Greater Green's function

In the complex realm of many-body quantum physics, understanding the intricate dance of countless interacting particles requires a specialized language. Simple questions about a single particle's position and momentum give way to more profound inquiries about correlations and responses across space and time. This is the challenge addressed by Green's functions, a powerful mathematical toolkit designed to navigate the dynamics of the quantum world. While fundamental, these concepts can often seem abstract, creating a knowledge gap between formal theory and practical application, particularly when systems are pushed away from simple equilibrium.

This article demystifies one of the key players in this framework: the Greater Green's function. It provides a conceptual guide to its meaning and use. The first chapter, ​​Principles and Mechanisms​​, will introduce the Greater Green's function alongside its counterparts—Lesser, Retarded, and Advanced—and explain what they reveal about particle populations, quantum coherence, and the fundamental laws governing both equilibrium and non-equilibrium states. The second chapter, ​​Applications and Interdisciplinary Connections​​, will then illustrate how these theoretical tools are applied to real-world problems, from interpreting spectroscopy experiments to calculating electrical and heat currents in nanoscale devices, showcasing the unifying power of the Green's function formalism.

In the bustling, jostling world of many-particle quantum systems, asking "Where is this particle right now?" is often the wrong question. A better, more fruitful question is, "If I do something here, what is the chance I'll see an effect there a little while later?" Quantum reality, especially in a crowd, is a story of connections, of correlations. The mathematical tools designed to tell this story, to track these intricate relationships across space and time, are known as ​​Green's functions​​. They are the grand chroniclers of the many-body world.

Imagine you can perform two acts in the quantum realm: you can inject a particle into a system, or you can detect and remove one. The Green's functions are essentially the quantum mechanical amplitudes for sequences of these two events. Let's meet the two most fundamental members of this family.

First, there is the ​​Greater Green's function, $G^>$​​. Suppose we create a particle at position $x'$ at time $t'$, and then at a later time $t$, we check for a particle at position $x$. The amplitude for this sequence of events is captured by $G^>(x,t; x',t')$. Mathematically, it looks like this:

$G^>(x,t; x',t') = -i \langle \hat{\psi}(x,t) \hat{\psi}^\dagger(x',t') \rangle$

Here, $\hat{\psi}^\dagger(x',t')$ is the operator that creates a particle, and $\hat{\psi}(x,t)$ is the operator that annihilates (or detects) one. The angle brackets $\langle \dots \rangle$ signify an average over all the possibilities the system allows. You can think of $G^>$ as describing the propagation of an added particle. It tells us about the available pathways, the empty states or "holes" into which this new particle can travel.

Its counterpart is the ​​Lesser Green's function, $G^<$​​. This function describes a different story. It answers: what is the amplitude that we detect a particle at $x$ at time $t$, given that it was part of a system where a particle was removed at $x'$ at time $t'$? Its form is:

$G^<(x,t; x',t') = i \langle \hat{\psi}^\dagger(x',t') \hat{\psi}(x,t) \rangle$

Notice the order of operators is flipped. $G^<$ is not about an extra particle we've added; it's about the particles that were already there. It tracks the dynamics of the system's existing occupants, telling us about the filled states.

These functions might seem abstract, but they become wonderfully concrete when we bring the two time points together, setting $t = t'$. What do they tell us now?

Let's look at the lesser function first. The quantity $-iG^<(x,t; x,t)$ turns out to be precisely the average number of particles at position $x$ at time $t$—the particle density. It's a quantum census taker! If you sum this quantity over all space, you get the total number of particles in the system: $N(t) = -i \int dx \, G^<(x,t; x,t)$. Even more, the off-diagonal elements, $-iG^<(x,t; x',t)$ for $x \neq x'$, measure the quantum ​​coherence​​ between different points, the delicate phase relationships that are the hallmark of quantum mechanics.

So, if $G^<$ counts the occupied states, what does $G^>$ do at equal times? As you might guess, $-iG^>(x,t; x,t)$ counts the unoccupied states, or the "holes". For fermions, like electrons, which obey the Pauli exclusion principle (only one particle per state), a state is either filled or empty. This leads to a beautiful and profound relationship derived directly from their fundamental anticommutation rules: the number of particles plus the number of holes must equal the total number of available states. This simple idea is encoded in a universal identity:

$G^>(t,t) - G^<(t,t) = -i\mathbf{I}$

where $\mathbf{I}$ is the identity matrix. The density of "what is" plus the density of "what could be" is a constant.

So far, we've discussed correlation functions that describe the system as it is. But physics is also about change. How does a system respond to a push or a pull? This question brings causality into the picture—the effect cannot come before the cause.

Neither $G^>$ nor $G^<$ on their own respect this principle; they are correlation functions, not response functions, and can be non-zero whether $t$ is before or after $t'$. To build a causal response, we must combine them. Enter the ​​Retarded Green's function, $G^r$​​:

$G^r(t,t') = -i\theta(t-t') \langle [\hat{A}(t), \hat{A}(t')] \rangle \propto \theta(t-t') (G^>(t,t') - G^<(t,t'))$

The crucial ingredient here is the Heaviside step function, $\theta(t-t')$, which is zero for $t < t'$ and one for $t > t'$. It enforces the arrow of time. The response is zero until after the perturbation. And what determines the response? It's the difference between the greater and lesser functions, which is related to the quantum mechanical commutator $[\hat{A}(t), \hat{A}(t')]$. This difference tells us the net "room" available for a quantum excitation to propagate—the balance between creating a particle in an empty state and destroying a particle in a filled one.

Along with its time-reversed twin, the ​​Advanced Green's function $G^a$​​, these four functions—greater, lesser, retarded, and advanced—form a complete toolkit. They are all interconnected by a deep and simple identity that holds in any system, whether in placid equilibrium or a violent non-equilibrium state: $G^r - G^a = G^> - G^<$. This elegant equation shows how the very structure of quantum field theory unifies the concepts of correlation ($G^>, G^<$) and causal response ($G^r, G^a$).

When a system is left alone long enough, it settles into thermal equilibrium. In this state of peaceful balance, profound new relationships emerge. The key insight is that any process that involves the system losing some energy $\hbar\omega$ (described by $G^>(\omega)$ in frequency space) must be balanced by a process where the system gains that same energy (described by $G^<(\omega)$).

This principle of detailed balance, when applied to quantum systems, is known as the ​​Kubo-Martin-Schwinger (KMS) condition​​. It states that the Greater and Lesser Green's functions are no longer independent, but are related by a simple factor determined by the temperature:

$G^>(\omega) = \exp(\beta \hbar \omega) G^<(\omega)$

where $\beta = 1/(k_B T)$ is the inverse temperature. This exponential factor is the universe's way of saying that it's much easier to find the energy to create low-energy excitations than high-energy ones at a finite temperature.

This single relation is the key that unlocks one of the most powerful results in all of physics: the ​​Fluctuation-Dissipation Theorem​​. "Fluctuations" are the spontaneous jitters and jives of a system in equilibrium, captured by the sum $G^> + G^<$. "Dissipation" describes how the system loses energy when perturbed, and we saw this is related to the difference $G^> - G^<$. The KMS condition provides an unbreakable link between them. It tells us that if we know how a system dissipates energy (something we can often measure by probing its response), we can precisely calculate the spectrum of its internal quantum and thermal fluctuations.

A beautiful example is a simple quantized LC electrical circuit, which behaves like a quantum harmonic oscillator. Its Green's function shows that fluctuations only happen at its resonant frequency, $\omega_0$. The part describing energy emission ($G^>$) is proportional to $1+n_B(\omega_0)$, accounting for both spontaneous emission (always possible) and stimulated emission (proportional to the number of existing excitations, $n_B$). The part for energy absorption ($G^<$) is proportional to just $n_B(\omega_0)$, as you can only absorb an excitation if one is there to begin with. The ratio is exactly $\exp(\beta\hbar\omega_0)$, a perfect illustration of the KMS condition at work.

The real world is rarely in perfect equilibrium. What happens when we drive a system, for example, by connecting a single molecule to the terminals of a battery? The simple KMS relation breaks down. $G^>$ and $G^<$ become unlinked, carrying independent information about two competing processes: the injection of electrons from the high-voltage source and their extraction by the low-voltage drain.

Yet, our Green's functions do not fail us. The state of the molecule can be described by the lesser function $G^<$, which tells us which molecular orbitals are occupied. This, in turn, is determined by the famous ​​Keldysh equation​​:

$G^<(\omega) = G^r(\omega) \Sigma^<(\omega) G^a(\omega)$

This equation has a wonderfully intuitive picture. The self-energy, $\Sigma^<(\omega)$, acts as a "source term", describing the rate at which electrons with energy $\hbar\omega$ are injected into the molecule from the electrical leads. This injected current then propagates through the molecule, a process governed by the retarded and advanced functions $G^r$ and $G^a$. This equation is the workhorse of modern nanoelectronics, allowing us to compute the flow of current through single-molecule devices.

Non-equilibrium doesn't only mean steady currents. Consider a ​​quantum quench​​: we take a system in its ground state and suddenly change its governing laws (its Hamiltonian). The system is now in a highly excited, time-evolving state. How does its Green's function look? It beautifully encodes the system's "memory". The function $G^>(\mathbf{k}, t, t')$ for a mode $\mathbf{k}$ will contain a prefactor $(1-n_\mathbf{k})$, where $n_\mathbf{k}$ is the occupation number of that mode before the quench. But its time-evolution part, $e^{-i\epsilon_{\mathbf{k}}(t-t')/\hbar}$, will be governed by the energy $\epsilon_\mathbf{k}$ of the Hamiltonian after the quench. The Green's function remembers where it came from but evolves according to its new reality.

Finally, what do these functions actually look like? For a simple one-dimensional gas of non-interacting electrons at zero temperature, the lesser Green's function at equal times has a strikingly simple and elegant form as a function of distance $r = x-x'$:

$G^<(r) \propto i \frac{\sin(k_F r)}{\pi r}$

This gentle, decaying ripple is a direct consequence of the sharp cut-off at the Fermi momentum, $k_F$, in the distribution of electrons. It's a real-space "photograph" of the Fermi sea, a phenomenon known as Friedel oscillations, perfectly captured by the Green's function. It is a testament to the power and beauty of these mathematical objects, which allow us to listen in on the intricate, correlated symphony of the quantum many-body world.

In the previous chapter, we took a deep dive into the formal machinery of the Keldysh formalism, introducing the cast of characters: the retarded, advanced, lesser, and greater Green's functions. You might be feeling a bit like a student who has just learned all the rules of chess—the moves, the captures, the special conditions—but has yet to see a real game. What is all this intricate framework for? What problems can it solve? What new physics does it reveal?

Now, we get to play the game. This chapter is a journey through the applications of this powerful formalism. We will see that these Green's functions are not just abstract mathematical objects; they are our most direct windows into the bustling, dynamic world of quantum particles. They are the tools we use to understand what happens when we disturb a system from its quiet equilibrium—by shining light on it, applying a voltage across it, or heating one end of it. We will discover that this language is remarkably universal, describing phenomena from the flow of electricity in a nanoscale transistor to the flow of heat in an insulating crystal.

Perhaps the most direct and satisfying application of our new tools is in understanding how we see quantum states. How can we experimentally verify the existence of the particle and hole states described by the lesser ($G^<$) and greater ($G^>$) Green's functions? The answer lies in the powerful techniques of modern spectroscopy.

Imagine a crystal, a vast city of electrons occupying various energy levels. We want to create a map of this city—which houses are occupied, and which are vacant? One way is to go door-to-door and knock an electron out. This is the essence of ​​Angle-Resolved Photoemission Spectroscopy (ARPES)​​. In an ARPES experiment, we bombard the material with high-energy photons. When a photon strikes an electron, it can give it enough energy to be ejected from the material entirely. We then catch this ejected electron and measure its energy and momentum. From this, we can deduce the energy and momentum it had inside the crystal.

Now, what is the probability that this process occurs? It depends on two things: first, that there was an electron there to be knocked out in the first place, and second, that an electronic state with that specific energy and momentum is allowed by the laws of quantum mechanics. The first condition—the occupation of the initial state—is precisely what the lesser Green's function, $G^<(\mathbf{k}, \omega)$, describes! In fact, under standard approximations, the measured intensity in an ARPES experiment is directly proportional to the density of occupied states, which for a system in thermal equilibrium is given by $I_{\mathrm{PES}}(\mathbf{k},\omega) \propto f(\omega) A(\mathbf{k},\omega)$, where $A(\mathbf{k},\omega) = i(G^>(\mathbf{k},\omega) - G^<(\mathbf{k},\omega))$ is the total spectral function and $f(\omega)$ is the familiar Fermi-Dirac distribution. In essence, ARPES directly measures the occupied part of the electronic spectrum, providing a stunning experimental visualization of $iG^<(\mathbf{k},\omega)$.

What about the vacant houses? To map those, we can't knock anyone out. Instead, we have to try to put someone in. This is the idea behind ​​Inverse Photoemission Spectroscopy (IPES)​​. Here, we shoot a beam of electrons at the material. If an electron finds an empty state it can drop into, it will do so, emitting a photon in the process. By measuring the energy of this emitted photon, we can work out the energy of the vacant state the electron just filled. The chance of this happening depends on the availability of empty states, which is exactly the information carried by the greater Green's function, $G^>(\mathbf{k}, \omega)$. The measured IPES intensity is proportional to the density of unoccupied states, given by $I_{\mathrm{IPES}}(\mathbf{k},\omega) \propto [1-f(\omega)] A(\mathbf{k},\omega)$.

Taken together, ARPES and IPES give us a complete picture of the single-particle excitations in a material. They are experimental manifestations of the lesser and greater Green's functions, transforming these theoretical concepts into tangible, measurable spectra. A simple but illuminating theoretical playground to understand the building blocks of such spectra is the single-site Hubbard model, where one can explicitly calculate $G^>(\omega)$ and see that it consists of sharp peaks corresponding to the distinct energies required to add an electron to the system, revealing the genesis of a spectrum from first principles.

While spectroscopy is about passive observation, the real power of the Keldysh formalism is unleashed when we actively drive a system out of equilibrium. The quintessential example of this is quantum transport—the study of how electrons flow through nanoscale structures.

Consider the quantum physicist's favorite toy: a ​​quantum dot​​, a tiny island of semiconductor material so small that it can hold just a handful of electrons. Let's place this dot between two large metallic contacts, a "source" and a "drain," and apply a voltage between them. Electrons will now flow from the source, through the dot, and into the drain. This is, in effect, the world's smallest transistor. How do we calculate the electrical current?

This is a classic non-equilibrium problem. The source and drain leads are each in their own thermal equilibrium, but at different chemical potentials, creating a steady flow of particles. The Meir-Wingreen formula, a cornerstone result of non-equilibrium physics, gives us the answer directly in the language of Green's functions. It states that the current flowing into the dot from a lead is a delicate balance between particles entering and particles leaving, expressed as:

There is a beautiful physical intuition here. The term $\mathbf{\Sigma}_\alpha^<(\omega)$ represents the rate at which the lead attempts to inject electrons into the dot, while $\mathbf{G}^>(\omega)$ represents the availability of empty states on the dot to receive them. The second term, $\mathbf{\Sigma}_\alpha^>(\omega)\mathbf{G}^<(\omega)$, represents the reverse process: electrons on the dot (described by $\mathbf{G}^<$) trying to escape into empty states in the lead (described by $\mathbf{\Sigma}_\alpha^>$). The net current is the result of this microscopic tug-of-war, integrated over all energies. Using this formalism, we can derive the famous Landauer formula for conductance, which connects a macroscopic property (current) to the microscopic quantum transmission characteristics of the dot.

We can ask more detailed questions. For instance, under a given voltage, what is the average number of electrons residing on the dot? This, too, can be found by integrating the lesser Green's function, $N = -i \int (d\omega/2\pi) G^<(\omega)$. This leads to a deeper concept: the ​​non-equilibrium distribution function​​. In equilibrium, the occupation of energy levels is given by the universal Fermi-Dirac function. But out of equilibrium, the situation is far more complex. We can define an effective, energy-dependent occupation $f_d(\omega)$ as the ratio of the occupied density of states to the total density of states, $f_d(\omega) = i G^<(\omega) / A(\omega)$. This function is no longer a simple step function; it is a complex landscape shaped by the voltage bias and, crucially, by the electron-electron interactions on the dot. It shows how inelastic scattering processes inside the dot redistribute electrons, creating a unique steady-state population that is a hallmark of the non-equilibrium condition.

The average current is not the whole story. Because electrons are discrete particles, their flow is not perfectly smooth; it is "shotty." There are random fluctuations around the average value, a phenomenon known as ​​shot noise​​. This noise is not just an experimental nuisance; it contains profound information about the nature of charge transport, such as the charge of the carriers (as in the fractional quantum Hall effect) and the correlations between them.

Can our formalism describe these fluctuations? Absolutely. The two-time correlation function of charge on the dot, $\langle \delta\hat{n}(t) \delta\hat{n}(0) \rangle$, which quantifies the noise, can be expressed directly in terms of Green's functions. The noise power spectrum, its Fourier transform, turns out to be a convolution of the lesser and greater functions: $S(\omega) \propto \int dE \, [G^<(E)G^>(E+\omega) + G^>(E)G^<(E-\omega)]$. This is a beautiful result. It tells us that noise arises from the correlated sequence of events of an electron arriving on the dot (related to $G^<$) and an empty state being available for it to leave (related to $G^>$), and vice-versa. The ability to calculate not just averages but also their fluctuations is a major triumph of the non-equilibrium Green's function approach.

So far, our discussion has centered on electrons. But here is where the true beauty and power of the formalism shines through. The story we just told is not just about electrons. It is a universal story about the transport of quantum particles out of equilibrium.

Consider the transport of heat. In many materials, particularly insulators, heat is not carried by electrons but by ​​phonons​​—quantized vibrations of the crystal lattice. Imagine we construct a "phononic" device by connecting two materials held at different temperatures, $T_L$ and $T_R$, via some central scattering region. A heat current will flow from the hotter end to the colder end.

How do we describe this? We simply replace our electron operators with phonon operators and our electron Green's functions with phonon Green's functions. The entire Keldysh machinery translates seamlessly. The expression for the steady-state heat current, known as the Caroli-Landauer formula, looks strikingly similar to the Meir-Wingreen formula for electrical current:

Here, $\mathcal{T}(\omega)$ is the phonon transmission function, and $n_B(\omega, T)$ is the Bose-Einstein distribution function for phonons. The transmission function itself is given by a trace over phonon Green's functions and self-energies, $\mathcal{T}(\omega) = \text{Tr}[\Gamma_L D^R \Gamma_R D^A]$, a perfect analogue of the electronic case. This reveals a deep unity in the physics of transport: the fundamental logic governing the flow of particles driven by a bias (be it chemical potential or temperature) is the same, regardless of whether the particles are fermions like electrons or bosons like phonons.

The formalism doesn't just describe different particle types in isolation; it excels at describing their interactions. For example, in a metal, the flowing electrons can scatter off the lattice vibrations, transferring energy and momentum. This process damps the phonons, limiting their lifetime. We can calculate this phonon damping rate by computing the phonon self-energy, which involves a "bubble" diagram of electron lesser and greater Green's functions. This gives us a quantitative understanding of the friction experienced by the lattice due to the surrounding sea of electrons. The self-energy components themselves acquire a direct physical meaning: for instance, the lesser self-energy $\Sigma^<(\omega)$ is directly proportional to the rate at which particles are scattered into a state with energy $\omega$ due to interactions, providing a vivid link between the formal theory and the system's kinetics.

From spectroscopy to nanoelectronics, from electrical current to heat flow, from charge noise to phonon damping—the Greater Green's function and its Keldysh partners provide a single, coherent, and profoundly beautiful framework. They are the language we speak when we wish to understand the rich, dynamic, and wonderfully complex dance of particles in a world away from the quiet slumber of equilibrium.