Interactive quantum physics · Student investigation 01 · V1.1

Quantum Measurement:
Predicting the Unpredictable

Can single photons behave randomly while their collective results follow an exact mathematical rule? Design a virtual experiment and find out.

Created by Boris Kiefer · New Mexico State University
Educational content and interactive simulator developed with AI assistance and expert validation.

QR code for Boris Kiefer's contact cardScan for contact

Source: Q12 Education, QIS Key Concepts for High School Physics (2025), pp. 18 and 22 ↗

1 · Question and prediction

Imagine a source that sends one vertically polarized photon at a time toward a polarization measurement device. At zero relative angle, the photon is accepted; at 90°, it is rejected. At angles between these extremes, individual outcomes can differ.

Make a prediction: At 45°, will exactly half of 20 photons be accepted every time? What about 1,000 photons? Record your reasoning before running the experiment.
Predict. Think about the individual outcomes and the average.
Investigate. Change the number of angular bins or photons per bin.
Explain. Compare sampled results with the mathematical prediction.

2 · Experiment A: Scan all measurement angles

Divide the measurement range from 0° to 90° into equal bins. In each bin, draw a new random angle for each photon, perform the simulated measurement, count accepted and rejected outcomes, and move to the next bin.

Exact accepted: cos² θ Exact rejected: sin² θ Sampled accepted Sampled rejected Point checker (if run)

Markers show the measured fraction in each bin at its midpoint. Exact curves are continuous; individual photon angles are randomized within each bin. Each bin contains the requested number of photons.

Why might bin markers differ slightly from the exact midpoint curve, even with many photons?

A photon in a bin can have any angle inside that bin. With many photons, its measured acceptance fraction approaches the average of cos² θ across the bin—not exactly the value at the midpoint. Increasing the number of bins narrows the bin and reduces this difference. Increasing the number of photons per bin reduces random fluctuations around the bin average.

Export includes counts, measured fractions, exact midpoint values, and exact bin averages.

3 · Experiment B: Test any angle

Pick one angle, draw a fixed number of photons at that exact angle, and compare your results with the theoretical probabilities. Repeat at the same angle: do you get identical measured fractions?

The point checker also places measured markers on the angular-scan graph. It uses the current seed followed by a new seed on each repeated run; each point run is independent of the angular scan.

4 · Interpret your experiment

Explore

  1. Repeat the scan with the same bins but 10 times more photons per bin. What changes?
  2. Keep the photon count per bin constant, but increase the number of bins. What changes?
  3. Use the point checker at 45°. Why aren't the accepted and rejected counts necessarily equal?
  4. Try 0° and 90°. What is special about these angles?

Your explanation

Reveal the science: probability and the measurement basis

For a vertically polarized input, a measurement axis rotated by θ has two possible outcomes. Their probabilities are Paccepted = cos² θ and Prejected = sin² θ. These add to one. Every photon produces one outcome; repetition reveals the probability. Changing the measurement angle changes the measurement basis and therefore the outcome probabilities.

Honors/AP extension: Connect the result to vectors and the Born rule

In an orthonormal polarization basis aligned with the measuring device, the initial polarization state has probability amplitudes cos θ and sin θ (up to a sign convention). The Born rule says to square their magnitudes, yielding cos² θ and sin² θ. The statistical uncertainty of a sampled accepted fraction is approximately √[p(1−p)/N] at a fixed angle. For random angles within a bin, replace p by the bin-averaged probability.

Educator notes: model boundaries and suggested evaluation

Idealized model: single photons prepared in the same pure polarization state, two mutually exclusive projective measurement outcomes, perfect preparation and detection, no dark counts or optical loss, independent trials. It is a sampling model—not a detailed hardware simulator. Suggested pilot: two prediction prompts before use, a short post-activity explanation, and one transfer question at an unfamiliar angle. NGSS links are supporting connections, not assertions that the entire performance expectations are met.