Skip to content

Sodium citrate kallitype developer

One salt, one weighing, one litre — and nothing in the bottle that can reduce silver. That is the first thing to understand about a kallitype developer and the thing that makes it unlike every other developer in this formulary. There is no developing agent here, no preservative, no alkali and no restrainer. There is a ligand in water.

The reducing agent was made by light, in the paper, hours ago. It is iron(II) oxalate, it is stuck where the light put it, and it is almost insoluble. All this bath does is dissolve it — and everything the print becomes follows from that one act of complexation.

IngredientQuantityForm the source specifies
Trisodium citrate dihydrate200 gKing writes "sodium citrate" and names no grade; the Formulary supplies a 300 g packet and names none either. The encyclopaedia's entry is the dihydrate at formula weight 294.10, and the anhydrous salt at 258.07 is sold beside it. Weighed as the dihydrate this is 0.680 mol of citrate per litre; weighed as the anhydrous salt it is 0.775 mol, 14 per cent more
Waterto make 1000 mLDistilled water, and a make-up volume rather than an added one: King's instruction is 200 g into 750 mL, dissolved, then water to 1000 mL, which is what makes the bottle a stated 20 per cent w/v. The Formulary starts its own make-up at 52 °C/125 °F; King states no temperature.

To dissolve the iron the light reduced, so that it can hand its electron to silver; and then to go on dissolving the iron the light never touched, so that it leaves with the developer instead of staying in the paper. Those are two different jobs with two different timescales, and the whole character of this bath is that the first is over in half a minute and the second is what you are waiting for.

Mike Ware’s account of the platinotype names the developer’s role exactly, and it is the definition this page is built on. Willis’s process needed three things: a noble-metal salt, a light-sensitive iron salt, and third, “a ‘developer’, potassium oxalate, to dissolve the insoluble ferrous oxalate produced by light, providing a strongly reducing solution” which could then reduce the metal. Not to reduce anything itself. To dissolve.

The kallitype is the same photochemistry with silver nitrate in place of the platinum, and this developer is that same bath with citrate in place of the oxalate. What it adds to the platinum version is a second duty that matters far more for a silver print than for a platinum one: citrate holds iron(III) as well as iron(II), so the unexposed iron comes out of the sheet in the same tray, and residual iron is precisely what kills a kallitype.

Any kallitype made with the two-solution sensitiser, where the colour is going to be decided by a toner rather than by the developer. This is King’s whole method in one sentence, and he states the principle plainly: instead of the dozens of developers recommended in the older texts, “I recommend just one.” His stated principles are “less is more” and that “the print should be processed for maximum permanence”; the practical argument he gives alongside them is that toning comes after development and clearing, “when it is fairly obvious if the print is a keeper or not”, so that no gold or palladium is spent on a print that was never going to work. A developer that also chooses the colour takes that decision earlier, and takes it in the dark.

Where you intend to tone, which on King’s argument is always. He is unequivocal that an untoned kallitype will eventually fade, because it is impossible to remove all residual iron and any that remains will oxidise the image silver. Toning replaces that silver with a metal that resists the oxidation. This developer suits that plan better than its rivals for a reason that is chemical rather than aesthetic: it is the bath that takes the most iron out before the print ever reaches the toner. See gold and thiocyanate and the platinum toner for what happens next.

As the first developer anyone should use for this process. It is one salt, it needs no temperature, it has no ratio to get wrong, and it fails visibly rather than subtly: an exhausted citrate developer announces itself as a stain in the paper that was masked, which you can see in the tray. King’s advice to a beginner is to stay with it — “there is really no reason to use any other developer unless you want an unusual color that cannot be rendered through toning with gold, platinum or palladium.”

As the bridge to a palladiotype. The same salt at the same strength, with acid added, is what Willis specified for his palladium papers; the course publishes that bath separately as the sodium citrate developer. A printer who has learned to judge development and clearing in citrate on a silver print has learned most of what a palladium print will ask, at a fraction of the cost of learning it in palladium.

  • When you want the developer to choose the colour. The Rochelle salt and borax developers exist because the ratio between them moves the print from sepia to black, and Photographers’ Formulary’s older kit sells all three mixtures from two stocks. This bath cannot do that. It gives one brown and hands the rest to the toner.
  • When the image silver has to be protected during processing above all else. Ware’s argument, set out under The mechanism, is that the kallitype family uses “alkaline-buffered developers of high pH, e.g. Borax” specifically to stop the nitrate anion dissolving the colloidal image silver. A borate bath sits higher on the pH scale than a plain citrate one does, and that is a real difference, with a real cost attached that this page also states.
  • When the metal is platinum rather than silver. Use Willis’s potassium oxalate developer. The redox table under The mechanism says why: Ware’s position is that the citrato-iron couple does not reduce platinum(II), and he notes that nobody has recently checked — using a palladiotype’s citrate developer on a platinotype “seems not to have been tested in modern times”. Palladium is the awkward case, because Willis’s own palladium developer was a citrate; that is discussed under The mechanism and published separately as the sodium citrate developer.
  • When there is no developer at all in the plan. The Van Dyke Brown sensitiser and Ware’s argyrotype print out in the frame and go straight to a wash. You give up shadow depth and you give up the control this tray represents, and you save a bath.
  • When you are reconstructing the process as Nicol first published it. In Nicol’s kallitype the silver is in the developer, not in the paper, and the bath below is not that formula and cannot be substituted for it.

Two hundred grams into seven hundred and fifty millilitres, dissolved, then made up to a litre. That is King’s instruction verbatim, and the order matters for the ordinary reason: a make-up volume is only a strength if the solid goes into part of the water and the rest is added afterwards. Two hundred grams of a salt occupies real volume, and 200 g dissolved in a litre is not a 20 per cent solution.

Photographers’ Formulary reaches the identical strength by a different route — a 300 g packet into 1100 mL, then water to 1500 mL — and warms its water to 52 °C/125 °F to speed the dissolving. King states no temperature. Neither is wrong, because saturation is nowhere near: Chemical Safety Card 1219 gives sodium citrate’s solubility as 77 g per 100 mL, and this bath asks for 20 g per 100 mL, roughly a quarter of what the water will hold. Warm water buys dissolving time and nothing else.

Use distilled water, as both sources specify, and there is a mechanism behind the instruction. Ware’s account of clearing an iron print gives the reaction that hard water sets off in this chemistry:

Ca2+ + Fe(C2O4)33− + 2 H2O → CaC2O4 + Fe(C2O4)2(H2O)2
Calcium displacing iron from its oxalate complex, after Ware

Calcium precipitates as insoluble calcium oxalate and pushes the iron down a chain of aquated species that ends in iron(III) hydroxide. Citrate will sequester calcium — PubChem’s ChEBI description names it a chelator and specifically a calcium chelator — but a developer whose ligand is busy holding your tap water’s calcium is a developer with less ligand for your print’s iron. The sources say “distilled”; this is why.

Nothing here needs a safelight. The developer contains no light-sensitive substance and can be mixed, stored and poured in room light. The tray it is poured into is a different matter, because the sheet going in carries silver nitrate and iron. For the bench work, follow the SOP for weighing a solid and the one for making a solution up to a final volume; for the darkroom, the sensitiser’s own page.

The image arrives before you have finished pouring, and then nothing much happens for ten minutes. King’s figures are the clearest statement of this anywhere: development is “visually complete in about 15-30 seconds, but a development time of 5-10 minutes is important for archival purposes.” The Formulary’s own description agrees on the first half — “the image will appear almost immediately” — and disagrees sharply on the second.

The two sheets that publish this developer disagree by a factor of three about how long to use it, and the disagreement is not sloppiness. King says 5 to 10 minutes and gives permanence as the reason: much of the residual ferric iron “is removed at this stage”. The Formulary says two to three minutes and gives density as the reason: develop for the full two minutes “to allow the developer to react with all available iron and silver in the sensitizer”. Read carefully, those are two different objectives. One is optimising the print in front of you; the other is optimising what the print will look like in fifty years. The course prints both times and does not average them, and the experiment at the foot of this page is how you settle it for your own paper.

The instruction is a century older than either sheet. Wall’s 1912 dictionary, describing a borax and Rochelle salt developer rather than a citrate one, gives the same reasoning in the same order: the print “must be left in the developer for about ten or fifteen minutes. This is necessary in order to ensure that the iron salts shall be all dissolved. The presence of the iron salts is a most prolific cause of failure.” He adds the reassurance a modern printer needs and rarely gets: “the prolonged immersion in the developer will not produce over-development, providing the exposure has been correctly timed.” A kallitype developer is not a silver-gelatin developer and does not run away with the print if you leave it in.

The developer has to arrive everywhere at once. Both sheets warn about it; the Formulary gives the technique. Pour the developer into the tray, tilt the tray so it pools at one end, hold the sheet face up clear of the liquid, and lower the tray so the solution runs across the paper in a single movement. Wall’s 1912 alternative for a large sheet is to put it in face down and turn it over at once, breaking any air bubbles with a fingertip. A print developed in two waves develops in two densities, and the watermark where the front stalled does not come out.

A black deposit forms next to the dark tones. The Formulary describes it exactly — “a residual black substance forming from and next to dark tones” — instructs you to agitate so that it does not settle on the highlights, and notes that it accumulates in the bath and can be filtered out through a coffee filter. No source read identifies what it is. The course’s reading, offered as a reading: the developer is a reducing solution of iron(II) citrate, and any silver nitrate that has diffused off the sheet into it meets that solution in bulk, so finely divided silver reduced in the liquid rather than in the paper is the obvious candidate. That is an inference from where it forms and what colour it is, not a finding, and the experiment at the foot of this page would test it.

Exhaustion shows up as a clearing problem, not as a development problem. This is the single most useful thing to know about the bath in use. A tired citrate developer still produces an image in fifteen seconds — the citrate is nowhere near used up — but the iron it has been accumulating makes the print harder and harder to clear afterwards. King: “if the developer is not replenished, the accumulation of ferrous iron will make it increasingly difficult to clear the print during processing”, which costs “not only an unpleasant stain in the masked areas of the print, but may also decrease permanence, because the stain consists in large part of residual ferrous iron.” The Formulary’s version of the same signal is visual: the bath “becomes clouded with the byproducts of development”.

The print gets weaker in the next tray and stronger in the two after that. King warns about it because a printer who has not been told will over-develop the next sheet: “The image will lighten considerably during clearing, but don’t worry because all the lost density will return during toning and fixing.” What you judge coming out of the developer is not the print.

It is a mild bath and it does not attack the paper. That is not a claim this page makes on its own authority; it is Willis’s stated reason for adopting citrate in the first place. His distributor’s notice says the palladiotype’s citrate developer and clearing baths “have no tendency to injure the beautiful surface of the paper employed for coating”, where prolonged immersion of a matt platinotype in the oxalate and acid baths does destroy its bloom. King’s separate practical criterion points the same way: he judges a paper by whether it clears in four to five minutes, and the Formulary says that a print taking more than ten minutes to clear means the paper is too absorbent and needs sizing. Both are statements about the paper, not about the developer.

Colour: one brown, and King does not pretend to like it. “After development in sodium citrate, the print will have a rather unpleasant brown color, but do not despair. Subsequent processing will change final image color quite dramatically.” That sentence is the honest summary of this developer’s aesthetic contribution. Where the tartrate bath gives sepia and the borate bath gives black from the same exposed sheet, this one gives a starting point.

Why a silver image is brown at all is particle size, not chemistry. Ware’s measurements on plain-paper silver prints put the image particles at colloidal dimensions — around twenty nanometres, far smaller than the wavelengths of visible light — and their colour comes from an absorption that depends on the particles’ “shape, size, state of aggregation and chemical environment” rather than on anything about silver as an element. The same metal in the much larger filaments of a developed silver-gelatin print reads as neutral black.

Why one kallitype developer should therefore give a different colour from another is a reading the course offers and no source read states. If image colour is set by particle size, and if the rate at which silver is reduced sets the balance between making new particles and growing existing ones, then a bath that changes that rate changes the colour. That is consistent with the one thing everybody reports — that the tartrate, borate and citrate baths give sepia, black and brown from one exposed sheet — and it is not a measurement. Nobody read has measured the particle sizes this bath produces, so the page gives the principle and no number.

Contrast: none is offered here, and the reason is a policy rather than an omission. Every contrast control any source publishes for this developer is a chromium(VI) salt added to the tray. The course does not use chromium(VI) at any level; the chromium policy sets that rule out in full and the sensitiser page records the specific figures the course declines to give. What is left is the negative. King’s target is a density range of about log 1.8, reachable in-camera by developing roughly 50 per cent longer than for a grade 2 silver paper, or set directly when the negative is made digitally.

Tonality: long, and set downstream. Because this developer contributes one colour and no contrast control, the tonal decisions all happen after it — in the clearing bath, which lightens the print; in the toner, which King says raises contrast by about a step in a gold bath through loss of density in the high values; and in the fixer, which takes density out of the highlights. A kallitype’s tonal scale is assembled across five baths, and this is the one that does the least to it.

Solarisation, and what the developer cannot fix. King reports that untoned kallitypes frequently show tone reversal in heavily exposed shadows — with increasing exposure the shadows get lighter — and that toning with gold, platinum or palladium counteracts it. Nothing about this developer causes or cures that, and no adjustment to it is offered as a remedy, because none is published.

Four steps, of which the developer performs exactly one.

Step one, in the printing frame: light makes an insoluble solid. Ultraviolet light decomposes iron(III) oxalate, reducing the iron at the expense of an oxalate ligand which leaves as carbon dioxide. Döbereiner saw this in 1831, more than forty years before Willis turned it into a printing process:

Fe2(C2O4)3 + UV → 2 FeC2O4 + 2 CO2
Döbereiner's photolysis, after Ware

The product, iron(II) oxalate, is barely soluble — Ware gives 0.022 g per 100 cc of water — which is why the kallitype prints out so faintly and why it needs a developer at all. The reducing agent exists, in the right places, in the right amounts, and it cannot move.

Step two, in the tray: citrate dissolves it. Ware states the general principle for the platinotype and it holds here: the photoproduct “cannot reduce platinum(II) or palladium(II) salts in aqueous solution to the metal unless it is solubilised by complexation”. The platinum developer complexes it with oxalate; this one complexes it with citrate.

FeC2O4 + C6H5O73− → [FeC6H5O7] + C2O42−
The developer's only reaction: an insoluble oxalate exchanged for a soluble citrate

Step three, still in the tray: the dissolved iron(II) reduces silver. Wall’s 1924 statement of the whole process is one sentence and it is still correct: “the ferrous salts dissolve in the developer, and reduce silver nitrate to the metallic state at the points where the ferric salt has been reduced by light.”

[FeC6H5O7] + Ag+ → FeC6H5O7 + Ag
One electron, one silver atom

The silver deposits where the ferrous ion was, which is where the light was. The image is metallic silver in the paper fibres, not in a binder, and it is the smallness of those particles that makes it brown.

Step four, over the next several minutes: the developer keeps working on iron that was never exposed. This is the part that separates a print that survives from one that does not, and it is why the tray time is twenty times the image time.

Deeper: what complexing the iron does to what it can reduce

Section titled “Deeper: what complexing the iron does to what it can reduce”

Complexing iron changes what it can reduce, and the numbers are worth having in one place. Ware’s figures, all against the standard hydrogen electrode:

Couple Potential What it means
Fe3+/Fe2+, uncomplexed +0.771 V iron(II) is a poor reducing agent
FeIII(C2O4)33−/FeII(C2O4)22− +0.02 V the oxalate developer: a strong reductant
FeIIICit/FeIICit +0.372 V this developer: a moderate one
Pd(II) as PdCl42− +0.62 V palladium
Pt(II) as PtCl42− +0.73 V platinum
Ag+/Ag +0.80 V silver
Au(III) as AuCl4 +1.00 V gold

A couple reduces a metal that sits above it, and the size of the gap is the driving force. Ware introduces the table as showing that “the ease of reduction is reflected in the relative values of the redox potentials”, and it is worth reading it as the ranking he says it is rather than as a threshold.

What the oxalate developer has is margin. At +0.02 V the oxalato-iron couple sits at least 0.60 V below every metal in the table. One bath therefore develops platinum, palladium, silver and gold alike, and that is the chemical reason the platinotype had to wait for ferric oxalate instead of using Herschel’s far cheaper citrate: ferrous citrate, Ware writes, “is not a sufficiently powerful reducing agent to enable a similar platinum printing process.”

What the citrate developer has is enough for silver. Its margin against silver is 0.80 − 0.372 = 0.43 V, and against gold 0.63 V, and Ware states the outcome directly: the citrate complex “does not reduce platinum(II) or palladium(II), although it will reduce gold(III) and silver(I), under the printing conditions.”

Be careful about how much those numbers settle on their own, because it is less than it looks. The margins against palladium (0.25 V) and platinum (0.36 V) are the two smallest in the table, but they are still positive: thermodynamics by itself does not forbid either reduction, and Ware’s qualifying phrase is “under the printing conditions”. He makes the general point in the same section — that a reduction the thermodynamics permits can still be too inert kinetically to yield an image “within the short time of a few minutes that is available”. So the potentials rank the four metals by how easily they are reduced; where the usable line actually falls is settled by kinetics and by the few minutes a developer gets. The sensitiser page works through the same table for the neighbouring question of why a Van Dyke prints out and a kallitype does not.

The consequence for this formulary is lopsided, and it is why this page exists at all. Silver is the second-easiest metal in the table to reduce, so a kallitype can be developed in citrate, in tartrate or in borate and each gives a different picture. Platinum is the second-hardest, so it gets one developer, and the arguments about it are only over whether to acidify.

Deeper: the iron that was never exposed, and why it decides whether the print lasts

Section titled “Deeper: the iron that was never exposed, and why it decides whether the print lasts”

Only a fraction of the iron on the sheet is reduced by light. The rest is iron(III), sitting in the paper, and it is the reason kallitypes have the reputation they have.

Ware sets out its behaviour in three stages. First, some of it chemisorbs — binds chemically to the hydroxyl groups of the cellulose — so it is not simply rinsed away. Second, above pH 4 iron(III) hydrolyses:

Fe3+ + 3 H2O → Fe(OH)3 + 3 H+
Hydrolysis of iron(III), which begins above pH 4

forming colloidal iron(III) hydroxide that lodges in the fibres. Third, and this is the sentence that matters, freshly formed iron(III) hydroxide “can be redissolved in dilute acids initially”, but if it is not removed it “slowly transforms irreversibly into a highly insoluble polymeric form, iron(III) oxyhydroxide, FeO(OH) — the mineral called Goethite — which is quite insoluble in dilute acids”. Ware’s conclusion is unambiguous: it is essential to remove all the iron(III) at the wet processing stage, before the print dries.

What that iron does if it stays is the second half of King’s permanence argument: “even very small quantities of residual ferrous iron will eventually oxidize the silver, and the image will fade.” Ware documents the parallel cycle for the paper itself rather than for the image: iron(III) trapped in a sheet can be photochemically reduced to iron(II), air can reoxidise it — “slowly in acidic and rapidly in basic solution” — and the intermediates include free radicals that cut cellulose chains. Whether the two mechanisms are one mechanism is not something the course has read an answer to; what both accounts agree on is that iron left in the paper is not inert. The Getty’s conservation atlas records that residual iron is still detectable by X-ray fluorescence in platinum prints after the best clearing procedures available, which tells you how hard it is to get out even when a professional laboratory is trying.

Against that background, look again at what a 20 per cent citrate developer is. It is 0.68 molar in a ligand that binds iron(III) strongly — the same ligand, in the form of ammonium iron(III) citrate, that carries iron(III) onto paper in the first place in the Van Dyke and the cyanotype. Every minute the print spends in it is a minute in which iron(III) is being pulled off the cellulose and into solution as a soluble complex, instead of hydrolysing into something that will not come off later. That is what King means by “much of the residual ferric iron… is removed at this stage”, and it is the reason his development time is five to ten minutes rather than thirty seconds.

There is one further mechanism at work, and it is a trade-off rather than a benefit, so the page states both halves.

Ware’s criticism of the whole iron-silver family begins with the anion nobody thinks about: “nitrate is an oxidising anion, and tends to dissolve the colloidal image silver during wet processing, especially under acidic conditions. To minimise this loss of image the Kallitype process employs alkaline-buffered developers of high pH, e.g. Borax.” So the historical reason kallitype developers are alkaline is silver protection.

He then names the price: “Alas, these create a new problem, because they cause hydrolysis of the excess iron(III) in the sensitizer and the deposition of insoluble ferric hydroxide in the image, which ultimately causes it to fade.”

A sodium citrate developer sits between the two horns of that dilemma, and the course reads it as a deliberate compromise even though no source says so in those words. The salt’s own solution is slightly alkaline — HSDB records the dihydrate’s aqueous solution as slightly alkaline to litmus at about pH 8 — which is on the protective side for the silver, if less so than a borate bath. And because citrate is a chelator, the iron(III) it liberates is captured as a soluble complex instead of hydrolysing to the ferric hydroxide Ware warns about. The bath is alkaline enough to help the silver and complexing enough to keep the iron in solution, which is a bargain neither borax nor plain water can strike.

Two honest qualifications. No source read publishes a measured pH for this bath, and the figure above is a property of the salt rather than a measurement of the developer. And the compromise is real: a borate developer is more alkaline than this one, so on Ware’s argument it protects the image silver better during wet processing, and the printers who use it are not making a mistake. What they are buying that protection with is the iron.

The last link in the chain is the rinse, and it is why King is so specific about it. His rule is that the first rinse after development must be neutral or slightly acidic, because “if the first rinse is alkaline, ferrous hydroxide compounds may be formed in the paper, making complete clearing difficult or impossible.” Everything the developer has just achieved by holding iron in solution can be undone in sixty seconds by rinsing in alkaline tap water, which precipitates it again inside the sheet. The citric acid clearing bath that follows is the same logic continued: acid suppresses hydrolysis, and citrate chelates what is left.

Trisodium citrate dihydrate, 200 g per litre — and it is the only thing in the bottle, so this section has to earn the whole of the page.

What it is. The sodium salt of citric acid with all three carboxyl protons removed. Three carboxylate groups and a hydroxyl on a three-carbon backbone, arranged so that the molecule can close a ring around a metal ion from several directions at once — PubChem’s ChEBI entry names its identity outright: a chelator. Formula weight 294.10 as the dihydrate, 258.07 as the anhydrous salt, CAS 6132-04-3 and 68-04-2 respectively. It is a food additive, E331.

Why it is here. Because the reducing agent that light made is insoluble and has to be got into solution before it can do anything, and because the iron that light did not touch has to be got out of the paper before the print dries. Citrate does both jobs with the same property. It is emphatically not a developing agent: it reduces nothing, it is not oxidised in the process, and there is no redox chemistry in the bottle at all until a sheet of paper goes into it.

What it does chemically. Three things, in this order.

  1. It exchanges ligands with iron(II) oxalate, converting a solid whose solubility Ware gives as 0.022 g per 100 cc into a dissolved complex that can diffuse and can meet a silver ion.
  2. It complexes the iron(III) that develops out of that reaction, and the iron(III) that was never exposed at all, holding both in solution rather than letting them hydrolyse into the colloidal hydroxide that becomes goethite on drying.
  3. It buffers, weakly and on the alkaline side. Citric acid’s third dissociation constant is pKa3 5.68 at 20 °C on the IUPAC compilation, so a solution of the fully deprotonated salt with no added acid sits well above that — about pH 8 on HSDB’s reading of the dihydrate. That is enough to keep the bath off the acidic side where, on Ware’s argument, the nitrate anion attacks the image silver hardest.

What photographic consequence follows. An image that appears in fifteen to thirty seconds, a brown that King calls unpleasant, no contrast control, and a print that clears in four or five minutes if the paper is right. The developer’s contribution to the finished picture is almost entirely negative in the photographic sense: it is what it removes — iron — that decides whether the print is still there in fifty years.

What happens with more or less.

  • Less citrate means less ligand for the same iron, and no source read says where the bath stops working, so this page states no floor. The arithmetic above says the stoichiometric margin is enormous: at half this strength the citrate would still outnumber one print’s iron by more than a hundred to one. Willis’s own comment on his citrate developer, in the 1880 patent, nonetheless pushes the other way — “the strength of the solution may be varied, but the best results will usually be obtained by a strong or saturated solution” — which is a careful nineteenth-century chemist saying that concentration buys something other than stoichiometry. What, he does not say, and neither does this page.
  • More citrate buys more ligand and more room before the dissolved iron builds up, and 20 per cent is nowhere near the 77 g per 100 mL the water will hold. What it costs is unknown, because nobody read published a stronger kallitype developer to compare against. It is the cheapest experiment on this page.
  • Acid added turns this bath into Willis’s palladiotype developer, which is a different formula with its own entry and its own consequences — see Variants.
  • The wrong grade weighed shifts the citrate by 14 per cent in either direction. Invisible here; not invisible if you are also adding acid.

What it interacts with. Iron in both oxidation states, which is the point; calcium, which is why the water is distilled; silver, which it does not complex in any way the sources discuss but which it is in the same tray with; and the citric acid of the clearing bath that follows, which is the same molecule with its protons back on and which continues the same chemistry at a lower pH.

With the sensitiser it develops. This bath is specified for the two-solution kallitype sensitiser — 10 per cent silver nitrate and 20 per cent ferric oxalate mixed in equal parts — and the pairing is not arbitrary. The sensitiser’s ligand is oxalate and the developer’s is citrate, and the exchange between them is the reaction the developer exists to perform. Put this developer on a Van Dyke, whose iron is already citrate and which prints out without a developer, and there is no insoluble photoproduct for it to dissolve.

With the water on the way in. Calcium in tap water competes for the developer’s ligand and, on Ware’s equation above, pushes iron toward hydrolysis. Both sources specify distilled water for the make-up. Neither says what to do about a tap-water rinse, and King’s rule that the first rinse must be neutral or slightly acidic is the only guidance published on the point.

With the rinse and the clearing bath on the way out. These three baths are one chemical operation in three stages, and the developer’s work is undone if the second stage is wrong. Development pulls iron into solution as a citrato complex; an alkaline rinse precipitates iron hydroxides back into the sheet; a 3 per cent citric acid bath dissolves what the developer left, at a pH low enough to suppress hydrolysis. The course’s iron-silver clearing sequence covers the more thorough EDTA and sulfite route that platinum practice uses.

With the toner, which comes before the fixer. King tones after clearing and before fixing, and one of his three reasons is that a toned print does not bleach in the fixer. Nothing about this developer changes that sequence, but the amount of silver the developer put down is what the toner has to work with: a thinly developed print has less silver to exchange and tones to a different place.

With the fixer, which takes some of it away. A kallitype loses density in sodium thiosulfate, and Ware gives the mechanism for plain-paper silver images generally: colloidal silver “is easily oxidised by air in the presence of thiosulphate ions”, so over-long immersion, too concentrated a bath, or oxidising impurities in the water all cost image density. Bostick & Sullivan state the practical consequence for the kallitype in one line: prints “will bleach in the fix, so either tone before fixing or overprint by a stop or two.” The dilute alkaline thiosulfate fixer is the course’s entry for that bath, and a 1 per cent sodium sulfite washing aid is King’s step after it.

With an oxidising agent, which is what a contrast control is. Ware explains what these additives do in a siderotype and the explanation is worth having even though the course publishes no dose: an oxidant “reoxidise[s] some of the iron(II) photoproduct, thereby making it unavailable for reducing” the metal, “so truncating the exposure scale of the process”. Contrast is bought by destroying the weakest parts of the latent iron image before the developer can reach them. Ware, writing about potassium chlorate in a platinotype sensitiser, notes that the route “can cause a deterioration in image quality and an increase in image graininess”; King, writing about potassium dichromate in this very developer, reports independently that too much of it makes the image “take on a granular look”. Two authors, two different oxidants, two different processes, the same artefact — which is a good reason to believe the artefact belongs to the mechanism rather than to the reagent.

With air, which is the slow version of the same reaction. Ware gives it as an equation, and it runs in the tray, in the bottle and in the finished print:

4 Fe2+ + O2 + 4 H+ → 4 Fe3+ + 2 H2O
Aerial reoxidation of iron(II), after Ware — slow in acid, rapid in base

In the developer this is a mild nuisance: iron(II) that reoxidises before it meets a silver ion is image density lost. In a finished print it is the whole permanence problem, because the iron(III) it produces goes on to oxidise the silver.

The acidified version, which is Willis’s palladiotype developer. Ware records that Willis specified for his Palladiotype papers “trisodium citrate (20% w/v) acidified with added citric acid (2% w/v)”, and in the chapter where he converts the original imperial units gives the acid as 2.2 per cent w/v against a citrate dihydrate of formula weight 294.10. Same salt, same strength, plus acid. The course publishes it separately as the sodium citrate developer, because it is a formula for a different metal with its own provenance and its own times.

What the acid does is move the bath down the pH scale toward citric acid’s third dissociation, which suppresses iron(III) hydrolysis further. What it costs, on Ware’s argument about nitrate, is some of the protection an alkaline bath gives to colloidal silver — which matters for a silver print and does not matter at all for a palladium one. That is a coherent reason for the two baths to differ, and it is the course’s reading rather than a statement any source makes. Nobody read has tested the acidified citrate developer on a kallitype, and this page does not recommend it; it is named because a printer who finds it in the platinum literature should know why it is not the formula above.

The contrast-controlled version, which this course does not publish. Both sheets add a chromium(VI) salt to this developer to raise contrast — King specifies a 5 per cent potassium dichromate solution at between 1 and about 16 mL per litre, with a stated effect on the range of negatives that can be printed, and the Formulary notes the same additive without a quantity. The sensitiser page sets out the figures the course declines to give and the chromium policy sets out the rule. Two things are worth adding here that belong to the developer rather than to the sensitiser.

First, the mechanism is not mysterious and does not require chromium: any oxidant that reaches the iron(II) before the silver does will truncate the exposure scale, which is why potassium chlorate and hexachloroplatinate(IV) appear in the same role in the platinum literature. The course offers none of them as a substitute, because no source read gives a quantity for any of them in a citrate kallitype developer, and Rule 6 forbids guessing at one.

Second, the most authoritative voice in this literature declines the whole category on grounds that have nothing to do with safety. Ware, writing about his own platinum and palladium work: “With the availability today of better controls for modern negative-making, especially by digital means, contrast-enhancing agents, such as chlorate, hexachloroplatinate(IV) or dichromate, become unnecessary when a correctly calibrated negative is made. In the present work, to avoid introducing yet another variable, the use of these undesirable image-degrading agents will be avoided.” The course’s position and Ware’s arrive at the same practice from different directions.

No safer course variant of the quantities is offered, and there is nothing to make safer: the bath is one unclassified salt in water. The variants that matter are the ones above, and both of them are somebody else’s formula rather than a modification of this one.

Level A, and this is one of the few baths in the formulary where that is not a close call. Against the course’s rubric: the substance is not classified as hazardous on the ECHA entry for the dihydrate, where 520 of 536 company reports find no criterion met; the bath is used at room temperature; nothing is heated beyond the 52 °C the Formulary suggests for dissolving; no vapour is produced; and the waste is a collectable liquid. There is no criterion in the Level B list that this solution meets.

What is not a hazard here, and why. Sodium citrate is a food additive. EH40 sets no workplace exposure limit for it, and EH40 says in terms that absence from its list does not indicate that a substance is safe — so the absence is reported rather than leaned on. There is no acid and no alkali strong enough to burn, no oxidiser, no solvent, nothing that off-gasses, and no incompatibility that produces a gas. The one hazard statement anywhere in the record comes from a second ECHA entry filed against the anhydrous salt under the looser name “Citric acid, sodium salt”, which reports Warning, H319, causes serious eye irritation in 72 per cent of 148 reports. That entry covers a less well-defined article than trisodium citrate, and the course reports it rather than adopting it — but it is why eye protection appears below for a substance whose own classification box is empty. Chemical Safety Card 1219 does the same thing from the other direction: an empty classification box and a text recording irritation to the eyes and the respiratory tract.

Personal protection. Nitrile gloves, on the general basis that a darkroom hand goes into several trays in a session and this is the least of them; safety glasses as a minimum, and splash goggles while mixing the 20 per cent stock, which is the control the encyclopaedia takes by preference on the strength of the one ECHA entry that does give H319. Weigh without raising dust — Card 1219 words its ventilation advice oddly and usefully, “use ventilation (not if powder)”, meaning that a draught over a powder is the wrong answer and containment is the right one.

The dry salt is stable in air, and Chemical Safety Card 1219 sets no storage requirement for it at all — which is more than can be said for several of its neighbours in a darkroom cupboard. It is combustible, and it decomposes at 150 °C — a figure the card is careful to describe as an apparent melting point caused by the loss of the water of crystallisation rather than a real one. Keep it dry, closed and labelled, in a container that has never held food.

The made-up stock has no published keeping time and the sources do not treat it as a keeping problem. Photographers’ Formulary’s instruction is to “save the contents of the developer after each printing session for reuse and discard after the contents of the kit have been used up” — a bath retired by what is in it, not by a date. King gives keeping times for the ferric oxalate stock and says nothing at all about the developer. The course does not supply a figure to fill the gap. What it will say is that the two published tests are both compositional: the Formulary retires a bath that has “become clouded with the byproducts of development”, and King retires the settled fraction at the bottom of the bottle every time he replenishes.

Store it in a bottle you can decant from without disturbing the sediment, because that is what King’s replenishment procedure requires: pour from the top, discard what has settled. A wide-necked container that has to be tipped hard defeats the method. Dark storage is not needed — nothing in the bottle is light-sensitive until iron and silver arrive in it — but a used bath is best kept closed, since aerial oxidation of the dissolved iron(II) is a real reaction and the equation for it is under Interactions.

Hard water and anything calcareous. Calcium is the practical incompatibility of this bath and the reason both sources specify distilled water: it precipitates as calcium oxalate on meeting the oxalate the print brings in, and it occupies citrate that the iron needs. Ware names two routes for it into the chemistry: hardness in the water, and the chalk buffer in the paper — a sheet buffered with calcium carbonate brings its own supply into the tray, which is one more reason a paper that suits a kallitype has to be found by testing rather than by reputation.

Strong oxidisers, which is the general incompatibility of any organic salt and the specific one here for a reason that is photographic as well as chemical: an oxidiser in this tray destroys the iron(II) that makes the picture. That is exactly what the dichromate contrast control is for, which is a useful reminder that “incompatible” and “useless” are not the same word.

Acid, in the sense that adding it makes a different formula. Citric acid does not react destructively with sodium citrate; it converts part of it to hydrogencitrate and gives a buffer at lower pH. That is Willis’s palladiotype developer, and the point is that a bath topped up with the clearing solution by accident is no longer the formula on this page.

Nothing else of consequence. Card 1219 records only that the aqueous solution is a weak base. There is no incompatibility here that produces a gas, a fire or a violent reaction, and saying so plainly is more useful than a list padded to look thorough.

One contamination route that is not chemistry. The tray, the tongs and the graduate that hold this developer will also have held silver nitrate on paper. Keep them to the process, not to the stop bath shelf of a silver-gelatin darkroom, and keep the brush that coated a kallitype away from anything that coated a cyanotype: ferricyanide and silver in one dish is a different picture entirely.

What leaves the tray is not what went into it. The spent bath carries dissolved iron in both oxidation states as citrato complexes, silver — some of it dissolved from the sheet as nitrate, some of it the black precipitate — and the citrate itself, which is a food additive. The metals are why the container is collected; the citrate is not the issue.

pH is the wrong test, and it is worth understanding why. Kodak’s J-52 guidance gives 5.6 to 9.4 as the pH window sewer codes most frequently set, and a citrate bath at about pH 8 sits comfortably inside it. A solution can pass that test and still be the wrong thing to pour away, because what makes this stream a waste stream is the metal it carries rather than its acidity, and no pH meter will tell you anything at all about the silver.

In practice. Collect the developer with the first rinse and the clearing baths, which carry the same metals; label the container with what is in it and the date; and follow the disposal caveat and your local authority’s route. ILFORD’s published route for domestic users in the United Kingdom is a household waste and recycling centre’s chemical cupboard. Local regulation governs, everywhere, and this page does not give a jurisdiction-specific instruction.

The black deposit is worth keeping separate. If the Formulary’s coffee-filter suggestion is followed, what stays on the filter is the most silver-rich thing in the session by concentration. It belongs in the solid waste you collect rather than in the bin, and if you are running enough prints to make silver recovery worth attempting, that filter is where to start.

A stain in the areas that were masked, which will not clear. The commonest fault, and King names the cause: an unreplenished developer, whose accumulated ferrous iron makes the print progressively harder to clear. Replenish at his rate, decant from the top, discard the settled fraction. If the stain persists in a fresh bath, the paper is the suspect rather than the developer — King’s own test is that a paper needing more than about four minutes to clear is unsuitable, and the Formulary’s threshold is ten minutes before it tells you to size the sheet.

A print that will not clear at all after an apparently normal development. Check what you rinsed in. An alkaline first rinse forms ferrous hydroxide compounds in the paper and King’s warning is that this makes complete clearing “difficult or impossible”. Neutral or slightly acidic water, for one to two minutes.

Streaks, tide marks or a band of different density. The developer did not arrive at once. Use the tilted-tray pour, or immerse a large sheet face down and turn it immediately. This fault is made at the moment of pouring and cannot be fixed later in the sequence.

Grey or black specks in the highlights. The residual black substance the Formulary describes, settled on the sheet. Agitate constantly, and filter the bath when it accumulates.

The bath has gone cloudy. The Formulary’s own trigger for pouring out and replacing all or part of the developer. It is a compositional signal, not an age one.

The image looks weak coming out of the developer. Before adjusting anything, remember what the next three trays do: the print lightens considerably during clearing and gets that density back during toning and fixing. King warns about this explicitly because the natural response — developing the next one longer — is the wrong one. If the print is genuinely thin after fixing, the exposure is the suspect: there is little print-out image to judge by, so exposure is found by test strip.

The highlights have bleached. That is the fixer, not the developer. Bostick & Sullivan give the two published answers — tone before fixing, or overprint by a stop or two — and note that warm water accelerates it; the Formulary adds that acidic water does too.

The image has a granular look and exposures have got much longer. If a dichromate contrast control is in the bath, King attributes exactly this pair of symptoms to too much of it, and Ware reports the same graininess for the equivalent oxidant in platinum printing. The course’s answer is not a smaller dose; it is not using it and calibrating the negative instead.

Development takes noticeably longer than fifteen to thirty seconds to show an image. Look upstream. The developer is a ligand in vast excess and is unlikely to be the limit; a degraded ferric oxalate stock, an under-exposed sheet, or a coating that was force-dried with heat are all documented causes of a weak or fogged print in the sources for the sensitiser.

Settle the two published times, and measure the half nobody measures. One negative, five identical sheets, one fresh bath: develop for 30 seconds, 1, 2, 5 and 10 minutes, then clear, tone, fix and wash all five identically. The sensitiser page proposes reading the maximum density from such a series; the reading that belongs to the developer is the second one — how long each sheet takes to clear afterwards, timed with a clock rather than judged. King’s argument predicts a clearing time that keeps falling long after the density has stopped rising; the Formulary’s implies there is nothing left to gain after two minutes. Record both columns, because the gap between them is the whole disagreement.

Find out what the black deposit is made of, or at least where it comes from. Coat two sheets identically. Expose one normally; give the other no exposure at all. Develop each in its own small volume of fresh developer, side by side, for the same time. If the deposit forms in both trays it is coming from unexposed silver diffusing off the sheet; if it forms only where there is an image it is coming from the developed print. Either answer is more than any source read publishes, and neither requires a reagent you do not already have.

Test the replenishment claim directly. Run twelve 8 by 10 prints from one negative through a single tray. Replenish at King’s rate for the first six and not at all for the second six, and record only one number for each print: the time it takes to clear. The arithmetic under The mechanism predicts that development time will not change while clearing time climbs, because the citrate is never the limiting reagent. If clearing time stays flat too, the arithmetic is wrong and you should say so.

A strength series, which nobody has published. Three baths from the same tin — 10, 20 and 30 per cent w/v — and three prints from one exposure, developed for the same time and cleared identically. Twenty per cent is the control because it is the only strength with provenance. Look for the difference where the mechanism says it should be: not in the speed at which the image appears, but in how long the print takes to clear afterwards. Record the result as your own, not as a formula.

Warm against cold. Willis said his citrate developer could be applied “cold, warm, or hot, but preferably hot”, and Photographers’ Formulary says of its own tartrate-and-borate black-tone developer that it “works best if it is warm, (around 38°C/100°F or higher)”. Nobody publishes a temperature for the citrate kallitype developer. Two prints, one bath at room temperature and one at 38 °C, everything else identical. Watch the highlights and the clearing time, and be aware that you are testing a variable the sources chose to leave out.

Compare it against the developer it replaced. One exposure, two sheets: this bath, and the Rochelle salt or borax developer of the older kit. Judge the colour wet, judge it again dry, and then judge it a third time after both have been toned identically. King’s claim is that after toning there is no reason to use anything but the citrate; that claim is testable, and the third comparison is the one that tests it.

Track a print for a year. Two prints from one negative, developed identically, one toned and one not, both cleared and washed to the same standard, both kept in the same conditions. King’s prediction is unambiguous — the untoned one will eventually fade, because residual iron oxidises the silver. Photograph both against a step wedge on the day they dry and again every three months. This is the experiment that tests the reason for everything else on this page, and it is the only one that cannot be hurried.

Sources for this page

17 cited · checked 2026-09-06

  1. 01Making Kallitype Prints: A Fresh Look at a Beautiful Printing ProcessSandy King§ Necessary Materials — 2) Developer, for the 200 g make-up; About My Method, for the one-developer principle; Notes on Image Permanence, for residual iron and the oxidation of image silver; and, on page two, Working Procedures step 4 Development, for the pour, the 5 to 10 minutes against 15 to 30 seconds, the replenishment rate, the dichromate contrast range and the "rather unpleasant brown color"; step 5 First Rinse, for the neutral-or-acid rule; step 6 Clearing; Refinements to the Process, for the recommendation to use no other developerunblinkingeye.com/Articles/Kallitype/kallitype.htmltier 2, specialist2026-09-06
  2. 02Photographers' Formulary New Kallitype Printing Kit, catalogue number 07-0075: instructionsPhotographers' Formulary, Inc.§ Chemicals contained in this kit; Mixing the solutions — Developer Solution, for the 300 g in 1500 mL make-up, the 52 °C water and the 500 mL replenisher; Development, for the two to three minutes, the tilted-tray pour, the black residue and the dichromate note; Clearing; Fixing; Final Notes and Suggestionsphotoformulary.homestead.com/07-0075_New_Kallitype.pdftier 1, primary2026-09-06
  3. 03Photographers' Formulary Kallitype Printing Kit, catalogue number 07-0070: instructionsPhotographers' Formulary§ Chemicals contained in this kit; The Developer Stock Solutions; Development, for the three Rochelle salt and borax developers this one replaces and their times and temperaturesfreestylephoto.com/pdf/product_pdfs/formulary/FormularyKallitype.pdftier 1, primary2026-09-06
  4. 04Traditional Kallitype Printing: Dick Stevens' FormulasBostick & Sullivan, Inc.§ Your Kit contains; Developing; Clearing; Fixing; Toning, for the third supplier's practice, which uses a proprietary black-tone developer rather than a citrate onebostick-sullivan.com/wp-content/uploads/2022/03/KallitypePrinting.pdftier 1, primary2026-09-06
  5. 05Platinomicon: A Technical Account of Photographic Printing in Platinum and PalladiumMike Ware, 2017§ 1.4 Willis's three components, for the developer defined as the bath that dissolves the photochemically formed ferrous oxalate, and for ferrous citrate being too weak a reductant for platinum; 2.8 and 2.9 Processing of Platinotype and Palladiotype, for Willis's trisodium citrate developer at 20 per cent w/v with citric acid, the 5 per cent potassium citrate demonstration of 1917, the potassium oxalate developer it replaced and the reason Willis gave for the change; 6.5 Agents for increasing contrast, for the oxidant mechanism, the graininess it causes and Ware's own refusal to use such agents; 10.3 Hydrolysis and precipitation; 10.5 Coordination by oxalate, for the oxalate formation constants and the redox potential of the oxalato couple; 10.9 Chelation of iron; 10.10 Chemistry of clearing siderotypes, for iron(III) hydrolysis above pH 4, chemisorption to cellulose, the irreversible transformation to goethite and the calcium oxalate equation; 11.1 Photochemistry of iron(III) oxalates, for the Döbereiner reaction and the 0.022 g per 100 cc solubility of iron(II) oxalate; 11.3 Siderotype by reduction of noble metals, for the citrato couple at +0.372 V and the table of noble-metal potentials; Appendix VII.3, the specification of Willis's British patent No 1117 of 15 March 1880, for the citrate of soda developermikeware.co.uk/downloads/Platinomicon.pdftier 2, specialist2026-09-06
  6. 06The Argyrotype ProcessMike Ware§ Structure and Stability of Silver Images; An Alternative Silver Salt, for nitrate as an oxidising anion that dissolves colloidal image silver, for the kallitype's answer of alkaline-buffered developers of high pH such as borax, and for the ferric hydroxide those developers deposit in the imagemikeware.co.uk/mikeware/Argyrotype_Process.htmltier 2, specialist2026-09-06
  7. 07Argyronomicon: Silver Photographs on Paper — Chemical History of their Invention, Deterioration, and ConservationMike Ware, 2019§ 7.5.4 Thiosulphate fixation, for the oxidation of colloidal image silver by air in the presence of thiosulphate; 9.3 Coating Weight and Particle Size, for the particle radius and the small fraction of applied silver that survives in a plain-paper printmikeware.co.uk/downloads/Argyronomicon.pdftier 2, specialist2026-09-06
  8. 08The Dictionary of Photography and Reference Book for Amateur and Professional Photographers, 9th editionE. J. Wall, edited by F. J. Mortimer, 1912§ Kallitype, for the four developers of 1912, for the ten to fifteen minutes in the developer "in order to ensure that the iron salts shall be all dissolved", for the statement that the presence of the iron salts is a most prolific cause of failure, and for the face-down immersion of a large sheetarchive.org/details/dictionaryofphot1912walltier 1, primary2026-09-06
  9. 09Photographic Facts and FormulasE. J. Wall, F.C.S., F.R.P.S., 1924§ Kallitype, in The Iron Processes, for the statement of the mechanism and for Hall's formulas, whose clearing bath is sodium citrate with citric acidarchive.org/details/photographicfact00walltier 1, primary2026-09-06
  10. 10The Atlas of Analytical Signatures of Photographic Processes: PlatinotypeDusan C. Stulik and Art Kaplan, 2013§ The steps of the platinotype process, for the developer's role and for contrast adjusted with potassium chlorate or dichromate; XRF detection of residual iron after the best clearing proceduresweb.archive.org/web/20220121043200id_/https://www.getty.edu/conservation/publications_resources/pdf_publications/pdf/atlas_platinotype_corrected.pdftier 1, primary2026-09-06
  11. 11PubChem compound summary: Sodium Citrate Dihydrate (CID 71474)National Center for Biotechnology Information§ Identity, computed properties, CAS and the aggregated ECHA C&L notificationspubchem.ncbi.nlm.nih.gov/compound/71474tier 1, primary2026-09-06
  12. 12PubChem compound summary: Trisodium citrate (CID 6224)National Center for Biotechnology Information§ Molecular weight and CAS of the anhydrous grade; the second ECHA entry filed as "Citric acid, sodium salt", which does carry a hazard statement; the HSDB entry giving the pH of the dihydrate solutionpubchem.ncbi.nlm.nih.gov/compound/6224tier 1, primary2026-09-06
  13. 13International Chemical Safety Card 1219: Sodium citrate dihydratePrepared by an international group of experts on behalf of the International Labour Organization and the World Health Organization, with the financial assistance of the European Commission§ Physical and chemical information, for solubility and for the solution being a weak base; chemical dangers; effects of short-term exposure; the empty classification and occupational-exposure-limit boxesinchem.org/documents/icsc/icsc/eics1219.htmtier 1, primary2026-09-06
  14. 14IUPAC Digitized pKa Dataset, high-confidence subset v2.3International Union of Pure and Applied Chemistry, Dissociation Constants project; digitised from the Serjeant and Dempsey and Perrin compilations, 2024§ Citric acid, pKa1 to pKa3 at 20 degrees C in 0.1 mol/L sodium perchlorategithub.com/IUPAC/Dissociation-Constantstier 1, primary2026-09-06
  15. 15EH40/2005 Workplace exposure limits, containing the list of workplace exposure limits for use with COSHHHealth and Safety Executive, 2005§ Table 1, searched for citrates; introduction, paragraph 6hse.gov.uk/pubns/priced/eh40.pdftier 1, primary2026-09-06
  16. 16Disposal of Small Volumes of Photographic-Processing Solutions, publication J-52Eastman Kodak Company, 1986§ Characteristics of photographic-processing effluents, for the sewer pH windowp2infohouse.org/ref/30/29045.pdftier 1, primary2026-09-06
  17. 17General health and safety adviceHARMAN technology Limited (ILFORD Photo)§ Waste disposal for photographic products, domestic usersilfordphoto.com/health-and-safetytier 1, primary2026-09-06

Formulas, hazard statements, historical dates and process descriptions on this page were checked against the sources above on the date shown. Safety data changes: obtain the current safety data sheet for the product you actually buy before you open it.