Sidon sets and perfect difference sets

DefinitionPerfect difference sets modulo \(v\)#✓fully proved

A set \(B \subseteq \mathbb{N}\) is a perfect difference set modulo a natural number \(v\) when every nonzero residue \(x\) modulo \(v\) satisfies \(x \equiv a - b \pmod{v}\) for exactly one ordered pair \((a, b)\) of elements of \(B\).

def IsPerfectDifferenceSetMod (B : Set ℕ) (v : ℕ) : Prop :=
  ∀ x : ZMod v, x ≠ 0 → ∃! ab : ℕ × ℕ, ab.1 ∈ B ∧ ab.2 ∈ B ∧ (ab.1 : ZMod v) - ab.2 = x
Used by 2

DefinitionErdős Problem 707#✓fully proved

The proposition that for every finite Sidon set \(A \subseteq \mathbb{N}\) there are a prime \(p\) and a set \(B \subseteq \mathbb{N}\) with \(A \subseteq B\) such that \(B\) is a perfect difference set modulo \(p^2 + p + 1\).

def sidonExtendsToPrimePerfectDifferenceSet : Prop :=
  ∀ A : Set ℕ, A.Finite → A.IsSidon →
    ∃ p : ℕ, p.Prime ∧ ∃ B : Set ℕ, A ⊆ B ∧ IsPerfectDifferenceSetMod B (p ^ 2 + p + 1)
Uses 2 · Used by 1

DefinitionErdős Problem 707, weak form#✓fully proved

The proposition that for every finite Sidon set \(A \subseteq \mathbb{N}\) there are a natural number \(m \ge 1\) and a set \(B \subseteq \mathbb{N}\) with \(A \subseteq B\) such that \(B\) is a perfect difference set modulo \(m^2 + m + 1\).

This weakens Erdős Problem 707 by allowing any \(m \ge 1\) in place of a prime \(p\).

def sidonExtendsToPerfectDifferenceSet : Prop :=
  ∀ A : Set ℕ, A.Finite → A.IsSidon →
    ∃ m : ℕ, 1 ≤ m ∧ ∃ B : Set ℕ, A ⊆ B ∧ IsPerfectDifferenceSetMod B (m ^ 2 + m + 1)
Uses 2 · Used by 1

TheoremErdős Problem 707 is false#●ready to prove

It is not the case that every finite Sidon set \(A \subseteq \mathbb{N}\) is contained in a perfect difference set modulo \(p^2 + p + 1\) for some prime \(p\): there is a finite Sidon set \(A \subseteq \mathbb{N}\) such that for every prime \(p\), no set \(B \subseteq \mathbb{N}\) with \(A \subseteq B\) is a perfect difference set modulo \(p^2 + p + 1\). Here \(A\) is Sidon when \(w + x = y + z\) with \(w, x, y, z \in A\) forces \(\{w, x\} = \{y, z\}\), and \(B\) is a perfect difference set modulo \(v\) when every nonzero residue \(r\) modulo \(v\) satisfies \(r \equiv a - b \pmod{v}\) for exactly one ordered pair \((a, b)\) of elements of \(B\).

Alexeev and Mixon proved that the Sidon set \(\{1, 2, 4, 8\}\) is such a set (their proof is formalized in Lean); erdosproblems.com/707 lists the problem as disproved.

public theorem_wanted sidonExtendsToPrimePerfectDifferenceSet_refuted :
    ¬ sidonExtendsToPrimePerfectDifferenceSet
Sources
  • T. F. Bloom, Erdős Problems, Problem #707, https://www.erdosproblems.com/707 (status DISPROVED (LEAN)); resolution: B. Alexeev and D. G. Mixon, Forbidden Sidon subsets of perfect difference sets, featuring a human-assisted proof, arXiv:2510.19804 (2025); M. Hall, Jr., Cyclic projective planes, Duke Math. J. 14 (1947), 1079-1090, abstract of arXiv:2510.19804: {1, 2, 4, 8, 13} is a finite Sidon set contained in no finite perfect difference set; erdosproblems.com/707: {1, 2, 4, 8} cannot be extended to a perfect difference set modulo p² + p + 1 for any prime p, and Hall's {1, 3, 9, 10, 13} cannot be extended to any perfect difference set, https://arxiv.org/abs/2510.19804.
Uses 1
Statement usesErdős Problem 707

TheoremThe weak form of Erdős Problem 707 is false#●ready to prove

It is not the case that every finite Sidon set \(A \subseteq \mathbb{N}\) is contained in a perfect difference set modulo \(m^2 + m + 1\) for some natural number \(m \ge 1\): there is a finite Sidon set \(A \subseteq \mathbb{N}\) such that for every \(m \ge 1\), no set \(B \subseteq \mathbb{N}\) with \(A \subseteq B\) is a perfect difference set modulo \(m^2 + m + 1\). Here \(A\) is Sidon when \(w + x = y + z\) with \(w, x, y, z \in A\) forces \(\{w, x\} = \{y, z\}\), and \(B\) is a perfect difference set modulo \(v\) when every nonzero residue \(r\) modulo \(v\) satisfies \(r \equiv a - b \pmod{v}\) for exactly one ordered pair \((a, b)\) of elements of \(B\).

Alexeev and Mixon proved that the Sidon set \(\{1, 2, 4, 8, 13\}\) is contained in no perfect difference set, and Hall (1947) proved the same for \(\{1, 3, 9, 10, 13\}\).

public theorem_wanted sidonExtendsToPerfectDifferenceSet_refuted :
    ¬ sidonExtendsToPerfectDifferenceSet
Sources
  • T. F. Bloom, Erdős Problems, Problem #707, https://www.erdosproblems.com/707 (status DISPROVED (LEAN)); resolution: B. Alexeev and D. G. Mixon, Forbidden Sidon subsets of perfect difference sets, featuring a human-assisted proof, arXiv:2510.19804 (2025); M. Hall, Jr., Cyclic projective planes, Duke Math. J. 14 (1947), 1079-1090, abstract of arXiv:2510.19804: {1, 2, 4, 8, 13} is a finite Sidon set contained in no finite perfect difference set; erdosproblems.com/707: {1, 2, 4, 8} cannot be extended to a perfect difference set modulo p² + p + 1 for any prime p, and Hall's {1, 3, 9, 10, 13} cannot be extended to any perfect difference set, https://arxiv.org/abs/2510.19804.
Uses 1
3 fully proved2 ready to prove

The dependency explorer needs JavaScript and its graph data.

Download the complete graph data.