Jps pt/Pd paper

An Investigation of

Platinum and Palladium Printing

Mike Ware

[First published in The Journal of Photographic Science 34 (5/6), 165-177 (1986).]
ABSTRACT. An alternative method for printing in platinum and palladium has been developed which offers some advantages over the historical platinotype and palladiotype process with regard to the reproducibility, economy and quality of the image. The innovations include:

(i) the use of simple sensitizers containing only ammonium salts,

(ii) an economical paper coating technique,

(iii) controlled drying of sensitized papers to prescribed relative humidities, which allows a choice of image colour and contrast,

(iv) optimising the conditions for a 'printing-out' process,

(v) the use of a chelating agent to clear the paper.

The sensitometric characteristics of platinum and palladium printing papers are given in the form of D/logH curves; differences in the behaviour of the two metals are explained in terms of the state of aquation of the complex in the sensitizer. The effect of additives such as salts of mercury(II), lead(II) and gold(III) is discussed and a mechanism is proposed for their mode of action.

1 INTRODUCTION

Although the platinum printing process invented by Willis¹ in 1873 seems now quite obsolete from the commercial standpoint, it still retains an interest for the maker of fine prints, to whom the cost of materials and the time expended in their use is not the main concern. Accounts of the history and aesthetics of photography acknowledge the qualities of the platinotype, and similar palladiotype, which have always been much esteemed². Users of these processes can control the colour, texture and gradation of their images and be assured of their archival permanence.

Since the 1930's, when commercial production of platinum and palladium papers ceased, these considerations persuaded a few devotees to continue the craft by hand-coating their own sensitized papers. In the 1970's a 'renaissance' of historical and non-silver processes saw the publication³ of methods for platinum and palladium that, in their essentials, followed the recipes evolved by Willis and Pizzighelli and Hübl⁴ in the nineteenth century. It is evident from the work of contemporary exponents that these historical methods are capable of excellent results in skilled hands, but the older recipes call for materials that are now difficult to obtain, and they tend to be profligate in the consumption of precious metal.

The purpose of this paper is twofold: to bring to light some aspects of the chemistry of the process, which has not been recently reviewed, and thereby to offer a reproducible and more economic method of platinum and palladium printing, using only readily available materials. The parameters governing image quality are also summarised here and discussed, in the hope of stimulating more research into this off-shoot from the main stream of photographic development.

2 CHEMISTRY OF THE PROCESS

2.1 Redox Reactions

The platinotype and palladiotype are 'iron-based' processes⁵ in which the photosensitive material is an oxalato-complex of iron(III), traditionally ferric oxalate itself. In the present work it has been found advantageous to use instead salts of the trisoxalatoferrate(III) anion, for reasons described below in §3.2. This complex undergoes a photochemically-induced redox reaction, yielding carbon dioxide and an oxalato-complex of iron(II). The overall stoicheiometry can be represented by:

h+ 2[Fe(C₂O₄)₃]^3-  2[Fe(C₂O₄)₂]^2- + C₂O₄^2- + 2CO₂ reaction <1>

It is evident from the standard redox potentials⁶ that reaction <1> should proceed spontaneously:

E^o([Fe(C₂O₄)₃]^3-/[Fe(C₂O₄)₂]^2-) = +0.02 V

E^o(2CO₂/C₂O₄^2-) = -0.49 V

but at ambient temperature there is a kinetic barrier to this process which is only overcome when the complex is photoexcited by absorption of ultra-violet light in the vicinity of its ligand-to-metal charge transfer band at max = 260 nm. This reaction has been the subject of much photochemical investigation; it is believed to proceed via a radical-anion mechanism⁷.

The iron(II) oxalato-complex so formed is quite a powerful reducing agent, as indicated by its E^o value; it can readily reduce compounds of metals having more positive potentials, to yield the metal itself which constitutes the final image. Such readily reducible metals belong to the category once described as "noble" e.g.:

platinum E^o([PtCl₄]^2-/Pt,4Cl^-) = +0.73 V

palladium E^o([PdCl₄]^2-/Pd,4Cl^-) = +0.62 V

silver E^o(Ag⁺/Ag) = +0.80 V

gold E^o([AuCl₄]^-/Au,4Cl^-) = +1.00 V

in the case of platinum, for example, the reaction is:

[PtCl₄]^2- + 2[Fe(C₂O₄)₂]^2-Pt + 2[Fe(C₂O₄)₂]^- + 4Cl^- reaction <2>

where the iron(III) oxidation product [Fe(C₂O₄)₂]^- will subsequently coordinate more ligands, such as (C₂O₄)^2- or H₂O. Similar equations can be written for the other metals. However, kinetic factors may make the rates of such reactions too slow to be useful unless the noble metal complex is sufficiently labile, (as are those quoted above). For instance, the chemical thermodynamics also permits the reduction of the hexachloroplatinate(IV) anion, because E^o([PtCl₆]^2-/Pt,6Cl^-) = +0.68 V, but this complex is too inert kinetically to yield a platinum image within the short time of a few minutes that is available for the reaction to take place. The same was found to be true of [Pt(NH₃)₄]²⁺, [Pd(NH₃)₄]²⁺, [IrCl₆]^3-, and [RhCl₆]^3-. Prolonging the reaction time excessively will only result in re-oxidation of the iron(II) complex by the air.

2.2 Physical State of the Sensitized Layer

The central problem in formulating a chemical model for the platinotype or palladiotype process lies in deciding what phase is appropriate to describe the state of the aqueous sensitizer solution after it has been imbibed into the fibres of a cellulose paper substrate and then dried to a prescribed degree. If it is fully dehydrated, as in the traditional method, it probably takes a microcrystalline form comprising more than one solid phase. (It has recently been found⁸ that when the palladium to iron ratio is very low, a single solid phase is obtained consisting of palladium-doped ferrioxalate crystals; the photochemical reaction may then be modelled by a solid state process involving the conduction band of this single phase. However, this description does not seem applicable to the conditions of the present work in which the molar ratio of palladium to iron takes the stoicheiometric value of 1:2.) Since mixed microcrystals probably predominate in a dry sensitizer, only reaction <1> -the photoreduction of iron(III)- can take place during the exposure. The resulting colour change is slight, with only the shadow tones of the image becoming discernable. Precipitation of the bulk of the platinum or palladium metal does not occur until the exposed paper is immersed in a "developer", i.e. an aqueous solution that mobilises the ions sufficiently for reaction <2> to proceed. In the traditional method of platinotype using a dried iron(III) oxalate sensitizer, the photoproduct is the insoluble iron(II) oxalate, FeC₂O₄; accordingly, the traditional developers contained alkali metal oxalates to render this soluble by complexation, and so permit reaction <2> to take place. Other chelating ligands that bind strongly to iron(III) and maintain a low iron(III)/iron(II) redox potential will also act as "developing" agents: in the present work, disodium ethylenediaminetetraacetate (Edta) was used in preference to oxalate for reasons described in §3.6. It provides a suitably reducing potential:

E^o(Fe(III)Edta/Fe(II)Edta) = +0.12 V

If the trisoxalatoferrate(III) anion is used for the sensitizer, rather than iron(III) oxalate, the photochemistry is somewhat different. Simple iron(II) oxalate, FeC₂O₄, is not the initial photoproduct⁹, but instead an iron(II) complex, such as [Fe(C₂O₄)₂(H₂O)₂]^2- or possibly a dimeric species such as [Fe₂(C₂O₄)₅]^6-, is formed, both of which are quite soluble in water. A "developer" as such is not strictly needed, and the presence of water alone suffices to bring about reaction <2>. Nonetheless there are advantages in retaining the use of a chelating agent such as disodium ethylenediaminetetraacetate in the wet processing procedure, since it is also very effective in removing excess unreacted iron(III) from the paper.

The description so far has ignored any role played by the cellulose paper substrate. It is likely that the components of the sensitizer will be wholly or partially chemisorbed onto the cellulose, especially those species with a strong propensity for hydrogen-bond formation such as [Fe(C₂O₄)₃]^3- or aquated derivatives thereof, like [Fe(C₂O₄)₂(H₂O)₂]^-. If the sensitized paper is not completely dried, but allowed to equilibrate at ambient relative humidity (between 40% and 80%, the latter figure being more typical of Manchester!) then it will also contain significant amounts of absorbed water, as indicated by the cellulose/water absorption isotherm¹⁰, shown in Fig.1. Under conditions of high relative humidity, there is sufficient water hydrogen-bonded within the amorphous regions of the cellulose structure (about ten molecules of water locally to each one of trisoxalatoferrate(III)) to confer a limited mobility on the sensitizer ions and allow reaction <2> to take place within the apparently dry paper. Thus a 'printing-out' process results, in which a complete, or nearly complete, image is formed during the exposure, and requires little or no subsequent development. Such a process has three advantages over the 'development' method:

(i) the ability to inspect the final image at any stage of the exposure does away with the need for prior test-strips;

(ii) there is a self-masking effect in print areas of high optical density, which proportionally resist further darkening; negatives of high density range are accommodated simply by extending the exposure;

(iii) there is no need for a developer.

The present methods have been evolved with a view to maximising this 'printing-out' effect; in particular, the composition of the sensitizer and the control of humidity are important. In this respect, the procedure departs from the method of platinotype generally practised. Although Pizzighelli and Hübl did report a 'print-out' method for platinum in a later revision of their work, this seems not to have been successfully exploited because of difficulties in controlling the humidity.

2.3 Quantitative aspects of the Photochemistry

The quantum yield, , for the photolysis of the trisoxalatoferrate(III) ion in aqueous solution has been determined reliably at several wavelengths of the mercury emission spectrum by a number of independent workers¹¹. All agree that a value of  slightly greater than unity (ca. 1.2) obtains over the wavelength range from 250 to 400 nm, but  falls off sharply to longer wavelengths, becoming insignificant in the yellow/green region of the spectrum and beyond, as shown in Table 1.

The determination of  for trisoxalatoferrate(III) in the solid state is beset with difficulties arising from geometrical-optical effects and the physical state of the sample. Values of  ranging from 0.15 to 1.3 at 365 nm have been reported¹² and the origins of these seemingly discordant results have been discussed in some depth; a value of  = 0.68 at 365 nm seems most reliable¹³. It is also significant that  has been found to be sensitive to the presence of oxygen¹⁴; it is therefore important that the exposing paper should not have uneven access to the atmosphere when in its printing frame.

A theoretical estimate of the relationship between the exposure time and the other parameters of the system can be made as follows:

the number of moles, m, of [Fe(C₂O₄)₃]^3- that are photolysed per unit area, A, of exposed surface may be written as:

m/A = Itf / (Nhc)

m = number of moles of trisoxalatoferrate(III) reacted

A = area of exposed sensitizer in m²

 = quantum yield of the reaction in moles of Fe/einstein

I = intensity of incident light in watts/m²

t = exposure time in seconds

 = wavelength of light in m

f = fraction of incident light absorbed by the photoactive species

N = Avogadro's number = 6.023 x 10²³ mol^-1

h = Planck's constant = 6.6262 x 10^-34 Joule seconds

c = speed of light = 3 x 10⁸ m/s

Inserting the values for the physical constants, we get:

m/A = 8.3612 Itf

In two-component systems, such as the sensitizers used here, f can be written:

f = (1 - 10^-Dl)_FeC_Fe/D

_Fe = decadic molar extinction coefficient of [Fe(C₂O₄)₃]^3- at 

C_Fe = concentration of [Fe(C₂O₄)₃]^3- in the coated layer in mol/dm³

_P = decadic molar extinction coefficient of the Pt or Pd complex at 

C_P = concentration of the Pt or Pd complex in the coated layer in mol/dm³

l = thickness of the sensitized layer in cm (optical path length)

D = _FeC_Fe + _PC_P

This assumes that:

(i) only the [Fe(C₂O₄)₃]^3- is photoactive, and that no energy transfer takes place from the platinum metal complex

(ii) the Beer-Lambert absorption law is valid in a heterogeneous system of metal ions absorbed on cellulose

(iii) there is no significant loss by scattering or absorption due to the other components of the sensitized paper.

The variation of f with  is dependent on the absorption spectra of the two components¹⁵. f will also vary with time, t, of exposure; but for simplicity we will here assume it constant. The expression for f is simplified by making use of the fact that, in the sensitizers used in this work,

C_Fe = 2C_P

and that the concentration of iron (mol/dm³) in the sensitized paper layer is:

C_Fe = 0.1w/l

where w = coating weight of Fe in moles/m² (see §3.3) and the factor of 0.1 is included to correct for the usual units of l in cm, as in the customary definition of .

The absorption spectra of solutions of [Fe(C₂O₄)₃]^3-, [PtCl₄]^2- and [PdCl₄]^2- were recorded from 700 to 200 nm. Values of f were calculated at the wavelengths of the principal mercury lamp emission lines and are included in Table 1.

Table 1. Quantum Yields for the Photolysis of Aqueous [Fe(C₂O₄)₃]^3-

Wavelength of Mercury Emission Line /nm	Quantum Yield (Fe) (moles/einstein)	Fraction f of Light Absorbed by (Fe) in Pt Sensitizer	Fraction f of Light Absorbed by (Fe) in Pd Sensitizer
254	1.25	0.94	0.76
313	1.24	0.99	0.89
365	1.18	0.92	0.90
405	1.14	0.35	0.32
436	1.05	0.073	0.063
468	0.93	0.0078	0.0022
509	0.86	0.00084	0.00043
546	0.15	0.00048	0.00048
579	0.013	0.0011	0.0011
620	<0.01	0.0022	0.0022

The expression for f contains a term for the 'internal filter effect' in the sensitizer due to the absorption of light by the non-photoactive species (the platinum metal salt), and a term for the incomplete absorption of the incident light by the entire sensitized layer. Both terms combine to give values of f that fall sharply with increasing wavelength; it will be noted that absorption in the visible region is negligible, but at 365 nm both sensitizer systems are absorbing about 90% of the incident light. Little is gained by going to even shorter wavelengths; indeed in the palladium system the value of f begins to fall again. These considerations are important in deciding the choice of light source, as will be seen in §3.5. An excitation wavelength of  = 365 nm is recommended as most convenient. At this value, the photochemical yield is given by

m/A = 3.052 x 10^-6 Itf

Taking the value of f as approximately unity, and the coating weight m/A = 0.018 mol/m² for complete photolysis of the sensitizer, we get:

It = 5898 J/m²

Thus, for a typical light source delivering a flux, I, of 50 W/m² to the sensitizer at 365 nm, an exposure time, t = 118 seconds is required. This prediction is born out well in practice for the palladium sensitizer which was found to require exposures of about two minutes, but the platinum sensitizer required about two to three times this calculated exposure. (A more exact calculation was also performed, in which the product If was integrated over the entire waveband of the light source emission, but it gave a result only slightly larger, ca.150 seconds, for the exposure.)