The determination of the combined water may often be made by simply igniting the substance from which the moisture has been removed. The quantity of water may be determined, either indirectly by the loss, or directly by collecting it in a calcium chloride tube, and weighing. In some cases, in which the loss on ignition does not give simply the proportion of combined water, it can be seen from the analysis to what else the loss is due; and, after a proper deduction, the amount of water can be estimated. For example, 1 gram of crystallised iron sulphate was found to contain on analysis 0.2877 gram of sulphuric oxide; and on igniting another gram, 0.2877 gram of ferric oxide was left. As the salt is known to be made up of ferrous oxide, sulphuric oxide, and combined water, the combined water can be thus calculated: 0.2877 gram of ferric oxide is equal to 0.2589 gram of ferrous oxide,[98] and consequently, the loss on ignition has been diminished by 0.0288 gram, which is the weight of oxygen absorbed by the ferrous oxide during calcining. The loss on ignition was 0.7123 gram, to which must be added 0.0288 gram; hence 0.7411 gram is the weight of the combined sulphuric oxide and water present. Deducting the weight of sulphuric oxide found, 0.2877 gram, there is left for combined water 0.4534 gram. The composition of 1 gram of the dry salt is then:—

Water 0.4534 Sulphuric oxide 0.2877 Ferrous oxide 0.2589 ——— 1.0000

The following is another example:—A sample of malachite lost on ignition 28.47 per cent., leaving a residue which was found on analysis to be made up of oxide of copper (equal to 70.16 per cent. on the mineral), and silica and oxide of iron (equal to 1.37 per cent.). Carbon dioxide and water (but nothing else) was found to be present, and the carbon dioxide amounted to 19.64 per cent.; deducting this from the loss on ignition, we have 8.82 as the percentage of water present. The analysis was then reported as follows:—

Cupric oxide 70.16 equal to 56.0% copper. Silica and ferric oxide 1.37 Carbon dioxide 19.64 Water 8.82 ——- 99.99



Direct Determination of Combined Water.—Transfer about 3 grams of the substance to a piece of combustion tube (8 or 10 inches long), attached (as in fig. 63) at one end to a U-tube containing sulphuric acid, and at the other end to a calcium chloride tube. The last is weighed previous to the determination. The tube should be warmed to ensure complete dryness, and must be free from a misty appearance. Aspirate a current of air through the apparatus, heat the mineral by means of a Bunsen burner, cautiously at first, and afterwards to redness (if necessary). The water is driven off and condenses in the calcium chloride tube, which is afterwards cooled and weighed. The increase in weight is due to the water. If the substance gives off acid products on heating, it is previously mixed with some dry oxide of lead or pure calcined magnesia.

EXAMINATION OF WATERS.

The assayer is occasionally called on to test water for the purpose of ascertaining the nature and quantity of the salts contained in it, and whether it is or is not fit for technical and drinking purposes.

In mineral districts the water is generally of exceptional character, being more or less charged, not only with earthy salts, but also frequently with those of the metals. Distilled water is only used by assayers in certain exceptional cases, so that by many it would be classed among the rarer oxides. Water of ordinary purity will do for most purposes, but the nature and quantity of the impurities must be known.

The following determinations are of chief importance:—

Total Solids at 100 C.—Where simply the amount is required, take 100 c.c. and evaporate on the water-bath in a weighed dish; then dry in the water-oven, and weigh.

Total Solids Ignited.—The above residue is very gently ignited (keeping the heat well below redness), and again weighed. A larger loss than 4 or 5 parts per 100,000 on the water requires an explanation.

Chlorine.—Take 100 c.c. of the water in a porcelain dish, add 2 c.c. of a 5 per cent. solution of neutral potassic chromate, and titrate with a neutral standard solution of nitrate of silver, made by dissolving 4.789 grams of crystallised silver nitrate in distilled water, and diluting to 1 litre. The addition of the nitrate of silver is continued until the yellow of the solution assumes a reddish tint. The reaction is very sharp. Each c.c. of nitrate of silver used is equal to 1 part by weight of chlorine in 100,000 of water. At inland places this rarely amounts to more than 1 in 100,000; but near the sea it may amount to 3 or 5. More than this requires explanation, and generally indicates sewage pollution.

Nitric Pentoxide (N{2}O{5}).—It is more generally reported under the heading, "nitrogen as nitrates." Take 250 c.c. of the water and evaporate to 2 or 3 c.c.; acidulate with a few drops of dilute sulphuric acid, and transfer to a nitrometer (using strong sulphuric acid to wash in the last traces). The sulphuric acid must be added to at least twice the bulk of the liquid. Shake up with mercury. The mercury rapidly flours, and nitric oxide is given off (if any nitrate is present). The volume of the nitric oxide (corrected to normal temperature and pressure), multiplied by 0.25, gives the parts of nitrogen per 100,000; or, multiplied by 0.965, will give the nitric pentoxide in parts per 100,000. In well and spring waters the nitrogen may amount to 0.3 or 0.4 parts per 100,000; or in richly cultivated districts 0.7 or 0.8 parts per 100,000. An excess of nitrates is a suspicious feature, and is generally due to previous contamination.

Ammonia.—Take 500 c.c. of the water and place them in a retort connected with a Liebig's condenser. Add a drop or two of a solution of carbonate of soda and distil over 100 c.c.; collect another 50 c.c. separately. Determine the ammonia in the distillate colorimetrically (with Nessler's solution, as described under Ammonia) and compare with a standard solution of ammonic chloride containing 0.0315 gram of ammonic chloride in 1 litre of water. One c.c. contains 0.01 milligram of ammonia. The second distillate will show little, if any, ammonia in ordinary cases. The amounts found in both distillates are added together, and expressed in parts per 100,000.

Waters (other than rain and tank waters) which contain more than 0.003 per 100,000 are suspicious.

Organic Matter.—The organic matter cannot be determined directly; but for ordinary purposes it may be measured by the amount of permanganate of potassium which it reduces, or by the amount of ammonia which it evolves on boiling with an alkaline permanganate of potassium solution.

A. Albuminoid Ammonia.—To the residue left after distilling the ammonia add 50 c.c. of a solution made by dissolving 200 grams of potash and 8 grams of potassium permanganate in 1100 c.c. of water, and rapidly boiling till the volume is reduced to 1 litre (this should be kept in a well stoppered bottle, and be occasionally tested to see that it is free from ammonia). Continue the distillation, collecting 50 c.c. at a time, until the distillate is free from ammonia. Three or four fractions are generally sufficient. Determine the ammonia colorimetrically as before. If the total albuminoid ammonia does not exceed 0.005 in 100,000, the water may be regarded as clean as regards organic matter; if it amounts to more than 0.015, it is dirty.

B. Oxygen Consumed.—A standard solution of permanganate of potash is made by dissolving 0.395 gram of the salt in water and diluting to 1 litre. Each c.c. equals 0.1 milligram of available oxygen. The following are also required:—1. A solution of sodium hyposulphite containing 1 gram of the salt (Na{2}S{2}O{3}.5H{2}O) in 1 litre of water. 2. Dilute sulphuric acid, made by adding one part of the acid to three of water, and titrating with the permanganate solution till a faint pink persists after warming for several hours. 3. Starch paste. 4. Potassium iodide solution.

Take 250 c.c. of the water in a stoppered bottle, add 10 c.c. of sulphuric acid and 10 c.c. of the permanganate, and allow to stand in a warm place for four hours. Then add a few drops of the solution of potassium iodide, and titrate the liberated iodine with "hypo," using starch paste towards the end as an indicator. To standardise the hyposulphite, take 250 c.c. of water and 10 c.c. of sulphuric acid, and a few drops of potassium iodide; then run in 10 c.c. of the "permanganate" solution, and again titrate; about 30 c.c. of the "hypo" will be used. The difference in the two titrations, divided by the last and multiplied by 10, will give the c.c. of permanganate solution used in oxidising the organic matter in the 250 c.c. of water. Each c.c. represents 0.04 parts of oxygen in 100,000.

Metals.—These may for the most part be estimated colorimetrically.

Lead.—Take 100 c.c. of the water in a Nessler tube, and add 10 c.c. of sulphuretted hydrogen water, and compare the tint, if any, against a standard lead solution, as described under Colorimetric Lead. Report in parts per 100,000.

Copper.—Proceed as with the last-mentioned metal; but, if lead is also present, boil down 500 c.c. to about 50 c.c., then add ammonia, filter, and estimate the copper in the blue solution, as described under Colorimetric Copper.

Iron.—Take 50 c.c., or a smaller quantity (if necessary), dilute up to the mark with distilled water, and determine with potassium sulphocyanate, as described under Colorimetric Iron.

Zinc.—Zinc is the only other metal likely to be present; and, since it cannot be determined colorimetrically, it must be separately estimated during the examination of the "total solids."

Examination of "Total Solids."—Evaporate 500 c.c. to dryness with a drop or two of hydrochloric acid. Take up with hydrochloric acid, filter, ignite, and weigh the residue as "silica." To the filtrate add a little ammonic chloride and ammonia, boil and filter, ignite, and weigh the precipitate as "oxide of iron and alumina." Collect the filtrate in a small flask, add a few drops of ammonium sulphide or pass sulphuretted hydrogen, cork the flask, and allow to stand overnight; filter, wash, and determine the zinc gravimetrically as oxide of zinc. If copper or lead were present, they should have been previously removed with sulphuretted hydrogen in the acid solution. To the filtrate add ammonic oxalate and ammonia, boil for some time, allow to stand, filter, wash, ignite, and weigh as "lime." Evaporate the filtrate with nitric acid, and ignite. Take up with a few drops of dilute hydrochloric acid, add baric hydrate in excess, evaporate, and extract with water. The residue contains the magnesia; boil with dilute sulphuric acid, filter, precipitate it with phosphate of soda and ammonia, and weigh as pyrophosphate. The aqueous extract contains the alkalies with the excess of barium. Add sulphuric acid in slight excess, filter, evaporate, and ignite strongly. The residue consists of the sulphates of the alkalies (which are separately determined, as described under Potash).

Sulphuric Oxide (SO_{3}).—Take 200 c.c. and boil to a small bulk with a little hydrochloric acid, filter (if necessary), add baric chloride solution in slight excess to the hot solution, filter, ignite, and weigh as baric sulphate.

Carbon Dioxide (free).—Carbon dioxide exists in waters in two forms, free and combined. The latter generally occurs as bicarbonate, although on analysis it is more convenient to consider it as carbonate, and to count the excess of carbon dioxide with the free. The method is as follows:—To determine the free carbon dioxide, take 100 c.c. of the water, place them in a flask with 3 c.c. of a strong solution of calcium chloride and 2 c.c. of a solution of ammonic chloride, next add 50 c.c. of lime-water. The strength of the lime-water must be known. Make up to 200 c.c. with distilled water, stop the flask, and allow the precipitate to settle. Take out 100 c.c. of the clear solution with a pipette, and titrate with the standard solution of acid.[99] The number of c.c. required, multiplied by two, and deducted from that required for the 50 c.c. of lime-water, and then multiplied by 0.0045, will give the carbon dioxide present other than as normal carbonates.

Carbon Dioxide combined as normal carbonate.—100 c.c. of the water are tinted with phenacetolin or lacmoid; then heated to near boiling, and titrated with standard acid. The number of c.c. used, multiplied by 0.0045, will give the weight in grams of the combined carbon dioxide.

Free Acid.—In some waters (especially those from mining districts) there will be no carbonates. On the contrary, there may be free mineral acid or acid salts. In these cases it is necessary to determine the amount of acid (other than carbon dioxide) present in excess of that required to form normal salts. This is done in the following way:—Make an ammoniacal copper solution by taking 13 grams of copper sulphate (CuSO{4}.5H{2}O), dissolving in water, adding solution of ammonia until the precipitate first formed has nearly dissolved, and diluting to 1 litre. Allow to settle, and decant off the clear liquid. The strength of this solution is determined by titrating against 10 or 20 c.c. of the standard solution of sulphuric acid (100 c.c. = 1 gram H{2}SO{4}). The finishing point is reached as soon as the solution becomes turbid from precipitated cupric hydrate. At first, as each drop falls into the acid solution, the ammonia and cupric hydrate combine with the free acid to form ammonic and cupric sulphates; but as soon as the free acid is used up, the ammonia in the next drop not only precipitates an equivalent of cupric hydrate from the solution, but also throws down that carried by itself. This method is applicable in the presence of metallic sulphates other than ferric. The standardising and titration should be made under the same conditions. Since sulphuric acid and sulphates are predominant in waters of this kind, it is most convenient to report the acidity of the water as equivalent to so much sulphuric acid.

Dissolved Oxygen.—For the gasometric method of analysing for dissolved oxygen, and for the Schtzenberger's volumetric method, the student is referred to Sutton's "Volumetric Analysis." The following is an easy method of estimating the free oxygen in a water:—Take 20 c.c. of a stannous chloride solution (about 20 grams of the salt with 10 c.c. of hydrochloric acid to the litre); add 10 c.c. of hydrochloric acid, and titrate in an atmosphere of carbon dioxide with standard permanganate of potassium solution (made by dissolving 1.975 gram of the salt in 1 litre of water: 1 c.c. equals 0.5 milligram of oxygen). A similar titration is made with the addition of 100 c.c. of the water to be tested. Less permanganate will be required in the second titration, according to the amount of oxygen in the water; and the difference, multiplied by 0.5, will give the weight of the oxygen in milligrams. Small quantities of nitrates do not interfere.

In REPORTING the results of the analysis, it is customary to combine the acids and bases found on some such principle as the following:—The sulphuric oxide is calculated as combined with the potash, and reported as potassic sulphate (K{2}SO{4}); the balance of the sulphuric oxide is then apportioned to the soda, and reported as sulphate of soda (Na{2}SO{4}); if any is still left, it is reported as calcium sulphate (CaSO{4}), and after that as magnesic sulphate (MgSO{4}). When the sulphuric oxide has been satisfied, the chlorine is distributed, taking the bases in the same order, then the nitric pentoxide, and lastly the carbon dioxide. But any method for thus combining the bases and acids must be arbitrary and inaccurate. It is extremely improbable that any simple statement can represent the manner in which the bases and acids are distributed whilst in solution; and, since different chemists are not agreed as to any one system, it is better to give up the attempt, and simply state the results of the analysis. This has only one inconvenience. The bases are represented as oxides; and, since some of them are present as chlorides, the sum total of the analysis will be in excess of the actual amount present by the weight of the oxygen equivalent to the chlorine present as chloride. The following is an example of such a statement:—

Parts per 100,000. Total solids, dried at 100 C. 28.73 Chlorine 1.70 Nitrogen as nitrate 0.03 Ammonia 0.001 Albuminoid ammonia 0.004 "Oxygen consumed" in 4 hours 0.01

The solids were made up as under:—

Per 100,000 of the Water. Potash 0.38 Soda 2.01 Magnesia 1.44 Lime 10.55 Ferric oxide 0.01 Silica 0.30 Sulphuric oxide 3.69 Nitrogen pentoxide 0.11 Carbon dioxide 8.38 Chlorine 1.70 Volatile and organic matter 0.66 ——- 29.23 Less oxygen equivalent to chlorine found 0.39 ——- 28.84

For the preparation of distilled water, the apparatus shown in fig. 64 is convenient for laboratory use. It consists of a copper retort heated by a ring gas-burner, and connected with a worm-condenser.



PRACTICAL EXERCISE.

A mineral, on analysis, gave the following results:—Water, 44.94 per cent.; sulphuric oxide, 28.72 per cent.; ferrous iron, 13.92 per cent.; ferric iron, 0.35 per cent.; copper, 6.1 per cent. The mineral was soluble in water, and showed nothing else on testing. How would you report the analysis? Calculate the formula for the salt.

THE HALOGENS.

There is a group of closely allied elements to which the name halogen (salt-producer) has been given. It comprises chlorine, bromine, iodine, and fluorine. These elements combine directly with metals, forming as many series of salts (chlorides, bromides, iodides, and fluorides), corresponding to the respective oxides, but differing in their formul by having two atoms of the halogen in the place of one atom of oxygen. For example, ferrous oxide is FeO and ferrous chloride is FeCl{2}, and, again, ferric oxide is Fe{2}O{3}, whilst ferric chloride is Fe{2}Cl{6}. These salts differ from the carbonates, nitrates, &c., in containing no oxygen. Consequently, it is incorrect to speak of such compounds as chloride of potash, fluoride of lime, &c., since potash and lime are oxides. It is important to bear this in mind in reporting analyses in which determinations have been made, say, of chlorine, magnesia, and potash, or of fluorine, silica, and alumina. It is necessary in all such cases to deduct from the total an amount of oxygen equivalent to the halogen found, except, of course, where the base has been determined and recorded as metal. Compounds containing oxides and fluorides, &c., do not lend themselves to the method of determining the halogen by difference. For example, topaz, which, according to Dana, has the formula Al{2}SiO{4}F{2}, would yield in the ordinary course of analysis—

Alumina 55.4% Silica 32.6 Fluorine 20.6 ——- 108.6

The oxygen equivalent to 20.6 per cent. fluorine may be found by multiplying the percentage of fluorine by 0.421; it is 8.7 per cent., and must be deducted. The analysis would then be reported thus:—

Alumina 55.4% Silica 32.6 Fluorine 20.6 ——- 108.6 Less oxygen equivalent to fluorine 8.7 ——- 99.9

Take as an illustration the following actual analysis by F.W. Clarke and J.S. Diller:—

Alumina 57.38% Silica 31.92 Fluorine 16.99 Potash 0.15 Soda 1.33 Water 0.20 ——— 107.97 Deduct oxygen equivalent 7.16 ——— 100.81

In calculating the factor for the "oxygen equivalent," divide the weight of one atom of oxygen (16) by the weight of two atoms of the halogen; for example, with chlorine it would be 16/71 or 0.2253; with bromine, 16/160 or 0.1000; with iodine, 16/254 or 0.063; and with fluorine, 16/38 or 0.421.

CHLORINE AND CHLORIDES.

Chlorine occurs in nature chiefly combined with sodium, as halite or rock salt (NaCl). With potassium it forms sylvine (KCl), and, together with magnesium, carnallite (KCl.MgCl{2}.6H{2}O). Of the metalliferous minerals containing chlorine, kerargyrite, or horn silver (AgCl), and atacamite, an oxychloride of copper (CuCl{2}.3Cu(HO){2}.) are the most important. Apatite (phosphate of lime) and pyromorphite (phosphate of lead) contain a considerable amount of it. Chlorine is a gas of a greenish colour, possessing a characteristic odour, and moderately soluble in water. It does not occur native, and is generally prepared by the action of an oxidising agent on hydrochloric acid. It combines directly with metals at the ordinary temperature (even with platinum and gold), forming chlorides, which (except in the case of silver) are soluble.

It is important in metallurgy, because of the extensive use of it in extracting gold by "chloridising" processes. It is also used in refining gold.

Detection.—Compounds containing the oxides of chlorine are not found in nature, because of the readiness with which they lose oxygen. By reduction they yield a chloride; the form in which chlorine is met with in minerals. In testing, the compound supposed to contain a chloride is boiled with water, or, in some cases, dilute nitric acid. To the clear solution containing nitric acid a few drops of nitrate of silver solution are added. If, on shaking, a white curdy precipitate, soluble in ammonia, separates out, it is sufficiently satisfactory evidence of the presence of chlorides.

Solution and Separation.—The chlorides are generally soluble in water, and are got into solution by extracting with warm dilute nitric acid. Or, if insoluble, the substance is fused with carbonate of soda, extracted with water, and the filtrate acidified with nitric acid. For the determination, it is not necessary to obtain the solution of the chloride free from other acids or metals. If tin, antimony, mercury, or platinum is present, it is best to separate by means of sulphuretted hydrogen. The chloride is determined in the solution after removal of the excess of the gas. Where traces of chlorides are being looked for, a blank experiment is made to determine the quantity introduced with the reagents. One hundred c.c. of ordinary water contains from 1 to 3 milligrams of chlorine. On the addition of nitrate of silver to the nitric acid solution, chloride of silver separates out. This is free from other substances, except, perhaps, bromide and iodide.

GRAVIMETRIC DETERMINATION.

Freely mix the solution containing the chloride with dilute nitric acid, filter (if necessary), and treat with nitrate of silver. Heat nearly to boiling, and, when the precipitate has settled, filter, and wash with hot distilled water. Dry, and transfer to a weighed Berlin crucible. Burn the filter-paper separately, and convert any reduced silver into chloride by alternate treatment with drops of nitric and of hydrochloric acid. Add the main portion to this, and heat cautiously till the edges of the mass show signs of fusing (about 260). Cool in the desiccator and weigh. The substance is chloride of silver (AgCl), and contains 24.73 per cent. of chlorine.

The precipitated chloride is filtered and washed as soon as possible after settling, since on exposure to light it becomes purple, and loses a small amount of chlorine.

VOLUMETRIC METHOD.

There are several volumetric methods; but that based on the precipitation of silver chloride in neutral solution, by means of a standard solution of silver nitrate (using potassium chromate as indicator), is preferred. Silver chromate is a red-coloured salt; and, when silver nitrate is added to a solution containing both chloride and chromate, the development of the red colour marks off sharply the point at which the chloride is used up. Silver chromate is decomposed and consequently decolorised by solution of any chloride. The solution for this method must be neutral, since free acid prevents the formation of the red silver chromate. If not already neutral, it is neutralised by titrating cautiously with a solution of soda. In a neutral solution, other substances (such as phosphates and arsenates) also yield a precipitate with a solution of nitrate of silver; and will count as chloride if they are not removed.

_The Standard Solution of Nitrate of Silver_ is made by dissolving 23.94 grams of the salt (AgNO_{3}) in distilled water, and diluting to 1 litre; 100 c.c. are equal to 0.5 gram of chlorine.

The indicator is made by adding silver nitrate to a strong neutral solution of yellow chromate of potash (K{2}CrO{4}), till a permanent red precipitate is formed. The solution is allowed to settle, and the clear liquid decanted into a stoppered bottle labelled "chromate indicator for chlorine."

Standardise the silver nitrate by weighing up 0.5 gram of pure sodium chloride (or potassium chloride). Transfer to a flask and dissolve in distilled water; dilute to 100 c.c. Fill an ordinary burette with the standard silver solution, and (after adjusting) run into the flask a quantity sufficient to throw down the greater part of the chlorine. Add a few drops of the chromate indicator and continue the addition of the silver nitrate until the yellow colour of the solution becomes permanently tinted red, after shaking. This shows that the chlorine is all precipitated, and that the chromate is beginning to come down. The further addition of a couple of drops of the silver solution will cause a marked difference in the tint. Read off the quantity run in, and calculate the standard. One gram of sodium chloride contains 0.6062 gram of chlorine; and 1 gram of potassium chloride contains 0.4754 gram.

For the determination of small quantities of chloride (a few milligrams), the same method is used; but the standard solution is diluted so that each c.c. is equal to 1 milligram of chlorine; and the chromate indicator is added before titrating. The standard solution is made by measuring off 200 c.c. of the solution described above, and diluting with distilled water to 1 litre.

BROMINE AND BROMIDES.

Bromine closely resembles chlorine in the nature of its compounds. It does not occur free in nature, but is occasionally found in combination with silver as bromargyrite (AgBr) and, together with chloride, in embolite. It mainly occurs as alkaline bromides in certain natural waters. Nearly all the bromine of commerce is derived from the mother liquors of salt-works—i.e., the liquors from which the common salt has been crystallised out. Bromine combines directly with the metals, forming a series of salts—the bromides. In ordinary work they are separated with, and (except when specially tested for) counted as, chlorides. They are detected by adding chlorine water to the suspected solution and shaking up with carbon bisulphide. Bromine colours the latter brown.

IODINE AND IODIDES.

Iodine does not occur in nature in the free state; and iodides are rare, iodargyrite or iodide of silver (AgI) being the only one which ranks as a mineral species. Iodates are found associated with Chili saltpetre, which is an important source of the element.

Iodine and Iodides are largely used in the laboratory, and have already been frequently referred to. It is used as an oxidising agent in a similar manner as permanganate and bichromate of potash, especially in the determinations of copper, arsenic, antimony, and manganese.

Iodine is not readily soluble in water; but dissolves easily in a concentrated solution of potassium iodide. Its solutions are strongly coloured; a drop of a dilute solution colours a large volume of water decidedly yellow; on the addition of starch paste, this becomes blue. The delicacy of this reaction is taken advantage of in titrations to determine when free iodine is present. The blue colour may be alternately developed and removed by the addition of iodine (or an oxidising agent) and hyposulphite of soda (or some other reducing agent). In decolorising, the solution changes from blue or black to colourless or pale yellow according to circumstances. Sometimes the solution, instead of remaining colourless, gradually develops a blue which recurs in spite of the further addition of the reducing agent. In these cases the conditions of the assay have been departed from, or (and this is more often the case) there is some substance present capable of liberating iodine.

Iodine forms a series of salts—the iodides—resembling in many respects the chlorides. These can be obtained by direct combination of the metals with iodine.

Detection.—Free iodine is best recognised by the violet vapours evolved from the solution on heating, and by the blue or black colour which it strikes on the addition of starch paste. Iodides are detected by boiling with strong solutions of ferric sulphate or chloride. Iodine is liberated, distilled over, and collected. Chlorine also liberates iodine from iodides; and this reaction is frequently made use of in assaying. A process based on this is described under Manganese. All substances which liberate chlorine on boiling with hydrochloric acid (dioxides, bichromates, permanganates, &c.) are determined in a similar way.

Solution and Separation.—Most iodides are soluble in water or dilute acids. The separation is effected by distilling the substance with solution of ferric sulphate, and collecting the vapour in a dilute solution of sulphurous acid or arsenite of soda. On the completion of the distillation, the iodine will be in the distillate as iodide; and the gravimetric determination is made on this.

GRAVIMETRIC DETERMINATION.

To the solution containing the iodine, as iodide, and which is free from chlorides (and bromides), add a little dilute nitric acid and nitrate of silver till no further precipitate is produced. Filter off, wash with hot water, and dry. Clean the filter-paper as much as possible, and burn it. Collect the ash in a weighed porcelain crucible, add the main portion, and heat to incipient fusion; cool, and weigh. The substance is silver iodide, and contains 54.03 per cent. of iodine.

VOLUMETRIC METHOD.

This is for the titration of free iodine, and is practically that which is described under Manganese. The substance to be determined is distilled with ferric sulphate, and the iodine is collected in a solution of potassium iodide, in which it readily dissolves. If flaky crystals separate out in the receiver, more potassium iodide crystals are added. When the distillation is finished, the receiver is disconnected, and its contents washed out into a beaker and titrated with "hypo." The standard solution of "hypo" is made by dissolving 19.58 grams of hyposulphite of soda (Na{2}S{2}O{3}.5H{2}O) in water and diluting to 1 litre; 100 c.c. are equal to 1 gram of iodine. To standardise the solution, weigh up 0.25 gram of pure iodine in a small beaker. Add 2 or 3 crystals of potassium iodide; cover with water; and, when dissolved, dilute to 50 or 100 c.c. Titrate, and calculate the standard.

FLUORINE AND FLUORIDES.

Fluorine is frequently met with as calcium fluoride or fluor-spar (CaF_{2}). It occurs less abundantly as cryolite (Na_{3}AlF_{6}), a fluoride of aluminium and sodium, which is used in glass-making. Certain other rarer fluorides are occasionally met with. Fluorine is also found in apatite, and in some silicates, such as topaz, tourmaline, micas, &c.

Hydrofluoric acid is used for etching glass and opening up silicates. It attacks silica, forming fluoride of silicon (SiF_{4}), which is volatile. Silica is by this means eliminated from other oxides, which, in the presence of sulphuric acid, are fixed. The commercial acid is seldom pure, and generally weak; and the acid itself is dangerously obnoxious. The use of ammonium fluoride (or sodium fluoride) and a mineral acid is more convenient. Determinations of this kind are made in platinum dishes enclosed in lead or copper vessels in a well-ventilated place. Fluor-spar is useful as a flux in dry assaying; it renders slags, which would otherwise be pasty, quite fluid. Fluorides generally are fusible, and impart fusibility to substances with which they form weak compounds. Their fluxing action does not depend on the removal of silicon as fluoride.

Detection.—Fluorides in small quantity are easily overlooked unless specially sought for. In larger amounts they are recognised by the property hydrofluoric acid has of etching glass. A watch-glass is warmed, and a layer of wax is melted over the convex side. When cold, some lines are engraved on the waxed surface with any sharp-pointed instrument. The substance to be tested is powdered; and moistened, in a platinum dish, with sulphuric acid. The watch-glass is filled with cold water and supported over the dish. The dish is then carefully warmed, but not sufficiently to melt the wax. After a minute or two, the glass is taken off, and the wax removed. If the substance contained fluorine, the characters will be found permanently etched on the glass. An equally good, but more rapid, test is to mix the powdered substance with some silica, and to heat the mixture in a test tube with sulphuric acid. Silicon fluoride is evolved, and, if a moistened glass rod is held in the tube, it becomes coated with a white deposit of silica, formed by the decomposition of the silicon fluoride by the water. This is also used as a test for silica; but in this case the substance is mixed with a fluoride, and the experiment must obviously be carried out in a platinum vessel.

Separation and Determination.—The determination of fluorine is difficult. In the case of fluorides free from silicates (such as fluor-spar), it is determined indirectly by decomposing a weighed portion with sulphuric acid, evaporating, igniting, and weighing the residual sulphate. The increase in weight multiplied by 0.655 gives the weight of fluorine.

In the presence of silica this method does not answer, because of the volatilisation of silicon fluoride. In these cases Whler adopted the following plan, which resembles that for the indirect determination of carbon dioxide. Mix the weighed substance in a small flask with powdered silica and sulphuric acid. The mouth of the flask is closed with a cork carrying a tube which is filled, the first half with calcium chloride and the second half with pumice coated with dried copper sulphate. The apparatus is weighed quickly, and then warmed till decomposition is complete. A current of dry air is aspirated for a minute or two; and the apparatus again weighed. The loss in weight gives that of the silicon fluoride (SiF_{4}), which, multiplied by 0.7307, gives the weight of fluorine.

Fresenius uses the same reaction; but collects and weighs the silicon fluoride. The finely powdered and dried substance is mixed with ten or fifteen times its weight of ignited and powdered silica. The mixture is introduced into a small dry flask connected on one side with a series of drying-tubes, and on the other with an empty tube (to condense any sulphuric acid). To this last is joined a drying-tube containing chloride of calcium and anhydrous copper sulphate. This is directly connected with a series of three weighed tubes in which the fluoride of silicon is collected. The last of these is joined to another drying-tube. The first weighed tube contains pumice and cotton wool, moistened with water; the second tube contains soda-lime as well as (in the upper half of the second limb) fused calcium chloride between plugs of wool; the third tube is filled half with soda-lime and half with fused calcium chloride. The distilling-flask containing the substance mixed with silica is charged with 40 or 50 c.c. of sulphuric acid, and placed on the hot plate. Alongside it is placed a similar dry flask containing a thermometer, and the temperature in this is kept at 150 or 160 C. A current of air is sent through the tubes during the operation, which takes from one to three hours for from 0.1 to 1 gram of the substance. A correction is made by deducting 0.001 gram for every hour the dried air has been passed through. The increase in weight of the three tubes gives the weight of the silicon fluoride.

Penfield uses a similar arrangement, but passes his silicon fluoride into an alcoholic solution of potassium chloride. Silica and potassium silico-fluoride are precipitated, and hydrochloric acid is set free.[100] The acid thus liberated is titrated, with a standard solution of alkali, in the alcoholic solution, and from the amount of free acid found the fluorine is calculated. The weight of hydrochloric acid (HCl) found, multiplied by 1.562, gives the weight of the fluorine. With this method of working, fewer U-tubes are required. The exit tube from the flask is bent so as to form a small V, which is kept cool in water; this is directly connected with the U-tube containing the alcoholic solution of potassium chloride. The flask with the assay is heated for about two hours, and a current of dry air is aspirated throughout the determination. Fluoride of silicon is a gas not easily condensed to a liquid: but is immediately decomposed by water or moist air.

FOOTNOTES:

[95] This will require two or three hours to thoroughly complete. It is best to powder the oxide first produced, and recalcine.

[96] No magnetic oxide was formed.

[97] For example:—

CaO + 2HCl = CaCl{2} + H{2}O.

PbO + H{2}SO{4} = PbSO{4} + H{2}O.

MgO + 2HNO{3} = Mg(NO{3}){2} + H{2}O.

Al_{2}O_{3} + 6HCl = Al_{2}Cl_{6} + 3H_{2}O.

Fe{2}O{3} + 3H{2}SO{4} = Fe{2}(SO{4}){3} + 3H{2}O.

[98] Fe{2}O{3}: 2FeO:: 0.2877: 0.2589.

[99] 100 c.c. contain 1 gram of sulphuric acid.

[100] 3SiF_{4} + 4KCl + 2H_{2}O = 2K_{2}SiF_{6} + SiO_{2} + 4HCl.



CHAPTER XVI.

SULPHUR AND SULPHATES.

Sulphur occurs native in volcanic districts, and is mined in Sicily, Italy, and California in considerable quantities. Combined with metals (sulphides), it is common in all mineral districts. Iron pyrites (FeS_{2}) is the most abundant source of this element. Sulphates, such as gypsum, are fairly common, but have no value so far as the sulphur in them is concerned. In coal it exists as an impurity, occurring partly as a constituent of organic compounds.

Sulphur, whether free or combined with metals, forms, on burning, sulphurous oxide (SO_{2}), which by the action of oxidising agents and water is converted into sulphuric acid. It forms two oxides, sulphurous (SO_{2}) and sulphuric (SO_{3}), which combine with bases to form sulphites and sulphates. Sulphites are of little importance to the assayer, and are converted into sulphates by the action of nitric acid and other oxidising agents.

The native sulphides, when acted on with hydrochloric acid, give off sulphuretted hydrogen; with nitric acid or aqua regia, sulphates are formed, and more or less sulphur separated.

Sulphur is detected in sulphides by the irritating odour of sulphurous oxide given off on roasting, by the evolution of sulphuretted hydrogen when treated with hydrochloric acid, or by a white precipitate of barium sulphate formed when barium chloride is added to the aqua regia solution.

Dry Assay.—There is no method of general application. Free or native sulphur may be volatilised, condensed, and weighed, but pyrites only gives up a portion of its sulphur when heated in a closed vessel, while most sulphides, and all sulphates, give up none at all.

In the determination of sulphur in brimstone, 10 grams of the substance are taken, placed in a small porcelain dish, heated over a Bunsen burner in a well-ventilated place, and ignited. When the sulphur has been completely burnt off, the residue (which consists chiefly of sand) is collected and weighed. In a separate portion the moisture and arsenic are determined; the amounts of these are deducted from the loss in the first experiment. The difference, multiplied by 10, gives the percentage of sulphur.

WET METHODS.

Solution.—All sulphates, excepting those of lead, barium, strontium, and lime, are soluble in water or dilute acid. All sulphides, except cinnabar, are converted into sulphates by the action of nitric acid at a gentle heat; or, better, by the action of a mixture of three volumes of nitric acid and one volume of hydrochloric acid. This last attacks cinnabar as well. With most substances it is difficult to convert the whole of the sulphur into sulphuric acid. The sulphur separates out at first as a dark spongy mass, which (on continued treatment) changes to light-coloured flakes. When the solution becomes concentrated and the temperature rises sufficiently, the sulphur fuses into one or more honey-coloured globules which, owing to the small surface they oppose to the acid, are very slowly oxidised. It is not desirable to assist the formation of these globules; therefore, the temperature is kept as low as possible, and strong nitric acid is used. When such globules form, it is best to allow the solution to cool, when the globules will solidify. They can then be filtered off and picked out from the insoluble residue, dried, weighed, ignited, and again weighed, the loss being counted as sulphur. With iron pyrites this difficulty seldom occurs.

Metallic sulphides when fused with an excess of nitre are completely oxidised. If the ore is rich in sulphur, some inert body (such as sodium chloride, or, better, sodium carbonate) is added to dilute the action. With pure sulphur, the action is so energetic as to cause an explosion, so that care should be taken. With burnt ores (incompletely calcined pyrites), there is sufficient oxide of iron present to prevent too rapid action.

These fusions with nitre are best conducted in a platinum dish covered with a piece of platinum foil. The ore is ground with the nitre to ensure complete mixing. The heat need not be excessive, so that a single Bunsen burner placed beneath the dish will suffice; if the bottom of the dish is seen to be red-hot, it is sufficient. On cooling and extracting with water, the sulphur will pass into solution as potassium sulphate, which is then filtered off from the insoluble oxides of iron, copper, &c. The filtrate, after having been treated with a large excess of hydrochloric acid, evaporated to dryness, and re-dissolved in water, is ready for the determination.

Lead sulphate may be dissolved by boiling with ammonium acetate. The insoluble sulphates of barium, strontium, and lime, are decomposed by fusing with 4 or 5 times their weight of "fusion mixture." The alkaline sulphates are then dissolved out with water, and filtered off from the insoluble residue. The filtrate is rendered acid with hydrochloric acid.

Separation.—The determination of the sulphuric acid in these solutions by precipitation with barium chloride also serves as a separation; but in hot acid solutions containing copper, and more especially iron salts, the baric sulphate has a strong tendency to carry down amounts of those bodies, varying, no doubt, with the conditions of the precipitation. Boiling hydrochloric acid fails to completely extract them. Moreover, the use of hot concentrated hydrochloric acid causes a loss by dissolving barium sulphate. Nitric acid and nitrates must be decomposed by prolonged boiling and evaporation with hydrochloric acid. The iron may be removed by adding a slight excess of ammonia to the faintly acid solution, filtering off, and washing the precipitated ferric hydrate with hot water. By slightly acidulating the filtrate with hydrochloric acid, it will be rendered ready for the determination.

GRAVIMETRIC METHOD.

This assay is one of those which strikingly shows the necessity of getting the assay solution under proper conditions, in order to obtain satisfactory results. The method has been repeatedly investigated, and the conclusion arrived at, "that it can be correct only by accident." Yet there are many chemists who get good results, and place considerable faith in its accuracy. This can only be due to differences in the manner of working. It is generally understood that nitric acid or nitrates must be absent; and our experience fully confirms this. Precipitations in nitrate solutions are worthless, as the following experiments show. In each experiment the bulk of the solution was 150 c.c. The solutions contained 10 grams of nitre, were freely acid with hydrochloric acid, and were precipitated (while boiling) with slight excess of baric chloride.

Sulphuric acid taken 0.020 gram 0.050 gram 0.100 gram " found 0.019 " 0.047 " 0.098 " " taken 0.500 " 1.004 " 1.000 " " found 0.526 " 1.126 " 1.126 "

All the precipitates were boiled with hydrochloric acid, and thoroughly washed before weighing. The results of some other experiments on this subject are given under "sulphur" in the "examination of commercial copper," page 207.

The solution having been obtained free from nitrates and chlorates (and containing but little free hydrochloric acid), is largely diluted, heated to boiling, and precipitated with a moderate excess of a solution of chloride of barium (8 parts of the crystallized barium chloride are sufficient for 1 of sulphur). It is allowed to settle for half-an-hour, and then decanted through a filter. The precipitate is shaken up with boiling water, rendered slightly acid, filtered, washed, dried, ignited, and weighed. The ignited precipitate, when pure, is white, and is not decomposed at a red heat; it is barium sulphate (BaSO{4}), and contains 13.73 per cent. of sulphur, or 34.33 per cent. of sulphuric oxide (SO{3}).

Determination of Sulphur in Pyrites.—Weigh up half a gram of the dried and powdered sample, and treat with 10 c.c. of a mixture of 3 volumes of nitric acid and 1 volume of hydrochloric acid, occasionally heating. Evaporate to dryness, treat with 5 c.c. of hydrochloric acid, and again evaporate; take up with 1 c.c. of hydrochloric acid and 100 c.c. of hot water, filter through a small filter, and wash. The residue may contain sulphates of lead, barium, or lime; it must be separately examined, if the total sulphur is wanted. The filtrate is heated, and rendered slightly alkaline with ammonia. Filter off the precipitated ferric hydrate through a quick filter, and wash with hot water. If necessary, evaporate the bulk to about 200 c.c., render faintly acid with hydrochloric acid, and add 20 c.c. of solution of barium chloride; allow to stand for half-an-hour, and decant through a filter. Wash with hot water, dry, ignite, and weigh. Pure pyrites contains 53.33 per cent. of sulphur.

VOLUMETRIC METHOD.

This is based upon the easy conversion of all sulphur compounds into sulphates by fusion with nitre or by oxidation with nitric acid; and on the determination of the sulphate formed by titration in an acetic acid solution with baric chloride.[101] The finishing point is determined by filtering off portions of the assay solution, and testing with sulphuric acid. A slight excess of baric chloride will cause a precipitate.

The process may be divided into—(1) the preparation of the solution, and (2) the titration.

Preparation of the Solution.—Weigh up from 1 to 5 grams of the dried and powdered substance, and mix intimately with 4 grams of powdered nitre; clean out the mortar with another gram of nitre, and add this as a cover. Heat in a platinum crucible for fifteen minutes at a low temperature; cool, and extract with water in an evaporating-dish about 9 inches across, and holding 700 or 800 c.c. Add 10 grams of sodium acetate and 10 c.c. of acetic acid, and dilute to half a litre. Boil. The solution is ready for titrating. Substances which lose sulphur on heating (such as pyrites) are thus treated:—Weigh up 1 gram, and evaporate nearly to dryness with 10 c.c. each of nitric and hydrochloric acids. Take up with 10 c.c. of hydrochloric acid, and again boil down to a small bulk; dilute and transfer to a 9-inch evaporating-dish; add 10 grams of sodium acetate and 5 c.c. of acetic acid, dilute to half a litre, and boil. The solution is ready for titrating. Sulphates may be dissolved up in the dish itself with the help of a c.c. or so of hydrochloric acid; sodium acetate and acetic acid are then added; and, after dilution and boiling, the solutions are at once titrated.

The solution before titration must contain no free mineral acid, but 5 or 10 c.c. of acetic acid should be present. It must contain 10 grams of sodium acetate, or sufficient to convert any free mineral acid into its corresponding sodic salt; or, if chlorides, nitrates or sulphates of the metals are present, sufficient to decompose them. If a precipitation occurs, as is the case with ferric salts, &c., the solution is titrated with the precipitate in it.

The Titration.—The standard solution of barium chloride is made by dissolving 76.25 grams of the crystallized salt (BaCl{2}.2H{2}O) in distilled water, and diluting to 1 litre. 100 c.c. will equal 1 gram of sulphur. As indicator, use dilute sulphuric acid. The strength of the solution may be checked by the titration of 5 grams of ferrous sulphate (oxidized with permanganate of potassium or a few drops of nitric acid), which should require 57.5 c.c. of the barium chloride solution; or any pure sulphate of known composition can be used; anhydrous salts should be preferred.



Fill an ordinary 100 c.c. burette with the solution of barium chloride. The evaporating dish containing the assay solution is placed on a round burner (as shown in fig. 65), and the solution is kept steadily boiling. An ordinary Bunsen-burner flame will cause bumping, and should not be used. Run in the standard solution in quantity known to be insufficient; then withdraw a portion of about 2 c.c., with a pipette, and filter through a fine filter-paper into a test tube. Run in another 0.5 c.c. of the standard solution, and withdraw and filter into a test tube another portion of 2 c.c.; and continue this operation until half-a-dozen or more portions have been drawn off. The test tubes should be arranged in order in a stand resting on a piece of paper, so that each test tube representing 0.5 c.c. of the standard baric chloride may have its value recorded beneath it (fig. 66). Add to each test tube 3 drops of dilute sulphuric acid; that which shows the first appearance of a precipitate marks the point at which the titration is complete. Suppose, for example, that the test tube marked 48.5 c.c. shows no precipitate, while that at 49.0 c.c. shows one, it is evident that the finishing point lies between these readings. With a little practice, one can judge from the appearance of the precipitate in the 49 c.c. tube, whether 1/4 c.c. should be deducted or not.



It is better to add dilute sulphuric acid, and to watch for the appearance of a precipitate in the test tube, than to add baric chloride and to look for its non-appearance; besides, baric chloride is much less likely to be present in a test tube as impurity than sulphates are. In this way the chance of error from what are termed "accidental causes" is diminished.

The following experiments show the effect of variation in the conditions of titration:—

Make _a standard solution of sulphuric acid_ by diluting 43.65 grams of sulphuric acid (sp. g. 1.6165) to 1 litre: 100 c.c. will contain 1 gram of sulphur. An equivalent solution may be made by dissolving 100.62 grams of sodium sulphate crystals (Na_{2}SO_{4}.10H_{2}O), or 86.88 grams of ferrous sulphate (FeSO_{4}.7H_{2}O), in water (oxidising the latter), and diluting to 1 litre.

The order in which these experiments are given is that in which they were made in an investigation into the conditions under which the titration could most accurately be effected.

Effect of Hydrochloric and Nitric Acids.—The titrations were performed in the manner already described, but sodic acetate and acetic acid were absent. Twenty c.c. of the standard solution of sulphuric acid were used.

Hydrochloric acid present 0.0 c.c. 1.0 c.c. 2.0 c.c. 5.0 c.c. "Baric chloride" required 20.0 " 20.0 " 19.7 " 12.5 "

Nitric acid present 0.0 c.c. 1.0 c.c. 2.0 c.c. 5.0 c.c. "Baric chloride" required 20.0 " 19.5 " 18.0 " 10.0 "

These show clearly the interference of free mineral acids, although very dilute hydrochloric acid (1 c.c. in 500 of water) has no effect.

Effect of Acetic and Citric Acids.—A similar series of experiments with these acids gave the following results:—

Acetic acid present 0.0 c.c. 5.0 c.c. 50.0 c.c. 100.0 c.c. "Baric chloride" required 20.0 " 20.0 " 20.0 " 20.0 "

Citric acid present 0 gram 1 gram 5 grams "Baric chloride" required 20.0 c.c. 20.0 c.c. 20.0 c.c.

These acids do not interfere.

Effect of Sodic Acetate and Acetic Acid.—In each of these experiments 5 c.c. of acetic acid was present.

Sodium acetate added 0 gram 1 gram 10 grams 50 grams "Baric chloride" required 20.0 c.c. 20.0 c.c. 20.0 c.c. 20.0 c.c.

As sodic acetate and acetic acid did not interfere, it became desirable to make some experiments on the finishing point. The first object sought for was the smallest amount of the standard baric chloride in 500 c.c. of water, required to give an indication when tested in the manner already described.

Baric Chloride Conditions of Assay Solution. required.

Water only 0.05 c.c. With 10 grams of sodium acetate and 5 c.c. of acetic acid 0.05 " The same with 5 grams of nitre 0.10 " Like the last, but with 5 grams of salt instead of nitre 0.10 "

These show that as small an amount of baric chloride solution as is equal to only 0.000002 gram of sulphur in the 2 c.c. of solution tested yields a decided precipitate on the addition of 3 drops of sulphuric acid.

To determine whether the same finishing point is obtained on testing the filtered portions in the test tubes with baric chloride as is obtained on testing with sulphuric acid, a titration was made with 20 c.c. of standard solution of sulphuric acid, together with the usual quantities of sodic acetate and acetic acid; and two lots of 2 c.c. each were filtered into two sets of test tubes after each addition of the standard baric chloride. To one series 3 drops of baric chloride solution were added, and to the other 3 drops of sulphuric acid. The results were—

With Dilute With Baric "Baric Chloride" added. Sulphuric Acid. Chloride Solution.

19.5 c.c. Clear Cloudy 19.75 " Clear Cloudy 20.0 " Finished Finished 20.25 " Cloudy Clear 20.5 " Cloudy Clear

The two methods of testing give the same result. But this balance is disturbed in the presence of much nitre, the indications with baric chloride being disturbed by an opalescence for some c.c. beyond the finishing point. In solutions containing free hydrochloric or nitric acid, a precipitate is obtained with either baric chloride or sulphuric acid.

Effect of Varying Sulphur.—In these and the subsequent experiments the titrations were performed in the presence of 10 grams of sodic acetate and 10 c.c. of acetic acid in the manner already described.

Standard sulphuric acid used 5.0 c.c. 10.0 c.c. 20.0 c.c. 50.0 c.c. 100.0 c.c.

"Baric chloride" required 5.0 " 10.0 " 20.0 " 50.0 " 100.0 "

Effect of Varying Temperature.—With 5 c.c. of standard sulphuric acid titrated at 15 C., 5 c.c. of baric chloride were required; but with larger quantities the results were altogether unsatisfactory when titrated cold.

Effect of Varying Bulk.—

Bulk 100.0 c.c. 200.0 c.c. 500.0 c.c. 1000.0 c.c. "Baric chloride" required 20.0 " 20.0 " 20.0 " 20.5 "

Considerable variation in bulk has no effect, but 500 c.c. is the most convenient volume to work with. It is well to occasionally replace the water boiled off during titration.

Effect of Foreign Salts.—In all these experiments 20 c.c. of "sulphuric acid" were used, and the titration was performed in the ordinary way.

Sodic chloride added 0 gram 5 grams 10 grams "Baric chloride" required 20.0 c.c. 20.0 c.c. 19.7 c.c.

Ammonic chloride added 0 gram 5 grams 10 grams "Baric chloride" required 20.0 c.c. 20.0 c.c. 19.5 c.c.

Calcic chloride added 0 gram 1 gram 2 grams 5 grams "Baric chloride" required 20.0 c.c. 20.0 c.c. 19.2 c.c. 19.0 c.c.

Zinc chloride added 0 gram 1 gram 3 grams 5 grams "Baric chloride" required 20.0 c.c. 20.0 c.c. 20.0 c.c. 20.0 c.c.

Ferrous chloride added 0 gram 1 gram 3 grams 5 grams "Baric chloride" required 20.0 c.c. 19.7 c.c. 19.5 c.c. 19.0 c.c.

Ferric chloride added 0 gram 1 gram 3 grams 5 grams "Baric chloride" required 20.0 c.c. 20.0 c.c. 20.0 c.c. 20.0 c.c.

Copper chloride added 0 gram 1 gram 3 grams 5 grams "Baric chloride" required 20.0 c.c. 20.0 c.c. 20.0 c.c. 20.0 c.c.

Potassic Nitrate added 0 gram 1 gram 5 grams 10 grams "Baric chloride" required 20.0 c.c. 20.0 c.c. 20.0 c.c. 19.0 c.c.

Potassic Nitrite added 0 gram 1 gram 5 grams "Baric chloride" required 20.0 c.c. 20.0 c.c. 20.0 c.c.

Sodic phosphate added 0 gram 1 gram "Baric chloride" required 20.0 c.c. 22.5 c.c.

Sodic arsenate added 0 gram 1 gram "Baric chloride" required 20.0 c.c. 20.5 c.c.

In the absence of ferric salts, phosphates and arsenates count as sulphur.

In two series of experiments for determining the effect of varying amounts of sulphur in the form of ferrous sulphate, we obtained the following results:—In the first series the assay solution was prepared in the manner we have described for Pyrites; and in the second series, by fusion with nitre.

Sulphur added 0.050 gram 0.100 gram 0.200 gram "Baric chloride" required (1) 5.0 c.c. 10.0 c.c. 20.0 c.c. " " (2) 4.7 " 10.0 " 20.0 "

Sulphur added 0.500 gram 1.000 gram "Baric chloride" required (1) 50.0 c.c. 100.0 c.c. " " (2) 50.0 " 100.0 "

More than 5 grams of nitre must not be used in an assay; and, since the requisite amount of nitre considerably exceeds that sufficient to oxidise the sulphur, not more than 0.5 gram of unoxidised sulphur should be present in the portion of the sample weighed up for determination. When the amount of sulphur present is not known within reasonable limits, the test portions may be tried with a drop of baric chloride solution instead of sulphuric acid, so that the diminishing quantity of precipitate may give warning of an approach to the finishing point.

Determination of Sulphur in Blende.—Weigh up 1 gram of dried and powdered blende, and mix and fuse with 5 grams of nitre in the manner described. Place the dish and its contents in the titrating-dish, extract with water, add 10 grams of sodium acetate and 10 c.c. of acetic acid, remove and wash the platinum-dish, and dilute to 500 c.c.; boil and titrate. In the example, duplicate determinations required (a) 32.0 c.c., (b) 32.25 c.c., giving an average of 32.1 per cent. of sulphur.

Determination of Sulphur in Chalcopyrite (Yellow Copper Ore).—Take 1 gram of the finely-powdered sample, and 5 grams of nitre. Sprinkle a little of the nitre in a small Wedgwood mortar, place the ore on it, and cover with 2 or 3 grams more of the nitre. Rub up together, and transfer to a small porcelain dish; clean out the mortar with the rest of the nitre, and add to the contents of the dish. Cover with a piece of platinum foil, and heat gently with a Bunsen burner till the nitre melts and the stuff shows signs of deflagrating; remove the heat, and allow the action to go on by itself for a minute or so, then heat over the Bunsen burner for 10 minutes. Cool; transfer the whole to the titrating-dish; boil with 500 c.c. of water; remove the small dish and foil; add sodic acetate and acetic acid, and titrate.

For example, 1 gram required 34.5 c.c. of "barium chloride" (standard = 1.005 gram S), which is equivalent to 34.7 per cent. sulphur. The theoretical percentage is 34.8.

Determination of Sulphur in Chalcocite (Grey Copper Ore).—Proceed as in the last experiment but, since the action with nitre is more moderate, no special precautions need be taken on heating. A platinum dish may be used.

An example which was heated for 30 minutes required 20.5 c.c. of the barium chloride solution. This is equivalent to 20.6 per cent. of sulphur. The theoretical yield is 20.2 per cent.

Determination of Sulphur in Pyrites.—Take 1 gram of the finely-powdered sample, cover with 10 c.c. of nitric acid, and, when action has ceased, evaporate to a small bulk. Add 3 or 4 c.c. of hydrochloric acid, and again evaporate to a paste. Take up with 1 or 2 c.c. of dilute hydrochloric acid, dilute with water, transfer to a titrating-dish, add 10 grams of sodic acetate and 5 c.c. of acetic acid, and dilute with water to 500 c.c. Boil and titrate.

An example with 1 gram of a pure crystallized pyrites required 52.7 c.c. of the barium chloride solution, which is equivalent to 53.0 per cent. of sulphur. Theory requires 53.3 per cent. of sulphur.

Determination of Sulphur in Mispickel.—Take 1 gram of the powdered ore and evaporate with 10 c.c. of nitric acid, and take up with 3 or 4 c.c. of hydrochloric acid. If any globules of sulphur remain, again evaporate with nitric acid. Dilute, and transfer to the titrating-dish. Add 10 grams of sodic acetate, dilute with water, boil, and titrate. The mispickel carries (according to theory) exactly sufficient iron to precipitate the arsenic as ferric arsenate in an acetic acid solution, so no more iron need be added. The ferric arsenate will separate out as a yellowish-white flocculent precipitate.

An example required, in duplicate experiment, 18.5 c.c. and 18.7 c.c. of barium chloride, equivalent to 18.7 per cent. of sulphur. The formula, FeS{2}.FeAs{2}, requires 19.6 percent., but the sulphur generally varies considerably from this amount.

Determination of Sulphur in Burnt Ores.—Take 5 grams of the dried and powdered ore, and rub up with 4 grams of nitre; transfer to the platinum-dish; clean out the mortar with another gram of nitre, and add this as a cover. Heat, and extract with water as before; add the sodium acetate and acetic acid; and titrate. Burnt ores carry from 2.5 to 5 per cent. of sulphur. A series of four determinations gave:—

"Baric Chloride" Required. Percentage of Sulphur. Gravimetric Results. 12.6 c.c. 2.52 % 2.45 % 29.9 " 5.98 " 5.84 " 18.1 " 3.62 " 3.53 " 22.0 " 4.40 " 4.43 "

For ores carrying less than 1 per cent. of sulphur, take 10 grams for the assay.

Determination of Sulphuric Oxide (SO{3}) in Sulphates.—When the sulphur exists in the sample received by the assayer in an oxidised state as sulphate, it is usual to report it in terms of sulphuric oxide (SO{3}). In this case, the metal must also be reported as oxide. For example, an analysis of copper sulphate would be thus reported:—

Oxide of copper (CuO) 31.8 % Sulphuric oxide (SO_{3}) 32.1 " Water 36.1 " ——- 100.0

The percentage of sulphur multiplied by 2.5 gives the percentage of sulphuric oxide. Thus a sample of copper sulphate containing 12.85 per cent. of sulphur will contain 12.85 2.5 or 32.12 per cent. of sulphuric oxide.

In minerals and metallurgical products, it is common to find the sulphur in both conditions—i.e., as sulphate and sulphide. Generally in these the percentage of sulphur only is wanted; but this will depend entirely on commercial requirements, and not on the fancy of the assayer. Soluble sulphates are determined separately by extracting with small quantities of cold water, so as to avoid the separation of basic sulphates, or, if the sulphides present are not at the same time attacked, by dilute hydrochloric acid. Lead sulphate may be extracted by boiling with ammonic acetate; whilst barium, strontium, and, perhaps, calcium sulphate, will be mainly found in the residue insoluble in acids.

Weigh up from 2 to 5 grams of the material according to the amount of sulphur judged to be present, and dissolve them in the titrating-dish with 1 c.c. of hydrochloric acid and 50 c.c. of water. Add 10 grams of sodic acetate, and 10 c.c. of acetic acid; dilute, boil, and titrate. In the case of ferric salts, half the quantity of acetic acid will be better, as then the ferric iron will be precipitated, and a colourless solution will be left, in which the end reaction is more readily distinguished.

Determined in this way, 5 gram samples of the following salts gave the results indicated below:—

"Barium Chloride" Salt. Required. Sulphuric Oxide. Copper sulphate 64.25 c.c. 32.12 % Magnesium sulphate 65.25 " 32.62 " Zinc sulphate 56.25 " 28.12 " Ferrous sulphate 58.25 " 29.12 " Sodium sulphate 51.25 " 25.60 "

Determination of Sulphuric Oxide in Barytes (Heavy spar).—Fuse 2 grams of the powdered mineral with 5 grams of "fusion mixture" for five minutes; and, when cold, extract with water. Filter, acidulate the filtrate with an excess of 10 c.c. of acetic acid, dilute, boil, and titrate. For example, a transparent crystallised sample required 27.0 c.c. of barium chloride, which is equivalent to 13.6 per cent. of sulphur, or 34.0 per cent. of sulphuric oxide. Theory requires 34.3 per cent. of the latter. Since both carbonate of soda and potash are liable to contain sulphates, a blank determination should be made on 5 grams of the "fusion mixture," and the amount found be deducted from that got in the assay.

PRACTICAL EXERCISES.

1. The price of sulphur in an ore being 4-1/2d. per unit in the northern markets, what would be the price of a ton of ore containing 49 per cent. of sulphur? What would be the effect on the price of an error of 0.25 per cent. in the assay?

2. Pyrites carries 50 per cent. of sulphur, and on calcining yields 70 per cent. of its weight of burnt ore. Supposing the burnt ore carries 3.5 per cent. of sulphur, what proportion of the sulphur will have been removed in the calcining?

3. How would blende compare with pyrites as a source of sulphur for sulphuric acid making?

4. How would you determine the percentage of sulphuric oxide in a sample of gypsum? What is sulphuric oxide, and what relation does it bear to sulphur?

5. A mineral contains 20.7 per cent. of water, 32.4 per cent. of lime, and 18.6 per cent. of sulphur. What is its probable composition? What experiment would you try to determine the accuracy of your conclusion?

SELENIUM

occurs in nature combined with copper, mercury, and lead, in certain rare minerals. In small quantities it is found in many ores. It is detected in solution by the red precipitate produced on boiling the acid solution with sodium sulphite. This reaction is used for its determination.

Solution.—The solution is effected by boiling with nitric acid or aqua regia, or by fusing with nitre. To separate the selenium, the solution is evaporated with an excess of hydrochloric acid and a little sodium or potassium chloride. This destroys any nitric acid that may be present, and reduces selenic acid (H{2}SeO{4}) to selenious (H{2}SeO{3}). The solution is diluted with water, and treated with a solution of sulphite of soda. It is warmed, and at last boiled. The selenium separates as a red precipitate, which (on boiling) becomes denser and black. It is collected on a weighed filter, washed with hot water, dried at 100 C., and weighed as pure selenium.

Selenium can be precipitated with sulphuretted hydrogen as a sulphide, which is readily soluble in ammonium sulphide. This sulphide may be oxidised with hydrochloric acid and chlorate of potash; and the selenium separated in the manner described.

TELLURIUM.

Tellurium occurs in nature, native, and in combination with gold, silver, bismuth and lead. It is sometimes met with in assaying gold ores. It may be detected by the purple colour it imparts to strong sulphuric acid when dissolved in the cold, and by the black precipitate of metallic tellurium which its solutions yield on treatment with a reducing agent. Telluric acid is reduced to tellurous (with evolution of chlorine) on boiling with hydrochloric acid.

Solution is effected by boiling with aqua regia, or by fusing with nitre and sodium carbonate.

Separation.—Tellurium closely resembles selenium in its reactions. It is separated and determined in the same way. Like it, it forms a sulphide soluble in ammonium sulphide. It is distinguished from selenium by the insolubility, in a solution of cyanide of potassium, of the metal precipitated by sodium sulphite; whereas selenium dissolves, forming a soluble potassic seleno-cyanide.[102]

For the determination, solution is effected by fusing with nitre and sodium carbonate, dissolving out the tellurate of potash with water, and boiling with hydrochloric acid. Tellurous compounds are formed, with evolution of chlorine; and the solution, on treating with a reducing agent (such as sulphurous acid or stannous chloride), yields metallic tellurium; which is washed, dried at 100 C., and weighed.

FOOTNOTES:

[101] BaCl{2} + Na{2}SO{4} = BaSO{4} + 2NaCl.

[102] Se + KCy = KCySe.



CHAPTER XVII.

ARSENIC, PHOSPHORUS, NITROGEN.

ARSENIC.

The chief source of the arsenic of commerce is arsenical pyrites, or mispickel, which contains about 45 per cent. of arsenic (As). Arsenic also occurs as a constituent of several comparatively rare minerals; and, as an impurity, it is very widely distributed. White arsenic is an oxide of arsenic, and is obtained by roasting arsenical ores, and refining the material (crude arsenic), which condenses in the flues. Arsenic itself is volatile, and many of its compounds have the same property. It forms two well-defined series of salts, corresponding to the oxides: arsenious oxide (As{2}O{3}), and arsenic oxide (As{2}O{5}). These combine with bases to form arsenites and arsenates respectively. Boiling with nitric acid converts the lower into the higher oxide; and powerful reducing-agents, such as cuprous chloride, have the opposite effect.

Arsenic may be detected by dissolving the substance in hydrochloric acid, or in aqua regia (avoiding an excess of nitric acid), and adding a little of this solution to the contents of a small flask in which hydrogen is being made by the action of zinc and hydrochloric acid. The ignited jet of hydrogen assumes a blue colour if arsenic is present, and a cold porcelain dish held in the flame (fig. 67) becomes coated with a dark deposit of metallic arsenic. Antimony produces a similar effect, but is distinguished by the insolubility of its deposit in a cold solution of bleaching-powder.



Arsenites are distinguished by the volatility of the chloride; by decolorising a solution of permanganate of potassium, and by immediately giving a yellow precipitate with sulphuretted hydrogen. Arsenates are distinguished (after converting into soda salts by boiling with carbonate of soda and neutralising) by giving with nitrate of silver a red precipitate, and with "magnesia mixture" a white crystalline one.

Dry Assay.—There is no dry assay which is trustworthy. The following method is sometimes used to find the proportion of arsenious oxide in "crude arsenic":—Weigh up 5 grams of the dried sample, and place them in a clean dry test-tube about 6 inches long. Tie a small filter-paper over the mouth of the tube, so as to prevent air-currents. Heat the tube cautiously so as to sublime off the white arsenic into the upper part of the tube. Cut off the bottom of the test-tube by wetting whilst hot. Scrape out the arsenic and weigh it. The weight gives an approximate idea of the quantity, and the colour of the quality, of the white arsenic obtainable from the sample. Some workers (sellers) weigh the residue, and determine the white arsenic by difference. In determining the percentage of moisture in these samples, the substance is dried on a water-bath or in a water-oven.

WET METHODS.

Solution.—Where, as in crude arsenic, the substance is arsenious oxide (As{2}O{3}) mixed with impurities, the arsenic is best got into solution by warming with caustic soda, and neutralising the excess with hydrochloric acid; it will be present as sodium arsenite. Metals and alloys are acted on by means of nitric acid; or the arsenic may be at the same time dissolved and separated by distilling with a strongly-acid solution of ferric chloride, in the way described under Volumetric Methods.

With minerals, mattes, &c., solution is thus effected:—The finely-powdered substance is mixed (in a large platinum or porcelain crucible) with from six to ten times its weight of a mixture of equal parts of carbonate of soda and nitre. The mass is then heated gradually to fusion, and kept for a few minutes in that state. When cold, it is extracted with warm water, and filtered from the insoluble residue. The solution, acidified with nitric acid and boiled, contains the arsenic as sodium arsenate. With mispickel, and those substances which easily give off arsenic on heating, the substance is first treated with nitric acid, evaporated to dryness, and then the residue is treated in the way just described.

When the arsenic is present as arsenite or arsenide, distillation with an acid solution of ferric chloride will give the whole of the arsenic in the distillate free from any metal except, perhaps, tin as stannic chloride. With arsenates, dissolve the substance in acid and then add an excess of soda. Pass sulphuretted hydrogen into the solution; warm, and filter. Acidulate the filtrate, and pass sulphuretted hydrogen. Decant off the liquid through a filter, and digest the precipitate with ammonic carbonate; filter, and re-precipitate with hydrochloric acid and sulphuretted hydrogen. Allow to stand in a warm place, and filter off the yellow sulphide of arsenic. Wash it into a beaker, clean the filter-paper (if necessary) with a drop or two of dilute ammonia; evaporate with 10 c.c. of dilute nitric acid to a small bulk; dilute; and filter off the globules of sulphur. The filtrate contains the arsenic as arsenic acid.

GRAVIMETRIC METHOD.

Having got the arsenic into solution as arsenic acid, and in a volume not much exceeding 50 c.c., add about 20 c.c. of dilute ammonia and 20 c.c. of "magnesia mixture." Stir with a glass rod, and allow to settle overnight. Filter, and wash with dilute ammonia, avoiding the use of large quantities of wash water. Dry, transfer the precipitate to a Berlin crucible, and clean the filter-paper thoroughly. Burn this paper carefully and completely; and add the ash to the contents of the crucible, together with 4 or 5 drops of nitric acid. Evaporate with a Bunsen burner, and slowly ignite, finishing off with the blow-pipe or muffle. Cool, and weigh. The ignited precipitate is pyrarsenate of magnesia (Mg_{2}As_{2}O_{7}), and contains 48.4 per cent. of arsenic (As).

Instead of igniting the precipitate with nitric acid, it may be collected on a weighed filter-paper, dried at 100 C., and weighed as ammonic-magnesic arsenate (2AmMgAsO{4}.H{2}O), which contains 39.5 per cent. of arsenic. The results in this case are likely to be a little higher. The drying is very tedious, and is likely to leave behind more water than is allowed for in the formula. In a series of determinations in which the arsenic was weighed in both forms, the results were:—

Ammonic-magnesic Arsenic Magnesium Pyrarsenate Arsenic Arsenate in grams. in grams. in grams. in grams. 0.0080 0.0032 0.0065 0.0031 0.0400 0.0158 0.0330 0.0160 0.0799 0.0316 0.0633 0.0306 0.1600 0.0632 0.1287 0.0623 0.4000 0.1580 0.3205 0.1551 0.7990 0.3156 0.6435 0.3114

VOLUMETRIC METHODS.

There are two methods: one for determining the arsenic in the lower, and the other in the higher state of oxidation. In the first-mentioned method this is done by titrating with a standard solution of iodine; and in the latter with a solution of uranium acetate. Where the arsenic already exists as arsenious oxide, or where it is most conveniently separated by distillation as arsenious chloride, the iodine method should be used; but when the arsenic is separated as ammonic-magnesic arsenate or as sulphide, the uranium acetate titration should be adopted.

IODINE PROCESS.

This is based on the fact that sodium arsenite in a solution containing an excess of bicarbonate of soda is indirectly oxidised by iodine to sodium arsenate,[103] and that an excess of iodine may be recognised by the blue colour it strikes with starch. The process is divided into two parts—(1) the preparation of the solution, and (2) the titration.

Preparation of the Solution.—For substances like crude arsenic, in which the arsenic is present as arsenious oxide, the method is as follows:—Take a portion which shall contain from 0.25 to 0.5 gram of the oxide, place in a beaker, and cover with 10 c.c. of sodic hydrate solution; warm till dissolved, put a small piece of litmus paper in the solution, and render acid with dilute hydrochloric acid. Add 2 grams of bicarbonate of soda in solution, filter (if necessary), and dilute to 100 c.c. The solution is now ready for titrating.



Where the arsenic has to be separated as arsenious chloride, the process is as follows:[104]—Weigh up 1 gram of the finely-powdered ore (metals should be hammered out into a thin foil or be used as filings), and place in a 16-ounce flask provided with a well-fitting cork, and connected with a U-tube, as shown in the drawing (fig. 68). The U-tube should contain 2 or 3 c.c. of water, and is cooled by being placed in a jar or large beaker of cold water. The water used for cooling should be renewed for each assay.

Pour on the assay in the flask 50 c.c. of a "ferric chloride mixture," made by dissolving 600 grams of calcium chloride and 300 grams of ferric chloride in 600 c.c. of hydrochloric acid, and making up to 1 litre with water.

Firmly cork up the apparatus, and boil over a small Bunsen-burner flame for fifteen or twenty minutes, but avoid evaporating to dryness. Disconnect the flask, and pour away its contents at once to prevent breakage of the flask by their solidification. The arsenic will be condensed in the U-tube, together with the greater part of the hydrochloric acid; transfer the distillate to a beaker washing out the tube two or three times with water; add a small piece of litmus paper; neutralise with ammonia; render faintly acid with dilute hydrochloric acid; add 2 grams of bicarbonate of soda in solution; and dilute to 250 c.c. The solution is now ready for titrating.

The arsenic comes over in the early part of the distillation, as will be seen from the following experiment, made on 1 gram of copper precipitate; in which experiment the distillate was collected in separate portions at equal intervals, and the arsenic in each portion determined:—

Time Iodine Equivalent to Arsenic Distilling. Required. in the Distillate.

5 minutes 12.0 c.c. 0.0450 gram 5 " 0.17 " 0.0005 " 5 " 0.0 " 5 " 0.0 " To dryness 0.0 "

The volume of each distillate was about 5 c.c.

In this operation the metals are converted into chlorides by the action of ferric chloride, which gives up a part of its chlorine, and becomes reduced to the ferrous salt. The calcium chloride does not enter into the chemical reaction, but raises the temperature at which the solution boils, and is essential for the completion of the distillation.[105] Two experiments with material containing 3.48 per cent. of arsenic gave—(1) with ferric chloride alone, 2.74 per cent.; and (2) with the addition of calcium chloride, 3.48 per cent.

It is always necessary to make a blank determination with 1 gram of electrotype copper, to find out the amount of arsenic in the ferric chloride mixture.[106] Unfortunately, a correction is always required. This amounts to about 0.15 per cent. of arsenic on each assay, even when the mixture has been purified; and this constitutes the weakness of the method, since, in some cases, the correction is as much as, or even greater than, the percentage to be determined.

The acid distillate containing the arsenious chloride may be left for an hour or so without much fear of oxidation; but it is safer to neutralise and then to add the bicarbonate of soda, as the following experiments show. Several portions of a solution, each having a bulk of 100 c.c., were exposed for varying lengths of time, and the arsenic in each determined.

Acid Solutions. Neutralised Solutions. Time Exposed. "Iodine" Arsenic Found. "Iodine" Arsenic Found. Required. Required 18.2 c.c. = 0.0136 gram 18.1 c.c. = 0.0136 gram 1 hour 18.2 " = 0.0136 " 18.2 " = 0.0136 " 2 hours 17.7 " = 0.0133 " 18.0 " = 0.0135 " 4 " 17.5 " = 0.0131 " 18.4 " = 0.0138 " 5 " 17.0 " = 0.0127 " 18.3 " = 0.0137 "

The Titration.—Make a standard solution of iodine by weighing up in a beaker 16.933 grams of iodine and 30 grams of potassium iodide in crystals; add a few c.c. of water, and, when dissolved, dilute to 1 litre: 100 c.c. will equal 0.500 gram of arsenic.

A solution of starch similar to that used in the iodide-copper assay will be required. Use 2 c.c. for each assay. Variations in the quantity of starch used do not interfere; but the solution must be freshly prepared, as after seven or eight days it becomes useless.

To standardise the iodine solution, weigh up 0.3 gram of white arsenic; dissolve in caustic soda; neutralise; after acidulating, add 2 grams of bicarbonate of soda and 2 c.c. of the starch solution, and dilute to 200 c.c. with cold water. Fill a burette having a glass stop-cock with the iodine solution, and run it into the solution of arsenic, rapidly at first, and then more cautiously, till a final drop produces a blue colour throughout the solution. Calculate the standard in the usual way. White arsenic contains 75.76 per cent. of arsenic.

The following experiments show the effect of variation in the conditions of the titration:—

Make a solution of arsenic by dissolving 6.60 grams of white arsenic in 100 c.c. of sodic hydrate solution; render slightly acid with hydrochloric acid; add 10 grains of bicarbonate of soda, and dilute to 1 litre: 100 c.c. will contain 0.50 gram of arsenic.

Effect of Varying Temperature.—The reaction goes on very quickly in the cold, and, since there is no occasion for heating, all titrations should therefore be carried out cold.

Effect of Varying Bulk.—In these experiments, 20 c.c. of arsenic solution were taken, 2 grams of bicarbonate of soda and 2 c.c. of starch solution added, and water supplied to the required bulk. The results were:—

Bulk 50.0 c.c. 100.0 c.c. 250.0 c.c. 500.0 c.c. "Iodine" required 20.0 " 20.0 " 20.0 " 20.0 "

Considerable variation in bulk does not interfere.

Effect of Varying Bicarbonate of Soda.—This salt must be present in each titration in considerable excess, to prevent the interference of free acid. The bicarbonate must be dissolved without heating, as neutral carbonates should be avoided.

Bicarbonate added 1 gram 2 grams 5 grams 10 grams "Iodine" required 20.1 c.c. 20.0 c.c. 20.1 c.c. 20.0 c.c.

These results show that large variation in the quantity of bicarbonate has no effect.

Effect of Free Acid.—In these experiments, the arsenic taken, the starch, and the bulk were as before, but no bicarbonate was added. In one case the solution was rendered acid with 5 c.c. of acetic acid, and in the other with 5 c.c. of hydrochloric acid; in both cases the interference was strongly marked, and no satisfactory finishing point could be obtained. This was much more marked with the hydrochloric acid.

Effect of Foreign Salts.—The process for getting the arsenic into solution will exclude all metals except tin, but the solution will be charged with sodium or ammonium salts in the process of neutralising, so that it is only necessary to see if these cause any interference. The alkaline hydrates, including ammonia, are plainly inadmissible, since no free iodine can exist in their presence. Monocarbonates similarly interfere, but to a much less extent; hence the necessity for rendering the assay distinctly acid before adding the bicarbonate of soda.

With 20 c.c. of arsenic solution; and with bulk, soda, and starch as before, the results obtained were:—

"Iodine" required. With 20 grams of ammonic chloride 20.0 c.c. " 20 grams of sodium chloride 20.0 " " 20 grams of sodium acetate 20.0 " " 0.050 gram of tin, as stannic chloride 19.6 " Without any addition 20.0 "

The interference of the stannic salt is probably mechanical, the precipitate carrying down some arsenious acid.

Effect of Varying Arsenic.—With bulk, starch, and soda as before, but with varying arsenic, the results were:—

Arsenic added 1.0 c.c. 10.0 c.c. 20.0 c.c. 50.0 c.c. 100.0 c.c. "Iodine" required 1.1 " 9.9 " 20.0 " 50.0 " 100.0 "

Determination of Arsenic in Metallic Copper.—Put 1 gram of the copper filings, freed from particles of the file with a magnet, into a 16-oz.-flask; and distil with the ferric chloride mixture, as above described. Neutralise the distillate; acidify; add bicarbonate of soda and starch; dilute; and titrate with the standard solution of iodine.[107] Make a blank determination with 1 gram of electrotype copper, proceeding exactly as with the assay; and deduct the amount of arsenic found in this experiment from that previously obtained.

Working in this way on a copper containing 0.38 per cent. of arsenic and 0.80 per cent. of antimony, 0.38 per cent. of arsenic was found.

Determination of White Arsenic in Crude Arsenic.—Weigh out 1 gram of the dried and powdered substance (or 0.5 gram if rich), and digest with 10 c.c. of a 10 per cent. solution of soda; dilute to about 50 c.c., and filter. Render faintly acid with hydrochloric acid, and filter (if necessary); add 2 or 3 grams of bicarbonate of soda in solution, then 5 c.c. of starch, and titrate the cold solution with the standard solution of iodine.

The following is an example:—

1 gram of crude arsenic required 53.7 c.c. "Iodine;" 100 c.c. "Iodine" = 0.6000 gram white arsenic; 100 : 53.7 :: 0.6 : 0.3222, or 32.2 per cent.

With the test-tube method of dry assaying, this same sample gave results varying from 33 to 35 per cent. of white arsenic, which (judging from its appearance) was impure.

URANIC ACETATE PROCESS.

This may be looked upon as an alternative to the gravimetric method. It is applicable in all cases where the arsenic exists in solution as arsenic acid or as arsenate of soda. The process may be considered in two parts: (1) the preparation of the solution, and (2) the titration.

Preparation of the Solution.—If the arsenic has been separated as sulphide, it is sufficient to attack it with 10 or 15 c.c. of nitric acid, and to heat gently till dissolved, avoiding too high a temperature at first. Afterwards continue the heat till the separated sulphur runs into globules, and the bulk of the acid has been reduced to 3 or 4 c.c. Dilute with 20 or 30 c.c. of water; put in a piece of litmus paper; and add dilute ammonia until just alkaline. Then add 5 c.c. of the sodium acetate and acetic acid solution (which should make the solution distinctly acid); dilute to 150 c.c., and heat to boiling. The solution is ready for titrating.

When the arsenic exists in a nitric acid solution mixed with much copper, it is separated in the way described under Examination of Commercial Copper (Arsenic and Phosphorus), pages 208, 209.

If the arsenic has been separated as ammonium-magnesium arsenate, and phosphates are known to be absent; dissolve the precipitate (after filtering, but without washing) in dilute hydrochloric acid. Add dilute ammonia till a slight precipitate is formed, and then 5 c.c. of the sodium acetate and acetic acid solution; dilute to 150 c.c., and heat to boiling. Titrate.

If phosphates are present (which will always be the case if they were present in the original substance, and no separation with sulphuretted hydrogen has been made), the phosphorus will count in the subsequent titration as arsenic (one part of phosphorus counting as 2.4 parts of arsenic). It will be necessary to dissolve the mixed arsenate and phosphate of magnesia in hydrochloric acid. Add about four or five times as much iron (as ferric chloride) as the combined phosphorus and arsenic present will unite with, and separate by the "basic acetate" process as described under PHOSPHORUS in the Examination of Commercial Copper, page 209. Obviously, when phosphates are present, it is easier to separate the arsenic as sulphide than to precipitate it with the "magnesia mixture."

The Titration.—The standard solution of uranium acetate is made by dissolving 34.1 grams of the salt (with the help of 25 c.c. of acetic acid) in water; and diluting to 1 litre. The water and acid are added a little at a time, and warmed till solution is effected; then cooled, and diluted to the required volume: 100 c.c. will equal 0.50 gram of arsenic.

The sodic acetate and acetic acid solution is made by dissolving 100 grams of sodic acetate in 500 c.c. of acetic acid, and diluting with water to 1 litre. Five c.c. are used for each assay.

The solution of potassic ferrocyanide used as indicator is made by dissolving 10 grams of the salt in 100 c.c. of water.

To standardise the solution of uranium acetate, weigh up a quantity of white arsenic (As{2}O{3}) which shall be about equivalent to the arsenic contained in the assay (0.1 or 0.2 gram); transfer to a flask, and dissolve in 10 c.c. of nitric acid with the aid of heat. Evaporate to a small bulk (taking care to avoid the presence of hydrochloric acid); dilute with water; add a small piece of litmus paper; render faintly alkaline with ammonia; then add 5 c.c. of the sodic acetate mixture; dilute to 150 c.c.; and heat to boiling.

Fill an ordinary burette with the uranium acetate solution, and run into the assay a quantity known to be insufficient. Again heat for a minute or two. Arrange a series of drops of the solution of ferrocyanide of potassium on a porcelain slab, and, with the help of a glass rod, bring a drop of the assay solution in contact with one of these. If no colour is produced, run in the uranium acetate, 1 c.c. at a time, testing after each addition, till a brown colour is developed. It is best to overdo the assay, and to count back. It is not necessary to filter off a portion of the assay before testing with the "ferrocyanide," since the precipitate (uranic arsenate) has no effect.

The following experiments show the effect of variation in the conditions of titration. Make a solution of arsenic acid by dissolving 4.95 grams of arsenious acid (As{2}O{3}) in a covered beaker with 35 c.c. of nitric acid; evaporate down to 7 or 8 c.c.; and dilute with water to 1 litre: 100 c.c. will contain 0.375 gram of arsenic. Use 20 c.c. for each experiment.

Effect of Varying Temperature.—It is generally recommended to titrate the boiling solution, since it is possible that the precipitation is only complete on boiling. Low results are obtained in a cold solution, the apparent excess of uranium acetate striking a colour at once; on boiling, however, it ceases to do so; consequently, the solution should always be boiled directly before testing.

In four experiments made in the way described, but with 20 c.c. of a solution of arsenic acid stronger than that given (100 c.c. = 0.5 gram As), the results at varying temperatures were:—

Temperature 15 C. 30 C. 70 C. 100 C. "Uranium" required 18.0 c.c. 18.5 c.c. 18.5 c.c. 18.7 c.c.

Effect of Varying Bulk.—These experiments were like those last mentioned, but were titrated boiling, and the volume was varied:—