Light as a Probability Wave — Question 9

Section 38-4 · Physics pages 6328–6329

Question 9 Fixed photon number produces binomial rather than Poisson partition noise Exactly N = 400 independent photons enter an interferometer. Each reaches detector D with probability p = .250. Detection is ideal, and K is the count at D. Use the binomial model. a b + − Schematic; labels, not drawing lengths, specify the data. (a) Find the expected count and its standard deviation. (b) Find the standard deviation of the estimated fraction bp = K/N and the covariance with the complementary detector count.
Show solutionHide solution
Question 9 - Solution Fixed photon number produces binomial rather than Poisson partition noise a b + − (a) Method and calculation ⟨K⟩= Np = 100 , Var K = Np(1 −p) = 75. σK = √ 75 = 8.66025 . (b) Interpretation and further calculation σbp = q p(1 −p)/N = .0216506 . L = N −K ⇒ Cov(K, L) = −Var K = −75 . Physical check The two output counts are anticorrelated because their sum is fixed. Poisson input fluctuations would define a different statistical experiment.

Question 9

Fixed photon number produces binomial rather than Poisson partition noise

Exactly N=400N=400 independent photons enter an interferometer. Each reaches detector D with probability p=.250p=.250. Detection is ideal, and K is the count at D. Use the binomial model.

Diagram shown in the original SVG worksheet.

  1. Find the expected count and its standard deviation.

  2. Find the standard deviation of the estimated fraction p̂=K/N\widehat p=K/N and the covariance with the complementary detector count.

Question 9 - Solution

Fixed photon number produces binomial rather than Poisson partition noise

Diagram shown in the original SVG worksheet.

(a) Method and calculation

⟨K⟩=Np=100,Var⁡K=Np(1−p)=75.\boxed{\langle K\rangle=Np=100},\quad\operatorname{Var}K=Np(1-p)=75.σK=75=8.66025.\boxed{\sigma_K=\sqrt{75}=8.66025}.

(b) Interpretation and further calculation

σp̂=p(1−p)/N=.0216506.\boxed{\sigma_{\widehat p}=\sqrt{p(1-p)/N}=.0216506}.L=N−K⇒Cov⁡(K,L)=−Var⁡K=−75.L=N-K\quad\Rightarrow\quad\boxed{\operatorname{Cov}(K,L)=-\operatorname{Var}K=-75}.

Physical check

The two output counts are anticorrelated because their sum is fixed. Poisson input fluctuations would define a different statistical experiment.