Consider the polynomial:
$$x^3-2x^2+3x+5=0$$
where $x_1, x_2$ and $x_3$ are the roots of the above polynomial. Now, consider the following determinant, which is defined using the above given roots:
$$\Delta = \begin{vmatrix} 1 & 1 & 1 \\ x_1 & x_2 & x_3 \\ x_1^2 & x_2^2 & x_3^2 \\ \end{vmatrix}$$
And what is asked of me is to find $\Delta^2$.
After a bit of manipulation I found the following:
$$\Delta = (x_2-x_1)(x_3-x_1)(x_3-x_2)$$
Interestingly enough, this type of matrix has a special name: Vandermonde matrix and instead of doing that bit of manipulation after which I arrived at the above expression for $\Delta$, I could've used the formula given on that wikipedia page. Anyways...
So, I have to find:
$$\Delta^2 = (x_2-x_1)^2(x_3-x_1)^2(x_3-x_2)^2$$
The problem is that I cannot find any of the roots. I used the rational root theorem and found that there are no rational roots. None of the divisors of the free term, $5$, give $0$ when plugged into the polynomial. I tried all options: $\{\pm 1, \pm 5 \}$ and they all give something $\ne 0$.
So then I used the notation:
$$f(x) = x^3-2x^2+3x+5$$
found the derivative:
$$f'(x) = 3x^2-4x+3$$
and I observed that $f'(x) > 0$ for all $x \in \mathbb{R}$. So the function $f$ is strictly increasing, so we can have at most one solution to $f(x) = 0$. Because of what I showed above, this solution cannot be rational. So I concluded that we have one rational root and two complex (and conjugate, since $f \in \mathbb{R}[X]$) roots. But this is as far as I got. I cannot find them. And I tried finding $\Delta^2$ without finding the roots, but I couldn't solve that either.
The 'easy way' to do this leans on power sums instead of elementary symmetric polynomials. We are implicitly working in $\mathbb C$ for this problem.
For any degree $n$ monic polynomial, first encode it in an $n$ x $n$ Companion matrix $C$.
now consider the power sum for $k\in\{1,2,3,...,\}$
$s_k := \lambda_1^k + \lambda_2^k +....+ \lambda_n^k = \text{trace}\big(C^k\big)$ and
$s_0:= n$
and $\lambda_i$ are the roots to your polynomial / the eigenvalues of $C$
now consider the matrix
$M_n := \begin{bmatrix} s_0 & s_1 & s_2 & \cdots & s_{n-1}\\ s_1 & s_2& s_3 & \cdots & s_n \\ s_2& s_3 & s_4 & \cdots & s_{n+1}\\ \vdots & \vdots & \vdots & \ddots & \vdots\\ s_{n-1} & s_{n} & s_{n+1} & \cdots & s_{2n-2} \end{bmatrix}$
for your problem here it is just
$M_3 := \begin{bmatrix} s_0 & s_1 & s_2\\ s_1 & s_2& s_3\\ s_2& s_3 & s_4 \\ \end{bmatrix}$
(note: matrix multiplication is only needed to get $s_2$. You get $s_0$ and $s_1$ immediately. Also $s_3$ and $s_4$ are obtainable by Cayley Hamilton.)
and
$\det\big(M_3\big) = \Delta^2 = (\lambda_2-\lambda_1)^2(\lambda_3-\lambda_1)^2(\lambda_3-\lambda_2)^2$
because
$M= V^TV \longrightarrow \det\big(M\big)=\det\big(V^TV\big)=\det\big(V^T\big)\det\big(V\big)=\det\big(V\big)^2$
(note that is a transpose, not conjugate transpose. This factorization turns out to be quite useful.)
where, for avoidance of doubt, $V$ is the Vandermonde matrix, shown below for the $n=3$ case
$V := \begin{bmatrix} 1 & \lambda_1 & \lambda_1^2 \\ 1 & \lambda_2& \lambda_2^2\\ 1 & \lambda_3 & \lambda_3^2 \\ \end{bmatrix}$