Introduction to the Matrix Representation of Optical Systems

When analyzing optical systems through traditional geometric optics, tracing light paths through complex assemblies usually relies on trigonometric relations and tedious algebraic calculations. As modern instruments grow increasingly intricate—incorporating multiple lenses, mirrors, and spacing elements—conventional ray-tracing methods quickly become cumbersome and prone to error. To systematically and structurally describe how light transforms as it travels through an optical layout, matrix representation, commonly known as matrix optics, was introduced to the field. By harnessing the power of linear algebra, this approach abstracts the propagation and refraction of light into straightforward matrix multiplication, offering an exceptionally efficient and standardized mathematical framework for optical design and analysis.

The core philosophy of matrix optics treats light rays as state vectors within a linear system, while individual optical components or propagation spaces act as linear operators modifying those vectors. Under the paraxial approximation—assuming light rays maintain sufficiently small angles relative to the optical axis so that trigonometric sines can be approximated by their radian values—the behavior of light strictly adheres to linear principles.
Within the meridian plane (the plane containing the optical axis), the trajectory of a light ray at any given point is uniquely determined by two fundamental parameters:

  • The height of the ray from the optical axis, denoted as $y$.
  • The angle (or the tangent of the angle) the ray makes with the optical axis, denoted as $\theta$.

These two parameters form the column vector representing the ray's state: $\begin{bmatrix} y \ \theta \end{bmatrix}$. As the ray traverses an optical element or propagates through space, its state changes linearly. This transformation is captured by a $2 \times 2$ transfer matrix $M$. Letting the input state be $\begin{bmatrix} y_1 \ \theta_1 \end{bmatrix}$ and the output state be $\begin{bmatrix} y_2 \ \theta_2 \end{bmatrix}$, their relationship is expressed as:

$$
\begin{bmatrix} y_2 \ \theta_2 \end{bmatrix} = M \begin{bmatrix} y_1 \ \theta_1 \end{bmatrix} = \begin{bmatrix} A & B \ C & D \end{bmatrix} \begin{bmatrix} y_1 \ \theta_1 \end{bmatrix}
$$

Here, the matrix $M = \begin{bmatrix} A & B \ C & D \end{bmatrix}$ is known as the Ray Transfer Matrix or the ABCD matrix. Each element carries distinct physical dimensions and meanings: $A$ and $D$ are dimensionless, representing linear magnification of height and angular magnification, respectively; $B$ has the dimension of length, indicating the dependence of ray height on the input angle; and $C$ has the dimension of inverse length, reflecting the dependence of the output angle on the input height, which is directly tied to the optical power of the component.

Fundamental Operations in Matrix Optics

The primary advantage of matrix optics lies in its ability to uniformly compare how different physical mechanisms affect light using a shared mathematical language. In geometric optics, basic light propagation relies on free-space transmission, refraction, and reflection. Within the matrix framework, these phenomena are abstracted into distinct characteristic ABCD matrices.

  • Free-Space Propagation: Light travels in a straight line through a uniform medium. After propagating over a distance $d$, the height increases by $d \cdot \theta$, while the angle remains unchanged. Its transfer matrix is:
    $$
    M_{space} = \begin{bmatrix} 1 & d \ 0 & 1 \end{bmatrix}
    $$
    The defining feature of this matrix is its lower-off-diagonal element $C = 0$, signifying that free space alone cannot alter ray angles—meaning it possesses no focusing or diverging power.

  • Refracting Interface: Light crosses a spherical refracting boundary with a radius of curvature $R$. Governed by paraxial refraction laws, the ray height remains unchanged during refraction, whereas the angle change relates to the surface's optical power. Its transfer matrix is:
    $$
    M_{refract} = \begin{bmatrix} 1 & 0 \ -\frac{n_2-n_1}{n_2 R} & \frac{n_1}{n_2} \end{bmatrix}
    $$
    Where $n_1$ and $n_2$ are the refractive indices of the incident and transmission media, respectively. This matrix features an upper-off-diagonal element $B = 0$, highlighting that refraction occurs instantaneously at a single spatial location.

  • Reflecting Interface: Light reflects off a spherical mirror. In matrix optics, reflection is treated as a special case of refraction by setting $n_2 = -n_1$. Its transfer matrix is:
    $$
    M_{reflect} = \begin{bmatrix} 1 & 0 \ -\frac{2}{R} & 1 \end{bmatrix}
    $$
    This unified formulation allows optical designers to seamlessly process mixed catadioptric systems containing both lenses and mirrors within a single computational pipeline.

Cascading Systems and Matrix Multiplication

Real-world optical instruments typically consist of multiple elements and intervals arranged in sequence. When handling such cascaded systems, matrix representation follows the rules of linear algebra: the total transfer matrix of the system is given by the product of the individual component matrices in reverse order.

Assuming an optical system comprises Element 1, Element 2, and Element 3 in sequence, with individual transfer matrices $M_1$, $M_2$, and $M_3$, a light ray passes through Element 1 first and Element 3 last. The total system matrix $M_{total}$ is calculated as:

$$
M_{total} = M_3 \cdot M_2 \cdot M_1
$$

This cascading property compresses highly complex optical trains into a single $2 \times 2$ matrix. Once obtained, this total matrix immediately reveals the macro-imaging properties of the system without the need to track intermediate ray paths step by step. For instance, if the $C$ element of the total matrix equals zero, the setup functions as a telescopic system; if the $B$ element is zero, the input and output planes form a pair of conjugate object-image planes.

System-Level Evaluation and Broad Applications

Beyond basic ray tracing, matrix representation serves as the foundation for deriving universal, system-level principles. Its deployment spans several critical domains:

  • Equivalent Focal Length Calculation: For multi-lens compound systems, calculating the $C$ element of the total matrix directly yields the system's equivalent focal length and back focal length, facilitating the evaluation of light-gathering capabilities.
  • Gaussian Beam Transformation: Although originating in geometric optics, the mathematical structure of matrix optics applies equally to paraxial Gaussian beam propagation in wave optics. Using the ABCD law, engineers can swiftly calculate the beam waist position and spot size of laser beams within complex resonators or delivery paths.
  • Optical Resonator Stability Criteria: In laser technology, cavities are formed by mirrors separated by free space. By deriving the round-trip transfer matrix, the magnitude of the matrix trace ($A+D$) determines cavity stability based on whether its absolute value is less than 2—representing one of the most classic system-level evaluations in matrix optics.

In summary, the matrix representation of optical systems abstracts the laws of light propagation, refraction, and reflection into the unified framework of ABCD matrices using linear algebra. By eliminating tedious intermediate geometric calculations, it empowers designers to evaluate and optimize the macro-characteristics of complex optical instruments from a high-level perspective, making it an indispensable foundation of modern optical engineering.