Let us therefore consider a one-dimensional Ising model with
spins and free boundary conditions, i.e. the first and the last spin can assume any value. Using the nearest neighbour interaction Hamiltonian we have:

and defining

and

for the sake of simplicity, the partition function of the system will be:

If we now set

, namely if the system is not subjected to any external field, then:

In order to compute

we use the so called
recursion method: from the expression of

we can deduce the expression of the partition function

of the same system with one additional spin added to the lattice:

However, the sum over

gives:

where we have used the evenness of the hyperbolic cosine, namely the fact that

. Therefore we have:

and iterating the relation

times we get:

Since
[1]:

we get the final result:

Now, following all the prescriptions we know, we get to:
![{\displaystyle F(T)=-k_{B}T\left[\ln 2+(N-1)\ln \left(2\cosh {\frac {J}{k_{B}T}}\right)\right]}](//restbase.wikitolearn.org/en.wikitolearn.org/v1/media/math/render/svg/fea77a682183d6b6c4c7bbbf688193177d1aa0eb)
and in the thermodynamic limit:

Let us note that

does indeed respect the properties we have seen in
Analytic properties of the Ising model. Furthermore, since the logarithm and the hyperbolic cosine are analytic functions we see that

is itself an analytic function of

: it is therefore impossible that the system will exhibit any kind of phase transition (at least for
[2]) even in the thermodynamic limit.
- ↑ This can be justified as follows:
is the partition function of a single-spin Ising model and we are considering two-spin interactions, so this single spin will not interact with anything. Therefore the Hamiltonian of the system in this case is null, and so
.
- ↑ In this case, in fact, there can be problems. From this fact we can state that the only "phase transition" that can happen in a one-dimensional Ising model occurs at
(which is of course an unphysical case), where all the spins are aligned.