Anodizing volts and laser power both grow the same nanometre-thin TiO₂ film on titanium — they just take completely different physical routes to get there. This page compares them side by side, then lets you drive both with a shared simulator.
Neither process adds pigment. Both grow a transparent titanium dioxide (TiO₂) layer a few tens to a few hundred nanometres thick on top of the bare metal. Light reflects off both the top of that film and the metal underneath it; the two reflections interfere, and whichever wavelengths cancel out determine what colour survives. Thickness is the only dial that matters — anodizing and lasering are just two different ways of turning that dial.
Electrochemical anodizing and laser marking reach the same colour space from opposite directions — one drives oxide growth with an electric field, the other with heat.
Switch modes to control oxide thickness with a bath voltage or with the real xTool F2 IR module (5W @ 1064nm) — the same laser used throughout Learning Lasers. Both feed the same TiO₂ interference model, so the same colour can come from either path.
Volts→thickness uses the widely-published ~2.5 nm/V titanium anodizing rule. Laser mode now shows real colours read from a physical titanium chip card made at this exact 1000 LPCM / 200mm/s reference, power-labelled 13-61% — Peak Temp is still the theoretical model, kept for context. Above ~41% the card itself desaturates into greys/mauves rather than deepening further — see the worked example and disclaimer below.
Every chip from the photographed card (13-61%), complete — with the theoretical model's Oxide/Volts/Power/Peak-Temp columns pulled in wherever a chip has a sensible theoretical counterpart. Where they land on wildly different power percentages for "the same" hue, that's the model's approximation talking, not the real card.
| Power (actual) | Colour (Eyes Verified) | Regime | Anodized Process | Oxide (nm) | Typical Volts | Power (theoretical) | Peak Temp (theoretical) |
|---|
Calibration status: Power, Colour, and Regime are read directly from the photographed chip card — real measurement. The four theoretical columns are the tuned Arrhenius model, shown only where a chip has a plausible theoretical counterpart; "—" means the real card produced a colour the model doesn't predict at all — mostly true above ~35%, where the real card runs sage-teal transitions and then desaturates into grey/mauve instead. The five darker "Not produced by laser" rows (62–96V) are the reverse case: published anodizing colours (Yellow-Green, 2nd-Order Gold, Orange, Pink/Magenta, Deep Green) with real volts and a computable theoretical laser-power equivalent, but no laser setting — on this card or otherwise — has ever actually produced them; the same 160–250nm range instead shows the real card's sage-to-grey transition above. Below 13% is untested (no colour); above 61% is untested (unknown). Colours are a careful visual read of a photograph, not a colorimeter measurement.
The same three-stage model as SS304's parameters section — energy in, peak temperature, then Arrhenius oxide growth — run with titanium's own numbers instead of steel's.
That temperature is not real. 4761°C is roughly 3000°C past titanium's boiling point. Matching the real chip card's colours with this simple model requires peak temperatures that cannot physically occur — a genuine finding, not an error, and the subject of the "Closing the Gap" section below.
Because titanium is lighter (roughly half SS304's density, so less thermal mass to heat) and absorbs more 1064nm light (see Power, Beam Shape & Heat Transfer for how absorptivity turns raw optical power into material-level heating), a given power setting pushes it far hotter than steel at the same exposure time — which is exactly why real titanium annealing markers run at a fraction of the power used for SS304 structural colour, and why power, not speed or density, is the dial that actually needs fine control.
A physical titanium colour chip card — made at this exact 1000 LPCM / 200mm/s reference, each chip labelled with its power% — beats any tuned model, so the simulator's swatch now reads directly off it (interpolated between chips) instead of the theoretical Arrhenius colour. Two things it shows that the model alone wouldn't predict:
Published anodizing charts put a rich, saturated Deep Green around ~90–110V — noticeably richer than the paler green earlier in the sequence. That's the ~250nm stop above; earlier "Learning Lasers" work on SS304 never reaches a colour this saturated at comparable thickness, because titanium's oxide keeps producing strong interference further into the 2nd order than chromium oxide does.
Neither anodizing nor laser marking can produce a true, saturated red on titanium. Thin-film interference at TiO₂'s refractive index just doesn't have a thickness where red reinforces without other wavelengths reinforcing alongside it — the closest you get is pink/magenta (a red-blue mix), never a clean red.
A second real chip card — a completely different laser recipe (200 mm/s, 834 LPCM, crosshatch fill, 12–91% power) — showed the same gap the anodizing chart already hints at above: plenty of gold, purple, blue and teal, then a drab olive, but never a rich, saturated green. We put the question to the physics.
A: Because green asks more of a simple thin-film than any other colour on this chart. Two-beam interference off a single TiO₂ layer produces one broad, cosine-shaped reflectance curve across the visible spectrum — one wavelength region reinforced, the opposite region cancelled. Gold, blue, and purple only need one end of the spectrum suppressed (say, blue cancelled to leave a warm gold, or red cancelled to leave a cool blue) — an easy, robust condition that survives small thickness variation. A clean, saturated green needs the interference to cancel both the red end and the blue end at once, leaving only a narrow window in the middle — a much sharper, more fragile condition that only exists briefly, at a very specific thickness, before sliding into the next hue.
The new chip card shows exactly where that fragile window would fall — and exactly why it never survives to become a real green:
| Power | Colour (sampled) | Hue | Saturation | What's happening |
|---|---|---|---|---|
| 33% | Purple | 241° | 36% | Vivid — smooth, coherent film |
| 37% | Blue | 210° | 66% | Peak saturation of the whole card |
| 50% | Teal | 175° | 27% | Closest approach from the blue side, already fading |
| 56% | Dark teal | 160° | 23% | Still short of true green (120°) — and dimming |
| 57–59% | Muted brown | ~32° | ~20% | Hue collapses straight through the green band without stopping |
| 60% | Drab olive | 98° | 18% | Closest approach from the yellow side — but barely saturated |
| 66% | Olive-brown | 61° | 31% | Sliding on into brown for the rest of the card |
Both approaches — from the blue side (44–56%) and from the yellow side (60–66%) — get within striking distance of green, and both lose saturation right at that point instead of gaining it. That's not a coincidence: it's the same thickness range where this laser-grown oxide starts to roughen from accumulated heat (the "Desaturating" regime documented in the Reference Charts below). The narrow, fragile reflectance notch a real green needs requires a still-smooth, still-coherent film — and this card's oxide stops being smooth right where green would need it to be.
Full digitized card (all 80 chips, RGB + derived hue/saturation/lightness): chip-card-200mms-834lpcm-crosshatch.json.
| Trait | 🔋 Anodizing | 📡 Laser |
|---|---|---|
| Uniformity | Very high — whole part colours evenly | Depends on scan pattern; can vary pixel to pixel |
| Patterning | Solid colour or masked regions only | Full raster/vector control — gradients, text, images |
| Speed per part | Seconds, once rig is set up | Seconds–minutes, scales with area, line density (LPCM), and speed |
| Equipment | Power supply, electrolyte bath, electrodes | IR fibre laser + galvo, no chemicals |
| Repeatability | High — voltage is a direct, stable control | Sensitive to focus, surface prep, and thermal history |
| Reversibility | Can often be stripped/re-anodized | Oxide can be removed by re-polishing; base metal unaffected |
Originally, theory said Blue needs ~7.7% power against a real ~23–27% — roughly 3× under. Step 1 below (re-fitting the model's two acknowledged-tuned constants against the real chip card) has since been done and is reflected throughout this page's Reference Charts and simulator. Here's the full derivation, what the refit changed, and what's still unresolved.
Seven real (power%, colour) anchors exist in the vivid range — the midpoint of each theoretical colour's real power band from the Reference Charts table (Pale Straw≈14%, Bronze≈17%, Purple=20%, Indigo≈21.5%, Blue≈25%, Cyan/Teal≈30%, Pale Green≈33%). A least-squares fit in log(nm) space, adjusting only the two constants already flagged as tuned/unverified — the Arrhenius pre-exponential factor A and the focal-skin depth — while holding the cited constants (α_Ti, η, E_a, ρ, c_p) fixed, gives:
| Constant | Before | After (refit) |
|---|---|---|
| A (pre-exponential) | 8.9×10¹² | 1.05×10⁶ |
| Skin depth | 40µm | 20µm |
The refit closes most of the gap — predicted power now lands within ~15% of every real anchor (down from a ~3× miss), and the Power (theoretical) column above reflects it. But getting there requires peak temperatures of 1500–6000°C across that same range — 1000–4000°C past titanium's actual boiling point (3287°C). That's not a bug in the refit; it's what the data demands under this model's functional form.
Terms used across this page, in the order they're most useful to know.
| Term | Meaning |
|---|---|
| TiO₂ (titanium dioxide) | The transparent oxide layer that grows on titanium's surface — the only thing actually producing colour, whether grown by voltage or by heat. |
| Thin-film interference | The optical effect behind every colour on this page: light reflecting off the top and bottom of the oxide film interferes with itself, cancelling some wavelengths and reinforcing others. No pigment involved. |
| Oxide thickness (nm) | The one physical dial that actually determines colour. Anodizing volts and laser power are just two different ways of setting it. |
| Interference order (1st / 2nd order) | How many half-wavelengths of light fit in the oxide's optical path before a colour reinforces. Thicker films repeat similar hues in progressively less saturated later orders. |
| Anodizing | Electrochemical process: a DC voltage across the part, immersed in an electrolyte bath, drives oxide growth directly. |
| xTool F2 (IR module) | The specific 5W, 1064nm fibre laser this page's real numbers — and every other Learning Lasers lesson — are built around. |
| LPCM (Lines Per Centimetre) | How densely a laser fill packs its lines — a separate setting from power or speed, held fixed at 1000 on this page per hands-on testing. |
| Absorptivity (α) | The fraction of incident laser light actually absorbed by the surface rather than reflected. Titanium absorbs 1064nm light markedly better than stainless steel (~55% vs. ~38%). |
| Peak surface temperature (T_peak) | The modelled highest temperature the surface reaches during a pass — the thing that (in this model) actually drives oxide growth. Shown for context; often exceeds what's physically possible, a finding in itself. |
| Parabolic growth law | How oxide films typically grow: thickness scales with the square root of time, because the existing film itself slows further growth (oxygen has to diffuse further to reach fresh metal). |
| Arrhenius equation | A standard formula describing how a reaction rate depends exponentially on temperature. Used here for the oxide growth-rate constant, k_p(T). |
| Activation energy (Eₐ) | The Arrhenius equation's temperature-sensitivity constant. A cited, published value for TiO₂ (155–205 kJ/mol) — left unchanged by the refit on this page. |
| Pre-exponential factor (A) | The Arrhenius equation's other constant, setting the overall scale of the growth rate. Not independently measured for this setup — re-fit directly against real chip-card data in "Closing the Gap," rather than tuned to a literature temperature window. |
| Rutile | The crystal structure TiO₂ forms into during thermal oxidation in the temperature range this page's literature sources describe. |
| Refit | This page's term for adjusting a model's tunable constants until it reproduces real measured data — as done in "Closing the Gap," versus the original constants, which were tuned only against a plausible-sounding temperature range. |
| Regime: Vivid / Desaturating | This page's own classification of the real chip card — "Vivid" where laser power still produces the gold-to-sage progression (13–41%), "Desaturating" where it fades into grey/mauve instead (43–61%). |