盖尔曼–劳定理 (英語:Gell-Mann and Low theorem )是量子场论 中的重要定理,它说明了有相互作用的多体系统的基态(真空态)与相应的无相互作用多体系统之间的关系。1951年由默里·盖尔曼 和弗朗西斯·劳 证明。该定理的重要意义在于,它将有相互作用系统的格林函数 和无相互作用系统的格林函数联系起来[ 1] 。尽管一般用于基态,盖尔曼–劳定理实际上可以应用于体系哈密顿量的任一个本征态。其原始证明[ 2] 用到微扰理论,它将多体系统中的相互作用视为微扰,并通过无限慢的过程(绝热过程)引入该微扰,从而将有相互作用的多体系统与对应的无相互作用的系统联系起来。
原始的论文是通过演化算符的戴森展开式来完成证明的,而Molinari则将其有效性推广到微扰论成立的范围之外。下面介绍Molinari的方法[ 3] 。在
H
ϵ
{\displaystyle H_{\epsilon }}
中令
g
=
e
ϵ
θ
{\displaystyle g=e^{\epsilon \theta }}
,由时间演化算符满足的薛定谔方程
i
ℏ
∂
t
1
U
ϵ
(
t
1
,
t
2
)
=
H
ϵ
(
t
1
)
U
ϵ
(
t
1
,
t
2
)
{\displaystyle i\hbar \partial _{t_{1}}U_{\epsilon }(t_{1},t_{2})=H_{\epsilon }(t_{1})U_{\epsilon }(t_{1},t_{2})}
及条件
U
ϵ
(
t
,
t
)
=
1
{\displaystyle U_{\epsilon }(t,t)=1}
,可以写出方程的形式解
U
ϵ
(
t
1
,
t
2
)
=
1
+
1
i
ℏ
∫
t
2
t
1
d
t
′
(
H
0
+
e
ϵ
(
θ
−
|
t
′
|
)
V
)
U
ϵ
(
t
′
,
t
2
)
.
{\displaystyle U_{\epsilon }(t_{1},t_{2})=1+{\frac {1}{i\hbar }}\int _{t_{2}}^{t_{1}}dt'(H_{0}+e^{\epsilon (\theta -|t'|)}V)U_{\epsilon }(t',t_{2}).}
先集中考虑
0
≥
t
2
≥
t
1
{\displaystyle 0\geq t_{2}\geq t_{1}}
的情形,换元后得到,
U
ϵ
(
t
1
,
t
2
)
=
1
+
1
i
ℏ
∫
θ
+
t
2
θ
+
t
1
d
t
′
(
H
0
+
e
ϵ
t
′
V
)
U
ϵ
(
t
′
−
θ
,
t
2
)
.
{\displaystyle U_{\epsilon }(t_{1},t_{2})=1+{\frac {1}{i\hbar }}\int _{\theta +t_{2}}^{\theta +t_{1}}dt'(H_{0}+e^{\epsilon t'}V)U_{\epsilon }(t'-\theta ,t_{2}).}
于是有
∂
θ
U
ϵ
(
t
1
,
t
2
)
=
ϵ
g
∂
g
U
ϵ
(
t
1
,
t
2
)
=
∂
t
1
U
ϵ
(
t
1
,
t
2
)
+
∂
t
2
U
ϵ
(
t
1
,
t
2
)
.
{\displaystyle \partial _{\theta }U_{\epsilon }(t_{1},t_{2})=\epsilon g\partial _{g}U_{\epsilon }(t_{1},t_{2})=\partial _{t_{1}}U_{\epsilon }(t_{1},t_{2})+\partial _{t_{2}}U_{\epsilon }(t_{1},t_{2}).}
将上式与前面提到的薛定谔方程及其伴式
−
i
ℏ
∂
t
1
U
ϵ
(
t
2
,
t
1
)
=
U
ϵ
(
t
2
,
t
1
)
H
ϵ
(
t
1
)
{\displaystyle -i\hbar \partial _{t_{1}}U_{\epsilon }(t_{2},t_{1})=U_{\epsilon }(t_{2},t_{1})H_{\epsilon }(t_{1})}
结合就有,
i
ℏ
ϵ
g
∂
g
U
ϵ
(
t
1
,
t
2
)
=
H
ϵ
(
t
1
)
U
ϵ
(
t
1
,
t
2
)
−
U
ϵ
(
t
1
,
t
2
)
H
ϵ
(
t
2
)
.
{\displaystyle i\hbar \epsilon g\partial _{g}U_{\epsilon }(t_{1},t_{2})=H_{\epsilon }(t_{1})U_{\epsilon }(t_{1},t_{2})-U_{\epsilon }(t_{1},t_{2})H_{\epsilon }(t_{2}).}
H
ϵ
I
{\displaystyle H_{\epsilon I}}
与
U
ϵ
I
{\displaystyle U_{\epsilon I}}
之间的关系式形式上与上式相同,事实上,将上式两边各左乘
e
i
H
0
t
1
/
ℏ
{\displaystyle e^{iH_{0}t_{1}/\hbar }}
,右乘
e
i
H
0
t
2
/
ℏ
{\displaystyle e^{iH_{0}t_{2}/\hbar }}
,并利用关系
U
ϵ
I
(
t
1
,
t
2
)
=
e
i
H
0
t
1
/
ℏ
U
ϵ
(
t
1
,
t
2
)
e
−
i
H
0
t
2
/
ℏ
.
{\displaystyle U_{\epsilon I}(t_{1},t_{2})=e^{iH_{0}t_{1}/\hbar }U_{\epsilon }(t_{1},t_{2})e^{-iH_{0}t_{2}/\hbar }.}
就可以得到
H
ϵ
I
{\displaystyle H_{\epsilon I}}
与
U
ϵ
I
{\displaystyle U_{\epsilon I}}
之间的关系式。
现在,令
t
1
=
0
,
t
2
=
+
∞
{\displaystyle t_{1}=0,t_{2}=+\infty }
,等式两边作用在
|
Ψ
0
⟩
{\displaystyle |\Psi _{0}\rangle }
上,并注意到
|
Ψ
0
⟩
{\displaystyle |\Psi _{0}\rangle }
是
H
ϵ
(
+
∞
)
=
H
0
{\displaystyle H_{\epsilon }(+\infty )=H_{0}}
的本征态,就有
(
H
ϵ
,
t
=
0
−
E
0
+
i
ℏ
ϵ
g
∂
g
)
U
ϵ
I
(
0
,
∞
)
|
Ψ
0
⟩
=
0.
{\displaystyle \left(H_{\epsilon ,t=0}-E_{0}+i\hbar \epsilon g\partial _{g}\right)U_{\epsilon I}(0,\infty )|\Psi _{0}\rangle =0.}
对于时间为负值的情况,证明完全类似,最后就得到,
(
H
ϵ
,
t
=
0
−
E
0
±
i
ℏ
ϵ
g
∂
g
)
U
ϵ
I
(
0
,
±
∞
)
|
Ψ
0
⟩
=
0.
{\displaystyle \left(H_{\epsilon ,t=0}-E_{0}\pm i\hbar \epsilon g\partial _{g}\right)U_{\epsilon I}(0,\pm \infty )|\Psi _{0}\rangle =0.}
下面以时间为负值为例继续证明,为清晰起见,先把算符写成简略形式,即将
U
ϵ
I
(
0
,
−
∞
)
{\displaystyle U_{\epsilon I}(0,-\infty )}
简写作
U
{\displaystyle U}
。
i
ℏ
ϵ
g
∂
g
(
U
|
Ψ
0
⟩
)
=
(
H
ϵ
−
E
0
)
U
|
Ψ
0
⟩
.
{\displaystyle i\hbar \epsilon g\partial _{g}\left(U|\Psi _{0}\rangle \right)=(H_{\epsilon }-E_{0})U|\Psi _{0}\rangle .}
下面计算
i
ℏ
ϵ
g
∂
g
|
Ψ
ϵ
−
⟩
{\displaystyle i\hbar \epsilon g\partial _{g}|\Psi _{\epsilon }^{-}\rangle }
,把
|
Ψ
ϵ
−
⟩
{\displaystyle |\Psi _{\epsilon }^{-}\rangle }
的定义式代入,并利用上面的关系式,可得,
i
ℏ
ϵ
g
∂
g
|
Ψ
ϵ
−
⟩
=
1
⟨
Ψ
0
|
U
|
Ψ
0
⟩
(
H
ϵ
−
E
0
)
U
|
Ψ
0
⟩
−
U
|
Ψ
0
⟩
⟨
Ψ
0
|
U
|
Ψ
0
⟩
2
⟨
Ψ
0
|
H
ϵ
−
E
0
|
Ψ
0
⟩
=
(
H
ϵ
−
E
0
)
|
Ψ
ϵ
−
⟩
−
|
Ψ
ϵ
−
⟩
⟨
Ψ
0
|
H
ϵ
−
E
0
|
Ψ
ϵ
−
⟩
=
[
H
ϵ
−
E
−
]
|
Ψ
ϵ
−
⟩
.
{\displaystyle {\begin{aligned}i\hbar \epsilon g\partial _{g}|\Psi _{\epsilon }^{-}\rangle &={\frac {1}{\langle \Psi _{0}|U|\Psi _{0}\rangle }}(H_{\epsilon }-E_{0})U|\Psi _{0}\rangle -{\frac {U|\Psi _{0}\rangle }{{\langle \Psi _{0}|U|\Psi _{0}\rangle }^{2}}}\langle \Psi _{0}|H_{\epsilon }-E_{0}|\Psi _{0}\rangle \\&=(H_{\epsilon }-E_{0})|\Psi _{\epsilon }^{-}\rangle -|\Psi _{\epsilon }^{-}\rangle \langle \Psi _{0}|H_{\epsilon }-E_{0}|\Psi _{\epsilon }^{-}\rangle \\&=\left[H_{\epsilon }-E^{-}\right]|\Psi _{\epsilon }^{-}\rangle .\end{aligned}}}
式中
E
−
=
E
0
+
⟨
Ψ
0
|
H
ϵ
−
H
0
|
Ψ
ϵ
−
⟩
{\displaystyle E^{-}=E_{0}+\langle \Psi _{0}|H_{\epsilon }-H_{0}|\Psi _{\epsilon }^{-}\rangle }
.
即
[
H
ϵ
−
E
−
−
i
ℏ
ϵ
g
∂
g
]
|
Ψ
ϵ
−
⟩
=
0.
{\displaystyle \left[H_{\epsilon }-E^{-}-i\hbar \epsilon g\partial _{g}\right]|\Psi _{\epsilon }^{-}\rangle =0.}
类似地可证明
|
Ψ
ϵ
+
⟩
{\displaystyle |\Psi _{\epsilon }^{+}\rangle }
的关系式,综合起来可写成:
[
H
ϵ
−
E
±
±
i
ℏ
ϵ
g
∂
g
]
|
Ψ
ϵ
±
⟩
=
0
{\displaystyle \left[H_{\epsilon }-E^{\pm }\pm i\hbar \epsilon g\partial _{g}\right]|\Psi _{\epsilon }^{\pm }\rangle =0}
然后取
ϵ
→
0
+
{\displaystyle \epsilon \rightarrow 0^{+}}
的极限,即可证明
|
Ψ
ϵ
±
⟩
{\displaystyle |\Psi _{\epsilon }^{\pm }\rangle }
是
H
{\displaystyle H}
的本征函数,本征值分别为
E
±
{\displaystyle E^{\pm }}
[ 3] 。
K. Hepp: Lecture Notes in Physics (Springer-Verlag, New York, 1969), Vol. 2.
G. Nenciu and G. Rasche: "Adiabatic theorem and Gell-Mann-Low formula", Helv. Phys. Acta 62, 372 (1989).
A.L. Fetter and J.D. Walecka: "Quantum Theory of Many-Particle Systems", McGraw–Hill (1971)