Relationship Between Coherence Length and Interference Visibility

In optical interference experiments, the crispness and definition of fringe patterns rely on more than just the physical intersection of two light beams; they require a sustained, stable phase relationship. The coherence length serves as the spatial manifestation of a light source's temporal coherence, whereas interference visibility acts as the direct, quantifiable metric of this coherence evaluated through fringe contrast. Grasping the interplay between these two concepts is fundamental for designing and troubleshooting advanced optical systems, including Michelson interferometers, optical coherence tomography (OCT) devices, and Fourier-transform spectrometers.

Real-world light sources deviate from the idealized model of infinite monochromatic waves; instead, they possess a finite spectral bandwidth. The emission can be conceptualized as a succession of finite wave trains possessing a mean duration known as the coherence time ($\tau_c$). In a vacuum, the coherence length ($L_c$) is formally expressed as:

$$
L_c = c\tau_c
$$

where $c$ denotes the speed of light in a vacuum. The coherence time is inversely proportional to the spectral frequency width ($\Delta\nu$):

$$
\tau_c \approx \frac{1}{\Delta\nu}
$$

Combining these relations yields the conventional engineering expression:

$$
L_c \approx \frac{c}{\Delta\nu} = \frac{\lambda_0^2}{\Delta\lambda}
$$

In this formulation, $\lambda_0$ represents the central wavelength, and $\Delta\lambda$ signifies the spectral bandwidth. This relationship demonstrates that narrower spectra yield longer coherence lengths, whereas broader spectra restrict coherence. While an idealized monochromatic source approaches $\Delta\lambda \to 0$ with an infinite coherence length, broadband emitters like white-light LEDs typically exhibit coherence lengths limited to the micrometer scale.
Interference visibility, frequently referred to as fringe contrast, quantifies how distinctly the bright and dark fringes are delineated:

$$
V = \frac{I_{\max} - I_{\min}}{I_{\max} + I_{\min}}
$$

For a general two-beam interference scheme where the respective arm intensities are $I_1$ and $I_2$, and the magnitude of the complex degree of coherence is denoted by $|\gamma|$, the visibility is given by:

$$
V = \frac{2\sqrt{I_1 I_2}}{I_1 + I_2}|\gamma|
$$

When the beam intensities are balanced ($I_1 = I_2$), the visibility simplifies directly to $V = |\gamma|$. Under these ideal equal-intensity conditions, interference visibility serves as a direct proxy for source coherence. A visibility of $V = 1$ indicates maximum fringe contrast, while $V = 0$ implies complete fringe washout.

Impact of Optical Path Difference on Visibility

Within a classic Michelson interferometer setup, the introduction of an optical path difference ($\Delta L$) between the two arms profoundly alters the fringe pattern. When $\Delta L = 0$, the interference patterns produced by all constituent wavelength components overlap constructively in phase, producing maximal visibility. As $\Delta L$ increases, differential phase shifts across distinct wavelengths lead to destructive superposition, washing out the distinct bright-and-dark transitions. Once $\Delta L$ approaches or exceeds the coherence length $L_c$, the fringes blur and eventually disappear.

The functional dependence of visibility on the path difference can be modeled as:

$$
V(\Delta L) = |\gamma(\Delta L/c)|
$$

The exact mathematical profile of $\gamma$ is dictated by the spectral lineshape of the source:

  • Rectangular Spectrum:
    $$
    V(\Delta L) = \left|\operatorname{sinc}\left(\frac{\pi \Delta\nu \Delta L}{c}\right)\right|
    $$
    Here, the first extinction zero occurs precisely at $\Delta L = c/\Delta\nu = L_c$.

  • Gaussian Spectrum:
    $$
    V(\Delta L) = \exp\left[-\left(\frac{\pi \Delta\nu \Delta L}{2\sqrt{\ln 2},c}\right)^2\right]
    $$
    In this scenario, visibility decays smoothly. Adopting the convention where the coherence length is marked by a drop to $1/e$, the Gaussian coherence length evaluates to roughly $0.53\lambda_0^2/\Delta\lambda$, though $\lambda_0^2/\Delta\lambda$ remains the standard order-of-magnitude estimator in engineering practice.

Practical Source Profiles

  • White-Light LEDs: Characterized by a central wavelength of roughly $550,\text{nm}$ and a broad spectral width of $30,\text{nm}$, yielding:
    $$
    L_c \approx \frac{550^2}{30},\text{nm} \approx 10,\mu\text{m}
    $$
    Consequently, white-light interference fringes remain observable only within an extremely narrow window around zero path difference.

  • Sodium Vapor Lamps: Emitting closely spaced doublet lines near $589.0,\text{nm}$ and $589.6,\text{nm}$ with an effective span of $0.6,\text{nm}$, producing:
    $$
    L_c \approx \frac{589^2}{0.6},\text{nm} \approx 0.58,\text{mm}
    $$
    In practice, the visibility envelope exhibits periodic beats due to the dual-line frequency mixing.

  • He-Ne Lasers: Operating with an ultra-narrow spectral width around $0.002,\text{nm}$ delivers:
    $$
    L_c \approx \frac{632.8^2}{0.002},\text{nm} \approx 0.2,\text{m}
    $$
    Highly stabilized single-frequency lasers can extend this coherence length to several meters or more.

Measurement Techniques and Applications

A Michelson interferometer offers a straightforward apparatus for measuring coherence length. By scanning the movable mirror to alter the optical path difference while monitoring the resulting fringe visibility, an operator can identify the threshold where visibility drops to a benchmark value (such as $1/e$ or $0.5$). The corresponding path offset yields $L_c$, which can then be inverted via $L_c \approx \lambda_0^2/\Delta\lambda$ to deduce the spectral bandwidth of the source.

In modern applications, coherence length dictates the spatial "coherence gate" of an optical setup. Optical coherence tomography (OCT), for instance, harnesses low-coherence light sources so that interference occurs exclusively when path matching falls within a microscopic depth window, enabling high-resolution cross-sectional biological imaging. Conversely, standard precision interferometry requires strict path matching where $\Delta L < L_c$; otherwise, fringe contrast vanishes, preventing accurate phase demodulation.

Common Misconceptions

  • The coherence length is not a physical dimension of the source emitter; rather, it is a statistical metric describing temporal coherence properties along the propagation axis.
  • A drop in fringe visibility does not signify a total loss of light intensity; it merely reflects a reduction in contrast between the maxima and minima of the interference pattern.
  • Extrinsic factors—such as unequal arm intensities, polarization mismatch, mechanical vibrations, and air turbulence—can severely degrade fringe visibility without altering the intrinsic coherence length of the source itself.
  • While narrower spectral widths generally promise longer coherence lengths, practical laser performance is frequently limited by frequency jitter, phase noise, and longitudinal mode structures beyond the nominal linewidth.

Ultimately, while the coherence length is fundamentally governed by the spectral profile of the light source, the interference visibility acts as the tangible, measurable bridge that translates this microscopic phase stability into macroscopic, observable fringe patterns.