Coherence of Light

Coherence is the property of waves being "in step" with one another, a fundamental concept that explains why a laser is a precision tool and a lightbulb is a simple illuminator. This ordered state of light, however, is not a single, simple idea. It manifests in two critical and distinct forms—temporal and spatial coherence—that govern what we can measure and build with light, from rulers of cosmic scale to instruments that can see inside the human body. The challenge lies in understanding how these two facets of coherence arise from the physical nature of light sources and how they dictate the limits and possibilities of optical technology.

This article demystifies the coherence of light by breaking it down into its core components. First, we will investigate the underlying "Principles and Mechanisms," exploring temporal coherence as a wave's "memory" and spatial coherence as a wavefront's "uniformity." Following this, the "Applications and Interdisciplinary Connections" chapter will reveal how these principles are not just theoretical curiosities but are the crucial ingredients enabling a vast range of technologies in fields from astrophysics and medicine to quantum physics.

Imagine a perfectly still pond. You dip your finger in, just once, creating a single, perfectly expanding ripple. Now imagine a whole marching band, every drummer striking their drum at the exact same instant, with the exact same force. In both cases, there is a kind of order, a predictability that you can sense. The ripple expands in a perfectly regular way. The sound from the drums arrives as a single, powerful, unified wavefront. This, in essence, is the heart of ​​coherence​​. It is the property of waves being in step with one another, either with themselves over time or with their neighbors across space.

In the world of light, this simple idea of "being in step" splits into two fascinating and crucial concepts: temporal coherence and spatial coherence. While they sound academic, they are the very reasons a laser can cut steel while a lightbulb can barely illuminate a book, and why we can measure the size of stars trillions of miles away. Let's take a journey to understand these principles, not as definitions to be memorized, but as fundamental truths about the nature of light.

Think of a "perfect" light wave, the kind you might draw in a textbook. It's an infinitely long sine wave, oscillating with a single, pure frequency. If you know the phase—the position on the up-and-down cycle—of this wave at one moment, you know its phase at any moment in the future or past. This wave has perfect ​​temporal coherence​​. It has a perfect memory of its own past.

But nature is not so simple. Real light isn't an infinite, perfect sine wave. It is generated by messy, chaotic physical processes. An atom, for instance, doesn't emit light forever. It emits a short burst, a finite wave train, before it stops. A real light source, like a glowing gas or a hot filament, is a collection of countless atoms, all starting and stopping their emissions at random times. The result is not one infinite wave, but a jumble of short wave trains.

Temporal coherence is the measure of how long, on average, one of these wave trains continues before its phase becomes unpredictable. We quantify this with two related ideas: the ​​coherence time​​ ($\tau_c$) and the ​​coherence length​​ ($L_c$). The coherence time is the average interval over which the wave's phase remains predictable. The coherence length is simply the distance light travels in that time, $L_c = c \tau_c$. It's the physical length of the "regular" part of the wave train.

This is not just an abstract concept; it has profound practical consequences. Consider a Michelson interferometer, a device designed to split a beam of light, send the two halves down different paths, and then recombine them to see interference fringes. These fringes—the bright and dark bands—only appear if the two recombining waves are coherent with each other. But since they are just two copies of the same original wave, this is really a test of the wave's self-coherence. If the difference in the path lengths is greater than the coherence length, then by the time the wave from the longer path arrives to meet its partner from the shorter path, the original phase relationship is lost. The wave has "forgotten" what it was doing. The two waves are like strangers, and the interference fringes vanish entirely. For a source with a very short coherence time, say 10 femtoseconds ($10^{-14}$ s), this "vanishing point" is reached at a path difference of just a few micrometers.

So, what determines the coherence time? Why is a laser's coherence length measured in meters or even kilometers, while the light from a glow-in-the-dark sticker is coherent for only a few micrometers? The answer lies in one of the most beautiful relationships in physics: the connection between the time domain and the frequency domain, linked by the Fourier transform.

A perfectly monochromatic wave, with its infinite coherence time, has a spectrum that is an infinitely sharp spike at a single frequency. In contrast, a wave that is restricted in time—a short pulse or a finite wave train—cannot be made of a single frequency. The very act of starting and stopping the wave requires a mixture, or a "superposition," of many different frequencies. The shorter the pulse in time, the broader the range of frequencies needed to build it.

This means that the coherence time $\tau_c$ is inversely proportional to the ​​spectral linewidth​​ $\Delta\nu$, which is the width of the band of frequencies present in the light. $\tau_c \approx \frac{1}{\Delta\nu}$ A narrow spectrum (like a laser's) means a long coherence time. A broad spectrum (like a lightbulb's or a glow-sticker's) means a very short coherence time. This gives us tools to calculate coherence length directly from a light source's spectrum, whether measured in frequency ($\Delta\nu$) or wavelength ($\Delta\lambda$). For a spectrum with a width $\Delta\lambda$ around a central wavelength $\lambda_0$, the coherence length is approximately $L_c \approx \lambda_0^2 / \Delta\lambda$.

This relationship is more than just an approximation; it holds a deep truth. The celebrated ​​Wiener-Khinchin theorem​​ formalizes this, stating that the temporal coherence function and the power spectral density of the light are a Fourier transform pair. This means that if we know the precise shape of the light's spectrum, we can predict the precise way its coherence fades with time. For a source with not just one broad peak, but a structured spectrum—say, two distinct spectral lines—the coherence doesn't just decay. It can decay, then reappear, and then decay again, in a phenomenal display of "quantum beats." The separation of the spectral lines dictates the frequency of these revivals in coherence, a beautiful and direct confirmation of the Fourier relationship.

To truly understand coherence, we must ask an even deeper question: why do light sources have a spectral width in the first place? Feynman would urge us to look at the machinery underneath. Light comes from atoms, and the "perfection" of the light they emit is limited by what's happening to those atoms.

An excited atom doesn't stay excited forever. It will spontaneously decay and emit its light in a process that has a characteristic lifetime. This finite lifetime fundamentally limits the length of the emitted wave train, a phenomenon called ​​natural broadening​​. But other things can disturb the atom, too. Imagine an atom trying to emit a perfectly smooth wave, but it's constantly being jostled by its neighbors in a hot gas. These collisions interrupt the phase of the emission, effectively cutting the wave train short. This is called ​​collisional broadening​​. The total spectral width is determined by all such processes that "dephase" the atomic oscillator. The coherence time of the emitted light, then, is a direct measure of how long an atom is left alone to do its job.

There's another, wonderfully intuitive source of broadening. In a gas discharge tube, the atoms emitting light are hot, which means they are whizzing about in all directions. Light from an atom moving towards you is Doppler blue-shifted. Light from one moving away is red-shifted. What you see is the sum of all these slightly different colors, resulting in a broadened spectral line known as ​​Doppler broadening​​. And because the atomic speeds increase with temperature, a hotter gas produces a broader spectrum and, therefore, a shorter coherence time.

Now let's turn our attention from the memory of a single wave train to the relationship between different points on a wavefront. This is ​​spatial coherence​​. If temporal coherence is one musician playing in time with themselves, spatial coherence is the entire orchestra playing in unison. It asks: if we pick two different points in space, are the light waves passing through them in sync?

This is the key to understanding why Young's double-slit experiment works so well with a laser but fails with a candle. To see interference, the light arriving at the two slits must be coherent with each other. For a laser, the beam is highly spatially coherent; the phase is uniform across the entire wavefront.

But what about a "normal" light source, like a frosted lightbulb or a distant star? These are ​​incoherent sources​​. You can think of them as being made up of millions of tiny, independent emitters, like a vast crowd of people talking at once. Each emitter sends out its own wave trains with no phase relationship to its neighbors.

How could such a jumbled source ever produce spatially coherent light? The magic happens with distance.

Imagine you are looking at an extended, incoherent source from very far away. Consider two points on your observation screen (or two slits in a barrier). For the light waves at these two points to be coherent, they must have originated from roughly the same region of the source and traveled roughly the same distance. If your two observation points are very close together, the light arriving at both will have come from a similar mix of emitters on the source. Their random phases will tend to add up in a similar way, and the waves will be correlated. But as you move your observation points further apart, they start "seeing" different parts of the incoherent source. The waves become uncorrelated, and the spatial coherence is lost.

This defines a ​​transverse coherence length​​, $l_c$, also known as the size of the "coherence area." Within this area, the light is effectively coherent. Outside it, it is not. This is why, to see fringes from a sodium lamp, the slits must be placed very close together—within its small coherence length.

Here comes the truly astounding part, formalized by the ​​van Cittert-Zernike theorem​​. It states that the spatial coherence pattern of the light from a distant, incoherent source is related to the Fourier transform of the source's shape and intensity distribution. This is a profound and powerful result. An immediate consequence is the formula for the transverse coherence length from a distant circular source of diameter $D_s$ at a distance $L$: $l_c \approx \frac{\lambda L}{D_s}$ (The exact formula includes a factor of 1.22 from the Fourier transform of a circle).

Notice the inverse relationship with the source diameter $D_s$. A larger source produces a smaller coherence area. A smaller source produces a larger coherence area. This is why placing a tiny pinhole in front of a messy, extended light source dramatically improves its spatial coherence! The pinhole effectively becomes a new, very small source, which generates a large area of coherent light downstream. The theorem also correctly predicts that longer wavelengths produce larger coherence areas.

The van Cittert-Zernike theorem is not just a mathematical curiosity; it is a fundamental tool for astronomers. If the coherence pattern is the Fourier transform of the source's shape, then by measuring the coherence, we can reconstruct the source's shape! This is the principle behind stellar interferometry. By building interferometers with detectors that can be moved apart, astronomers can measure the spatial coherence of light from a distant star. Finding the detector separation at which the interference fringes first vanish allows them to calculate the star's angular size with incredible precision.

The Fourier relationship reveals even more subtle details. Imagine observing a distant, elliptical nebula. The theorem predicts that the coherence area on your detector will also be an ellipse, but with a twist: it will be rotated by 90 degrees. The long axis of the nebula corresponds to the short axis of the coherence area, and vice-versa. This inverse relationship is a direct and beautiful hallmark of the Fourier transform. By measuring the aspect ratio of the coherence area, astronomers can directly deduce the aspect ratio of the distant nebula itself.

From the inner workings of an atom to the measurement of distant stars, the concept of coherence provides a unifying thread. It is a story of order and disorder, of memory and forgetfulness, of the deep and beautiful symmetry between time and frequency, space and shape. It is a reminder that even in the seemingly random flicker of a distant light, there is a hidden order waiting to be discovered.

Now that we have wrestled with the rather abstract ideas of temporal and spatial coherence, you might be excused for asking, "That's all very clever, but what is it good for?" The answer, it turns out, is astonishingly broad. Coherence is not some esoteric detail found only in the dusty corners of a textbook. It is a fundamental property of light that, in a very real sense, determines what we can see and measure in the universe. It is the secret ingredient that enables some of our most powerful technologies, and the fundamental limit that challenges our ingenuity.

In this chapter, we will embark on a journey to see how this one concept—how well a wave "agrees" with itself across time and space—weaves a golden thread through an incredible range of scientific and engineering disciplines. We will see how it lets us build rulers of breathtaking precision, how it decodes the chemical composition of distant stars, and how it allows us to peer inside living tissue without making a single cut. We will discover that the same principle used to measure stars can be used to probe the strange world of materials on the verge of a phase transition. And finally, we will catch a glimpse of the future, where coherence takes on an even deeper meaning in the quantum realm, becoming a resource we can actively control and manipulate.

Let’s begin with temporal coherence—the measure of a light wave's phase memory, or how long it can "remember" the rhythm of its own oscillations. This "memory," quantified by the coherence time $\tau_c$ or coherence length $L_c = c \tau_c$, is the absolute prerequisite for any instrument based on interference.

​​Precision Rulers and The Limits of Measurement​​

The simplest and most profound application of interference is to measure distance. An interferometer, in essence, is a ruler made of light. It works by splitting a beam, sending the two halves along different paths, and then recombining them. A change in the path length difference shows up as a shift in the interference fringes. But how long can this ruler be?

Here, coherence provides the answer. To see interference, the path length difference between the two arms, $\Delta L$, cannot be much greater than the coherence length of the light source. If you try to compare a wave with a copy of itself that is too delayed, the original wave will have "forgotten" its phase, and the stable interference pattern washes out. This simple rule dictates the design of any high-precision interferometric sensor. If you need your device to handle a large path difference, you absolutely must use a source with a long coherence time. This is why the monumental LIGO detectors, which measure spacetime ripples by tracking minuscule changes in a 4-kilometer-long path, must use extraordinarily coherent lasers.

The same principle appears in more common settings. Ever seen the beautiful, swirling colors on a soap bubble or a thin film of oil on water? Those are interference fringes. But you won’t see such fringes by reflecting light off a thick piece of window glass. Why? Because for the thick glass, the path difference between light reflecting from the front surface and light reflecting from the back surface is huge. An ordinary light bulb has a pathetically short coherence length, so the fringes are completely washed out. However, if you illuminate that same thick plate with a highly monochromatic source—like a sodium lamp or a laser—the long coherence length allows the fringes to reappear, sharp and clear. The visibility of these fringes is a direct measure of the light's temporal coherence, decaying exponentially as the path difference, and thus the plate's thickness, increases. Similarly, when looking at the interference from a wedge-shaped film of air, the number of fringes you can count is finite, limited directly by the point where the wedge thickness creates a path difference that exceeds the source's coherence length.

​​The Spectroscopist's Secret​​

How do we know the Sun is made of hydrogen and helium? We look at its spectrum—the unique barcode of colors it emits. The instrument that performs this magic is a spectrometer. But what might come as a surprise is that a spectrometer's resolving power—its ability to distinguish two very similar colors—is also governed by coherence.

A diffraction grating, the heart of many spectrometers, works by creating interference from thousands of tiny parallel paths. To separate two nearby wavelengths, say $\lambda$ and $\lambda + \Delta\lambda$, the instrument must create a large enough maximum path difference between the interfering rays. And, as we know, this maximum path difference is fundamentally limited by the coherence length of the light itself! It reveals a beautiful and deep reciprocity, rooted in the mathematics of the Fourier transform: to see very fine details in the frequency spectrum ($\Delta\lambda$ is small), you need to observe the wave over a very long time or path (coherence length $L_c$ is large). In fact, the theoretical best resolution any such instrument can achieve is given by a wonderfully simple relation: $\Delta\lambda_{min} = \lambda^2 / L_c$. To build a high-resolution spectrometer for astrophysics, one needs a large grating precisely to generate this long path difference, pushing the instrument's capabilities to the coherence limit of the starlight it analyzes. This same principle dictates that if you need a highly coherent beam for a lab experiment, a good way to create it is by passing broadband light through a filter, like a Fabry-Perot etalon, which removes other frequencies and thus increases the coherence length of the light that gets through.

​​Seeing with "Bad" Light: Optical Coherence Tomography​​

So, is high coherence always a good thing? Not at all! In a beautiful example of turning a "bug" into a billion-dollar "feature," the medical imaging technique known as Optical Coherence Tomography (OCT) harnesses low temporal coherence to see inside the human body.

Imagine trying to get a high-resolution image of the layers of the retina at the back of the eye. You can't cut it open. An OCT system sends a beam of light into the eye and analyzes the faint reflections coming back from the different tissue layers. The challenge is depth perception: how do you know which echo came from which layer?

The ingenious solution is to use light with a very short coherence length—that is, a source with a very broad range of colors, like a superluminescent diode. In the OCT instrument, this light is split. One beam goes into the eye, and the other travels along a reference path of controllable length. Interference can only happen when the path length of a reflected echo from a specific depth in the retina exactly matches the reference path length, to within one coherence length. Since the coherence length is tiny (micrometers!), only reflections from a very thin slice of tissue will produce an interference signal at any given moment. By systematically changing the length of the reference path, the machine scans this "coherent gate" through the tissue, building up a breathtakingly detailed, cross-sectional image, layer by layer. For OCT, high axial resolution requires a short coherence length, which in turn demands a light source with a very large spectral bandwidth, $\Delta \lambda$.

This elegant reversal of intuition—using "temporally messy" light to create beautifully ordered images—has revolutionized ophthalmology and is finding use in dozens of other medical fields.

Finally, as our technology for manipulating light advances, the concept of coherence follows. In the field of nonlinear optics, processes like second-harmonic generation can convert infrared laser pulses into green light. What happens to the coherence time? Intuitively, if the frequency of the wave is doubled, its "clock" ticks twice as fast. Any uncertainty in the original frequency is also doubled. This means the spectral bandwidth $\Delta\nu$ doubles, and because coherence time $\tau_c$ is inversely related to bandwidth, the coherence time of the new green light is halved.

Let’s now turn our attention sideways. Spatial coherence asks how the phase of a wavefront at one point relates to the phase at another point nearby. It's a measure of the "smoothness" or "uniformity" of a wavefront, and it carries information not about the light's color, but about the size and shape of its source.

​​Measuring the Stars​​

Even in the most powerful telescopes, most stars appear as dimensionless points of light. So how can we possibly measure their diameter? The answer, first demonstrated by Albert A. Michelson, lies in measuring the spatial coherence of starlight.

A truly point-like, infinitely distant source would produce perfectly plane waves, which have perfect spatial coherence. But a star is not a point; it’s a disc of a certain tiny angular size. It acts as an extended, incoherent source—each point on its surface radiates independently. The celebrated van Cittert-Zernike theorem tells us something remarkable: as light from such a source travels across the vastness of space, it becomes more and more spatially coherent. Here on Earth, the starlight exhibits partial coherence over a characteristic "patch size." The size of this patch is inversely proportional to the star's angular diameter.

A stellar interferometer works by using two separated telescopes to sample the light at two points. If the telescopes are close together (within a coherence patch), the two signals interfere strongly, producing clear fringes. As the distance between the telescopes increases, they begin to sample points that are less and less correlated, and the fringe visibility drops. The separation at which the fringes disappear gives a direct measure of the star's size.

There is a deep and beautiful mathematical duality at play here: the spatial coherence function of the light arriving on Earth is the Fourier transform of the brightness distribution of the source. This is precisely the same mathematical relationship that describes Fraunhofer diffraction, where the far-field diffraction pattern (like an Airy disk) is the Fourier transform of the aperture's shape. This isn't a coincidence; it reflects a profound unity in the theory of waves. You can even calculate the coherence between two points by mapping the problem onto the diffraction from an aperture with the same dimensions as the source.

​​Peeking into Murky Materials​​

This powerful idea—using coherence to map a source—is not limited to astrophysics. We can turn it around. Start with a perfectly coherent laser beam and shine it through a translucent material, like a fluid near its critical point, where it fluctuates wildly between liquid and gas. The scattering process scrambles the light, and the perfectly coherent incident beam emerges with a complex pattern of partial spatial coherence.

The coherence properties of this scattered light are now a map—a hologram, if you will—of the scattering medium itself. By measuring the spatial coherence of the light scattered in different directions, physicists can deduce the statistical properties of the density fluctuations inside the fluid. It allows them to measure quantities like the correlation length $\xi$, which describes the typical size of the fluctuating regions. In essence, the scattered light's coherence pattern is the Fourier transform of the material's internal correlation function. Light becomes a non-invasive probe, allowing us to "see" the statistical dance of atoms and molecules in complex systems, a fantastic bridge between optics and statistical mechanics.

The story of coherence does not end with classical waves. In the world of quantum mechanics, its meaning becomes even deeper and more powerful. Here, coherence refers to the existence of definite phase relationships between different quantum states in a superposition. An electron can be in a superposition of being in two places at once; an atom can be in a superposition of its ground and excited states. The phase relationship between these possibilities is its quantum coherence.

This is not just an analogy. The quantum mechanical "coherence" between two states, captured by the off-diagonal elements of a system's density matrix, is the direct mathematical and physical precursor to the classical coherence we've been discussing.

Consider two quantum emitters—two artificial atoms, perhaps—placed close together. We can prepare the system so that only one of them is excited. If the emitters can interact, the excitation can be exchanged back and forth between them. The system will then exist in a quantum superposition of "emitter 1 is excited" and "emitter 2 is excited". The phase relationship between these two components of the superposition is a measure of the system's quantum coherence. This very coherence dictates whether the light subsequently emitted by the two atoms will interfere in the far field. The fringe visibility is a direct readout of the quantum coherence between the emitters.

What is truly exciting is that we are now learning to engineer this coherence. In a remarkable setup, two emitters can be embedded in a tiny ring. By applying a magnetic field through the center of the ring, the Aharonov-Bohm effect allows us to tune the coherent coupling between them. A change in the magnetic flux changes the quantum phase accumulated by an excitation as it travels around the ring, thereby strengthening or weakening the coherent superposition. This means we can control the fringe visibility of the emitted light simply by turning a magnetic knob.

This is a profound shift. We are no longer just passive observers or beneficiaries of coherence as provided by nature. We are becoming architects of coherence at the most fundamental level, building and manipulating quantum superpositions. This ability is at the very heart of the ongoing revolutions in quantum computing, quantum sensing, and quantum communication.

From measuring galaxies to imaging a single cell to programming a quantum computer, the concept of coherence is an indispensable part of our toolkit for understanding and shaping the world. It is a testament to the power of physics that a single, elegant idea can connect so many disparate parts of our reality, revealing over and over again the hidden unity of nature.