Dice Question Streamline Icon: https://streamlinehq.com

Existence of a Hadamard matrix of order 668

Determine whether a real Hadamard matrix H ∈ M_{668}(±1) with pairwise orthogonal rows exists.

Information Square Streamline Icon: https://streamlinehq.com

Background

Concrete instances of the Hadamard Conjecture can be studied order by order. Known constructions cover all multiples of four up to N=664, but the existence at N=668 remains unresolved. This specific case serves as a benchmark open instance of the general conjecture and highlights the current limits of explicit constructions.

References

At higher $N$ things become more technical, and in order to discuss this, we will need:

\begin{theorem} The determinant of the $4\times4$ matrices is given by \begin{eqnarray*} &&\begin{vmatrix}a_1&a_2&a_3&a_4\ b_1&b_2&b_3&b_4\ c_1&c_2&c_3&c_4\ d_1&d_2&d_3&d_4\end{vmatrix}\ &=&a_1b_2c_3d_4-\ldots \end{eqnarray*} ... At higher $N$ things become more technical, and the continuation of it, we have:

\begin{conjecture}[Hadamard] There is an Hadamard matrix of order $N$, $$H\in M_N(\pm1)$$ for any $N\in4\mathbb N$. \end{conjecture} ... At higher $N$ things become more technical, and more complicated constructions, along the lines of those of Paley and Williamson, are needed. Quite curiously, as of now, early 21th century, the human knowledge stops at the number of the beast, namely: $$\mathfrak N=666$$ That is, explicit examples of Hadamard matrices have been constructed for all multiples of four $N\leq 664$, but no such matrix is known so far at $N=668$.

Advanced linear algebra (2506.18666 - Banica, 23 Jun 2025) in Chapter 11b. Hadamard matrices