The Coherence Function: Understanding the Correlation of Light

While often pictured as perfectly orderly waves, light from most real-world sources is a far more chaotic phenomenon. Yet, from this chaos, predictable patterns like interference fringes and focused images emerge. The key to understanding this behavior lies in the concept of coherence—the measure of correlation within a wave's fluctuations. This article demystifies coherence, addressing the gap between simplified models of light and its true statistical nature. First, in "Principles and Mechanisms," we will delve into the mathematical foundations of coherence, exploring the distinct concepts of temporal and spatial coherence and the two profound theorems that govern them: the Wiener-Khinchin and van Cittert-Zernike theorems. Subsequently, in "Applications and Interdisciplinary Connections," we will witness how these principles are applied across diverse fields, from practical optical engineering and astronomy to the frontiers of quantum physics, revealing coherence as a unifying concept in science.

In our introduction, we touched upon the idea that light isn't always the perfectly orderly, infinitely long train of waves we imagine in high school. Real light, from a star or a lightbulb, is a much wilder, more chaotic affair. And yet, from this chaos, order emerges. The beautiful rainbow colors of an oil slick, the iridescent sheen of a butterfly's wing, the very ability of a telescope to form an image—all depend on a subtle and profound property of light called ​​coherence​​. But what is it, really? How do we describe it? And where does it come from? Let's take a journey into the heart of this concept, and we'll find that it's all about one simple idea: correlation.

Imagine you're watching waves in a pond. If you drop two pebbles in at the same time, the circular ripples expand and interfere, creating a beautiful pattern of high crests and calm troughs. This is interference, and it's easy to understand because the two sources are perfectly synchronized. Light waves do the same thing in the classic Young's double-slit experiment. But what if the "pebbles" were dropped randomly, with no connection to each other? The resulting mess of ripples would have no stable pattern. The sources would be incoherent.

Coherence is the physicist's way of talking about the "relatedness" or correlation between the undulations of a wave at different points in space and time. To make this precise, we define a quantity called the ​​mutual coherence function​​, written as $\Gamma(\mathbf{r}_1, \mathbf{r}_2, \tau)$. This function is a workhorse; it asks a very specific question: "If I measure the electric field of the light wave at point $\mathbf{r}_1$ and time $t$, how much does it tell me about the field at point $\mathbf{r}_2$ at a later time $t+\tau$?" It's a measure of the statistical similarity. If the fields are strongly correlated, $\Gamma$ is large. If they are completely random with respect to each other, $\Gamma$ is zero.

This single function elegantly splits into two main flavors:

1. ​​Temporal Coherence​​: Here, we stay at one point in space ($\mathbf{r}_1 = \mathbf{r}_2$) and ask how the field at time $t$ is related to the field at time $t+\tau$. This is described by $\Gamma(\tau)$. It's a measure of how long a wave "remembers" its own phase.
2. ​​Spatial Coherence​​: Here, we look at a single instant in time ($\tau=0$) and ask how the field at point $\mathbf{r}_1$ is related to the field at point $\mathbf{r}_2$. This is described by $\Gamma(\mathbf{r}_1, \mathbf{r}_2, 0)$, or just $J(\mathbf{r}_1, \mathbf{r}_2)$.

The beauty of this is that interference patterns are a direct visualization of coherence. In a Young's double-slit experiment, the path difference from the two slits to a point on the screen creates a time delay $\tau$. The visibility of the interference fringes—how much contrast there is between the bright and dark bands—is given directly by the magnitude of the normalized temporal coherence function, $|\gamma(\tau)|$. If the light's "memory" of its phase fades quickly, fringes will only be visible for very small path differences. If the light has a long memory (long coherence time), you will see sharp fringes over a large area.

So, what determines this "memory" of a wave? What gives it temporal coherence? The answer is one of the most beautiful and unifying principles in all of physics: the ​​Wiener-Khinchin theorem​​. It states that the temporal coherence function, $\Gamma(\tau)$, and the power spectral density (the spectrum), $S(\omega)$, of the light are a Fourier transform pair.

Think about what this means. The spectrum tells you the mixture of "colors" or frequencies in your light. The coherence function tells you about correlations in time. The theorem says these are two sides of the same coin! You tell me the spectrum, and I can tell you its entire temporal correlation structure, and vice versa.

Let's look at some examples to get a feel for this.

• A hypothetical, perfectly monochromatic laser emits a single frequency $\omega_0$. Its spectrum is a sharp spike, a Dirac delta function. The Fourier transform of a delta function is a pure, unending cosine wave. This means the light is perfectly correlated with itself for all time—it has an infinite coherence time.
• Now consider a more realistic source, like the light from glowing gas in a neon sign. The atoms are constantly being interrupted, so they emit short bursts of waves. The spectrum of such light is often a ​​Lorentzian profile​​, which is a peak with a certain width. The Wiener-Khinchin theorem tells us that the coherence function for this light is an ​​exponentially decaying cosine wave​​,. The wider the spectral peak, the faster the coherence decays. A broad spectrum implies a short memory.
• What if the light source is more complex, containing two distinct atomic transitions? The spectrum will have two Lorentzian peaks at frequencies $\omega_1$ and $\omega_2$. What does the coherence function look like? Exactly what your intuition might guess: it's the sum of two decaying cosine waves, one for each spectral line. This superposition creates "beats" in the coherence function—a rapid oscillation modulated by a slower one. By measuring the coherence, you can deduce the structure of the spectrum.

This Fourier relationship is a profound link between the time domain and the frequency domain. It's the "uncertainty principle" of classical optics: a light pulse that is very short in time must be made of a broad range of frequencies. A light source that is very pure in color (narrow spectrum) must be coherent over a long time.

Now we turn to spatial coherence. Where does it come from? You might think that to get a spatially coherent field, you need a highly ordered source. But here, nature has a wonderful surprise for us, a piece of magic known as the ​​van Cittert-Zernike theorem​​.

The theorem states something astonishing: an extended, perfectly spatially incoherent source, like the hot filament of a lightbulb or the surface of a star, will produce a light field that is partially spatially coherent at a distance. The chaos of the source organizes itself through the simple act of propagation.

And the form of this emerging coherence is truly elegant. The theorem says that the complex degree of spatial coherence, $\mu(x_1, x_2)$, in a distant observation plane is given by the Fourier transform of the source's intensity distribution.

Does that sound familiar? It should! It's the same mathematical relationship that describes Fraunhofer diffraction. This means the spatial coherence pattern from an incoherent source has the exact same shape as the diffraction pattern from an aperture of the same size and shape as the source.

Let's see this in action with some beautiful examples:

• Consider an incoherent linear source, like a long, thin fluorescent bulb. If you were to treat this as an aperture and shine a laser through it, you'd get a ​​sinc function​​ ($\sin(x)/x$) diffraction pattern. The van Cittert-Zernike theorem tells us that the spatial coherence function produced by the bulb itself is also a sinc function.
• Imagine a distant binary star system—two incoherent points of light. The diffraction pattern from two point-like slits is a ​​cosine function​​ (Young's interference fringes). And indeed, the spatial coherence function generated by the binary star is a cosine wave across the sky. By measuring the spacing of these "coherence fringes" on Earth, astronomers can determine the separation of the two stars.
• Let's say a star is not a point but a disk. But for a more exotic example, what if we had an incoherent source shaped like a ring? The diffraction pattern from an annular aperture is famously described by a ​​Bessel function​​, $J_0$. Sure enough, the van Cittert-Zernike theorem predicts that the spatial coherence pattern from the ring source is also a Bessel function. This isn't just a textbook curiosity; it's the principle behind stellar interferometers that measure the angular size of distant stars.

This theorem unifies the seemingly separate concepts of coherence and diffraction. Coherence is not some intrinsic property stamped onto light at its creation; it can be born from incoherence and shaped by geometry. A small, distant star produces light that is highly spatially coherent over a large area on Earth, which is why stars twinkle—the coherent wavefront is easily distorted by our atmosphere. A large, nearby object like the sun produces light that is only coherent over very small distances.

We've seen how coherence works in the pure cases of time and space. Now, let's weave them together into a richer, more complete picture. The world is rarely so simple.

The most complete descriptor of a stationary field's second-order statistics is the ​​cross-spectral density​​, $W(\mathbf{r}_1, \mathbf{r}_2, \omega)$. It tells you the correlation between two points, but for one specific frequency component $\omega$ at a time. Integrating it over all frequencies gives you back the mutual coherence function.

With this powerful tool, we can explore even subtler phenomena:

• ​​How Coherence Evolves​​: The van Cittert-Zernike theorem is a far-field approximation. What happens closer to the source, in the Fresnel or near-field regime? The coherence doesn't just appear fully formed; it evolves as the light propagates. Starting with an incoherent source, the mutual coherence function at a distance $z$ can be found by propagating the initial coherence, not just the field. This involves a more complex convolution integral, but it provides a complete picture from the source to the observer. This shows us that we can define a ​​transverse coherence length​​ that grows with distance from an incoherent source.

• ​​Correlation-Induced Spectral Shifts​​: Here's a mind-bending effect. We said the spectrum determines the temporal coherence. But it turns out that spatial coherence can, in turn, affect the spectrum you measure! If you measure the spectrum of light from a broadband source at a single point, you'll get one result. But if you measure the spectrum associated with the correlation between two different points, you might find that the peak of the spectrum has shifted. This happens because the process of generating spatial coherence doesn't treat all frequencies equally. The geometry itself acts as a kind of spectral filter.

• ​​The World of Ultrashort Pulses​​: Our discussion so far has mostly assumed "stationary" light, where statistical properties don't change over time. But what about an ultrashort laser pulse, which is over in a flash? This is a ​​non-stationary​​ field. Its coherence function $\Gamma(t_1, t_2)$ depends not just on the time difference $t_2 - t_1$, but also on the average time $(t_1+t_2)/2$, capturing the rise and fall of the pulse. When we Fourier transform this, we get a two-frequency cross-spectral density, $W(\omega_1, \omega_2)$, which reveals correlations between different frequencies. For such a pulse, the various frequency components are not independent; they are phase-locked to form the short pulse, a phenomenon beautifully captured by these more general functions.

• ​​Into the Near Field​​: The van Cittert-Zernike theorem deals with propagating waves. But very close to a source—at distances smaller than a wavelength—there are also ​​evanescent waves​​. These are "frustrated" fields that cling to the surface and decay exponentially with distance. Do they have coherence? Absolutely. A full theory of coherence propagation must include these near-field components. When we do, we find new terms in our coherence function that describe these sub-wavelength correlations, often involving exotic functions like the modified Bessel function. This isn't just theory; it's the foundation of near-field optical microscopy (NSOM), which can image features much smaller than the diffraction limit.

From the simple observation of interference fringes, we have constructed a powerful and comprehensive framework. Coherence is the measure of correlation in a wave field, a bridge between the temporal and spectral domains (Wiener-Khinchin), and a property that can emerge from chaos through propagation (van Cittert-Zernike). It is a concept that reveals the deep, underlying unity in the behavior of waves, not just light, but all waves—from the quantum mechanical wavefunction of an electron to the seismic waves of an earthquake. It is one of nature's most subtle and beautiful stories.

Now that we have acquainted ourselves with the principles and mechanisms of the coherence function, you might be asking, "So what?" It's a fair question. Why have we gone to the trouble of developing this sophisticated mathematical tool? The answer, I think you will find, is delightful. The concept of coherence is not some esoteric detail; it is a fundamental thread that weaves together vast and seemingly unrelated tapestries of the physical world. From the practical design of microscopes and telecommunication systems to the profound nature of the cosmos and the quantum vacuum, the coherence function is our guide. It tells us not just if waves will interfere, but how, and what the consequences are when they don't do so perfectly. Let's embark on a journey to see this beautiful idea in action.

The most direct and classic consequence of coherence is interference. Imagine dropping two pebbles into a still pond. Where the crests of the ripples from each pebble meet, you get a higher crest; where a crest meets a trough, the water becomes flat. This is interference. For light, the same principle holds, and the famous Young's double-slit experiment is its quintessential demonstration.

But what if the sources of the ripples are not perfectly synchronized? What if our light source is not a perfect, single-frequency plane wave, but something a bit more... jumbled? This is where the coherence function enters the stage. If we illuminate a double-slit, the clarity of the interference pattern—the difference between the brightest brights and the darkest darks—is a direct measurement of the spatial coherence of the light between the two slits. If the light at one slit is perfectly in step with the light at the other, we get perfect, crisp fringes. If their phase relationship is completely random, the pattern washes out entirely. For anything in between, we get a "greyed-out" pattern, and the visibility of those fringes gives us the numerical value of the degree of coherence. It's a beautifully direct link between a mathematical concept and a measurable reality.

This isn't just true for two points. Any object that diffracts light, like a single opening or a lens, can be thought of as a continuous collection of countless tiny sources. The final pattern we see is the grand interference of all of them. If the illuminating light is only partially coherent, it means that light passing through one part of the opening is not perfectly in step with light passing through a distant part. This has the effect of "blurring" the diffraction pattern. The sharp, intricate features you would expect from a perfectly coherent source get smeared out, as if you were looking at them through a frosted glass. The characteristic size of the "clear" patches on this frosted glass is precisely the coherence length of the light.

Engineers must wrestle with this constantly. A diffraction grating, for instance, which is used in spectrometers to split light into its constituent colors, is essentially a vast array of slits. Its ability to produce sharp, well-separated spectral lines (its resolution) depends critically on the light retaining its coherence across many of these slits. If the coherence length is shorter than the size of the grating, it is as if different parts of the grating are working independently, washing out the fine details and limiting the instrument's performance.

The formation of an image in a microscope or a camera is, at its heart, a magnificent feat of controlled interference. A lens works by gathering light waves from an object and bending them in such a way that they interfere constructively at the image plane, recreating the object. It will come as no surprise, then, that the coherence of the light used to illuminate the object plays a starring role in the quality of the final image.

In microscopy, for instance, the illumination is never perfectly coherent. We often use a broad, diffuse source, and its size determines the degree of coherence at the object plane. The theory developed by Hopkins shows how the final image intensity is a complex interplay between the coherence of the illumination, the properties of the object, and the point spread function (PSF) of the imaging lens—which describes how the lens renders a single point of light. By changing the coherence of our illumination (for example, by adjusting a diaphragm in the condenser), a microscopist can dramatically alter the contrast and resolution of the final image, highlighting different features of a specimen.

You might think that tracking how coherence evolves as it propagates through a complex system of lenses would be a nightmare. But here, nature has been kind. For a very common and useful model of partially coherent light—the Gaussian Schell-model beam—something wonderful happens. As this beam passes through a standard paraxial optical system (one that can be described by simple ray-transfer matrices), it remains a Gaussian Schell-model beam!. Its fundamental character is preserved. The beam's size and its coherence length simply transform according to rules determined by the system's magnification and geometry. This provides a powerful and elegant predictive tool, showing a deep unity in how geometric optics and the statistics of light work hand-in-hand.

The power of the coherence function truly shines when we see its principles surface in other, seemingly unrelated fields of science. The same mathematical language describes a fantastic variety of phenomena.

​​Fiber Optics and the Temporal Domain:​​ So far, we have mostly spoken of spatial coherence—correlations across space. But what about temporal coherence—correlations in time? Imagine sending a short pulse of light down an optical fiber. A fiber is a dispersive medium: different frequencies of light travel at slightly different speeds. This has a fascinating effect. Much like how propagating through a random medium degrades spatial coherence, propagation through a dispersive medium alters temporal coherence. A pulse that starts with a certain duration and temporal coherence will emerge stretched and with its coherence properties modified by the fiber's group velocity dispersion. This is of paramount importance in modern telecommunications, where billions of pulses travel through optical fibers every second, and in ultrafast science, where the shape and coherence of femtosecond pulses are everything. The parallel is striking: diffraction is to space what dispersion is to time, and the coherence function is the master key to understanding both.

​​Probing Matter with X-Rays:​​ Let's journey from the scale of global networks to the scale of atoms. In X-ray crystallography, scientists probe the structure of materials by bombarding them with X-rays and observing the resulting diffraction patterns, known as Bragg peaks. The positions of these peaks reveal the arrangement of atoms in the crystal lattice. But what about the shape of the peaks? It turns out that the finite spatial coherence of the X-ray beam is one of the factors that causes these peaks to broaden. The width of a Bragg peak is inversely related to the coherence length of the illuminating X-rays. A more coherent beam yields sharper peaks, allowing for a more precise determination of the atomic structure. Once again, our tool from classical optics finds a crucial application in the heart of condensed matter physics.

​​Looking to the Heavens:​​ Have you ever wondered why stars twinkle? It is not the star's doing, but our own atmosphere's. A star is so far away that the light arriving at Earth is, for all practical purposes, a perfect plane wave—spatially coherent across vast distances. But then it must pass through the Earth's turbulent atmosphere. Pockets of hot and cold air, with slightly different refractive indices, drift around, acting like a giant, ever-changing, random phase screen. This process catastrophically scrambles the phase of the incoming wavefront, destroying its coherence. By the time the starlight reaches a telescope on the ground, its spatial coherence length has been reduced from thousands of kilometers to, on a typical night, a mere 10 centimeters!. This coherence length, known as Fried's parameter, dictates the fundamental limit on the resolution of any ground-based telescope, no matter how large its mirror. This is why stars look like blurry blobs instead of pinpoints, and why astronomers go to such great lengths to build adaptive optics systems (which try to undo the atmospheric scrambling) or to place telescopes in space, above the turbulent mess.

We end our journey with the most mind-bending application of all, one that connects classical optics to the deepest ideas of modern physics: quantum field theory and relativity. What, you might ask, is the most perfectly coherent state imaginable? Perhaps the vacuum of empty space itself. In quantum field theory, the vacuum is not truly empty but seethes with virtual particle fluctuations. The correlations between these field fluctuations at different points in spacetime are described by a mathematical object called the Wightman function.

Now, let's do a thought experiment. It turns out that this Wightman function is, for all intents and purposes, the quantum and relativistic cousin of our mutual coherence function. For an observer at rest in empty space, the vacuum fluctuations have a very specific coherence structure. But what if you are accelerating? An accelerating observer experiences the world differently. And when we calculate the coherence function of the quantum vacuum as seen from the perspective of a uniformly accelerating observer, we find something absolutely astonishing. The mathematical form of the correlation function transforms into the very same formula that describes the coherence of light from a hot, glowing object—a thermal blackbody!.

This is the famous Unruh effect. It means that an accelerating observer will perceive the empty vacuum as a warm bath of particles, with a temperature directly proportional to their acceleration. In a very real sense, acceleration makes the vacuum glow. The key that unlocks this profound discovery is none other than the coherence function, revealing an unbelievable unity between the statistics of light in a laboratory and the very fabric of spacetime. From a simple double slit to the quantum structure of reality, the coherence function has proven to be an indispensable tool, forever revealing the intricate and beautiful connections that underlie our universe.