Atomic Weight of Gold, history

In the early years of the nineteenth century the subject of the ratio of gold to oxygen in gold oxide was investigated with divergent results by several chemists, including Richter, Prout, Oberkampf, Dalton, and Thomsen. Dalton regarded the value of the atomic weight as being between 140 and 200 (O=7). In 1813 Berzelius formulated aurous oxide as AuO, and auric oxide as AuO₃, the corresponding atomic weight being Au=2486 (O=100) or 2×198.8 (O = 16). This figure is about twice the modern value Au = 197.2. In 1826 Berzelius was induced by a knowledge of the existence of oxides of the type R₂O₃ to assign to gold the atomic weight Au = 1243 (O = 100), corresponding with Au = 198.8 (O = 16), and to formulate the oxides of gold as Au₂O and Au₂O₃. Later he substituted for these formulae the so-called " equivalent formulae " written with letters having a central, horizontal stroke, Au^I and Au^III, corresponding with AuO and AuO₃. He took 2458.3 as the equivalent, and 1229.165 (O = 100) as the atomic weight of gold. In 1817 Meinecke gave the value Au=200, and in 1826 Gmelin gave as the equivalent Au = 66. Later, the equivalent was assumed to be identical with the atomic weight Au = 196 to 199, the corresponding formulae for the oxides being written as AuO and AuO₃, and for the chlorides as AuCl and AuCl₃.

The principal types of gold compounds are AuX and AuX₃, those of the formula AuX₂ being not improbably formed by combination of the other two forms. The atomic weight Au = 197.2 is supported by cryoscopic and ebullioscopic observations of solutions of gold in other metals, indicating the monatomicity of the element. The vapour-density of gold has not been determined, nor that of any of its compounds. The atomic heat calculated from the specific heat and the atomic weight 197.2 has the normal value 6.4. The element is isomorphous with copper and silver. Its properties and those of its compounds are functions of the atomic weight 197, belonging to an element of the eleventh row of Group I. or of the tenth row of Group VIII. of the periodic system of Mendeleeff.

In the subjoined numerical account the results have been recalculated, employing the antecedent data

O=16.000; Br=79.916; Ba = 137.37; Ag=107.880; K=39.100; Hg=200.6; Cl=35.457; S=32.065;

In 1813 Berzelius precipitated metallic gold from auric-chloride solution by the action of mercury, and as the mean of two experiments found that

3Hg: 2Au = 100: 65.68,
whence
Au = 197.64.

In 1819 Pelletier found Au=238 by analysing aurous iodide. Two years later, Javal found Au=201 by analysing auric oxide, and Au = 104 by analysing potassium aurichloride. In 1823 Figuier found Au = 179 from the analysis of sodium aurichloride. These very inaccurate results are in striking contrast to Berzelius's value.

Five analyses of potassium aurichloride by reduction with hydrogen were made in 1844 by Berzelius, the loss in weight and the quantities of potassium chloride and gold produced being determined. His mean value for the ratio Au: KCl leads to Au = 196.63, a less accurate value than that obtained by him in 1813.

In 1850 Levol converted a weighed quantity of gold into the trichloride, reduced this derivative with sulphurous acid, precipitated as barium sulphate the sulphuric acid formed, and weighed the precipitate. The mean of two values for the ratio 2Au: 3BaSO₄ leads to Au = 196.49.

Thomsen determined in 1876 the proportion of gold and bromine in the compound HAuBr₄,5H₂O, the result being

Au: 4Br=32.11: 52,
whence
Au = 197.39.

Kruss published in 1887 the first modern determinations of any value. The gold employed was carefully purified, special attention being devoted to the elimination of silver and the metals of the platinum group. In his analyses of auric chloride, an aqueous solution of that salt was first prepared. With one weighed portion of the solution the gold was estimated by reduction with sulphurous acid; with another the chlorine was determined by conversion into silver chloride. Eight experiments gave the mean result

3AgCl: Au = 100: 45.824,
whence
Au=197.048.

Kruiss's analyses of carefully purified and dried potassium auribromide, KAuBr₄, were more elaborate. In some experiments the percentage of gold in the salt was determined by reduction with sulphurous acid, in others by heating the auribromide in hydrogen. With sulphurous acid the bromine in the filtrates from the precipitated gold was estimated as silver bromide; with hydrogen the loss in weight on heating in this gas, proportional to BBr, was ascertained, and the potassium bromide dissolved from the residue by water was recovered and weighed, in addition to the gold. The results were

KAuBr₄: Au=100: 35.461, whence Au=197.123*;
Au: 4AgBr=100: 381.021, whence Au=197.150*;
Au: 3Br=100: 121.678, whence Au=197.035*;
Au: KBr=100: 60.390, whence Au= 197.374*.

The work of Thorpe and Laurie was published in 1887. Like Kruss, they employed potassium auribromide as the basis of their research. By heating, the salt was decomposed into gold and potassium bromide, the mixture weighed and extracted with water, and the residual gold weighed. The amount of potassium bromide was ascertained by difference, and the filtrate containing this salt was analysed for bromine by titrating it against silver according to the procedure of Stas. The silver bromide produced was also collected and weighed. The results were

Au: KBr=100: 60.331, whence Au=197.272*;
Ag: Au = 100: 182.827, whence Au=197.234*;
Au: AgBr=100: 95.208, whence Au=197.248*.

The elaborate research of Mallet appeared in 1889. Extreme care was exercised in this investigation, an example of the refinement of the methods employed being the substitution of quartz-sand for filter-paper, to avoid reduction of the gold salts to the metal during filtration.

All Mallet's analyses of auric chloride, auric bromide, and potassium auribromide were made by the same method. The gold in one sample of the compound was determined by reducing with sulphurous acid, collecting the precipitate, heating it in the vacuum of a Sprengel pump, cooling, and weighing. From another sample the halogen was precipitated by a slight excess of silver nitrate prepared from a known weight of silver, and the excess of silver was determined by titration with a standard solution of hydrobromic acid. To avoid weighing the gold salts, they were dissolved in water, and weighed portions of the solutions were employed for the analyses. The results were

(AuCl₃) 3Ag:Au = 100: 60.910, whence Au=197.129*;
(AuBr₃) 3Ag:Au = 100: 60.927, whence Au=197.184*;
(KAuBr₄) 4Ag:Au=100: 45.689, whence Au=197.157*.

A further series of five experiments was made to determine the ratio of the electrochemical equivalents of gold and silver, by passing the same quantity of electricity through solutions of potassium aurocyanide, KAu(CN)₂, and potassium silver cyanide, KAg(CN)₂, and comparing the amounts of gold and silver deposited:

Ag:Au = 100:182.808, whence Au=197.213*.

Mallet also determined the ratio N(CH₃)₃HAuCl₄: Au by estimating the gold left on ignition of trimethylammonium aurichloride, but the values obtained were too high. The same is true of two series of determinations of the ratio H: Au. One of them was carried out by determining the electrochemical equivalent of the metal with respect to hydrogen, both elements being liberated by the same electric current. The other experimental method consisted in the determination of the equivalent of hydrogen with respect to zinc, and also that of gold with respect to this metal by precipitation from auric chloride.

According to Brauner, the results marked with an asterisk (*) and obtained by Kruss, Thorpe and Laurie, and Mallet, are the most reliable. Their arithmetic mean is Au=197.208. Employing the value 107.883 for the atomic weight of silver, Brauner has calculated from them a mean value for the atomic weight of gold, Au = 197.20.

The International Atomic Weight Committee's current table gives

Au = 197.2.