Find the mechanism before the formula
A wall changes the local critical behaviour of a fluid or magnet. The bulk still supplies a diverging correlation length, but order and response measured at the boundary can follow different exponents. A useful analysis must therefore scale both distance and amplitude, then test whether profiles from different temperatures collapse onto one boundary scaling function.
We will separate bulk and surface observables, state the ordinary-transition scaling relations with their assumptions, calculate two correlation lengths and amplitudes, construct a changed-temperature profile collapse, and show how surface fields, microscopic bins and finite film thickness break a careless collapse.
This guide belongs to the Gold Standard of Thermodynamics, the Gold Standard of Education and the How X Works hub.
Choose a reading route
Start with the question closest to the decision you need to make. Every route returns to the same worked evidence and transfer checks.
| Route | Chapters | Use it when |
|---|---|---|
| 1 | 1–3 | You need bulk versus boundary observables and the exponent dictionary. |
| 2 | 4–5 | You want the correlation-length and amplitude calculations. |
| 3 | 6–8 | You need profile collapse, boundary relations and crossover checks. |
| 4 | 9–10 | You must fit evidence and respect finite-size and correction limits. |
| 5 | 11–12 | You want FAQs and a reproducible surface-scaling audit. |
Chapter contents
Open the group you need, then use the chapter links. Teaching sections remain open on the page.
Build the model
Test the mechanism
1. Define where the observable lives
The bulk order parameter mb can vanish as |t|β below the critical temperature, while the surface order parameter m1 at an ordinary boundary vanishes as |t|β1. The exponents β and β1 need not be equal because translation symmetry ends at the wall.
Surface-surface and surface-bulk susceptibilities also differ. Define the measurement depth, boundary condition, surface field, enhancement and reduced temperature t=(T−Tc)/Tc before assigning an exponent.
- Name the bulk and boundary observables.
- State the surface universality class.
- Define the wall position and extrapolation length.
- Report film thickness and surface-field history.
2. Surface exponents describe boundary correlations and response
Common exponents include β1 for the surface order parameter, γ11 for response of the surface layer to a surface field, and γ1 for a surface-to-bulk response. The correlation exponents η∥ and η⊥ describe decay parallel to the boundary and between boundary and bulk.
A negative γ11 can describe a finite cusp rather than a divergence. The sign is not a failed fit. It is one reason to keep the observable definition attached to the exponent name instead of copying a bulk critical-exponent table.
3. Boundary scaling relations are conditional consistency checks
For a short-range isotropic semi-infinite system in the standard scaling regime, relations include β1=ν(d−2+η∥)/2, γ11=ν(1−η∥), γ1=ν(2−η⊥) and η⊥=(η+η∥)/2. They reduce the number of independent exponents.
These equations do not erase boundary universality classes. Long-range forces, anisotropy, quenched disorder, edges, a nonzero surface field or dimensional crossover can change the applicable relations or add crossover variables. Use them to test one stated model, not to force every surface into the same numbers.
| Quantity | Scaling role | Evidence needed |
|---|---|---|
| ν | ξ=ξ₀|t|⁻ν | bulk correlation-length range |
| β | bulk order amplitude | bulk composition or magnetisation |
| β₁ | surface order amplitude | boundary-specific observable |
| η∥, η⊥ | boundary correlation decay | resolved parallel and normal correlations |
4. Worked example: calculate the growing correlation length
Take ν=0.63 and ξ0=0.20 nm. At |t|=0.01, ξ=0.20×0.01−0.63≈3.64 nm. At |t|=0.001, ξ≈15.5 nm. One decade closer to criticality expands the distance scale by about 4.27.
The calculation supplies the horizontal scaling variable x=z/ξ. A profile measured 5 nm from the wall sits at x≈1.37 for |t|=0.01 but x≈0.323 for |t|=0.001. Comparing the two raw values at the same z therefore compares different scaled positions.
5. Worked example: scale the bulk order amplitude
Take the three-dimensional Ising-like bulk value β=0.326 for an illustrative fluid or magnet in that universality class. The amplitude factor |t|β is approximately 0.223 at |t|=0.01 and 0.105 at |t|=0.001.
A profile form m(z,t)=|t|βP(z/ξ) therefore suggests dividing the measured order parameter by those amplitudes before comparing temperatures. The normalisation uses bulk β for this profile form; it does not assert that the wall order parameter itself follows β rather than β1.
6. Transfer check: change temperature and collapse the profile
For each temperature, plot m(z,t)/|t|β vertically against z/ξ horizontally. The point z=5 nm moves from x≈1.37 to x≈0.323, so it should be compared with a different physical distance in the other profile. Interpolate the scaled curves rather than matching raw z bins.
If the boundary condition and asymptotic regime are the same, the two rescaled profiles should approach a common P(x) within uncertainty. A successful collapse is evidence for the scaling form; it does not uniquely determine the exponents because nearby exponent choices and backgrounds can mimic a short-range collapse.
7. Near the wall, the scaling function has its own power law
At bulk criticality, an order-parameter profile can behave as m(z) proportional to z−β/ν in the normal or strongly symmetry-breaking case over the scaling range. In the finite-temperature form, this appears through the small-x behaviour of P(x).
The first molecular layers are not a continuum scaling regime. Surface roughness, an extrapolation length and instrument resolution shift the effective origin. Exclude or model those bins explicitly; moving z=0 until a collapse looks tidy is not an independent test.
8. Changing the boundary condition introduces crossover variables
At the ordinary transition, h1=0 and the surface remains disordered at bulk criticality. A nonzero surface field drives crossover toward the normal class. Surface enhancement can tune toward the special transition. Profiles then depend on additional scaled variables, not only z/ξ.
Repeat the collapse for more than one surface treatment. If a single curve works only after mixing different surface fields, it may be compensating for crossover rather than revealing universality. The companion critical-adsorption analysis supplies the surface-field length scale that this simple two-variable plot omits.
9. Fit a collapse with held-out temperatures and visible residuals
Estimate Tc, analytic backgrounds and ξ independently where possible. Choose exponents on a training subset, then predict profiles at held-out temperatures and thicknesses. Plot residuals against z/ξ, surface treatment and instrument resolution.
Quantify collapse quality rather than relying on an overplotted image. Bootstrap the profiles or propagate covariance from the reflectometry, scattering or simulation estimator. Strong correlations among Tc, ν, β and the distance origin should remain in the reported uncertainty.
10. Limits and misconceptions
- Bulk β and surface β₁ describe different observables.
- A visual collapse over a narrow range does not identify unique exponents.
- Boundary scaling relations require a stated universality class and interaction range.
- The first microscopic layers need corrections or exclusion.
- A nonzero surface field introduces crossover toward the normal class.
- When ξ approaches film thickness, two boundaries and finite-size rounding enter.
With L=10 nm in the worked example, ξ already exceeds L at |t|=0.001. That profile cannot be treated as semi-infinite. Gravity, wetting layers, long-range wall forces, surface disorder and finite resolution can all create systematic deviations that are scientifically meaningful rather than merely bad data.
11. Frequently asked questions
Why divide by |t| to a power?
Critical scaling changes both the spatial reach and the order-parameter amplitude. Scaling only z by ξ leaves curves vertically separated even when one universal profile describes them.
Can β₁ be inferred from a bulk profile collapse?
Not automatically. β₁ belongs to a specifically defined surface observable. A bulk-normalised profile can constrain boundary behaviour, but the surface order amplitude and its corrections need their own analysis.
Does failure to collapse disprove criticality?
No. It may reveal crossover, finite size, an incorrect Tc, backgrounds, long-range forces or resolution effects. The residual pattern should decide which model to test next.
12. A reproducible surface-scaling checklist
| Stage | Evidence to retain |
|---|---|
| Observables | bulk, surface and profile definitions |
| Boundary | universality class, h₁, enhancement and roughness |
| Scales | Tc, ξ₀, ν, β and distance origin |
| Collapse | scaled axes, excluded bins and uncertainty |
| Transfer | held-out temperatures, surfaces and thicknesses |
| Limits | finite size, crossover, backgrounds and long-range forces |
The gold standard is to calculate the horizontal and vertical rescaling, predict a changed-temperature profile, and then test the same exponents under a changed surface or film thickness. A boundary scaling law is useful when it exposes where universality holds and where another length scale has entered.
Reading routes · Previous chapter · Contents · Continue reading
Continue through the Gold Standard ecosystem
Return to the Gold Standard of Thermodynamics, the Gold Standard of Education parent, or the How X Works hub.
Read next in this subject
- Critical Adsorption, Surface Fields and Normal-Transition Crossover
- Critical Exponents, Scaling Laws and the Limits of Mean-Field Equations of State
- Critical Casimir Forces, Confinement and Finite-Size Scaling
Authoritative reading
- Pleimling, critical phenomena at perfect and non-perfect surfaces
- Diehl and Dietrich, field-theoretical approach to static critical phenomena in semi-infinite systems
- Diehl, critical adsorption and normal versus extraordinary surface transitions
- Ciach and Ritschel, critical adsorption with weak surface fields