下面将写出欧拉对上式的证明中缺失的严格论证的部分,即对连乘积公式的证明部分,而不涉及最终的系数比较
首先考虑当n为奇数时,将
z
n
−
a
n
{\displaystyle z^{n}-a^{n}}
分解为连乘积形式。
事实上,容易发现上式的全部复根为
a
,
a
e
2
π
i
1
n
,
a
e
2
π
i
2
n
,
.
.
.
,
a
e
2
π
i
n
−
1
n
{\displaystyle a,ae^{2\pi i{\frac {1}{n}}},ae^{2\pi i{\frac {2}{n}}},...,ae^{2\pi i{\frac {n-1}{n}}}}
由于n为奇数,所以可以将除了z=a外的其他根及其共轭一一配对,即 将共轭的根一一配对
a
e
2
π
i
k
n
,
a
e
2
π
i
n
−
k
n
=
a
e
−
2
π
i
k
n
{\displaystyle ae^{2\pi i{\frac {k}{n}}},ae^{2\pi i{\frac {n-k}{n}}}=ae^{-2\pi i{\frac {k}{n}}}}
看做一对, 则通过二次方程的韦达定理 可以还原出每对根的最小多项式:
按照韦达定理,有
x
1
+
x
2
=
−
a
1
a
0
=
a
e
2
π
i
k
n
+
a
e
−
2
π
i
k
n
=
cos
(
2
π
k
n
)
+
cos
(
−
2
π
k
n
)
=
2
cos
(
2
π
k
n
)
{\displaystyle x_{1}+x_{2}=-{\frac {a_{1}}{a_{0}}}=ae^{2\pi i{\frac {k}{n}}}+ae^{-2\pi i{\frac {k}{n}}}=\cos \left(2\pi {\frac {k}{n}}\right)+\cos \left(-2\pi {\frac {k}{n}}\right)=2\cos \left({\frac {2\pi k}{n}}\right)}
x
1
x
2
=
a
2
a
0
=
a
e
2
π
i
k
n
a
e
−
2
π
i
k
n
=
a
2
{\displaystyle x_{1}x_{2}={\frac {a_{2}}{a_{0}}}=ae^{2\pi i{\frac {k}{n}}}ae^{-2\pi i{\frac {k}{n}}}=a^{2}}
由于最小多项式首项系数为1,故
a
0
=
1
{\displaystyle a_{0}=1}
,由此得到这对根最小多项式为
a
0
z
2
+
a
1
z
+
a
2
=
z
2
−
2
cos
(
2
π
k
n
)
z
+
a
2
{\displaystyle a_{0}z^{2}+a_{1}z+a_{2}=z^{2}-2\cos \left({\tfrac {2\pi k}{n}}\right)z+a^{2}}
注意到k的取值上限为
n
−
1
2
{\displaystyle {\tfrac {n-1}{2}}}
,将每一对根的最小多项式相乘,
还有z=a这个根的最小多项式
z
−
a
{\displaystyle z-a}
,乘在一起,得到
z
n
−
a
n
=
(
z
−
a
)
∏
k
=
1
n
−
1
2
(
z
2
−
2
a
z
cos
2
k
π
n
+
a
2
)
{\displaystyle z^{n}-a^{n}=(z-a)\prod _{k=1}^{\frac {n-1}{2}}\left(z^{2}-2az\cos {\frac {2k\pi }{n}}+a^{2}\right)}
令
z
=
1
+
x
N
,
a
=
1
−
x
N
,
N
=
n
{\displaystyle z=1+{\frac {x}{N}},a=1-{\frac {x}{N}},N=n}
,代入上式,有:
(
1
+
x
N
)
N
−
(
1
−
x
N
)
N
=
[
(
1
+
x
N
)
−
(
1
−
x
N
)
]
∏
k
=
1
N
−
1
2
[
(
1
+
x
N
)
2
−
2
(
1
+
x
N
)
(
1
−
x
N
)
cos
(
2
π
k
N
)
+
(
1
−
x
N
)
2
]
=
2
x
N
∏
k
=
1
N
−
1
2
[
2
+
2
x
2
N
2
−
2
(
1
−
x
2
N
2
)
cos
(
2
π
k
N
)
]
=
2
x
N
∏
k
=
1
N
−
1
2
[
2
+
2
x
2
N
2
−
2
cos
(
2
π
k
N
)
+
2
x
2
N
2
cos
(
2
π
k
N
)
]
=
4
x
N
∏
k
=
1
N
−
1
2
(
(
1
−
cos
(
2
π
k
N
)
)
+
(
1
+
cos
(
2
π
k
N
)
)
x
2
N
2
)
=
4
x
N
∏
k
=
1
N
−
1
2
{
[
1
−
cos
(
2
π
k
N
)
]
[
1
+
1
+
cos
(
2
π
k
N
)
1
−
cos
(
2
π
k
N
)
x
2
N
2
]
}
{\displaystyle {\begin{aligned}\left(1+{\frac {x}{N}}\right)^{N}-\left(1-{\frac {x}{N}}\right)^{N}&=\left[\left(1+{\frac {x}{N}}\right)-\left(1-{\frac {x}{N}}\right)\right]\prod _{k=1}^{\frac {N-1}{2}}\left[\left(1+{\frac {x}{N}}\right)^{2}-2\left(1+{\frac {x}{N}}\right)\left(1-{\frac {x}{N}}\right)\cos \left({\frac {2\pi k}{N}}\right)+\left(1-{\frac {x}{N}}\right)^{2}\right]\\&={\frac {2x}{N}}\prod _{k=1}^{\frac {N-1}{2}}\left[2+{\frac {2x^{2}}{N^{2}}}-2\left(1-{\frac {x^{2}}{N^{2}}}\right)\cos \left({\frac {2\pi k}{N}}\right)\right]\\&={\frac {2x}{N}}\prod _{k=1}^{\frac {N-1}{2}}\left[{2+{\frac {2{x^{2}}}{N^{2}}}-2\cos \left({\frac {2\pi k}{N}}\right)+{\frac {2{x^{2}}}{N^{2}}}\cos({\frac {2\pi k}{N}})}\right]\\&={\frac {4x}{N}}\prod _{k=1}^{\frac {N-1}{2}}\left({(1-\cos({\frac {2\pi k}{N}}))+(1+\cos({\frac {2\pi k}{N}})){\frac {x^{2}}{N^{2}}}}\right)\\&={\frac {4x}{N}}\prod _{k=1}^{\frac {N-1}{2}}\left\{\left[1-\cos \left({\frac {2\pi k}{N}}\right)\right]\left[{1+{\frac {1+\cos \left({\frac {2\pi k}{N}}\right)}{1-\cos({\frac {2\pi k}{N}})}}{\frac {x^{2}}{N^{2}}}}\right]\right\}\\\end{aligned}}}
此时,上述乘积中的
4
N
∏
k
=
1
N
−
1
2
(
1
−
cos
(
2
π
k
N
)
)
{\displaystyle {\frac {4}{N}}\prod _{k=1}^{\frac {N-1}{2}}(1-\cos({\frac {2\pi k}{N}}))}
仅和N有关,记作
C
(
N
)
{\displaystyle C(N)}
,上式变为
(
1
+
x
N
)
N
−
(
1
−
x
N
)
N
=
C
(
N
)
x
∏
k
=
1
N
−
1
2
(
1
+
1
+
cos
(
2
π
k
N
)
1
−
cos
(
2
π
k
N
)
x
2
N
2
)
{\displaystyle \left(1+{\frac {x}{N}}\right)^{N}-\left(1-{\frac {x}{N}}\right)^{N}={C(N)}x\prod _{k=1}^{\frac {N-1}{2}}\left({1+{\frac {1+\cos({\frac {2\pi k}{N}})}{1-\cos({\frac {2\pi k}{N}})}}{\frac {x^{2}}{N^{2}}}}\right)}
而利用二项式定理,将等式左边展开:
(
1
+
x
N
)
N
=
∑
k
=
0
N
C
N
k
x
k
N
k
{\displaystyle {(1+{\frac {x}{N}})^{N}}=\sum _{k=0}^{N}{C_{N}^{k}{\frac {x^{k}}{N^{k}}}}}
(
1
−
x
N
)
N
=
∑
k
=
0
N
(
−
1
)
k
C
N
k
x
k
N
k
{\displaystyle {(1-{\frac {x}{N}})^{N}}=\sum _{k=0}^{N}{{{(-1)}^{k}}C_{N}^{k}{\frac {x^{k}}{N^{k}}}}}
两式相减,考虑一次项,为
C
N
1
x
N
−
(
−
1
)
C
N
1
x
N
=
2
C
N
1
x
N
=
2
x
{\displaystyle C_{N}^{1}{\frac {x}{N}}-(-1)C_{N}^{1}{\frac {x}{N}}=2C_{N}^{1}{\frac {x}{N}}=2x}
这正是等式的左边的一次项
而等式右边的一次项只能是连乘积中的全部1与连乘积外的C(n)x相乘,为使两边相等,必须有
C
(
N
)
=
2
{\displaystyle C(N)=2}
,于是上式变为
(
1
+
x
N
)
N
−
(
1
−
x
N
)
N
=
2
x
∏
k
=
1
N
−
1
2
(
1
+
1
+
cos
(
2
π
k
N
)
1
−
cos
(
2
π
k
N
)
x
2
N
2
)
{\displaystyle \left(1+{\frac {x}{N}}\right)^{N}-\left(1-{\frac {x}{N}}\right)^{N}=2x\prod _{k=1}^{\frac {N-1}{2}}\left({1+{\frac {1+\cos({\frac {2\pi k}{N}})}{1-\cos({\frac {2\pi k}{N}})}}{\frac {x^{2}}{N^{2}}}}\right)}
另一方面,令
θ
=
2
π
k
N
{\displaystyle \theta ={\frac {2\pi k}{N}}}
,有
cos
(
θ
)
=
1
−
θ
2
2
+
O
(
θ
3
)
{\displaystyle \cos(\theta )=1-{\frac {\theta ^{2}}{2}}+\mathrm {O} (\theta ^{3})}
于是,代入上式,得到
(
1
+
x
N
)
N
−
(
1
−
x
N
)
N
=
2
x
∏
k
=
1
N
−
1
2
[
1
+
1
+
cos
(
2
π
k
N
)
1
−
cos
(
2
π
k
N
)
x
2
N
2
]
=
2
x
∏
k
=
1
N
−
1
2
{
1
+
1
+
[
1
−
θ
2
2
+
O
(
θ
3
)
]
1
−
[
1
−
θ
2
2
+
O
(
θ
3
)
]
x
2
N
2
}
=
2
x
∏
k
=
1
N
−
1
2
[
1
+
2
−
θ
2
2
+
O
(
θ
3
)
θ
2
2
+
O
(
θ
3
)
x
2
N
2
]
=
2
x
∏
k
=
1
N
−
1
2
(
1
+
(
4
−
θ
2
+
O
(
θ
3
)
)
x
2
(
θ
2
+
O
(
θ
3
)
)
N
2
)
=
2
x
∏
k
=
1
N
−
1
2
(
1
+
(
4
−
(
2
k
π
N
)
2
+
O
(
(
2
k
π
N
)
3
)
)
x
2
(
(
2
k
π
N
)
2
+
O
(
(
2
k
π
N
)
3
)
)
N
2
)
=
2
x
∏
k
=
1
N
−
1
2
(
1
+
(
4
−
(
2
k
π
N
)
2
+
O
(
(
2
k
π
N
)
3
)
)
x
2
(
2
k
π
)
2
+
O
(
(
2
k
π
)
3
N
)
)
{\displaystyle {\begin{aligned}\left(1+{\frac {x}{N}}\right)^{N}-\left(1-{\frac {x}{N}}\right)^{N}&=2x\prod _{k=1}^{\frac {N-1}{2}}\left[1+{\frac {1+\cos \left({\frac {2\pi k}{N}}\right)}{1-\cos \left({\frac {2\pi k}{N}}\right)}}{\frac {x^{2}}{N^{2}}}\right]\\&=2x\prod _{k=1}^{\frac {N-1}{2}}\left\{1+{\frac {1+\left[1-{\frac {\theta ^{2}}{2}}+\mathrm {O} \left(\theta ^{3}\right)\right]}{1-\left[1-{\frac {\theta ^{2}}{2}}+\mathrm {O} \left(\theta ^{3}\right)\right]}}{\frac {x^{2}}{N^{2}}}\right\}\\&=2x\prod _{k=1}^{\frac {N-1}{2}}\left[1+{\frac {2-{\frac {\theta ^{2}}{2}}+\mathrm {O} \left(\theta ^{3}\right)}{{\frac {\theta ^{2}}{2}}+\mathrm {O} \left(\theta ^{3}\right)}}{\frac {x^{2}}{N^{2}}}\right]\\&=2x\prod _{k=1}^{\frac {N-1}{2}}\left(1+{\frac {\left(4-\theta ^{2}+\mathrm {O} \left(\theta ^{3}\right)\right)x^{2}}{\left(\theta ^{2}+\mathrm {O} \left(\theta ^{3}\right)\right)N^{2}}}\right)\\&=2x\prod _{k=1}^{\frac {N-1}{2}}\left(1+{\frac {\left(4-\left({\frac {2k\pi }{N}}\right)^{2}+\mathrm {O} \left(\left({\frac {2k\pi }{N}}\right)^{3}\right)\right)x^{2}}{\left(\left({\frac {2k\pi }{N}}\right)^{2}+\mathrm {O} \left(\left({\frac {2k\pi }{N}}\right)^{3}\right)\right)N^{2}}}\right)\\&=2x\prod _{k=1}^{\frac {N-1}{2}}\left(1+{\frac {\left(4-\left({\frac {2k\pi }{N}}\right)^{2}+\mathrm {O} \left(\left({\frac {2k\pi }{N}}\right)^{3}\right)\right)x^{2}}{(2k\pi )^{2}+\mathrm {O} \left({\frac {(2k\pi )^{3}}{N}}\right)}}\right)\end{aligned}}}
令N→∞,则右端大O符号的诸项都变为无穷小。另一方面,左端可写为:
lim
N
→
∞
(
1
+
x
N
)
N
−
(
1
−
x
N
)
N
=
e
x
−
e
−
x
{\displaystyle \lim _{N\to \infty }(1+{\frac {x}{N}})^{N}-(1-{\frac {x}{N}})^{N}=e^{x}-e^{-x}}
于是上式变为
e
x
−
e
−
x
=
2
x
∏
k
=
1
∞
(
1
+
(
4
+
o
(
1
)
)
x
2
(
2
k
π
)
2
+
o
(
1
)
)
=
2
x
∏
k
=
1
∞
(
1
+
(
1
+
o
(
1
)
)
x
2
k
2
π
2
+
o
(
1
)
)
=
2
x
∏
k
=
1
∞
(
1
+
x
2
k
2
π
2
)
{\displaystyle {\begin{aligned}e^{x}-e^{-x}&=2x\prod _{k=1}^{\infty }\left({1+{\frac {(4+o(1)){x^{2}}}{{{(2k\pi )}^{2}}+o(1)}}}\right)\ \\&=2x\prod _{k=1}^{\infty }\left({1+{\frac {(1+o(1)){x^{2}}}{{k^{2}}{\pi ^{2}}+o(1)}}}\right)\ \\&=2x\prod _{k=1}^{\infty }\left({1+{\frac {x^{2}}{{k^{2}}{\pi ^{2}}}}}\right)\ \\\end{aligned}}}
此时,只需比较左右两端展开式的三次项系数,即可得出结果。
对左式进行级数展开,可得:
e
x
−
e
−
x
=
−
∑
k
=
0
∞
(
−
x
)
k
−
x
k
k
!
{\displaystyle e^{x}-e^{-x}=-\sum _{k=0}^{\infty }{\frac {(-x)^{k}-x^{k}}{k!}}}
其中当
k
=
3
{\displaystyle k=3}
时可提取左式的三次项为
2
x
3
3
!
{\displaystyle {\frac {2x^{3}}{3!}}}
。
同时展开右式可得右式的三次项为
(
∑
k
=
1
∞
2
k
2
π
2
)
x
3
{\displaystyle \left(\sum _{k=1}^{\infty }{\frac {2}{k^{2}\pi ^{2}}}\right)x^{3}}
由于等式左右端相等,所以左右式三次项系数必须相等, 因此可得:
2
3
!
=
∑
k
=
1
∞
2
k
2
π
2
{\displaystyle {\frac {2}{3!}}=\sum _{k=1}^{\infty }{\frac {2}{k^{2}\pi ^{2}}}}
化简可得:
π
2
6
=
∑
k
=
1
∞
1
k
2
=
ζ
(
2
)
{\displaystyle {\frac {\pi ^{2}}{6}}=\sum _{k=1}^{\infty }{\frac {1}{k^{2}}}=\zeta \left(2\right)}