Coherency Matrix

While the pristine, orderly state of a laser beam can be described by a simple vector, most light we encounter—from the sun, a candle, or a lightbulb—is a statistical jumble of polarizations. Describing this "fuzzy" or partially polarized light requires a more sophisticated tool that captures not just a snapshot, but the average behavior and internal correlations of the light wave. This complexity represents a significant challenge in classical optics, where a complete description of any light beam is essential for both fundamental understanding and practical application.

This article introduces the ​​coherency matrix​​, the elegant mathematical framework developed to solve this problem. It provides a complete statistical portrait of any light beam's polarization state, from perfectly ordered to completely random. We will explore the principles behind this powerful tool and see how it unifies various concepts in the study of light. The following chapters will guide you through this exploration. First, under "Principles and Mechanisms," we will deconstruct the matrix itself, learning how to interpret its elements to understand intensity, correlation, and the Degree of Polarization, and uncovering its deep connection to quantum mechanics. Following that, in "Applications and Interdisciplinary Connections," we will see the coherency matrix in action, demonstrating its power to solve practical problems in optics, interference, and material science.

Imagine trying to describe a perfectly choreographed ballet. You could track the precise position of each dancer at every moment. This is what we do for perfectly polarized light, like the beam from an ideal laser—we use a simple arrow, the ​​Jones vector​​, to describe its unwavering state. But now, picture trying to describe the chaotic energy of a crowded dance floor at a wedding. Dancers are moving every which way, some in pairs, some alone, some just shuffling randomly. This is the world of most light we encounter—the light from the sun, from a lightbulb, from a flickering candle. It is a statistical jumble. To capture the character of this "fuzzy" light, we need a more sophisticated tool, one that goes beyond a single snapshot to describe the dance's overall patterns and correlations over time.

The tool we are looking for is a beautiful mathematical object called the ​​coherency matrix​​, denoted by $J$. It’s a compact $2 \times 2$ matrix that provides a complete statistical portrait of the light's polarization. Its elements are defined by time averages:

At first glance, this might seem abstract, but its meaning is quite physical.

The diagonal elements, $J_{xx} = \langle |E_x|^2 \rangle$ and $J_{yy} = \langle |E_y|^2 \rangle$, represent the average intensity of the light's electric field oscillating along the x-axis and y-axis, respectively. If you add them together, you get the sum of the diagonal elements, a quantity known as the ​​trace​​ of the matrix. This trace, $\text{Tr}(J) = J_{xx} + J_{yy}$, is nothing more than the total ​​intensity​​ of the light beam—its overall brightness. This gives us a direct, measurable anchor for our matrix.

The real magic, however, lies in the off-diagonal elements, $J_{xy}$ and $J_{yx}$. These terms measure the ​​correlation​​ between the x and y components of the electric field. Think back to our dance floor. If two dancers are performing a perfectly synchronized waltz, their movements are highly correlated; knowing where one is tells you a lot about where the other is. For light, this corresponds to a predictable phase relationship between the x and y oscillations, as in circularly or elliptically polarized light. In this case, the off-diagonal terms have a significant value. But if the dancers are all moving randomly, with no regard for one another, their motions are uncorrelated. On average, there's no fixed relationship. For light, this is the unpolarized state, and these correlation terms, $J_{xy}$ and $J_{yx}$, average to zero. This elegant $2 \times 2$ matrix packages all this rich statistical information—intensity, and the degree and nature of correlation—into just four numbers.

With this tool in hand, we can now classify any state of polarization, no matter how messy. Let's look at the three fundamental cases.

1. ​​Completely Unpolarized Light​​: This is the state of maximum randomness, like the thermal glow from an incandescent bulb. The electric field oscillates equally in all directions with no preferred orientation and no stable phase relationships. This means the intensity is evenly split, $J_{xx} = J_{yy}$, and the correlations are zero, $J_{xy} = J_{yx} = 0$. The coherency matrix takes on a particularly simple form, becoming proportional to the identity matrix: $J_{un} = \frac{I}{2} \begin{pmatrix} 1 0 \\ 0 1 \end{pmatrix}$.

2. ​​Completely Polarized Light​​: This is the state of perfect order, like our ideal laser beam. The electric field traces a stable, predictable ellipse (with linear and circular polarization as special cases). Because the state is not random, we can describe it with a single Jones vector $\mathbf{E}$, and the coherency matrix is simply $J_p = \mathbf{E}\mathbf{E}^{\dagger}$. This pure state has a remarkable mathematical signature: its ​​determinant​​ is always zero. That is, $\det(J_p) = 0$. This provides a sharp, unambiguous test for perfect polarization purity.

3. ​​Partially Polarized Light​​: This is the most general case, a mixture of order and randomness. The light coming from the sky on a clear day is a classic example. The incredible insight of this formalism is that any state of partial polarization can be uniquely described as an incoherent superposition—a simple sum—of a completely polarized part and a completely unpolarized part: $J = J_p + J_{un}$. It’s as if the light consists of a perfectly ordered beam E superimposed on a background of completely random light.

This decomposition allows us to ask—and answer—a crucial question: exactly how polarized is a given beam of light? The answer is a single number called the ​​Degree of Polarization​​, $P$. It varies from $P=0$ for unpolarized light to $P=1$ for fully polarized light. And we can calculate it directly from any coherency matrix using the wonderfully compact formula:

You can immediately see how this works. For a fully polarized beam, we know $\det(J) = 0$, so the formula gives $P=1$. For a fully unpolarized beam, a quick calculation shows that the fraction becomes 1, giving $P=0$. Using this, we can analyze any complex beam—for instance, one formed by incoherently mixing a circularly polarized beam with a linearly polarized one—and immediately quantify its overall degree of polarization. This single number, derived from the matrix, cleanly extracts the "amount of order" from the total fuzziness.

The coherency matrix, with its complex numbers, is mathematically pristine but can sometimes feel abstract. For decades, long before the coherency matrix was formalized, astronomers and physicists used an alternative (and entirely equivalent) language to describe polarization: a set of four real numbers called the ​​Stokes parameters​​.

We can think of these parameters as answers to four intuitive questions about the light beam:

• $S_0$: What is the total intensity? (This is simply our old friend, $\text{Tr}(J)$.)
• $S_1$: What is the preference for $+45^\circ$ versus $-45^\circ$ (or $135^\circ$) linear polarization?
• $S_2$: What is the preference for right-hand versus left-hand circular polarization?
• $S_3$: What is the preference for horizontal ($0^\circ$) versus vertical ($90^\circ$) linear polarization?

This set of four real numbers, $(S_0, S_1, S_2, S_3)$, provides a complete and practical description of the polarization state. You can translate between the two languages, converting coherency matrix elements into Stokes parameters and vice versa. This translation is not just a bookkeeping exercise; it reveals a profound connection between the two formalisms. One of the most elegant results from this translation is an expression for the determinant of the coherency matrix:

S_k = \mathrm{Tr}(J \sigma_k), \quad \text{for } k \in {0, 1, 2, 3}

J = \frac{1}{2} (S_0\sigma_0 + S_1\sigma_1 + S_2\sigma_2 + S_3\sigma_3)

Now that we have acquainted ourselves with the machinery of the coherency matrix, it's fair to ask a simple question: What is it good for? It may seem like we've built a rather elaborate mathematical contraption just to describe polarized light. But the true beauty of a great physical tool is not in its complexity, but in the breadth and elegance of the problems it can solve. The coherency matrix is exactly such a tool. It is a key that unlocks a remarkable range of phenomena, a Rosetta Stone that translates the subtle language of light's polarization and coherence into a simple, powerful $2 \times 2$ matrix.

Let's embark on a journey to see this matrix in action. We'll find it at the optician's bench, in the heart of interferometers, and as a messenger from distant stars and microscopic worlds.

Our first stop is the world of practical optics, where we regularly want to manipulate a beam of light. The coherency matrix gives us a precise way to predict what will happen.

Imagine you have a beam of completely unpolarized light—the chaotic, equal mix of all polarization states that comes from a typical light bulb. Its coherency matrix, up to an intensity factor, is the simple identity matrix, $J \propto \begin{pmatrix} 1 0 \\ 0 1 \end{pmatrix}$, reflecting perfect symmetry. Now, you pass this light through a real-world linear polarizer, like in a pair of polarized sunglasses. An ideal polarizer would be a perfect gate, letting one polarization through and completely blocking the orthogonal one. But real polarizers are more like leaky gates. They have a "high-transmittance" axis that lets through a fraction $p_1$ of the light polarized along it, and a "low-transmittance" axis that still leaks a small fraction $p_2$ of the light polarized along that direction.

How polarized is the light that emerges? The coherency matrix formalism gives a swift and beautiful answer. After transforming the input matrix by the Jones matrix of the polarizer, we find that the degree of polarization of the output light is given by a wonderfully simple expression:

This result from the principles in is perfectly intuitive! The polarization "purity" depends on the difference between the transmittances, normalized by the total amount of light that gets through. If the polarizer is perfect ($p_2=0$), we get $P_{out}=1$. If it's completely useless ($p_1=p_2$), we get $P_{out}=0$. The matrix did the hard work, and left us with simple physical insight.

Now, let's consider a different kind of tool: a retarder, or wave plate. Unlike a polarizer, an ideal retarder doesn't absorb light; it simply "retards" one polarization component with respect to the other, introducing a phase shift. It can change linearly polarized light into circularly polarized light, for instance. Its action is described by a unitary matrix, which is the mathematical way of saying no energy is lost. What happens to the degree of polarization when light passes through a retarder? Let's take a partially polarized beam and send it through. We apply the matrix transformation $J_{out} = M J_{in} M^\dagger$. Because the matrix $M$ for the retarder is unitary, it turns out that both the trace and the determinant of the coherency matrix remain unchanged. Since the degree of polarization is calculated from these two quantities, $P = \sqrt{1 - 4 \det(J) / (\text{Tr}(J))^2}$, an amazing thing happens: nothing! The degree of polarization $P_{out}$ is identical to $P_{in}$.

This is a profound result. A retarder can twist and turn the state of polarization in countless ways, but it cannot change its fundamental degree of polarization. It can't make unpolarized light polarized, nor can it make partially polarized light completely pure. To increase the degree of polarization, you must discard a portion of the light—you must use a filter, like our polarizer. The coherency matrix formalism beautifully separates these two actions: non-unitary transformations (like polarizers) that can alter the degree of polarization, and unitary ones (like retarders) that can only alter its form.

Light beams rarely travel alone. What happens when they meet? The coherency matrix is our guide to light's social behavior.

First, consider the simple case of mixing two completely independent, or incoherent, beams of light, like shining two separate flashlights on the same spot. Because there is no stable phase relationship between them, their fluctuations are entirely uncorrelated. In this case, the coherency matrix of the combined beam is simply the sum of the individual matrices: $J_{total} = J_1 + J_2$. It's like mixing two crowds of people; the properties of the final crowd are just the sum of the parts.

But the real magic happens when the beams are coherent, meaning they originate from the same source and maintain a stable phase relationship. Now, we can't just add the matrices. We must account for the cross-correlations between the fields, which are stored in the off-diagonal elements of the matrix. This leads us to one of the most stunning applications of our formalism: understanding interference.

Let's return to the classic Young's double-slit experiment. The appearance of bright and dark fringes is the hallmark of wave interference. The "visibility" of these fringes—how sharp the contrast is between bright and dark—is a direct measure of the coherence between the light arriving from the two slits. But what happens if we tamper with the polarization?

Suppose we place a half-wave plate, a type of retarder, in the path of the light from just one of the slits. This device can rotate the plane of linear polarization. Let's say our incident light is vertically polarized. We can orient the wave plate to rotate the polarization of the light from slit 2 by 90 degrees, making it horizontally polarized. Now, at the screen, we have vertically polarized light from slit 1 trying to interfere with horizontally polarized light from slit 2. The result? The interference pattern completely vanishes! The visibility drops to zero. Why? Because the two beams are now orthogonally polarized. From the light's perspective, they have become distinguishable. The polarization state acts as a "which-slit" marker; if you can tell which path the light took, you can't see interference. The coherency matrix formalism predicts this perfectly, showing that the fringe visibility depends directly on the correlation between the polarization components arriving from each slit. To see interference, the fields must not only be phase-coherent but also share a common polarization component.

Perhaps the most powerful application of the coherency matrix is using light as a probe. By sending light with a known polarization state out into the world and carefully measuring the coherency matrix of the light that comes back, we can deduce extraordinary things about the matter it has interacted with.

Think about the glare you see reflecting off a pond. That reflected light is strongly polarized. This is a general phenomenon: when unpolarized light reflects from a surface like water or glass, it becomes partially polarized. This is because the reflection efficiency, described by the Fresnel equations, is different for light polarized parallel to the plane of incidence ($p$-polarization) and perpendicular to it ($s$-polarization). The surface acts as a natural, if imperfect, polarizer. By measuring the full coherency matrix of the reflected light—not just its intensity—we can work backward to determine the precise properties of the surface, such as its refractive index or the thickness of any thin films on it. This is the principle behind the incredibly sensitive technique of ellipsometry, used everywhere from semiconductor manufacturing to biology.

The story doesn't end at surfaces. Light also carries information from the volume of a substance through scattering. The blue, polarized light of the daytime sky is a perfect example. Unpolarized sunlight scatters off air molecules. Because of the nature of dipole radiation, the light scattered at 90 degrees to the sun's direction becomes strongly polarized. The coherency matrix allows us to quantify this process rigorously. More than that, the polarization state of the scattered light is a sensitive fingerprint of the scattering particles themselves. If we scatter light from a dilute suspension of particles, the degree to which the scattered light is depolarized tells us about the particles' shape. Spherical particles will produce a different polarization signature than rod-shaped or anisotropic particles. This principle is used to characterize everything from atmospheric aerosols and colloidal solutions in chemistry to interstellar dust clouds in astronomy. By reading the polarization of the light, we are reading the structure of the matter it has touched.

From the design of an LCD screen (a sandwich of polarizers and liquid crystal retarders) to the analysis of light that has traveled millions of light-years, the coherency matrix provides a single, unified language. It beautifully marries the vector nature of polarization with the statistical nature of coherence, giving us a complete and powerful portrait of the character of light. It stands as a profound example of how an elegant mathematical idea can give us a clearer and deeper view of the physical world.