Abstract
The relation between the statistics of photons in input light on a detector and the measured photocounts by the detector is discussed. The averages and variances of the photocounts are compared with the averages and variances of input photons on practical detectors having quantum efficiencies of $\mu$. This comparison was made for three kinds of inputs which include Fock state light, coherent light, and thermal light. The calculations were carried out based on the combined operator model for a detector having less-than-unit quantum efficiency.
Introduction
In quantum optics and precision sensing, the accuracy of a measurement is fundamentally limited by the statistical nature of light and the imperfections of the detection system. A primary imperfection is the quantum efficiency ($\mu$), representing the probability that an incident photon triggers a detection event. When $\mu < 1$, the detector effectively "loses" information, which not only reduces the signal strength but also fundamentally alters the noise characteristics of the measured light.
This paper provides a streamlined mathematical framework to map input photon statistics to output photocounts. By employing a "combined operator model," the authors derive how detector inefficiency affects the mean and variance of three fundamental states of light: Fock states, coherent light, and thermal light.
Methodology: The Combined Operator Model
To simplify the derivation of photocount statistics, the authors model a practical detector as a composite system. Instead of directly calculating the effects of loss using complex Bernoulli distributions, they introduce an equivalent optical circuit that treats the detector's inefficiency as a spatial loss mechanism.
As shown in the figure above, the model consists of:
- A Beam Splitter: An imaginary beam splitter with a power transmission coefficient equal to the quantum efficiency $\mu$. This component "splits" the input mode $\hat{a}$ into a transmitted mode $\hat{a}'$ and a reflected "lost" mode representing the vacuum state $\hat{v}$.
- An Ideal Detector: A theoretically perfect detector ($\mu = 1$) that measures the transmitted mode $\hat{a}'$.
By defining the output annihilation operator $\hat{a}' = \sqrt{\mu}\hat{a} + \sqrt{1-\mu}\hat{v}$, the authors leverage standard quantum commutation rules to derive a general expression for the variance of photocounts, known as the Burgess variance theorem: $(\Delta n')^2 = \mu^2 (\Delta n)^2 + \mu(1-\mu)\langle \hat{n} \rangle$.
Statistical Analysis of Light States
The model is applied to three distinct types of light, each representing different statistical behaviors:
- Fock State Light: A non-classical state with a fixed number of photons ($n$). Here, the input variance is zero. The output variance becomes $(\Delta n')^2 = \mu(1-\mu)n$. The uncertainty is maximized at $\mu = 0.5$, reflecting the maximum ambiguity between detecting or missing a photon.
- Coherent Light: Typical of laser light, where variance equals the mean ($\langle \hat{n} \rangle = |\alpha|^2$). The model shows that the output remains Poissonian: $(\Delta n')^2 = \mu|\alpha|^2 = \langle \hat{n}' \rangle$. This explains why laser light is a robust reference; its statistical distribution type is invariant to detector loss.
- Thermal Light: Characterized by "photon bunching," where photons arrive in clusters. The input variance is $m + m^2$ (where $m$ is the mean photon number).
Key Results and Discussion
The impact of quantum efficiency on noise is most pronounced in thermal light. Unlike coherent light, which maintains a linear relationship between variance and efficiency, the variance of photocounts for thermal light increases quadratically with the quantum efficiency $\mu$.
The authors use Mandel’s Q-parameter ($Q = \frac{(\Delta n')^2 - \langle \hat{n}' \rangle}{\langle \hat{n}' \rangle}$) to quantify deviations from Poissonian statistics. For Fock states, $Q = -\mu$, indicating that lower efficiency "dilutes" the sub-Poissonian (non-classical) nature of the light. Conversely, for thermal light, $Q = \mu m$, showing that the "bunched" nature is preserved but scaled by the efficiency.
Conclusion
The combined operator model offers a mathematically efficient and intuitive way to understand the transformation of photon statistics in practical detection systems. While the model currently assumes a single-mode steady state and ignores factors like dark counts or dead time, its operator-based approach is highly extensible. Future work could apply this framework to more complex quantum states, such as squeezed light, to further explore the limits of quantum-enhanced sensing.
Summary updated