Seven Haven

Find the sum of all positive integers a a such that a 2 ( 2 a a 3 ) + 1 a^2(2^a-a^3)+1 is a power of 7 7 .


The answer is 10.

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.

5 solutions

Calvin Lin Staff
Apr 17, 2014

As most would have realized, by testing small values, a = 10 a=10 seems to be the only solution. However, it is harder to show that this is indeed the only possible solution.


First of all, if a a is odd, then both a 2 a^2 and 2 a a 3 2^a-a^3 are odd, so a 2 ( 2 a a 3 ) + 1 a^2(2^a-a^3)+1 is even, thus it is not a power of 7. 7. So a = 2 c a=2c for some integer c . c.

Suppose a 2 ( 2 a a 3 ) + 1 = 7 b a^2(2^a-a^3)+1=7^b . Replacing a a by 2 c , 2c, we get 4 c 2 ( 4 c 8 c 3 ) = 7 b 1 4c^2(4^c-8c^3)=7^b-1 Modulo 3 , 3, we get c 2 ( 1 + c ) 0 ( m o d 3 ) , c^2(1+c)\equiv 0\pmod 3, so either c 0 c\equiv 0 or c 2 c\equiv 2 modulo 3. 3.

Now we consider this equation modulo 7 7 . We know that 4 3 1 ( m o d 7 ) 4^3 \equiv 1 \pmod{7} . Note that if c 0 ( m o d 3 ) , c\equiv 0 \pmod 3, then 4 c 1 ( m o d 7 ) ; 4^c\equiv 1 \pmod 7; if c 2 ( m o d 3 ) , c\equiv 2 \pmod 3, then 4 c 2 ( m o d 7 ) . 4^c\equiv 2 \pmod 7. Looking at all different residues of c c modulo 7 7 we get that the only possibility is that c 2 ( m o d 3 ) c\equiv 2\pmod 3 and c 5 ( m o d 7 ) . c\equiv 5 \pmod 7. This means c 5 ( m o d 21 ) , c\equiv 5 \pmod {21}, so c = 5 , 26 , 47 , . . . c=5,26,47,... It is easy to check that c = 5 c=5 works, and it corresponds to a = 10. a=10. We will prove that this is the only possibility.

Now considering the same equation modulo 8 , 8, we easily conclude that b b is even, say b = 2 d . b=2d. Then note that 4 c 2 4 c 4c^2\cdot 4^c and 7 b = 7 2 d 7^b=7^{2d} are both perfect squares. Because for c 26 c\geq 26 , 4 c 2 ( 4 c 8 c 3 ) + 1 4c^2\cdot (4^c-8c^3)+1 is less than 4 c 2 4 c , 4c^2\cdot 4^c, it must be less than or equal to ( 2 c 2 c 1 ) 2 . (2c\cdot 2^c-1)^2. This implies 32 c 5 2 2 c 2 c 32c^5\geq 2\cdot 2c\cdot 2^c . This simplifies to 2 c 8 c 4 2^c\leq 8c^4 It is easy to see that this inequality is false for c 26 , c\geq 26, thus c = 5 c=5 is the only solution.

So the answer is 10.

wow!

Sagnik Saha - 7 years, 1 month ago
Cody Johnson
Dec 8, 2013

First observe that a a must be even. This means

a 2 ( 2 a a 3 ) + 1 1 ( m o d 3 ) a 0 , 1 ( m o d 3 ) a 0 , 4 ( m o d 6 ) a^2\left(2^a-a^3\right)+1\equiv1\pmod{3}\implies a\equiv0,1\pmod{3}\implies a\equiv0,4\pmod{6}

For a 0 ( m o d 6 ) a\equiv0\pmod{6} , we see that this is impossible after plugging in a = 6 k a=6k and testing all 0 k 6 0\le k\le6 modulo 7 7 . For a 4 ( m o d 6 ) a\equiv4\pmod{6} , we see that a = 6 k + 4 a=6k+4 is only a residue for k = 7 k + 1 k=7k+1 modulo 7 7 . Hence, a 10 ( m o d 42 ) a\equiv10\pmod{42} . But we see that if k 1 k\ge1 , then a = 42 k + 10 a=42k+10 is a nonresidue, meaning that k = 0 k=0 is the only solution: 10 \boxed{10} .

You've ended your proof very abruptly. I did not understand even a single step of it.

Led Tasso - 7 years, 6 months ago

Few questions here:

Why does the fact that a a is even imply that it is congruent to 0 , 1 ( m o d 3 ) 0, 1 \pmod 3 ? In particular, why must a 2 ( 2 a a 3 ) a^2(2^a - a^3) be divisible by three and why do you know that it is divisible by three when a 1 ( m o d 3 ) a \equiv 1 \pmod 3 ?

How can you find the residue modulo 7 so easily when you test a = 6 k a = 6k and similarly a = 6 k + 4 a = 6k + 4 ?

What do you mean by "if k 1 k \geq 1 then a = 42 k + 10 a = 42k + 10 is a nonresidue?"

Maybe I missed this, but how can you be sure that k = 0 k = 0 even yields a solution? I see that it does by checking, but was there something you said in your proof that I missed?

Thanks

Mike Kong - 7 years, 6 months ago

Log in to reply

a 2 ( 2 a a 3 ) a^2(2^a-a^3) is divisible by 3 because

a 2 ( 2 a a 3 ) = 7 k 1 = ( 7 1 ) ( 7 k 1 + . . . . + 1 ) = 6 ( 7 k 1 + . . . + 1 ) a^2(2^a-a^3)=7^k-1=(7-1)(7^{k-1}+....+1)=6(7^{k-1}+...+1)

Therefore,it is divisible by both 3 and 2..But I am still at a loss to comprehend the other parts of the proof.

Rahul Saha - 7 years, 6 months ago

How do you know that "if k 1 , k\geq1, , then a = 42 k + 10 a=42k+10 is a nonresidue?"

David Altizio - 7 years, 6 months ago

Can you fill in the missing parts of the solution? There are so many gaps in it- I dont see how you go from one to the next.

Nahom Yemane - 7 years, 6 months ago

Doesn't a being congruent to 0 or 1 mod 3 mean a is congruent to 0, 1, 3, or 4 mod 6, not just 0 or 4?

Edit: Also, a can be congruent to 2 mod 3. Take a=5 or a=11, for instance.

Alexander Xue - 7 years, 6 months ago
Soham Zemse
Dec 10, 2013

a^2(2^a - a^3) + 1 = 7^n.

a^2(2^a - a^3) = 7^n - 1.

7^n - 1 is always even.

=> a^2(2^a - a^3) is also always even.

2^a is even so if a were odd, odd x (even - odd) =odd x odd = odd.

and thats bad.

so a is even.

as 7^n is exponential is is never negative.

and 7^n - 1 is positive ( provided n > 0 ).

so we want a^2(2^a - a^3) to be positive.

for that 2^a - a^3 > 0.

2^a > a^3.

and this happens when a =< 10.

so now we have all tools ready, we are going to brute now.

in a^2(2^a - a^3) + 1 = 7^n.

a=10 then a^2(2^a - a^3) + 1 = 10^2(2^10 - 10^3) + 1 = 100(1024 - 1000) + 1 = 2401 = 7^4.

and for a > 10 we get no value as seven exponentials.

we have only a=10 as solution.

precisely

∑a = 10.

how do you justify "and for a > 10 we get no value as seven exponentials."?

Naren Manoj - 7 years, 6 months ago

Log in to reply

Soham says that 2 a > a 3 2^a > a^3 occurs at a 10 a \leq 10 but in fact it occurs at a 10 a \geq 10 , so Soham only checked a 10 a \leq 10 .

Mike Kong - 7 years, 6 months ago

Log in to reply

I also used some random values for a >=10 and accidentally guessed the answer right in the first try. Can you tell me how did you justify that we don't get any value as seven exponential.

Aamira Shabnam - 7 years, 6 months ago
Bingo Maddy
Dec 12, 2013

a=10 is the only value which equals a power of 7; a=10 gives 2401 which is 7^4

Why is that???How can you be sure that values greater than 10 will not be a solution???

Eddie The Head - 7 years, 5 months ago
Sai Arvind
Dec 10, 2013

for the above equation to be true 2^1 must be greater than a^3 . so if u check for values of a for which the above eqn is true ull find that a>=10 satisfies the eqn . so substituting a=10 we get 2401=7^3 The PROBLEM CAN ALSO BE SOLVED IN ANOTHER METHOD power of 7's -->units digit-->can be ----->1,7,3,9 and so we can conclude sayin that a^2(2^a-a^3) is even becoz wen u subtract the 1 in the ques with the units digit we infer that its even so a^2 has to be even becoz if a^2 is odd then a^3 is also odd and we will get a even quantity which can never be a power of 7. Therefore 2^a-a^3 shud be even Checkin for values only possibility a =10 hence PROVED ;)

we can prove that a<=44, let a^2(2^a-a^3)+1=7^n, it' obvious that a is even, let a=2 m ,and a>=10 .both side mod 4, we have 1 =7^n , thus n is even, let n=2 k, then we have a^2 (2^a)-(7^k)^2=a^5-1 then a^5-1=[a (2^(a/2))-7^k][a (2^(a/2))+7^k] [a (2^(a/2))-7^k] is a positive integer so [a (2^(a/2))-7^k>=1 then we have a^5-1>=[a (2^(a/2))+7^k], obviously, a^5>a^5-1 and a (2^(a/2))+7^k>a (2^(a/2)) so a^5>a*(2^(a/2)) ,thus a^4>2^(a/2) then we get a<=44 (which is very loose),

Jiaxin Zhang - 7 years, 6 months ago

Log in to reply

this soln. can be completed if we also note that a 3 ( m o d 7 ) a \equiv 3 \pmod{ 7 } and a 1 ( m o d 3 ) a \equiv 1 \pmod{3} . This can be proved as follows:

  • a square is congruent to 0,1,4,2 mod 7
  • a cube is congruent to 0,1,6 mod 7
  • a power of 2 is congruent to 1,2,4 mod 7 Now, ( 0 , 1 , 4 , 2 ) ( ( 2 , 4 , 1 ) ( 0 , 1 , 6 ) ) + 1 0 ( m o d 7 ) (0,1,4,2)((2,4,1)-(0,1,6))+1 \equiv 0 \pmod {7} . This leads to the following cases:
  • a 2 4 ( m o d 7 ) a^{2} \equiv 4 \pmod{7} , 2 a 4 ( m o d 7 ) 2^{a} \equiv 4 \pmod{7} and a 3 6 ( m o d 7 ) a^{3}\equiv 6 \pmod{7} . The above case implies a 5 ( m o d 7 ) a\equiv 5 \pmod{7} and a 2 ( m o d 3 ) a\equiv 2 \pmod{3} .
  • a 2 2 ( m o d 7 ) a^{2}\equiv 2 \pmod{7} , 2 a 2 ( m o d 7 ) 2^{a}\equiv 2 \pmod{7} and a 3 6 ( m o d 7 ) a^{3}\equiv 6 \pmod{7} . The above case implies a 3 ( m o d 7 ) a\equiv 3 \pmod{7} and a 1 ( m o d 3 ) a\equiv 1 \pmod{3} .
  • a 2 2 ( m o d 7 ) a^{2}\equiv 2 \pmod{7} , 2 a 4 ( m o d 7 ) 2^{a}\equiv 4 \pmod{7} and a 3 1 ( m o d 7 ) a^{3}\equiv 1 \pmod{7} . The above case implies a 4 ( m o d 7 ) a\equiv 4 \pmod{7} and a 2 ( m o d 3 ) a\equiv 2 \pmod{3} .

The first and third case are not possible since a 1 ( m o d 3 ) a\equiv 1 \pmod{3} is the only possibility.(putting a 2 ( m o d 3 ) a\equiv 2 \pmod{3} in the expression given gives 0 mod 3.)

Thus a can take only values of the form 10+21k,k=0,1,... Since a 44 a\leq 44 , the only possibilities are 10,31. a=10 is fine but the unit digit when we put a=31 is 8 which is not possible for a power of 7. Hence the only possibility is a=10.

Utkarsh Agrawal - 7 years, 6 months ago

Log in to reply

At first I thought of that third bullet as well. However, I realized it was inaccurate. Cubing doesn't exactly cycle like that . It is 0, 1, 1, 6, 1, 6, 6. It cycles mod 7, not like powers of 2, which cycles every 3. So you would have to take 21 cases, which if you are like me, would be too lazy to do. Plus it wouldn't lead to the desired result, probably. Furthermore, you could have found this mistake because a can be congruent to 0 mod 3, not just 1. Edit: a can also be congruent to 2 mod 3.

Alexander Xue - 7 years, 6 months ago

Log in to reply

@Alexander Xue It is just coincidence that for square and power of 2 i wrote it in cycles. But it wasn't my intention because it just doesn't matter what the cycle is to write down the third bullet.

All you need to see is what are the various remainders you get when you modout a square, a power of 2 and a cube by 7 and then find out the correct combinations for which the third bullet is congruent to 0 mod 7.

For example when you take 2 from the first bracket, then the combinations

2 ( 2 6 ) + 1 2(2-6)+1 and 2 ( 4 1 ) + 1 2(4-1)+1 only give 0 mod 7.

Similarly you can check for other values 0,1 and 4. You should just get the above 3 cases in which a 0 ( m o d 3 ) a \equiv 0 \pmod{3} doesn't occur because it implies 2 a 1 ( m o d 7 ) 2^{a} \equiv 1 \pmod{7} which is not possible.

Utkarsh Agrawal - 7 years, 6 months ago

Log in to reply

@Utkarsh Agrawal Ah I see. Thanks. But a can be congruent to 2 mod 3. Take 5 for instance. Or 11. For a being congruent to 2 mod 3 and being odd, the equation becomes 1(2-8)+1 which is 1 mod 3, not 0. Also, in that last paragraph, where you said 2^a being congruent to 1 mod 7 is not possible, can you clarify on that? I can't see how it is not possible.

Alexander Xue - 7 years, 6 months ago

Log in to reply

@Alexander Xue a congruent to 2 mod 3 is not possible because then the given expression becomes divisible by 3, which we don't want.

2^a congruent to 1 mod 7 is not possible because it doesn't occur in any(and clearly only) of the 3 cases I've listed above.

Utkarsh Agrawal - 7 years, 6 months ago

Log in to reply

@Utkarsh Agrawal A being congruent to 2 mod 3 does not mean the given expression is divisible by 3...Just check 5 or 11 as I said in my previous comment. You may have been confused by 2^2 being congruent to 1 mod 3, but 2^(x), where x is congruent to 2 mod 3, is not necessarily congruent to 1 mod 3. The order of 2 mod 3 is 2, not 3, so it makes the cases a bit confusing--a^2 and a^3 will cycle in 3 but 2^a cycles in 2.

Alexander Xue - 7 years, 6 months ago

Log in to reply

@Alexander Xue Yes , but actually we have a is even . so 2^(a)is always congruent to 1 mod 3, so if a is congruent to 2 mod 3, a^2(2^a-a^3)+1=a^2(1-a^3)+1=0 mod 3 .

Jiaxin Zhang - 7 years, 6 months ago

@Alexander Xue But a is even as well.

Utkarsh Agrawal - 7 years, 6 months ago

Both solutions are incomplete.You have just proven that a must be even for the entire expression to be odd.You have not shown that no other solution exists apart from a=10.

Rahul Saha - 7 years, 6 months ago

Log in to reply

I'm afraid I also don't know how to prove that there are no solutions a > 10 a>10 .

I considered the possibility that there are simply no a a values such that the expression is a multiple of 7 5 7^5 (in which case the result would follow without much difficulty). It turns out, however, that there are a a values such that the expression is a multiple of 7 5 7^5 ... a = 10 + 3 5 7 4 = 36025 a=10+3\cdot 5 \cdot 7^4 =36025 for example.

Peter Byers - 7 years, 6 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...