← All problems

Erdős Problem #406 — Powers of two with restricted ternary digits

47 years old · traced to

For each integer n≥0n\geq 0, let L(n)L(n) and aj(n)∈{0,1,2}a_j(n)\in\{0,1,2\} be the unique ternary digits of 2n2^n, so that

2n=∑j=0L(n)aj(n)3j,aL(n)(n)≠0.2^n=\sum_{j=0}^{L(n)}a_j(n)3^j,\qquad a_{L(n)}(n)\ne 0.

Is the set

{n∈Z≥0:aj(n)∈{0,1} for every 0≤j≤L(n)}\bigl\{n\in\mathbb Z_{\geq 0}:a_j(n)\in\{0,1\}\text{ for every }0\leq j\leq L(n)\bigr\}

finite? Equivalently, are there only finitely many powers of 22 whose base-33 expansion omits the digit 22? The known exponents are 0,2,80,2,8, giving 11, 4=1+34=1+3, and 256=1+3+32+35256=1+3+3^2+3^5, respectively. Erdős Problem #406

Background

Erdős posed a stronger exact-examples conjecture: that the only exponents giving a power of 22 expressible as a sum of distinct integral powers of 33 are 0,2,80,2,8. That conjecture is distinct from the question here: proving that there are finitely many such exponents would not by itself show that the three known examples are the only ones. Narkiewicz records that the question was motivated by Gupta’s numerical check, which found no other examples through exponent 47344734. Narkiewicz (1980) Stoll, CIRM slides (2019)

A later computation reported that every 2n2^n for 16≤n≤5.9×102116\leq n\leq 5.9\times10^{21} contains all three ternary digits. Such a finite range cannot decide either the finiteness question or the exact-examples conjecture, since it leaves larger exponents unexamined. Erdős Problem #406

References

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.