Jump to content

Green's function (many-body theory): Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
Ququ (talk | contribs)
Undid revision 914449359 by Mcirma44 (talk)
 
(10 intermediate revisions by 6 users not shown)
Line 1: Line 1:
{{Short description|Correlators of field operators}}
In [[many-body theory]], the term '''Green's function''' (or '''Green function''') is sometimes used interchangeably with [[Correlation function (quantum field theory)|correlation function]], but refers specifically to correlators of [[field operator]]s or [[creation and annihilation operators]].
In [[many-body theory]], the term '''Green's function''' (or '''Green function''') is sometimes used interchangeably with [[Correlation function (quantum field theory)|correlation function]], but refers specifically to correlators of [[field operator]]s or [[creation and annihilation operators]].


The name comes from the [[Green's functions]] used to solve inhomogeneous [[differential equations]], to which they are loosely related. (Specifically, only two-point 'Green's functions' in the case of a non-interacting system are Green's functions in the mathematical sense; the linear operator that they invert is the [[Hamiltonian (quantum mechanics)|Hamiltonian operator]], which in the non-interacting case is quadratic in the fields.)
The name comes from the [[Green's functions]] used to solve inhomogeneous [[differential equations]], to which they are loosely related. (Specifically, only two-point "Green's functions" in the case of a non-interacting system are Green's functions in the mathematical sense; the linear operator that they invert is the [[Hamiltonian (quantum mechanics)|Hamiltonian operator]], which in the non-interacting case is quadratic in the fields.)


==Spatially uniform case==
==Spatially uniform case==
Line 10: Line 11:


The [[Heisenberg picture|Heisenberg operators]] can be written in terms of [[Schrödinger picture|Schrödinger operators]] as
The [[Heisenberg picture|Heisenberg operators]] can be written in terms of [[Schrödinger picture|Schrödinger operators]] as
<math display="block">\psi(\mathbf{x},t) = e^{i K t} \psi(\mathbf{x}) e^{-i K t},
:<math>
</math>and the creation operator is <math>\bar\psi(\mathbf{x},t) = [\psi(\mathbf{x},t)]^\dagger</math>, where <math>K = H - \mu N</math> is the [[Grand canonical ensemble|grand-canonical]] Hamiltonian.
\psi(\mathbf{x},t) = \mathrm{e}^{\mathrm{i} K t} \psi(\mathbf{x}) \mathrm{e}^{-\mathrm{i} K t},
</math>
and the creation operator is <math>\bar\psi(\mathbf{x},t) = [\psi(\mathbf{x},t)]^\dagger</math>, where <math>K = H - \mu N</math> is the [[Grand canonical ensemble|grand-canonical]] Hamiltonian.


Similarly, for the [[Imaginary time|imaginary-time]] operators,
Similarly, for the [[Imaginary time|imaginary-time]] operators,
<math display="block">\psi(\mathbf{x},\tau) = e^{K \tau} \psi(\mathbf{x}) e^{-K\tau}</math>
:<math>
\psi(\mathbf{x},\tau) = \mathrm{e}^{K \tau} \psi(\mathbf{x}) \mathrm{e}^{-K\tau}
<math display="block">\bar\psi(\mathbf{x},\tau) = e^{K \tau} \psi^\dagger(\mathbf{x}) e^{-K\tau}.</math>
</math>
:<math>
\bar\psi(\mathbf{x},\tau) = \mathrm{e}^{K \tau} \psi^\dagger(\mathbf{x}) \mathrm{e}^{-K\tau}.
</math>
[Note that the imaginary-time creation operator <math>\bar\psi(\mathbf{x},\tau)</math> is not the [[Hermitian conjugate]] of the annihilation operator <math>\psi(\mathbf{x},\tau)</math>.]
[Note that the imaginary-time creation operator <math>\bar\psi(\mathbf{x},\tau)</math> is not the [[Hermitian conjugate]] of the annihilation operator <math>\psi(\mathbf{x},\tau)</math>.]


In real time, the <math>2n</math>-point Green function is defined by
In real time, the <math>2n</math>-point Green function is defined by
<math display="block"> G^{(n)}(1 \ldots n \mid 1' \ldots n') = i^n \langle T\psi(1)\ldots\psi(n)\bar\psi(n')\ldots\bar\psi(1')\rangle, </math>
:<math>
G^{(n)}(1 \ldots n \mid 1' \ldots n')
= i^n \langle T\psi(1)\ldots\psi(n)\bar\psi(n')\ldots\bar\psi(1')\rangle,
</math>
where we have used a condensed notation in which <math>j</math> signifies <math>(\mathbf{x}_j, t_j)</math> and <math>j'</math> signifies <math>(\mathbf{x}_j', t_j')</math>. The operator <math>T</math> denotes [[time ordering]], and indicates that the field operators that follow it are to be ordered so that their time arguments increase from right to left.
where we have used a condensed notation in which <math>j</math> signifies <math>(\mathbf{x}_j, t_j)</math> and <math>j'</math> signifies <math>(\mathbf{x}_j', t_j')</math>. The operator <math>T</math> denotes [[time ordering]], and indicates that the field operators that follow it are to be ordered so that their time arguments increase from right to left.


In imaginary time, the corresponding definition is
In imaginary time, the corresponding definition is
<math display="block"> \mathcal{G}^{(n)}(1 \ldots n \mid 1' \ldots n') = \langle T\psi(1)\ldots\psi(n)\bar\psi(n')\ldots\bar\psi(1')\rangle, </math>
:<math>
where <math>j</math> signifies <math>\mathbf{x}_j, \tau_j</math>. (The imaginary-time variables <math>\tau_j</math> are restricted to the range from <math>0</math> to the inverse temperature <math display="inline">\beta = \frac{1}{k_\text{B} T}</math>.)
\mathcal{G}^{(n)}(1 \ldots n \mid 1' \ldots n')
= \langle T\psi(1)\ldots\psi(n)\bar\psi(n')\ldots\bar\psi(1')\rangle,
</math>
where <math>j</math> signifies <math>\mathbf{x}_j, \tau_j</math>. (The imaginary-time variables <math>\tau_j</math> are restricted to the range from <math>0</math> to the inverse temperature <math>\beta=\frac{1}{k_B T}</math>.)


'''Note''' regarding signs and normalization used in these definitions: The signs of the Green functions have been chosen so that [[Fourier transform]] of the two-point (<math>n=1</math>) thermal Green function for a free particle is
'''Note''' regarding signs and normalization used in these definitions: The signs of the Green functions have been chosen so that [[Fourier transform]] of the two-point (<math>n=1</math>) thermal Green function for a free particle is
<math display="block"> \mathcal{G}(\mathbf{k},\omega_n) = \frac{1}{-i\omega_n + \xi_\mathbf{k}}, </math>
:<math>
\mathcal{G}(\mathbf{k},\omega_n) = \frac{1}{-\mathrm{i}\omega_n + \xi_\mathbf{k}},
</math>
and the retarded Green function is
and the retarded Green function is
<math display="block">G^{\mathrm{R}}(\mathbf{k},\omega) = \frac{1}{-(\omega+i\eta) + \xi_\mathbf{k}},</math>
:<math>
where <math display="block">\omega_n = \frac{[2n+\theta(-\zeta)]\pi}{\beta}</math> is the [[Matsubara frequency]].
G^{\mathrm{R}}(\mathbf{k},\omega) = \frac{1}{-(\omega+\mathrm{i}\eta) + \xi_\mathbf{k}},
</math>
where
:<math>
\omega_n = {[2n+\theta(-\zeta)]\pi}/{\beta}
</math>
is the [[Matsubara frequency]].


Throughout, <math>\zeta</math> is <math>+1</math> for [[boson]]s and <math>-1</math> for [[fermion]]s and <math>[\ldots,\ldots]=[\ldots,\ldots]_{-\zeta}</math> denotes either a [[commutator]] or anticommutator as appropriate.
Throughout, <math>\zeta</math> is <math>+1</math> for [[boson]]s and <math>-1</math> for [[fermion]]s and <math>[\ldots,\ldots]=[\ldots,\ldots]_{-\zeta}</math> denotes either a [[commutator]] or anticommutator as appropriate.
Line 59: Line 40:


The Green function with a single pair of arguments (<math>n=1</math>) is referred to as the two-point function, or [[propagator]]. In the presence of both spatial and temporal translational symmetry, it depends only on the difference of its arguments. Taking the Fourier transform with respect to both space and time gives
The Green function with a single pair of arguments (<math>n=1</math>) is referred to as the two-point function, or [[propagator]]. In the presence of both spatial and temporal translational symmetry, it depends only on the difference of its arguments. Taking the Fourier transform with respect to both space and time gives
<math display="block">\mathcal{G}(\mathbf{x}\tau\mid\mathbf{x}'\tau') = \int_\mathbf{k} d\mathbf{k} \frac{1}{\beta}\sum_{\omega_n} \mathcal{G}(\mathbf{k},\omega_n) e^{i \mathbf{k}\cdot(\mathbf{x}-\mathbf{x}')-i\omega_n (\tau-\tau')},</math>
:<math>
\mathcal{G}(\mathbf{x}\tau\mid\mathbf{x}'\tau') = \int_\mathbf{k} d\mathbf{k} \frac{1}{\beta}\sum_{\omega_n} \mathcal{G}(\mathbf{k},\omega_n) \mathrm{e}^{\mathrm{i} \mathbf{k}\cdot(\mathbf{x}-\mathbf{x}')-\mathrm{i}\omega_n (\tau-\tau')},
</math>
where the sum is over the appropriate [[Matsubara frequency|Matsubara frequencies]] (and the integral involves an implicit factor of <math>(L/2\pi)^{d}</math>, as usual).
where the sum is over the appropriate [[Matsubara frequency|Matsubara frequencies]] (and the integral involves an implicit factor of <math>(L/2\pi)^{d}</math>, as usual).


In real time, we will explicitly indicate the time-ordered function with a superscript T:
In real time, we will explicitly indicate the time-ordered function with a superscript T:
<math display="block">G^{\mathrm{T}}(\mathbf{x} t\mid\mathbf{x}' t') = \int_\mathbf{k} d \mathbf{k} \int \frac{d\omega}{2\pi} G^{\mathrm{T}}(\mathbf{k},\omega) e^{i \mathbf{k}\cdot(\mathbf{x} -\mathbf{x} ')-i\omega(t-t')}.</math>
:<math>
G^{\mathrm{T}}(\mathbf{x} t\mid\mathbf{x}' t') = \int_\mathbf{k} d \mathbf{k} \int \frac{\mathrm{d}\omega}{2\pi} G^{\mathrm{T}}(\mathbf{k},\omega) \mathrm{e}^{\mathrm{i} \mathbf{k}\cdot(\mathbf{x} -\mathbf{x} ')-\mathrm{i}\omega(t-t')}.
</math>


The real-time two-point Green function can be written in terms of 'retarded' and 'advanced' Green functions, which will turn out to have simpler analyticity properties. The retarded and advanced Green functions are defined by
The real-time two-point Green function can be written in terms of 'retarded' and 'advanced' Green functions, which will turn out to have simpler analyticity properties. The retarded and advanced Green functions are defined by
<math display="block">G^{\mathrm{R}}(\mathbf{x} t \mid \mathbf{x}' t') = -i\langle[\psi(\mathbf{x} ,t),\bar\psi(\mathbf{x} ',t')]_{\zeta}\rangle\Theta(t-t')</math>
:<math>
G^{\mathrm{R}}(\mathbf{x} t\mid\mathbf{x} 't') = -\mathrm{i}\langle[\psi(\mathbf{x} ,t),\bar\psi(\mathbf{x} ',t')]_{\zeta}\rangle\Theta(t-t')
</math>
and
and
<math display="block">G^{\mathrm{A}}(\mathbf{x} t\mid\mathbf{x} 't') = i\langle[\psi(\mathbf{x} ,t),\bar\psi(\mathbf{x}', t')]_{\zeta}\rangle \Theta(t'-t),</math>
:<math>
G^{\mathrm{A}}(\mathbf{x} t\mid\mathbf{x} 't') = \mathrm{i}\langle[\psi(\mathbf{x} ,t),\bar\psi(\mathbf{x} ',t')]_{\zeta}\rangle\Theta(t'-t),
</math>
respectively.
respectively.


They are related to the time-ordered Green function by
They are related to the time-ordered Green function by
<math display="block">G^{\mathrm{T}}(\mathbf{k},\omega) = [1+\zeta n(\omega)]G^{\mathrm{R}}(\mathbf{k},\omega) - \zeta n(\omega) G^{\mathrm{A}}(\mathbf{k},\omega),</math>
:<math>
G^{\mathrm{T}}(\mathbf{k},\omega) = [1+\zeta n(\omega)]G^{\mathrm{R}}(\mathbf{k},\omega) - \zeta n(\omega) G^{\mathrm{A}}(\mathbf{k},\omega),
</math>
where
where
<math display="block">n(\omega) = \frac{1}{e^{\beta \omega}-\zeta}</math>
:<math>
n(\omega) = \frac{1}{\mathrm{e}^{\beta \omega}-\zeta}
</math>
is the [[Bose–Einstein statistics|Bose–Einstein]] or [[Fermi–Dirac statistics|Fermi–Dirac]] distribution function.
is the [[Bose–Einstein statistics|Bose–Einstein]] or [[Fermi–Dirac statistics|Fermi–Dirac]] distribution function.


====Imaginary-time ordering and ''&beta;''-periodicity====
====Imaginary-time ordering and ''β''-periodicity====


The thermal Green functions are defined only when both imaginary-time arguments are within the range <math>0</math> to <math>\beta</math>. The two-point Green function has the following properties. (The position or momentum arguments are suppressed in this section.)
The thermal Green functions are defined only when both imaginary-time arguments are within the range <math>0</math> to <math>\beta</math>. The two-point Green function has the following properties. (The position or momentum arguments are suppressed in this section.)


Firstly, it depends only on the difference of the imaginary times:
Firstly, it depends only on the difference of the imaginary times:
<math display="block">\mathcal{G}(\tau,\tau') = \mathcal{G}(\tau - \tau').</math>
:<math>
\mathcal{G}(\tau,\tau') = \mathcal{G}(\tau - \tau').
</math>
The argument <math>\tau - \tau'</math> is allowed to run from <math>-\beta</math> to <math>\beta</math>.
The argument <math>\tau - \tau'</math> is allowed to run from <math>-\beta</math> to <math>\beta</math>.


Secondly, <math>\mathcal{G}(\tau)</math> is (anti)periodic under shifts of <math>\beta</math>. Because of the small domain within which the function is defined, this means just
Secondly, <math>\mathcal{G}(\tau)</math> is (anti)periodic under shifts of <math>\beta</math>. Because of the small domain within which the function is defined, this means just
<math display="block">\mathcal{G}(\tau - \beta) = \zeta \mathcal{G}(\tau),</math>
:<math>
\mathcal{G}(\tau - \beta) = \zeta \mathcal{G}(\tau),
</math>
for <math>0 < \tau < \beta</math>. Time ordering is crucial for this property, which can be proved straightforwardly, using the cyclicity of the trace operation.
for <math>0 < \tau < \beta</math>. Time ordering is crucial for this property, which can be proved straightforwardly, using the cyclicity of the trace operation.


These two properties allow for the Fourier transform representation and its inverse,
These two properties allow for the Fourier transform representation and its inverse,
<math display="block">\mathcal{G}(\omega_n) = \int_0^\beta d\tau \, \mathcal{G}(\tau)\, e^{i\omega_n \tau}.</math>
:<math>
\mathcal{G}(\omega_n) = \int_0^\beta \mathrm{d}\tau \, \mathcal{G}(\tau)\, \mathrm{e}^{\mathrm{i}\omega_n \tau}.
</math>


Finally, note that <math>\mathcal{G}(\tau)</math> has a discontinuity at <math>\tau = 0</math>; this is consistent with a long-distance behaviour of <math>\mathcal{G}(\omega_n) \sim 1/|\omega_n|</math>.
Finally, note that <math>\mathcal{G}(\tau)</math> has a discontinuity at <math>\tau = 0</math>; this is consistent with a long-distance behaviour of <math>\mathcal{G}(\omega_n) \sim 1/|\omega_n|</math>.
Line 114: Line 77:
===Spectral representation===
===Spectral representation===
The [[propagator]]s in real and imaginary time can both be related to the spectral density (or spectral weight), given by
The [[propagator]]s in real and imaginary time can both be related to the spectral density (or spectral weight), given by
<math display="block">\rho(\mathbf{k},\omega) = \frac{1}{\mathcal{Z}}\sum_{\alpha,\alpha'} 2\pi \delta(E_\alpha-E_{\alpha'} - \omega) |\langle \alpha \mid \psi_\mathbf{k}^\dagger \mid \alpha'\rangle|^2 \left(e^{-\beta E_{\alpha'}} - \zeta e^{-\beta E_{\alpha}}\right),</math>
:<math>
where {{math|{{!}}''α''⟩}} refers to a (many-body) eigenstate of the grand-canonical Hamiltonian {{math|''H'' − ''μN''}}, with eigenvalue {{math|''E<sub>α</sub>''}}.
\rho(\mathbf{k},\omega) = \frac{1}{\mathcal{Z}}\sum_{\alpha,\alpha'} 2\pi \delta(E_\alpha-E_{\alpha'}-\omega) |\langle \alpha \mid \psi_\mathbf{k}^\dagger\mid \alpha'\rangle|^2 \left(\mathrm{e}^{-\beta E_{\alpha'}} - \zeta\mathrm{e}^{-\beta E_{\alpha}}\right),
</math>
where |{{mvar|α}}⟩ refers to a (many-body) eigenstate of the grand-canonical Hamiltonian {{math|''H''&nbsp;−&nbsp;''μN''}}, with eigenvalue {{math|''E<sub>α</sub>''}}.


The imaginary-time [[propagator]] is then given by
The imaginary-time [[propagator]] is then given by
:<math>
<math display="block">
\mathcal{G}(\mathbf{k},\omega_n) = \int_{-\infty}^\infty \frac{\mathrm{d}\omega'}{2\pi}
\mathcal{G}(\mathbf{k},\omega_n) = \int_{-\infty}^\infty \frac{d\omega'}{2\pi} \frac{\rho(\mathbf{k},\omega')}{-i\omega_n+\omega'}~,
\frac{\rho(\mathbf{k},\omega')}{-\mathrm{i}\omega_n+\omega'}~,
</math>
</math>
and the retarded [[propagator]] by
and the retarded [[propagator]] by
<math display="block">G^{\mathrm{R}}(\mathbf{k},\omega) = \int_{-\infty}^\infty \frac{d\omega'}{2\pi} \frac{\rho(\mathbf{k},\omega')}{-(\omega+i\eta)+\omega'},</math>
:<math>
where the limit as <math>\eta \to 0^+</math> is implied.
G^{\mathrm{R}}(\mathbf{k},\omega) = \int_{-\infty}^\infty \frac{\mathrm{d}\omega'}{2\pi}
\frac{\rho(\mathbf{k},\omega')}{-(\omega+\mathrm{i}\eta)+\omega'},
</math>
where the limit as <math>\eta\rightarrow 0^+</math> is implied.


The advanced propagator is given by the same expression, but with <math>-\mathrm{i}\eta</math> in the denominator.
The advanced propagator is given by the same expression, but with <math>-i\eta</math> in the denominator.


The time-ordered function can be found in terms of <math>G^{\mathrm{R}}</math> and <math>G^{\mathrm{A}}</math>. As claimed above, <math>G^{\mathrm{R}}(\omega)</math> and <math>G^{\mathrm{A}}(\omega)</math> have simple analyticity properties: the former (latter) has all its poles and discontinuities in the lower (upper) half-plane.
The time-ordered function can be found in terms of <math>G^{\mathrm{R}}</math> and <math>G^{\mathrm{A}}</math>. As claimed above, <math>G^{\mathrm{R}}(\omega)</math> and <math>G^{\mathrm{A}}(\omega)</math> have simple analyticity properties: the former (latter) has all its poles and discontinuities in the lower (upper) half-plane.
Line 138: Line 95:


The spectral density can be found very straightforwardly from <math>G^{\mathrm{R}}</math>, using the [[Sokhatsky–Weierstrass theorem]]
The spectral density can be found very straightforwardly from <math>G^{\mathrm{R}}</math>, using the [[Sokhatsky–Weierstrass theorem]]
<math display="block">\lim_{\eta \to 0^+} \frac{1}{x\pm i\eta} = P\frac{1}{x} \mp i\pi\delta(x),</math>
:<math>
\lim_{\eta\rightarrow 0^+}\frac{1}{x\pm\mathrm{i}\eta} = {P}\frac{1}{x}\mp i\pi\delta(x),
</math>
where {{mvar|P}} denotes the [[Cauchy principal part]].
where {{mvar|P}} denotes the [[Cauchy principal part]].
This gives
This gives
<math display="block">\rho(\mathbf{k},\omega) = 2\operatorname{Im} G^{\mathrm{R}}(\mathbf{k},\omega).</math>
:<math>
\rho(\mathbf{k},\omega) = 2\mathrm{Im}\, G^{\mathrm{R}}(\mathbf{k},\omega).
</math>


This furthermore implies that <math>G^{\mathrm{R}}(\mathbf{k},\omega)</math> obeys the following relationship between its real and imaginary parts:
This furthermore implies that <math>G^{\mathrm{R}}(\mathbf{k},\omega)</math> obeys the following relationship between its real and imaginary parts:
<math display="block">\operatorname{Re} G^{\mathrm{R}}(\mathbf{k},\omega) = -2 P \int_{-\infty}^\infty \frac{d\omega'}{2\pi} \frac{\operatorname{Im} G^{\mathrm{R}}(\mathbf{k},\omega')}{\omega-\omega'},</math>
:<math>
\mathrm{Re}\, G^{\mathrm{R}}(\mathbf{k},\omega) = -2 P \int_{-\infty}^\infty \frac{\mathrm{d}\omega'}{2\pi}
\frac{\mathrm{Im}\, G^{\mathrm{R}}(\mathbf{k},\omega')}{\omega-\omega'},
</math>
where <math>P</math> denotes the principal value of the integral.
where <math>P</math> denotes the principal value of the integral.


The spectral density obeys a sum rule,
The spectral density obeys a sum rule,
<math display="block">\int_{-\infty}^\infty \frac{d\omega}{2\pi} \rho(\mathbf{k},\omega) = 1,</math>
:<math>
\int_{-\infty}^{\infty} \frac{\mathrm{d}\omega}{2\pi} \rho(\mathbf{k},\omega) = 1,
</math>
which gives
which gives
<math display="block">G^{\mathrm{R}}(\omega)\sim\frac{1}{|\omega|}</math>
:<math>
as <math>|\omega| \to \infty</math>.
G^{\mathrm{R}}(\omega)\sim\frac{1}{|\omega|}
</math>
as <math>|\omega| \rightarrow \infty</math>.


====Hilbert transform====
====Hilbert transform====
The similarity of the spectral representations of the imaginary- and real-time Green functions allows us to define the function
The similarity of the spectral representations of the imaginary- and real-time Green functions allows us to define the function
<math display="block">G(\mathbf{k},z) = \int_{-\infty}^\infty \frac{dx}{2\pi} \frac{\rho(\mathbf{k},x)}{-z+x},</math>
:<math>
G(\mathbf{k},z) = \int_{-\infty}^\infty \frac{\mathrm{d} x}{2\pi} \frac{\rho(\mathbf{k},x)}{-z+x},
</math>
which is related to <math>\mathcal{G}</math> and <math>G^{\mathrm{R}}</math> by
which is related to <math>\mathcal{G}</math> and <math>G^{\mathrm{R}}</math> by
<math display="block">\mathcal{G}(\mathbf{k},\omega_n) = G(\mathbf{k}, i\omega_n)</math>
:<math>
\mathcal{G}(\mathbf{k},\omega_n) = G(\mathbf{k},\mathrm{i}\omega_n)
</math>
and
and
<math display="block">G^{\mathrm{R}}(\mathbf{k},\omega) = G(\mathbf{k},\omega + i\eta).</math>
:<math>
G^{\mathrm{R}}(\mathbf{k},\omega) = G(\mathbf{k},\omega + \mathrm{i}\eta).
</math>
A similar expression obviously holds for <math>G^{\mathrm{A}}</math>.
A similar expression obviously holds for <math>G^{\mathrm{A}}</math>.


Line 184: Line 124:


We demonstrate the proof of the spectral representation of the propagator in the case of the thermal Green function, defined as
We demonstrate the proof of the spectral representation of the propagator in the case of the thermal Green function, defined as
<math display="block">\mathcal{G}(\mathbf{x} , \tau\mid\mathbf{x} ',\tau') = \langle T\psi(\mathbf{x} ,\tau)\bar\psi(\mathbf{x} ', \tau') \rangle.</math>
:<math>
\mathcal{G}(\mathbf{x} ,\tau\mid\mathbf{x} ',\tau') = \langle T\psi(\mathbf{x} ,\tau)\bar\psi(\mathbf{x} ',\tau')\rangle.
</math>


Due to translational symmetry, it is only necessary to consider <math>\mathcal{G}(\mathbf{x} ,\tau\mid\mathbf{0},0)</math> for <math>\tau > 0</math>, given by
Due to translational symmetry, it is only necessary to consider <math>\mathcal{G}(\mathbf{x} ,\tau\mid\mathbf{0},0)</math> for <math>\tau > 0</math>, given by
:<math>
<math display="block">
\mathcal{G}(\mathbf{x},\tau\mid\mathbf{0},0) = \frac{1}{\mathcal{Z}}\sum_{\alpha'} \mathrm{e}^{-\beta E_{\alpha'}}
\mathcal{G}(\mathbf{x},\tau\mid\mathbf{0},0) = \frac{1}{\mathcal{Z}}\sum_{\alpha'} e^{-\beta E_{\alpha'}}
\langle\alpha' \mid \psi(\mathbf{x},\tau)\bar\psi(\mathbf{0},0) \mid \alpha' \rangle.
\langle\alpha' \mid \psi(\mathbf{x},\tau)\bar\psi(\mathbf{0},0) \mid \alpha' \rangle.
</math>
</math>
Inserting a complete set of eigenstates gives
Inserting a complete set of eigenstates gives
:<math>
<math display="block">
\mathcal{G}(\mathbf{x} ,\tau|\mathbf{0},0) = \frac{1}{\mathcal{Z}}\sum_{\alpha,\alpha'} \mathrm{e}^{-\beta E_{\alpha'}}
\mathcal{G}(\mathbf{x} ,\tau\mid\mathbf{0},0) = \frac{1}{\mathcal{Z}}\sum_{\alpha,\alpha'} e^{-\beta E_{\alpha'}}
\langle\alpha' \mid \psi(\mathbf{x} ,\tau)\mid\alpha \rangle\langle\alpha \mid \bar\psi(\mathbf{0},0) \mid \alpha' \rangle.
\langle\alpha' \mid \psi(\mathbf{x} ,\tau)\mid\alpha \rangle\langle\alpha \mid \bar\psi(\mathbf{0},0) \mid \alpha' \rangle.
</math>
</math>


Since <math>|\alpha \rangle</math> and <math>|\alpha' \rangle</math> are eigenstates of <math>H-\mu N</math>, the Heisenberg operators can be rewritten in terms of Schrödinger operators, giving
Since <math>|\alpha \rangle</math> and <math>|\alpha' \rangle</math> are eigenstates of <math>H-\mu N</math>, the Heisenberg operators can be rewritten in terms of Schrödinger operators, giving
:<math>
<math display="block">
\mathcal{G}(\mathbf{x} ,\tau|\mathbf{0},0) = \frac{1}{\mathcal{Z}}\sum_{\alpha,\alpha'} \mathrm{e}^{-\beta E_{\alpha'}}
\mathcal{G}(\mathbf{x} ,\tau|\mathbf{0},0) = \frac{1}{\mathcal{Z}}\sum_{\alpha,\alpha'} e^{-\beta E_{\alpha'}}
\mathrm{e}^{\tau(E_{\alpha'} - E_\alpha)}\langle\alpha' \mid \psi(\mathbf{x} )\mid\alpha \rangle\langle\alpha \mid \psi^\dagger(\mathbf{0}) \mid \alpha' \rangle.
e^{\tau(E_{\alpha'} - E_\alpha)}\langle\alpha' \mid \psi(\mathbf{x} )\mid\alpha \rangle \langle\alpha \mid \psi^\dagger(\mathbf{0}) \mid \alpha' \rangle.
</math>
</math>
Performing the Fourier transform then gives
Performing the Fourier transform then gives
:<math>
<math display="block">
\mathcal{G}(\mathbf{k},\omega_n) = \frac{1}{\mathcal{Z}} \sum_{\alpha,\alpha'} \mathrm{e}^{-\beta E_{\alpha'}} \frac{1-\zeta \mathrm{e}^{\beta(E_{\alpha'} - E_\alpha)}}{-\mathrm{i}\omega_n + E_\alpha - E_{\alpha'}} \int_{\mathbf{k}'} d\mathbf{k}' \langle\alpha \mid \psi(\mathbf{k}) \mid \alpha' \rangle\langle\alpha' \mid \psi^\dagger(\mathbf{k}')\mid\alpha \rangle.
\mathcal{G}(\mathbf{k},\omega_n) = \frac{1}{\mathcal{Z}} \sum_{\alpha,\alpha'} e^{-\beta E_{\alpha'}} \frac{1-\zeta e^{\beta(E_{\alpha'} - E_\alpha)}}{-i\omega_n + E_\alpha - E_{\alpha'}} \int_{\mathbf{k}'} d\mathbf{k}' \langle\alpha \mid \psi(\mathbf{k}) \mid \alpha' \rangle\langle\alpha' \mid \psi^\dagger(\mathbf{k}')\mid\alpha \rangle.
</math>
</math>


Momentum conservation allows the final term to be written as (up to possible factors of the volume)
Momentum conservation allows the final term to be written as (up to possible factors of the volume)
<math display="block">|\langle\alpha' \mid\psi^\dagger(\mathbf{k})\mid\alpha \rangle|^2,</math>
:<math>
|\langle\alpha' \mid\psi^\dagger(\mathbf{k})\mid\alpha \rangle|^2,
</math>
which confirms the expressions for the Green functions in the spectral representation.
which confirms the expressions for the Green functions in the spectral representation.


The sum rule can be proved by considering the expectation value of the commutator,
The sum rule can be proved by considering the expectation value of the commutator,
<math display="block">1 = \frac{1}{\mathcal{Z}} \sum_\alpha \langle\alpha \mid e^{-\beta(H-\mu N)}[\psi_\mathbf{k},\psi_\mathbf{k}^\dagger]_{-\zeta} \mid \alpha \rangle,</math>
:<math>
1 = \frac{1}{\mathcal{Z}} \sum_\alpha \langle\alpha \mid \mathrm{e}^{-\beta(H-\mu N)}[\psi_\mathbf{k},\psi_\mathbf{k}^\dagger]_{-\zeta} \mid \alpha \rangle,
</math>
and then inserting a complete set of eigenstates into both terms of the commutator:
and then inserting a complete set of eigenstates into both terms of the commutator:
: <math>
<math display="block">
1 = \frac{1}{\mathcal{Z}} \sum_{\alpha,\alpha'} \mathrm{e}^{-\beta E_\alpha} \left(
1 = \frac{1}{\mathcal{Z}} \sum_{\alpha,\alpha'} e^{-\beta E_\alpha} \left(
\langle\alpha \mid \psi_\mathbf{k} \mid \alpha' \rangle\langle\alpha' \mid \psi_\mathbf{k}^\dagger \mid \alpha \rangle - \zeta \langle\alpha \mid \psi_\mathbf{k}^\dagger \mid \alpha' \rangle\langle\alpha' \mid \psi_\mathbf{k}\mid\alpha \rangle
\langle\alpha \mid \psi_\mathbf{k} \mid \alpha' \rangle\langle\alpha' \mid \psi_\mathbf{k}^\dagger \mid \alpha \rangle - \zeta \langle\alpha \mid \psi_\mathbf{k}^\dagger \mid \alpha' \rangle\langle\alpha' \mid \psi_\mathbf{k}\mid\alpha \rangle
\right).
\right).
Line 227: Line 161:


Swapping the labels in the first term then gives
Swapping the labels in the first term then gives
:<math>
<math display="block">
1 = \frac{1}{\mathcal{Z}} \sum_{\alpha,\alpha'}
1 = \frac{1}{\mathcal{Z}} \sum_{\alpha,\alpha'}
\left(\mathrm{e}^{-\beta E_{\alpha'}} - \zeta \mathrm{e}^{-\beta E_\alpha} \right)
\left(e^{-\beta E_{\alpha'}} - \zeta e^{-\beta E_\alpha} \right)
|\langle\alpha \mid \psi_\mathbf{k}^\dagger \mid \alpha' \rangle|^2 ~,
|\langle\alpha \mid \psi_\mathbf{k}^\dagger \mid \alpha' \rangle|^2 ~,
</math>
</math>
which is exactly the result of the integration of {{mvar|ρ}}.
which is exactly the result of the integration of {{mvar|ρ}}.


====Non-interacting case====
====Non-interacting case====
In the non-interacting case, <math>\psi_\mathbf{k}^\dagger\mid\alpha' \rangle</math> is an eigenstate with (grand-canonical) energy <math>E_{\alpha'} + \xi_\mathbf{k}</math>, where <math>\xi_\mathbf{k} = \epsilon_\mathbf{k} - \mu</math> is the single-particle dispersion relation measured with respect to the chemical potential. The spectral density therefore becomes
In the non-interacting case, <math>\psi_\mathbf{k}^\dagger\mid\alpha' \rangle</math> is an eigenstate with (grand-canonical) energy <math>E_{\alpha'} + \xi_\mathbf{k}</math>, where <math>\xi_\mathbf{k} = \epsilon_\mathbf{k} - \mu</math> is the single-particle dispersion relation measured with respect to the chemical potential. The spectral density therefore becomes
:<math>
<math display="block">
\rho_0(\mathbf{k},\omega) = \frac{1}{\mathcal{Z}}\,2\pi\delta(\xi_\mathbf{k} - \omega) \sum_{\alpha'}\langle\alpha' \mid\psi_\mathbf{k}\psi_\mathbf{k}^\dagger\mid\alpha' \rangle(1-\zeta \mathrm{e}^{-\beta\xi_\mathbf{k}})\mathrm{e}^{-\beta E_{\alpha'}}.
\rho_0(\mathbf{k},\omega) = \frac{1}{\mathcal{Z}}\,2\pi\delta(\xi_\mathbf{k} - \omega) \sum_{\alpha'}\langle\alpha' \mid\psi_\mathbf{k}\psi_\mathbf{k}^\dagger\mid\alpha' \rangle(1-\zeta e^{-\beta\xi_\mathbf{k}})e^{-\beta E_{\alpha'}}.
</math>
</math>


From the commutation relations,
From the commutation relations,
:<math>
<math display="block">
\langle\alpha' \mid \psi_\mathbf{k}\psi_\mathbf{k}^\dagger\mid\alpha' \rangle =
\langle\alpha' \mid \psi_\mathbf{k}\psi_\mathbf{k}^\dagger\mid\alpha' \rangle =
\langle\alpha' \mid(1+\zeta\psi_\mathbf{k}^\dagger\psi_\mathbf{k})\mid\alpha' \rangle,
\langle\alpha' \mid(1+\zeta\psi_\mathbf{k}^\dagger\psi_\mathbf{k})\mid\alpha' \rangle,
</math>
</math>
with possible factors of the volume again. The sum, which involves the thermal average of the number operator, then gives simply <math>[1 + \zeta n(\xi_\mathbf{k})]\mathcal{Z}</math>, leaving
with possible factors of the volume again. The sum, which involves the thermal average of the number operator, then gives simply <math>[1 + \zeta n(\xi_\mathbf{k})]\mathcal{Z}</math>, leaving
<math display="block">\rho_0(\mathbf{k},\omega) = 2\pi\delta(\xi_\mathbf{k} - \omega).</math>
:<math>
\rho_0(\mathbf{k},\omega) = 2\pi\delta(\xi_\mathbf{k} - \omega).
</math>


The imaginary-time propagator is thus
The imaginary-time propagator is thus
<math display="block">\mathcal{G}_0(\mathbf{k},\omega) = \frac{1}{-i\omega_n + \xi_\mathbf{k}}</math>
:<math>
\mathcal{G}_0(\mathbf{k},\omega) = \frac{1}{-\mathrm{i}\omega_n + \xi_\mathbf{k}}
</math>
and the retarded propagator is
and the retarded propagator is
<math display="block">G_0^{\mathrm{R}}(\mathbf{k},\omega) = \frac{1}{-(\omega+i \eta) + \xi_\mathbf{k}}.</math>
:<math>
G_0^{\mathrm{R}}(\mathbf{k},\omega) = \frac{1}{-(\omega+\mathrm{i} \eta) + \xi_\mathbf{k}}.
</math>


====Zero-temperature limit====
====Zero-temperature limit====


As {{mvar|β}}→∞, the spectral density becomes
As {{math|''β'' → ∞}}, the spectral density becomes
:<math>
<math display="block">
\rho(\mathbf{k},\omega) = 2\pi\sum_{\alpha} \left[ \delta(E_\alpha - E_0 - \omega)
\rho(\mathbf{k},\omega) = 2\pi\sum_{\alpha} \left[ \delta(E_\alpha - E_0 - \omega)
|\langle\alpha \mid \psi_\mathbf{k}^\dagger\mid 0 \rangle|^2
\left|\left\langle \alpha \mid \psi_\mathbf{k}^\dagger \mid 0 \right\rangle\right|^2
- \zeta \delta(E_0 - E_{\alpha} - \omega)
- \zeta \delta(E_0 - E_{\alpha} - \omega)
|\langle0 \mid \psi_\mathbf{k}^\dagger\mid\alpha \rangle|^2\right]
\left|\left\langle 0 \mid \psi_\mathbf{k}^\dagger \mid \alpha \right\rangle\right|^2\right]
</math>
</math>
where {{mvar}} = 0 corresponds to the ground state. Note that only the first (second) term contributes when {{mvar|ω}} is positive (negative).
where {{math|1=''α'' = 0}} corresponds to the ground state. Note that only the first (second) term contributes when {{mvar|ω}} is positive (negative).


==General case==
==General case==
Line 275: Line 203:


We can use 'field operators' as above, or creation and annihilation operators associated with other single-particle states, perhaps eigenstates of the (noninteracting) kinetic energy. We then use
We can use 'field operators' as above, or creation and annihilation operators associated with other single-particle states, perhaps eigenstates of the (noninteracting) kinetic energy. We then use
<math display="block">\psi(\mathbf{x} ,\tau) = \varphi_\alpha(\mathbf{x} ) \psi_\alpha(\tau),</math>
:<math>
\psi(\mathbf{x} ,\tau) = \varphi_\alpha(\mathbf{x} ) \psi_\alpha(\tau),
</math>
where <math>\psi_\alpha</math> is the annihilation operator for the single-particle state <math>\alpha</math> and <math>\varphi_\alpha(\mathbf{x} )</math> is that state's wavefunction in the position basis. This gives
where <math>\psi_\alpha</math> is the annihilation operator for the single-particle state <math>\alpha</math> and <math>\varphi_\alpha(\mathbf{x} )</math> is that state's wavefunction in the position basis. This gives
:<math>
<math display="block">
\mathcal{G}^{(n)}_{\alpha_1\ldots\alpha_n|\beta_1\ldots\beta_n}(\tau_1 \ldots \tau_n | \tau_1' \ldots \tau_n')
\mathcal{G}^{(n)}_{\alpha_1\ldots\alpha_n|\beta_1\ldots\beta_n}(\tau_1 \ldots \tau_n | \tau_1' \ldots \tau_n')
= \langle T\psi_{\alpha_1}(\tau_1)\ldots\psi_{\alpha_n}(\tau_n)\bar\psi_{\beta_n}(\tau_n')\ldots\bar\psi_{\beta_1}(\tau_1')\rangle
= \langle T\psi_{\alpha_1}(\tau_1)\ldots\psi_{\alpha_n}(\tau_n)\bar\psi_{\beta_n}(\tau_n')\ldots\bar\psi_{\beta_1}(\tau_1')\rangle
Line 288: Line 214:


These depend only on the difference of their time arguments, so that
These depend only on the difference of their time arguments, so that
:<math>
<math display="block">
\mathcal{G}_{\alpha\beta}(\tau\mid \tau') = \frac{1}{\beta}\sum_{\omega_n}
\mathcal{G}_{\alpha\beta}(\tau\mid \tau')
= \frac{1}{\beta}\sum_{\omega_n}
\mathcal{G}_{\alpha\beta}(\omega_n)\,\mathrm{e}^{-\mathrm{i}\omega_n (\tau-\tau')}
\mathcal{G}_{\alpha\beta}(\omega_n)\,e^{-i\omega_n (\tau-\tau')}
</math>
</math>
and
and
:<math>
<math display="block">
G_{\alpha\beta}(t\mid t') = \int_{-\infty}^{\infty}\frac{\mathrm{d}\omega}{2\pi}\,
G_{\alpha\beta}(t\mid t')
= \int_{-\infty}^{\infty}\frac{d\omega}{2\pi}\,
G_{\alpha\beta}(\omega)\,\mathrm{e}^{-\mathrm{i}\omega(t-t')}.
G_{\alpha\beta}(\omega)\,e^{-i\omega(t-t')}.
</math>
</math>


Line 301: Line 229:


The same periodicity properties as described in above apply to <math>\mathcal{G}_{\alpha\beta}</math>. Specifically,
The same periodicity properties as described in above apply to <math>\mathcal{G}_{\alpha\beta}</math>. Specifically,
<math display="block">\mathcal{G}_{\alpha\beta}(\tau\mid\tau') = \mathcal{G}_{\alpha\beta}(\tau-\tau')</math>
:<math>
\mathcal{G}_{\alpha\beta}(\tau\mid\tau') = \mathcal{G}_{\alpha\beta}(\tau-\tau')
</math>
and
and
<math display="block">\mathcal{G}_{\alpha\beta}(\tau) = \mathcal{G}_{\alpha\beta}(\tau + \beta),</math>
:<math>
\mathcal{G}_{\alpha\beta}(\tau) = \mathcal{G}_{\alpha\beta}(\tau + \beta),
</math>
for <math>\tau < 0</math>.
for <math>\tau < 0</math>.


Line 313: Line 237:


In this case,
In this case,
:<math>
<math display="block">
\rho_{\alpha\beta}(\omega) = \frac{1}{\mathcal{Z}}\sum_{m,n} 2\pi \delta(E_n-E_m-\omega)\;
\rho_{\alpha\beta}(\omega) = \frac{1}{\mathcal{Z}}\sum_{m,n} 2\pi \delta(E_n-E_m-\omega)\;
\langle m \mid \psi_\alpha\mid n \rangle\langle n \mid \psi_\beta^\dagger\mid m \rangle
\langle m \mid \psi_\alpha\mid n \rangle\langle n \mid \psi_\beta^\dagger\mid m \rangle
\left(\mathrm{e}^{-\beta E_m} - \zeta \mathrm{e}^{-\beta E_n}\right) ,
\left(e^{-\beta E_m} - \zeta e^{-\beta E_n}\right) ,
</math>
</math>
where <math>m</math> and <math>n</math> are many-body states.
where <math>m</math> and <math>n</math> are many-body states.


The expressions for the Green functions are modified in the obvious ways:
The expressions for the Green functions are modified in the obvious ways:
<math display="block"> \mathcal{G}_{\alpha\beta}(\omega_n) = \int_{-\infty}^{\infty} \frac{d\omega'}{2\pi} \frac{\rho_{\alpha\beta}(\omega')}{-i\omega_n+\omega'}</math>
:<math>
\mathcal{G}_{\alpha\beta}(\omega_n) = \int_{-\infty}^{\infty} \frac{\mathrm{d}\omega'}{2\pi}
\frac{\rho_{\alpha\beta}(\omega')}{-\mathrm{i}\omega_n+\omega'}
</math>
and
and
<math display="block">G^{\mathrm{R}}_{\alpha\beta}(\omega) = \int_{-\infty}^{\infty} \frac{d\omega'}{2\pi} \frac{\rho_{\alpha\beta}(\omega')}{-(\omega+i\eta)+\omega'}.</math>
:<math>
G^{\mathrm{R}}_{\alpha\beta}(\omega) = \int_{-\infty}^{\infty} \frac{\mathrm{d}\omega'}{2\pi}
\frac{\rho_{\alpha\beta}(\omega')}{-(\omega+\mathrm{i}\eta)+\omega'}.
</math>


Their analyticity properties are identical. The proof follows exactly the same steps, except that the two matrix elements are no longer complex conjugates.
Their analyticity properties are identical to those of <math>\mathcal{G}(\mathbf{k},\omega_n)</math> and <math>G^{\mathrm{R}}(\mathbf{k},\omega)</math> defined in the translationally invariant case. The proof follows exactly the same steps, except that the two matrix elements are no longer complex conjugates.


====Noninteracting case====
====Noninteracting case====


If the particular single-particle states that are chosen are `single-particle energy eigenstates', i.e.
If the particular single-particle states that are chosen are 'single-particle energy eigenstates', i.e.
<math display="block">[H-\mu N,\psi_\alpha^\dagger] = \xi_\alpha\psi_\alpha^\dagger,</math>
:<math>
[H-\mu N,\psi_\alpha^\dagger] = \xi_\alpha\psi_\alpha^\dagger,
</math>
then for <math>|n \rangle</math> an eigenstate:
then for <math>|n \rangle</math> an eigenstate:
<math display="block">(H-\mu N)\mid n \rangle = E_n \mid n \rangle,</math>
:<math>
(H-\mu N)\mid n \rangle = E_n \mid n \rangle,
</math>
so is <math>\psi_\alpha \mid n \rangle</math>:
so is <math>\psi_\alpha \mid n \rangle</math>:
<math display="block">(H-\mu N)\psi_\alpha\mid n \rangle = (E_n - \xi_\alpha) \psi_\alpha \mid n \rangle,</math>
:<math>
(H-\mu N)\psi_\alpha\mid n \rangle = (E_n - \xi_\alpha) \psi_\alpha \mid n \rangle,
</math>
and so is <math>\psi_\alpha^\dagger\mid n \rangle</math>:
and so is <math>\psi_\alpha^\dagger\mid n \rangle</math>:
<math display="block">(H-\mu N)\psi_\alpha^\dagger \mid n \rangle = (E_n + \xi_\alpha) \psi_\alpha^\dagger \mid n \rangle.</math>
:<math>
(H-\mu N)\psi_\alpha^\dagger \mid n \rangle = (E_n + \xi_\alpha) \psi_\alpha^\dagger \mid n \rangle.
</math>


We therefore have
We therefore have
<math display="block">\langle m \mid \psi_\alpha\mid n \rangle\langle n \mid \psi_\beta^\dagger\mid m \rangle =\delta_{\xi_\alpha, \xi_\beta} \delta_{E_n, E_m + \xi_\alpha} \langle m \mid \psi_\alpha\mid n \rangle\langle n \mid \psi_\beta^\dagger \mid m \rangle.</math>
:<math>
\langle m \mid \psi_\alpha\mid n \rangle\langle n \mid \psi_\beta^\dagger\mid m \rangle =
\delta_{\xi_\alpha,\xi_\beta}\delta_{E_n,E_m+\xi_\alpha}\langle m \mid \psi_\alpha\mid n \rangle\langle n \mid \psi_\beta^\dagger\mid m \rangle.
</math>


We then rewrite
We then rewrite
:<math>
<math display="block">
\rho_{\alpha\beta}(\omega) = \frac{1}{\mathcal{Z}}\sum_{m,n} 2\pi \delta(\xi_\alpha-\omega)
\rho_{\alpha\beta}(\omega) = \frac{1}{\mathcal{Z}}\sum_{m,n} 2\pi \delta(\xi_\alpha-\omega)
\delta_{\xi_\alpha,\xi_\beta}\langle m \mid \psi_\alpha\mid n \rangle\langle n \mid \psi_\beta^\dagger \mid m \rangle
\delta_{\xi_\alpha,\xi_\beta}\langle m \mid \psi_\alpha\mid n \rangle\langle n \mid \psi_\beta^\dagger \mid m \rangle
\mathrm{e}^{-\beta E_m}\left(1 - \zeta \mathrm{e}^{-\beta \xi_\alpha}\right),
e^{-\beta E_m} \left(1 - \zeta e^{-\beta \xi_\alpha}\right),
</math>
</math>
therefore
therefore
:<math>
<math display="block">
\rho_{\alpha\beta}(\omega) = \frac{1}{\mathcal{Z}}\sum_m 2\pi \delta(\xi_\alpha-\omega)
\rho_{\alpha\beta}(\omega) = \frac{1}{\mathcal{Z}}\sum_m 2\pi \delta(\xi_\alpha-\omega)
\delta_{\xi_\alpha,\xi_\beta}\langle m \mid \psi_\alpha\psi_\beta^\dagger\mathrm{e}^{-\beta (H-\mu N)}\mid m \rangle
\delta_{\xi_\alpha,\xi_\beta}\langle m \mid \psi_\alpha\psi_\beta^\dagger e^{-\beta (H-\mu N)}\mid m \rangle
\left(1 - \zeta \mathrm{e}^{-\beta \xi_\alpha}\right),
\left(1 - \zeta e^{-\beta \xi_\alpha}\right),
</math>
</math>
use
use
<math display="block">\langle m \mid \psi_\alpha \psi_\beta^\dagger\mid m \rangle = \delta_{\alpha,\beta}\langle m \mid \zeta \psi_\alpha^\dagger \psi_\alpha + 1 \mid m \rangle</math>
:<math>
\langle m \mid \psi_\alpha \psi_\beta^\dagger\mid m \rangle = \delta_{\alpha,\beta}\langle m \mid \zeta \psi_\alpha^\dagger \psi_\alpha + 1 \mid m \rangle
</math>
and the fact that the thermal average of the number operator gives the Bose–Einstein or Fermi–Dirac distribution function.
and the fact that the thermal average of the number operator gives the Bose–Einstein or Fermi–Dirac distribution function.


Finally, the spectral density simplifies to give
Finally, the spectral density simplifies to give
<math display="block">\rho_{\alpha\beta} = 2\pi \delta(\xi_\alpha - \omega)\delta_{\alpha\beta},</math>
:<math>
\rho_{\alpha\beta} = 2\pi \delta(\xi_\alpha - \omega)\delta_{\alpha\beta},
</math>
so that the thermal Green function is
so that the thermal Green function is
<math display="block">\mathcal{G}_{\alpha\beta}(\omega_n) = \frac{\delta_{\alpha\beta}}{-i\omega_n + \xi_\beta}</math>
:<math>
\mathcal{G}_{\alpha\beta}(\omega_n) = \frac{\delta_{\alpha\beta}}{-\mathrm{i}\omega_n + \xi_\beta}
</math>
and the retarded Green function is
and the retarded Green function is
<math display="block">G_{\alpha\beta}(\omega) = \frac{\delta_{\alpha\beta}}{-(\omega+i\eta) + \xi_\beta}.</math>
:<math>
G_{\alpha\beta}(\omega) = \frac{\delta_{\alpha\beta}}{-(\omega+\mathrm{i}\eta) + \xi_\beta}.
</math>
Note that the noninteracting Green function is diagonal, but this will not be true in the interacting case.
Note that the noninteracting Green function is diagonal, but this will not be true in the interacting case.


Line 397: Line 296:
*[[Propagator]]
*[[Propagator]]
*[[Correlation function (quantum field theory)]]
*[[Correlation function (quantum field theory)]]
*[[Numerical analytic continuation]]


== References ==
== References ==
Line 412: Line 312:


==External links==
==External links==
* [http://www.cond-mat.de/events/correl14/manuscripts/pavarini.pdf Linear Response Functions] in Eva Pavarini, Erik Koch, Dieter Vollhardt, and Alexander Lichtenstein (eds.): DMFT at 25: Infinite Dimensions, Verlag des Forschungszentrum Jülich, 2014 {{ISBN|978-3-89336-953-9}}
* [https://web.archive.org/web/20160304030119/http://www.cond-mat.de/events/correl14/manuscripts/pavarini.pdf Linear Response Functions] in Eva Pavarini, Erik Koch, Dieter Vollhardt, and Alexander Lichtenstein (eds.): DMFT at 25: Infinite Dimensions, Verlag des Forschungszentrum Jülich, 2014 {{ISBN|978-3-89336-953-9}}


[[Category:Quantum field theory]]
[[Category:Quantum field theory]]

Latest revision as of 03:22, 15 October 2024

In many-body theory, the term Green's function (or Green function) is sometimes used interchangeably with correlation function, but refers specifically to correlators of field operators or creation and annihilation operators.

The name comes from the Green's functions used to solve inhomogeneous differential equations, to which they are loosely related. (Specifically, only two-point "Green's functions" in the case of a non-interacting system are Green's functions in the mathematical sense; the linear operator that they invert is the Hamiltonian operator, which in the non-interacting case is quadratic in the fields.)

Spatially uniform case

[edit]

Basic definitions

[edit]

We consider a many-body theory with field operator (annihilation operator written in the position basis) .

The Heisenberg operators can be written in terms of Schrödinger operators as and the creation operator is , where is the grand-canonical Hamiltonian.

Similarly, for the imaginary-time operators, [Note that the imaginary-time creation operator is not the Hermitian conjugate of the annihilation operator .]

In real time, the -point Green function is defined by where we have used a condensed notation in which signifies and signifies . The operator denotes time ordering, and indicates that the field operators that follow it are to be ordered so that their time arguments increase from right to left.

In imaginary time, the corresponding definition is where signifies . (The imaginary-time variables are restricted to the range from to the inverse temperature .)

Note regarding signs and normalization used in these definitions: The signs of the Green functions have been chosen so that Fourier transform of the two-point () thermal Green function for a free particle is and the retarded Green function is where is the Matsubara frequency.

Throughout, is for bosons and for fermions and denotes either a commutator or anticommutator as appropriate.

(See below for details.)

Two-point functions

[edit]

The Green function with a single pair of arguments () is referred to as the two-point function, or propagator. In the presence of both spatial and temporal translational symmetry, it depends only on the difference of its arguments. Taking the Fourier transform with respect to both space and time gives where the sum is over the appropriate Matsubara frequencies (and the integral involves an implicit factor of , as usual).

In real time, we will explicitly indicate the time-ordered function with a superscript T:

The real-time two-point Green function can be written in terms of 'retarded' and 'advanced' Green functions, which will turn out to have simpler analyticity properties. The retarded and advanced Green functions are defined by and respectively.

They are related to the time-ordered Green function by where is the Bose–Einstein or Fermi–Dirac distribution function.

Imaginary-time ordering and β-periodicity

[edit]

The thermal Green functions are defined only when both imaginary-time arguments are within the range to . The two-point Green function has the following properties. (The position or momentum arguments are suppressed in this section.)

Firstly, it depends only on the difference of the imaginary times: The argument is allowed to run from to .

Secondly, is (anti)periodic under shifts of . Because of the small domain within which the function is defined, this means just for . Time ordering is crucial for this property, which can be proved straightforwardly, using the cyclicity of the trace operation.

These two properties allow for the Fourier transform representation and its inverse,

Finally, note that has a discontinuity at ; this is consistent with a long-distance behaviour of .

Spectral representation

[edit]

The propagators in real and imaginary time can both be related to the spectral density (or spectral weight), given by where |α refers to a (many-body) eigenstate of the grand-canonical Hamiltonian HμN, with eigenvalue Eα.

The imaginary-time propagator is then given by and the retarded propagator by where the limit as is implied.

The advanced propagator is given by the same expression, but with in the denominator.

The time-ordered function can be found in terms of and . As claimed above, and have simple analyticity properties: the former (latter) has all its poles and discontinuities in the lower (upper) half-plane.

The thermal propagator has all its poles and discontinuities on the imaginary axis.

The spectral density can be found very straightforwardly from , using the Sokhatsky–Weierstrass theorem where P denotes the Cauchy principal part. This gives

This furthermore implies that obeys the following relationship between its real and imaginary parts: where denotes the principal value of the integral.

The spectral density obeys a sum rule, which gives as .

Hilbert transform

[edit]

The similarity of the spectral representations of the imaginary- and real-time Green functions allows us to define the function which is related to and by and A similar expression obviously holds for .

The relation between and is referred to as a Hilbert transform.

Proof of spectral representation

[edit]

We demonstrate the proof of the spectral representation of the propagator in the case of the thermal Green function, defined as

Due to translational symmetry, it is only necessary to consider for , given by Inserting a complete set of eigenstates gives

Since and are eigenstates of , the Heisenberg operators can be rewritten in terms of Schrödinger operators, giving Performing the Fourier transform then gives

Momentum conservation allows the final term to be written as (up to possible factors of the volume) which confirms the expressions for the Green functions in the spectral representation.

The sum rule can be proved by considering the expectation value of the commutator, and then inserting a complete set of eigenstates into both terms of the commutator:

Swapping the labels in the first term then gives which is exactly the result of the integration of ρ.

Non-interacting case

[edit]

In the non-interacting case, is an eigenstate with (grand-canonical) energy , where is the single-particle dispersion relation measured with respect to the chemical potential. The spectral density therefore becomes

From the commutation relations, with possible factors of the volume again. The sum, which involves the thermal average of the number operator, then gives simply , leaving

The imaginary-time propagator is thus and the retarded propagator is

Zero-temperature limit

[edit]

As β → ∞, the spectral density becomes where α = 0 corresponds to the ground state. Note that only the first (second) term contributes when ω is positive (negative).

General case

[edit]

Basic definitions

[edit]

We can use 'field operators' as above, or creation and annihilation operators associated with other single-particle states, perhaps eigenstates of the (noninteracting) kinetic energy. We then use where is the annihilation operator for the single-particle state and is that state's wavefunction in the position basis. This gives with a similar expression for .

Two-point functions

[edit]

These depend only on the difference of their time arguments, so that and

We can again define retarded and advanced functions in the obvious way; these are related to the time-ordered function in the same way as above.

The same periodicity properties as described in above apply to . Specifically, and for .

Spectral representation

[edit]

In this case, where and are many-body states.

The expressions for the Green functions are modified in the obvious ways: and

Their analyticity properties are identical to those of and defined in the translationally invariant case. The proof follows exactly the same steps, except that the two matrix elements are no longer complex conjugates.

Noninteracting case

[edit]

If the particular single-particle states that are chosen are 'single-particle energy eigenstates', i.e. then for an eigenstate: so is : and so is :

We therefore have

We then rewrite therefore use and the fact that the thermal average of the number operator gives the Bose–Einstein or Fermi–Dirac distribution function.

Finally, the spectral density simplifies to give so that the thermal Green function is and the retarded Green function is Note that the noninteracting Green function is diagonal, but this will not be true in the interacting case.

See also

[edit]

References

[edit]

Books

[edit]
  • Bonch-Bruevich V. L., Tyablikov S. V. (1962): The Green Function Method in Statistical Mechanics. North Holland Publishing Co.
  • Abrikosov, A. A., Gorkov, L. P. and Dzyaloshinski, I. E. (1963): Methods of Quantum Field Theory in Statistical Physics Englewood Cliffs: Prentice-Hall.
  • Negele, J. W. and Orland, H. (1988): Quantum Many-Particle Systems AddisonWesley.
  • Zubarev D. N., Morozov V., Ropke G. (1996): Statistical Mechanics of Nonequilibrium Processes: Basic Concepts, Kinetic Theory (Vol. 1). John Wiley & Sons. ISBN 3-05-501708-0.
  • Mattuck Richard D. (1992), A Guide to Feynman Diagrams in the Many-Body Problem, Dover Publications, ISBN 0-486-67047-3.

Papers

[edit]
[edit]