LTE Challenge Problems: Amir Hossein Parvardi December 2017
LTE Challenge Problems: Amir Hossein Parvardi December 2017
LTE Challenge Problems: Amir Hossein Parvardi December 2017
Abstract
These problems were solved by Fedja Nezarov, who helped me a lot during writing the Lifting The
Exponent Lemma article.
Problem 1. Let k be a positive integer. Find all positive integers n such that 3n |2n − 1.
Solution. 2|n otherwise 2n − 1 ≡ 1 mod 3. If n = 2m, we should have
v3 (4m − 1) = v3 (m) + 1 ≥ k,
ap ≡ 1 (mod pn ).
Prove that
a≡1 (mod pn−1 ).
Solution. By Fermat, a ≡ ap ≡ 1 mod p, so
vp (a − 1) = vp (ap − 1) − 1 ≥ n − 1.
Problem 3. Show that the only positive integer value of a for which 4(an + 1) is a perfect cube for all
positive integers n, is 1.
Solution. If a > 1, a2 + 1 is not a power of 2 (because it is > 2 and either 1 or 2 modulo 4). Choose some
odd prime p|a2 + 1. Now, take some n = 2m with odd m and notice that vp (4(an + 1)) = vp (a2 + 1) + vp (m)
but vp (m) can be anything we want modulo 3.
Problem 4. Let k > 1 be an integer. Show that there exists infinitely many positive integers n such that
n|1n + 2n + 3n + · · · + k n .
Solution. If 1 + k is not a power of 2, choose an odd prime p|1 + k and take n = pm . Then, for each j not
divisible by p, we have
vp (j n + (k + 1 − j)n ) = vp (k + 1) + vp (n) ≥ m + 1.
Also, if p|j (and, thereby, p|k + 1 − j), then n|pm |pn |j n so the sum in question is divisible by pm = n. If 1 + k
is a power of 2, then take an odd prime divisor p of k and repeat the above argument with k − 1 instead of
k (the last term k n is, obviously, not a problem)
Problem 5. Let p be a prime number, and a and n positive integers. Prove that if
2p + 3p = an .
then n = 1.
1
Solution. 22 + 32 = 13, so assume that p is odd. Then 2p + 3p ≡ 2 mod 3, so it cannot be a square. But
v5 (2p + 3p ) = 1 + v5 (p) ≤ 2.
Problem 6. Find all positive integers n for which there exist positive integers x, y and k such that gcd(x, y) =
1, k > 1 and 3n = xk + y k .
Solution. k must be odd since the sum of 2 squares is divisible by 3 only if both squares are. If p|x + y,
then p is odd and vp (3n ) = vp (xk + y k ) = vp (k) + vp (x + y), which means that p = 3, so x + y = 3m and
n = v3 (k) + m. Now it is just cases. a) m > 1. Then v3 (k) ≤ k − 2 for all k > 1, and M = max(x, y) ≥ 5 so
1 m k−1
xk + y k ≥ M k > 3 5 > 3m 5k−2 ≥ 3m+k−2 ≥ 3m+v3 (k) = 3n ,
2
which gives an immediate contradiction. b) m = 1. Then x = 1, y = 2 (or vice versa) and we get 31+v3 (k) =
1 + 2k , meaning k ≤ 2(1 + v3 (k)) whence v3 (k) = 1, so n = 2 giving the only solution 32 = 13 + 23 .
Problem 7. Let x, y, p, n, k be positive integers such that n is odd and p is a prime. Prove that if xn + y n =
pk , then n is a power of p.
Solution. x + y|xn + y n = pk , so x + y = pm . Now divide x and y by the highest power of p they contain
(it has to be the same). This may change k and m but not n in our condition. Then use the LTE to get
m + vp (n) = k, so xn + y n = (x + y)pvp (n) . If n 6= pvp (n) , we get n ≥ 2pvp (n) ≥ 2, so
n
M n < xn + y n ≤ (x + y) ≤ nM
2
1
and M < n n−1 , which is less than 2 for odd n ≥ 3 but the case M = 1 is impossible.
Problem 8. Let p be a prime number. Solve the equation ap − 1 = pk in the set of positive integers.
Solution. By Fermat, a − 1 ≡ ap − 1 ≡ 0modp. a) p is odd. Then vp (a − 1) = k − 1, so a > pk−1 and, if
k > 1, then ap > pp(k−1) > p3(k−1) > pk . Thus, in this case k = 1 and a = (p + 1)1/p , which is never an
integer (because it is strictly between 1 and 2). b) p = 2. Then a2 − 1 = 2k but, unless a = 3, either a − 1
or a + 1 is not a power of 2. So, the only solution is 32 − 1 = 23 .
Problem 9. Find all solutions of the equation
(n − 1)! + 1 = nm
in positive integers.
Solution. n must be 1 or prime (otherwise any nontrivial divisor d of n will divide both (n − 1)! and nm .
Now, n = 2, m = 1 is a solution, n = 1 is not, so it suffices to consider the case when n is an odd prime. If
m is odd, then v2 (nm − 1) = v2 (n − 1) < v2 ((n − 1)!) if n > 3. n = 3 is not a solution, so we can consider
only even m. Then
v2 ((n − 1)!) = v2 (m) − 1 + v2 (n + 1) + v2 (n − 1)
or
v2 ((n − 2)!) = v2 (n + 1) + v2 (m) − 1.
n−3
But the left hand side is at least 2 (just count evens up to n − 2), so m ≥ 2(n−1)/2 /(n + 1), which is at
least n for n ≥ 18. It remains to note that nn > (n − 1)! + 1 for n > 1. Now comes the remaining finite trial
and error part: 4! + 1 = 52 is good 6! + 1 = 721 is bad n = 11 gives v2 (m) = 9 which is far too large. n = 13
gives v2 (m) = 10, which is too large too n = 17 gives some big v2 (m) as well.
Problem 10. For some positive integer n, the number 3n − 2n is a perfect power of a prime. Prove that n
is a prime.
Solution. Assume n = ab, a, b > 1. Then 3a − 2a = pm with m ≥ 1. Thus, by LTE, vp (3n − 2n ) = m + vp (b)
and
pm+vp (b) = 3ab − 2ab > (3a − 2a )b = pmb
but for b > 1, one has mb ≥ m + b − 1 ≥ m + vp (b), which gives a contradiction.
2
Problem 11. Let m, n, b be three positive integers with m 6= n and b > 1. Show that if prime divisors of
the numbers bn − 1 and bm − 1 be the same, then b + 1 is a perfect power of 2.
Solution. I failed to find a way to use the LTE here. The way I solved it is as follows. Let u = gcd(m, n).
Then bu − 1 and b2u − 1 have the same prime divisors (this uses the proof of the principle I mentioned rather
than the statement itself: when you repeat going from m, n to m − n, n, you stop when you get two equal
numbers). But then each prime dividing bu + 1 has to divide bu − 1 whence bu + 1 is a power of 2. If u were
even, the remainder of the LHS modulo 4 would be 2, so u is odd. Then b + 1|bu + 1 and must be a power
of 2 too.
Problem 12. Find the highest degree k of 1991 for which 1991k divides the number
1992 1990
19901991 + 19921991 .
Solution. Using p = 11 and p = 181, we get
1992
vp (19901991 + 1) = 1 + 1992 = 1993
and 1990
vp (19921991 − 1) = 1 + 1990 = 1991.
Since vp (a + b) = min(vp (a), vp (b)) when vp (a) 6= vp (b), the answer is 1991.
Problem 13. Let p be a prime number and m > 1 be a positive integer. Show that if for some positive
integers x > 1, y > 1 we have m
xp + y p
x+y
= ,
2 2
then m = p.
p p
x+y p
Solution. Since x +y
2 ≥ 2 , we must have m ≥ p. Now, factoring out d = gcd(x, y) and writing
x = dx1 , y = dy1 , we get
2m−1 (xp1 + y1p ) = dm−p (x1 + y1 )m .
Assume that p is odd. Take any prime divisor q|x1 + y1 and let v = vq (x1 + y1 ). If q is odd, we get
v + 1 ≥ v + vq (p) ≥ mv whence m ≤ 2 and p ≤ 2, giving an immediate contradiction. If q = 2, we get
m − 1 + v ≥ mv, so v ≤ 1 and x1 + y1 = 2, i.e., x = y, which immediately implies m = p. If p = 2, we just
notice that 2 3
x2 + y 2
x+y x+y
<2 ≤
2 2 2
if x + y ≥ 4, so m = 2 is the only possibility unless {x, y} = {1, 2}, which is easy to outrule.
Problem 14. Find all positive integers x, y such that px − y p = 1, where p is a prime.
Solution. The case p = 2 is easily done mod 4 (y 2 +1 has remainder 2, so 21 −12 = 1 is the only possibility),
so assume that p is odd. Then y + 1 = pm and, by LTE, y p + 1 = pm+1 . But (pm − 1)p + 1 > pm+1 unless
p = 3, m = 1, which gives the second solution.
Problem 15. Let x and y be two positive real numbers such that for each positive integer n, the number
xn − y n is a positive integer. Show that x and y are both positive integers.
Problem 16. Let x and y be two positive rational numbers such that for infinitely many positive integers
n, the number xn − y n is a positive integer. Show that x and y are both positive integers.
Solution. They are very much alike, so I’ll combine the solutions. In 17, start with the observation that
2
−y 2
x − y and x + y = xx−y are rational, so x, y ∈ Q. Now we are in the conditions of 18. Write x = ac , y = cb
where c is the least common denominator of x, y. If c > 1, take any prime divisor p of c. Then pn |an − bn
for infinitely many n and p cannot divide a or b (otherwise it would divide them both and we could reduce
the fractions). Led u be the least power such that p|au − bu . Then all those n’s are mu for some integer m
and we get n ≤ vp (an − bn ) ≤ vp (m) + vp (au − bu ) + vp (au + bu ) (just not to consider 2 separately). But
vp (m) grows only logarithmically in n, so we get a contradiction.
3
Problem 17. Does there exist a positive integer n such that n has exactly 2000 prime divisors and n divides
2n + 1?
m m
Problem 18. Note that 3m |23 +1 by LTE. Thus, if 23 +1 has 1999 distinct prime divisors q1 , . . . , q1999 > 3,
m m+1
n = 3m q1 . . . q1999 will work. Note that each divisor of 23 + 1 is also a divisor of 23 + 1, so the set of
prime divisors either grows without bound or saturates to some finite set P . In the latter case, we have
m mp mp
vp (23 + 1) ≤ (m − mp ) + vp (23 + 1) ≤ m + Cp where mp is the least integer such that p|23 + 1. Thus,
m
23 + 1 ≤ CAm where A is the product of all primes in P , which is absurd.
m
Problem 19. Suppose that m and k are non-negative integers, and p = 22 + 1 is a prime number. Prove
m+1 k
that 22 p
≡ 1 (mod pk+1 ); 2m+1 pk is the smallest positive integer n satisfying the congruence equation
2 ≡ 1 (mod pk+1 ).
n
m k m m k m+1 k
Solution. vp (22 p + 1) = k + vp (22 + 1) = k + 1 by LTE and 22 p + 1|22 p
− 1. Furthermore,
m+1 m
2 2 n m+1
p|2 − 1 but not 2 − 1, so if p|2 − 1, we have 2 |n (powers of 2 have only divisors that are powers
of 2 themselves). If n = Q2m+1 , then
m+1
vp (2n − 1) = vp (Q) + vp (22 − 1) = vp (Q) + 1,
so vp (Q) ≥ k.
Problem 20. Let p ≥ 5 be a prime. Find the maximum value of positive integer k such that
Solution. Let p − 1 = 2s m. Then, since vp ((p − 2)2 + (p − 4)) = vp (p2 − 3p) = 1, vp ((p − 2)2 − (p − 4)) =
vp (p2 − 5p + 8) = 0, and vp (2s−1 m) = 0, we get
Problem 21. Find all positive integers a, b which are greater than 1 and
ba |ab − 1.
Solution. Let p be the least prime divisor of b. Let m be the least positive integer for which p|am − 1. Then
m|b and m|p − 1, so any prime divisor of m divides b and is less than p. Thus, not to run into a contradiction,
we must have m = 1. Now, if p is odd, we have avp (b) ≤ vp (a − 1) + vp (b), so
which is impossible. Thus p = 2, b is even, a is odd and av2 (b) ≤ v2 (a − 1) + v2 (a + 1) + v2 (b) − 1 whence
a ≤ (a − 1)v2 (b) + 1 ≤ v2 (a − 1) + v2 (a + 1), which is possible only if a = 3, v2 (b) = 1. Put b = 2B with odd
B and rewrite the condition as 23 B 3 |32B − 1. Let q be the least prime divisor of B (now, surely, odd). Let
n be the least positive integer such that q|3n − 1. Then n|2B and n|q − 1 whence n must be 2 (or B has a
smaller prime divisor), so q|32 − 1 = 8, which is impossible. Thus B = 1.
Problem 22. Let a, b be distinct real numbers such that the numbers
a − b, a2 − b2 , a3 − b3 , . . .
4
Solution. Assume p is odd. Let x−1 = pm . Then n = m+vp (r), so pm+vp (r) = (pm +1)r −1 ≥ pmr , implying
m + vp (r) ≥ mr ≥ m + r − 1 and vp (r) ≥ r − 1, which is impossible for r > 1. Thus p = 2. If r is odd, we get
n = v2 (xr −1) = v2 (x−1) so r = 1, which is outruled. Thus, r is even and n = v2 (x−1)+v2 (x+1)+v2 (r)−1.
If m = 1, we get the usual 32 − 1 = 23 . Otherwise v2 (x + 1) = 1, so n = m + v2 (r) and we can finish just as
we started.
Problem 24. Let a > b > 1 be positive integers and b be an odd number, let n be a positive integer. If
n
bn | an − 1, then show that ab > 3n .
Solution. Let P be the set of all prime divisors of b. For p ∈ P , let sp be the Q least integer
Q such that
p|asp − 1. We have sp |n, sp |p − 1 and n ≤ vp (asp − 1) + vp (n/sp ). Now note that b ≥ p∈P p > p∈P sp = S
and that asp − 1|aS − 1 for all p ∈ P . Thus,
Y Y
ab > aS − 1 ≥ pn p−vp (n) ≥ 3n /n.
p∈P p∈P
Problem 25. Let p be a prime number, p 6= 3, and integers a, b such that p | a + b and p2 | a3 + b3 . Prove
that p2 | a + b or p3 | a3 + b3 .
Solution. If p|a, b, then p3 |a3 + b3 . Otherwise LTE applies and vp (a + b) = vp (a3 + b3 ) ≥ 2.
Problem 26. Let m and n be positive integers. Prove that for each odd positive integer b there are infinitely
many primes p such that pn ≡ 1 (mod bm ) implies bm−1 | n.
Solution. Let q be any prime divisor of b. If vq (pq−1 − 1) = 1, we get mvq (b) ≤ vq (n) + 1 (using, as before,
that the minimal s satisfying q|ps − 1 divides both n and q − 1) and vq (n) ≥ mvq (b) − 1 ≥ (m − 1)vq (b).
Thus we just need to show that there are infinitely many primes p satisfying the condition vq (pq−1 − 1) = 1
for all prime divisors of b. If one knows Dirichlet, it is simple: just consider p ≡ 1 + q mod q 2 for all primes
q dividing p. If not, I’m stuck for now.
Problem 27. Determine all integers n > 1 such that
2n + 1
n2
is an integer.
v
Solution. Let p be a prime divisor on n, v = vp (n), Q = p−v n. We have p2v |2p Q + 1. We also know that
v
2p ≡ 2 mod p by Fermat. Thus p|2Q + 1 and, by LTE pv |2Q + 1. Let m be the least positive integer such
that pv |2m + 1. Then m|Q, pv−1 (p − 1). If p is the least prime divisor of n, we conclude that m = 1 (because
Q consists of primes larger than p) and p = 3, v = 1. Now, if p is the second smallest prime divisor in n,
then m can be only 3 (the only factor in Q that can occur in p − 1). But 23 + 1 = 9 has no prime divisors
greater than 3 and we are stuck in our attempt to acsend. Thus, the answer is n = 1 or n = 3.
Problem 28. Find all positive integers n such that
2n−1 + 1
n
is an integer.
vj vj
m −v
Solution. Let n = pv11 . . . pm v . As before, put Qj = pj j n. Since pj |2pj Qj −1 +1 and 2pj ≡ 2 (mod pj ) by
Fermat, we get pj |2Qj −1 + 1. Let mj be the least positive integer such that pj |2mj + 1. Then mj divides both
p −1 Q
Qj − 1 and j2 and Qj − 1 is an odd multiple of mj . But the power of 2 in Qj − 1 = k6=j (1 + (pk − 1)) − 1
is at least mink6=j v2 (pk − 1) while v2 (mj ) ≤ v2 (pj − 1) − 1, so choosing pj so that v2 (pj − 1) is minimal, we’ll
get a contradiction.
(5p − 2p )(5q − 2q )
Problem 29. Find all primes p, q such that is an integer.
pq
5
Solution. Assume p ≤ q. If 5p − 2p ≡ 5 − 2 = 3 mod p by Fermat, so if p|5p − 2p , then p = 3. In
particular, if p = q, then p = q = 3. If p < q and p|5q − 2q , then p|5p−1 − 2p−1 by Fermat, so the least
m > 0 such that p|5m − 2m must divide both p − 1 and q, i.e., m = 1, so p = 3 again. Now, either q = 3, or
q|53 − 23 = 117 = 9 · 13, so (3, 13) is the only other solution.
Problem 30. For some natural number n let a be the greatest natural nubmer for which 5n − 3n is divisible
by 2a . Also let b be the greatest natural number such that 2b ≤ n. Prove that a ≤ b + 3.
Solution. If n is odd, a = 1 and there is nothing to prove. If n is even, a = v2 (5n − 3n ) = v2 (5 − 3) + v2 (5 +
3) + v2 (n) − 1 = 3 + v2 (n). But, clearly, b ≥ v2 (n).
Problem 31. Find all surjective functions f : N → N such that for every m, n ∈ N and every prime p, the
number f (m + n) is divisible by p if and only if f (m) + f (n) is divisible by p.
Solution. We start with the (well-known?) observation that every subset S of positive integers that is
closed under addition is an eventual arithmetic progression. More precisely, there exists d ≥ 1 (which
actually is just the greatest common divisor of the elements of S) and N such that for n ≥ N we have
n ∈ S ⇐⇒ d|n. Now, for prime p, let Sp = {n : p|f (n). Let dp be the corresponding difference. Thus
p|f (n) ⇐⇒ dp |n if n is large. Now take any n and take a huge m divisible by dp (so p|f (m)). Then
dp |m ⇐⇒ dp |n + m ⇐⇒ p|f (n + m) ⇐⇒ p|f (n), so the equivalence holds without the requirement that
n is large too. The next step is to show that the remainder of n modulo dp determines the remainder of
f (n) modulo p and vice versa. Let’s take A = {1, 2, . . . dp − 1. For every n there exists the unique a ∈ A
such that dp |n + a determined by n mod dp . But then p|f (n) + f (a) determining f (n) mod p uniquely.
Conversely, let B = {b1 , . . . , bp } so that f (bj ) = j (here is where we use surjectivity). Then once we know
f (n) mod p, we know j such that p|f (n) + j ≡ f (n + bj ) mod p whence we know that dp |n + bj . This
one-to one correspondence between remainders implies that p = dp and that
p|n − m ⇐⇒ p|f (n) − f (m).
In particular, f (1) = 1 and if f (n + 1) − f (n) = ±1 for all n. Now take a huge odd prime P and note that
we can have P |f (P ) only if all ±1 up to P are actually 1. Since P is arbitrarily large, f (n) = n for all n.
Problem 32. Determine all sets of non-negative integers x, y and z which satisfy the equation
2x + 3 y = z 2 .
Solution. This is just a casework: If x = 0, we get 3y = (z − 1)(z + 1), but 1 and 3 are the only two powers
of 3 differing by 2, so y = 1, z = 2. If y = 0, then 2x = (z − 1)(z + 1) giving z = 3, x = 3 in the same way.
If x, y > 0, then x is even (z 2 cannot be 2 mod 3) whence y is even (z 2 cannot be 3 mod 4), so, letting
x = 2X, y = 2Y , we get 32Y = (z − 2X )(z + 2X ). Thus, we must have z = 2X + 1 and 32Y − 1 = 2X+1 . But
then X + 1 = v2 (Y ) + 3 by the LTE, so 2Y ≥ 2X−1 > X + 1 if X ≥ 4. X = 1 gives nothing, X = 2 gives
Y = 1, and X = 3 gives nothing.
Problem 33. Find all positive integer solutions of equation x2009 + y 2009 = 7z .
Solution. 7|2009 so 7|x + y by Fermat. Removing the highest possible power of 7 from x, y, we get
v7 (x2009 + y 2009 ) = v7 (x + y) + v7 (2009) = v7 (x + y) + 2,
so x2009 + y 2009 = 49(x + y) but the left hand side is much larger than the right hand one if max(x, y) > 1.
n
Problem 34. Let n be an odd positive integer. Prove that ((n − 1)n + 1)2 divides n(n − 1)(n−1) +1
+ n.
Solution. n|(n − 1)n + 1, so for every p|(n − 1)n + 1, we have
n
vp ((n − 1)(n−1) +1
+ 1)
(n − 1)n + 1 + 1
= vp ((n − 1)n + 1) + vp
n
= 2vp ((n − 1)n + 1) − vp (n),
which is just what we need in terms of prime divisors.
6
Problem 35. Find all positive integers n such that 3n − 1 is divisible by 2n .
Solution. n ≤ v2 (3n − 1) ≤ 3 + v2 (n), so n ≤ 4. 1, 2, 4 work, 3 doesn’t.
Problem 36. Let p be a prime and a, b be positive integers such that a ≡ b (mod p). Prove that if px ka − b
and py kn, then px+y kan − bn .