← All problems

Uniform separation of small sums of roots of unity

For every fixed k∈Nk\in\mathbb{N} and every ε>0\varepsilon>0, does there exist p0=p0(k,ε)p_0=p_0(k,\varepsilon) such that, for every prime p≥p0p\ge p_0 and every pairwise distinct a1,…,ak∈Z/pZa_1,\ldots,a_k\in\mathbb{Z}/p\mathbb{Z}, one has

∑j=1kexp⁡ ⁣(2πiajp)≠0⟹∣∑j=1kexp⁡ ⁣(2πiajp)∣≥exp⁡(−εp)?\sum_{j=1}^k\exp\!\left(\frac{2\pi i a_j}{p}\right)\ne 0\quad\Longrightarrow\quad\left|\sum_{j=1}^k\exp\!\left(\frac{2\pi i a_j}{p}\right)\right|\ge \exp(-\varepsilon p)?
References

Progress summary

Refreshed
Open · progress reported

A new preprint reports partial progress on separating nonzero sums of roots of unity, but the strongest bound remains conditional and the claims are unverified.

The problem seeks uniform lower bounds for nonzero sums of roots of unity. No proposer or original date is identified in the retrieved sources.

Known results

  • Barber obtained an upper bound of O(n−4/3)O(n^{-4/3}) for five-term sums; the bound improves to O(n−7/3)O(n^{-7/3}) infinitely often.
  • For general kk, nearby work gives f(k,N)=O ⁣(N−⌊log⁡k/log⁡6⌋)f(k,N)=O\!\left(N^{-\lfloor\log k/\log 6\rfloor}\right); when 6∣N6\mid N, it gives f(5,N)≪N−2f(5,N)\ll N^{-2}.
  • Myerson is credited with earlier work on the minimum nonzero magnitude of sums of roots of unity.

October 2026 claimed progress

Qi Cheng's preprint claims a polylogarithmic exceptional-set estimate for symmetric five-term sums and a conditional subexponential lower bound for sums of prime-order roots of unity under the number-field abcabc conjecture. This is a substantive advance if correct, but no independent mathematical assessment was found.

Current status (as of October 2026): The classical quantitative results are established, while Cheng's new exceptional-set and conditional-separation claims remain unverified and the general uniform-separation problem remains open.

Sources

Solutions 0

No solutions have been posted yet.