Pan Chengdong "Number Theory Foundation" Selected Solutions: Chapter 3
This article is translated by AI.
1
Prove ∏pp2−1p2=6π2.
Proof: Consider the Euler product formula of the Riemann zeta function ζ(s):
ζ(s)=n=1∑∞ns1=p∏(1−ps1)−1
where the product traverses all prime numbers p, and this formula holds for Re(s)>1.
Let s=2, we know ζ(2)=∑n=1∞n21=6π2. Substituting s=2 into the Euler product formula, we get:
ζ(2)=p∏(1−p21)−1=6π2
Now consider the infinite product given in the problem:
p∏p2−1p2=p∏p2p2−11=p∏1−p211=p∏(1−p−2)−1
By comparison, this product is exactly the Euler product representation of ζ(2). Therefore,
p∏p2−1p2=ζ(2)=6π2
Q.E.D.
2
Prove that the series ∑pp1 diverges.
Proof: Use proof by contradiction. Assume that the series ∑pp1 converges. According to the Cauchy criterion, this means that for any ϵ>0, there exists N such that for all n>m≥N, ∑k=m+1npk1<ϵ. In particular, this means that the tail series ∑k=K+1∞pk1 converges for any K (as the limit of n→∞).
Let K be any positive integer. Consider the set of positive integers MK={m∈N∣∀p s.t. p∣m,p>pK} whose prime factors are all greater than pK. Since ∑k=K+1∞pk1 converges (by assumption), according to the relationship between infinite products and series, the infinite product ∏k=K+1∞(1−pk1) converges to a positive value PK′>0.
Therefore, the Euler product ∑m∈MKm1=∏k=K+1∞(1−pk1)−1=PK′1 converges to a finite value VK.
Now, let PK=p1p2⋯pK be the product of the first K prime numbers. Consider integers of the form 1+qPK, where q=1,2,3,….
Any prime factor p of 1+qPK must satisfy p∤PK, otherwise p∣qPK and p∣(1+qPK), which means p∣1, which is impossible.
Therefore, all prime factors of 1+qPK are greater than pK, that is, 1+qPK∈MK holds for all q≥1.
Thus, we have the series inequality:
q=1∑∞1+qPK1≤m∈MK∑m1=VK
This shows that if ∑pp1 converges, then the series ∑q=1∞1+qPK1 must converge.
However, we use the limit comparison test to compare the series ∑q=1∞1+qPK1 with the divergent harmonic series ∑q=1∞q1:
q→∞limq11+qPK1=q→∞lim1+qPKq=PK1
Since PK=p1⋯pK≥2, the limit value PK1 is a positive finite constant.
Because the harmonic series ∑q=1∞q1 diverges, according to the limit comparison test, the series ∑q=1∞1+qPK1 must also diverge.
This contradicts the conclusion "∑q=1∞1+qPK1 converges" derived from the assumption "∑pp1 converges".
Therefore, the initial assumption "∑pp1 converges" must be wrong. This means that the series ∑pp1 does not satisfy the Cauchy criterion, so the series diverges. Q.E.D.
3
Prove that the sequence {6n−1} contains infinitely many prime numbers.
Proof: Use proof by contradiction. Assume that there are only finitely many prime numbers of the form 6n−1, let them be p1,p2,…,pr. Consider the integer N=6(p1p2⋯pr)−1.
First, N>1. In the prime factorization of N, all prime factors p must satisfy p∤6, that is, p cannot be 2 or 3. Therefore, any prime factor p of N must be of the form 6k+1 or 6k−1.
Note that N=6(p1p2⋯pr)−1≡−1(mod6).
If all prime factors of N are of the form 6k+1, then their product N must also be of the form 6k+1. (Because (6k1+1)(6k2+1)=36k1k2+6k1+6k2+1=6(6k1k2+k1+k2)+1≡1(mod6)) This contradicts N≡−1(mod6).
Therefore, N must have at least one prime factor of the form 6k−1, let it be p.
We prove that p is not equal to any of p1,p2,…,pr. If p=pi holds for some i∈{1,2,…,r}, then pi∣N and pi∣6(p1p2⋯pr). Therefore pi must divide their difference, that is, pi∣(6(p1p2⋯pr)−N), which is pi∣1. This is impossible.
So, p is a prime number of the form 6k−1, but it is not in our assumed finite list p1,p2,…,pr. This contradicts our initial assumption (all prime numbers of the form 6n−1 are in this list).
Therefore, the assumption is wrong, and there are infinitely many prime numbers of the form 6n−1. Q.E.D.
5
Using ∏p⩽x(1−p1)−1⩽∏K=2π(x)+1(1−K1)−1, prove: (1) π(x)>logx−1; (2) pn<3n+1 (pn is the n-th prime number).
Proof:
(1) Prove π(x)>logx−1
We know that for x≥1, ∑n≤xn1>logx. At the same time, we have
n≤x∑n1≤n∈Sx∑n1=p≤x∏(1−p1)−1
where Sx is the set of positive integers whose prime factors are all ≤x.
Combining the above inequality and the inequality given in the problem, we get:
logx<n≤x∑n1≤p≤x∏(1−p1)−1≤K=2∏π(x)+1(1−K1)−1
Calculate the product on the right:
K=2∏π(x)+1(1−K1)−1=K=2∏π(x)+1(KK−1)−1=K=2∏π(x)+1K−1K=12⋅23⋅34⋯π(x)π(x)+1=π(x)+1
Therefore, we get logx<π(x)+1. Rearranging gives π(x)>logx−1.
(2) Prove pn<3n+1
From (1) we know π(x)>logx−1. Let x=pn, where pn is the n-th prime number. Then π(x)=π(pn)=n. Substituting into the inequality, we get:
n>logpn−1
Rearranging gives:
logpn<n+1
Taking the exponent on both sides (with natural logarithm base e):
pn<en+1
Since e≈2.718<3, we have en+1<3n+1. Therefore,
pn<3n+1
Q.E.D.
References
https://zhuanlan.zhihu.com/p/543010816
https://zhuanlan.zhihu.com/p/544792225
https://zhuanlan.zhihu.com/p/545595129
https://zhuanlan.zhihu.com/p/546385318
https://zhuanlan.zhihu.com/p/547333984
https://zhuanlan.zhihu.com/p/548061254
