1 Goal
Our objective is to determine the structure of every finite abelian group completely. Specifically, we will show that every finite abelian group is isomorphic to a direct product of cyclic groups, and that this decomposition is essentially unique.
2 Basic Properties of Cyclic Groups
If gcd(m,n)=1, then Z/mnZ≅Z/mZ×Z/nZ.
Define φ:Z/mnZ→Z/mZ×Z/nZ by φ(a)=(amodm,amodn). This is a homomorphism. If φ(a)=(0,0), then m∣a and n∣a; since gcd(m,n)=1, we get mn∣a, so a=0 in Z/mnZ. Thus φ is injective, and since ∣Z/mnZ∣=mn=∣Z/mZ×Z/nZ∣, it is an isomorphism. □
If n=p1a1p2a2⋯prar is the prime factorization of n, then Z/nZ≅Z/p1a1Z×Z/p2a2Z×⋯×Z/prarZ.
3 Decomposition ofp-Groups
Let G be a finite abelian p-group (i.e., ∣G∣=pn). If g∈G is an element of maximal order, then ⟨g⟩ is a direct factor of G: there exists a subgroup H such that G=⟨g⟩×H.
The proof of this lemma proceeds by induction on ∣G∣. The key idea is as follows.
Let ∣g∣=pk. We induct on ∣G∣=pn. If k=n, then G=⟨g⟩ and the result is trivial. Suppose k<n.Let π:G→G/⟨g⟩ be the natural projection. By induction, G/⟨g⟩ decomposes as a direct product of cyclic groups.For each cyclic direct factor ⟨hˉ⟩ of G/⟨g⟩ (where hˉ=π(h)), we can choose the lift h so that ord(h)=ord(hˉ); this is possible because g has maximal order. Repeating this process yields G=⟨g⟩×H. □
Every finite abelian p-group G is isomorphic to a direct product of cyclic p-groups: G≅Z/pa1Z×Z/pa2Z×⋯×Z/parZ,a1≥a2≥⋯≥ar≥1.
By induction on ∣G∣. The preceding lemma gives G=⟨g⟩×H where ∣g∣=pa1 and ∣H∣<∣G∣. Applying the induction hypothesis to H yields H≅Z/pa2Z×⋯×Z/parZ. Since g has maximal order, a1≥a2. □
4 Primary Decomposition
Let G be a finite abelian group with ∣G∣=p1n1p2n2⋯prnr. Writing Gpi for the Sylow pi-subgroup of G, G=Gp1×Gp2×⋯×Gpr.
Since G is abelian, each Sylow subgroup Gpi is unique, hence normal. We have ∣Gpi∣=pini. Since gcd(pini,∏j=ipjnj)=1, it follows that Gpi∩∏j=iGpj={e}. As ∏i∣Gpi∣=∣G∣, we conclude G=Gp1×⋯×Gpr. □
5 The Fundamental Theorem of Finite Abelian Groups
Combining the results of the previous sections:
Every finite abelian group G is isomorphic to a direct product of cyclic groups: G≅Z/p1a1Z×Z/p2a2Z×⋯×Z/psasZ
where the pi are primes (not necessarily distinct) and ai≥1. This decomposition is unique up to reordering. The prime powers p1a1,…,psas are called the elementary divisors of G.
By the primary decomposition theorem, G=Gp1×Gp2×⋯×Gpr where each Gpi is the Sylow pi-subgroup. Each Gpi is a finite abelian pi-group, so by the decomposition theorem for p-groups, Gpi≅Z/piai,1Z×Z/piai,2Z×⋯×Z/piai,riZ.
Collecting these factors across all primes gives the desired decomposition. Uniqueness follows from the uniqueness of the decomposition of each p-component (proved below). □
Every finite abelian group G can be written as G≅Z/d1Z×Z/d2Z×⋯×Z/dtZ,d1∣d2∣⋯∣dt,di≥2.
This decomposition is unique. The integers d1,…,dt are called the invariant factors of G.
Starting from the elementary divisor decomposition, we construct the invariant factors as follows. For each prime p, list its elementary divisors in decreasing order: pa1≥pa2≥⋯≥parp where a1≥a2≥⋯≥arp≥1.Set t=maxprp (the maximum number of factors across all primes). For j=1,…,t, define dj=p∏paj(p)
where aj(p) is the j-th exponent for the prime p (with aj(p)=0 when j>rp).Since a1(p)≥a2(p)≥⋯ for each p, we have d1∣d2∣⋯∣dt. By the Chinese Remainder Theorem (gcd(m,n)=1 implies Z/mnZ≅Z/mZ×Z/nZ), Z/djZ≅p∏Z/paj(p)Z,
so ∏jZ/djZ agrees with the original elementary divisor decomposition. Uniqueness follows from that of the elementary divisors. □
The elementary divisors and invariant factors carry the same information in different packaging. To pass from elementary divisors to invariant factors, one groups the prime powers by prime and multiplies them together in a specific pattern; the reverse conversion is obtained by factoring each invariant factor into prime powers.
6 Proof of Uniqueness
The elementary divisor decomposition of a finite abelian group is unique up to reordering.
Let p be a prime. From the group G, one can read off intrinsic invariants using the subgroups G[p]={g∈G∣gp=e} and Gp={gp∣g∈G}.If G≅Z/pa1Z×⋯×Z/parZ with a1≥⋯≥ar≥1, then ∣G[p]∣=pr, which determines r (the number of cyclic factors in the p-component).Moreover, G/Gp≅(Z/pZ)r, and the Sylow p-subgroup of Gp is isomorphic to Z/pa1−1Z×⋯×Z/par−1Z (with factors of exponent 0 dropped). Iterating this procedure determines each ai uniquely. □
7 Classification in Practice
We have ∣G∣=8=23. The partitions of 3 are (3), (2,1), and (1,1,1). The corresponding abelian groups are:
Z/8Z (elementary divisors {8}, invariant factors {8}).
Z/4Z×Z/2Z (elementary divisors {4,2}, invariant factors {2,4}).
Z/2Z×Z/2Z×Z/2Z (elementary divisors {2,2,2}, invariant factors {2,2,2}).
We have 36=22⋅32. The partitions of 2 are (2) and (1,1), for both the 2-part and the 3-part. This gives 2×2=4 groups:
Z/4Z×Z/9Z≅Z/36Z (invariant factors {36}).
Z/4Z×Z/3Z×Z/3Z (invariant factors {3,12}).
Z/2Z×Z/2Z×Z/9Z (invariant factors {2,18}).
Z/2Z×Z/2Z×Z/3Z×Z/3Z≅Z/6Z×Z/6Z (invariant factors {6,6}).
We have 72=23⋅32. There are 3 partitions of 3 and 2 partitions of 2, giving 3×2=6 groups:
Z/8Z×Z/9Z≅Z/72Z.
Z/8Z×Z/3Z×Z/3Z.
Z/4Z×Z/2Z×Z/9Z.
Z/4Z×Z/2Z×Z/3Z×Z/3Z.
Z/2Z×Z/2Z×Z/2Z×Z/9Z.
Z/2Z×Z/2Z×Z/2Z×Z/3Z×Z/3Z.
8 Exponent and Rank
For a finite abelian group G:
The exponentexp(G) is the least common multiple of the orders of all elements of G. In the invariant factor form, exp(G)=dt (the largest invariant factor).
The rank of G is the number of factors in the elementary divisor decomposition.
G is cyclic ⇔exp(G)=∣G∣⇔G has exactly one invariant factor.
9 Summary
Every finite abelian group decomposes as a direct product of cyclic groups (in both the elementary divisor form and the invariant factor form).
The decomposition is unique, providing a complete classification of finite abelian groups.
The classification reduces to the prime factorization of the group order: for each prime power, one enumerates partitions.
The theory of abelian groups is, in this sense, "complete"— a striking contrast to the vastly more complex structure theory of non-abelian groups.