Odd Numbers not Sum of Prime and Power
Theorem
The sequence of odd numbers which cannot be expressed as the sum of a perfect power and a prime number begins:
- $1, 5, 1549, 1 \, 771 \, 561, \ldots$
This sequence is A119747 in the On-Line Encyclopedia of Integer Sequences (N. J. A. Sloane (Ed.), 2008).
It is not known if there are any more terms.
Proof
The cases $1$ and $5$ are trivial.
Now we show that $1549 - a^b$ is never prime for $a \ge 1$ and $b \ge 2$.
It suffices to show the result for prime values of $b$.
We first prove:
\(\text {(1)}: \quad\) | \(\ds 2\) | \(\divides\) | \(\ds a\) | |||||||||||
\(\text {(2)}: \quad\) | \(\ds 3\) | \(\divides\) | \(\ds a\) | \(\ds \text {if } b = 2\) | ||||||||||
\(\text {(3)}: \quad\) | \(\ds a^b\) | \(\not \equiv\) | \(\ds 4\) | \(\ds \pmod {10}\) |
For $(1)$:
Suppose $a$ is odd.
Then so is $a^b$.
Therefore $1549 - a^b$ is even.
Hence $1549 - a^b$ is prime only if it is equal to $2$.
But $1547$ is not a perfect power.
$\Box$
For $(2)$:
Suppose $b = 2$ and $a$ is not divisible by $3$.
By Corollary to Square Modulo $3$:
- $3 \divides \paren {a^2 - 1}$
Hence:
- $3 \divides \paren {1548 - a^2 + 1} = 1549 - a^b$
So $1549 - a^b$ is prime only if it is equal to $3$.
But $1546$ is not a perfect square.
$\Box$
For $(3)$:
Suppose $a^b \equiv 4 \pmod {10}$.
Then $1549 - a^b \equiv 5 \pmod {10}$
So $1549 - a^b$ is prime only if it is equal to $5$.
But $1544$ is not a perfect power.
$\Box$
Hence we only need to check the cases:
- For $b = 2: \: 6^2, 12^2, 18^2, 24^2, 30^2, 36^2$ since $42^2 > 1549$
- For $b \ne 2: \: 2^3, 2^5, 2^7, 4^3, 4^5, 6^3, 8^3, 10^3$ since $2^{11}, 4^7, 6^5 > 1549$
of which $12^2 \equiv 18^2 \equiv 4^3 \equiv 4^5 \equiv 4 \pmod {10}$, so they are rejected.
We have:
\(\ds 1549 - 6^2\) | \(=\) | \(\ds 1513\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 17 \times 89\) | ||||||||||||
\(\ds 1549 - 24^2\) | \(=\) | \(\ds 973\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 7 \times 139\) | ||||||||||||
\(\ds 1549 - 30^2\) | \(=\) | \(\ds 649\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 11 \times 59\) | ||||||||||||
\(\ds 1549 - 36^2\) | \(=\) | \(\ds 253\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 11 \times 23\) | ||||||||||||
\(\ds 1549 - 2^3\) | \(=\) | \(\ds 1541\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 23 \times 67\) | ||||||||||||
\(\ds 1549 - 2^5\) | \(=\) | \(\ds 1517\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 37 \times 41\) | ||||||||||||
\(\ds 1549 - 2^7\) | \(=\) | \(\ds 1421\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 7^2 \times 29\) | ||||||||||||
\(\ds 1549 - 6^3\) | \(=\) | \(\ds 1333\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 31 \times 43\) | ||||||||||||
\(\ds 1549 - 8^3\) | \(=\) | \(\ds 1037\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 17 \times 61\) | ||||||||||||
\(\ds 1549 - 10^3\) | \(=\) | \(\ds 549\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 3^2 \times 61\) |
and none of the above are prime.
$\Box$
For $11^6$, note the factorizations from Difference of Two Powers:
- $11^6 - m^2 = \paren {11^3 - m} \paren {11^3 + m}$
- $11^6 - n^3 = \paren {11^2 - n} \paren {1 + 11^2 n + 11^4 n^2}$
therefore if $11^6 - a^b$ is a prime, $b$ cannot be divisible by $2$ or $3$, unless $11^3 - m$ or $11^2 - n$ is equal to $1$.
For $m = 11^3 - 1$ and $n = 11^2 - 1$:
- $11^6 - m^2 = 2661 = 3 \times 887$
- $11^6 - n^2 = 1757161 = 7 \times 173 \times 1451$
and both are not primes.
Hence none of the above cases hold.
$\Box$
Since $\log_2 \paren {11^6} \approx 20.8$, we must have $b \le 19$.
By the conditions above, $b$ can only be one of $5, 7, 11, 13, 17, 19$.
Moreover $a$ cannot be odd unless $11^6 - a^b = 2$.
However $11^6 - 2 = 1771559$ is prime, and thus not a perfect power.
The cases we need to check are:
- $a^b = 2^5, 6^5, 10^5, 12^5, 14^5, 2^7, 6^7, 2^{11}, 2^{13}, 2^{17}, 2^{19}$
since $18^5, 10^7, 6^{11} > 11^6$, and $4, 8, 16$ are square/cube.
We have:
\(\ds 11^6 - 2^5\) | \(=\) | \(\ds 1771529\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 23 \times 77023\) | ||||||||||||
\(\ds 11^6 - 6^5\) | \(=\) | \(\ds 1763785\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 5 \times 352757\) | ||||||||||||
\(\ds 11^6 - 10^5\) | \(=\) | \(\ds 1671561\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 3^2 \times 79 \times 2351\) | ||||||||||||
\(\ds 11^6 - 12^5\) | \(=\) | \(\ds 1522729\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 13 \times 117133\) | ||||||||||||
\(\ds 11^6 - 14^5\) | \(=\) | \(\ds 1233737\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 311 \times 3967\) | ||||||||||||
\(\ds 11^6 - 2^7\) | \(=\) | \(\ds 1771433\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 31 \times 57143\) | ||||||||||||
\(\ds 11^6 - 6^7\) | \(=\) | \(\ds 1491625\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 5^3 \times 11933\) | ||||||||||||
\(\ds 11^6 - 2^{11}\) | \(=\) | \(\ds 1769513\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 17 \times 104089\) | ||||||||||||
\(\ds 11^6 - 2^{13}\) | \(=\) | \(\ds 1763369\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 41^2 \times 1049\) | ||||||||||||
\(\ds 11^6 - 2^{17}\) | \(=\) | \(\ds 1640489\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 31 \times 52919\) | ||||||||||||
\(\ds 11^6 - 2^{19}\) | \(=\) | \(\ds 1247273\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds 17 \times 73369\) |
and none of the above are prime.
$\Box$
![]() | This theorem requires a proof. In particular: In addition to the above, we also need to demonstrate that for $5 < n < 1549$, an odd number does have such an expression. You can help $\mathsf{Pr} \infty \mathsf{fWiki}$ by crafting such a proof. To discuss this page in more detail, feel free to use the talk page. When this work has been completed, you may remove this instance of {{ProofWanted}} from the code.If you would welcome a second opinion as to whether your work is correct, add a call to {{Proofread}} the page. |
Sources
- 1986: David Wells: Curious and Interesting Numbers ... (previous) ... (next): $1549$
- 1997: David Wells: Curious and Interesting Numbers (2nd ed.) ... (previous) ... (next): $1549$