For any $m\in\mathbb N$, let $S(m)$ be the digit sum of $m$ in the decimal system.

For example, $S(1234)=1+2+3+4=10, S(2^5)=S(32)=5$.

Question 1 :Is the following true? $$\lim_{n\to\infty}S(3^n)=\infty.$$

Question 2 :How about $S(m^n)$ for $m\ge 4$ except some trivial cases?

Remark : This question has been asked previously on math.SE without receiving any complete answers, where spin proved that $\lim_{n\to\infty} \sup S(m^n )=\infty$ when $m$ is not a power of $10$.


Motivation : I've got the following : $$\lim_{n\to\infty}S(2^n)=\infty.$$

Proof : The point of this proof is that there exists a non-zero number between the ${m+1}^{th}$ digit and ${4m}^{th}$ digit.

If $$2^n=A\cdot{10}^{4m}+B, B\lt {10}^m, 0\lt A,$$ then $2^n\ge {10}^{4m}\gt 2^{4m}$ leads $n\gt 4m$. Hence, the left side can be divided by $2^{4m}$. Also, $B$ must be divided by $2^{4m}$ because ${10}^{4m}=2^{4m}\cdot 5^{4m}$. However, since $$B\lt {10}^m\lt {16}^m=2^{4m},$$ $B$ can not be divided by $2^{4m}$ if $B\not=0$. If $B=0$, then the right side can be divided by $5$ but the left side cannot be divided by $5$. Hence, we now know that there is a non-zero number between the ${m+1}^{th}$ digit and ${4m}^{th}$ digit. Since $2^n$ cannot be divided by $5$, the first digit is not $0$. There exists non-zero number between the second digit and the fourth digit. Again, there exists non-zero number between $5^{th}$ digit and ${16}^{th}$ digit. By the same argument as above, if $2^n$ has more than $4^k$ digits, then $S(n)\ge {k+1}$. Hence, $$n\log {2}\ge 4^k-1\ \ \Rightarrow \ \ S(n)\ge k+1.$$ Now we know that $$\lim_{n\to\infty}S(2^n)=\infty$$ as desired. Now the proof is completed.

However, I've been facing difficulty for the $m=3$ case. I've got $\lim\sup S(3^n)=\infty$.

Proof : Suppose that $3^n$ has $m$ digits. Letting $l=\varphi({10}^m)+n$, then $$3^l-3^n=3^n(3^{\varphi({10}^m)}-1).$$ Since this can be divided by ${10}^m$, we know that the last $m$ digits of $3^l$ are equal to those of $3^n$. Hence, we get $\lim\sup S(3^n)=\infty$.

However, I can't get $\lim_{n\to\infty}\inf S(3^n)$. Can anyone help?

  • $\begingroup$ I like your argument. Note you can replace 4m by ceil(mlog_2 10), giving a slightly tighter location of nonzero digits. For 3, maybe you can argue that a large string of 0's near the least significant bits gives that many nonzeros further up. $\endgroup$ Sep 26, 2013 at 15:28
  • $\begingroup$ @TheMaskedAvenger: Thank you for pointing it out. $\endgroup$
    – mathlove
    Sep 26, 2013 at 15:35
  • $\begingroup$ Btw this sequence is A004166 in the OEIS oeis.org/A004166. It has no much information however; maybe you could add there the link to this question... $\endgroup$ Sep 26, 2013 at 18:40
  • $\begingroup$ See this paper of Blecksmith, Filaseta, and Nicol for some closely related results: math.sc.edu/~filaseta/papers/blockpaper.pdf $\endgroup$ Sep 27, 2013 at 1:01

3 Answers 3


This follows from W. M. Schmidt's Subspace theorem, which is a deep theorem in diophantine approximations generalizing Roth's to several variables. A full account of this theorem and its proof, as well as some of its striking applications, can be found in chapter 7 of Heights in Diophantine Geometry by Bombieri and Gubler. The following result, the finiteness of the number of non-degenerate solutions to the so-called "$S$-unit equation," is a straightforward application of Schmidt's theorem. (See Theorem 7.4.2 in [HIDG]):

Let $S$ be a finite set of prime numbers, and fix $n \in \mathbb{N}$. Consider $\mathcal{X}$ the set of solutions to $x_1 + \cdots + x_n = 1$ in rational numbers $x_i$ of the form $\pm \prod_{p \in S} p^{a_p}$, $a_p \in \mathbb{Z}$, such that no proper subsum of $x_1+\cdots+x_n$ vanishes. Then $\mathcal{X}$ is a finite set.

This implies your question immediately upon considering $S := \{2,3,5,7\}$.

However, the proof of the Subspace theorem is not effective, and this only shows $S(3^n) \to +\infty$ without any lower estimate on the rate of growth. An effective lower bound on $S(3^n)$ (going to infinity with $n$) is available through Baker's method; it is due to Stewart, and the google search led me to the old MathOverflow post linked to in my comment below. I will just copy the references which Gerry Myerson supplied there:

C. L. Stewart, On the representation of an integer in two different bases, J Reine Angew Math 319 (1980) 63-72, MR 81j:10012; H G Senge, E G Straus, PV-numbers and sets of multiplicity, Proceedings of the Washington State University Conference on Number Theory (1971) 55-67, MR 47 #8452.

  • $\begingroup$ You are welcome! And actually, upon looking at a paper of Stewart from 1980, I realize that I was wrong about effectivity: an explicit lower bound is actually possible with Baker's method. I will edit. $\endgroup$ Sep 26, 2013 at 15:54
  • 1
    $\begingroup$ Actually, see here: mathoverflow.net/questions/38971/… $\endgroup$ Sep 26, 2013 at 15:56
  • 2
    $\begingroup$ The only problem with copying my references is that you copied my typo. I can't fix it in the comments at 38971, but I did fix it here. Incidentally, Senge & Straus also published their result in another paper of the same title, Period. Math. Hungar. 3 (1973) 93–100, MR0340185 (49 #4941). $\endgroup$ Sep 26, 2013 at 23:26
  • 2
    $\begingroup$ What is $7$ achieving in $S$? Is $\{2,3,5\}$ insufficient? $\endgroup$
    – Will Sawin
    Jan 11, 2014 at 18:51
  • 1
    $\begingroup$ @WillSawin: Your are right, it is enough to take $S = \{2,3,5\}$. (I was thinking of accomodating the prime factors of every digit in $\{1,\ldots,9\}$, but this is unnecessary since e.g. we may split $7 \cdot 10^k \cdot 3^{-m}$ into $7$ new $\{2,3,5\}$-unit variables $= 10^k \cdot 3^{-m}$.) $\endgroup$ Jan 11, 2014 at 19:52

There is more simple agrument which solves the problem: powers $m^n$ can start with arbitrary string of digits.

  • $\begingroup$ I don't think that works, that just proves that the lim sup is infinite, but not that the limit exists. $\endgroup$
    – R.P.
    Nov 2, 2020 at 11:53

I just want to give some other proof to your second question. Let $SS(m^n)=\lbrace S(m^n) ; n\geq1\rbrace$. We show that this set is unbounded. First a Lemma:

$Lemma:$ Let $a$ be any arbitrary integer unless your trivial cases. For any arbitrary integer $b$, there is an integer $t$ such that $a^t$ begins with $b$.

Now by contrary, suppose $Max$ $SS(m^{n_1})=M$ for some $n_1\in N$, and just look to this sequence of integer:

$\overline {m^{n_1}1}, \overline {m^{n_1}11}, \overline {m^{n_1}111}, \ldots$

this sequence is infinite and by previous lemma, for each of them you must have a suitable $t_i$, $i$ is the position of each number in the above sequence, that $m^{t_i}$ begins with $i$-th term in the sequences. Obviously, each of them has the digit sum greater than $M$ and it is a contradiction. This completes the proof.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.