How Science Works | How Thermal Radiation Reduces Matter-Wave Interference

In 2004, Hackermüller and colleagues varied the internal heating of C70 molecules before matter-wave interference and compared fringe visibility with a microscopic thermal-emission model. Contrast decreased as heating increased. Counts initially rose because detection efficiency depended on internal energy, then fell under stronger heating. The researchers also checked that relevant ions and fragments from the heating region were not recorded by the downstream detector. Original Nature study, author manuscript.

The interesting question is therefore not simply whether a detector registered “more” or “less.” It is whether the distribution retained its modulation across the scanned interference pattern, and what physical process explained the change.

An interference pattern is assembled statistically from many detections. A single molecular registration is a local event; the pattern is a property of the distribution accumulated under a preparation and measurement procedure. We should not describe one event as carrying a miniature visible fringe pattern.

For a reading diagram, draw three boxes: molecular preparation, interferometric evolution and detection. Put a separate arrow from the evolving molecules to emitted radiation. This separates a change in the prepared ensemble, a coupling during evolution and a change in detector response.

That separation is the organising principle of the article. Any explanation of a heating-dependent curve should say which box it concerns and which measurement constrains it. Otherwise one changed control setting can silently stand in for several different physical changes.

Contents · Next chapter

Fringe visibility is commonly written V = (Nmax − Nmin)/(Nmax + Nmin), when maximum and minimum counts have comparable exposure and the appropriate detector and background treatment. It summarises modulation relative to the overall signal scale.

Our original teaching exercise uses a bright-position count of 120 and a dark-position count of 40. The difference is 80, the sum is 160, and V = 0.50. These numbers are fictional and are not digitised measurements from the C70 paper.

For the exercise, assume equal exposure at the two positions, stable preparation, phase-independent detection efficiency and no background. Also assume the positions correspond to genuine maxima and minima rather than selecting the largest and smallest noisy bins after looking at the data.

The pair’s sum is a convenient local intensity-scale indicator. It is not automatically the total number detected in a complete scan. A full experimental count-rate analysis must use all relevant positions, exposures and acquisition times.

Now record two columns in a paper worksheet: the normalised modulation and the overall signal scale. A mechanism can change either, both or neither. Keeping separate columns prevents “the curve is lower” from becoming an ambiguous statement about coherence.

The visibility formula itself is simple. The scientific work lies in making sure that the counts being inserted describe comparable conditions and that the chosen summary answers the question we actually want to ask.

Contents · Previous chapter · Next chapter

Suppose a hypothetical phase-independent loss process removes half the molecules at every position, while all other assumptions remain unchanged. The expected bright and dark counts become 60 and 20. Their difference halves to 40 and their sum halves to 80, so V remains 0.50.

The beam has become weaker, but the normalised modulation has not changed. This is why a reduced count rate alone does not demonstrate loss of interference contrast.

Now consider a different original pair: 90 and 70. Its sum is still 160, but its difference is only 20, giving V = 0.125. The contrast has fallen even though this pair’s overall scale matches the first one.

These two transformations answer different questions. Multiplying the entire pattern by a constant is attenuation. Reducing the modulated part relative to its baseline changes visibility. A real heating experiment can involve both, and the detector can introduce additional changes.

Uniformity is the crucial assumption. If losses depend on position, velocity class or another variable correlated with the interference response, the surviving ensemble can have a different contrast. The half-count example does not prove that every form of beam loss preserves visibility.

A useful control would therefore test whether the detection and transmission changes are independent of the scanned fringe position and relevant molecular subensembles. The result should be stated at that level of specificity rather than reduced to “loss is never decoherence” or “loss always washes out fringes.”

Contents · Previous chapter · Next chapter

Although uniform attenuation preserves the ideal visibility, it usually reduces statistical precision for a fixed measurement duration. Fewer independent counts provide less information about the underlying rates.

For an original large-count approximation, treat bright and dark counts A and B at known phases as independent Poisson variables. Propagating their count variances through V = (A − B)/(A + B) gives Var(V) approximately 4AB/(A + B)³.

For 120 and 40, the estimated standard uncertainty is about 0.0685. For 60 and 20, it is about 0.0968. The central visibility is still 0.50, but halving the counts increases this approximate uncertainty by √2.

For the 90-and-70 pair, the same approximation gives about 0.0784 around V = 0.125. Reporting more decimal places in the calculated ratio does not make the evidence more precise.

This formula is not a complete error budget for a fitted experimental scan. It ignores background uncertainty, correlated drift, fitted phase, detector nonlinearity and selection effects. At low counts or near boundaries, the ratio’s sampling distribution can make simple symmetric approximations unreliable.

It is particularly important not to select the noisiest high and low bins and then treat them as independently known extrema. Fitting a justified model to the full scan, preserving residuals and repeating measurements can provide a more defensible estimate.

The lesson for the thermal-radiation case is that count changes matter even when they do not explain a visibility change: they affect how strongly that change can be established.

Contents · Previous chapter · Next chapter

Add a hypothetical uniform background of 20 counts to each member of the original 120-and-40 pair. The observed values become 140 and 60. Their difference remains 80, but their sum is now 200, so the apparent visibility is 0.40.

If the background is independently known and correctly subtracted, the original 0.50 is recovered. In practice, estimating the background adds uncertainty, and assuming a constant background when it varies can bias the result. Subtraction is a modelled correction with evidence, not a cosmetic operation.

A phase change requires another distinction. A deterministic shift of a complete ideal fringe pattern changes the location of its maxima and minima, while leaving its visibility unchanged. Sampling only the old bright and dark positions can falsely suggest a reduced contrast after the shift.

If repeated scans have different unrecorded phases and are averaged without alignment, the averaged pattern can have lower visibility. For an original example, average equally weighted patterns with phase offsets +θ and −θ. Their cosine modulation is multiplied by cosθ, so visibility is multiplied by |cosθ| under equal-baseline assumptions.

At θ = 60°, an initial visibility of 0.50 becomes 0.25 in the average, although each individual ideal scan still has visibility 0.50.

This is not a claim that every process called dephasing is categorically separate from decoherence. It identifies one specific averaging mechanism that must be considered when interpreting data. Which physical variables fluctuated, whether they were recorded and how the analysis averages over them determine the appropriate account.

Contents · Previous chapter · Next chapter

A thermal photon leaving a molecule can carry information related to where the emission occurred. If different molecular spatial alternatives produce distinguishable environmental states, interference available in the molecular signal can be reduced when those environmental degrees of freedom are not retained in the description.

No person needs to watch the photon. The relevant issue is the physical correlation between the molecule and its environment, not conscious attention or the decision to read a display.

In a simplified two-alternative picture, label the corresponding environmental states E₁ and E₂. Their quantum overlap describes how distinguishable they are in this model. Identical environmental states preserve the interference term; orthogonal states remove that term from the reduced molecular description.

A real interferometer has a richer geometry than two fixed paths. The original theoretical treatment of Talbot–Lau interference explicitly connects environmental coupling with observable pattern changes while accounting for the interferometer. It also compares quantum and classical descriptions rather than treating any periodic signal as automatic proof of quantum interference. Hornberger, Sipe and Arndt, Physical Review A.

For a learner’s diagram, show two spatial alternatives and two outgoing radiation states. Annotate whether the radiation states strongly overlap or are readily distinguishable. Do not draw every photon as an ideal camera with unlimited spatial resolution.

The wavelength relative to the relevant separation, the emission geometry and the emission time matter. Radiation that scarcely distinguishes the alternatives need not destroy the whole pattern merely because energy was emitted.

Contents · Previous chapter · Next chapter

Consider an original deliberately simplified model. Start with visibility V₀ = 0.50. Assume each independent emission multiplies the relevant interference term by a real factor g between zero and one, and changes neither the baseline counts nor the phase. The factor represents retained coherence for this exercise.

After exactly n emissions, the visibility is Vn = V₀gⁿ. With g = 0.80 and exactly two emissions, Vn = 0.50 × 0.80² = 0.32. This is not a fitted molecular emission spectrum.

Now let the emission number fluctuate according to a Poisson distribution with mean ν. Averaging gⁿ over that distribution gives V = V₀ exp[−ν(1 − g)]. With ν = 2 and g = 0.80, the result is approximately 0.3352.

The value differs from the exactly-two-emission result because the ensemble includes zero, one, two and larger emission counts. Replacing a fluctuating count by its mean inside a nonlinear expression does not generally give the correct average.

If g = 0, a single emission completely removes the relevant interference term in this toy model. The ensemble still has visibility V₀ exp(−ν), because a fraction exp(−ν) experiences no emission. At ν = 2, the visibility is approximately 0.0677, rather than exactly zero.

If g = 1, emissions leave this interference term unchanged regardless of ν. These limits explain why photon number alone cannot determine contrast loss. Real photons have varying wavelengths and emission positions, so a single constant g is only an instructional reduction of a more detailed environmental calculation.

Contents · Previous chapter · Next chapter

The C70 study’s model accounted for emission wavelengths, emission position, molecular cooling and path separation; it did not treat the molecules as simple fixed-temperature equilibrium blackbodies. Heating and temperature were connected through calibration and modelling. The authors compared different velocity/heating arrangements and noted alignment-related variability. These details limit how a heating-dependent curve can be generalised. Hackermüller and colleagues, author manuscript.

A useful evidence map has three links. First connect the external preparation setting with the molecular ensemble that actually reaches the interferometer. Next connect the ensemble and radiation coupling with a predicted contrast change. Finally connect the molecular distribution with the detector’s registered counts.

Each link can have independent uncertainty. A good fit at the final step is more informative when intermediate quantities are constrained by other observations instead of being freely adjusted to follow the same curve.

For an original critical-reading question, imagine two theories that both predict a lower visibility as a control setting increases. Theory A predicts the same reduction for every transit time; theory B predicts a time-dependent reduction tied to the opportunity for emission. Comparing several independently characterised transit conditions can distinguish predictions that one curve cannot.

This is a proposed reasoning comparison, not a summary of an unreported experiment or a claim that every environmental process depends on time in that simple way. Its purpose is to move beyond qualitative agreement.

The strongest conclusion is the one the evidence supports: thermal radiation can reduce the observable molecular matter-wave interference in the studied setting. It is not a claim that the molecules cease to obey quantum mechanics.

Contents · Previous chapter · Next chapter

Use four columns on paper: observation, candidate cause, additional check and remaining limitation. Place the fictional count pairs in the first column, keeping them separate from published observations.

For 120/40 changing to 60/20, a candidate explanation is phase-independent attenuation. A useful check is whether the whole scan scales uniformly and whether exposure and detector response remain comparable. The remaining limitation is that selective loss among unresolved subensembles could change more than the overall scale.

For 120/40 changing to 90/70, the observation is lower normalised contrast at the same two-point scale. Candidate causes include environmental correlation, unrecorded phase fluctuations, background changes or altered preparation. The arithmetic does not choose among them.

A complete scan with stable phase references can test the averaging explanation. Background measurements can constrain additive counts. Independent preparation information can test whether the same molecular ensemble is compared. A quantitative environmental model can then be checked against the residual pattern and its controlled dependence.

Keep systematic and statistical uncertainties distinct. More counts can reduce random uncertainty while leaving an incorrect background model or a drifting phase reference untouched. Repetition should include deliberate changes that expose these possibilities, rather than merely collecting the same biased measurement longer.

Finally, avoid treating one successful environmental explanation as a solution to every philosophical question about measurement. The experiment tests a physical coupling and a distribution of outcomes. It does not require consciousness as a mechanism, nor does this visibility analysis select a unique interpretation of quantum theory.

Contents · Previous chapter · Next chapter

Begin with the fictional 120/40 pair and apply three separate transformations: multiply both by one-half, add 20 to both, and replace them with 90/70. The resulting visibilities are 0.50, 0.40 and 0.125. State the assumptions for each transformation and explain why the same word “loss” would be ambiguous.

Then use the independent-Poisson approximation to compare the uncertainties before and after attenuation. Explain why unchanged ideal visibility does not imply unchanged evidential strength. Do not use the approximation to claim a precise confidence interval for an actual fitted fringe scan.

Next compare exactly two toy emissions with a Poisson mean of two. The answers 0.32 and about 0.3352 demonstrate the effect of averaging a nonlinear response. Write down which assumptions would fail if the photon wavelengths varied substantially or emission changed the detected population.

Finish by writing a measured statement and an inferred statement separately. “The normalised contrast decreased” describes the selected observable. “Thermal radiation caused the decrease” requires the environmental model and controls against alternatives. The second statement can be well supported, but it contains more reasoning than the first.

There is no universal temperature at which all objects lose interference. The system, spatial scale, emission properties, transit history and measurement all matter. Heating power cannot be converted into molecular temperature without the appropriate calibration and model.

Use supplied data and paper calculations. Lasers, heated molecular sources, vacuum systems and chemical materials require specialist facilities; this lesson supplies no apparatus procedure.

The central skill is to preserve the chain from counts to contrast to mechanism. Return to Quantum Mechanics or the How Science Works master guide with a precise conclusion: emitted radiation can change accessible interference by carrying correlations, and the experimental case is strongest when detector response, phase behaviour and statistical uncertainty are tested alongside it.

Contents · Previous chapter · Continue to Quantum Mechanics