Wavefront shaping controls the spatial phase, amplitude or both of an optical field to produce a chosen output. Through a scattering material, that control can make contributions arrive in phase at a target and form a focus from an otherwise spread-out pattern.
The gold standard names the desired output, the controlled input and the measurement that establishes improvement. A brighter target spot is useful evidence for that target; claims about total transmission, imaging or a changing material need their own checks.
eduKateSG · Wavefront Shaping
Find your next learning step
Begin with two interfering fields, follow a controlled phase change and test what a successful focus actually demonstrates.
Full chapter index · Practice and transfer · Education root · How X Works
Full chapter index
Define the control · Chapters 1–2
Work the interference · Chapters 3–4
Build a reliable experiment · Chapters 5–7
Model and transfer · Chapters 8–10
Choose further reading · Chapters 11–12
CHAPTER 1 OF 12 · Define the control
1. Choose the output before optimising the input
An optical control problem needs a stated objective. Examples include maximising intensity at one detector location, distributing light across several targets or reproducing a field pattern. These objectives can favour different inputs.
Record the wavelength, illuminated area, selected polarization and target definition. A system adjusted for one location or wavelength should be evaluated again when those conditions change.
For a learner, sketch an input plane, the optical system and the measured output plane. Mark which quantities can change and which are held fixed. That drawing is the first check on whether the claimed improvement answers the original question.
In coherent illumination, contributions following different paths can combine with stable relative phases over the relevant measurement period. Their interference produces a structured intensity pattern, often called speckle.
Vellekoop and Mosk demonstrated focusing through strongly scattering samples by controlling the incident wavefront. The experiment shows that a complicated path distribution can remain controllable under suitable conditions.
Do not treat a complicated pattern as a list of independent brightness values. The relative phases are part of the explanation. Changing an input phase can redistribute output intensity even when the input amplitude is unchanged.
Take two equally scaled coherent field contributions at a target: E₁ = 1 and E₂ = exp(iφ). The local intensity in these arbitrary units is I = |E₁ + E₂|² = 2 + 2cosφ.
At φ = 0, I = 4. At φ = π/2, I = 2. At φ = π, I = 0. The contributions have the same individual magnitudes in all three cases; their relative phase changes the sum.
These values describe one target location, not the total output power of a complete optical system. Include that boundary when explaining constructive interference. The calculation does not create extra energy.
CHAPTER 4 OF 12 · Work the interference
4. Worked example: unequal contributions limit cancellation
Now let the field magnitudes be 1 and 0.5. Then I = |1 + 0.5exp(iφ)|² = 1.25 + cosφ. The maximum is 2.25 and the minimum is 0.25.
Opposite phase does not give zero because the magnitudes differ. The smallest possible resultant has magnitude 1 − 0.5 = 0.5, whose square is 0.25.
This gives a useful diagnostic: an intensity minimum alone does not establish perfect phase control or equal amplitudes. Before blaming an optimiser, check which fields are actually being combined and what the model predicts for their magnitudes.
CHAPTER 5 OF 12 · Build a reliable experiment
5. Feedback connects an adjustment with a measured response
A feedback approach varies the controlled input, measures a target signal and selects a setting that improves the stated objective. For phase control, the experiment must connect commanded settings to the optical phase actually applied.
Use a consistent measurement interval and account for changes in illumination or detector response. Otherwise a brighter reading could reflect source drift or a changed acquisition rather than a better wavefront.
A practical teaching exercise uses supplied intensity measurements rather than optical equipment: compare several phase settings, identify the largest repeatable target response and explain which uncontrolled change could imitate the result.
CHAPTER 6 OF 12 · Build a reliable experiment
6. A phase setting needs an optical calibration
A spatial light modulator applies a controllable optical change across parts of a beam. Its response depends on the device and operating conditions. A numerical command is not, by itself, a guaranteed phase in radians.
The useful calibration relates the commands, wavelength and polarization to the response used in the experiment. Also distinguish phase-only control from control of both phase and amplitude; their accessible inputs differ.
Make the explanation auditable by naming the controlled variables. If the optimisation changes both brightness and phase, a comparison described as fixed-power phase-only control would need additional evidence.
| Stage | Record | Check |
|---|---|---|
| Objective | Target position or pattern | What counts as success? |
| Control | Phase, amplitude and input power | Which variables changed? |
| Evidence | Target signal and comparison | Was the measurement comparable? |
| Transfer | New target or later time | Does the earlier calibration still apply? |
CHAPTER 7 OF 12 · Build a reliable experiment
7. The medium must stay relevant to its calibration
A measured relation between input and output can become outdated when a scattering sample moves or the optical alignment changes. Research on fast wavefront shaping explicitly addresses changing media.
Consider the equal-field example after a successful φ = 0 setting. If an uncontrolled relative phase change of π/2 develops, the target intensity drops from 4 to 2 even though the saved command is unchanged.
Separate acquisition time, computation time and application time. The useful question is whether the optical relation remains valid across that whole interval. A high device update rate alone does not establish successful tracking of a changing specimen.
For a fixed linear coherent system, a transmission matrix relates input field components to output components. Consider just one normalised target row: y = (x₁ + ix₂)/√2.
For an input x = (1,1)/√2, the input squared norm is 1 and y = (1 + i)/2, so |y|² = 0.5. Choose x = (1,−i)/√2 instead. Its squared norm remains 1, while y = 1 and |y|² = 1.
The second input compensates the relative phase in this toy row. The calculation demonstrates the target result under the given model. A real measured matrix also has finite sampling, noise and a defined set of accessible input and output modes.
CHAPTER 9 OF 12 · Model and transfer
9. Enhancement and efficiency need different denominators
Suppose a target measurement rises from a reference mean of 12 counts to 180 counts with matched exposure and illumination. The target enhancement is 180/12 = 15. Define how the reference mean was obtained.
That ratio does not say that 15 times the incident power crossed the specimen. To estimate a power fraction, measure compatible power quantities and define the target collection region. For invented values of 15 mW in that region and 100 mW incident, the fraction is 0.15.
When comparing experiments, retain the target size, background definition, power and exposure conditions. A large enhancement against a small baseline can coexist with a modest collected power fraction.
First, use equal unit field magnitudes with φ = 2π/3. Second, find the maximum and minimum intensities for magnitudes 2 and 1. Third, explain why a phase pattern for one output location may fail at another.
Checks: I = 2 + 2cos(2π/3) = 1. For magnitudes 2 and 1, the extrema are (2 + 1)² = 9 and (2 − 1)² = 1. The new target generally has different input-to-output phase relations.
Transfer the reasoning to a moving scattering sample. Identify which measured relation could have changed, then propose a fresh comparison. The strongest answer traces the failure to a specific assumption rather than simply requesting more controlled pixels.
CHAPTER 11 OF 12 · Choose further reading
11. Connected optics and mathematical guides
Build the field explanation
- Wavefront Sensing — distinguish measurement of a wavefront from control of an output.
- Optics — revisit interference, propagation and optical performance.
Follow control and reconstruction
- Photonics — connect sources, modulators, detectors and optical systems.
- Gerchberg–Saxton phase retrieval — compare phase estimation under constraints with an optical control objective.
Choose the sensing route when the missing step is how an optical error was measured. Choose the reconstruction route when intensity information must be connected with an inferred phase.
CHAPTER 12 OF 12 · Choose further reading
12. Primary research and the next learning step
- Vellekoop and Mosk: focusing through scattering media — the original experimental demonstration of controlled focusing.
- Popoff and colleagues: the transmission-matrix approach — field mapping, phase conjugation and reconstruction assumptions.
- Tzang and colleagues: fast wavefront shaping — an experimental approach addressing dynamic scattering media.
The numerical examples are original, idealised teaching cases. Their arbitrary field units and small matrices explain interference and control; they are not specifications for a particular modulator or specimen.
Explain one successful control result in four steps: the target, the adjustable field, the measured improvement and the conditions under which it persists. Continue through the Education root to practise transferring that chain to other systems.