Number Theory

1
 Number Theory - Wei LI 2

§1 Divisibility and GCD (I)

n
1. For any n ≥ 0, let Fn = 22 + 1 which is called Fermat number. Prove that

(i) (Fm , Fn ) = 1 for any m 6= n.

(ii) Fn+1 = Fn · · · F0 + 2 for any n ≥ 0.
(iii) The first conclusion implies that there are infinitely many primes in N.

2. Set A1 = 1 and An+1 = A2n − An + 1 for n ≥ 1. Prove that

(i) (Am , An ) = 1 for any m 6= n.

(ii) An+1 = An · · · A1 + 1 for any n ≥ 1.
(iii) The first conclusion implies that there are infinitely many primes in N.

3. If 2n + 1 is a prime, then n = 2k for some k.

4. If 2n − 1 is a prime, then n is also a prime.

5. Let a, m and n are positive integers with m 6= n. Prove that


1, 2 | a,
m n
(a2 + 1, a2 + 1) =
2, 2 ∤ a.

6. Prove that (n! + 1, (n + 1)! + 1) = 1.

7. Assume m > 1 and m | (m − 1)! + 1, prove that m is a prime.

8. Let m > 4 be not prime, prove that m | (m − 1)!.

9. Assume n 6= 1, prove that (n − 1)2 | nk − 1 ⇐⇒ (n − 1) | k.

10. Prove that for any k ∈ N there exists k consecutive positive integers that are all
composite.

11. Let n > 1 be odd. Prove that n is a prime ⇐⇒ n can not be a sum of some
consecutive positive integers a1 , . . . , ak with k ≥ 3.

12. Assume a, b and c are positive integers satisfying (a, b) = 1 and ab = ck for some
k ∈ N. Prove that a = uk and b = v k for some positive integers u and v. Moreover,
u = (a, c) and v = (b, c).
 Number Theory - Wei LI 3

§2 Divisibility and GCD (II)

1. Assume k ≥ 1 is odd, prove that:

(i) n + 2 does not divide 1k + 2k + · · · + nk for any n ≥ 1.

(ii) 1 + 2 + · · · + n divides 1k + 2k + · · · + nk for any n ≥ 1.

2. For any given n > 1, show that there exits a ≥ 1 such that any number in {aa + 1,
a
aa + 1, . . . } can be divided by n.

3. A = {a1 , . . . , an } are distinct integers where n ≥ 2. The difference between the

product of any n − 1 numbers in A and the remainder one can be divided by n. Prove
that n also divides the quadratic sum of a1 , . . . , an .

4. Assume m, n ≥ 1 and mn | (m2 + n2 ). Prove m = n.

5. Prove n4 + 4n is not prime for any n ≥ 2.

6. Assume a, b, c, d are positive integers satisfying ab = cd. Prove a + b + c + d is not

prime.
ab
7. a, b, c ≥ 1, (a, b, c) = 1 and = c. Prove a − b is a square.
a−b
8. Assume a, m, n are positive integers and a > 1, prove that (am −1, an −1) = a(m,n) −1.

9. a, b ∈ Z satisfies 2a2 + a = 3b2 + b. Prove a − b and 2a + 2b + 1 are both squares.

10. Compute (2002 + 2, 20022 + 2, 20023 + 2, . . . ).

11. Let a and b be distinct positive integers such that ab(a + b) is divisible by a2 + ab + b2 .
√
Prove that |a − b| > 3 ab.

12. Find all primes p and q such that p + q = (p − q)3 .

13. Consider the prime numbers n1 < n2 < · · · < n31 . Prove that if 30 divides n41 + n42 +
· · · + n431 , then among these numbers one can find three consecutive primes.

14. Let n, a, b, c be positive integers with (b, c) = 1. Prove that if c | n! then

c | a(a + b)(a + 2b) · · · (a + (n − 1)b).

 Number Theory - Wei LI 4

§3 Fermat’s method of infinite descent

Let us begin with the following problem.

a2 + b2
Proposition 3.1. Assume a, b ∈ N and = k is a positive integer. Prove that k is
ab + 1
a square.

Proof. We will provide two distinct proofs which are based on a same method essentially.

• First Proof: We will prove the result by Fermat’s method of infinite descent. WLOG,
assume a ≥ b.

2a2
– If a = b, then k = a2 +1
< 2. So k = 1 is a square.
– If a > b, by the division of a remainder, there exists two integers q and r such
that
a = qb − r, 0 ≤ r < b.

It is easy to see q ≥ 2. Then

a2 + b2 q 2 b2 − 2qr + r 2 + b2 −qbr + r 2 + b2 − q
k= = = q + .
ab + 1 qb2 − rb + 1 qb2 − rb + 1
On the one hand, we have

−qbr + r 2 + b2 − q < 2b2 + −rb + 1 < qb2 − rb + 1.

One the other hand, rb < b2 yields rb + 1 ≤ b2 . Combining with r 2 + b2 ≥ 2rb

gives

−qbr + r 2 + b2 −q = (−q)(br + 1) + r 2 + b2 ≥ (−q)b2 + 2rb > −qb2 +rb−1.

Finally, we have
−qbr + r 2 + b2 − q
−1 < < 1.
qb2 − rb + 1
−qbr+r 2 +b2 −q −qbr+r 2 +b2 −q
Since qb2 −rb+1
= k − q is an integer, there must hold qb2 −rb+1
= 0. Hence

b2 + r 2
−qbr + r 2 + b2 − q = 0 =⇒ = q = k.
br + 1
Summary: If a > b, we get some r such that b > r ≥ 0 and

a2 + b2 b2 + r 2
k= = .
ab + 1 br + 1
 Number Theory - Wei LI 5

If r 6= 0, applying this process again to b and r, we can get a third integer r1 such
that r > r1 ≥ 0 and
a2 + b2 b2 + r 2 r 2 + r12
k= = = .
ab + 1 br + 1 rr1 + 1
Finally, by repeating the process we get an infinite sequence satisfying (denote
r0 = r)
a > b > r0 > r1 > · · · > rn > · · · ≥ 0,
and
2
a2 + b2 b2 + r02 r02 + r12 rn−1 + rn2
k= = = = ··· = = ··· .
ab + 1 br0 + 1 r0 r1 + 1 rn−1 rn + 1
This is impossible: any such sequence of integers must sooner or later contain 0.
2
Thus there must be some rn = 0. Then k = rn−1 is a square.
a2 + b2
• Second Proof: Obviously = k is equivalent to
ab + 1
a2 + b2 = k(ab + 1). (3.1)

Claim: There is no positive integer pair (a, b) satisfying equation (3.1) provided k is
not a square.
We will prove the claim by contradiction. Let (a0 , b0 ) be the positive integer pair
satisfy (3.1) such that a0 + b0 attains the minimum of the sums of all positive integer
pairs satisfying (3.1). WLOG, let a0 ≥ b0 . a0 is a root of the following quadratic
equation
x2 − kb0 x + b20 − k = 0.
Let a1 be the other root of the above equation. By Vieta’s formulas, we have
(
a0 + a1 = kb0 ,
a0 a1 = b20 − k.
First, a1 = kb0 − a0 is an integer. Next we will show a1 is positive.

– If a1 = 0, then k = b20 which is a contradiction.

– If a1 < 0, then a21 + b20 = k(a1 b0 + 1) ≤ 0 since k > 0 and a1 b0 + 1 ≤ 0. Thus
a1 = b0 = 0 which is also a contradiction.

Hence a1 is a positive integer.

b20 − k b2 − 1 a2 − 1
a1 = ≤ 0 ≤ 0 < a0 .
a0 a0 a0
Therefore a1 + b0 < a0 + b0 . But (a1 , b0 ) is also a positive integer pair satisfying (3.1)
which yields a contradiction to the minimality of a0 + b0 . This finish the proof.
 Number Theory - Wei LI 6

The argument used in the above proofs has become known as Fermat’s method of
infinite descent: to show that a given Diophantine equation has no positive solutions,
we show that any such solution gives rise to a smaller one, and hence (by iteration) to an
infinite decreasing sequence of positive integer solutions, which is impossible.
Fermat used this technique many times for Diophantine equations. Two typical examples
are showing the non-solvability of the Diophantine equation x4 + y 4 = z 2 and proving
Fermat’s theorem on sums of two squares, which states that any prime p such that p ≡ 1
mod 4 can be expressed as a sum of two squares.
Next we focus on the non-solvability of the Diophantine equation x4 + y 4 = z 2 . Let us
recall some basic conclusion about Pythagorean equation at the first.
Pythagorean equation is defined by

x2 + y 2 = z 2 , (3.2)

where x, y and z are integers.

• (z, 0, ±z) and (0, z, ±z) are trivial solutions to (3.2) for any z ∈ Z.

• For finding non-trivial solutions, we only need to consider the solution of positive
integers, which is called Pythagorean triple.

• For a Pythagorean triple (x, y, z), the triple ( xd , yd , dz ) also solves (3.2) for any common
factor d of x, y, z.

• A Pythagorean triple (x, y, z) is called to be primitive if (x, y, z) = 1.

Theorem 3.2. Let (x, y, z) be a primitive Pythagorean triple, then

• (x, y) = (y, z) = (z, x) = 1.

• 2 ∤ x + y, i.e. x and y have opposite parity.

Theorem 3.3. Let (x, y, z) be a primitive Pythagorean triple with 2 | y, then

x = r 2 − s2 , y = 2rs, z = r 2 + s2 , (3.3)

where r and s are arbitrary integers satisfying

r > s > 0, (r, s) = 1, 2 ∤ r + s. (3.4)

Exercises: Use Fermat’s method of infinite descent to solve the following problems.

1. x4 + y 4 = z 2 has no solution with xyz 6= 0.

 Number Theory - Wei LI 7

Corollary 3.4 (a special case of FLT). x4 + y 4 = z 4 has no solution with xyz 6= 0.

2. Assume m, n ∈ N and mn | m2 + n2 + 1, prove that

m2 + n2 + 1 = 3mn.
 Number Theory - Wei LI 8

§4 Order and its application

Theorem 4.1. Assume m ≥ 2 and (a, m) = 1.

• There exists a positive integer d ≤ m − 1 such that m | (ad − 1), i.e.

ad ≡ 1 mod m. (4.1)

• Let d0 be the smallest positive integer satisfying (4.1). d0 is called the order of a
modulo m. Let h be a positive integer, then

m | (ah − 1) ⇐⇒ d0 | h.

Proof.

• For any fixed j = 1, 2, . . . , m, by the division of a remainder, there exists two integers
qj and rj (0 ≤ rj ≤ m − 1) such that

aj = qj m + rj .

Noting that the m remainders take m − 1 distinct values at most, the pigeonhole
principle indicates that there exists i 6= j such that ri = rj , i.e.

ai − qi m = aj − qj m.

Let i > j ≥ 0, we have

aj (ai−j − 1) = (qi − qj )m.
Hence, m | aj (ai−j − 1). By (aj , m) = 1 (since a and m are coprime), we get
m | (ai−j − 1), i.e.
ai−j ≡ 1 mod m.
Set d = i − j, then 0 < d ≤ m − 1 and ad ≡ 1 mod m.

• By the division of a remainder again, we have h = qd0 + r for some integers q and r
with q ≥ 0, 0 ≤ r ≤ d0 − 1.

ah − 1 = (ah − ar ) + (ar − 1) = ar (akd0 − 1) + (ar − 1).

Obviously (ad0 − 1) | (akd0 − 1) since k ≥ 0. On the other hand, by the definition of

d0 , we have m | (ad0 − 1). Hence m | (akd0 − 1). Therefor

m | (ah − 1) ⇐⇒ m | (ar − 1).

By the minimality of d0 , m | (ar − 1) ⇐⇒ r = 0 ⇐⇒ d0 | h. This finishes the

proof.
 Number Theory - Wei LI 9

Exercises:

1. n > 1 and n | (2n + 1), prove 3 | n.

2. Show that there exists infinitely many positive integers n such that n | (2n + 1).

3. n > 1, prove n ∤ (2n − 1).

4. n > 1 is odd, prove n ∤ (mn−1 + 1) for any m ≥ 1.

p2p + 1
5. p is an odd prime, D is a positive divisor of , prove D ≡ 1 mod 4p.
p2 + 1
n
6. n > 1 and q is a prime divisor of 22 + 1, prove q ≡ 1 mod 2n+2 .

+
7. Following the assumption and notation of Theorem 4.1. Denote by δm (a) = d0 the
order of a modulo m. Moreover, if there exists some positive integer d′ such that
′
m | ad + 1, let δm
−
(a) be the minimal such integer. (Note: δm
−
(a) does NOT exist at
all time, e.g. a = 4, m = 3.) Prove that:

(i) when m = 2, then δ2+ (a) = δ2− (a) = 1;

+
(ii) when m > 2, then δm (a) = 2δm
−
(a);
(iii) when m > 2, then m | ah + 1 ⇐⇒ h = qδm
−
(a) with q is an odd integer;
(iv) if m | ah − 1 or m | ah + 1, we have δm
−
(a) | h.
k−1
8. (i) For any positive integer k, prove that 3k k (23 + 1);
(ii) Let k and s be positive integers, prove that 3k | (2s + 1) if and only if

2 ∤ s and 3k−1 | s.
 Number Theory - Wei LI 10

§5 Fermat-Euler theorem

Theorem 5.1 (Fermat’s Little Theorem). Let a be a positive integer and p be a prime.
Then
ap ≡ a mod p.
Moreover, when (a, p) = 1, we have

ap−1 ≡ 1 mod p.

Definition 5.2 (Euler’s function). For any positive integer m we denote by ϕ(m) the
number of all positive integers less than m that are coprime to m. The function ϕ is
called Euler’s function.

• For small m, we have

m = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, . . .

ϕ(m) = 1, 1, 2, 2, 4, 2, 6, 4, 6, 4, 10, 4, . . .

• ϕ(m) = m − 1 ⇐⇒ m is a prime.

• 2 | ϕ(m) for any m ≥ 3.

Definition 5.3 (Reduced set of residue classes). A set R of integers is called a reduced
set of residue classes modulo m if

• (r, m) = 1 for any r ∈ R;

• For any a ∈ Z satisfying (a, m) = 1, there exists a unique element r ∈ R such that

a≡r mod m.

Proposition 5.4.

1. {r : (r, m) = 1, 1 ≤ r ≤ m}, which consists of ϕ(m) elements, is a reduced set of

residue classes modulo m.

2. If R is a reduced set of residue classes modulo m, then

♯R = ϕ(m).

3. Let r1 , . . . , rϕ(m) be ϕ(m) integers satisfying (ri , m) = 1 for each i = 1, . . . , ϕ(m) and

ri 6≡ rj mod m, ∀i 6= j,

then {r1 , . . . , rϕ(m) } is a reduced set of residue classes modulo m.

 Number Theory - Wei LI 11

4. Let m ≥ 1 and (a, m) = 1. Assume that R is a reduced set of residue classes modulo
m. Then
aR , {ar : r ∈ R}
is also a reduced set of residue classes modulo m.

Proof. TBA.

Theorem 5.5 (Euler’s Theorem). Let a and m be coprime positive integers. Then

aϕ(m) ≡ 1 mod m.

Proof. TBA.

Remark 5.6. Fermat’s Little Theorem is a special case of Euler’s Theorem when m is a
prime.

Exercises:

1. Find n such that 2n k 31024 − 1.

2. Compute (2002 + 2, 20022 + 2, 20023 + 2, . . . ).

3. Let a and b be distinct positive integers such that ab(a + b) is divisible by a2 + ab + b2 .

√
Prove that |a − b| > 3 ab.

4. Find all primes p and q such that p + q = (p − q)3 .

5. Consider the prime numbers n1 < n2 < · · · < n31 . Prove that if 30 divides n41 + n42 +
· · · + n431 , then among these numbers one can find three consecutive primes.

6. Let p be a prime. Prove that p | (abp − bap ) for all integers a and b.

7. Let p ≥ 7 be a prime. Prove that the number |11 {z

. . . 1} is divisible by p.
p−1

8. Let p > 5 be a prime. Prove that p8 ≡ 1 mod 240.

9. Prove that (n2 − 1) | (2n! − 1) for any even positive integer n.

10. Let n, a, b, c be positive integers with (b, c) = 1. Prove that if c | n! then

c | a(a + b)(a + 2b) · · · (a + (n − 1)b).

11. Find the last two decimal digits of 3406 .

12. Let m > n ≥ 1, find the minimum value of m + n such that 1000 | 1978m − 1978n .
 Number Theory - Wei LI 12

§6 Further discussion on Euler’s function

TBA
 Number Theory - Wei LI 13

§7 Congruences and Chinese remainder theorem

TBA