Arsenic Trisulphide, As₂S₃

Arsenic Trisulphide (Arsenious Sulphide, Orpiment), As₂S₃, occurs in Nature as a yellow mineral which was well known in early times. It may be formed artificially by heating together arsenious oxide and sulphur in the requisite proportions, when it sublimes; or by melting together realgar and sulphur in suitable proportions. It may be produced as a precipitate by the action of hydrogen sulphide or an alkali sulphide on a solution in hydrochloric acid of arsenious oxide, an arsenite or an arsenate. In the latter case the reaction is slow but may be accelerated by the presence of a soluble iodide, which facilitates the reduction of the arsenate probably in the following manner:

H₃AsO₄ + 2HI = H₃NaO₃ + H₂O + I₂
H₂S + I₂ = 2HI + S

With arsenious oxide the product has a high degree of purity if precipitation is not allowed to proceed to completion. The presence of hydrochloric acid is not necessary and the arsenious oxide may be suspended in water or in a salt solution, in which case the liquid should be warmed as the hydrogen sulphide is passed in.

The arsenious sulphide may be formed as a crystalline precipitate under favourable conditions. When hydrogen sulphide is passed into a 0.2N solution of arsenious oxide in water, golden-yellow leaflets are formed in small quantity, and the amount increases with the concentration up to 0.3N. Beyond this concentration the amount of crystalline sulphide produced diminishes whilst, with very dilute solutions (0.05N), traces only are formed. The crystalline form is also obtained by heating in a sealed tube a mixture of arsenious oxide, ammonium thiocyanate and hydrochloric acid. If arsenic acid is used in place of the oxide, sulphur is also precipitated. A convenient method for obtaining the crystalline sulphide is first to obtain the double sulphide, As₂S₃.C₄H₁₀N₂.H₂S, which is formed by the prolonged action of hydrogen sulphide on the solution obtained by boiling arsenious oxide with piperazine; if this is treated with cold dilute hydrochloric acid or sodium hydroxide, crystalline arsenious sulphide is formed.

Sodium hydrosulphite reduces arsenates and arsenites, yielding precipitates containing sulphides the composition of which varies with conditions. In strongly acid solutions arsenic trisulphide is the main product. Sodium thiosulphate also precipitates arsenic as the trisulphide from acid solutions, but the amount of precipitation depends on the nature and concentration of the acid present. Thus, with hydrochloric, perchloric or sulphuric acid, the precipitation reaches a maximum of 50 to 80 per cent, for 0.1N acid, and above this concentration the amount of precipitation falls to zero with hydrochloric acid but passes through a minimum with perchloric acid at N concentration and with sulphuric acid at 2 to 3N concentration. At still higher acid concentrations precipitation becomes almost quantitative.

If arsenopyrite is allowed to stand in aqueous hydrochloric acid for some time, the formation of arsenic trisulphide may be observed.

When yellow arsenious sulphide, obtained by precipitation from the colloidal solution by addition of an electrolyte, is heated in an air oven at 100° C., it is converted into a red vitreous mass. If hydrochloric acid is present, this must first be removed before the change will take place. The fed form may also be obtained by evaporation of the colloidal solution on a water-bath, or by freezing the colloidal solution, when a mixture of ice and the red form separates. The red form is gradually transformed to the yellow by exposure to air for 5 to 6 weeks at the ordinary temperature, or the change may be brought about more rapidly by heating for some time at 150° to 160° C. When the yellow precipitated arsenious sulphide is dried in a current of dry air, the product is pale yellow and, according to Spring, has the composition As₂S₃.6H₂O. Its density is 1.8806 at 25.6° C. It is decomposed by a pressure of 6000 to 7000 atm. into the sulphide and water. Spring also observed that the precipitated sulphide becomes micro- crystalline if kept for some days at 150° C.

Arsenic trisulphide crystallises in short octahedral prisms which were at first described as rhombic but which Groth described as monoclinic. The crystals are greasy and possess a lustre which is nacreous at the plane of cleavage. The cleavage on the (010)-face is perfect, while that on the (100)-face shows in traces. The (001)-face is a gliding plane. The optical character is negative. The colour of the crystals varies from lemon-yellow to deep orange, whilst the precipitated sulphide, when dried, may be of any shade from yellow, through orange, to red according to the conditions of precipitation. The lighter shades result when hydrogen sulphide acts upon a solution of arsenious oxide or an arsenite containing sufficient acid or other electrolyte to cause immediate precipitation. The darker shades are obtained when the tervalent arsenic is first converted to arsenious sulphide sol and this is subsequently coagulated by the addition of an excess of an electrolyte. The red colour has been variously attributed to polymorphism, to the presence of realgar or of a red thioarsenite. Weiser, however, considers that the variation in colour is due to differences in the physical nature of the precipitated sulphide. Direct precipitation in the presence of a foreign electrolyte yields a flocculent precipitate of relatively large particles or loose aggregates which disintegrate on drying to an impalpable yellow to orange-yellow powder. On the other hand, coagulation of a sol gives a gelatinous precipitate consisting of aggregates of ultramicroscopic particles which on drying coalesce to give a red glassy mass. The latter, by grinding or by heating below the sintering temperature, disintegrates and becomes yellow, whilst heating the yellow sulphide at about 175° C. causes it to sinter, contract and assume an orange to brown colour, depending on the temperature and time of heating. The density of arsenic trisulphide varies from 3.44 to 3.48 according to its origin. The hardness of natural orpiment is 1.5 to 2, and of the artificial vitreous form, 3.

The suggestion of Winter that two distinct forms of the sulphide do exist was supported by the work of Borodowski. The former gave the transition point as 150° to 160° C. If the yellow form, α-As₂S₃, is heated for some time at 170° in carbon disulphide vapour, it changes to the red β-form, but there is no change below that temperature. Weiser observed, however, that the red form is stable at ordinary temperatures in the dark and when thoroughly dry is not affected by light. The presence of light and moisture causes a superficial chemical disintegration and the red sulphide becomes coated with a yellow film of sulphur and sulphide. This photochemical action is similar to that on As₂S₃ sols.

When heated, arsenious sulphide readily sublimes and fusion occurs, according to Borgstrom, at 320° C. Earlier determinations have put the melting point at 310°, 300° and 325° C. Air must be excluded or oxidation occurs. Some degree of volatilisation may be observed at the ordinary temperature. According to Schuller, when heated in a vacuum volatilisation begins only after melting. In the vacuum of the cathode light, sublimation begins at a temperature just above 220° C. and the sulphide distils unchanged. Vapour density determinations between 820° and 1150° C. indicate that dissociation occurs and that the vapour probably contains molecules of As₂S₃, As₂S₂, As₄, As₂ and S₂.

The crystals exhibit pleochroism, being greenish and reddish-yellow in the direction of the a and c axes, respectively; on gentle heating, the pleocroism resembles that of realgar but, if the temperature is not allowed to exceed 150° C., the original state is recovered on cooling. A suggested explanation of this is that the reversible change

3As₂S₃ ⇔ 2As₂S₂ + As₂S₅

occurs.

Arsenic trisulphide does not conduct electricity at the ordinary temperature, but if heated above 60° C. conductivity becomes appreciable. The incidence of light appears to be without effect. The sulphide exhibits no fluorescence in ultraviolet light; 7 it is opaque to X-rays.

The molar heat of formation from solid arsenic and rhombic sulphur has been calculated to be 34,700 calories.

Arsenic trisulphide may be reduced to arsenic by heating in a stream of hydrogen; the arsenic sublimes. The reaction begins at about 300° C., but proceeds more readily if the sulphide is first fused with an alkali carbonate. A similar reduction occurs when the sulphide is heated with a mixture of charcoal and alkali carbonate or lime; when heated with potassium cyanide, an oxalate or with a metal such as silver or iron; the latter if in excess yields arsenide.

In moist air the sulphide undergoes slow oxidation. As the temperature is raised, oxidation is appreciable at about 200° C. and complete at 750° C., arsenious oxide and sulphur dioxide being produced. When ignited, the sulphide burns with a pale lilac flame. With the sulphide in neutral or acid solution, oxidation by means of atmospheric oxygen under pressure proceeds only very slowly, but in alkaline media oxidation occurs more readily - under such conditions the process of dissolution of the sulphide in aqueous sodium hydroxide is extremely complex and consists of a number of successive reactions which at 100° to 110° C. may be represented summarily by the equation

2As₂S₃ + 8NaOH + 5O₂ = 4NaAsO₂ + 2Na₂S₂O₃ + 2S + 4H₂O
and at 150° to 300° C. by
As₂S₃ + 10NaOH + 7O₂ = 2Na₂HAsO₄ + 3Na₂SO₄ + 4H₂O

with Na₂HAsS₃ in both cases as an intermediate product. The reaction is catalysed by copper sulphate to an extent increasing with rise of temperature. If a suspension of the sulphide in 4N-sodium hydroxide is heated for two hours at 150° C. and 25 to 50 atm., colloidal sulphur is formed, which may be precipitated by the addition of aqueous carbon dioxide or sulphuric acid, the amount obtained being 50 to 75 per cent, of that originally present as arsenious sulphide.

The sulphide is only slightly attacked by water at the ordinary temperature, and even on prolonged boiling only a little arsenious oxide passes into solution, whilst a trace of hydrogen sulphide is evolved. With the freshly precipitated sulphide the hydrolysis is more rapid, possibly owing to the physical condition allowing greater contact with the water. It has been suggested that the accelerated reaction may be due to the more ready formation of an intermediate hydroxysulphide by the fresh precipitate; the presence of arsenious oxide, which forms an oxysulphide, retards the reaction, however. According to Regnault, steam reacts to form an arsenic oxysulphide of variable composition. If the sulphide is boiled with water in vacuo, decomposition commences at 22° C.

The solubility of arsenic trisulphide in pure water at 0° C. has been determined by digesting the mixture for several days, filtering through an ultra-filter, and estimating the dissolved arsenic iodometrically; a value of 0.89 mg. As₂S₃ per litre was obtained. In the presence of 0.002 per cent, of hydrogen sulphide the solubility was reduced to 0.23 mg. per litre. With higher concentrations of hydrogen sulphide the solubility increased, but this increase did not occur in the presence of hydrochloric acid and the trisulphide may be precipitated quantitatively by saturation of the solution in hydrochloric acid with hydrogen sulphide. Arsenic trisulphide dissolves readily in solutions of alkali hydroxides, carbonates or sulphides. Thus with alcoholic sodium hydroxide, arsenate and mono- and di-thioxyarsenates result. With alcoholic sodium hydroxide, arsenate, mono- and di-thioxyarsenates are formed. A solution in sodium carbonate, saturated at 80° C., yields crystals of the trisulphide on cooling, but if heated to 100° C. a thioarsenite is formed in solution. In yellow ammonium sulphide, ammonium thioarsenate is formed -

3(NH₄)₂S + 2S + As₂S₃ = 2(NH₄)₃AsS₄

while in an aqueous solution of sodium sulphide a thioarsenite is produced:

Na₂S + As₂S₃ = 2NaAsS₂

In ammoniacal solution, arsenic trisulphide is oxidised by hydrogen peroxide to arsenic and sulphuric acids.

Arsenic trisulphide reacts readily with the halogens. When exposed to chlorine, considerable heat is evolved and a liquid product containing arsenic trichloride and sulphur dichloride is obtained. In an aqueous medium oxidation to quinquevalent arsenic occurs. Oxidation also results when the sulphide is heated with hydrochloric acid and potassium chlorate, some arsenic trichloride being vaporised during the process. Bromine water, or a solution of bromine in hydrochloric acid or in aqueous potassium bromide, reacts similarly, quinquevalent arsenic and arsenic tribromide being formed in solution. Iodine in carbon disulphide solution reacts with the freshly precipitated sulphide, though not with natural orpiment, to form arsenic triiodide and sulphur. The same products result when a mixture of the sulphide and iodine is gently heated, but at a higher temperature the reaction is reversible:

As₂S₃ + 3I₂ ⇔ 2AsI₃ + 3S

When arsenic trisulphide is exposed to dry hydrogen chloride or hydrogen bromide, it liquefies at the ordinary temperature and on heating complete volatilisation occurs. It is not readily attacked by halogen acids. When boiled with concentrated hydrochloric acid it is decomposed, but with great difficulty, and the hydrogen sulphide and arsenious chloride evolved reproduce arsenious sulphide in the receiver. A similar reaction occurs when heated with a chloride in the presence of concentrated sulphuric acid, but the decomposition is incomplete. The reaction is facilitated by the presence of cuprous chloride or ferric chloride. Only a slight reaction is observed with dilute acid, and the trisulphide is quite insoluble in, and unacted upon by, aqueous hydrochloric acid of density 1.16, providing the liquid is saturated with hydrogen sulphide. Metallic chlorides, notably mercuric chloride, react directly on heating to produce arsenic trichloride.

When finely divided arsenic trisulphide is exposed to gaseous ammonia, the latter is slowly absorbed until, after about three weeks, the composition of the product corresponds with As₂S₃.NH₃; this loses ammonia on exposure to air, whilst water converts it to ammonium arsenite and thioarsenite. The trisulphide dissolves readily in aqueous ammonia and is slightly soluble in liquid ammonia. It is decomposed by nitric acid or aqua regia. A few drops of fuming nitric acid on melted orpiment produce a deflagration, whilst nitric acid of density 1.42 causes separation of sulphur, which melts and may form a protective film on some of the sulphide particles, thus preventing complete oxidation. The presence of hydrazine effectively retards the oxidising action of nitric acid on arsenic trisulphide, both with acid of concentration 1 to 4N at boiling temperature, and with 6 to 10N acid (density 1.2 to 1.3) at 20° to 22° C. The reaction velocity depends upon the amount of sulphide in solution, and a gradual oxidation occurs which, even in the presence of large amounts of hydrazine, reaches equilibrium after several days, during which the components are destroyed in the proportions of 1 mole of hydrazine to 4 equivalents of arsenic trisulphide. In the presence of oxygen under a pressure of 20 atm. and of twice its weight of 40 per cent, nitric acid, orpiment is completely oxidised at 120° C. to arsenic and sulphuric acids in 15 minutes; this amount of nitric acid, as in the case of realgar, is less than that theoretically needed in the absence of oxygen. Oxygen alone acts on an aqueous suspension of the sulphide to oxidise some of the sulphur to sulphuric acid, but no arsenic acid is formed.

Arsenic trisulphide is not dissolved by dry liquid hydrogen sulphide. The precipitated sulphide obtained by the action of hydrogen sulphide on solutions of arsenious oxide in aqueous hydrochloric or acetic acid is found to contain an amount of sulphur in excess of that required by the formula As₂S₃, but which cannot be extracted with carbon disulphide. This has been attributed to the formation of a hydrosulphide since, if the precipitate is dried in a vacuum and then heated at 115° C., hydrogen sulphide is evolved. The products vary in composition, but that produced in presence of hydrochloric acid approximates to 16As₂S₃.H₂S and that from acetic acid solutions to 8As₂S₃.H₂S.

The trisulphide reacts with sulphur dioxide at temperatures between 300° and 800° C. to form sulphur and a sulphate, whilst if the sulphide is digested with an aqueous solution of sulphur dioxide, or of potassium hydrogen sulphite, it dissolves and the solution when boiled evolves sulphur dioxide, and arsenate, thiosulphate and free sulphur are formed. Concentrated sulphuric acid also dissolves the trisulphide to form arsenious oxide and sulphur dioxide. Sulphur monochloride, when heated with the sulphide, yields a molten mixture of arsenic trichloride and sulphur; decomposition is complete at about 140° C. Sulphur iodides react similarly. Thionyl chloride also attacks the trisulphide when the mixture is heated in a sealed tube at 150° C.

Arsenic trisulphide is insoluble in benzene or carbon disulphide; it is soluble in an aqueous solution of citric acid or of an alkali citrate. An aqueous solution of borax (2 per cent.) dissolves the sulphide slowly in the cold, more rapidly on heating.

Arsenious sulphide reacts with metallic sulphides as an acid thioanhydride and forms a series of complex salts known as thioarsenites. These may be considered to be derived from the following hypothetical acids:

Orthothioarsenious acid, H₃AsS₃
Metathioarsenious acid, HAsS₂
Pyrothioarsenious acid, H₄As₂S₅
Metathiotriarsenious acid, HAs₃S₅
Orthothiotetrarsenious acid, H₆As₄S₉
Metathiotetrarsenious acid, H₂As₄S₇
Metathio-octo-arsenious acid, H₂As₈S₁₃
Metathioennea-arsenious acid, HAs₉S₁₄
Metathiododeca-arsenious acid, H₂As₁₂S₁₉
These acids are supposed to be formed from H_3nAs_nS_3n by loss of H₂S; thus the last of the series is equivalent to H₃₆As₁₂S₃₆-17H₂S.

Many of these compounds have been described in the literature, but in only a few cases have the conditions been such as to produce pure compounds. Many thioarsenites occur in Nature.

The salts are usually prepared by the interaction of arsenious sulphide with the metallic sulphide, hydrosulphide or carbonate, taking care to exclude air to prevent the formation of thioarsenates. Thus, Nilson obtained the salts of the alkali and alkaline earth metals by dissolving arsenious sulphide in the aqueous solutions of the respective hydrosulphides and concentrating in vacuo.

Wiinschendorff, using carefully purified arsenious sulphide and the alkali sulphide, prepared the following compounds:

Potassium metathioarsenite, KAsS₂, red rhombic crystals;
Potassium metathiotriarsenite, KAs₃S₅.1.5H₂O, an insoluble red granular powder;
Potassium metathiotetrarsenite, K₂As₄S₇.2H₂O, red crystalline needles.

The ortho- and pyro-thioarsenites could not be obtained in the solid condition, their solutions decomposing into the orthothioarsenate and arsenic when concentrated. Berzelius described products obtained by precipitation with alcohol from solutions of arsenious sulphide in alkali sulphides as orthothioarsenites, R₃AsS₃ (R = K, Na, NH₄); and pyrothioarsenites, R₄As₂S₅, were said to be obtained by heating the corresponding pyrothioarsenates. Nilson obtained the orthothiotetrarsenite, K₆As₄S₉.8H₂O, as a blood-red gelatinous mass by evaporation of the mother liquor from the metathiotetrarsenite, after separating the latter from a solution of arsenious sulphide in potassium hydrosulphide.

Five sodium salts may be prepared by Wunschendorff's method:

Sodium metathioarsenite, NaAsS₂, brown prismatic crystals;
Sodium metathiotriarsenite, NaAs₃S₅.3H₂O, brown spherites;
Sodium metathiotetrarsenite, Na₂As₄S₇.2H₂O, brown prisms;
Sodium pyrothioarsenite, Na₄As₂S₅.H₂O, dark orange prisms;
Sodium orthothioarsenite, Na₃AsS₃.

The last two compounds are extremely unstable and decompose rapidly to form arsenic and sodium orthothioarsenate.
Only two ammonium salts have been prepared by the above method:

Ammonium metathioarsenite, NH₄AsS₂, yellow needles, very unstable;
Ammonium metathiotetrarsenite, (NH₄)₂As₄S₇, red needles, stable.

The ortho-salt was reported by Berzelius, and Nilson obtained ammonium metathiotriarsenite, NH₄As₃S₅.2H₂O, by evaporation of a saturated solution of arsenious sulphide in ammonium hydrosulphide.

The following other salts have been obtained by Wlinschendorff:

Ca(AsS₂)₂.8H₂O, yellow prisms;
Ca₂As₂S₅.9H₂O, yellow triclinic crystals, unstable;
Sr(AsS₂)₂.2H₂O, yellow amorphous powder;
Sr₂As₂S₅.7H₂O, orange triclinic crystals;
Sr₃(AsS₃)₂.6H₂O, yellowish-white scales;
Ba₃As₄S₉.6H₂O, brownish-yellow crystals;
Ba₂As₂S₅.5H₂O, yellow crystals;
Ba₃(AsS₃)₂.8H₂O, yellow prisms;
Ba(AsS₂)₂.0.5H₂O, an insoluble brown precipitate;
AgAsS₂; Ag₃AsS₃; KAg₂AsS₃; M₃(AsS₃)₂ and KMAsS₃ (M = Zn, Pb, Mn). The corresponding salts of iron, cobalt and nickel were apparently formed, but were unstable and could not be purified.

Thioarsenites of the heavy metals may be prepared in a dry way by heating together arsenious sulphide and the metallic chloride in suitable proportions. For example, a mixture containing 3AgCl:As₂S₃ yields the ortho-salt Ag₃AsS₃, the reaction commencing at 150° C.; the product is a brittle reddish-black lustrous mass. From 3AgCl:2As₂S₃ reddish-black crystals of the meta-salt, AgAsS₂, are obtained, the reaction commencing at 170° C. The pyro-salt, Ag₄As₂S₅, is obtained from a mixture of composition 12AgCl:5As₂S₃; it is a lustrous black solid. In a similar manner the following lead salts have been obtained: Pb(AsS₂)₂; Pb₂As₂S₅ and Pb₃(AsS₃)₂. Copper thioarsenites of definite composition cannot be obtained by this method, although Sommerlad obtained a product which approximated to Cu₄As₂S₅.

Thioarsenite solutions may be employed in the purification of coal-gas for the removal of hydrogen sulphide. When saturated with the latter the purifier may be recovered by aeration, until precipitation of sulphur ceases, and saturation of the filtered solution with carbon dioxide, which yields a yellow precipitate containing most of the arsenic; this is redissolved in sodium hydroxide or carbonate, with aeration, and the solution returned to the purifying circuit.

In experiments with proustite, Ag₃AsS₃, Coblentz observed that at temperatures from +20° to -50° C. the spectrophotoelectric sensitivity curve showed a slight maximum at about 6100 A. and a marked sensitivity with a maximum in the extreme violet. As the temperature is lowered to -100° C. the maximum in the violet is more or less obliterated by a new maximum (the 6100 A. band) which occurs at about 5800 A. The position of this new maximum remains quite constant as the temperature is further lowered to -170° C. No photoelectric sensitivity is observed for radiation stimuli of wavelengths extending from 10,000 to 20,000 A. in the infra-red.

Few oxythioarsenites have been prepared. A sodium compound of composition Na₈As₁₈O₇S₂₄.30H₂O has been obtained by several methods, for example: by boiling a mixture of arsenious sulphide and aqueous sodium carbonate; by boiling a mixture of arsenious oxide and sodium hydrosulphide, adding alcohol to the filtered solution and allowing the alcoholic extract to crystallise; by evaporation of the mother liquor from sodium sulphite produced by interaction of sodium thiosulphate and sodium dihydrogen orthoarsenite. It crystallises as deep red hexagonal plates, which decompose in the presence of water, acids and alkalis. A barium salt, Ba₅As₄O₂S₉.6H₂O, has also been described.

It was observed by Berzelius that arsenious sulphide, obtained by precipitation from aqueous arsenious oxide with hydrogen sulphide, after it had been washed with cold water, dissolved to a slight extent in hot water forming a yellow solution. Water containing hydrogen sulphide did not dissolve it. On keeping, the yellow solution gradually deposited the sulphide. Moreover, whilst hydrogen sulphide immediately and almost completely precipitates the sulphide from a saturated solution of arsenious oxide, if the gas or its aqueous solution is added to a dilute aqueous solution of the oxide a clear yellow solution results which, after excess of hydrogen sulphide has been removed by passing oxygen or hydrogen through it, gives on addition of hydrochloric acid complete precipitation of the arsenic as trisulphide. Thus, although the hydrogen sulphide causes no precipitation, the arsenic is quantitatively converted to trisulphide, which remains in colloidal solution. This hydrosol has been the subject of much classical investigation, especially as regards the conditions governing its stability and coagulation, for it was early observed that the sulphide separated in yellow flakes when the liquid was heated or frozen, or on adding to it certain electrolytes or even insoluble powders such as charcoal, copper oxide, glass powder or powdered Iceland spar.

In order to prepare the hydrosol free from electrolytes, pure arsenious oxide should be dissolved in "conductivity water" which is kept boiling and the solution obtained allowed to flow into a saturated solution of hydrogen sulphide through which a current of the gas is continuously passing. The uncombined hydrogen sulphide is subsequently removed by passing a current of hydrogen, preferably with exclusion of light, and the liquid is finally filtered. Or, hydrogen sulphide gas may be passed into the saturated solution of arsenious acid until the latter can no longer be detected in the filtrate after precipitation with an electrolyte. There is a limit to the concentration of the sols thus prepared, owing to the sparing solubility of arsenious oxide; but by passing hydrogen sulphide and then dissolving more arsenious oxide, and so repeating several times, a sol containing as much as 37.46 per cent. As₂S₃ has been obtained. Much of the water may be eliminated under reduced pressure and any large particles removed by energetic centrifuging. Such a sol has the appearance of an intensely yellow milk, but is transparent under the microscope. Dilute hydrosols of arsenious sulphide prepared from more concentrated sols by dilution are more turbid than dilute sols of the same concentration prepared directly, and are more yellow than the latter, which have a reddish- yellow tint.

According to Gazzi, the most highly purified sols contain an excess of arsenious oxide since, on analysis of the hydrosol, the quantity of oxide obtained is always much greater than that which could result from the complete hydrolysis of soluble arsenic trisulphide. Gazzi found the solubility of the precipitated trisulphide to be 0.5166 mg. per litre, and this quantity would yield 0.4154 mg. of the trioxide per litre. In sols prepared by passing hydrogen sulphide through aqueous solutions of arsenious oxide, and after addition of two drops of dilute sulphuric acid, filtering and removing excess of hydrogen sulphide either by dialysis, boiling or passing hydrogen with exclusion of light, the arsenious oxide content was 1.5 to 3.0 mg. per litre for the dialysed sols, and with the others the amount increased with the duration of boiling or passing hydrogen. Chaudhury and Kundu found the atomic ratio of As:S in sols containing excess of arsenious acid to be 1:1.46, agreeing with the composition As₂S₃, but in sols which were purified from arsenious acid and hydrogen sulphide the ratio was 1:2, possibly corresponding with As₂S₃.H₂S or As₂S₃.As₂S₅, the former being the more probable.

In sols purified by electrodecantation the concentration of hydrogen ion calculated from the results of electrometric titration is greater than that given by the conductivity, whereas the converse is true of the liquid separated from a coagulum obtained by freezing the sol. The coagulum has the composition As₂S₃ but retains adsorbed stabilising groups from which H + may be liberated by treatment with a barium salt. Both the original sol and the intermicellar liquid may be shown by conductometric titration and by analysis to contain the acid H₃AsO₄ and a salt, probably H₂(AsO)AsO₄, which are formed by oxidation of the S-containing stabilising complex. Sulphur is absent from the intermicellar liquid.

The arsenic trisulphide hydrosols, if carefully protected from air and light, are very stable, and may be kept for considerable periods with little deposition; Linder and Picton record no change in a 2 per cent, sol over three years, and Dumanski made the observation that during four years the rate of fall, due to gravity, of the particles of arsenic trisulphide was on the average 0.031 cm. per day. The sols, however, undergo oxidation in the presence of atmospheric oxygen, the products after prolonged action being arsenious acid, free sulphur and sulphuric acid. The last-named has a precipitative influence on the colloid present and may be a disturbing factor in experimental determinations carried out in the presence of air. The presence in the sol of electrolytes, non-electrolytes and protective colloids, has each a marked effect on the stability of the colloid and is discussed below.

The colour of arsenic trisulphide hydrosols varies from pale yellow to orange-red, dependent to some extent apparently on the size of the colloid particles. The coarser suspensions are usually orange-red, whilst with sols in the highest degree of fine division there is only slight milkiness by transmitted light. According to Menon, light refracted by freshly prepared sols is almost completely plane-polarised. Bhatnagar, however, does not agree with the view that the difference in colour is due only to difference in physical character, but states that the reddish precipitate separated from a sol has a more complex composition than is indicated by the formula As₂S₃. The size of the colloid particles increases with the concentration of the arsenious oxide solution employed in the preparation of the sol; thus when the concentration was 10^-2 N, Borjeson found the mean radius of the sulphide particles to be 39 μμ; with 5×10^-4 N arsenious oxide, the mean radius was 16 μμ, and with 10⁴ N the value was 11 μμ. According to Boutaric and Semelet, an orange-coloured fine-grained sol is obtained by a rapid flow of hydrogen sulphide through aqueous arsenious acid, whereas with a slow supply a yellow coarse-grained sol is produced. The mean magnitude also increases to a slight extent with rise in temperature and may be further increased by protracted boiling of the sol at constant volume. The colour of the sol usually darkens on boiling, the opacity increases and precipitation occurs. The Brownian movement may be observed in suspensions where the radius of the particles is less than 2.5 μ. From the examination of freshly prepared sols by means of X-rays and the ultramicroscope, it has been concluded that the particles of arsenic trisulphide are amorphous and nearly spherical.

The density of the hydrosols varies linearly with the concentration up to about 9 per cent. As₂S₃, but beyond this the increase is more rapid. Linder and Picton showed that at low concentrations the density could be calculated by the law of mixtures, thus:

As2S3 per cent.	D (obs.).	D (calc.).
4.4	1.033810	1.033810
2.2	1.016880	1.016905
1.1	1.008435	1.008440
0.01719	1.000137	1.000134

The solid sulphide obtained from the sol by rapid centrifuging was found by Dumanski to have density 2.938. The viscosity also depends on the concentration of the sol and, according to Boutaric and Simonet, if η and η₀ represent respectively the viscosities of the sol and of the dispersive medium, both at 20° C., and φ the ratio of the volume of the disperse substance to that of the suspension, the value of

k = (η - η₀)η₀φ (Einstein's equation)

approaches 2.5 as dilution approaches infinity. The addition of an electrolyte generally causes a change in the viscosity; thus, small quantities of potassium chloride or cadmium chloride cause an increase to a maximum, after which the viscosity falls off with further addition. Such changes are not observed, however, on addition of mercuric chloride. The surface tension of the hydrosol is the same as that of water. The diffusibility of the sol into water was studied by Linder and Picton, who confirmed Graham's view that colloids, no less than electrolytes, diffuse considerably, although the rate is very slow. The dialysis is influenced by the presence of other substances; thus the presence of a soluble tartrate accelerates, while the presence of a gel retards, the speed of diffusion, the extent of the effect depending on the concentration of the added substance. The molecular weight derived from the diffusion constant is greater than 6000. Osmotic pressure measurements give very variable values which are always small, and the sol has no effect on the freezing point of water.

The particles of the colloid are electronegatively charged, so that during cataphoresis they are transported towards the anode. The velocity of migration of particles suspended in a liquid is, according to Smoluchowski, given by the formula ζHε/4πη, where ζ is the potential difference of the double layer, H the fall of potential (volts per cm.), ε the dielectric constant and η the viscosity; putting ζ = 0.05 volt, the value for glass and water, Smoluchowski calculated the velocity under a potential fall of 1 volt per cm. to be 34×10^-5 cm. per sec. For an arsenic trisulphide sol containing particles of diameter 50 μμ the value 22×10^-5 was obtained by Linder and Picton, while Kruyt and van der Willigen determined the velocity to be 31×10^-5 cm. per sec. The electrical conductivity has been determined as 136×10^-6 mho. The cataphoretic speed is influenced by the presence of electrolytes, the effect usually being a fall with low concentrations, but a gradual increase as the amount of electrolyte increases; on the other hand, the decrease in cataphoretic speed may continue to high concentrations of the electrolyte11 until the flocculation point is reached. With acids and alkalies the migration velocity may be at a high value, and in some cases almost equal to that in the original sol, when precipitation occurs. This is probably due to the high adsorption at the surface of the particles increasing the dielectric constant, and while the cataphoretic velocity remains high, the critical potential at which precipitation results is lowered. The velocity increases with increase in concentration of univalent cations and does not pass through a maximum; this again appears to depend on high adsorbability. With the alkali chlorides at 0.0002N concentration the cataphoretic speeds indicate the following order of adsorption: K⁺ > Na⁺ > Li⁺ but the order of adsorption of cations varies with conditions. The addition of arsenious oxide to the hydrosol decreases the cataphoretic speed, the effect depending on the amount of oxide added. If potassium chloride is already present, the arsenious oxide causes a decrease in speed, but the effect for increasing oxide passes through a maximum; in the presence of barium chloride the decrease is greater with increasing concentration of the latter. It will readily be understood that the migration velocity, depending as it does on the charge on the colloid particle, will vary considerably with the method of preparation of the sol, the nature of the ions present, and their relative adsorbability. Acids with smaller dielectric constants, such as acetic and formic acids, are more highly adsorbed and lower the charge to a greater extent than acids of higher dielectric constant, such as oxalic and hydrochloric acids. The coagulating power of these acids is the reverse of their capacity to diminish the charge. The cataphoretic speed varies with the period of dialysis and has been observed to decrease up to 8 days, increase up to 28 days and then to decrease; the effect is attributed to changes in the composition of the sol during dialysis.

The fact that the addition of acids or salts to the hydrosol caused coagulation of the arsenic trisulphide particles was first recorded in 1832 by Boutigny, who observed that the mineral acids were most effective, but that weak organic acids, such as oxalic and acetic acids, and even carbonic acid, caused some precipitation. Such weak acids as boric, tartaric, benzoic and salicylic acids, when added in cold solution, do not cause precipitation. Salts which are strong electrolytes readily cause precipitation, and Schulze observed that the concentration required depended on the ion whose charge was of opposite sign to that of the colloid, the coagulating power of the ion being greater the higher the valency. This was confirmed in an extensive series of experiments carried out by Freundlich, some results of which are given in the table opposite. The precipitation values were compared by determining the concentration of the salt solution that, within a given time and under otherwise equal conditions, caused a separation of flocks large enough to be completely kept back by a filter of standard type.

There is a limiting concentration necessary for complete precipitation, and also a limiting concentration below which no coagulation occurs even after a long interval of time. Thus, with a hydrosol containing 9.57 millimoles of arsenious sulphide per litre, to portions of which potassium chloride was added in concentrations of 1.22, 2.44 and 3.90 millimoles per litre, respectively, no precipitation had occurred in the first two cases after 340 days, whereas in the third almost complete precipitation took place in that time.

The anion is not completely without influence on the precipitating power of the electrolyte. Solutions of chlorides, bromides, iodides and nitrates, if the cations are of equivalent concentrations, show the same coagulating power, and the same relation obtains for the free acids.

Precipitation values of electrolytes for arsenious sulphide hydrosol (7.54 millimoles per litre)

Electrolyte.	Concentration (millimoles per litre).
Univalent Cations:
HCl	30.8
½H2SO4	30.1
½K2SO4	65.5
KCl	49.5
KNO3	50.0
NaCl	51.0
LiCl	58.5
K formate	86
K acetate	110
½K3 citrate	240
Bivalent Cations:
MgSO4	0.810
MgCl2	0.717
CaCl2	0.649
SrCl2	0.635
BaCl2	0.691
Ba(NO3)2	0.687
ZnCl2	0.685
Tervalent Cations:
AlCl3	0.093
Al(NO3)3	0.095

With anions of higher valency, however, the salt concentration necessary to precipitate the trisulphide within a given time increases with the valency of the anion, and the effect is more marked with fairly complex anions, such as in benzoates and ferrocyanides. This is shown in the table of results which were obtained with sols containing 39.8 millimoles of arsenic trisulphide per litre; a is the dilution of the electrolyte in litres containing 1 gram-equivalent after mixing; b is the concentration of the cation in gram-ions per litre at 18° C. (or * at 25° C.); the relative times of coagulation were observed by passing a definite current (0.2 amp.) through a straight filament 4-volt lamp placed at a fixed distance from the cell containing the colloid. As coagulation proceeded, the sol became more and more opaque and the light of the lamp viewed through the cell diminished in intensity; the time at which the filament became invisible was determined, at least six observations being made in each case.

The influence of the anion, however, is relatively unimportant and the valency has little effect; the complexity appears to be an influencing factor.

The process of coagulation is greatly affected by the quality and concentration of the sol and by the concentration of the electrolyte. It may be studied photometrically by periodically measuring the coefficient of absorption, which first increases rapidly and then reaches a limiting value. The speed of the flocculation by alkali chlorides and by aluminium chloride is greatly retarded by the presence of a slight excess of hydrogen sulphide in the sol, but precipitation with calcium, strontium and barium chlorides is accelerated, whilst there is little effect on the precipitation with manganese or magnesium chloride. The presence of an excess of arsenious oxide, on the other hand, increases the flocculating power of univalent and bivalent ions, thus sensitising the sol, but the effect is less than that caused by hydrogen sulphide. In comparing the effects of electrolytes on the sol, the latter should be carefully freed from both these impurities. Hydrogen sulphide itself is liable to cause considerable coagulation if a sol containing it is heated to 180° C., whilst if the hydrogen sulphide is first removed no precipitation is apparent at this temperature. The rapidity of coagulation with potassium and barium chlorides diminishes as the colloid particles increase in size, but the inverse is the case with aluminium and thorium chlorides. The increase of particle size by boiling makes sols more stable at first towards potassium and barium chlorides and then less stable; the stability towards aluminium chloride is practically unaffected. The sols on dialysis become more stable towards potassium chloride and less stable towards barium and aluminium chlorides. A smaller minimum quantity of thorium chloride is necessary for precipitation of a coarse-grained than for a fine-grained sol.

Influence of anion on time of coagulation of As₂S₃ sol.

Electrolyte	a.	b.	Time
(i) Potassium salts:
Chloride.	24	0.0373	7 min. 29 sec.
Sulphate.	22	0.035	25 min. 20 sec.
Oxalate	16	0.049	11 min. 43 sec.
Benzoate.	16	. . .	100 min. 0 sec.
(ii) Potassium salts:
Chloride	14	0.061	8 min. 7 sec.
Sulphate	14	0.053	28 min. 11 sec.
Benzoate	8	. . .	68 min. 0 sec.
Ferrocyanide	8	0.069*	29 min. 59 sec.
(iii) Acids:
HCl	34	0.0275	10 min. 52 sec.
½H2SO4	24.4	0.0271	9 min. 44 sec.
½H2C2O4	8	0.0345	7 min. 2 sec.
CH2Cl.COOH	2	0.075	7 min. 14 sec.
CCl3.COOH	26	0.0361	1 min. 4 sec.

With constant quantities of electrolyte and colloid, the velocity of flocculation diminishes at first as dilution of the electrolyte increases, but tends towards a limit when the dilution reaches a certain value.

In comparing velocities of flocculation by different electrolytes it is therefore necessary to ensure in every case that the electrolyte is sufficiently diluted to give the limiting velocity. When the amount of added electrolyte is varied, but the concentration is kept constant, the velocity of flocculation increases with the amount of electrolyte used. The velocity of flocculation increases as the concentration of the colloid increases with potassium, barium, magnesium and manganese chlorides, but diminishes with aluminium and cadmium chlorides. Moreover, the amount of the electrolyte necessary to cause flocculation varies with the concentration of the colloid and also with the valency of the cation. Thus, with sols containing, respectively, 0.027 and 0.00337 g. As₂S₃ per c.c. the concentration of the electrolyte required to coagulate a given amount of disperse phase, for univalent ions, K⁺, Li⁺ increased with decreasing concentration of the colloid; for bivalent ions, Mg⁺⁺, Ba⁺⁺, the concentration necessary was almost constant and independent of the concentration of the colloid; for tervalent ions, Al⁺⁺⁺, La⁺⁺⁺, the concentration necessary varied almost directly with that of the colloid, and for quadrivalent ions, Zr⁺⁺⁺⁺ and Ce⁺⁺⁺⁺, it decreased much more rapidly than the concentration of the colloid. Hazel and McQueen showed that the position of some ions in the lyotropic series, i.e. the order of precipitating power, was altered in going from high to low concentrations of the sol; thus, for high concentrations the order is Th > Cr > Al > Fe > Ba > K and for low concentrations Th > Cr > Fe > Al > Ba > K. The order of the coagulating power of the alkali sulphates is Cs > Rb > K > Na > Li. Mukherjee and Ganguly found that dilution of arsenious sulphide sols, whether arsenious oxide was present or not, stabilised the sol towards hydrochloric acid or lithium, potassium and barium chorides; but, as regards the last-named, this is not in agreement with the observations of Burton, Dhar and their co-workers. Ghosh and Dhar ascribe the stabilisation towards univalent ions on dilution to hydrolysis and to the peptising effect of the hydrogen sulphide thus formed being greater than the coagulating effect of arsenious acid. Rossi and Marescotti agree that dilution of the sol increases greatly the degree of dispersion of the arsenious sulphide.

The temperature also affects the process of coagulation. With the chlorides of potassium, sodium, lithium and ammonium the velocity of floeculation varies inversely as the temperature; with the chlorides of barium, strontium, calcium, magnesium and cadmium the velocity varies directly as the temperature; with aluminium chloride it is independent of the temperature. Heating thus stabilises the sol towards univalent cations but diminishes the stability towards bivalent ions. The stability is also diminished towards hydrochloric and sulphuric acids. The effects of temperature are small compared with valency and concentration effects, and with aluminium sulphate and thorium nitrate both increase and decrease in stability may be met with, according to the quality of the sol and the concentration of the electrolyte. The influence of temperature depends also on the concentration of the sol when the concentration of the coagulating electrolyte is constant. The addition of the electrolyte lowers the critical temperature of stability, that is, the temperature below which the sol is indefinitely stable, by weakening the repulsive forces.

The age of the sol may also affect the observations, the sol generally becoming less stable on ageing. Thus the precipitation value of barium chloride decreases with time; but ageing may cause either an increase or a decrease of stability towards a particular electrolyte, according as the micelles contain excess of arsenious oxide or of hydrogen sulphide. Ageing, especially under the influence of light, results in a decrease in the amount of disperse phase, an increase in the amount of arsenious acid in the dispersion medium, and the production of colloidal sulphur. These factors influence the stability of the sol to an extent which varies with different electrolytes.

Since the precipitating influence of an electrolyte is mainly determined by the electric charge on the ion with opposite charge to that on the colloid particle, the effect appears to be a consequence of the reduction or elimination of the potential difference between the disperse phase and the medium. If it is assumed that the negative charge on the arsenious sulphide particle is due to adsorption of anions, then the neutralisation of this charge, and consequently precipitation, can be brought about by adsorption of cations. As seen in the table, the hydrogen ion, which is readily adsorbed, has a greater coagulating power than other univalent ions, and this is generally true, any readily adsorbable ion having a lower precipitating value than other ions of the same valency. The organic ions are generally readily adsorbed and their precipitating values, given opposite, for an arsenic trisulphide sol containing 7.54 millimoles per litre should be compared with the values for inorganic ions given in the table.

Moreover, the adsorbability of an ion is generally greater the greater the valency. Matsuno used the precipitating values of cobalt - ammines to determine the valency of the complex ions, employing the equation, deduced from Freundlich's adsorption hypothesis, SN = S/N⁴ where S_N is the equivalent concentration of an N-valent ion, N being the valency of the complex ion, and S the precipitating value of a univalent ion. The results confirmed those obtained by spectroscopic and conductivity methods.

The way in which the precipitating electrolyte is added has a notable effect, and the slower the addition, the longer the time and the greater the quantity of the reagent required for complete precipitation. The hydrosol may thus appear to become "acclimatised" to the coagulant. The first effect of adding the electrolyte is the neutralisation of the colloid charge, and the delay in precipitation on slow addition is due to adsorption of the precipitating ion by the neutralised particles. It was suggested by Mines that the precipitant reverses the sign of part of the disperse phase, which then mutually precipitates with uncharged particles; when slowly added, time is afforded for all the particles to be equally affected and there is therefore no precipitation. According to Ghosh and Dhar, the phenomenon is to be traced to adsorption of ions carrying the same charge as the sol particles. Krestinskaja and Jakovleva studied the effect of slowly adding barium chloride solution to the sol, and concluded that the Ba⁺⁺-ions react with the hydrolysis products of the arsenious sulphide, so quickening the hydrolysis. The hydrolysis products are arsenious acid and hydrogen sulphide, and an outer layer of the latter in the micelle gives a negative charge to it. The barium sulphide formed is adsorbed by the colloid and consequently the Ba⁺⁺- ion concentration of the solution is diminished; thus the critical amount of barium chloride required for coagulation is increased. This explanation is supported by the fact that hydrochloric acid, which does not accelerate the hydrolysis of arsenious sulphide, does not show the "acclimatisation" phenomenon. If this acid is added to the hydrosol in amounts insufficient to cause coagulation, a stabilising or a destabilising effect may result as regards the precipitating action of the hydrochloric acid itself. The effect depends on the quantity added and is accompanied by a change in the degree of dispersion of the sol - an increase or a decrease according as the sol is stabilised or destabilised. When the addition of acids is spread over several days, small amounts being added at intervals, the quantity required to effect coagulation is less than that required when added all at once. This phenomenon, which has been termed "negative acclimatisation," is more marked in dilute solutions. It appears to originate from the checking by the acid of hydrolysis of the sol, and is observed only where adsorption of oppositely charged ions is very high and that of similarly charged ions is negligible. The " positive acclimatisation " described above involves the adsorption of similarly charged ions. The " negative acclimatisation " also occurs when univalent electrolytes such as potassium chloride or nitrate are added to arsenious sulphide sols in the presence of strong acids, the diminution of hydrolysis due to the latter resulting in a decrease of peptisation. Also, smaller quantities of crystal violet, strychnine or quinine hydrochloride are required to coagulate the sol if the electrolyte is added slowly than if rapidly.

The above explanation of " acclimatisation " is supported by the results obtained on adding mixtures of electrolytes to the hydrosol. Hydrolysis of the coagulating electrolyte has a pronounced influence and the presence of one electrolyte may diminish the coagulating power of another. More magnesium chloride is required to coagulate a sol containing lithium chloride than is required in the absence of lithium chloride. Also in the presence of sodium benzoate or sodium nitrite, more than the calculated quantity of potassium or barium chloride is required for precipitation.

An examination of the changes in hydrogen ion concentration during coagulation of the hydrosol led Rabinowitsch to suggest that the latter behaved as a fairly strong complex acid, ionising as follows:

(As₂S₃)_n.SH₂ ⇔ (As₂S₃)_n.SH^- + H⁺ ⇔ (As₂S₃)_n.SH⁼ + 2H⁺

the second ionisation constant being less than the first. The hydrogen ion concentrations of the sol and of the filtrate after precipitation by means of barium chloride were measured and the latter was found to be the greater after correcting for dilution. The increase in acidity rises with concentration of the sol. The coagulum contained Ba⁺⁺-ions but no Cl^--ions. When an arsenious sulphide sol is titrated with barium chloride, the conductivity increases, and the increase is due in part to the presence of the added electrolyte and partly to the liberation of the more mobile H⁺-ions. This increase in acidity appears to be general, and the addition of an electrolyte to the sol causes at first a rapid increase in conductivity, which gradually slows down and becomes linear. The initial increase is ascribed to the displacement of the H⁺-ions by the cations of the electrolyte, a process which ceases as the conductivity curve becomes linear. In order to produce coagulation, a cation must partly or completely displace the H⁺ -ions which are attached to the colloid particles; a certain excess of cations is also necessary, which is the greater the lower the valency of the ion. The adsorption of cations of different valencies by the colloid particles reaches a maximum at approximately the same equivalent concentration, but the quantities adsorbed vary. The process of coagulation thus takes place in two stages: (i) the exchange of added cations with the H⁺-ions of the sol particles; (ii) visible clotting of the particles. The action of dilute solutions of potassium, barium and aluminium chlorides on arsenious sulphide hydrosol enclosed in sealed tubes was observed over a period of 4 to 9 months, and the time noted at which precipitation suddenly increased. The concentration required to cause coagulation was not equivalent to the concentration of H⁺-ion liberated from the sol; the K⁺-ion displaces the H⁺-ion only slowly, while the Al⁺⁺⁺-ion tends to cause coagulation before the exchange adsorption is complete; the displacing power and consequently the precipitating power of the Ba⁺⁺-ion is intermediate between these two. The second stage, the clotting of the particles, is ascribed by Rabinowitsch to electrostatic compression. Weiser and Gray followed the changes in hydrogen ion concentration by means of the glass electrode during the stepwise addition of metallic chlorides to the hydrosol, and showed that the H⁺-ion displacement curve resembles an adsorption curve. The displacement is relatively greater at lower concentrations and reaches a maximum at or below the precipitation value. The total displaced H⁺-ion amounts to 20 to 40 per cent, of the total hydrogen ion concentration of the supernatant liquid after coagulation, the actual proportion depending on the conditions of formation of the sol. The amount of H⁺-ion displaced is less than the amount of precipitating ion adsorbed. The order of displacing power of the chlorides examined was Al > Ba or Sr > Ca > NH₄, which is also the order of their precipitating power and the order of their adsorption below the precipitation value. Ghosh found, however, that the order of the precipitating values of the alkali and alkaline earth metals, determined by measuring the velocity of increase in turbidity, was the inverse of their adsorption values, thus: Li > Na > K > NH₄ > H; and Mg > Ca > Sr > Ba.

The pH value at which an arsenious sulphide sol flocculates infinitely slowly is approximately the same for different strong acids. Thus for a sol containing 1.55 g. As₂S₃ per litre the pH was 1.22, but the value varies with the concentration of the sol. Weak acids fail to cause precipitation of the sol of the concentration mentioned; with more concentrated sols weak acids cause precipitation, but the limiting pH shows a minimum value, less acid being required for very dilute and for very concentrated sols than for sols of intermediate concentrations. On dilution or neutralisation by potassium hydroxide, the pH value of an arsenious sulphide sol varies in a similar manner to the pH under similar conditions in the case of a weak acid, such as acetic acid, except that equilibrium is attained only after 1 to 3 days, indicating an evolution in the structure of the micelles.

Coagulation is supposed to occur when the potential of the electrical double layer is decreased to a certain critical value. The potential is directly proportional to the cataphoretic speed, but Mukherjee and Raichoudhuri assert that there is no critical potential at which coagulation takes place.

Changes in conductivity, pH yalue and stability result on filtering arsenious sulphide sols, probably owing to dissolution of electrolytes from the paper, or to adsorption of H⁺-ions by the paper. Similar changes occur also on dialysis, the conductivity showing at first a sharp decrease, followed by a slow increase probably due to slow ionisation.

When arsenious sulphide is precipitated from a solution containing a barium salt, the adsorbed barium cannot be removed from the precipitate by washing with aqueous sodium or potassium chloride, but if washed with a solution of a tervalent metal, such as iron, aluminium or chromium, interchange of barium with the metal takes place. After precipitating an arsenious sulphide hydrosol with barium chloride, Pauli and Semler found that the precipitate contained 4 equivalents of barium for each equivalent of H⁺-ion found in the sol, and suggested as a possible composition of the colloid, [xAs₂S₃.H₂As₂S₄.HAs₂S₄]H, one hydrogen atom only being ionised in solution, but all four being replaceable on precipitation.

It was observed by Kruyt and van Duin that the addition of an alcohol or phenol to the hydrosol influenced the coagulation by electrolytes, sensitising the sol (i.e. diminishing the limiting concentration of electrolytes necessary for coagulation) towards uni- and ter-valent cations, and stabilising the sol towards bi- and quadri-valent cations. The observations of subsequent workers indicate that the behaviour of the coagulating ion in the presence of a non-electrolyte cannot be predicted from its valency, nor does the adsorbability of the ion run parallel to the tendency to coagulation. Small amounts of alcohols sensitise the sol towards sodium chloride, the effect increasing with increased molecular weight of the alcohol. Higher concentrations caused stabilisation until a maximum was reached, when further alcohol sensitised the sol again. Ethyl alcohol stabilises the sol towards barium chloride over a wide range of concentration. With aluminium chloride, the coagulation concentration diminishes with increasing concentration of alcohol. With sodium sulphate, a large concentration of ethyl alcohol compared with that of the electrolyte sensitises the sol; whilst a little alcohol stabilises the sol towards eerie and thorium chlorides. Methyl alcohol sensitises the sol towards barium chloride, and both alcohols sensitise it towards hydrochloric acid and reduce the cataphoretic velocity. Weak organic acids act in a similar manner to alcohols. The effects appear to be due (i) to a decrease in the dielectric constant of the medium, so that coagulation takes place at a higher particle charge and the sol is sensitised, and (ii) to a change in the interfacial tension, which also affects the potential at which coagulation takes place. That the change in dielectric constant alone is insufficient to account for the influence of non-electrolytes on coagulation is the conclusion arrived at by Bikermann after a study of the coagulation of organosols of arsenic trisulphide by certain electrolytes; the electrokinetic potential at which coagulation occurred was almost independent of the nature of the dispersion medium, and its value did not vary more than 20 per cent, in sols whose dielectric constants showed a 5-fold variation.

When two colloids of opposite electric charges are mixed, mutual precipitation may occur. Thus the gradual addition of ferric hydroxide or aluminium hydroxide hydrosol to arsenic trisulphide hydrosol causes instability in the system, the cataphoretic speed is lowered and a point is reached at which complete precipitation occurs. Mixtures containing higher proportions of the positive sol move towards the cathode during cataphoresis, so that the mobilities of the mixtures lie between those of the pure colloids. Billiter made the following observations of the appearance and behaviour of mixtures of ferric hydroxide and arsenic trisulphide hydrosols:

Behaviour of mixed sols of arsenic trisulphide and ferric hydroxide

Composition of Mixture (mg. In 10 c.c.)		Cataphoresis.	Coagulating Effect.
As2S3	Fe2O3
20.3	0.61	To anode	Opalescence
16.6	6.08	To anode	Immediate precipitation
14.5	9.12	No movement	Complete precipitation
10.4	15.2	To cathode	Immediate precipitation
4.14	24.3	To cathode	Slight opalescence
2.07	27.4	To cathode	No change

The amount of a positive sol necessary for complete precipitation is not the chemical equivalent of the arsenic trisulphide present and the requisite amounts of various positive sols differ considerably. This is shown in the following list of optimum quantities required for the precipitation of a hydrosol containing 24 mg. of arsenic trisulphide:

Positive sol	Fe2O3	ThO2	CeO2	ZrO2	Al2O3	Cr2O3
Milligrams	13	6	4	2	2	0.5

The decreased stability of the system results from mutual adsorption by the two colloids, with consequent unequal redistribution of the total charges round the aggregates; chemical reactions between the stabilising ions do not appear to have an important effect.

The addition of a protective colloid, such as gelatin or agar-agar, to a hydrosol of arsenious sulphide results in stabilisation of the sol towards the precipitative action of electrolytes. If only a trace of gelatin is added, however, precipitation is facilitated, that is, a smaller quantity of the electrolyte effects precipitation. This appears to be due to the interaction of the gelatin with the traces of stabilisers present in the hydrosol, resulting either in adsorption of the stabilisers by the gelatin, or in diminution of the gelatin concentration by the precipitative action of the stabilisers.4 Larger quantities of gelatin hinder precipitation, probably owing to the formation of a film of gelatin particles round the particles of arsenic sulphide, thus preventing their coalescence. The following example illustrates this protective action. To coagulate 300 c.c. of a hydrosol containing 0.2872 g. of arsenious sulphide, 12.87 millimoles of hydrochloric acid were necessary; but after addition of 12 g. of gelatin, 300 millimoles of hydrochloric acid were necessary. If the gelatin is added slowly, drop by drop, to the hydrosol, the latter is less stable towards electrolytes than when the protective colloid is added all at once; this is a parallel of the "acclimatisation" phenomenon. Protective action is also exerted by soaps, and with these it is generally increased by rise in temperature.

On exposure to light, arsenic trisulphide hydrosols show an increase in electrical conductivity, the rate of change increasing with decreasing concentration of the sol.8 The charge on the colloid particle decreases and the conductivity increases with an increase in the period of exposure. After a short exposure the hydrosol is stabilised towards uni- and bi-valent electrolytes, but it becomes unstable on prolonged exposure. According to Joshi and his co-workers, there is no preliminary stabilisation towards magnesium chloride. Ganguly and Dhar observed that the sols coagulated on exposure to tropical sunlight. The hydrosols are photochemically active, the oxidation of certain coloured compounds, such as eosin and malachite green, being sensitised by them in light. Peskoff observed that in the light the addition of anthracene to the sol caused precipitation after a few hours, but in the dark there was no change after 17 days. The increase in electrical conductivity and the photochemical activity are attributed to increased hydrolysis in light of the arsenic trisulphide, free arsenious acid and hydrogen sulphide being formed. The latter is oxidised to sulphur and a thionic acid by the sensitising action of the micelles of arsenic trisulphide. A reaction between the thionic acid and hydrogen sulphide, which provide stabilising ions for the micelles of arsenic trisulphide and sulphur, results in their removal and the destabilisation of the two colloids, which are consequently precipitated. The greater activity of the dilute sol is attributed to greater dispersion. When colloidal sulphur is added to colloidal arsenic trisulphide, the mixture is unstable, although both sols are negatively charged. The particles rapidly increase in size and the precipitation values of hydrochloric acid and aluminium nitrate are less than half the values for the individual sols. The instability is due probably to the reaction between pentathionic acid contained in the sulphur sol with hydrogen sulphide in the arsenic sulphide sol; so that the formation of sulphur micelles in colloidal arsenic trisulphide sols will always tend towards instability.

The earlier stages of the process of coagulation may be studied by measurement of the rate of change of the intensity of the scattered and transmitted light. By this means evidence has been obtained that after addition of traces of the precipitating electrolyte, series of equilibrium states may be set up, so that coagulation occurs by stages. Boutaric and Bouchard have examined the effect of visible and ultraviolet light on the rate of coagulation by electrolytes of arsenious sulphide hydrosols in fluorescent media, such as fluorescein, eosin and erythrosin. In all cases illumination decreased the time required for coagulation, the effect being greater in ultraviolet light than in visible light freed from infrared. Thus the time was lowered by 15 to 30 minutes on exposure to light, and by 35 to 63 minutes by ultraviolet rays. Sulphuric acid and potassium sulphate, which inhibit the fluorescence of the dyes, also suppress the effect of light. The difference, Δt. between the time required for flocculation in the dark and in daylight or ultraviolet, is approximately proportional to the logarithm of the fluorescing power of the mixture of fluorescein and the electrolyte. Sulphuric acid gives the greatest inhibiting effect; lithium chloride has no effect, but tannin, hydroquinone, phenol or cresol, in the presence of lithium chloride, reduce both the fluorescing power and Δt. Eosin and erythrosin, which have absorption bands close to that of fluorescein, reduce the fluorescence of the latter and also reduce its effect on the flocculation. The addition of sucrose and glycerol to an arsenic trisulphide sol containing fluorescein and potassium chloride reduces the fluorescence by increasing the viscosity of the sol, and the effect on the flocculation time is consequently diminished.

Exposure to light, including ultraviolet and infra-red, has little effect either on the cataphoretic speed or the rate of coagulation of arsenious sulphide sols. The absorption spectrum of the hydrosol has been studied and shows that there is simple absorption which increases with the size of the particles, and also a selective absorption in the region 6200 A. which is due to reflection of the incident rays from the surfaces of the colloid particles, and which decreases with increase in the size of the particles. The product of the refractive index and the specific volume is a linear function of the concentration.

If a mixture containing sols of silver and arsenic trisulphide is kept in the dark, the colour changes from golden-brown through greenish- brown to lilac, whilst in the light the colour goes through green to golden- yellow. The former change appears to be due to direct interaction between the sol particles, but the latter involves oxygen, with the probable formation of a silver thioarsenite. Both changes are prevented by the addition of gelatin gel.

Periodic coagulation of arsenic trisulphide has been effected by diffusing a solution of ferric chloride or of aluminium sulphate into the sol contained in an agar gel.

Many attempts have been made to elucidate the constitution of the colloidal aggregates in the arsenic trisulphide hydrosols. Linder and Picton observed that the sols contained an excess of hydrogen sulphide which was not removable by hydrogen, and considered this to be an essential constituent of the colloid, to which they assigned the formula 8As₂S₃.H₂S. When coagulated by an electrolyte some of the metal is carried down with the precipitate, replacing the hydrogen, thus -

mAs₂S₃.H₂S + 2KCl ⇔ mAs₂S₃.K₂S + 2HCl

but Pauli and Semler found that four equivalents of the metal were removed for each equivalent of H⁺-ion present in the sol and gave the formula [xAs₂S₃.H₂As₂S₄.HAs₂S₄]H, while Rabinowitsch considered it to be a fairly strong acid which he formulated (As₂S₃)_n.SH₂. Bhatnagar and Rao stated that when hydrogen sulphide was removed from the sol the composition was more nearly (AsS)_x, and that if the red colloidal solution is heated it yields the yellow sol and a precipitate of sulphur, the change involving oxygen, thus:

As₂S₂.xH₂S + xO = As₂S₃ + xH₂O + (x – 1)S

If the hydrogen sulphide content is small, no sulphur is precipitated. These authors considered it probable that the red variety was identical in properties with realgar and the yellow with orpiment, and that the action of light and heat consists mainly in the transformation of one variety into the other, thus:

As₂S₂ + H₂S + O = As₂S₃ + H₂O

Chaudhury and his co-workers showed that the composition of the hydrosol varied according to the method of its formation, but that arsenious acid was usually present. This they considered to be a normal constituent of the sol, and the many irregularities observed in different sols were attributed to interaction of H + -ions and polythionic acids present. The constitution of a sol containing excess of arsenious acid was given as As₂S₃, but in a sol containing no free arsenious acid or hydrogen sulphide the probable composition was thought to be As₂S₃.H₂S or As₂S₃.As₂S₅. Murphy and Mathews suggested that a complex compound of the type As₂O₃(As₂S₃)_n existed in the solution; Gazzi, however, maintained that all the arsenious oxide may be removed by dialysis, and that the oxide and sulphide are present as a mixture.

Reversal of the electric Charge on the colloid particle may be accomplished by the adsorption of a large quantity of positive ions, and by employing a thorium salt a positively charged arsenic trisulphide sol has been obtained.

The addition of alkali to the arsenic trisulphide hydrosol causes a solvent action on the sol particles, the rate of dissolution depending on the alkali used and increasing in the order Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺. The addition of ammonium hydroxide, however, greatly decreases the rate of dissolution. The coagulated sulphide is dissolved by alkalis in exactly the same order as above, so that the action is manifest with all degrees of dispersion.

Various organosols of arsenic trisulphide have been described. Bikermann prepared stable sols in nitrobenzene and ethyl acetoacetate by passing dry H₂S through solutions of arsenic trichloride in the anhydrous liquids, the excess of hydrogen sulphide and the hydrogen chloride formed being removed by a current of dry air. Concentrations up to 29 millimoles As₂S₃ per litre were obtained. The sols were precipitated by dissolved salts at a definite potential which was practically independent of the solvent and the concentration of the sol; also the valency rule as to precipitating power held as for hydrosols. Concentrated sols in pure or aqueous glycerine, preferably in the presence of a protective colloid, are obtained by the action of hydrogen sulphide on arsenious oxide dissolved in the medium.

Stable sols of arsenic trisulphide in concentrated acids, including sulphuric, phosphoric, acetic and trichloracetic acids, have been prepared. In such sols coagulation occurs if a certain degree of ionisation of the acid is reached.