Minimal polynomial (linear algebra)

In linear algebra, the minimal polynomial of an matrix over a field is the monic polynomial over of least degree such that. Any other polynomial with is a multiple of.
The following three statements are equivalent:
  1. is a root of,
  2. is a root of the characteristic polynomial of,
  3. is an eigenvalue of matrix.
The multiplicity of a root of is the largest power such that strictly contains. In other words, increasing the exponent up to will give ever larger kernels, but further increasing the exponent beyond will just give the same kernel.
If the field is not algebraically closed, then the minimal and characteristic polynomials need not factor according to their roots alone, in other words they may have irreducible polynomial factors of degree greater than. For irreducible polynomials one has similar equivalences:
  1. divides,
  2. divides,
  3. the kernel of has dimension at least.
  4. the kernel of has dimension at least.
Like the characteristic polynomial, the minimal polynomial does not depend on the base field, in other words considering the matrix as one with coefficients in a larger field does not change the minimal polynomial. The reason is somewhat different from for the characteristic polynomial, namely the fact that the minimal polynomial is determined by the relations of linear dependence between the powers of : extending the base field will not introduce any new such relations.
The minimal polynomial is often the same as the characteristic polynomial, but not always. For example, if is a multiple of the identity matrix, then its minimal polynomial is since the kernel of is already the entire space; on the other hand its characteristic polynomial is . The minimal polynomial always divides the characteristic polynomial, which is one way of formulating the Cayley–Hamilton theorem.

Formal definition

Given an endomorphism on a finite-dimensional vector space over a field, let be the set defined as
where is the space of all polynomials over the field. is a proper ideal of. Since is a field, is a principal ideal domain, thus any ideal is generated by a single polynomial, which is unique up to units in. A particular choice among the generators can be made, since precisely one of the generators is monic. The minimal polynomial is thus defined to be the monic polynomial which generates. It is the monic polynomial of least degree in.


An endomorphism of a finite dimensional vector space over a field is diagonalizable if and only if its minimal polynomial factors completely over into distinct linear factors. The fact that there is only one factor for every eigenvalue means that the generalized eigenspace for is the same as the eigenspace for : every Jordan block has size. More generally, if satisfies a polynomial equation where factors into distinct linear factors over, then it will be diagonalizable: its minimal polynomial is a divisor of and therefore also factors into distinct linear factors. In particular one has:
These cases can also be proved directly, but the minimal polynomial gives a unified perspective and proof.


For a vector in define:
This definition satisfies the properties of a proper ideal. Let be the monic polynomial which generates it.



Define to be the endomorphism of with matrix, on the canonical basis,
Taking the first canonical basis vector and its repeated images by one obtains
of which the first three are easily seen to be linearly independent, and therefore span all of. The last one then necessarily is a linear combination of the first three, in fact
so that:
This is in fact also the minimal polynomial and the characteristic polynomial : indeed divides which divides, and since the first and last are of degree and all are monic, they must all be the same. Another reason is that in general if any polynomial in annihilates a vector, then it also annihilates , and therefore by iteration it annihilates the entire space generated by the iterated images by of ; in the current case we have seen that for that space is all of, so. Indeed one verifies for the full matrix that is the null matrix: