Sodium citrate developer

William Willis spent forty years selling one developer. He patented it in 1873, printed a strength for it in 1880, and his company supplied it with every sheet of Platinotype paper it ever sold. Then in 1917, when the platinum embargo forced him to launch a palladium paper, he did not put potassium oxalate in the box. He put in an acidified solution of citrate of soda.

That decision is the whole of this page. The two baths look interchangeable — one salt of a dicarboxylic or tricarboxylic acid, dissolved in water, doing a job that involves no reduction at all — and they are not. One of them fogs a palladium print and the other does not.

Ingredient	Quantity	Form the source specifies
Trisodium citrate dihydrate	232.5 g	Wall writes "Sodium citrate" and names no grade. Ware, converting the manufacturer's own instructions, names it: trisodium citrate dihydrate, Na3C6H5O7·2H2O, formula weight 294.10, which is the encyclopaedia's entry and the article normally sold. The anhydrous salt, formula weight 258.07, carries 14 per cent more citrate per gram and is not what this figure assumes. Weighed as the dihydrate, 232.5 g is 0.79 mol of citrate
Citric acid (anhydrous)	23.25 g	A tenth of the citrate by weight in both printings. Wall names no grade; the encyclopaedia's entry is the anhydrous acid at 192.12, and the monohydrate at 210.14 would need 1.094 times the weight for the same acidity. Weighed as the anhydrous acid, 23.25 g is 0.121 mol
Water	1000 mL, added	"Water 1000 ccm" as Wall prints it, with no "to make", so it is recorded as water added and the finished bath is more than a litre. Distilled or deionised water: calcium in hard water is sequestered by citrate, which is useful, but it also arrives in a bath whose whole job is to carry iron away, and Ware's clearing chemistry shows what calcium does to an iron oxalate system. Wall states no mixing temperature for this bath.

Purpose

To dissolve the iron(II) that light has made, so it can reach the palladium salt and reduce it, and to do that without reducing anything itself. A siderotype developer is not a developer in the sense the rest of this formulary uses the word. There is no developing agent in the bottle, no restrainer, no alkali and no preservative. The reducing agent was made by light, in the paper, during the exposure. It is iron(II) oxalate, it sits exactly where the photons put it, and it is almost insoluble — 0.022 g per 100 cc, on Ware’s figure. Until something dissolves it, it cannot react with anything.

That is the first job, and it is over in seconds. The second job takes minutes: the same ligand has to pick up the iron(III) that was never exposed, and the iron(III) the reaction has just produced, and carry both of them out of the sheet before the print dries. Iron left in a finished print is the whole permanence problem of the process, and the developer is the first and largest of the three baths that address it.

The third purpose is the one that makes this a separate formula rather than a footnote to the potassium oxalate developer: to do all of that without putting down any palladium of its own. Oxalate is itself a reducing agent for palladium(II). Citrate, on the evidence of the same author’s redox table, is very much less so. What follows from that single difference is a print with clean high values, and it is the reason this bath exists.

Recommended uses

Developing a palladium print — the palladiotype — where the high values matter. This is the use the formula was published for and the one every piece of evidence read for this page supports. Ware’s own comparative tests put the two developers side by side on palladium and photograph the result: the oxalate bath gives “slightly higher densities and warmer image colour” but “does cause a perceptible chemical fog in the high values, which is worsened by humid conditions”, while “the citrate development bath recommended by Willis leaves clear highlights”. If you are printing in palladium and your highlights are veiled, this bath is the first thing to change.

Printing on a paper whose surface you want to keep. Willis’s own published reason for the change was not about the image at all. His distributor’s notice, quoted by Ware from Abel’s Weekly, says: “The Palladiotype developer and clearing baths have no tendency to injure the beautiful surface of the paper employed for coating. An undue immersion of matt Platinotypes in the oxalate and acid baths does tend to destroy their natural bloom.” That is a statement about sizing and surface finish, made by a manufacturer who had watched his own product being spoiled, and it applies with more force to a modern hand-coated sheet than it did to his machine-made stock.

Working with a paper that throws black specks. Ware’s account of the “plague of black spots” that afflicts development-process platinum and palladium prints notes that they “appear to increase in number and density with the vigour of the developer (i.e. oxalate produces a worse effect than citrate)”, and separately, writing about one particular modern paper, that “under oxalate development, but less markedly so under citrate, the sensitized area randomly generates small black specks”. If a paper is speckling, the citrate bath is a cheaper experiment than a new paper.

Reproducing a commercial Palladiotype from the 1917–1930 period. Every palladium print made on Willis’s own paper according to Willis’s own instructions passed through this bath at this strength. If the object of the exercise is to understand what those prints are, the developer is not a detail.

Reading the period literature. Wall’s Photographic Facts and Formulas gives palladiotype a single paragraph and assumes the reader knows what the numbers in it mean. Knowing that 232.5 g and 23.25 g to the litre is a ten-to-one citrate buffer, and that the 7 to 16 °C is a deliberate instruction rather than a room temperature, makes that paragraph legible.

Not, on the evidence read here, developing platinum. See the next section, and the note under The mechanism, which is the most interesting unresolved thing on this page.

When another formula is preferable

• When the metal is platinum. Ware’s redox table gives the citrato-iron couple at +0.372 V and states in terms that it “does not reduce platinum(II) or palladium(II)… under the printing conditions”. Palladium is developed in citrate anyway, by Willis’s instruction and in Ware’s own tests, which is a tension this page does not pretend to resolve. Platinum is the harder case on both counts — a higher potential at +0.73 V and, on Ware’s account of the third-row metal, much slower kinetics. No source read for this page reports a platinum print developed in a cold citrate bath. Willis’s own 1880 patent offers citrate of soda as a platinum developer and says it should be applied “cold, warm, or hot, but preferably hot”, which is exactly the concession the kinetics would demand. For platinum, use the potassium oxalate developer.
• When you want the warmest palladium the process will give. Oxalate gives a warmer colour and higher speed, and Ware says so plainly. The cost is the fog in the high values, and the American tradition’s answer was to accept the fog and etch it back with 1:200 hydrochloric acid — two errors which, as Ware notes, are “somewhat self-compensating”. That is a defensible working practice with a century of prints behind it, and it is not the same print as a citrate-developed one.
• When you have bought a kit. Neither surviving supplier ships citrate. Photographers’ Formulary packs 227 g of potassium oxalate in the palladium kit; Bostick and Sullivan supply a ready-made oxalate bath in a 32-ounce bottle. Their instructions, times and clearing sequences are written for that bath, and are covered on the Photographers’ Formulary and Bostick and Sullivan pages.
• When the print is a kallitype and the image is silver. Use the sodium citrate kallitype developer, which is the same salt at almost the same strength without the acid, for a reason set out on that page: an acid bath is the wrong place for colloidal silver. The two formulas are near neighbours and are not interchangeable.
• When you are printing out rather than developing. Ware’s print-out platino-palladiotype forms its image during the exposure and needs clearing rather than developing; its first bath is disodium EDTA.
• When you need contrast from the chemistry rather than from the negative. This bath offers none that the course will publish. The period answer is under Variants and it is a chromium(VI) salt.

Mixing

Weigh, dissolve, acidify, and do not warm it unless you mean to. Wall gives no mixing instruction at all beyond the three lines of the formula, and no mixing temperature, which is itself informative: both solids dissolve readily in cold water and there is nothing in the bottle that needs heat to get in.

The order in mixingOrder is the course’s and not Wall’s, and it is worth a sentence because no source states one. The citrate goes in first because it is 91 per cent of the solid load and because a saturated-ish citrate solution is what the acid is being added to; putting the acid in first would give a moment of pH 2 solution that does nothing useful. Neither order is dangerous. Neither order changes the finished bath, because a buffer does not remember how it was assembled.

🧮What Wall's litre actually holds, and why Ware's figure is smaller

Wall prints “Water 1000 ccm”. He does not print “water to make 1000 ccm”, and the difference between those two phrasings is a convention of the period’s published formulas rather than a reading of punctuation: a make-up volume fixes the strength of the finished solution, an added volume does not. This course records what each source states, and Wall’s line is water added. So the bath that comes out of the beaker is a litre of water plus 232.5 g of citrate and 23.25 g of citric acid — 255.75 g of dissolved solid — and it occupies more than a litre.

The consequence is arithmetical and it runs one way only. The nominal figure, grams per litre of water, is 23.25 per cent. The true figure, grams per 100 cm³ of finished solution, is lower, because the denominator grew. How much lower depends on how much volume the dissolved salts take up, and no source read gives that number for this pair of solids, so no number is offered here.

What can be said is that the direction closes the gap to Ware, who converts the manufacturer’s own instructions to 20 per cent w/v of the made-up solution — his definition, stated on the page: “the number of grams of solute that are present in 100 cubic centimeters of the made-up solution”. Two printings that differ by a seventh in nominal strength, in the same direction the two conventions predict, holding the same ten-to-one ratio of salt to acid, are better evidence for each other than either would be alone.

What the course does not do is average them. If you mix Wall’s figures you have made Wall’s bath. If you want Ware’s, weigh 20 g of the dihydrate and 2.2 g of citric acid and make the total up to 100 cm³ with water, which is what “20% w/v” means and the only unambiguous way to hit it.

⚗Where the pH lands, as an estimate and not as a measurement

No source read publishes a measured pH for this bath. What follows is an estimate from published dissociation constants, shown because the reasoning is more useful than the number.

Trisodium citrate carries no acidic protons on the citrate: HSDB records a solution of the dihydrate as “slightly alkaline to litmus, pH about 8”. Citric acid carries three, at IUPAC’s pK_a1 2.87, pK_a2 4.35 and pK_a3 5.68 (20 °C, 0.1 mol/L sodium perchlorate, all assessed approximate). Mixing the two makes a buffer whose pH depends only on how many protons there are per citrate.

In this bath: 232.5 g of the dihydrate is 0.79 mol of citrate; 23.25 g of anhydrous citric acid is 0.121 mol, carrying 0.363 mol of acidic protons. Total citrate 0.911 mol, total protons 0.363 mol —

0.363 ÷ 0.911 = 0.40 protons per citrate

Average protonation of the citrate in this bath

which puts the equilibrium between the fully deprotonated citrate and the singly protonated hydrogencitrate, the step governed by pK_a3. Henderson–Hasselbalch then gives pH = 5.68 + log(0.60/0.40) ≈ 5.9.

Three reasons to distrust the third figure and not the first two. The pK values were measured in 0.1 mol/L perchlorate and this bath is above 2 mol/L in sodium; activity coefficients at that ionic strength are nowhere near unity. The calculation ignores the iron, palladium and chloride the print brings in. And nobody has measured it. What survives the caveats is the useful part: this is an acidified bath, but only just — somewhere in the fives, not the twos. That is the window the chemistry needs, and the next section says why.

Behaviour

The image appears almost at once and then you wait. The Getty’s atlas describes the developed platinum-type process in one sentence — “the platinotype image appears almost immediately” — and Bostick and Sullivan’s modern instructions for the oxalate bath say that “the print will develop immediately. Development is complete within a few seconds.” Ware’s transcription of Willis and Clements’ palladiotype instruction is “treat for as long as necessary to develop the image: 4-5 minutes”, and the arithmetic of those two statements is the same as it is for every siderotype developer: the picture is finished long before the tray time is. What the remaining minutes buy is iron removal, and it is bought in the developer because the developer is the largest and least loaded bath the print will meet.

It is a cold bath, and that is deliberate. Wall’s instruction for black palladiotype paper is to hold the developer between 7 and 16 °C. Nothing else in this formulary is worked that cold. Set beside the platinotype practice of the same era — a bath at 60 to 77 °C before 1892 — it is a striking inversion, and it is the clearest single indication that in this process image colour is a developer variable rather than a paper variable. Wall gives the other end of the same lever in the next paragraph: the sepia paper’s developer, four or five times more dilute, is heated to 38 °C, “not beyond”, for warmer tones.

Nothing about it is vigorous. Bostick and Sullivan say of the oxalate bath that it “can be used from room temperature up to the boiling point” and that the print colour changes across that range; the citrate bath’s published range is 7 to 38 °C, and the top of it belongs to a different, weaker solution. This is a mild reagent used gently, and most of its published faults are faults of under-immersion rather than over.

📷The photography

What you actually see, and what it tells you. The sheet goes in yellow-brown — that is unreduced ferric oxalate plus whatever faint print-out image the exposure produced — and the picture arrives in the first few seconds as the dissolved iron(II) reaches the palladium salt. For the remaining minutes what changes is not the image but the background: the yellow drains out of the unexposed areas as citrate lifts the untouched iron(III) into solution.

That drain is the thing to watch. A print pulled at one minute looks finished and is not; the iron it still holds will hydrolyse in the wash, chemisorb to the cellulose, and turn up months later as a yellow highlight stain that no clearing bath will now shift. Ware demonstrates this deliberately — one to two minutes in each bath, to reproduce the staining attributed to Stieglitz’s impatience — which means the shortened process is a documented way of ruining a print rather than a shortcut.

The developer darkens with use, and it stays photosensitive. Everything the print brings out with it — iron, chloride, unreduced palladium salt — accumulates. Ware’s warning about the oxalate bath is about what the print brings in rather than about which salt the bath is, so it transfers: “any re-used platinotype developer will always contain excess sensitizer reagents in solution, and will therefore still be photosensitive; if it is not totally protected from actinic light, it will steadily precipitate particulate platinum metal, which if left in suspension could find its way onto the surface of subsequent prints.” He wrote it of platinum; nothing in the reasoning is specific to the metal. No source read publishes a capacity or a reuse regime for the citrate bath specifically, and the course does not borrow the oxalate bath’s.

Image characteristics

Colour. Cooler than the same sheet in oxalate. Ware’s tests are unambiguous about the direction — “oxalate also produces a warmer colour and higher speed” — and give the mechanism in passing: faster reactions produce greater supersaturation and hence smaller particles in the precipitate, “with a consequent warmer colour”. A gentler developer makes a coarser, cooler deposit. Within the citrate bath itself the same lever is available in the two directions Wall prints: cold and concentrated for black, dilute and warmed to 38 °C for sepia. The AIC’s conservation account adds the general rule that “the higher the pH, the lighter the print and conversely, the lower the pH, the darker the print”, and that more acidic processing gives browner tones.

Highlights. This is the characteristic that justifies the formula. Ware’s figure 6.7 sets four palladium prints side by side — developed in oxalate and cleared three ways, and developed in citrate and cleared in citrate — and his text on the last of them is that the citrate bath “leaves clear highlights”. The fog it avoids is not iron stain: it is “a brownish grey fogging of the highlights by palladium metal, quite distinct from the yellow stain of iron”, and it is worse in humid conditions.

Maximum density and speed. Slightly lower than oxalate, on the same evidence, and unquantified in every source read. If you change developers you will need to re-test your exposure; nobody publishes by how much.

Contrast. Not a property of this bath. The AIC’s overview of contemporary practice says contrast “is achieved mostly by exposure, not the developer”, and the exposure scale of an unmodified ferric-oxalate sensitiser is long — Ware measures it at no more than about 2.4 and typically nearer 2.0. A negative with a density range under about 2 will print flat in any of these developers.

Tonal reversal, which belongs to the sensitiser and not to this bath. Ware lists the factors that promote the partial reversal seen in some Stieglitz palladiotypes, and the developer appears in the list only as a modifier: “oxalate developer is slightly more effective than citrate, but tends to cause fogging in palladium”. The dominant factors are low humidity at exposure, the paper, and a contrasty negative.

Staining. None of its own. The stain risk in this process is residual iron left by insufficient development and insufficient clearing, and that is a process fault rather than a property of the formula.

The mechanism

One: light makes an insoluble salt, and leaves it where it is

The sensitiser is ferric oxalate plus a palladium(II) salt, dried into the paper. Ultraviolet light drives the reaction Döbereiner first saw in 1831:

Fe₂(C₂O₄)₃ + light → 2 FeC₂O₄ + 2 CO₂

Photolysis of ferric oxalate, after Ware — carbon dioxide leaves, iron(II) oxalate stays

The product is iron(II) oxalate, which Ware gives as soluble to only 0.022 g per 100 cc. It is a solid, sitting in the fibres, in a pattern that is the latent image. In Ware’s words it “cannot reduce platinum(II) or palladium(II) salts in aqueous solution to the metal unless it is solubilised by complexation”.

Nothing has been reduced except the iron. There is no palladium image yet.

Two: the bath dissolves it

The developer’s whole function is that one word, solubilised. In the oxalate bath the mechanism is Ware’s own, and it is simply mass action on a sparingly soluble salt:

FeC₂O₄ + C₂O₄²⁻ → [Fe(C₂O₄)₂]²⁻

How the oxalate developer mobilises the photolytic iron(II), after Ware

Citrate is a stronger and more versatile chelating agent than oxalate — three carboxylates and a hydroxyl on one small molecule, able to close a ring round an iron ion from several directions — so at 0.79 mol per litre it does the same job by a different route: ligand exchange rather than common-ion dissolution.

FeC₂O₄ + C₆H₅O₇³⁻ → [Fe(C₆H₅O₇)]⁻ + C₂O₄²⁻

Schematic: citrate displacing oxalate from the photolytic iron(II). The course's construction on the pattern of Ware's oxalate equation; no source read prints it

⁇The one thing on this page the sources do not settle

Read Ware’s chemistry chapters straight through and you arrive at a contradiction with his own history chapters.

What §11.3 says. The redox potential of the citrato-iron couple is E(Fe^IIICit/Fe^IICit) = +0.372 V, against +0.02 V for the oxalato couple. Ware’s conclusion is explicit: this “shows it to be a weaker reducing agent than the oxalato-complex… and it does not reduce platinum(II) or palladium(II), although it will reduce gold(III) and silver(I), under the printing conditions.” Elsewhere he gives the same reason for why Herschel’s citrate-sensitised siderotypes made gold, silver and mercury images but never a platinum one: “ferrous citrate is not a sufficiently powerful reducing agent to enable a similar platinum printing process.”

What §2.9, §6.16 and §6.17 say. Willis specified an acidified trisodium citrate developer for Palladiotype; Willis and Clements published it; Wall printed it; and Ware himself developed palladium prints in it and photographed the results, reporting only that they have cleaner highlights and slightly lower density and speed than the oxalate ones.

Both statements are in the same document, and this page reports both. What the course can add is that the two are not necessarily about the same system, and the distinction is worth understanding because it is the distinction between a sensitiser and a developer.

Ware’s §11.3 statement is about iron(II) citrate formed in situ from an iron(III) citrate sensitiser — the Herschel case, where the reduction has to happen in a nearly dry paper layer, during or just after the exposure, at whatever local concentration the coating provides. This bath is a different situation on every axis: 0.9 molar in citrate, wet, at four to five minutes, and against palladium(II) rather than platinum(II), which Ware repeatedly describes as the more reactive and kinetically faster of the two metals.

Thermodynamics does not forbid it either way: palladium’s E(PdCl₄²⁻/Pd) is +0.62 V, which is above +0.372 V, so the reaction is downhill even for the citrato couple — by about a quarter of a volt, where oxalate has six-tenths. Read that way, “does not reduce” is a statement about rate under one set of conditions rather than about thermodynamic possibility, and the citrate developer works because it is given far more time and far more ligand than a print-out sensitiser ever gets.

That reading is the course’s and not Ware’s. It is offered because a reader who meets the contradiction deserves a hypothesis rather than a silence, and it is labelled because nothing read for this page states it. What is not in doubt is the observation: palladium prints developed in citrate exist, they have been made recently under controlled conditions, and they are photographed in the literature.

Three: the mobile iron(II) reduces the palladium

Once the iron is in solution it can travel the fraction of a micron to the nearest palladium(II) ion and hand over an electron. Ware writes the platinum case; the palladium case is the same equation with the other metal:

2 [Fe(C₂O₄)₂]²⁻ + [PdCl₄]²⁻ → 2 [Fe(C₂O₄)₂]⁻ + Pd + 4 Cl⁻

Reduction of tetrachloropalladate(II) by the mobilised iron(II), after Ware's platinum equation

The palladium comes out as metal, as nanoparticles lodged in the paper fibres, and that is the image. There is no binder, no silver, no grain in the photographic sense — the tonal scale is made of differing amounts of metal in the same fibre matrix, which is why these prints look the way they do.

Deeper: why the developer must be acidified, and only slightly

Two constraints pull in opposite directions and the citric acid is the compromise between them.

Pulling down. Ware’s clearing chemistry establishes that above pH 4 iron(III) hydrolyses:

Fe³⁺ + 3 H₂O → Fe(OH)₃ + 3 H⁺

Hydrolysis of iron(III), which begins above pH 4

and the product does not stay as an obliging hydroxide. Left in the sheet 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”. That is the yellow highlight stain, and once it has formed the print cannot be rescued. A plain trisodium citrate bath at pH about 8 is on the wrong side of that line.

Pulling up. The complementary rule comes from the clearing baths rather than the developer: the hydrochloric acid used at 1:60 for platinum “dissolves palladium and bleaches the image partially”, which is why Paul Anderson insisted on 1:200 for palladiotypes. A strongly acid bath in contact with a freshly formed palladium image costs density.

The compromise. A tenth as much citric acid as citrate lands the bath in the region of citric acid’s third dissociation — the estimate under Mixing puts it near pH 6 — which is far below the pH 8 of the unacidified salt and far above the pH under 1 of the clearing acids. And the constraint about hydrolysis is softened further by the ligand itself: iron(III) at pH 6 hydrolyses when it is free, and in this bath it is not free, because 0.9 molar citrate is holding it. The acid is insurance, not the mechanism.

Deeper: why oxalate fogs a palladium print and citrate does not

This is the difference the formula exists for, and it follows from the same potential.

At +0.02 V the oxalato-iron couple is a strong enough reductant that the oxalate system as a whole can put down palladium where no light fell. Ware’s tests state the observation directly: “the oxalate developer can reduce the Pd(II) to some extent, thereby causing a brownish grey fogging of the highlights by palladium metal, quite distinct from the yellow stain of iron.” At +0.372 V the citrato system has about a quarter of a volt of driving force against palladium(II) instead of six-tenths, and the fog does not appear.

The general principle is worth extracting, because it recurs everywhere in this course. A developer that is more energetic than the job requires does not develop better; it develops indiscriminately. The same trade sits behind every restrainer in a silver developer and behind every chemical fog problem in this formulary. Here the trade is made not by adding a restrainer but by choosing a weaker ligand, which is a more elegant solution and a rarer one.

The corollary is that the American practice — develop in oxalate, then clear the fog off with hydrochloric acid — is not equivalent. Ware calls it “a case of two errors in processing being somewhat self-compensating”, and notes that the acid takes image density with the fog. The prints are different objects.

Deeper: what the citrate does after the image is finished

Both developers dissolve iron(II). Only some of them hold iron(III) well, and that matters for the minutes after the picture has arrived.

Citrate holds both oxidation states. So does oxalate — Ware’s overall formation constant for trisoxalatoferrate(III) is about 10²⁰, against about 10⁵ for the iron(II) complex, and it is precisely that difference in binding that drags the redox potential down from +0.771 V for the aquated couple to +0.02 V for the oxalato one. What citrate offers is the same service at a milder potential, plus the property Willis actually cited: it does not attack the paper’s surface the way the oxalate-and-mineral-acid sequence does.

The practical consequence is that the developing tray is where most of the unexposed iron leaves. The clearing sequence that follows removes what is left, and the less there is left, the longer those baths last and the less likely a stain becomes.

Function of every ingredient

Trisodium citrate dihydrate — 232.5 g

What it is. Na₃C₆H₅O₇·2H₂O, formula weight 294.10, the sodium salt of citric acid with all three carboxyl groups deprotonated. A food additive (E331), sold as colourless crystals, granules or powder, soluble to 77 g per 100 mL. Wall writes only “Sodium citrate”; Ware, converting the manufacturer’s instructions, names the dihydrate explicitly and gives its formula weight, which is why this page does too. The anhydrous salt is a different article at 258.07 and carries 14 per cent more citrate per gram; weighing 232.5 g of it would make a bath 14 per cent stronger than the formula.

Why it is there. It is the entire developing action. Nothing else in the bottle touches the iron.

What it does chemically. Three things, all of them complexation. First, it dissolves the almost-insoluble iron(II) oxalate that light made, by offering the iron a better ligand than the oxalate it is sitting in. Second, it holds the iron(III) produced when that iron(II) reduces the palladium, keeping it in solution instead of letting it hydrolyse. Third, over the remaining minutes it lifts the iron(III) that was never exposed out of the paper altogether. The first job makes the picture; the second and third decide whether the print survives.

What follows on the print. A palladium image with clean high values, at a slightly lower density and speed than the same sheet would give in oxalate, and in a slightly cooler colour.

More of it. Wall’s black-paper bath is already the strong end of the published range: 23.25 per cent nominal against Ware’s 20. Ware records for the platinum process that the trade sometimes ran developers “nearly saturated”, and Wall’s own note about the platinotype sepia developers is that “the stronger the developer, the more rapid its action and the softer the print”. Nothing read for this page tests that on the citrate bath. What can be said is that more citrate means more free ligand, faster dissolution and a shorter time to a visible image — and that the iron removal, which is the slow half, is limited by diffusion out of the paper rather than by the concentration in the tray.

Less of it. Wall prints the four-or-five-times-weaker bath himself, for the sepia paper, and Ware records the manufacturer recommending “ca. 4x” dilution for the same stock. So a weaker bath is not a mistake; it is the sepia formula. Weaker still and the dissolution slows, less iron leaves in the developing tray, and the clearing baths inherit the difference.

What it interacts with. Calcium, which it sequesters — useful, but it means hard water arrives in the tray carrying a competitor for the ligand. The paper’s alkaline buffer, if it has one, for the same reason and worse. And citric acid, with which it forms the buffer that is the second half of this formula.

Citric acid — 23.25 g

What it is. C₆H₈O₇, formula weight 192.12 anhydrous, a triprotic organic acid with pK_a values of 2.87, 4.35 and 5.68 — three protons close enough together that the solution buffers broadly rather than sharply. The monohydrate at 210.14 is also sold and would need 1.094 times the weight for the same acidity. Wall gives it as exactly a tenth of the citrate by weight, and so, within his own rounding, does Ware.

Why it is there. To move the pH, and for nothing else. It is not a reagent in this bath: it is the same chemical species as the citrate, one to three protons up.

What it does chemically. A solution of trisodium citrate alone sits at about pH 8. Adding citric acid protonates part of the citrate and drops the bath to the region of pK_a3 — the estimate under Mixing puts it near pH 6, with the caveats stated there. That matters because iron(III) hydrolyses above pH 4 and its hydrolysis product converts irreversibly to goethite, which no clearing bath removes. The acid buys margin against a stain that cannot be undone.

What follows on the print. Cleaner, whiter high values in the long term rather than on the day the print is made. The AIC’s account adds a same-day consequence too: lower pH gives a darker print and browner tones.

More of it. The bath moves toward pK_a2 and then pK_a1. There is no published figure for where it starts to cost image density, but the direction of travel is toward the clearing acids, and those demonstrably dissolve palladium: Anderson’s 1:200 hydrochloric acid rule exists because 1:60 “dissolves palladium and bleaches the image partially”. Willis’s own clearing stock uses the same two chemicals at 9 per cent acid — four times the developer’s proportion — and it is a clearing bath rather than a developer, which is the useful calibration.

Less of it, or none. Take the acid out and you have the sodium citrate kallitype developer, at about pH 8, which is the correct bath for a silver image and the wrong one for iron removal. For a palladium print, leaving the acid out means iron(III) hydrolysing as it comes out of the fibres, and the yellow highlight stain that follows.

What it interacts with. The citrate, as its conjugate; metals, which Chemical Safety Card 0855 records it as attacking, so no metal trays, funnels or tongs; and the palladium image itself, at strengths far above this one.

Water — 1000 ccm

Distilled or deionised, and this is not a counsel of perfection. Calcium is the specific problem: Ware gives the equation by which calcium ions precipitate insoluble calcium oxalate from an iron oxalate system and “promote aquation and hydrolysis of the Fe(III) complex”, which is the first step on the road to goethite. Citrate will sequester the calcium, but it is doing that instead of holding iron. Wall states no water temperature and no grade; both solids dissolve cold.

Interactions

With the sensitiser, which is where the oxalate comes from. This developer is not oxalate-free in use. Every sheet brings ferric oxalate into the tray, and the customary 2 per cent of free oxalic acid in the sensitiser comes with it. The bath therefore becomes a mixed citrate–oxalate system within a few prints, and no source read for this page reports what that does. It is one of the open questions under Experiments.

With the palladium salt, which is the point. The three-solution drop system sets the metal and its proportion; this bath develops whatever that system laid down. Ware’s kinetic contrast between the two metals — second-row palladium fast, third-row platinum slow — is why a developer that works for one may not work for the other, and it is why this page’s recommendations stop at palladium.

With the clearing baths, which Willis designed as a set. Willis’s own clearing agent for Palladiotype is this same pair of chemicals at a different ratio: a stock of 20 per cent citrate with 9 per cent citric acid, diluted eightfold to about 2.5 per cent citrate and 1.1 per cent acid, three baths of 10, 15 and 20 minutes, then a thirty-minute wash. Wall prints the same stock at 232.5 g citrate and 93 g citric acid to the litre, diluted 1 part in 7, with the same three times. Developer and clearing bath are the same chemistry at two acidities, which is why the transition between them costs the print nothing. The course’s modern clearing sequence is a different and better-tested route, and the comparison belongs on that page.

With hydrochloric acid, which is the alternative clearing tradition and does not belong with this developer. Ware is explicit that Willis “did not advise clearing Palladiotypes in hydrochloric acid like Platinotypes, but recommended the milder acidified citrate solutions. If hydrochloric acid is used to clear Palladiotypes one can expect some loss of image densities compared with citrate clearing.” Using the acid to compensate for oxalate fog is the American practice; using it after a citrate developer removes density and compensates for nothing.

With an oxidising agent, which is what a contrast control is. Ware’s account of the mechanism is the same one that appears throughout the iron processes: 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”. It costs highlight gradation and, in his words, “can cause a deterioration in image quality and an increase in image graininess”. The course publishes no dose; see Variants.

With air, slowly. The same reoxidation runs without any added oxidant:

4 Fe²⁺ + O₂ + 4 H⁺ → 4 Fe³⁺ + 2 H₂O

Aerial reoxidation of iron(II) — slow in acid, faster as pH rises

In the tray this is a minor loss of density. In the bottle, over months, it is one reason a working bath that has developed prints is not the solution it was.

With light, which is why an old bath is not inert. See Storage: a used bath holds dissolved sensitiser and will precipitate palladium under actinic light.

Variants

The 1880 patent’s version, which is the ancestor and is not this bath. William Willis’s British patent No 1117 of 15 March 1880 offers an “improved developing solution” as an alternative to potassium oxalate: “I use for this solution the tartrate or citrate of soda, of potash, or of ammonia… I prefer however to use the citrate of soda. I make a solution containing one hundred and twenty (120) grains of citrate of soda in one ounce of water.” At Ware’s conversion factor of 0.228 per cent w/v per grain per fluid ounce, 120 grains to the ounce is a nominal 27 per cent — the same order as the bath above. Two differences matter. There is no acid in it; and it is a platinum developer, for paper coated with a platinous salt and ferric oxalate, to be applied “cold, warm, or hot, but preferably hot”. The official Abridgments of Specifications records the same claim independently, in the Patent Office’s own words: a developer “may also consist of tartrate, citrate, or acetate of sodium potassium or ammonium… Sodium citrate is preferably employed.”

The 1917 demonstration version, at a quarter of the strength and possibly the other cation. Ware records that in March 1917 the Platinotype Company’s technical manager, W. H. Smith, demonstrated the new Palladiotype to the Camera Club of London and “showed that both developing and clearing of Palladiotype were accomplished simply with baths of 5% potassium citrate solution acidified with 0.26% citric acid.”

⁇Sodium or potassium, at five per cent?

Wall’s sepia palladiotype developer is sodium citrate 50 g and citric acid 2.62 g to 1000 ccm — 5 per cent and 0.262 per cent. Ware’s account of the 1917 demonstration is potassium citrate at 5 per cent with 0.26 per cent citric acid. The strengths agree to three figures. The cation does not.

Three readings are available and this page picks none of them. The 1917 report may have said potassium where the company meant sodium. The company may genuinely have demonstrated the potassium salt and published the sodium one, which would be unremarkable — the 1880 patent lists the sodium, potassium and ammonium salts of citric and tartaric acid as alternatives in the same sentence. Or the two are simply different formulations that happen to share a strength.

What the coincidence does establish is that the 5 per cent, 0.26 per cent bath is a real published formulation and not a transcription slip in either source, since two independent documents arrive at it. The course records it as the weak end of this developer’s published range, and notes that potassium citrate and ammonium citrate have encyclopaedia entries of their own for readers who want to follow the substitution.

The sepia version, which is this bath diluted and warmed. Wall’s second palladiotype developer, for sepia paper: sodium citrate 50 g, citric acid 2.62 g, water 1000 ccm, heated to 38 °C, “not beyond”. Ware records the manufacturer’s own instruction as “more dilute solutions (ca. 4x)… for the Sepia Vellum Palladiotype paper”. It is recorded in dilutions above rather than as a separate entry, because both sources present it as the same formula at another strength.

The ammonium version, which is where this chemistry lives today. Modern platinum and palladium practice knows a citrate developer chiefly as ammonium citrate, and that entry is where the modern evidence belongs. One incompatibility is worth carrying across from Ware now: hexachloroplatinate(IV), used as a contrast agent, “cannot be employed with any chemistry involving ammonium cations, in sensitizer or developer, because ammonium hexachloroplatinate(IV)… has a very low solubility and will crystallise out.” That restriction does not apply to the sodium bath on this page.

☠The contrast addition, and why no dose appears here

Wall’s instruction is “for brilliant prints add from 4 to 6 drops of a 10 per cent solution of potassium bichromate to the developer”. Ware records that Willis and Clements recommended a very small amount of potassium dichromate for the same purpose — in the oxalate bath in his account, which is a discrepancy with Wall worth noticing — and that Paul Strand and Ned Scott used it on flat negatives.

The course publishes no dose. Potassium dichromate is notified as Danger with H350, may cause cancer, and H340, may cause genetic defects, along with sensitisation statements and an aquatic hazard; the chromium policy governs, and the figures above are printed as what the period sources said rather than as an instruction to follow.

The mechanism does not require chromium, and neither does the print. Any oxidant that reaches the iron(II) first will truncate the scale, which is why potassium chlorate and hexachloroplatinate(IV) appear in the same role. Ware, who has tested more of these than anyone writing today, declines the whole category on grounds that have nothing to do with safety: “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… the use of these undesirable image-degrading agents will be avoided.”

The mercury variants belong to platinum and are covered elsewhere. Warm-toned platinum was made by adding mercuric chloride to the sensitiser or the developer; the course’s mercuric sepia platinotype entry treats that at Level D. The modern substitute for a warm-toned noble-metal print is palladium, which is the metal this developer was published for.

No safer course variant of the quantities is offered, because there is nothing in the formula to make safer. It is two food-grade solids in water. Its hazards belong to the sensitiser, the print and the waste, and they are unchanged by anything that could be done to this bath.

Safety

Level A, and unusually for this formulary that is not a close call for the bath itself — though it is emphatically a statement about this solution and not about palladiotype printing, which involves noble-metal salts, a ferric-oxalate sensitiser and an ultraviolet source, each with its own classification.

The two solids. Trisodium citrate is not classified: the ECHA notifications aggregated on PubChem for the dihydrate return 520 of 536 company reports finding no GHS criterion met, and Chemical Safety Card 1219 leaves its classification box empty while recording that the substance is irritating to eyes and respiratory tract. A second ECHA entry filed against the anhydrous record as “Citric acid, sodium salt” does carry Warning, H319, causes serious eye irritation, in 72 per cent of its reports; the course reports that rather than adopting it, and it is why eye protection appears below for a substance whose own entry is empty. Citric acid is classified Warning with H319, H335 may cause respiratory irritation, and, on the Safe Work Australia entry, H315 causes skin irritation.

PPE for mixing and use. Nitrile gloves; safety glasses at minimum and splash goggles when handling the concentrated solids or the mixed bath; dust control when weighing, which Card 1219 words unusually well — “use ventilation (not if powder)”, meaning that a draught over a powder is the wrong answer and containment is the right one. No fume extraction is needed: nothing here is volatile and nothing evolves a gas.

What is genuinely hazardous in the session, and it is not this tray. The sensitiser carries palladium and platinum salts, which are occupational sensitisers; the exposure unit is an ultraviolet source; and if you are working in the historical manner the clearing baths may be dilute hydrochloric acid. The safety library and the process page carry those. The reason this page says so is that a Level A developer sitting in the middle of a higher-level process is a classic way to relax at the wrong moment.

The one comparison worth drawing. The developer this one replaces is not Level A. Potassium oxalate is Level B, and Photographers’ Formulary prints an explicit warning with its palladium kit that the salt is a poison and an anticoagulant and that tongs or gloves must be used. Swapping oxalate for citrate removes the only systemically toxic reagent from the developing tray, and that is a real and rarely mentioned advantage of Willis’s formula.

EH40 sets no workplace exposure limit for citrates or for citric acid. 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.

Storage

Mixed bath, between sessions: in a full, closed, labelled bottle, in the dark. The darkness is not a generic precaution. Ware’s warning about the platinotype developer applies to any bath that has developed prints, because it depends on what the print brings with it: a used developer holds dissolved sensitiser and “will therefore still be photosensitive; if it is not totally protected from actinic light, it will steadily precipitate particulate platinum metal, which if left in suspension could find its way onto the surface of subsequent prints.” He is writing about platinum; the argument turns on what the print dissolves into the bath and applies to a palladium one unchanged. A brown bottle in a cupboard, not a tray on the bench under a window.

Unused solution. No source read publishes a keeping time. It is two stable organic solids in water with no reducing agent to oxidise, so the limits in practice are biological growth and contamination rather than chemistry — but that is reasoning, not evidence, and no figure is offered.

Solids. Trisodium citrate dihydrate is stable in air, unlike several of its neighbours; Card 1219 gives it no storage requirement at all beyond keeping it dry and closed. It decomposes at 150 °C, which the card notes is an apparent melting point caused by loss of the water of crystallisation. Citric acid is stored dry and well closed, away from strong oxidants, strong bases, metal nitrates and metals; the monohydrate effloresces in dry air, so an old open jar no longer weighs what its label says. Neither container is ever one that has held food.

Label the bottle with the date, the strength and the word palladium. A used bath carries dissolved noble metal and is a waste stream, not a drain-load.

Incompatibilities

• Metals. Chemical Safety Card 0855 records citric acid as attacking metals. No metal tray, funnel, tongs or stirring rod. This is a plastic or glass bath.
• Hard water and calcium. Citrate sequesters calcium, which sounds convenient and is not: the ligand doing that is not holding iron, and Ware’s clearing chemistry shows calcium promoting “aquation and hydrolysis of the Fe(III) complex”. Distilled water.
• Alkaline-buffered paper. The chalk or calcium carbonate reserve in most modern fine papers is destructive to this chemistry generally — Ware calls it “destructive towards the iron chemistry, causing decomposition of the sensitizer” — and it arrives in the developer as well. Unbuffered cotton rag stock is a requirement of the process, not a preference of this bath.
• Strong oxidisers, the general incompatibility of any organic salt, and the specific one for any siderotype developer: an oxidant destroys the iron(II) the picture is made of.
• Ammonium chemistry plus hexachloroplatinate(IV), which is not an incompatibility of this bath but is the reason the sodium version and the ammonium version are not freely interchangeable. See Variants.
• Cross-contamination between processes. A tray that has held thiosulfate, sulfide or a silver solution has no business here. The general rule for alternative-process benches — one set of trays per process — is stricter than usual in platinum and palladium work because the quantities of metal are small and the prints are expensive.
• Nothing else of consequence. There is no acid or alkali here strong enough to burn, no reducing agent to oxidise, and no gas-evolving pair.

Waste

Bottle it. The citrate is a food additive; the metal is why the bath is collected.

A spent developer from this process carries dissolved iron in quantity and traces of palladium or platinum — unreduced metal salt washed out of the sheet, and, in a bath that has stood in the light, precipitated metal in suspension. Noble metal is the reason this is not a drain-load, and it is also the reason a printer with any volume of work should be collecting rather than discarding: the metal is the most expensive thing in the darkroom.

Segregate it. The alternative-process waste rule this course follows keeps three streams apart — silver-bearing, iron-bearing and noble-metal-bearing — because they have different routes and different recovery options. This bath is in the third.

pH is the wrong test and it is worth understanding why. Kodak’s J-52 guidance gives 5.6 to 9.4 as the window sewer codes most frequently set, and this bath sits inside it. That tells you nothing useful: what makes it a waste stream is the metal it carries, and no pH meter will report that. ILFORD’s advice for domestic users is to bottle wastes separately, label them, and take them to a household waste and recycling centre. Local regulation governs, and this course does not know your jurisdiction.

The rinse water counts too, at least the first one. The tray rinse immediately after development carries most of what the developer had in it.

Troubleshooting

Veiled, grey-brown highlights on a palladium print. If the bath is oxalate, this is Ware’s chemical fog — palladium metal reduced where no light fell — and it is “worsened by humid conditions”. Changing to this developer is the direct fix. If the bath is already citrate, the veil is more likely iron stain (see the next entry) or fog from an over-old sensitiser.

Yellow highlight stain, appearing days or months later. Residual iron, hydrolysed and then converted to goethite, which Ware describes as “quite insoluble in dilute acids” — by the time you see it, the clearing baths can no longer remove it. The causes are all about time and pH: too short a development, too short a clearing sequence, alkaline wash water, a buffered paper, or a print allowed to dry between the developer and the clearing baths. Ware reproduces the fault deliberately by cutting the developer and each clearing bath to one or two minutes. The cure is prevention: the full four to five minutes in the developer even though the image arrived in seconds, and the full clearing sequence wet.

Random black specks, appearing only after development. Ware’s “plague of black spots”. The cause is not settled — his own hypothesis is microcrystals of iron(II) oxalate precipitated in the coating by reducing impurities in the paper, invisible until the developer dissolves them and they meet the metal salt — but two observations are practical. The specks “increase in number and density with the vigour of the developer (i.e. oxalate produces a worse effect than citrate)”, so this bath is already the better choice; and Ian Leake finds that filtering the developer helps while filtering the sensitiser does not. Ware also notes the fault is paper-dependent and has never been seen with the print-out process.

Black specks that look like comet tails following the coating direction. A different fault: metal particles in the paper, which Ware says should be visible under a lens in the untreated sheet and can usually be lifted out with a scalpel point. Not a developer problem.

Image weaker than the same negative gave in oxalate. Expected, and not a fault. Ware: oxalate gives “slightly higher densities” and “higher speed”. Re-test your exposure rather than extending the development; a palladium print, like a platinum one, is not made denser by leaving it in longer once the iron has all dissolved.

Print bleaches or loses density during clearing. Hydrochloric acid at platinum strength. Anderson’s rule is 1:200 for palladium, not the 1:60 used for platinum, and Ware’s position is that Willis never advised hydrochloric acid for palladiotype at all: “if hydrochloric acid is used to clear Palladiotypes one can expect some loss of image densities compared with citrate clearing.”

Prints from a bath that has stood in the light acquire specks. Filter the bath, and thereafter keep it dark. See Storage.

Nothing develops at all. Check that the paper was sensitised with a ferric oxalate-based sensitiser and not a print-out one, and that the sensitiser has not decomposed — a chalk-buffered paper will do that on its own, before the exposure. This bath cannot develop what light did not make.

Experiments

1. The developer comparison, which is the experiment this page exists for. Coat and expose two identical sheets from the same sensitiser batch through the same step tablet. Develop one in this bath and one in potassium oxalate at its published strength, clear both identically, and read the two step tablets on a densitometer. You are looking for three things Ware reports and nobody has published numbers for: how much speed the citrate bath costs, how much maximum density, and — the important one — the minimum density in the unexposed steps, which is where the oxalate fog lives. Record the relative humidity, because Ware says the fog is worse when it is humid.

2. The acid series. Mix five litres of the citrate bath at 232.5 g per litre with citric acid at 0, 5, 10, 23.25 and 46 g per litre. Print the same negative in each, clear identically, and keep the prints for a year in the same conditions. The day-one result should be a small colour and density shift; the one-year result is the experiment. This is a slow test and it is the only honest way to see what the acid is buying, because what it buys is the absence of a stain that has not appeared yet.

3. The temperature pair Wall prints. Develop one print in the full-strength bath held between 7 and 16 °C and one in the four-times-diluted bath at 38 °C, and compare the image colour. Wall claims cold and strong gives black, dilute and warm gives sepia. Measure both with a reflection densitometer and, if you can, a colorimeter; the claim is over a century old and this course has read no modern test of it.

4. What happens as the bath ages. Every print puts oxalate, iron and chloride into a citrate developer. Keep one litre in continuous use, in the dark, and print a step tablet through it after 1, 5, 10 and 20 8×10 prints, clearing identically each time. Nobody publishes a capacity for this bath; this is how you would find yours. Filter a sample of the aged bath and look at what comes out on the paper.

5. The paper qualification test, which is not optional. Before any of the above, test your paper: unbuffered, alum-rosin or AKD sized rather than gelatin sized if you can get it, and free of iron contamination. Coat and develop an unexposed strip; if it speckles or greys, the paper is reducing your sensitiser and no developer will fix it.

📓Lab notebook

Record, for every sheet: the sensitiser batch and drop counts; the relative humidity at coating, at drying and at exposure; the exposure; the developer’s strength, temperature and age in prints; the clearing sequence and its times; and the wash. Then measure D_max and D_min and write them down. Ware’s own tests turn on humidity and on developer choice, and a print without those two numbers beside it cannot be compared with anything.