English (unofficial) translations of posts at kexue.fm
Source

[Euler Mathematics] Two Proofs of the Infinitude of Primes

Translated by DeepSeek V4 Pro. Translations can be inaccurate, please refer to the original post for important stuff.

Prime numbers are the basic units of numbers, much like bricks in a skyscraper. Clearly, the fact that there are infinitely many primes is a prerequisite for the value of number theory research. Otherwise, the study of numbers would be limited to a finite set of primes, and many numbers would lose their charm. It would be like having only a finite number of bricks, which would make it impossible to create architectural wonders. Below, I will introduce two classic proofs regarding the infinitude of primes. One is Euclid’s proof, which is the most original and simplest method; I believe many readers have already learned it, but it is still worth mentioning here. The other is one I saw in How to Solve It, originally by Euler, which is also a very beautiful proof. Of course, this article emphasizes the ideas, and the argumentation process may have some informalities; readers are welcome to refine them. ^_^

Euclid’s Proof

The idea of this proof is very simple: the product of several primes plus 1 produces a new prime factor. If there were only n primes, what would happen if we multiplied them all together and added 1? If the result is a prime number, it contradicts the assumption that there are only n primes. If it is a composite number, it is not divisible by any of the original n primes (the remainder is always 1), so it must possess a new prime factor, which again contradicts the assumption of only n primes. In either case, assuming the number of primes is finite leads to a contradiction. Therefore, the number of primes must be infinite.

Euler’s Classic Proof

This proof requires some preparation. Its core idea is: Geometric Series + Generating Functions.

First, we have the formula S(p)=1+p^{-1}+p^{-2}+\dots+p^{-n}=\frac{1-p^{-n-1}}{1-p^{-1}}. This is the sum formula for a geometric series. When |p| > 1 and n \to \infty, then p^{-n-1} \to 0, and we have: S(p)=\sum_{n=0}^{\infty} p^{-n}=\frac{1}{1-p^{-1}}=\frac{p}{p-1}

Next, we attempt to let p range over all prime numbers, i.e., 2, 3, 5, 7, \dots, and multiply each S(p) together to get (denoted as K): \begin{aligned}K &= S(2) \cdot S(3) \cdot S(5) \dots \\ &= \left(1+\frac{1}{2}+\frac{1}{2^2}+\frac{1}{2^3} \dots\right)\left(1+\frac{1}{3}+\frac{1}{3^2}+\frac{1}{3^3} \dots\right)\left(1+\frac{1}{5}+\frac{1}{5^2}+\frac{1}{5^3} \dots\right) \dots \\ &= \frac{2}{2-1} \cdot \frac{3}{3-1} \cdot \frac{5}{5-1} \cdot \dots\end{aligned}

What is special about K? Notice that here the powers of each prime are multiplied with each other exactly once. This is identical to how natural numbers are generated: By multiplying the powers of several primes, one can obtain any natural number. Thus, we can write (not without foundation): K=1+\frac{1}{2}+\frac{1}{3}+\frac{1}{4}+\frac{1}{5}+\dots

As we know, the series 1+\frac{1}{2}+\frac{1}{3}+\frac{1}{4}+\frac{1}{5}+\dots (the harmonic series) diverges. Therefore, it is impossible for there to be only a finite number of primes. If there were only finitely many primes, then the product K = \frac{2}{2-1} \cdot \frac{3}{3-1} \cdot \frac{5}{5-1} \cdot \dots would necessarily be finite, leading to a contradiction. Thus, primes are infinite.

A Connection to the Prime Number Theorem

The Prime Number Theorem tells us that the number of primes not exceeding n, denoted by \pi(n), is approximately equal to \frac{n}{\ln n}.

From the discussion above, we obtained: K = \frac{2}{2-1} \cdot \frac{3}{3-1} \cdot \frac{5}{5-1} \cdot \dots = 1+\frac{1}{2}+\frac{1}{3}+\frac{1}{4}+\frac{1}{5}+\dots

Using calculus, it can be proven that the part on the right side of the second equals sign, 1+\frac{1}{2}+\frac{1}{3}+\frac{1}{4}+\frac{1}{5}+\dots+\frac{1}{n}, can be approximated as \ln n. By inverting the expression, we get: \frac{2-1}{2} \cdot \frac{3-1}{3} \cdot \frac{5-1}{5} \cdot \dots \approx \frac{1}{\ln n}

For a large n, the number of even numbers not exceeding n is clearly about n/2. Among the remaining n/2 odd numbers, the multiples of 3 are about 1/3, so the non-multiples of 3 are about n \cdot 1/2 \cdot 2/3. Among the remaining numbers, the non-multiples of 5 are about n \cdot 1/2 \cdot 2/3 \cdot 4/5. Following this logic, we get: \pi(n) \approx n \cdot 1/2 \cdot 2/3 \cdot 4/5 \cdot 6/7 \cdot \dots \approx \frac{n}{\ln n}

Of course, this is not a rigorous proof at all; at most, it is a strained but luckily successful piece of reasoning. However, even though it cannot withstand the test of strict logic, it is quite interesting as an elementary piece of thinking to appreciate. ^_^

When reposting, please include the original address: https://kexue.fm/archives/1484

For more detailed information on reposting, please refer to: Scientific Space FAQ