Order in the court!

Find the number of primes p p such that q = 2 p 1 q=2p-1 is a prime and the following is true for all integers a a from 1 1 to p 1 p-1 :

For every positive integer k k , if q q divides a k 1 a^k-1 , then p p also divides a k 1 a^k-1 .


The answer is 2.

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.

4 solutions

C Lim
Jul 28, 2013

First p=2 is clearly a solution so let's assume p is odd. Thus, q 1 ( m o d 4 ) q \equiv 1 \pmod 4 so the Legendre symbol gives ( 1 q ) = + 1 \left(\frac {-1} q\right) = +1 and there exists

1 < x q 1 2 = p 1 , x 2 + 1 0 ( m o d q ) . 1 < x \le \frac{q-1}2 = p-1, \quad x^2 + 1 \equiv 0 \pmod q.

Since this means x 4 1 ( m o d q ) x^4 \equiv 1 \pmod q , by the condition of the problem, we also have x 4 1 ( m o d p ) x^4 \equiv 1 \pmod p . This means p divides x 4 1 = ( x 1 ) ( x + 1 ) ( x 2 + 1 ) x^4 - 1 = (x-1)(x+1)(x^2 + 1) so either x = 1, p -1 or x 2 + 1 x^2 + 1 is a multiple of p .

Case 1 : x = 1, p -1. We already ruled out x=1. For x=p-1, we have ( p 1 ) 2 1 ( m o d q ) (p-1)^2 \equiv -1 \pmod q . Since 2 ( p 1 ) = q 1 2(p-1) = q-1 , this gives ( q 1 ) 2 4 ( m o d q ) (q-1)^2 \equiv -4 \pmod q . Expanding, we get 1 4 ( m o d q ) 1 \equiv -4 \pmod q so q=5. This gives p=3.

Case 2 : x 2 1 ( m o d p ) x^2 \equiv -1 \pmod p . Thus x 2 + 1 x^2 + 1 is a multiple of p and of q, and hence of pq. But this is impossible since

1 x p 1 x 2 + 1 ( p 1 ) 2 + 1 < p q . 1 \le x \le p-1 \implies x^2 + 1 \le (p-1)^2 + 1 < pq.

Thus, the only solutions are ( p , q ) = ( 2 , 3 ) , ( 3 , 5 ) (p, q) = (2, 3), (3, 5) .

Doesnt p=19 work???

avinash iyer - 7 years, 3 months ago
Thomas Beuman
Jul 31, 2013

The cases p = 2 p=2 and p = 3 p=3 are quite trivial; they are solutions. We now assert p > 3 p>3 .

Since p p is odd, we have q 1 ( m o d 4 ) q \equiv 1 \ (\mathrm{mod}\ 4) . By virtue of the first supplement to the law of quadratic reciprocity, 1 -1 is a quadratic residue modulo q q , i.e. there is a positive integer x < q x < q such that x 2 1 ( m o d q ) x^2 \equiv -1 \ (\mathrm{mod}\ q) . Since q x q-x has the same property, and either x x or q x q-x must be smaller than q / 2 q/2 , we can assert that x p 1 x \leq p-1 , so x x is an integer that satisfies the condition for a a as stated in the problem.

We now have that q q divides x 4 1 = ( x 1 ) ( x + 1 ) ( x 2 + 1 ) x^4-1 = (x-1)(x+1)(x^2+1) . If p p is to meet the condition of the problem statement, we must have that p p also divides x 4 1 x^4-1 . In order for that to be true, it must divide one of the factors given above.

If p p were to divide x 1 x-1 , we'd have x = 1 x=1 , which is obviously impossible given the definition of x x . If p p were to divide x 2 + 1 x^2+1 , as q q does, then x 2 + 1 x^2+1 would be divisible by p q pq . However, since x < p < q x < p < q , we obviously have p q > x 2 + 1 pq > x^2+1 , so this is impossible. Therefore, the only remaining possibility is that p p divides x + 1 x+1 , which implies x = p 1 x=p-1 .

By the definition of x x , we'd then necessarily have ( p 1 ) 2 1 ( m o d 2 p 1 ) (p-1)^2 \equiv -1 \ (\mathrm{mod}\ 2p-1) . Since p p is odd, 1 2 ( p 1 ) \frac12(p-1) is integer, and we have

1 ( p 1 ) 2 ( p 1 ) 2 1 2 ( p 1 ) ( 2 p 1 ) 1 2 ( 3 p 1 ) ( m o d 2 p 1 ) . -1 \equiv (p-1)^2 \equiv (p-1)^2 - \frac12(p-1)(2p-1) \\ \quad \equiv \frac12(3p-1) \ (\mathrm{mod}\ 2p-1).

But for p > 3 p>3 , we have 1 < 1 2 ( 3 p 1 ) < 2 p 2 -1 < \frac12(3p-1) < 2p-2 , so this is never true.

Therefore, there are no solutions with p > 3 p>3 . The only solutions are thus p = 2 \boxed{p=2} and p = 3 \boxed{p=3}

I made a mistake: towards the end, the two instances of 1 2 ( p 1 ) \frac12(p-1) should be replaced by 1 2 ( p 3 ) \frac12(p-3) .

By the way, it's funny how I thought my proof was quite laborious, yet it is almost a carbon copy of the other two solutions (so far).

Thomas Beuman - 7 years, 10 months ago

Log in to reply

What seems clear in one's head can often be laborious when we try to write it down..:)

After playing around with it for a while I convinced myself that if q = 4 k + 1 q = 4k + 1 , where k>1, then the fact that ( b k ) 4 1 ( m o d q ) {(b^{k})}^{4} \equiv 1 \pmod q , where b b is a primitive root mod q, exists, means that there are no such solutions (i.e. all solutions are q <= 5). It just seems you can get 1 out of b k b^{k} so much more often than there can possibly be roots of unity mod p.

However I couldn't put my conviction in to a formal argument. I still feel there's got to be a more elegant solution than what's come so far so I will keep working on it :)

Matt McNabb - 7 years, 10 months ago
Lucas Reis
Jul 29, 2013

If p = 2 q = 3 p=2\Rightarrow q=3 and for every a a integer from 1 1 to 2 1 = 1 2-1=1 we have 3 1 k 1 = 0 2 1 k 1 = 0 3|1^k-1=0\Rightarrow 2|1^k-1=0 .

Now, be p > 2 p>2 , p p is a odd prime, p = 2 d + 1 q = 4 d + 1 p=2d+1\Rightarrow q=4d+1 and q q is a prime number.

How we know, for every prime number of the form P = 4 d + 1 P=4d+1 , 1 -1 is a quadractic residue ( m o d P ) \pmod P .

Thus there exists a integer number 1 x q 1\le x\le q such that x 2 1 ( m o d q ) x^2\equiv -1\pmod {q} .

WLG, suppose that x p 1 x\le p-1 because if x > p 1 x p x>p-1\Rightarrow x\ge p , and we have 2 p 1 x 2 p 1 p = p 1 2p-1-x\le 2p-1-p=p-1 and ( 2 p 1 x ) 2 ( x ) 2 x 2 1 ( m o d q ) (2p-1-x)^2\equiv (-x)^2\equiv x^2\equiv -1\pmod{q} .

Thus there exists 1 m p 1 1\le m\le p-1 such that m 2 1 ( m o d q ) m 4 1 ( m o d q ) m^2\equiv -1\pmod {q}\Rightarrow m^4\equiv 1\pmod q and then q m 4 1 q|m^4-1 .

By the enunciate we have p m 4 1 = ( m 2 + 1 ) ( m + 1 ) ( m 1 ) p|m^4-1=(m^2+1)(m+1)(m-1) as well.

If p m 2 + 1 p|m^2+1 , we have p q m 2 + 1 pq|m^2+1 , where p 2 + 1 > m 2 + 1 p q = p ( 2 p 1 ) > p ( p + 1 ) > p 2 + 1 p^2+1>m^2+1\ge pq=p(2p-1)>p*(p+1)>p^2+1 since that p > 2 p>2 , which is a contradiction.

And how m 1 < p m-1<p we can conclude that p m + 1 m p 1 m = p 1 p|m+1\Rightarrow m\ge p-1\Rightarrow m=p-1 since that m p 1 m\le p-1 .

Thus 2 p 1 m 2 + 1 = ( p 1 ) 2 + 1 = p 2 2 p + 2 2p-1|m^2+1=(p-1)^2+1=p^2-2p+2 and 2 p 1 2 p 2 p 2p-1|2p^2-p so we have 2 p 1 2 p 2 p 2 ( p 2 2 p + 2 ) ( 2 p 1 ) = p 3 2p-1|2p^2-p -2*(p^2-2p+2)-(2p-1)=p-3 and since that 0 p 3 < 2 p 1 0\le p-3<2p-1 we must have p 3 = 0 p = 3 , q = 5 p-3=0\Rightarrow p=3, q=5 .

In fact, for every positive integers k , 1 a 3 1 = 2 k, 1\le a\le 3-1=2 , if 5 1 k 1 = 0 , 2 k 1 5|1^k-1=0, 2^k-1 we have 2 k 1 ( m o d 5 ) o r d 5 2 k 4 k , k = 4 M 2^k\equiv 1\pmod 5\Rightarrow ord_{5}2|k\Rightarrow 4|k, k=4M and thus 2 k 1 2 4 M 1 1 6 M 1 1 M 1 0 ( m o d 3 ) 2^k-1\equiv 2^{4M}-1\equiv 16^M-1\equiv 1^M-1\equiv 0\pmod 3 .

Thus 3 1 k 1 = 0 , 2 k 1 3|1^k-1=0, 2^{k}-1 and so we have 2 2 prime numbers with this property.

Really cool ,,my solution is not as elegant as yours!

Bryan Andrade - 7 years, 10 months ago
Damiann Mangan
May 20, 2014

While the most elegant way to provide a solution is by using the general case, giving an extreme case trial could gives us a rough solution, in some cases.

We could deduce that 2 p 1 a 1 2p - 1 \mid a - 1 , for k = 1 k = 1 , the lowest positive integer. While using a = 1 a = 1 will always provide problem requirement, the other extreme case, a = p 1 a = p - 1 , gives us 2 p 1 p 2 2p - 1 \mid p - 2 .

From the last definition, we have either 2 p 1 p 2 2p - 1 \leq p - 2 , because p p is a prime therefore it is or bigger than 2 2 which is positive, or p 2 = 0 p - 2 = 0 , because 0 0 is the only number that dividable by anything.

The latter is the solution, because the other gives us p 1 p \leq -1 which gives us a contradiction. Therefore, p = 2 p = 2 is and the only solution.

How many mistakes are made by this student in understanding the problem?

You may submit solutions to mathematics@brilliant.org.

Calvin Lin Staff - 7 years ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...