Carbon print

☠Level D: chromium(VI), never used at any level

The only light-sensitive component of a carbon print is the one substance this course will not use at any level. The chromium policy is absolute, and this entry gives no procedure.

The carbon print is the permanence argument of this whole atlas in its purest form: take silver out of the image entirely and every silver failure mode goes with it. It is also the plainest case of the course’s own limit, because the price of that permanence is chromium(VI).

The chemistry

Light hardens the gelatin in proportion to exposure, and warm water then removes the rest, taking its pigment with it.

The photochemistry is photoglyphic engraving’s and the potassium dichromate entry carries it: chromium(VI) takes electrons from the colloid around it and falls to chromium(III), and the reduced chromium binds the colloid so that the exposed layer no longer swells and dissolves. The course has not found a description of that cross-link at the level of which groups are bridged, and does not assert one.

Two consequences follow and both are structural.

The image substance is the pigment. So a carbon print’s colour, permanence and grain are the pigment’s, not silver’s. There is no silver anywhere in the finished print.

The image is a relief. More light means more hardened gelatin means a thicker layer, so the print is genuinely thicker in the shadows. That is what makes it a continuous-tone original a printing process can be built from, and it is why the carbon print is the ancestor of the ink processes — the Woodburytype casts from exactly such a relief.

⚗A related hardening the course does have a source for

Gelatin’s tanning by an oxidised developing agent is the same kind of effect reached by a different route, and the timeline records it: in 1881 Leon Warnerke reported that only the unexposed parts of a pyrogallol-developed gelatine film remain soluble in warm water, the exposed and tanned parts being insoluble — and for two generations after that, the tanning was the basis of a family of processes.

It is worth knowing because it shows that “harden a colloid where the light went, then wash the rest away” is a general strategy rather than a property of chromium, and because staining and tanning developers are written in this course while the dichromate route is not.

Historical workflow

Eder’s chronology of the family is short and this process is its centre: Ponton, 1839, on paper; Talbot, 1852, on the colloid; Poitevin made the carbon process from it in 1855, and Swan made that workable in 1864.

The workflow in outline, and no further: a tissue carrying pigment in dichromated gelatin is exposed under a negative; the exposed tissue is developed in warm water, which dissolves the gelatin light did not harden and carries its pigment away; and because the hardening starts at the surface the light reached first, the finished image is normally transferred to a final support rather than developed in place. That transfer step is why the process has a reputation for difficulty, and this course does not describe it because it does not describe any of it.

The image

The pigment’s colour, which is a genuine freedom no silver process has: the maker chooses it when the tissue is made.

A relief in continuous tone, which is both its aesthetic signature and its industrial importance.

No silver failure modes at all — no tarnish, no mirroring, no sulfiding, no yellowed highlights from retained silver. Set that against the albumen print, where a small amount of silver remains in every area of a fixed sheet and turns to silver sulfide whatever the wash does, and the contrast is the whole argument for pigment processes.

Permanence

The permanence of the pigment rather than of a metal, which is a stronger claim than any toner can make, and one this atlas can state without hedging because it is a claim about what is absent.

What remains is the gelatin, the paper, the mount and the storage climate — the same four things IPI names as outranking everything else for a silver print. A carbon print is not immortal; it simply has one fewer thing that can go wrong, and it happens to be the thing that goes wrong most often in the rest of this atlas.

One caution about analysis. A pigment-in-gelatin image will show none of the silver failures, so a conservator’s first job with a suspected carbon print is to establish which family the object belongs to. Knowing that decides which failures to look for, and the same point governs the Woodburytype and collotype, where the image is ink or pigment in gelatin and none of the silver diagnoses apply.

Hazards

Chromium(VI), and nothing else on the page comes close. Potassium dichromate’s notified classification runs to twelve hazard statements; HSE’s EH40 sets chromium(VI) compounds at 0.01 mg/m³ as chromium; and the Level D policy names dichromates and chromic acid among the families it covers.

The chromium policy settles it once for the whole course: chromium(VI) is never used, at any level, anywhere. Chromium(III), as chrome alum, is permitted at Level B — which is worth stating in the same breath, because a reader who takes “chromium is forbidden” away from this page will misread half of Part V.

Gelatin and the pigments are, in themselves, benign; the tissue is not.

Why we study it and do not reproduce it

Because it is the permanence argument in its purest form. Every other entry in this atlas that argues about permanence is arguing about how to protect silver. This one removes the silver.

Because it is the honest statement of the course’s own limit. The chromium policy is a decision the course made and can defend, and the cost of that decision is exactly this: a beautiful, durable, well-documented process that cannot be taught. Saying so is better than pretending the process is difficult or obscure.

And because it is the parent of the ink processes. A relief image in continuous tone is what the Woodburytype casts from and what photogravure etches through, so a reader who skips carbon printing arrives at those two without the mechanism.

Where the course carries it, and it carries no steps. Part XXVI owns the light-hardened colloids, in Light-Hardened Colloids: Bitumen, Dichromate and the Pigment Processes. That lesson treats the whole family as the second family of photochemistry — no latent image, nothing amplified, no fixer, and an image made of whatever colloid the light left standing — which is precisely why this entry’s argument about removing the silver is not an eccentricity but the family’s defining property.

For this row in particular, two of its sections do the work. The chemistry, stated to the resolution the evidence supports stops the mechanism where the sources stop rather than rounding it up. And the hazard that stops us is where the classification is reasoned out at length instead of announced, which is what this entry’s claim to be “the honest statement of the course’s own limit” requires somebody to have done.

And the part draws a line this entry could not. The safety rubric puts chromium(VI) at Level C — specialist or supervised laboratory — in any quantity, and Part XXVI nonetheless presents these processes at Level D. It names that as the course’s own decision rather than the rubric’s, for one reason: a course for a reader at home cannot assume extraction, a waste contract or a second trained person, so it withholds the procedure from everybody. Level D is not a prohibition on the practice; it is an admission about what a page can assume.

The dichromate-free question is answered, and the answer is “not yet, and here is the bar”. The lesson records that substitutes exist and are named in the literature — Ware refers in passing to the Chibatype, which hardens pigmented colloids with an ammonium ferric citrate sensitiser, and to diazidostilbenes — and then says plainly that a mention is not a formulation. What would have to exist before the course taught one is set out as four conditions: a published reproducible formulation from a source meeting the research standard; a dated hazard record for the sensitiser, because an unclassified substance is uncharacterised rather than mild; a waste route a domestic reader can use; and ideally an independent account of somebody working it. And the lesson rules out the shortcut explicitly: it will not take a dichromated formula, swap the crosslinker and publish the result as carbon printing, because a substituted formulation is a new and untested process wearing a historical name.

Where institutions and conservators work with the process under controlled conditions, the Level D policy directs a page to point at their published accounts rather than paraphrase them into something that reads like instructions. The lesson does exactly that, naming three routes — a supervised workshop, a department with extraction and a waste contract, and a verified dichromate-free route if one arrives — and one honest trade it can offer today: if what you want is a print whose image substance is not silver, cyanotype and palladium are in this course with procedures, because their controls are ones a reader can assemble. Neither is a pigment process and neither gives you a relief, which the lesson says rather than glosses.