Partial Coherence

In the world of physics, we often think in absolutes: the perfect, orderly waves of a laser represent coherence, while the chaotic flicker of a candle flame embodies incoherence. Yet, most light in the universe, from the twinkle of a star to the glow of a computer screen, exists in a state somewhere between these two extremes. This is the realm of partial coherence, a concept that is both fundamental to understanding the nature of light and crucial for a vast array of technologies. The challenge lies in moving beyond intuitive notions of order and chaos to develop a precise framework for this 'in-between' state: how can we quantify it, what physical laws govern it, and what are its practical consequences?

This article provides a comprehensive overview of partial coherence, bridging theory and application. The first chapter, ​​Principles and Mechanisms​​, will demystify the concept by exploring how coherence is measured, introducing its two faces—temporal and spatial—and explaining the two profound theorems that form its theoretical bedrock. Following this, the chapter on ​​Applications and Interdisciplinary Connections​​ will reveal how partial coherence is not a nuisance but a powerful parameter that is deliberately engineered in fields ranging from microscopy and semiconductor manufacturing to astronomy and even the study of analogue black holes. To embark on this exploration, we first need to understand the fundamental principles that govern this fascinating phenomenon.

Imagine dropping a single pebble into a silent, still pond. A perfect, ordered pattern of circular waves expands outwards. If you were to place two tiny corks on the water, they would bob up and down in a predictable, synchronized rhythm. Their motion is correlated. This is the essence of ​​coherence​​. Now, imagine a hailstorm pelting the surface of the pond. The water is a chaotic mess of crisscrossing, interfering ripples. The motion of two nearby corks would be random and utterly uncorrelated. This is ​​incoherence​​.

The light from a laser is like that single pebble—an orderly, continuous train of waves. The light from a candle flame or a star is more like that hailstorm—a jumble of countless tiny bursts of light, emitted randomly by trillions of excited atoms. Most light in the universe, and in our daily lives, is neither perfectly coherent nor perfectly incoherent. It lies somewhere in between. It is ​​partially coherent​​. But how do we get a grip on this "in-between" state? How do we measure it, and what does it mean?

The classic way to reveal the wavelike nature of light is Young's double-slit experiment. When coherent light passes through two narrow slits, it creates a pattern of bright and dark bands, or ​​fringes​​, on a distant screen. The bright fringes are where the waves from the two slits arrive in step (constructive interference), and the dark fringes are where they arrive out of step (destructive interference).

The "clarity" or "contrast" of this fringe pattern is a direct measure of the coherence of the light at the two slits. If the light is perfectly coherent, we get sharp, distinct fringes—dark bands that are perfectly black and bright bands that are maximally bright. If the light is completely incoherent, the patterns from the two slits just add up, and we see a uniform wash of light with no fringes at all.

For partially coherent light, we see fringes, but they are washed out. The dark bands aren't completely dark, and the bright bands aren't as bright as they could be. We can put a number on this by measuring the ​​fringe visibility​​, defined as:

where $I_{max}$ and $I_{min}$ are the maximum and minimum intensities in the fringe pattern. A visibility of $V=1$ means perfect contrast, while $V=0$ means no fringes.

This visibility is governed by a fundamental quantity we call the ​​complex degree of coherence​​, denoted by $\gamma_{12}$. This number's magnitude, $|\gamma_{12}|$, ranges from $1$ (perfectly coherent) to $0$ (perfectly incoherent). In the ideal case where the slits are providing equal amounts of light to the screen, the visibility is simply equal to the magnitude of the degree of coherence, $V = |\gamma_{12}|$. In the real world, factors like unequal beam intensities or stray background light can reduce the visibility, even if the underlying coherence of the light is high. The degree of coherence remains the fundamental property of the light itself, dictating its potential to interfere.

So, coherence is about correlation. But correlation can happen in two different ways, giving coherence two "faces": temporal and spatial.

​​Temporal coherence​​ is about correlation in time. Imagine a single, infinitely long wave train of a pure red color. If you look at the wave at one point in space, its shape now is perfectly predictable from its shape a moment ago, or a second ago, or an hour ago. It's correlated with itself across time. This is perfect temporal coherence. Now think of white light from an incandescent bulb. It’s a mashup of countless tiny wave packets of all colors, emitted at random times. The wave's shape at one instant has almost no relation to its shape a tiny fraction of a second later. It has very poor temporal coherence. The time duration over which a wave remains correlated with itself is called the ​​coherence time​​, $\tau_c$. For our pure red wave, $\tau_c$ is infinite; for the white light, it is fantastically short.

​​Spatial coherence​​, on the other hand, is about correlation in space. This time, imagine taking a snapshot of the wave at a single instant. How is the wave's value at point A related to its value at point B? For light coming from a distant, tiny point source (an idealized star), the waves expand in perfect spheres. Any two points on the same expanding wavefront will be perfectly in step. This field has high spatial coherence. Now consider a large, extended source like a frosted lightbulb. Light from the left side of the bulb is emitted independently from light on the right side. The overall wavefront is a jumble. If you pick two points close together, the waves might be in step, but if you move them farther apart, the correlation disappears. The region within which the field stays reasonably correlated is called the ​​coherence area​​.

Let's look closer at temporal coherence. A light source like a bulb or a star is a collection of agitated atoms. Each atom, when it de-excites, emits a tiny packet of light—a short wave train that lasts for only a brief period. The total light field is the sum of trillions of these random, overlapping wave packets. This is why the coherence time is so short.

There is a deep and beautiful connection here, a universal rule of nature known as the ​​Wiener-Khinchin theorem​​. It states that the temporal coherence of a light field and its frequency spectrum are inextricably linked—specifically, they are a ​​Fourier transform pair​​.

Think of sound. A quick, sharp sound like a clap contains a very wide range of frequencies. You can't hum the "note" of a clap. A long, sustained, pure note from a flute, however, has a very narrow frequency spectrum. The same is true for light. A short pulse of light (short coherence time) must, by its very nature, be composed of a broad range of frequencies (a wide spectrum). A wave that is nearly one pure color (a narrow spectrum) must be long and continuous (long coherence time).

This isn't just a qualitative idea; it's a precise mathematical law. If you know the power spectral density $S(\omega)$ of a light source—essentially its "color palette"—you can calculate its temporal coherence function $\gamma(\tau)$ by taking the Fourier transform. For instance, if a source has a spectrum consisting of two sharp, distinct frequency peaks (a spectral doublet), its coherence function turns out to be an oscillating cosine wave that dies away exponentially. The oscillation corresponds to the "beating" between the two frequencies, while the decay is related to the width of the spectral lines. Conversely, if you measure a coherence function that has this decaying cosine shape, you can deduce that the source spectrum must be a doublet of broadened lines. This powerful two-way relationship allows us to understand the nature of light sources by studying the interference they produce, and vice-versa. Even an abstract spectrum, like a one-sided exponential decay, has a corresponding and unique coherence function.

Now we come to one of the most surprising and elegant ideas in all of optics: spatial coherence can arise spontaneously from a completely incoherent source. This is the magic of the ​​van Cittert-Zernike theorem​​.

Let's go back to our hailstorm on the pond. Close to where the hail is falling, the water surface is pure chaos. But if you go very, very far away, the jumble of tiny, sharp ripples starts to smooth out. From a great distance, the disturbance begins to look more like a single, large, slowly undulating wave. The same thing happens with light.

Consider a star. It's a giant ball of roiling, independent emitters—a textbook example of a spatially incoherent source. Yet, the light from that star, by the time it reaches Earth, is spatially coherent enough to produce interference fringes in a telescope. How?

Let's build a simple model. Imagine two tiny, independent lightbulb filaments, A and B, sitting side-by-side. They flicker randomly, with no correlation between them. A great distance away, we place a screen and pick two points, $P_1$ and $P_2$. The light reaching $P_1$ is the sum of the fields from A and B ($E_{A1} + E_{B1}$). The light at $P_2$ is likewise $E_{A2} + E_{B2}$. To find the coherence between $P_1$ and $P_2$, we're checking for a correlation between $(E_{A1} + E_{B1})$ and $(E_{A2} + E_{B2})$. Because A and B are independent, the cross-correlations (like between $E_{A1}$ and $E_{B2}$) average to zero over time.

We are left with two terms: the self-correlation from source A between the two points, and the self-correlation from source B. The light from a single point source is spatially coherent with itself, so each of these terms is non-zero. But—and here is the key—the path from source A to $P_1$ and $P_2$ is different from the path from source B to $P_1$ and $P_2$. This introduces a path-dependent phase shift. When we add the two correlation terms together, these phases can add up or cancel out. The result is an interference pattern in the coherence itself. The coherence between $P_1$ and $P_2$ is found to vary as a beautiful cosine function, depending on the separation of the points and the separation of the original sources.

The van Cittert-Zernike theorem generalizes this profound result: ​​the complex degree of spatial coherence of the field produced by a distant, incoherent source is given by the Fourier transform of the source's intensity distribution.​​ For our two point sources, the intensity distribution is two delta functions, and its Fourier transform is a cosine—exactly what we found. For a uniformly lit, incoherent line source of width $a$, the intensity distribution is a rectangle function. Its Fourier transform is a sinc function, $\sin(x)/x$. And indeed, the theorem correctly predicts that the spatial coherence in the far field will follow a sinc function pattern. This remarkable theorem is the principle behind stellar interferometry, which allows astronomers to measure the size of distant stars by, in effect, measuring the Fourier transform of their disk as written into the coherence of their light.

We have seen that the magnitude of the complex degree of coherence, $|\gamma_{12}|$, tells us the visibility or contrast of the interference fringes. But what about the fact that it's a complex number? We can write it as $\gamma_{12} = |\gamma_{12}| e^{i\alpha}$, where $\alpha$ is the phase. Does this phase have a physical meaning?

It absolutely does. It tells us the position of the fringes.

In a standard, symmetric Young's double-slit experiment, the central bright fringe appears right on the central axis ($y=0$), equidistant from the two slits. This is the spot where the path lengths are equal. This happens when the light reaching the two slits is perfectly in phase, meaning $\alpha=0$.

But suppose the light reaching one slit is slightly delayed relative to the other. This could happen if the illuminating source is not centered, or if we deliberately place a thin slip of glass over one slit. This introduces a phase difference $\alpha$ between the two slits. The condition for maximum constructive interference is no longer just zero path difference. The path difference must now compensate for this initial phase $\alpha$. The result? The entire fringe pattern is shifted sideways. The amount of the shift is directly proportional to the phase $\alpha$.

So, the full story of coherence is written in a complex number. Its magnitude tells you how strong the correlation is, determining the contrast of the interference. Its phase tells you about the relative timing or alignment of the waves, determining the position of the interference. Together, they provide a complete statistical description of the relationship between light waves at different points in space and time, turning what seems like a random, chaotic process into a predictable and beautifully structured phenomenon.

Now that we have journeyed through the principles and mechanisms of coherence, a fair question to ask is: what is it all good for? Is this concept of "partial coherence" merely a mathematical refinement, a footnote to the idealized world of perfectly coherent waves? The answer, you might be delighted to find, is a resounding no. The real world is a world of partial coherence. This seemingly subtle idea is not an academic curiosity; it is a fundamental aspect of nature that shapes what we see, what we can build, and how we probe the deepest secrets of the universe. From the twinkling of a distant star to the creation of a computer chip, the fingerprints of partial coherence are everywhere. Let us now explore some of these fascinating applications and connections.

The most direct and fundamental consequence of partial coherence is its effect on interference, the beautiful phenomenon that first convinced us of the wave nature of light. Imagine you are performing Thomas Young’s classic double-slit experiment. In a perfect world, with an idealized, perfectly "in-sync" light source, you’d see a striking pattern of sharp bright and dark stripes, or fringes. But what happens with a real-world light source, like a glowing filament or even a simple gas discharge lamp? The light arriving at one slit is not perfectly correlated with the light arriving at the other.

As the two slits are moved farther apart, this correlation drops. The result? The fringes become washed-out, losing their crisp contrast, or "visibility." In fact, this visibility is a direct, measurable embodiment of the complex degree of coherence, $\gamma_{12}$, between the two points. If the distance $d$ between the slits becomes significantly larger than the source's transverse coherence length, $L_c$, the interference pattern vanishes completely, smeared into a uniform glow. The intensity we observe on the screen is no longer a simple pattern of perfect nulls and peaks, but is modulated by the degree of coherence, following a rule like $I(\theta) \propto 1 + |\gamma_{12}| \cos(\phi)$, where $\phi$ is the phase difference from the path lengths. The "realness" of the source imprints itself directly onto the "perfection" of the fringes.

This principle extends far beyond two simple slits. Any time a wave passes through an aperture, it diffracts, creating a pattern. The shape of this diffraction pattern is not determined by the aperture alone. It is a dialogue between the object and the light illuminating it. Light from a source with limited spatial coherence will "smear out" the fine details of the diffraction pattern, broadening its features in a way that is mathematically analogous to how a blurry lens would affect an image. Furthermore, as a beam of partially coherent light, such as that described by the workhorse Gaussian Schell-model, propagates through space, its coherence properties evolve. A key consequence is that a partially coherent beam spreads out, or diverges, more rapidly than a fully coherent laser beam of the same initial size. This is a critical consideration in fields ranging from fiber optics to long-distance laser communication.

For an engineer, partial coherence is not just a nuisance to be overcome; it is a powerful parameter to be controlled and exploited. Nowhere is this more evident than in the field of optical microscopy. You might think that to get the sharpest possible image of a tiny biological cell, you would want to illuminate it with the most coherent light possible, say, from a pristine laser. The truth is quite the opposite! Illuminating a sample with perfectly coherent light often creates a host of problems, such as spurious fringes and a grainy pattern called "speckle," which obscure the very details you wish to see.

Modern high-resolution microscopes, particularly those using Köhler illumination, are marvels of "coherence engineering." By carefully adjusting the size of an aperture in the illumination system (the condenser diaphragm), a microscopist deliberately sets the spatial coherence of the light at the sample plane. The beautiful physics behind this is captured by the van Cittert-Zernike theorem, which tells us that the coherence function at the sample is essentially the Fourier transform of the light source's intensity distribution. By tuning the source size, one can find a "sweet spot" of partial coherence that minimizes artifacts while maintaining high resolution. It is a wonderful paradox: to see things more clearly, we must sometimes make the light a little less perfect. This control extends through the entire imaging system, where lenses and other components transform the coherence properties of the light in a predictable way, a process elegantly described by matrix optics.

Another spectacular example comes from the multi-trillion-dollar semiconductor industry. To manufacture modern computer chips, intricate patterns are etched onto silicon wafers using a process called photolithography. A layer of light-sensitive material, or photoresist, is exposed to ultraviolet light passing through a mask. A persistent problem arises because the silicon wafer underneath is highly reflective. The light reflects off the wafer and interferes with the incoming light, creating "standing waves" within the photoresist. This causes the resist to be exposed in an uneven, layered pattern, ruining the fidelity of the tiny circuits being printed.

The ingenious solution involves not spatial, but temporal coherence. By deliberately using a light source that is not perfectly monochromatic—that is, one with a finite spectral bandwidth $\Delta\lambda$—engineers can suppress these unwanted standing waves. The reflected light is delayed by its round-trip journey through the resist. If this time delay is greater than the source's coherence time (which is inversely proportional to $\Delta\lambda$), the reflected wave is no longer "in-sync" with the incoming wave, and the interference is washed out. It is a stunning application where "dirtiness" in the light's color spectrum is precisely engineered to ensure "cleanliness" in the final product.

The reach of partial coherence extends from the factory floor to the cosmos. Anyone who has looked up at the night sky has observed the twinkling of stars. This familiar and poetic phenomenon is a direct consequence of the loss of coherence. A plane wave of light from a distant star arrives at the top of our atmosphere in a state of near-perfect spatial coherence. But as this wave travels through the turbulent atmosphere, with its swirling pockets of air at different temperatures and densities, different parts of the wavefront are randomly delayed.

The perfectly coherent plane wave is corrupted into a chaotic, partially coherent mess by the time it reaches a telescope on the ground. This is what astronomers call "atmospheric seeing," and it is the primary reason why ground-based telescopes blur the images of stars into dancing blobs instead of sharp points. The entire field of adaptive optics is a heroic effort to measure this corruption of coherence in real-time and undo it with deformable mirrors, restoring the starlight to its pristine, coherent state.

Zooming from the largest scales to the smallest, the concept of coherence is also vital in an entirely different way of "seeing": X-ray diffraction. Scientists use beams of X-rays to probe the arrangements of atoms in crystals. The X-rays scatter off the orderly planes of atoms, creating a diffraction pattern of sharp spots known as Bragg peaks. The positions and intensities of these peaks reveal the crystal's structure. However, the width of these peaks also contains information—about the size of the crystal or the presence of defects. But there's a catch: the width of the measured peak also depends on the spatial coherence of the X-ray beam itself! An X-ray beam with a finite coherence length $\xi$ will cause an intrinsic broadening of the Bragg peaks, which can be mistaken for a physical property of the sample material. Understanding the coherence of the source is therefore absolutely essential for a correct interpretation of the data, especially when using advanced sources like synchrotrons and free-electron lasers.

Perhaps the most profound illustration of the power of the concept of coherence is its appearance in one of the most exotic frontiers of modern physics: the study of analogue black holes. In recent years, physicists have realized that they can create "mock" black holes in the laboratory using systems like flowing water or ultracold atoms in a Bose-Einstein condensate (BEC). In a BEC, for example, one can create a region where the fluid of atoms is flowing faster than the local speed of sound. This region creates an "acoustic event horizon"—a point of no return for sound waves (phonons), just as a real event horizon is a point of no return for light.

One of the most startling predictions of general relativity is that black holes are not truly black, but emit a faint thermal glow known as Hawking radiation. Theories predict that these acoustic horizons should do the same, emitting a thermal bath of phonons. This "radiation" would manifest as a peculiar type of statistical correlation between the quantum phonon field on either side of the horizon. And how do physicists describe this correlation? With the mutual coherence function, $\Gamma(x_1, x_2)$. The very same mathematical tool we use to quantify the washed-out fringes in a Young's slit experiment is used to hunt for the echoes of Hawking radiation from a simulated black hole in a cloud of atoms cooled to near absolute zero.

This is a breathtaking testament to the inherent unity of physics. The single, simple idea of how well two waves are synchronized with each other—their coherence—provides a continuous thread connecting the mundane to the magnificent. It bridges the gap between classical optics and quantum fields, between engineering and cosmology. Partial coherence is not just a correction to an ideal theory; it is a fundamental part of the rich and intricate tapestry of the physical world.