PrevNext

Overview

The Lifting the Exponent Lemma, often abbreviated as LTE, is a powerful tool in olympiad number theory.

It helps determine how many times a prime divides expressions such as

an−bna^n-b^n

or

an+bn.a^n+b^n.

In other words, LTE is useful when we want to compute the exponent of a prime inside a number.

For example, instead of just asking whether 33 divides 106−110^6-1, we can ask:

What is the largest power of 33 that divides 106−110^6-1?

LTE gives a fast and systematic way to answer questions like this.

Motivation

Consider

106−1.10^6-1.

We can factor it as

106−1=(103−1)(103+1),10^6-1=(10^3-1)(10^3+1),

or even further using difference of powers. However if we only care about powers of 33, expanding or fully factoring the expression is inefficient.

Instead notice that

10≡1(mod3),10 \equiv 1 \pmod{3},

so 33 divides 10−110-1.

LTE tells us that under suitable conditions,

v3(106−1)=v3(10−1)+v3(6).v_3(10^6-1)=v_3(10-1)+v_3(6).

Since

v3(9)=2v_3(9)=2

and

v3(6)=1,v_3(6)=1,

we get

v3(106−1)=3.v_3(10^6-1)=3.

So 333^3 divides 106−110^6-1, but 343^4 does not.

This kind of computation appears frequently in olympiad divisibility problems.

Prime Valuation

Before stating LTE, we need one piece of notation.

For a prime pp and a positive integer NN, define

vp(N)v_p(N)

to be the exponent of pp in the prime factorisation of NN.

In other words, vp(N)v_p(N) is the largest integer kk such that

pk∣N.p^k \mid N.

Examples

Since

72=23×32,72=2^3 \times 3^2,

we have

v2(72)=3v_2(72)=3

and

v3(72)=2.v_3(72)=2.

Also,

v5(72)=0,v_5(72)=0,

because 55 does not divide 7272.

This notation allows us to discuss divisibility by prime powers far more simply.

Main Statement: Difference Form

The most common version of LTE is the difference form.

Lifting the Exponent Lemma

Let pp be an odd prime. Suppose that p∣a−bp\mid a-b and p∤abp\nmid ab. Then for every positive integer nn,

vp(an−bn)=vp(a−b)+vp(n).v_p(a^n-b^n) = v_p(a-b)+v_p(n).

This is the version used most often in olympiad problems.

Why the Conditions Matter

The assumptions are important.

  • We require pp to be an odd prime.
  • We require p∣a−bp\mid a-b.
  • We require p∤abp\nmid ab.

The condition p∣a−bp\mid a-b means that

a≡b(modp).a\equiv b \pmod p.

The condition p∤abp\nmid ab means that neither aa nor bb is divisible by pp. This prevents degenerate cases where both powers contain extra factors of pp for unrelated reasons.

Intuition Behind LTE

The expression

an−bna^n-b^n

always factors as

(a−b)(an−1+an−2b+⋯+abn−2+bn−1).(a-b)(a^{n-1}+a^{n-2}b+\cdots+ab^{n-2}+b^{n-1}).

So a−ba-b is always a factor.

LTE says something more significant. If an odd prime pp divides a−ba-b, then the additional powers of pp that appear in an−bna^n-b^n come from the exponent nn.

That is why the formula is

vp(an−bn)=vp(a−b)+vp(n).v_p(a^n-b^n) = v_p(a-b)+v_p(n).

The first term comes from the base difference a−ba-b, and the second term comes from the exponent nn.

Proof Sketch

We can outline the idea behind the proof for odd primes.

Write

a=b+h,a=b+h,

where p∣hp\mid h.

Then

an−bn=(b+h)n−bn.a^n-b^n=(b+h)^n-b^n.

Using the binomial theorem,

(b+h)n−bn=nbn−1h+(n2)bn−2h2+⋯+hn.(b+h)^n-b^n = nb^{n-1}h+\binom{n}{2}b^{n-2}h^2+\cdots+h^n.

The first term is

nbn−1h.nb^{n-1}h.

Since p∤bp\nmid b, the pp-adic contribution of this term is essentially

vp(n)+vp(h).v_p(n)+v_p(h).

The later terms contain higher powers of hh, so they are more divisible by pp. Under the hypotheses of LTE, these later terms do not change the final valuation.

Since

h=a−b,h=a-b,

this gives

vp(an−bn)=vp(a−b)+vp(n).v_p(a^n-b^n)=v_p(a-b)+v_p(n).

This proof sketch explains why the formula has the shape it does. A fully rigorous proof requires checking that the later binomial terms have large enough pp-adic valuation.

Worked Examples

Example 1

Find the largest power of 33 dividing 106−110^6-1.

We use LTE with

p=3,a=10,b=1,n=6.p=3,\qquad a=10,\qquad b=1,\qquad n=6.

Since

3∣10−13\mid 10-1

and

3∤10,3\nmid 10,

the conditions are satisfied.

Therefore,

v3(106−1)=v3(10−1)+v3(6).v_3(10^6-1)=v_3(10-1)+v_3(6).

Now

v3(10−1)=v3(9)=2v_3(10-1)=v_3(9)=2

and

v3(6)=1.v_3(6)=1.

Thus

v3(106−1)=3.v_3(10^6-1)=3.

So the largest power of 33 dividing 106−110^6-1 is 333^3.

Part 2: Check the part 2 for continuation

Module Progress:

Join the Discord Community!

Stuck on a problem, or don't understand a module? Join the Discord and get help with your doubts while making more math friends.

PrevNext