The History of Atomic Weight of Beryllium

Anode-ray analysis indicates that beryllium is a " simple element " with an atomic weight of 9.0 ± 0.1 (Na = 23). Other modern supplemental methods of verifying atomic weights, such as the transparency of beryllium oxide to X-rays, indicate a similar value. Modern methods of checking valency, such as the precipitation of colloidal arsenic sulphide by beryllium salts, also indicate that the element is divalent.

Up to 1885 opinion swayed between the trivalency and divalency of beryllium. Berzelius, controverting Vauquelin, wrote its oxide Be₂O₃ and its atomic weight 13.7 (O = 16), i.e. three times its equivalent. Awdejew and Debray supported the earlier view. Berzelius considered beryllia to be an analogue of alumina because ammonia precipitated its hydroxide from solutions of its salts, because it was insoluble in acids after calcination, and because alumina could be replaced by beryllia in some mineral species. When beryllium chloride was discovered by Rose it appeared to be very similar to aluminium chloride, and when Wohler isolated the metal it seemed to be analogous to aluminium.

Analogies between beryllium and aluminium constantly appear - their band spectra, for example, are very similar. Humpidge noted, in 1883, that the two metals both formed stable double fluorides with potassium and sodium. He also commented on the tendency of beryllium to form basic compounds.

Awdejew failed to obtain a double sulphate of beryllium and potassium analogous to potash alum, and concluded that the composition of many minerals containing beryllium indicated the formula BeO for beryllia. Debray, reviewing the evidence, decided that, though it was indecisive, beryllia should be regarded as similar in constitution to magnesia rather than to alumina.

Since an atomic weignt of about 9 fitted beryllium comfortably into the periodic system of Mendeleef and Lothar Meyer, subsequent chemists usually accepted this value, and Reynolds, in 1876, confirmed it by a determination of the atomic heat. Atterberg, two or three years before, had urged the determination of the specific heat of the metal, since the formulation of some beryllium compounds seemed to require the formula BeO for beryllium oxide.

It was, however, difficult to obtain pure beryllium, and experimenters found specific heats which pointed alternately to about 9 and 13.5 as the atomic weight. The results of Nilson and Pettersson, in 1878, favoured the higher value. Brauner, convinced that Mendeleeff's Periodic Law fixed the atomic weight of beryllium at the lower value, suggested a rise of the specific heat with temperature, and noted the importance of obtaining the vapour density of some volatile compound of beryllium. Reynolds repeated his own determination, and defended his results against Nilson and Pettersson. Humpidge, however, summed up in their favour: reviewing the available data which bore on the atomic weight of beryllium, and arguing that experimental results must outweigh Lothar Meyer's "theoretical conceptions arising from a consideration of Mendeleeff's periodic law." Humpidge subsequently determined the specific heat of beryllium, confirmed the higher value for its atomic weight, and, continuing to prefer practical results to theoretical conceptions, remarked that the result was "unfortunate for the periodic law."

Two years after, in 1885, Humpidge agreed to accept the number which " satisfies the requirements of the periodic law," and recognised the " due importance " of the " periodic arrangement." He discovered that the variation in the specific heat of beryllium with the temperature threw doubt on any deduction of its atomic weight. The specific heat of beryllium increases rapidly to 400° C., and then remains very constant at about 0.62 between 400° C. and 500° C. This constant value agrees, according to the Law of Dulong and Petit, more closely with an atomic weight of about 9.1 than with one of about 13.5. The vapour densities of beryllium chloride and bromide decided the issue, and Humpidge, after determining them, represented the two compounds as BeCl₂ and BeBr₂, "in which the metal is a dyad, and has the atomic weight 9.1." Nilson and Pettersson had, the year before, obtained a similar result for the chloride, and after they verified their result, beryllium may be said to have been confirmed in its divalency and its atomic weight of about 9.1. Three years after, in 1889, Mendeleeff publicly established beryllium in its present position in the Periodic Table.

Rosenheim and Woge confirmed the results of the vapour-density method by determining the molecular weight of beryllium chloride in pyridine. They also described a number of reactions which pointed to the divalency of beryllium. Glassmann has also confirmed its divalency through the molecular weight of the anhydrous picrate.

Mallard observed, in 1889, that crystallised beryllia was isomorphous with crystallised zinc oxide - an indication that beryllium is divalent like zinc. The evidence from isomorphism is, however, not uniform: beryllium silicotungstates, for example, are isomorphous with aluminium silicotungstates. Retgers denied isomorphism between beryllium salts and the salts of the metals of the magnesium group.

In 1895 Lebeau implied trivalency for beryllium by comparing its carbide to that of aluminium and formulating it as Be₄C₃. Henry promptly defended the formula Be₂C, and his conclusion, with the implication of divalency for beryllium, is now generally accepted.

In 1904 Pollok suggested that an undiscovered element in Limoges beryl, with a higher equivalent than beryllium and similar properties to it, had confused many of the results in determining the atomic weight of beryllium. This suggestion, however, has not matured.

Tanatar, about the same time, suggested quadrivalency for beryllium and an atomic weight of 18.2, because some of its compounds appear to have the constitution R3:BeOBe:R3, and because its specific heat at low temperatures is more suited to an atomic weight of 18.2 than to a lower value. Glassmann criticised his conclusion, and such dissent may be said to be now negligible. The vapour densities of beryllium acetylacetonate and basic beryllium acetate confirm the atomic weight as about 9.

The earlier determinations were usually made from the ratio BeO:BaSO₄. Debray analysed the double oxalate, Be(NH₄)₂.(C₂O₄)₂, and obtained the ratio 4CO₂:BeO = 100:14.41. This result, corrected for the atomic weights O = 16 and C = 12.003, gives an atomic weight of 9.36.

Previous to 1922, the most reliable determinations, recalculated from the fundamental values O = 16.000, S = 32.065, C = 12.003, and H = 1.00762, were as follows: -

1. Nilson and Pettersson, ignition of sulphate, BeSO₄.4H₂O:BeO = 100:14.169 (mean of four determinations), atomic weight 9.113.
2. Kruss and Moraht, ignition of sulphate, BeSO₄.4H₂O:BeO = 100:14.144 (mean of sixteen determinations), atomic weight 9.061.
3. Parsons, ignition of acetylacetonate, Be(C₅H₇O₂)₂:BeO = 100:12.112 (mean of seven determinations), atomic weight 9.101.
4. Parsons, ignition of basic acetate, Be₄O(C₂H₃O₂)₆:4BeO = 100:24.698 (mean of nine determinations), atomic weight 9.108.

Concordant results for the equivalent of beryllium have formerly been difficult to obtain from the chloride, because it is difficult to purify completely. In 1922, however, Honigschmid and Birckenbach worked with pure beryllium chloride prepared by heating the highly purified oxide with carbon in a current of chlorine. They obtained the ratio BeCl₂:2Ag = 0.370465:1 from thirteen determinations, and the ratio BeCl₂:2AgCl = 0.278825:1 from five. Both ratios gave Be = 9.018, if Ag = 107.88 and Cl = 35.457.

The International Committee on Atomic Weights for 1925 has adopted the value

Be = 9.02.