Titanium anodizing colours: the voltage relationship and why batches drift
Titanium colour comes from oxide thickness set by voltage, not dye. The approximate voltage-to-colour map, why red is impossible, and what makes a batch come out uneven.
Titanium anodizing is the only finish in EDC where the colour is not a pigment. It is physics: a transparent oxide layer grown to a precise thickness, producing colour by thin-film interference in exactly the way an oil film on water does.
That has two consequences worth understanding before you specify it.
Colour is set by voltage
Thicker oxide, different interference, different colour. Oxide thickness is controlled almost entirely by the voltage held during anodizing, not by time in the bath. Time affects how completely the layer forms; voltage decides what colour it forms at.
The approximate sequence, low to high voltage:
| Approx. voltage | Colour |
|---|---|
| 10–15 V | Light bronze / tan |
| 18–22 V | Deep bronze / brown |
| 25–30 V | Violet / purple |
| 30–40 V | Blue, deepening to royal |
| 45–55 V | Light blue, then silver-grey |
| 55–70 V | Yellow / gold |
| 75–90 V | Magenta / rose |
| 95–110 V | Green |
Treat these as indicative. The exact voltage for a given colour shifts with electrolyte concentration, temperature, alloy and surface preparation. Every shop has its own calibrated chart, which is why you ask for a physical sample rather than naming a number.
Why there is no true red
The interference sequence simply does not pass through a saturated red. Between magenta and gold you get rose and pink; you do not get fire-engine red. Anyone offering red anodized titanium is using paint, dye, PVD or a coating — which is fine, but it is not anodizing and it does not wear like anodizing.
Grade changes the result
Commercially pure Grade 2 gives deeper, more saturated colour than Grade 5. The alloying elements in Ti-6Al-4V make the oxide slightly less uniform, so the same voltage yields a flatter, chalkier version of the same hue.
If colour is the entire point of the part and it carries no load, Grade 2 is the better specification. See Grade 2 vs Grade 5.
Surface preparation decides sharpness
The oxide layer is transparent, so whatever is underneath shows through. A bead-blasted part anodizes to a soft satin colour; a polished part anodizes to a deep, almost wet-looking colour; a stonewashed part gets colour that varies slightly with the swirl, which is often the desired effect.
This also means surface defects are not hidden by anodizing — they are magnified by it. Any tool witness line, scratch or handling mark that survives to the anodize tank will be visible in the finished part.
Why batches drift
Three causes, in order of frequency:
Racking. Every part must sit at the same electrical potential. Loose contact, too many parts per rack, or parts shadowing each other produce voltage drop across the load and a visible gradient from one end of the batch to the other.
Two loads. A 200-piece order run as two tank loads on different days, with electrolyte that has aged between them, will produce two slightly different colours. Ask whether your batch goes through in one load.
Inconsistent prep. Mixed surface finishes going into the same tank come out as mixed colours, because the substrate is doing half the work.
What to write in the RFQ
- The colour, described in words, plus a reference photo if you have one
- That you want a physical sample set approved before production
- Whether the whole batch must match, and that it should run in one load if so
- The surface finish under the anodize, because it changes the result
That is four lines and it prevents the most common reorder dispute in EDC manufacturing. Send the part and the reference and the sample step gets priced in from the start.
Electrolyte and process variables
The voltage chart is the headline, but three other variables move the result:
Electrolyte composition. TSP (trisodium phosphate), ammonium sulphate and proprietary mixes are all in use. Each has a slightly different voltage-to-colour mapping, which is why a chart from one shop does not transfer to another.
Bath temperature and age. An electrolyte that has run all week behaves differently from a fresh one. This is the mechanism behind most batch-to-batch drift.
Ramp rate. How quickly voltage is brought up affects the uniformity of the layer. A fast ramp on a part with varying section can produce a visible difference between thick and thin areas.
None of these are things you specify. They are reasons to specify an approved physical sample and one tank load instead.
Surface preparation in detail
The oxide layer is transparent, so the substrate does all the visual work. Preparation is genuinely half the finish.
Degreasing is mandatory. Any oil, fingerprint or coolant residue produces a blotch. Parts should be handled with gloves from the final clean onward.
Pickling or etching in a mild acid removes the natural oxide that formed after machining, giving the anodize a clean start. Skipping it is a common cause of patchy colour.
Mechanical finish sets the character:
- Polished substrate gives deep, almost liquid colour
- Bead-blasted gives soft satin colour, and hides machining marks
- Stonewashed gives colour that varies subtly with the swirl
- As-machined shows every tool path in the finished colour
Decide which you want and state it, because the same voltage over three different substrates produces three visibly different parts.
Fading, wear and repair
Titanium anodize does not fade from UV the way dyed aluminium does — there is no dye to fade. It wears mechanically: the oxide layer is thin, and abrasion at contact points thins it, which shifts the interference colour before removing it entirely.
That is why a well-carried titanium scale develops colour variation at the high points. Some customers consider this a feature. If yours will not, either specify a substrate that hides it, or choose a lower-voltage colour where a small thickness change shifts the hue less.
Anodize can be stripped and reapplied. Unlike aluminium, stripping titanium anodize removes essentially no base metal, so a colour that comes out wrong is genuinely recoverable. That is a useful thing to know before you scrap a batch.
Two-tone and masked effects
Because colour is voltage-dependent, a part can be anodized twice at different voltages with masking between, producing crisp two-colour parts. Brush anodizing — applying the electrolyte with a conductive brush — allows gradient and freehand effects.
Both are manual processes. They are viable for limited editions in small numbers and they are not consistent enough for a repeating production part. Price them as a craft operation, per part, not per batch.
Anodizing and marking sequence
For titanium, mark before anodizing so the colour flows over the mark and sits flush. Marking after leaves a bare-metal scar through the colour.
The exception is when you deliberately want a bright mark on a coloured part, which is a legitimate design choice — just make it deliberately. See laser marking vs engraving vs etching.
FAQ
Can I match a Pantone colour? No. Interference colours occupy their own gamut and it does not map onto print colour systems. Work from a physical sample.
Why do two parts from the same batch look different in different light? Interference colour is genuinely angle- and light-dependent. A part that looks bronze in daylight can look pink under warm indoor lighting. This is inherent, not a defect. Photograph your product in the lighting your customers will see it in.
Does anodizing add thickness? Very little — the layer is a fraction of a micron to a few microns. Unlike Type III on aluminium, it does not meaningfully change fit.
Is colour anodized titanium food safe or skin safe? Titanium oxide is inert and titanium is widely used in implants and body jewellery. The anodized surface is the same oxide the metal forms naturally, just thicker.
Can Grade 5 achieve the same colours as Grade 2? The same hues, less saturated. If a customer is comparing a Grade 5 part to a Grade 2 reference photo they will see the difference. Be explicit about the grade in your product copy.
How many samples should I ask for? Three to five, at the voltages either side of your target. It is a small cost and it converts a subjective conversation into a physical decision. Ask for them with the quote.