4
$\begingroup$

Let $k$ be any positive integer then how to prove that the sequence $$Q_k=10^{\lceil k \log_{10}(n) \rceil }+n^{k-1}$$ Contains infinitely many primes?

It seems like because if you look at some small values of $k$ we can think of infinude of primes $$Q_2=\{11,12,13,104,105,106,107,108,109,1010,1011,1012,1013 \cdots \}$$ $$Q_3=\{11,14,109,116,1025,10251036,1049,1064,1081,10100 ,111331 \cdots\}$$

Is this true? If false then how many terms are prime in the general $Q_n$ sequence?


Where this came from: Consider the sequence of of numbers formed by concatenating (keeping side by side) $n$ and $n^k$ . For $k=2$. The numbers would be $11 ,24,39,416,525,636 ,749,864 \cdots$ Let they form a sequence $s(n,k)$ . Thus $$s(n,2)= \{ 11 ,24,39,416,525,636 ,749,864 \cdots \}$$
Similarly $$s(n,3)=\{11,28,327,464,5125 \cdots \}$$ We can easily see that $$s(n,k)=n \cdot 10^{\lceil k \log(n) \rceil }+n^k$$ The sequence in question $Q_k$ is just $$Q_k= \frac{s(n,k)}{n}$$

$\endgroup$
6
  • $\begingroup$ this seems interesting, how this question came up? Is it a subproblem of something else? or just experimenting with sequences of numbers? $\endgroup$ – clark May 21 '14 at 4:53
  • $\begingroup$ @clark just experimenting with sequences of numbers $\endgroup$ – Shivam Patel May 21 '14 at 4:54
  • 1
    $\begingroup$ Could be, but there are, in essence, no sequences that depend on a single variable ( your $n$ ) where it can be proved that infinitely many primes occur. Plenty of functions of two variables, any integer quadratic form that cannot be factored for example, as $x^2 + y^2$ for example. $\endgroup$ – Will Jagy May 21 '14 at 5:08
  • $\begingroup$ Is there any significant to writing Qn in this form? It seems for straightforward to simplify to Qn = $n^k$ + n $\endgroup$ – beanshadow May 21 '14 at 5:08
  • $\begingroup$ @beanshadow there is a meaning ,so let me edit the question $\endgroup$ – Shivam Patel May 21 '14 at 5:18
1
$\begingroup$

The exponent $\lceil k\log_{10}n\rceil$ is constant for large stretches of $n$, so the question can be rephrased:

For each integer $k>0$ are there infinitely many integers $e$ with at least one prime $10^e+n^{k-1}$ with $10^{e/k}\le n<10^{(e+1)/k}$?

For $k=1$ this is

Are there infinitely many integers $e$ with $10^e+1$ prime?

which is probably false (these are generalized Fermat primes with base 10, so $e$ must be a power of 2 and the expected number of primes is just 2, 11 and 101). For $k=2$ this is

Are there infinitely many integers $e$ with at least one prime $10^e+n$ with $10^{e/2}\le n<10^{(e+1)/2}$?

which is true under stronger-than-currently-known results on prime gaps.

For larger $k$ this is surely unknown, since it requires finding primes in quadratic or higher polynomials.

$\endgroup$
9
  • $\begingroup$ thankyou for throwing light on the question. Can you tell me which is the closest result we have got to this? $\endgroup$ – Shivam Patel May 22 '14 at 3:22
  • $\begingroup$ @ShivamPatel: The conjecture on generalized Fermat numbers is folklore. the best current result on prime gaps is Baker-Harman-Pintz 2001, "The difference between consecutive primes, II" (which is 'nearly' good enough, though even RH does not suffice). Relevant to the $k>2$ case is the Bateman–Horn–Stemmler conjecture. $\endgroup$ – Charles May 22 '14 at 11:07
  • $\begingroup$ I am not able to understand the -The Bateman–Horn–Stemmler conjecture " analogue here? $\endgroup$ – Shivam Patel May 23 '14 at 6:10
  • $\begingroup$ @ShivamPatel: Read the Bateman, Horn, and Stemmler paper (or the shorter Bateman and Horn paper), or at least the Wikipedia article, and compare to what I wrote in my answer. $\endgroup$ – Charles May 23 '14 at 13:35
  • 1
    $\begingroup$ You need only $f_1$, which is your $Q_k$. $\endgroup$ – Charles May 23 '14 at 14:27

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.