Colour drift in decorative PVD: Why most defect reports get the root cause wrong

Two distinct physical mechanisms create colour in PVD coatings. Misidentifying which one is active in your process will send you chasing the wrong parameter, burning production time, every single time.

Picture made by AI

The complaint that hides the real problem

Faucet handles rejected. Watch cases returned. A batch of door hardware that looks slightly different from last week's run, with no obvious change in the process log.

The defect report says: colour drift, cause unknown.

In most cases, the cause is not unknown. It is just rarely examined at the right level. Colour in PVD coatings does not come from one mechanism. It comes from two fundamentally different physical processes. Treating them as the same problem, and applying the same corrective action, is where production efficiency quietly disappears.

This article explains both mechanisms in plain terms, shows how to distinguish them in practice, and describes what each one demands from your process and your materials.

🌈Mechanism 1: Intrinsic colour (absorption-based)

The first colour mechanism is rooted in the electronic structure of the coating material itself.

In transition metals, the outermost electrons occupy what chemists call d-orbitals. These orbitals are not always full. When a metal is oxidised or combined with nitrogen or carbon to form a ceramic compound, the remaining d-electrons interact with incoming light. Specific wavelengths of visible light are absorbed, and the complementary colour is what the eye perceives.

This is intrinsic colour. It belongs to the material, not to the thickness of the film.

A well-documented example is the titanium oxide series (Carlotto, J.A., "The Chemistry of Thin Film Colour," 43rd SVC Annual Technical Conference, 2000):

• TiO (Ti²⁺): black/grey. Most visible wavelengths absorbed.

• Ti₂O₃ (Ti³⁺): blue. Absorption shifts to the lower-energy end of the spectrum.

• TiO₂ (Ti⁴⁺): colourless. All d-shells filled by oxygen, no selective absorption.

• TiO₃ (Ti⁴⁺ peroxide): yellow.

The same logic applies to other transition metals. Vanadium compounds span yellow, blue and green. Chromium oxide (Cr₂O₃) produces green. Tantalum nitride (Ta₃N₅) delivers a red tone (Carlotto, 2000), one of the rarest achievable in PVD. The colour is not arbitrary. It follows directly from the electronic configuration of the element and how it bonds with the reactive gas.

Key characteristics of intrinsic colour:

• Colour is defined by compound stoichiometry and oxidation state

• Changing thickness changes intensity, not hue (for semitransparent films)

• Adjusting reactive gas flow shifts the oxidation state and therefore the colour

• High reproducibility is achievable when stoichiometry is tightly controlled

What this means for production:

In the intrinsic regime, colour stability depends on controlling the chemistry of the reaction. Reactive gas flow, partial pressure, target surface condition and plasma density all influence which oxidation state forms at the substrate. A small drift in oxygen or nitrogen partial pressure can shift the compound phase, and with it, the colour.

The thickness tolerances are relatively forgiving. A film at 50 nm and the same film at 80 nm with identical stoichiometry will appear the same colour, at different saturation levels.

🌈Mechanism 2: Interference colour (thickness-based)

The second mechanism has nothing to do with electronic structure. It is pure optics.

When light strikes a thin, partially transparent film, part of it reflects from the top surface and part travels through the film, reflects from the substrate below, and travels back out. If the film thickness is on the order of visible wavelengths, these two reflected beams interfere with each other. Some wavelengths cancel through destructive interference. The colour you see is what remains.

This is the so-called λ/4 effect, described by the relationship:

t = λₘ / 4n

Where t is the film thickness, n is the refractive index of the transparent layer, and λₘ is the wavelength being cancelled. The apparent colour is the complement of λₘ.

Experimental measurements of this effect were reported for TiOₓ/TiN bilayer coatings deposited by reactive cathodic arc evaporation (García et al., "Decorative Electric Arc PVD Coatings," 48th SVC Annual Technical Conference, 2005):

| Colour | λₘ (nm) | Oxide thickness |

| ------ | ------- | --------------- |

| Purple | 570 nm  | ~52.7 nm        |

| Blue   | 620 nm  | ~57.4 nm        |

| Green  | 660 nm  | ~61.1 nm        |

The difference between purple and blue: 4.7 nm. Achieved in 31 seconds of additional O₂ exposure under the test conditions described. The difference between blue and green: 3.7 nm.

Key characteristics of interference colour:

• Thickness is the primary control parameter

• Chemistry plays a secondary role, mainly through the refractive index

• Single-digit nanometre variations produce visible colour shifts

• Colour uniformity on three-dimensional parts is inherently difficult

What this means for production:

In the interference regime, colour stability depends on controlling film thickness with nanometre precision across the entire substrate surface, including complex geometries. A gas distribution asymmetry of a few percent, a variation in substrate rotation, or uneven plasma density can produce visible colour gradients. These defects typically appear as rainbow banding on convex surfaces, or as colour mismatch between parts at different fixture positions in the chamber.

This is the regime where defect reports most often say "process drift" when the actual issue is plasma uniformity.

🌈How to identify which regime you are operating in

The diagnostic is straightforward in principle, more difficult in practice. Start with two questions:

1. Does the colour change when you change thickness, while keeping chemistry constant?

If yes: you are in the interference regime. Thickness is driving colour.

2. Does the colour change when you change reactive gas flow, at constant thickness?

If yes: you are in the intrinsic regime. Stoichiometry is driving colour.

In practice, both mechanisms can operate simultaneously. A TiOₓ layer deposited on a TiN sublayer produces interference effects because TiN acts as a reflective base, while the TiOₓ introduces optical path length differences. The colour output is therefore sensitive to both the TiOₓ stoichiometry (which determines the refractive index) and its thickness.

This overlap is precisely where misdiagnosis happens. Engineers adjust chemistry, but thickness is the dominant variable. Or they adjust deposition time, not realising the stoichiometry has shifted and changed the refractive index. Both interventions produce unpredictable results.

🌈Target material selection and its role in colour control

The choice of coating material does not only determine deposition rate and film hardness. It shapes which colour mechanisms are accessible, and how controllable each one is in production.

Materials with multiple accessible oxidation states, such as titanium, vanadium, niobium and molybdenum, offer a range of intrinsic colours through controlled reactive deposition. The transition between these states is engineered through reactive gas partial pressure, with the composition of the source material establishing which colour range is reachable in the first place.

Niobium deserves particular attention. As a so-called valve metal, niobium can be anodised after deposition to produce a highly dense, hard oxide layer. The colour produced this way is interference-based, but the post-deposition anodisation step gives far greater thickness uniformity than reactive deposition inside a vacuum chamber. The result is a reproducible, full-spectrum palette: champagne, dark blue, antique gold, pink, purple, turquoise and green, accessible through voltage control rather than process gas management (Hovsepian et al., "High Performance Colourful PVD Coatings," 47th SVC Annual Technical Conference, 2004). Reported hardness for these anodised Nb₂O₅ layers reaches HV 1500.

This approach effectively decouples the vacuum process from the colour-defining step, which improves batch-to-batch reproducibility significantly for interference colours.

🌈Production checklist: before you adjust anything

After a colour deviation, identify the active mechanism first.

Intrinsic colour regime:

• Stabilise reactive gas partial pressure (not just flow rate)

• Monitor target surface condition and erosion uniformity

• Check plasma density distribution (unbalanced magnetron configuration, power stability)

• Investigate target poisoning if colour suddenly shifts to "washed out"

Interference colour regime:

• Map plasma density and film thickness uniformity across the chamber

• Review substrate rotation speed and fixture geometry

• Verify reactive gas inlet distribution, especially for large or asymmetric chambers

• Use spectrophotometric measurement rather than visual inspection alone

🌈Final message

Colour drift in decorative PVD is rarely mysterious once you ask the right question: is this a chemistry problem or a thickness problem?

Intrinsic colour responds to stoichiometry control. Interference colour responds to thickness control. Confusing the two leads to process interventions that make things worse, not better.

Understanding which mechanism governs your specific coating system is the starting point for meaningful process optimisation. It also determines which source materials and process routes give you the most reliable path to colour reproducibility at industrial scale.

🌈Work With Us

If you are developing or scaling a decorative PVD process and are uncertain which colour regime you are operating in, we are happy to discuss your specific application in detail.

Avaluxe supplies PVD coating materials across the full range of elements and alloys used in decorative and functional thin film processes. We also offer technical consultation for companies looking to establish or optimise their PVD process, from equipment setup, to material selection through to process architecture.

🌈Contact us to discuss your target requirements or to explore our process know-how support.

References

1. Carlotto, J.A. "The Chemistry of Thin Film Colour." 43rd SVC Annual Technical Conference Proceedings, Denver, 2000.

2. García, J.A., Martínez, R., Rico, M., Bueno, R., Fuentes, G.G., Rodríguez, R.J. "Decorative Electric Arc PVD Coatings." 48th SVC Annual Technical Conference Proceedings, 2005.

3. Hovsepian, P.Eh. et al. "High Performance Colourful PVD Coatings." 47th SVC Annual Technical Conference Proceedings, Dallas, 2004.

4. Eerden, M., Tietema, R., Krug, T., Hovsepian, P.Eh. "Current Progress in Large Scale Manufacturing of PVD Coatings for Decorative Applications." 48th SVC Annual Technical Conference Proceedings, 2005.

5. Schulz, S.W. "Whither Decorative Coatings?" 44th SVC Annual Technical Conference Proceedings, Philadelphia, 2001.