125 and 512; 256 and 625

For how many positive integers n < 1000 n<1000 , do there exist integers a , a, b , b, c , c, and d , d, that satisfy the following conditions:

{ 100 a + b = c n 10 b + a = d n g c d ( c , d ) = 1 ? \begin{cases} 100a+b =c^n\\ 10b+a=d^n\\ gcd(c,d)=1? \end{cases}

Details and assumptions

There are no restrictions on the values of a a and b b .


The answer is 888.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Mark Hennings
Dec 3, 2013

I claim that the following three statements are equivalent for a positive integer n n :

  1. There exist integers a , b , c , d a,b,c,d with ( c , d ) = 1 (c,d) = 1 such that 100 a + b = c n 100a + b = c^n and 10 b + a = d n 10b + a = d^n .

  2. There exists an integer u u such that u n 10 u^n \,\equiv \, 10 modulo 999 999 .

  3. There exists an integer v v such that v n 100 v^n \,\equiv \, 100 modulo 999 999 .

If Condition (1) is true, then 10 c n d n = 999 a 0 10c^n - d^n = 999a \,\equiv\, 0 modulo 999 999 . If ( c , 999 ) > 1 (c,999) > 1 , let p p be a common prime factor of c c and 999 999 (either 3 3 or 37 37 ). Then p p also divides d n d^n , and hence p p divides d d , so that c c and d d are not coprime. Thus we deduce that c c and 999 999 are coprime, so there exists an integer e e such that c e 1 ce \, \equiv\, 1 modulo 999 999 , and then ( d e ) n 10 (de)^n \,\equiv\,10 modulo 999 999 . Hence Condition (2) is true.

If Condition (2) is true, so that u n 10 u^n \equiv 10 modulo 999 999 for some u u , then ( u 2 ) n = ( u n ) 2 100 (u^2)^n \,=\, (u^n)^2 \,\equiv\, 100 modulo 999 999 , and hence Condition (3) is true.

If Condition (3) is true, so that v n 100 v^n \equiv 100 modulo 999 999 for some v v , put c = v c=v and d = 1 d=1 . Certainly c c and d d are coprime. Then 10 c n d n = 10 v n 1 10 × 100 1 0 100 d n c n = 100 v n 0 \begin{array}{rcl} 10c^n - d^n & = & 10v^n - 1 \; \equiv \; 10\times100 - 1 \; \equiv \; 0 \\ 100d^n - c^n & = & 100 - v^n \; \equiv \; 0 \end{array} modulo 999 999 , and hence we can find integers a , b a,b such that 10 c n d n = 999 a 100 d n c n = 999 b 10c^n - d^n \; = \; 999a \qquad 100d^n - c^n \; = \; 999b Thus 999 ( 100 a + b ) = 100 ( 10 c n d n ) + ( 100 d n c n ) = 999 c n 999 ( 10 b + a ) = 10 ( 100 d n c n ) + 10 c n d n = 999 d n \begin{array}{rcl} 999(100a + b) & = & 100(10c^n - d^n) + (100d^n - c^n) \; = \; 999c^n \\ 999(10b + a) & = & 10(100d^n - c^n) + 10c^n - d^n \; = \; 999d^n \end{array} so that 100 a + b = c n 100a+b = c^n and 10 b + a = d n 10b+a = d^n , and Condition (1) is satisfied.

Since 1 0 3 = 1000 1 10^3 \,=\, 1000 \,\equiv\, 1 modulo 999 999 , we see that 1 0 3 a + 1 10 10^{3a+1} \,\equiv\,10 for all a 0 a \ge 0 , and hence Condition (2) is satisfied for all integers n 1 n \equiv 1 modulo 3 3 . Also 1 0 3 a + 2 100 10^{3a+2} \,\equiv\, 100 for all a 0 a \ge 0 , and hence Condition (3) is satisfied for all integers n 2 n \equiv 2 modulo 3 3 .

Since 62 5 3 10 625^3 \,\equiv\, 10 modulo 999 999 (thank you, Excel!), we deduce that 62 5 9 a + 3 10 625^{9a+3} \,\equiv\, 10 and 62 5 9 a + 6 100 625^{9a+6} \,\equiv\, 100 modulo 999 999 , and so Conditions (2) and (3) are satisfied for all integers n 3 n \equiv 3 and n 6 n \equiv 6 modulo 9 9 respectively.

So far, we have shown that Condition (1) holds for all integers n n except multiples of 9 9 . I claim that Condition (1) does not hold if n n is a multiple of 9 9 . Suppose that Condition (1) held for some n = 9 a n = 9a . Then we can find an integer u u such that u 9 a 10 u^{9a} \equiv 10 modulo 999 999 . This would imply that ( u a ) 9 10 (u^a)^9 \equiv 10 modulo 999 999 , and hence that v 9 10 v^9 \equiv 10 modulo 999 999 for some integer v v . But (thank you, Excel, again) we can easily check that 10 10 is not a ninth power modulo 999 999 .

Thus we deduce that Condition (1) holds precisely for those integers which are not multiples of 9 9 . There are 111 111 multiples of 9 9 between 1 1 and 999 999 , and hence there are 888 888 integers n n between 1 1 and 999 999 inclusive for which Condition (1) holds.

There's an easy way to see without calculator to see that 10 10 is not a ninth power mod 999 999 : factor 999 999 into 3 3 37 3^3 \cdot 37 . Now clearly 1 0 3 1 ( m o d 999 ) 10^3 \equiv 1 \pmod{999} , so the order of 10 10 mod 3 3 3^3 and mod 37 37 is exactly 3 3 (certainly by the first observation it divides into 3 3 , which leaves only one case to rule out).

Now compare this with ϕ ( 3 3 ) = 18 \phi(3^3) = 18 and ϕ ( 37 ) = 36 \phi(37) = 36 , and use the fact that odd prime powers possess primitive roots. The equivalence of (2) and " n n is not divisible by 9 9 " falls out pretty quickly without needing Excel.

Erick Wong - 7 years, 6 months ago

Log in to reply

Of course, if v 9 10 v^9 \equiv 10 modulo 999 999 then v 27 1 v^{27} \equiv 1 modulo 999 999 , and hence modulo 27 27 and 37 37 as well. Since ϕ ( 27 ) = 18 \phi(27)=18 and ϕ ( 37 ) = 36 \phi(37)=36 , and since ( 27 , 18 ) = ( 27 , 36 ) = 9 (27,18) = (27,36) = 9 , we deduce that v 9 1 v^9 \equiv 1 modulo both 27 27 and 37 37 , and hence 10 v 9 1 10 \equiv v^9 \equiv 1 modulo 999 999 , which is impossible. This shows what I wanted, and that Condition (2) implies that n n is not a multiple of 9 9 . Put that together with what I have written, and we are done.

Having found the equivalence of (1), (2) and (3), and after finding 625 625 as a cube root of 10 10 , I already had Excel set up to investigate ninth powers (just a matter of dragging one column of formulae). It was a totally defined, finite calculation, so why not? You could argue that I did not need to find 625 625 , since the pair 125 125 , 512 512 in the question title shows that n = 3 n=3 works.

Mark Hennings - 7 years, 6 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...