ON THE CRITICAL GROUP OF HINGE GRAPHS
AREN MARTINIAN AND ANDR´
ES R. VINDAS-MEL´
ENDEZ
Abstract.
Let
G
be a finite, connected, simple graph. The critical group
K
(
G
), also known as
the sandpile group, is the torsion subgroup of the cokernel of the graph Laplacian
cok
(
L
). We
investigate a family of graphs with relatively simple non-cyclic critical group with an end goal of
understanding whether multiple divisors, i.e., formal linear combinations of vertices of
G
, generate
K
(
G
). These graphs, referred to as hinge graphs, can be intuitively understood by taking multiple
base shapes and “gluing” them together by a single shared edge and two corresponding shared
vertices. In the case where all base shapes are identical, we compute the explicit structure of the
critical group. Additionally, we compute the order of three special divisors. We prove the structure
of the critical group of hinge graphs when variance in the number of vertices of each base shape is
allowed, generalizing many of the aforementioned results.
1. Introduction
In this paper we study a finite abelian group associated to a finite connected graph
G
, known as
the critical group of
G
. The critical group goes by different names (e.g., the Jacobian group, sandpile
group, component group) and is studied in various mathematical areas (e.g., algebraic geometry,
statistical physics, combinatorics) [16]. We focus on the combinatorial definition of the critical
group involving chip-firing operations and its connections to graph-theoretic trees. In particular, for
a finite connected graph, the order of the critical group equals the number of spanning trees of the
graph. For the interested reader, we recommend the survey paper by Glass and Kaplan [16] as an
introduction to the study of critical groups and chip-firing, as well as the books by Klivans [20] and
Cory and Perkinson [14] for comprehensive considerations of chip-firing.
There are many results on the group structure of the critical group and the relationship with the
structure of an associated graph, see for instance [2,7,9,13]. Determining the critical group for certain
families of graphs continues to be an active area of research. There exists work where the critical
group has been partially determined for some families of graphs, for instance see [5,11,17,22,29
–
31].
Additionally, there is a growing body of work where the complete critical group structure for families
of graphs is determined, see for instance [8,10,12,18,21,23–28].
The family of graphs that we study are those which we call hinge graphs. These are graphs that
can be intuitively understood by taking multiple base shapes and “gluing” them together on a single
shared edge and two corresponding shared vertices. In [13], Cori and Rossin show that the critical
group of a planar graph
G
is isomorphic to the critical groups of the dual of
G
. It so happens that
hinge graphs are dual to a family of graphs known as thick cycle graphs, which are cycle graphs
where multiple edges are allowed, and were studied in [1,4]. Furthermore, thick cycle graphs can be
seen as specializations of outerplanar graphs studied in [3].
The study of hinge graphs arose independently and was motivated primarily in attempt to answer
the question proposed in [16] on how divisors generate critical groups in cases where the group
is non-cyclic. Hinge graphs, especially those containing identical copies of the same base shape,
are some of the simplest examples of graphs with non-cyclic critical groups, and whose behavior
can be thoroughly studied. In addition, the study of hinge graphs has led to observations about
proving linear equivalence and the order of divisors which provides a streamlined approach to the
investigation of divisors and critical groups of graphs more generally. As mentioned before, hinge
1
arXiv:2210.01793v3 [math.CO] 20 Sep 2023