Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated

Problem 479

../

claims/: The 1 claim page of Problem 479, one per claimant's result; the problem's standing derives from them.


Statement. Is it true that, for all k≠1k\neq 1, there are infinitely many nn such that 2n≡k(modn)2^n\equiv k\pmod{n}?

Status. Open: the site's label. The one claim page, Tang 2025, is a pending partial claim for the cases k=2ik=2^i, so the problem is open.

Source. erdosproblems.com/479, accessed 2026-09-04. Cite as: T. F. Bloom, Erdős Problem #479, https://www.erdosproblems.com/479.

Formalization. Statement in formal-conjectures.

Current assessment

The assertion for every fixed integer k≠1k\neq1 remains open on the catalog snapshot accessed 2026-09-04 and the searches of 2026-09-05 and 2026-09-06. The 2026-09-05 searches covered the site page, Tang's note and a web search for a proof of the full congruence question, and a bibliographic search for the Graham–Lehmer–Lehmer publication; they found Tang's power-of-two family, no universal proof and no primary publication. The 2026-09-06 searches covered a secondary report of Tang's note, OEIS A036236 for the historical reference, and the site page. The site's commentary records the cases k=2ik=2^i (i≥1i\geq1) and k=−1k=-1, which Erdős and Graham attribute to Graham, Lehmer and Lehmer, and credits Tang's note with a proof for k=2ik=2^i. Tang's note (Section 2.2) and his forum comment of 2 December 2025 list k∈{0,−1,−2}∪{2i:i≥1}k\in\{0,-1,-2\}\cup\{2^i:i\geq1\} as the values for which infinitely many nn are known, and call every other fixed kk open. The historical source and the manuscript discussed below have the narrower scopes stated here.

Selected source statements and bounded applications are author-recorded; no independent review of them is recorded in this repository. This page records no new mathematical status, publication acceptance or formal verification.

Known Results

Quanyu Tang's A Note on Erdős Problem #479: Infinitude of the Sets A(2i)A(2^i) and Related Results (2 December 2025) is an unpublished author manuscript, recorded on its claim page. Its p. 1 explicitly disclaims novelty of the underlying number-theoretic statements and describes the power-of-two argument as expository and independent; it may or may not coincide with the unpublished Graham–Lehmer–Lehmer proof. No accepted or published version was located by the searches recorded above.

Theorem 3.1, pp. 4–5, states that for every integer i≥1i\geq1 there are infinitely many positive integers nn with 2n≡2i(modn)2^n\equiv2^i\pmod n. This is a strict partial family of the problem.

The source's short construction can be checked as follows. Put n=ipn=ip for an odd prime p∤ip\nmid i. Fermat's theorem gives p∣2ip−2ip\mid 2^{ip}-2^i. Write i=2s∏jqjeji=2^s\prod_j q_j^{e_j} with the qjq_j odd. The factor 2s2^s divides 2i2^i since s≤is\leq i. Set

dj=ord⁡qjej(2),mj=djgcd⁡(dj,i),L=lcm⁡jmj,d_j=\operatorname{ord}_{q_j^{e_j}}(2),\qquad m_j=\frac{d_j}{\gcd(d_j,i)},\qquad L=\operatorname{lcm}_j m_j,

where L=1L=1 for an empty list. If p≡1(modL)p\equiv1\pmod L, then dj∣i(p−1)d_j\mid i(p-1) for every jj, so every odd prime-power factor of ii divides 2ip−2i2^{ip}-2^i. Hence i∣2ip−2ii\mid2^{ip}-2^i. Dirichlet's theorem supplies infinitely many primes p≡1(modL)p\equiv1\pmod L after discarding 22 and the finitely many divisors of ii. As gcd⁡(i,p)=1\gcd(i,p)=1, the two divisibilities combine to give the desired congruence modulo ipip, for distinct unbounded n=ipn=ip. This construction check takes Fermat's theorem, multiplicative orders and Dirichlet's theorem as named external dependencies, and it is not a complete audit of the source proof.

Erdős–Graham (1980), printed p. 96, already records the cases k=2ik=2^i for i≥1i\geq1 and k=−1k=-1. It reports D. H. and Emma Lehmer's finite search finding n<5⋅109n<5\cdot10^9 for every ∣k∣≤100|k|\leq100, k≠1k\neq1. For k=3k=3 it gives

4700063497=19⋅47⋅52632294700063497=19\cdot47\cdot5263229

as the then-smallest solution with n>1n>1 and the only one then known. This is a dated computational report, not a current minimum claim or an infinitude proof. The underlying Graham–D. H. Lehmer–Emma Lehmer source remains unlocated. Tang's partial family receives no universal-solution, acceptance, formal, or complete-proof credit here.

Linked library material

These entries are derived from explicit links on library pages. They are navigation only and do not by themselves record mathematical progress.