by
$[S^a]$
) are called the ath power GCD matrix on S and the ath power LCM matrix on S, respectively. The set S is said to be factor closed (written as FC for short) if
$(x\in S$
,
$d>0$
,
$d|x)\Rightarrow d\in S$
. We say that S is gcd closed if S contains
$(x_i, x_j)$
for all integers i and j with
$1\le i,j\le n$
. Obviously, any FC set contains 1 and is gcd closed but the converse is not true. For instance, if the set S is FC, then for any integer
$x>1$
, the set
$xS:=\{xy| y\in S\}$
is gcd closed but not FC since
$1\not \in xS$
. One hundred and fifty years ago, Smith [Reference Smith26] proved that
if S is FC, where
$\varphi $
is the Euler’s totient function and
$\pi $
is the multiplicative function defined for the prime power
$p^r$
by
$\pi (p^r)=-p$
. After that, many generalizations of Smith’s determinant (1.1) and related results were published (see, e.g., [Reference Apostol1–Reference McCarthy25] and [Reference Tan and Li27–Reference Zhu and Yu35]). In particular, an elegant result was gotten by Hong et al. [Reference Hong, Hu and Lin10] stating that for any integer
$n\ge 2$
, one has
where
$\mu $
is the M
$\ddot {\mathrm {o}}$
bius function and an integer
$x\ge 1$
is called square free if x is not divisible by the square of any prime number.
Hong [Reference Hong12] introduced the concept of greatest-type divisor when he solved the famous Bourque–Ligh conjecture [Reference Bourque and Ligh4]. For any integer
$x\in S$
, if
$(y<x, y|z|x \ \mathrm {and}\ y,z\in S)\Rightarrow z\in \{y,x\},$
then y is called a greatest-type divisor of x. Let
$G_S(x):=\{y\in S: y \ \text {is a greatest-type divisor of} \ x \ \mathrm {in} \ S\}$
. Let
$M_n({\mathbb Z})$
stand for the ring of
$n\times n$
matrices over the integers. In 1992, Bourque and Ligh [Reference Bourque and Ligh4] proved that
$(S)$
divides
$[S]$
in the ring
$M_n(\mathbb Z)$
if S is FC. Namely, there exists a
$B\in M_n({\mathbb Z})$
such that
$[S]=B(S)$
or
$[S]=(S)B$
. In 2002, Hong [Reference Hong14] showed that such a factorization is not true when S is gcd closed and
$\max _{x \in S}\{|{G_S(x)}|\}=2$
.
On the other hand, Hong [Reference Hong17] initiated the investigation of divisibility among power LCM matrices. It was proved in [Reference Hong17] that
$[S^a]|[S^b]$
if
$a|b$
and S is a divisor chain (i.e.,
$x_{\sigma (1)}|\dots |x_{\sigma (n)}$
for a permutation
$\sigma $
of
$\langle n\rangle $
). Evidently, a divisor chain is gcd closed but not conversely. Tan and Li [Reference Tan and Li27] generalized Hong’s result [Reference Hong17] by proving that
$[S^a]|[S^b]$
holds in the ring
$M_{|S|}({\mathbb Z})$
if
$a|b$
and S consists of finitely many coprime divisor chains with
$1\in S$
, and that such divisibility relation is not true if
$a\nmid b$
. Zhu and Li [Reference Zhu and Li33] confirmed the conjecture of Hong raised in [Reference Hong17] stating that if
$a|b$
and S is a gcd-closed set with
$\max _{x\in S}\{|G_S(x)|\}=1$
, then
$[S^a]\mid [S^b]$
in the ring
$M_n({\mathbb Z})$
. In 2026, Hong [Reference Hong18] established the same divisibility result when S is FC. As in [Reference Hong18], for any set S of positive integers and for any
$x\in S$
with
$|G_S(x)|\ge 2$
, we say that the two distinct greatest-type divisors
$y_1$
and
$y_2$
of x in S satisfy the condition
$\mathcal {G}$
if
$[y_1,y_2]=x$
and
$(y_1,y_2)\in G_S(y_1)\cap G_S(y_2)$
. We say that x satisfies the condition
$\mathcal {G}$
if any two distinct greatest-type divisors of x in S satisfy the condition
$\mathcal {G}$
. Moreover, we say that a set S of positive integers satisfies the condition
$\mathcal {G}$
if any element x in S satisfies that either
$|G_S(x)|\le 1$
, or
$|G_S(x)|\ge 2$
and x satisfies the condition
$\mathcal {G}$
. In [Reference Hong18], Hong showed that any FC set is a gcd-closed set satisfying the condition
$\mathcal {G}$
. Furthermore, Hong raised the following conjecture in Section 3 of [Reference Hong18].
Conjecture 1.1 [Reference Hong18, Conjecture 3.4]
Let a and b be positive integers with
$a|b$
and let S be a gcd-closed set satisfying the condition
$\mathcal G$
. Then
$[S^a]\mid [S^b]$
in the ring
$M_{|S|}(\mathbb Z)$
.
For the case
$\max _{x\in S}\{|G_S (x)|\}=1$
, by the theorem of Zhu and Li [Reference Zhu and Li33], we know that this conjecture is true. For the case
$\max _{x\in S}\{|G_S (x)|\}=2$
, by [Reference Wan and Zhu28], one knows that Conjecture 1.1 holds. Furthermore, Zhu et al. [Reference Zhu, Luo and Wan34] confirmed Conjecture 1.1 for the case
$\max _{x\in S}\{|G_S (x)|\}=3$
. But Conjecture 1.1 is still kept open when S is the general gcd-closed set satisfying the condition
$\mathcal {G}$
.
In this article, our main goal is to study the divisibility among power LCM matrices. We introduce a new method to investigate Conjecture 1.1. The main result of this article can be stated as follows.
Theorem 1.2 Let S be a gcd-closed set satisfying the condition
$\mathcal {G}$
and let a and b be positive integers such that
$a|b$
. Then the ath power LCM matrix
$[S^a]$
divides the bth power LCM matrix
$[S^b]$
in the ring
$M_{|S|}(\mathbb Z)$
.
This article is organized as follows. After introducing some necessary preliminary results, we proceed to the core of our argument. In fact, we first recall and show in Section 2 some results related to the arithmetic properties of gcd-closed sets that satisfy the condition
$\mathcal G$
and then present a formula for the inverse of power LCM matrices. Finally, in Section 3, by proving a key lemma, we present the proof of Theorem 1.2.
The divisibility of matrices under consideration is invariant under relabeling the elements of S. Specifically, for any permutation
$\sigma $
on
$\langle n\rangle $
, we have
$[S^a]\mid [S^b]$
if and only if
$[S^a_{\sigma }]\mid [S^b_{\sigma }]$
. Therefore, without loss of generality, we may assume throughout the rest of this article that the set
$S=\{x_1,\ldots ,x_n\}$
is ordered such that
$x_1<\dots <x_n$
.
2 Preliminary lemmas
Throughout this section, let
$x_m \in S$
with
$G_S(x_m)=\{x_{m_1}, \ldots , x_{m_s}\}$
,
$2\le m\le n=|S|$
and
$1\le s\le m-1$
. Let
$1\le k\le s$
and for any integers
$i_1,\ldots , i_k$
with
$1 \le i_1<\cdots <i_k \le s$
, we define
Then
$x_{m_{1\dots s}}=(x_{m_1}, \ldots , x_{m_s})$
. In what follows, we recall and show several results that are needed in the proof of our main result. For any
$j\in \langle s\rangle $
, define
$P_j:=\frac {x_m}{x_{m_j}}$
. We begin with a known result which is due to Zhu and Yu [Reference Zhu and Yu35] and is about the arithmetic property of the gcd-closed set S which satisfies the condition
$\mathcal {G}$
.
Lemma 2.1 [Reference Zhu and Yu35]
Let S be a gcd-closed set satisfying the condition
$\mathcal {G}$
and
$x_m \in S$
with
$G_S(x_m)=\{x_{m_1}, \ldots , x_{m_s}\}$
. Let k be an integer with
$1 \le k \le s$
. Then for any integers
$i_1,\ldots ,i_k$
with
$1\le i_1<\cdots <i_k\le s$
, we have
Lemma 2.2 [Reference Zhu32]
Let S be a gcd-closed set satisfying the condition
$\mathcal {G}$
and
$x_l, x_m\in S$
with
$G_S(x_m)=\{x_{m_1},\ldots ,x_{m_s}\}$
. Let
$1\le k\le s-1$
. Assume that
$(x_l,x_m)\mid x_{m_{1\dots k}}$
and
$(x_l,x_m) \nmid x_{m_j}$
for all integers j with
$k+1 \le j \le s$
. Let
$1 \le i_1<\cdots < i_h \le k$
with
$1\le h\le k$
and
$k+1 \le j_1<\cdots <j_t \le s$
with
$1\le t\le s-k$
. Then each of the following is true:
(i).
$(x_l, x_{m_{i_1\dots i_h}})=(x_l, x_m).$
(ii).
$[x_l,x_{m_{j_1\dots j_t}}]=[x_l,x_m].$
(iii).
$[x_{l}, x_{m_{i_1\dots i_hj_1\dots j_t}}]=[x_l,x_{m_{i_1\dots i_h}}].$
Lemma 2.3 [Reference Bourque and Ligh6, Theorem 3]
If S is gcd-closed set of n positive integers and
$(f(S))$
is nonsingular, then
with
and
Lemma 2.4 [Reference Hong16]
Let S be a gcd-closed set and let
$\alpha _{f}(x_k)$
be given as in (2.1). Then
Lemma 2.5 [Reference Bourque and Ligh6]
If S is gcd closed, then
with
$\alpha _{f}(x_k)$
being given as in (2.1).
Lemma 2.6 [Reference Zhu and Yu35]
Let S be a gcd-closed set satisfying the condition
$\mathcal G$
and let
$x_m\in S$
be such that
$\{x_{m_1},\ldots ,x_{m_s}\}\subseteq G_S(x_m)$
. For any real number e, define
Then
For any real number e, the reciprocal arithmetic function
$\frac {1}{\xi _e}$
is defined for any positive integer x by
$\frac {1}{\xi _e}(x):=\frac {1}{x^e}$
. In fact, one has
$\frac {1}{\xi _e}=\xi _{-e}$
.
Lemma 2.7 Let S be a gcd-closed set satisfying the condition
$\mathcal G$
and
$x_m\in S$
with
$G_S(x_m)=\{x_{m_1},\ldots ,x_{m_s}\}$
. Then
Proof Setting
$f=\frac {1}{\xi _a}$
in Lemma 2.4, by Lemmas 2.4 and 2.6, we get that
as required. So Lemma 2.7 is proved.
For any real number e and for any gcd-closed set S satisfying the condition
$\mathcal {G}$
, Zhu and Yu [Reference Zhu and Yu35] proved that the eth power LCM matrix
$[S^e]$
is nonsingular when
$e\ne 0$
which confirmed another conjecture of Hong raised in [Reference Hong18]. In the following, we present the explicit formulas for the inverses of the eth power GCD matrix
$(S^e)$
and the eth power LCM matrix
$[S^e]$
.
Lemma 2.8 Let S be a gcd-closed set of n positive integers and let S satisfy the condition
$\mathcal G$
. Then for any nonzero real number e, the eth power GCD matrix
$(S^e)$
and the eth power LCM matrix
$[S^e]$
are both nonsingular, and for any integers i and j with
$1\le i,j\le n$
, we have
and
where
$c_{ik}$
and
$c_{jk}$
are defined as in (2.2).
Proof Since
we have
where
$B:=\big (\frac {1}{\xi _e}(x_i,x_j)\big )=(\xi _{-e}(S))$
is the
$(-e)$
th power GCD matrix defined on S. We take the determinants on both sides of (2.3) and obtain that
Since the two diagonal matrices in (2.3) are nonsingular, the nonsingularity of the eth power LCM matrix
$[S^e]$
is equivalent to that of B. Thus, we first prove that B is nonsingular.
By Lemma 2.5, the determinant of B is equal to
Hence, it suffices to show that each factor
$\alpha _{\frac {1}{\xi _e}}(x_k)$
is nonzero. First, from Lemma 2.4, we have
Next, for any
$2\le m\le n=|S|$
, let
$G_S(x_m)=\{x_{m_1},\ldots , x_{m_s}\}$
with
$s=|G_S(x_m)|\ge 1$
. For each
$x_{m_j}\in G_S(x_m)$
(
$1\le j\le s$
), since
$-e\ne 0$
and
$x_m\ne x_{m_j}$
, we have
Applying Lemma 2.7 then yields that
Combining (2.5)–(2.7), we conclude that
$\det B\ne 0$
, so B is nonsingular. Replacing
$-e$
by e, this gives that the eth power GCD matrix
$(S^e)$
is nonsingular since
$e\ne 0$
. Moreover, together with (2.4), this implies that
$[S^e]$
is nonsingular.
Furthermore, we apply Lemma 2.3 to the function
$f=\xi _e$
and get the formula for
$(S^e)^{-1}$
. Letting
$f=\frac {1}{\xi _e}$
in Lemma 2.3 gives the entrywise formula for
$B^{-1}$
as follows:
Finally, combining (2.3) and (2.8) yields the desired result immediately.
This completes the proof of Lemma 2.8.
Lemma 2.9 [Reference Zhu32]
Let S be a gcd-closed set satisfying the condition
$\mathcal G$
with
$x_m \in S$
. For any integer r with
$1\le r\le m$
, if
$c_{rm}$
is defined as in (2.2) and
$G_S(x_m)=\{x_{m_1},\ldots , x_{m_s}\}$
, then
3 A key lemma and proof of Theorem 1.2
In this section, with the aid of the lemmas presented in Section 2, we prove Theorem 1.2. For any integers l and m with
$1\le l,m\le |S|$
, we define the two-variable function H as follows:
First of all, we prove a key lemma that plays a crucial role in the proof of Theorem 1.2.
Lemma 3.1 Let S be a gcd-closed set satisfying the condition
$\mathcal {G}$
with
$x_l,x_m\in S$
, and let
$G_S(x_m)=\{x_{m_1},\ldots , x_{m_s}\}$
with
$s\ge 1$
. Let a and b be positive integers. Then
Proof By (3.1) and Lemma 2.9, we have
Since
$(x_l,x_m)|x_m$
, it follows that either
$(x_l,x_m)|x_{m_{1\dots s}}$
, or
$(x_l,x_m)=x_m $
, or there exist an integer k with
$1\le k\le s-1$
and k integers
$i_1,\ldots , i_k$
with
$1\le i_1<\cdots <i_k\le s$
such that
$(x_l,x_m)|x_{m_{i_1\dots i_k}}$
and
$(x_l,x_m) \nmid x_{m_j} $
for all
$j \in \langle s\rangle \setminus \{ i_{1 },\ldots , i_{k}\}$
. So we only need to consider the following three cases.
Case 1.
$(x_l,x_m)|x_{m_{1\dots s}}$
. For
$1 \le j_1<\cdots <j_t\le s$
with
$1\le t\le s$
, since
$x_{m_{1\dots s}}| x_{m_{j_1\dots j_t}}|x_m$
, we have
$(x_l,x_m)|x_{m_{j_1\dots j_t}}|x_m$
and
Thus, by (3.2) and Lemma 2.6, we have
Then it follows from Lemmas 2.1 and 2.7 that
as expected.
Case 2.
$(x_l,x_m)=x_m$
. Then
$x_m|x_l$
. Clearly,
$x_{m_{j_1\dots j_t}}|x_m|x_l$
for
$1 \le j_1<\cdots <j_t\le s$
. This implies that
Therefore, by (3.2) and Lemmas 2.6 and 2.7, we have
and so
$H(l,m)=x_l^b$
as desired.
Case 3. There is an integer k with
$1\le k\le s-1$
and k integers
$i_1,\ldots , i_k$
with
$1\le i_1<\cdots <i_k\le s-1$
such that
$(x_l,x_m)|x_{m_{i_1\dots i_k}}$
and
$(x_l,x_m) \nmid x_{m_j} $
for all
$j \in \langle s\rangle \setminus \{i_{1}, \ldots , i_{k}\}$
. WLOG, one may let
$(x_l,x_m)\mid x_{m_{1\dots k}}$
and
$(x_l,x_m) \nmid x_{m_j} $
for all
$k+1 \le j \le s$
. By (3.2), we get that
where
and
First of all, we compute
$E_1$
. By Lemmas 2.2(i) and 2.6, we obtain that
Subsequently, we calculate
$E_2$
. From Lemmas 2.1, 2.2(i)–(iii), and 2.6, we can deduce that
Thus,
and so by Lemmas 2.1 and 2.7, we have
as asserted.
This finishes the proof of Lemma 3.1.
Finally, we conclude this article by giving the proof of Theorem 1.2.
Proof of Theorem 1.2
Let S be a gcd-closed set which satisfies the condition
$\mathcal {G}$
and let a and b be positive integers with
$a\mid b$
. For any integers i and j with
$1 \le i,j \le |S|$
, by Lemma 2.8, we have
with
$H(i,t)$
being defined as in (3.1).
In what follows, we prove that
for all integers i and j with
$1\le i,j \le |S|$
. By the definition of
$c_{jt} $
as in (2.2), we know that
$ c_{jt} \in \mathbb {Z}$
. So we need only to show that for any positive integer t with
$x_j\mid x_t$
, we have
This will be shown in what follows.
For any given integers i and t with
$1 \le i,t\le |S|$
and
$x_j|x_t$
, we consider the following two cases.
Case 1.
$t=1$
. Then
$x_t=x_1=\gcd (S)$
. At this moment, we must have
$x_j=x_1$
. By Lemmas 2.4 and 2.9, we obtain that
as one expects.
Case 2.
$t\ge 2$
. Since S is gcd closed, one has
$x_1\mid x_t$
and so
$|G_S(x_t)|\ge 1$
. Write
$G_S(x_t):=\{x_{t_1},\ldots , x_{t_s}\}$
with
$s\ge 1$
. Since
$a\mid b$
and S satisfies the condition
$\mathcal {G}$
as well as
$x_j\mid x_t$
, by Lemma 3.1, we have
It then follows that
as required.
This concludes the proof of Theorem 1.2.
Acknowledgements
The author is grateful to the anonymous referee for insightful and invaluable suggestions and comments that improved greatly the presentation of the article.





















