--------+--------++--------+--------++--------+-------- 1.838 100.0 1.568 66.0 1.247 33.0 1.840 99.0 1.557 65.0 1.239 32.0 1.841 98.0 1.545 64.0 1.231 31.0 1.841 97.0 1.534 63.0 1.223 30.0 1.840 96.0 1.523 62.0 1.215 29.0 1.838 95.0 1.512 61.0 1.206 28.0 1.836 94.0 1.501 60.0 1.198 27.0 1.834 93.0 1.490 59.0 1.190 26.0 1.831 92.0 1.480 58.0 1.182 25.0 1.827 91.0 1.469 57.0 1.174 24.0 1.822 90.0 1.458 56.0 1.167 23.0 1.816 89.0 1.448 55.0 1.159 22.0 1.809 88.0 1.438 54.0 1.151 21.0 1.802 87.0 1.428 53.0 1.144 20.0 1.794 86.0 1.418 52.0 1.136 19.0 1.786 85.0 1.408 51.0 1.129 18.0 1.777 84.0 1.398 50.0 1.121 17.0 1.767 83.0 1.388 49.0 1.113 16.0 1.756 82.0 1.379 48.0 1.106 15.0 1.745 81.0 1.370 47.0 1.098 14.0 1.734 80.0 1.361 46.0 1.091 13.0 1.722 79.0 1.351 45.0 1.083 12.0 1.710 78.0 1.342 44.0 1.075 11.0 1.698 77.0 1.333 43.0 1.068 10.0 1.686 76.0 1.324 42.0 1.061 9.0 1.675 75.0 1.315 41.0 1.053 8.0 1.663 74.0 1.306 40.0 1.046 7.0 1.651 73.0 1.297 39.0 1.039 6.0 1.639 72.0 1.289 38.0 1.032 5.0 1.627 71.0 1.281 37.0 1.025 4.0 1.615 70.0 1.272 36.0 1.019 3.0 1.604 69.0 1.264 35.0 1.013 2.0 1.592 68.0 1.256 34.0 1.006 1.0 1.580 67.0 --------+--------++--------+--------++--------+--------
APPENDIX B.
ESTIMATION OF SMALL QUANTITIES OF GOLD.[124]
In the case of small buttons of gold the weight can be determined more easily and accurately by measuring with the help of a microscope than by the actual use of a balance. Moreover, the method of measurement is applicable to the determination of quantities of gold too minute to affect even the most delicate balance.
For quantities of gold of from .5 to .005 milligram a microscope with 1/2 inch objective and B eyepiece is suitable. The measurements are made with the help of a scale engraved (or, better, photographed) on a circular piece of glass which rests on the diaphragm of the eyepiece.
This scale and the object upon the stage can be easily brought into focus at the same time. The button of gold obtained by cupelling is loosened from the cupel by gently touching with the moistened point of a knife; it generally adheres to the knife, and is then transferred to a glass slide. The slide is placed on the stage of the microscope, illuminated from below; and the button is brought into focus, and so placed that it apparently coincides with the scale. The diameters in two or three directions (avoiding the flattened surface) are then read off: the different directions being got by rotating the eyepiece. The mean diameter is taken. The weight of the button is arrived at by comparing with the mean diameter of a _standard prill_ of gold of known weight.
The weights are in the proportion of the cubes of the diameters. For example, suppose a prill has been obtained which measures 12.5 divisions of the scale, and that a standard prill weighing 0.1 milligram measures 11.1 divisions. The weight will be calculated as follows:
11.1^{3} : 12.5^{3} :: 0.1 : _x_
0.112.512.512.5 _x_ = -------------------- = 0.143 milligram.
11.111.111.1
The calculations are simplified by the use of a table of cubes. The standard prills used in the comparison should not differ much in size from the prills to be determined. They are prepared by alloying known weights of gold and lead, so as to get an alloy of known composition, say one per cent. gold. Portions of the alloy containing the weight of gold required (say 0.1 milligram) are then weighed off and cupelled on small smooth cupels, made with the finest bone-ash. Care must be taken to remove the cupels as soon as cupellation has finished. Several standard prills of the same size should be made at the same time, and their mean diameter calculated. The lead for making the gold-lead alloy is prepared from litharge purified by reducing from it about 10 per cent. of its lead by fusion with a suitable proportion of flour; the purified litharge is powdered, mixed with sufficient flour and reduced to metal.
In determining the gold contained in small buttons of silver-gold alloy obtained in assaying (and in which the silver is almost sure to be in excess of that required for parting), transfer the button from the cupel to a small clean porcelain crucible; pour on it a drop or two of nitric acid (diluted with half its bulk of water), and heat gently and cautiously until action has ceased. If the residual gold is broken up, move the crucible so as to bring the particles together, so that they may cohere. Wash three or four times with distilled water, about half filling the crucible each time and decanting off against the finger. Dry the crucible in a warm place; and when dry, but whilst still black, take the gold up on a small piece of pure lead. Half a grain of lead is sufficient, and it is best to hold it on the point of a blunt penknife, and press it on the gold in the crucible. The latter generally adheres.
Transfer to a small smooth cupel and place in the muffle. When the cupellation has finished, the button of gold is measured as already described.
PRACTICAL NOTES ON THE IODIDE PROCESS OF COPPER ASSAYING.
For the following remarks and experiments we are indebted to Mr. J.W.
Westmoreland, who has had considerable experience with the process.
Having dissolved the ore he converts the metals into sulphates by evaporating with sulphuric acid. The copper is then separated as subsulphide by means of hyposulphite of soda, and the precipitate is washed, dried, and calcined. The resulting oxide of copper is then dissolved in nitric acid; and to the concentrated solution, a saturated solution of carbonate of soda is added in sufficient quantity to throw down a considerable proportion of the copper. Acetic acid is added to dissolve the precipitate, and when this is effected more of the acid is poured on so as to render the solution strongly acid. To this potassium iodide crystals are added in the proportion of ten parts of iodide to each one part of copper supposed to be present. The solution is then titrated with "hypo" as usual.
For the examination of technical products experiments made in sulphuric acid solutions have no value, since arsenic acid, which is generally present to a greater or less extent, affects the end reaction. In such solutions bismuth may also interfere.
The solution best suited for the assay is one containing acetate of soda and free acetic acid. The presence of acetate of soda counteracts the interference of arsenic and of bismuth.
The return of the blue colour after titration is due to the excessive dilution of the assay, or to an insufficiency of potassium iodide, or to the presence of nitrous fumes. The interference of an excess of sodium acetate is avoided by adding more iodide crystals to the extent of doubling the usual amount.
The interference of lead can be avoided by the addition of sulphuric acid or of phosphate of soda to the acid solution containing the copper, and before neutralising with carbonate of soda. The end reaction is, however, with care distinguishable without this addition. The following experiments, each containing .0648 gram of lead, were made by him in illustration:
---------------+-------------------+---------------+--------------------- Copper taken. Reagent added. Copper found. End reaction.
---------------+-------------------+---------------+--------------------- .2092 gram -- .2077 gram fairly satisfactory .2101 " -- .2092 " "
.2167 " sulphuric acid .2152 " "
.2117 " " .2108 " "
.2109 " phosphate of soda .2092 " good, colourless .2205 " " .2174 " rather yellow ---------------+-------------------+---------------+---------------------
_Effect of Sodium Acetate._--Each solution contained .3343 gram of copper.
a.b.c. d. e. f. g.
grams. grams. grams. grams. grams.
"Acetate" added -- 16.2 16.2 16.2 16.2 "Iodide" added 3.5 3.5 7.0 3.5 7.0 Copper found .3343 .3324 .3351 .3269 .3356
In these experiments, except with the excessive quantities of acetate of soda and the insufficiency of potassium iodide in the cases of c and f, there was no difficulty with the after-blueing.
METHOD OF SEPARATING COBALT AND NICKEL.
The following method of separating and estimating cobalt and nickel has been described by Mr. James Hope,[125] with whom it has been in daily use for several years with completely satisfactory results.
The quantity of ore taken should contain about .5 gram of the mixed metals. It is dissolved in hydrochloric acid or aqua regia, and the solution evaporated to dryness. The residue is taken up with dilute hydrochloric acid and hot water. The solution is filtered off from the silica, freed from second group metals by treatment with sulphuretted hydrogen and filtered, and after oxidation with nitric acid is separated from iron and alumina by the basic acetate method (page 233). The precipitate is redissolved in a little hydrochloric acid, and again precipitated by sodium acetate. The two filtrates are mixed and treated with a little acetic acid, and the cobalt and nickel are then precipitated as sulphides by a current of sulphuretted hydrogen. The precipitate is filtered off, washed, dried, and calcined, and the resulting oxides are weighed to get an idea as to the quantity of the two metals present.
The calcined precipitate is dissolved in a small covered beaker in aqua regia with the help of a few drops of bromine to remove any separated sulphur, and the solution evaporated to dryness with a few drops of sulphuric acid. The residue is dissolved in hot water, diluted to about 50 c.c., and heated to boiling. About 2 grams (four times the quantity of mixed metals present) of ammonium phosphate (AmH_{2}PO_{4}) are weighed off, dissolved in the smallest possible quantity of water, and boiled for a minute or two with a few c.c. of dilute sulphuric acid.
This is added to the boiling-hot solution of cobalt and nickel, which is then treated cautiously with dilute ammonia until the precipitate partially dissolves. The addition of the ammonia is continued drop by drop with constant stirring, until the cobalt comes down as a pink precipitate of ammonium cobalt phosphate (AmCoPO_{4}). The beaker is placed on the top of a water bath with occasional stirring for five or ten minutes. The blue liquid containing the nickel is decanted through a small filter and the precipitate is dissolved with a few drops of dilute sulphuric acid. The resulting solution is treated with a small excess of ammonium phosphate and the cobalt again precipitated by the cautious addition of ammonia exactly as before. The precipitate containing the whole of the cobalt is filtered off and washed with small quantities of hot water. The filtrate is added to the previous one containing the greater part of the nickel.
The ammonium cobalt phosphate is dried, transferred to a platinum crucible, and ignited over a Bunsen flame for fifteen or twenty minutes.
A purple coloured cobalt pyrophosphate (Co_{2}P_{2}O_{7}) is thus formed, and is weighed. It contains 40.3 per cent. of cobalt.
The mixed filtrates containing the nickel are placed in a tall beaker, and dilated if necessary to about 200 c.c. Ten c.c. of strong ammonia are added, and the solution, heated to 70 C., is ready for electrolysis. A battery of two 1-1/2 pint Bunsen cells is used. This is found capable of depositing from .15 to .20 gram of nickel per hour, and from two to three hours is generally sufficient for the electrolysis.
The electrode with the deposited nickel is washed with distilled water, afterwards with alcohol as described under copper, and is then dried and weighed.
The following results obtained with this method by Mr. Hope illustrate the accuracy of the method. They were obtained by working on solutions containing known weights of the two metals:
-------------------------+------------------------- Taken. Found.
------------+------------+------------+------------ Cobalt. Nickel. Cobalt. Nickel.
------------+------------+------------+------------ .1236 gram .1155 gram .1242 gram .1155 gram .1236 " .0577 " .1232 " .0575 "
.2472 " .0577 " .2449 " .0585 "
.3708 " .0577 " .3701 " .0580 "
.0618 " .3465 " .0619 " .3454 "
.0618 " .2310 " .0625 " .2295 "
.0618 " .1155 " .0621 " .1155 "
FOOTNOTES:
[124] For fuller information see a paper on "The Estimation of Minute Quantities of Gold," by Dr. George Tate; read before the Liverpool Polytechnic Society, Nov. 1889.
[125] _Journal of the Society of Chemical Industry_, No 4, vol. ix.
April 30, 1890.
APPENDIX C.
A LECTURE ON THE THEORY OF SAMPLING.
The problem of the sampler is essentially the same as that of the student of statistics. One aims at getting a small parcel of ore, the other a number of data, but each hopes to obtain what shall represent a true average applicable to a much larger mass of material. Ignoring the mechanical part of the problems, the sampling errors of the one and the deviations from the average of the other are the same thing.