J . Chew. SOC., Faraday Trans. 1, 1988, 84(1), 57-64 Thermal Decomposition of Silver Squarate Andrew K. Galwey" and M. Abdel Aziz Mohamedt Chemistry Department, Queen's University, Berfast BT9 5AG, Northern Ireland Michael E. Brown Chemistry Department, Rhodes University, Grahamstown 6140, South Africa The kinetics of the thermal decomposition (473-510 K) of crystalline silver(1) squarate, Ag,C,O,, under reduced pressure in an accumulatory gas apparatus, have been studied for comparison with results obtained for the decompositions of the nickel(@ and copper(I1) salts. No melting was observed and the overall products of decomposition were solid silver particles in a carbonaceous residue, pseudomorphic with the original reactant crystallites, and gaseous CO and CO,. The isothermal a us.time curves were mainly deceleratory and approximated, for a < 0.5, to either the contracting-area or the contracting-volume rate equations, with an apparent activation energy of 190 & 8 kJ mol-l. Silver oxide powder was found to catalyse the decomposition and this, together with the presence of CO, in the gaseous products, led us to suggest that Ag,O is a reaction intermediate, which reacts further with product CO gas. This sequence of reactions in a solid-state decomposition and our failure to detect any recognisable reactant/product interface in electron microscopic studies of partially decomposed material are the central features of this study. The relationship between these results and those for silver oxalate and for nickel and copper squarates is discussed.The metal salts of organic acids provide a group of related substances which permit comparison of the effects of variations of either cation or of anion on thermal stability. The thermal decomposition of many of the simple carboxylates, such as formates and oxalates, have been investigated,' but the patterns of behaviour and proposed reaction mechanisms have not, as yet, been generally agreed. The thermal decomposition of silver squarate has not, to our knowledge, been previously investigated. This salt is particularly suitable for study because it can be prepared in the anhydrous form, so that anion decomposition is not preceded by a dehydration. In this it resembles silver oxalate2* and silver mal~nate,~ but contrasts with the reaction of nickel sq~arate,~ where the dehydration step immediately precedes decomposition and was identified as exerting control over the kinetics of anion breakdown.Silver squarate was also selected to permit comparisons to be made with the thermal reactions of copper squarate.'* ' The squarate anion contains no carboxyl group, and so its study is complementary to the more extensive work1 on the reactions of formates, oxalates and other carboxylates. Experimental Preparation of Silver Squarate Silver squarate was prepared by slowly mixing, at 320 K with continuous stirring, equal volumes of 0.045 mol dm-3 aqueous squaric acid and 0.090 mol dm-3 silver nitrate. The t Permanent address : Chemistry Department, Assiut University, Qena, Egypt. 5758 Thermal Decomposition of Silver Squarate white precipitate stood for 14 h at 298 K, was washed with distilled water and dried.Elemental analysis gave a composition approximating to the acid salt AgHC,04 H,O. The preparation was further treated by warming to 320 K with excess ethanol for 3 h, followed by filtration. The resultant yellow-green solid was dried and stored in the dark as a precaution against possible photolysis (e.g. silver oxalate2. 3). The analysis of the treated salt agreed with the theoretical composition for anhydrous silver squarate, Ag2C404* Kinetic Measurements Isothermal (k 1 K) kinetic studies were based on measurements at known times of the pressure of evolved gas in a constant volume, initially evacuated Pa) glass apparatus. A manually operated McLeod gauge and Baratron diaphragm gauge were used.Output from the Baratron gauge was recorded8 and values of pressures, times and temperatures were stored in a Sinclair Spectrum microcomputer for later kinetic analysis. The gaseous products (CO and CO + CO,) could be studied selectively by using cold traps (78 or 175 K) between the heated reactant and the gauge. Electron Microscopy Jeol 35CF and JSM840 scanning electron microscopes were used. Samples were precoated with a thin film of Au/Pd except for back-scattering studies in the JSM840 instrument. Samples of reactant, of product and of salt, partially decomposed to various known extents, were examined. Results Reaction Stoichiometry The reaction stoichiometry, determined from measurements of the masses of reactant, of the solid residue and of the pressures of gases evolved on completion of decomposition in the known-volume apparatus using either the 78 K trap (to measure CO only) or the 175 K trap (CO + CO,), was : Ag2C404(s) -+ 2Ag(s) + 1.44CO(g) + 0.96C02(g) + [C1.60,.61(s)~ The residual product was shown by X-ray diffraction to contain silver metal but no Ag20.Ag20 is known' to be reduced readily to Ag by CO well below the present reaction temperature. The carbon content of the residue, shown above in square brackets, as determined by combustion analysis, is in satisfactory agreement with that expected (by difference) from the reactant composition and the yields of the volatile products. Mass spectrometry of the gaseous products confirmed the evolution of CO and CO,. Smaller responses were observed at m/e ratios: 114 (possibly traces of squaric acid), 86 and 82 (unidentified), 80 (possibly C,02) and 58 (C20,H, or acid fragment).There was no indication of oxygen (0,) formation. This absence is strong evidence that the decomposition of silver(1) oxide, which yields molecular oxygen" and occurs' below the temperature of silver squarate decomposition (> 470 K), is not a contributory step in the present reaction. A check was made on whether the relative proportions of evolved CO and CO, varied systematically with time, temperature or reactant mass, by making duplicate, but otherwise identical kinetic measurements with either a 78 K or a 175 K trap. Results showed that reaction rates were not detectably different when CO, was allowed to accumulate as a gaseous product and when it was continuously and irreversibly condensed.After both types of experiment the final yields of both products wereJ . Chern. SOC., Faraday Trans. 1, Vol. 84, part 1 Plates 1 and 2 Plate 1. Scanning electron micrographs showing typical crystals of undecomposed reactant (a = 0.0) silver squarate (scale bar = 1.0 pm). Plate 2. Silver squarate crystallites decomposed in vacuum to a = 0.3 at 500 K (scale bar = 1.0 pm). A. K. Galwey, M. A. A. Mohamed and M. E. Brown (Facing p . 58)J. Chem. SOC., Faraday Trans. 1, Vol. 84, part 1 Plate 3 Plate 3. Completely decomposed (a = 1.00) crystallites of silver squarate (both scale bars = 1.0 pm). A. K. Galwey, M. A. A. Mohamed and M. E. BrownA . K. Galwey, M. A .A . Mohamed and M. E. Brown I 59 tlmin Fig. 1. Representative a us. time curves for the isothermal decomposition of silver squarate at different temperatures (K). identical. A further decomposition was completed at 493 K; the 175 K trap present was replaced by a 78 K trap, to condense product CO,, and the residual products were maintained at ca. 500 K for 15 h. The CO pressure was not reduced and the CO, pressure, (175 K trap), was also unchanged. The composition of the final gaseous products was thus not influenced by removal of CO, during or after reaction and the residual silver-carbon mixture does not catalyse the reaction 2co -b CO, + c. Electron Microscopy From electron microscopic studies at various extents of reaction, we could find no evidence of melting8* 11-15 or sintering during silver squarate decomposition. The final residual particles (fractional reaction, a = 1.00) were pseudomorphic with those of the original reactant.The appearance of typical reactant crystallites (a = 0.00) is shown in plate 1. Linear dimensions of individual particles were usually ca. 1 pm, although sometimes with one longer dimension. Surfaces were approximately planar, some included pits, and crystallite corners were usually rounded. Aggregates of many small crystallites were also present. Decomposition (a = 0.3 at 500 K) was accompanied by the appearance of small protuberances on the surfaces, often located at crystallite edges, (plate 2). The decomposition residue (a = 1.00) (plate 3) is comprised of rounded particles (identified from back-scattering electron microscopic measurements as silver crystallites) embedded in a coherent matrix [a carbonaceous polymeric (?) material], pseudomorphic with the original reactant crystallites.This residue is similar in appearance to that from silver malonate decomp~sition.~ Similar aggregation of metal as small rounded particles on an apparently immobile carbonaceous coherent matrix was also described8 for the decomposition of copper(1) malonate, where again the residual particles were pseudomorphic with the reactant crystallites.60 1 .o 0.8 ar 0.6 0.4 0.2 0 Thermal Decomposition of Silver Squarate 0 0 + ' 0 0 t 0 + 0 0 + o o + 0 t 0 0 + 0 + o o + o + o + o + o + o t o + o t: O + o 4* I 20 40 60 80 100 120 140 160 tlmin 0 Fig. 2. Effect of variation of sample mass on the thermal decomposition of silver squarate at 500K: 0, 1 1 ; +, 35mg.Kinetic Measurements Isothermal a us. time curves (473-510 K) for silver squarate decomposition (fig. 1) were predominantly deceleratory with no induction period. Initially (0 < a < 0.45) the reaction rate was almost constant (zero-order), but diminished when a > 0.45, reaching a minimum between 0.55 < a < 0.60. This was followed by an acceleratory process, which became deceleratory when a > 0.85. The intermediate acceleratory region was not observed at the lower end of the temperature range studied (490 K) and, even when present, kinetic behaviour was not very reproducible, depending to some extent on sample mass. Fig. 2 shows two representative a vs. time plots for identical decompositions at 500 K using different reactant masses (1 1 and 35 mg).The mid- reaction acceleratory process was absent when the lower reactant mass was used. This behaviour suggested that the kinetics of this reaction are pressure-dependent. Comparative experiments were thus made in which the total accumulated product gases were measured throughout, and in which the gaseous products were removed by evacuation (for 3 min) at a = 0.45 before the onset of the acceleratory reaction. This kinetic comparison is shown in fig. 3, in the form of a differential plot of Ap/At vs. time, where Ap is the pressure difference between consecutive readings made at constant time intervals, At, for two such experiments at 500 K. The later acceleratory process was eliminated, or very significantly reduced, following evacuation of product gases at a = 0.45.The kinetic data for the initial process (a < 0.5) obeyed, with equal acceptability," both the contracting-area and the contracting-volume equations.' The latter fitted the data over the wider range (fig. 4). The kinetic data could,also be satisfactorily expressed by the first-order expression for 0.30 < a < 0.70, but at higher values of a( > 0.7) obedience was poor. Reaction is thus best represented by a contracting-volume model.' It is concluded that decomposition is initiated at the original crystallite surfaces and that chemical changes proceed within a zone that progressively advances inwards. The activation energy for the first part of reaction (a < 0.5) was 190+8 kJ mol-' (483-508 K).A .K. Galwey, M. A . A . Mohamed and M. E. Brown aB, 00 c 0 0 0 0 0 0 0 0 OO OO 0 0 0 I I 40 80 1: tlmin 61 I Fig. 3. Plots of Ap/At vs. time for two experiments at 500 K: 0, product gases evacuated (3 min) at a = 0.45; 0, in the presence of the continued accumulation of the product gases. 0.2 ””, I I 3 v 3 0.1 0 20 40 60 80 100 120 140 160 tlmin Fig. 4. Test for conformity of the data shown in fig. 1 to the contracting-volume equation.] Pre-crushing the reactant appreciably increased the rate of decomposition. This is evidence that the reaction rate was influenced by the surface area of the original reactant crystallites. Mixing silver squarate with 10% (by mass) of silver metal powder did not change the kinetics of decomposition, the gaseous product yields (at a = 1.00), or the CO/CO, ratio.Addition, with crushing, of 10% (by mass) of silver oxide powder62 Thermal Decomposition of Silver Squarate (Ag,O) to the reactant accelerated the breakdown of silver squarate so strongly that reaction was too vigorous to permit kinetic measurements within the present temperature interval. Ag,O is thus identified as an excellent catalyst for the decomposition of silver squarate, and may also occur as an active reaction intermediate. Discussion Reaction Mechanism Discussion will be concerned with the two complementary aspects of behaviour that require''' l2 consideration in the formulation of the mechanism of a solid-state reaction : the reaction geometry and the chemistry of the changes occurring within the reaction zone.The absence of autocatalysis or of any acceleration of decomposition on mixing with silver, together with the absence from the electron micrographs of any recognizable interface, lead us to conclude that there is no well defined reactant-product-active contact zone within which the chemical changes occur preferentially. Our observations (plate 3) show that metallic particles tend to be generated and to grow on the external surfaces of the crystallites,' whereas the carbonaceous residue is effectively immobile, preserving the sizes and shapes of the original reactant particles. Such development of silver particles at sites remote from reaction is evidence that there is no catalytically active silver-silver squarate contact interface and that silver metal does not participate in the anion-breakdown step.This contrasts with the mechanism described for silver malonate., The kinetic obedience to the contracting-volume expression, taken with the increase in rate caused by reactant crushing, suggests that the decomposition zone progressively proceeds inwards from the original crystallite surfaces. No well demarcated interface, within which an autocatalytic chemical change was completed, could be recognized. The evidence was that silver metal migrated beyond the zone at which it was generated. Intracrystalline Chemistry In formulating a detailed reaction mechanism, we first eliminate several possible intermediates. (i) If the initial steps in reaction were electron transfers, this could yield cyclobutanetetraone : Ag,C,04 + 2Ag + c40,.This mechanism was excluded because there is no obvious or plausible reorganization whereby the (CO), intermediate could give the significant yields of product CO, observed. (ii) The unsaturated anion might plausibly rearrange to silver acetylide. This reaction was excluded because silver acetylide is unstable'' at reaction temperature and would decompose rapidly. (iii) It is most improbable that silver would be oxidized to Ag2+: Ag,C,O, + Ag i- Ag2+C40:- under the present predominantly reducing reaction conditions. The formation of CO, provides an important insight into the reaction mechanism. This requires transfer of oxygen between carbon atoms and can be most satisfactorily explained here through the intervention of Ag20.In contrast, CO, is formed in only a relatively small yield during nickel squarate decomp~sition,~ where the nickel residue is a more active heterogeneous catalyst than our silver product.A . K. Galwey, M. A . A . Mohamed and M. E. Brown 63 Our proposed reaction mechanism is as follows: [C,O,] + 2.5CO + [C,.,O,.,] (polymeric residue) [Ag,O] + CO + 2Ag + CO,. Cherall this is Ag,C,O, + 2Ag + CO, + 1.5CO + (c1.50,)5)n which is in satisfactory agreement with the stoichiometric data above and also explains the formation of ca. 1 mol of CO, per mol of salt decomposed. The reduction of Ag,O by CO is very rapid under reaction conditions, so that this intermediate is short- lived.' Reaction proceeding through two consecutive steps accounts for the absence of a reaction interface, since the silver metal product is not in direct contact with the solid reactant.Breakdown of the postulated intermediate, [C,O,], can be expected to yield CO, and the polymeric residue containing some oxygen. The mid-reaction acceleratory process, observed during decompositions with product- gas accumulation, is ascribed to changes in the participating chemical steps rather than interface geometry.'l? l2 We suggest that the product CO, within the residual carbonaceous layer, interacts directly with the undecomposed salt, or promotes its breakdown, when the prevailing pressure is sufficient. The overall chemical change is unaltered : Ag,C,O, + CO + 2Ag + CO, + [C,O,] but the first step and oxide reduction are accelerated. The subsequent decomposition of [C,O,] is unchanged because there is no CO involvement.It is improbable that this mid- reaction acceleratory process arises through self-heating. Moreover, while this pattern of changes of reaction rate is formally similar to Smith-Topley behaviour,' it seems improbable that the present products (CO and CO,, but not H,O) promote the textural changes that explain this characteristic behaviour of hydrates. Two central features of this proposed mechanism differ from the behaviour usually regarded as characteristic of solid-state decompositions or crystolysisl' reactions. These are the absence of an active reactant-product interface and the occurrence of a secondary reaction of an initial product (Ag,O) with a product gas (CO). Thus, as is often found in this field,'Y l9 the thermal reactions of silver squarate differ from those of related reactants containing common constituents, e.g. silver malonate, or nickel squarate., Although silver malonate decomposes in a similar temperature interval,, the generation and development of a reaction interface during this nucleation and growth process involves a quite different sequence of chemical changes and controls from those discussed here.This difference may be the consequence of the hydrogen in the reactant anion, CH,(CO,),, which permits an entirely different surface chemistry involving reactive chemisorbed intermediates. The decomposition of silver squarate also differs from that of silver oxalate,'-, which occurs in a lower temperature range and is a nucleation and growth process exhibiting a sigmoid a us.time curve. The first step in Ag,C,O, breakdown has been identified as electron transfer. This decomposition, however, yields Ag and CO, products only, thus reducing the possibility of surface deactivation of the silver metal by deposited carbon or chemisorbed CO, and consequently the residual product silver readily promotes anion breakdown. The reaction of silver squarate shows several points of dissimilarity with the decomposition of nickel squarate,, which, however, proceeds in a similar temperature range. The kinetics of decomposition of NiC,O, .2H,O are dominated by the precursor dehydration step and this salt yields a higher proportion of product CO. The reaction of silver squarate did, however, exhibit several points of similarity with copper 3 FAR I64 Thermal Decomposition of Silver Squarate squarate,'.? which will be considered in detail in the context of formulating a reaction mechanism for this latter compound, discussed in a forthcoming paper.? The authors thank Mr J.McCrae and his staff and Mr R. H. M. Cross for helpful advice in obtaining the electron micrographs. M.A.M. thanks the Egyptian Government and the ORS Award Scheme for Scholarships held during the period of this work. M.E.B. acknowledges financial support from the South African CSIR. References 1 M. E. Brown, D. Dollimore and A. K. Galwey, Comprehensive Chemical Kinetics, Vol. 22. Reactions in 2 A. Finch, P. W. M. Jacobs and F. C. Tompkins, J. Chem. SOC., 1954, 2053. 3 A. G. Leiga, J. Phys. Chem., 1966,70, 3254; 3260. 4 A. K. Galwey and M. A. Mohamed, J. Chem. SOC., Faraday Trans. I , 1985, 81, 2503. 5 A. K. Galwey and M. E. Brown, J. Chem. SOC., Faraday Trans. I , 1982, 78, 41 1. 6 M. E. Brown, A. K. Galwey and M. W. Beck, Zsr. J. Chem., 1982, 22,215. 7 A. K. Galwey, M. A. Mohamed, S. Rajam and M. E. Brown, to be published. 8 N. J. Carr and A. K. Galwey, Proc. R. SOC. London, Ser. A, 1986, 404, 101. 9 I. Nakamori, H. Nakamura, T. Hayano and S. Kagawa, Bull. Chem. SOC. Jpn, 1974,47, 1827. the Solid State (Elsevier, Amsterdam, 1980). 10 G. V. Malinin and Yu. M. Tolmachev, Russ. Chem. Rev., 1975,44, 392. 11 A. K. Galwey, Proc. 7th Znt. Con$ Thermal Analysis, Kingston, Ontario (Wiley, New York, 1982), 12 A. K. Galwey, Thermochim. Acta, 1985, 96, 259. 13 A. K. Galwey, R. Spinicci and G. G. T. Guarini, Proc. R. SOC. London, Ser. A, 1981, 378, 477. 14 A. K. Galwey and L. Poppl, Philos. Trans. R. SOC. London, Ser. A, 1984, 311, 159. 15 A. K. Galwey, L. Poppl and S. Rajam, J. Chem. SOC., Faraday Trans. I , 1983, 79, 2143. 16 M. E. Brown and A. K. Galwey, Thermochim. Acta, 1979, 29, 129. 17 J. D. McCowan, Trans. Faraday SOC., 1963, 59, 1860. 18 N. J. Carr and A. K. Galwey, Thermochim. Acta, 1984, 79, 323. 19 D. A. Young, Decomposition of Sofia3 (Pergamon, Oxford, 1966). p. 38. Paper 612314; Received 1st December, 1986