Phi-ratio

Find the sum of all integers n n such that 2 n 100 2 \leq n \leq 100 and n ϕ ( n ) \dfrac{n}{\phi (n)} is an odd integer, where ϕ ( n ) \phi (n) denotes the number of positive integers less than n n coprime to n n .


The answer is 366.

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.

2 solutions

Michael Tang
Jan 2, 2014

Let the distinct prime factors of n n be p 1 , , p k . p_1, \ldots, p_k. Then,

ϕ ( n ) = n ( 1 1 p 1 ) ( 1 1 p k ) = n ( p 1 1 p 1 ) ( p k 1 p k ) , \phi(n) = n\left(1-\dfrac{1}{p_1}\right)\ldots\left(1-\dfrac{1}{p_k}\right) = n\left(\dfrac{p_1-1}{p_1}\right)\ldots\left(\dfrac{p_k-1}{p_k}\right), so

n ϕ ( n ) = ( p 1 p 1 1 ) ( p k p k 1 ) . \dfrac{n}{\phi(n)} = \left(\dfrac{p_1}{p_1-1}\right)\ldots\left(\dfrac{p_k}{p_k-1}\right).

Since this is an integer, it must be that ( p 1 1 ) ( p k 1 ) (p_1-1)\ldots(p_k-1) divides p 1 p k . p_1\ldots p_k. We investigate this very strong condition.

First, if all of the p i p_i are odd, then ( p 1 1 ) ( p k 1 ) (p_1-1)\ldots(p_k-1) is even, while p 1 p k p_1\ldots p_k is odd, so the condition cannot hold. Thus, one of the p i , p_i, say p 1 , p_1, must equal 2 , 2, and the other p i p_i must be odd. Our condition becomes

( 2 1 ) ( p 2 1 ) ( p k 1 ) 2 p 2 p k (2-1)(p_2-1)\ldots(p_k-1) \mid 2p_2\ldots p_k or

( p 2 1 ) ( p k 1 ) 2 p 2 p k . (p_2-1)\ldots(p_k-1) \mid 2p_2\ldots p_k. Now, note that the right-hand side has exactly one factor of 2 , 2, since all of p 2 , , p k p_2, \ldots, p_k are odd; since the left-hand side is composed of even factors, we can only have one factor on the left-hand side. That is, our condition reduces to just

p 2 1 2 p 2 . p_2-1 \mid 2p_2. Then there must be some positive integer a a such that 2 p 2 = a ( p 2 1 ) , 2p_2 = a(p_2-1), or 2 p 2 = a p 2 a . 2p_2 = ap_2 - a. Taking mod p 2 , p_2, we get a 0 ( m o d p ) . a \equiv 0 \pmod p. So, we let a = b p 2 a = bp_2 for some positive integer b , b, and the equation becomes

2 p 2 = ( b p 2 ) p 2 b p 2 2p_2 = (bp_2)p_2 - bp_2 which gives

2 = b ( p 2 1 ) . 2 = b(p_2-1). This means that p 2 1 p_2-1 is either 1 1 or 2 2 ; in the first case, we get p 2 = 2 , p_2 = 2, which contradicts the assumption that the p i p_i are distinct; therefore the second case must hold, so p 2 = 3. p_2 = 3.

Thus, n = 2 c 3 d n = 2^c3^d for some positive integers c , d . c, d. This implies that ϕ ( n ) = n 1 2 2 3 = n 3 , \phi(n) = n \cdot \dfrac12 \cdot \dfrac23 = \dfrac n3, so n ϕ ( n ) = 3. \dfrac{n}{\phi(n)} = 3. (This is an interesting discovery - if n ϕ ( n ) \dfrac{n}{\phi(n)} is an integer, then it is an odd integer.) So, all n n in the form 2 c 3 d 2^c3^d work. Going through all cases, we get the possible values

n = 6 , 18 , 54 , 12 , 36 , 24 , 72 , 48 , 96 , n = 6, 18, 54, 12, 36, 24, 72, 48, 96, the sum of which is 366 . \boxed{366}.

MSTang, your statement is wrong:

If n ϕ ( n ) \frac{n}{\phi(n)} is an integer, then it is an odd integer.

Don't forget when its only prime factor is 2 2 .

Cody Johnson - 7 years, 5 months ago

Log in to reply

Yeah, when n = 2 k , n φ ( n ) = 2 n=2^k, \frac{n}{\varphi(n)}=2 .

Bogdan Simeonov - 7 years, 5 months ago

Log in to reply

oh, yeah. Sorry about that.

Michael Tang - 7 years, 5 months ago

Well, the interesting discovery would be that whenever n ϕ ( n ) \dfrac{n}{\phi (n)} is an integer, it must be either 2 2 or 3 3 . Hmm, I should have stored this for Proofathon...

Sreejato Bhattacharya - 7 years, 5 months ago

Let π 2 ( n ) \pi_2(n) denote the highest power of 2 2 that divides n n . For example, 40 = 2 3 × 5 40= 2^3 \times 5 , so π 2 ( 40 ) = 3 \pi_2(40)= 3 . Observe that for n ϕ ( n ) \dfrac{n}{\phi (n)} to be an odd integer, we must have π 2 ( n ) = π 2 ( ϕ ( n ) ) \pi_2(n)= \pi_2(\phi (n)) .

*Lemma: * If gcd ( a , 2 ) = 1 \gcd (a, 2)= 1 and b b is any integer, π 2 ( a b ) = π 2 ( b ) \pi_2(ab)= \pi_2(b) .

*Proof: * Trivial

*Lemma: * For all a 1 , a 2 , , a n a_1, a_2, \cdots , a_n , π 2 ( a 1 a 2 a n ) = π 2 ( a 1 ) + π 2 ( a 2 ) + + π 2 ( a n ) \pi_2 (a_1 a_2 \cdots a_n)= \pi_2(a_1) + \pi_2(a_2) + \cdots + \pi_2(a_n) .

*Proof: * Also trivial; left to the reader

*Claim: * If π 2 ( n ) = π 2 ( ϕ ( n ) ) \pi_2 (n)= \pi_2 (\phi (n)) , then n = 2 m p j n= 2^m p^j for some positive integers m , j m, j , where p p is a prime of the form 4 k + 3 4k+3 .

*Proof: * Let n = 2 a 0 × p 1 a 1 p 2 a 2 p n a n n= 2^{a_0} \times p_1^{a_1} p_2^{a_2} \cdots p_n^{a_n} , where p 1 , p 2 , , p n p_1, p_2, \cdots , p_n are odd primes and a 0 , a 1 , , a n a_0, a_1, \cdots , a_n are positive integers. We then have π 2 ( n ) = a 0 \pi_2(n)= a_0 . Since ϕ ( n ) \phi (n) is always even, a 0 a_0 cannot be zero. Again, since π 2 ( ϕ ( 2 m ) ) = π 2 ( 2 m ( 1 1 2 ) ) = m 1 \pi_2 \left ( \phi (2^m) \right ) = \pi_2 \left ( 2^m \left ( 1 - \dfrac{1}{2} \right ) \right ) = {m-1} , n 0 n \neq 0 . Now, note that ϕ ( n ) = n ( 1 1 2 ) ( 1 1 p 1 ) ( 1 1 p n ) = 2 a 0 1 p 1 a 1 1 p 2 a 2 1 p n a n 1 p 1 a 1 1 p 2 a 2 1 p n a n 1 ( p 1 1 ) ( p 2 1 ) ( p n 1 ) \begin{array}{lcl} \phi (n) &=& n \left ( 1 - \frac{1}{2} \right ) \left ( 1 - \frac{1}{p_1} \right ) \cdots \left ( 1 - \frac{1}{p_n} \right ) \\ & = & 2^{a_0-1} p_1^{a_1-1} p_2^{a_2-1} \cdots p_n^{a_n-1} p_1^{a_1-1} p_2^{a_2-1} \cdots p_n^{a_n-1} (p_1-1)(p_2-1)\cdots (p_n-1)\end{array} Since gcd ( 2 , p 1 a 1 1 p 2 a 2 1 p n a n 1 ) = 1 \gcd (2, p_1^{a_1-1} p_2^{a_2-1} \cdots p_n^{a_n-1} )= 1 , π 2 ( ϕ ( n ) ) = a 0 1 + π 2 ( ( p 1 1 ) ( p 2 1 ) ( p n 1 ) ) = a 0 1 + π 2 ( p 1 1 ) + π 2 ( p 2 1 ) + + π 2 ( p n 1 ) \begin{array}{lcl} \pi_2(\phi (n))&= &a_0-1 + \pi_2((p_1-1)(p_2-1)\cdots (p_n-1)) \\ & = & a_0-1 + \pi_2 (p_1-1) + \pi_2 (p_2-1) + \cdots + \pi_2(p_n-1) \end{array} This has to be equal to a 0 a_0 , so π 2 ( p 1 1 ) + π 2 ( p 2 1 ) + + π 2 ( p n 1 ) = 1 \pi_2 (p_1-1) + \pi_2 (p_2-1) + \cdots + \pi_2(p_n-1)= 1 Observe that p 1 1 , p 2 1 , , p n 1 p_1-1, p_2-1, \cdots , p_n - 1 are all even, so π 2 ( p i 1 ) 1 \pi_2(p_i-1) \geq 1 for all i i . Then, i = 1 n π 2 ( p i 1 ) n \sum \limits_{i=1}^{n} \pi_2 (p_i-1) \geq n Since this has to be equal to 1 1 , we find out n = 1 n= 1 . So, n = 2 m p n n= 2^m p^n for some prime p p . Also, π 2 ( p 1 ) = 1 \pi_2 (p-1)= 1 . If p p is of the form 4 k + 1 4k+1 , 4 p 1 4 \mid p-1 , and hence π 2 ( p 1 ) > 1 \pi_2 (p-1) > 1 . So, p p must be a prime of the form 4 k + 3 4k+3 . QED

Now, let n = 2 m ( 4 k + 3 ) n n= 2^m (4k+3)^n . We then find out ϕ ( n ) = 2 m 1 ( 4 k + 3 ) n 1 ( 4 k + 2 ) = 2 m ( 4 k + 3 ) n 1 ( k + 1 ) \phi (n)= 2^{m-1} (4k+3)^{n-1} (4k+2) = 2^m (4k+3)^{n-1}(k+1) . Then, n ϕ ( n ) = 2 m ( 4 k + 3 ) n 2 m ( 4 k + 3 ) n 1 ( k + 1 ) = 4 k + 3 k + 1 = 4 1 k + 1 \begin{array}{lcl} \frac{n}{\phi (n)}&= & \frac{2^m (4k+3)^n}{2^m (4k+3)^{n-1}(k+1)} \\ & = & \frac{4k+3}{k+1} \\ & = & 4- \frac{1}{k+1} \end{array} The only possible solution is k = 0 k= 0 , so n = 2 m 3 n n= 2^m3^n for some positive integers m m and n n . Indeed, we can verify that in this case, n ϕ ( n ) = 3 \dfrac{n}{\phi (n)}= 3 . Quickly listing all possible such numbers from 2 2 to 100 100 , we find out that the set of valid n n is { 6 , 12 , 18 , 24 , 36 , 48 , 54 , 72 , 96 } \{6,12,18,24,36,48,54,72,96\} . The sum is 366 \boxed{366} .

The standard notation for your π \pi function is v 2 ( n ) . v_2(n).

Michael Tang - 7 years, 5 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...