men in treatment of those under your charge-with fairness, frankness, firmness. Most men will go a long way to meet you when they think you are square, and will respond to frankness. For example, we check all our conductors every month. We tell them they are going to be checked and explain why-so that we can know they are all right as the handlers of money. They don't know when they are being checked but they can always see the results afterwards. They understand it and no longer worry about spotters spying on them. It is all right when they understand they are simply audited unawares or without knowing just when, just as a bank teller or a post office clerk is.

Remember that it is up to you to set an example of and to teach sportsmanship to those who haven't had your training and advantages.

We haven't needed any for a long time, but we used to have a number of so-called grievance arbitration cases. In one of those a certain minister of the highest character was agreed on by both sides as arbitrator. There was really no doubt about the guilt of the man whose case was in dispute, and the arbitrator so decided. But the head of the guilty man's organisation could see only his own side-to protect cne of his constituents and in mailing to the arbitrator a check for his organisation's half of the fee, he enclosed with it a note reading, "God help you as a minister!" Wel!, that showed bad sportsmanship; but he didn't know any better. Afterwards this labour leader learned better and expressed regret for some things he had said and done.

So you must have patience with these things, continually watch that your skirts are clear, and results will come.

own

The most successful and most lasting results come where the dealing is square. That's where, for example, the shining reputation of the A., T. & S. F. Ry. organisation comes from-the outstanding sportsmanship (fairplay, courage, imperturbability) of the late E. P. Riley, impressed on his associates and carried on down through the entire organisation.

So what I preach to you is to carry into your business and professional and public. careers the spirit of sportsmanship which has become second-nature to you in your games. No matter how foggy the conditions are, stick to the rules. If you find a

HYDROLYSIS OF ALUMINUM SALTS. By Wilder D. Bancroft.

re

(From the Journal of Physical Chemistry, June, 1922.)

Over a century ago, Gay-Lussac' found that if one heats a solution of aluminum acetate, it soon becomes turbid and a large amount of alumina precipitates. If the solution is allowed to cool, the precipitate dissolves slowly and becomes transparent. On the reheating and recooling these changes repeat themselves, and this can be kept up indefinitely. With a dilute solution of aluminum acetate the turbidity begins at about 50° and a precipitate forms at a little higher temperature. The precipitate must change gradually because it dissolves more slowly, when the solution is cooled, the longer the heating has lasted. With a more concentrated solution of aluminum acetate the temperature must be raised somewhat higher before turbidity occurs; but this solution also clears up when cooled.

"To determine the amount of alumina precipitated from an acetate solution by heating and the variation with the temperature, there were taken two equal portions of aluminum acetate made by mixing in the cold two solutions of alum and of lead acetate. One of these portions was raised to the boiling-point and filtered once, while the other portion was precipitated by ammonia. The two precipitates were washed and dried; after which it appeared that the first weighed about half as much as the second.

Annales de Chimie, 74, 193 (1810).

at

A

"These observations may be very important for the makers of dyed fabrics, for they use the hot solutions of alum and lead acetate in order to get as concentrated solutions of the mordant as possible. great deal of alumina must precipitate, and the loss will be considerable if the solution is filtered at once. To avoid this it is necessary to let the solution cool completely before filtering or decanting off the mother liquor, and it is also necessary to stir vigorously so as to be certain that ail the alumina redissolves. Unless these precautions are taken, the aluminum acetate will be very acid, which is probably the reason for usually adding chalk. It is nevertheless easy to prevent the precipitation of alumina when an aluminium acetate solution is heated, by adding alum. As is well known, alum dissolves alumina and therefore keeps the solution of aluminum acetate from becoming turbid. A large excess of acid will accomplish the same result."

"The precipitation of alumina on boiling and the redissolving at a lower temperature are facts which are of interest to the general theory of chemistry, and which are rather exceptional. If the precipitation were due to volatilization of the acetic acid, the alumina would not redissolve when the temperature is lowered. As a matter of fact similar changes can ie observed in a strongly acid solution or in hermetically sealed flasks. Since the precipitation does not depend on the volatilization of the acid, it is evident that it is due to the heat which wrenches apart the molecules of acid and alumina, carrying each out of the sphere of action of the other, and causing their separation. With less heat the same molecules come within each other's sphere of activity and combine."

Gal-Lussac is practically saying in other words what we now designate as reversible hydrolysis. In 1854 Crum' prepared colloidal alumina from a weaker and more basic solution than that used by GayLussac. "By the continued action of heat on a weak solution of binacetate of alumina, Al,O,(CH,CO2),(OH),, a permanent separation of the constituents of the salt takes place, although no acid escapes and no alumina is precipitated. The properties of the alumina are at the same time materially changed. A solution of binace

'Jour. Chem. Soc., 6, 225 (1854).

tate of alumina diluted so as to contain not more than one part of alumina in two hundred of water, was placed in a closed vessel which was immersed to the neck in boiling water, and kept in that state day and night for ten days. It had then nearly lost the astringent taste of alum, and acquired the taste of acetic acid. Being afterwards boiled in an open capsule, acetic acid was freely given off, and when the boiling had continued about five hours (the loss of water being continually restored), the liquid was found to have retained not more than one-eleventh of its original quantity of acetic acid, or about one equivalent to five and a half of alumina."

The theory of this is very simple. Water will hydrolyze any salt until the product of the concentrations of the hydrogen and hydroxyl ions reaches a value of about 10 14. If either base or acid is very sparingly soluble, the hydrolysis will run farther than if both are strong electrolytes. Equilibrium will be reached much rapidly if the solution is heated. Whether the insoluble base precipitates or remains in colloidal solution will depend on the conditions of the experiment. That hydrolysis has taken place can be shown in a number

more

of different ways. Gay-Lussac deduced it

from seeing the precipitated alumina, and Crum from the change in the taste of the solution. One could measure the change in acidity in other ways than by the sense of taste. Debray showed that if a dilute solution of ferric chloride is heated to 70° it no longer reacts with potassium ferrocyanide to form Prussian blue. The reason for this is that there is no more ferric salt in solution, it having been converted completely into colloidal ferric oxide. Colloidal ferric oxide is not blackened by hydrogen sulphide. On the other hand it is precipitated from apparent solution either by sodium sulphate or by sulphuric acid, a behaviour which is distinctly not characteristic of ferric chloride. Wiedemann has used the magnetic properties as a means of following the hydrolysis, because the atomic magnetism of the iron in colloidal ferric oxide is only one-fifth that of the iron in a strongly acid solution of ferric chloride. Colloidal solutions of the hydrous oxides can be obtained by hydrolysis of the

acetates, nitrates, or chlorides; but not in general by hydrolysis of sulphates, because sulphuric acid precipitates the colloidal solutions more readily than does hydrochloric, nitric, or acetic acid.

In 1883 Liechti and Suida published some work on the hydrolysis of solutions of aluminum salts. Their original paper is not in the John Crerar Library, the Library of Congress, or the Library of the Franklin Institute. I have therefore made use of the lengthy abstract by J. J. Hummel.' With aluminum sulphate (Al(SO,),.18H2O, 200 g per litre) there was no visible hydrolysis on heating or on diluting. With Al2(SO4)2(OH), made from 200 ge Al2(SO), .18H2O+31.82 g Na, CO, per litre, there was no visible change on heating; but a precipitate formed when the solution was diluted fourteenfold. With a solution threequarters as concentrated, there was no visible change on heating, but dissociation took place on diluting tenfold. Since threefourths of fourteen is ten and a half, this shows the error in determining the beginning of visible hydrolysis. With time. The general results are that hydrolysis takes place more readily on heating and AL(SO)(OH), made from 200 g Al(SO,),.18H,0+45.7g NaHCO_per litre a jelly was formed on heating and a precipitate remained on cooling. Diluting to one-half caused a precipitate to appear in the cold. With AL(SO)(OH),, made from 300 g Al(SO,),.18H,0-151.3g NaHCO per litre, the solution kept only a short on diluting, the more basic the solutions. The statement is made that sodium sulphate accelerates the dissociation. If this is true, it must be because the sodium sulphate decreases the hydrogen ion concentration by reacting with the sulphuric acid to form acid sodium sulphate. Since sodium sulphate coagulates colloidal alumina, it may be that it has practically no effect on the hydrolysis; but causes the precipitate to become visible sooner. In other places Liechti and Schwitzer say that they have proved that sodium sulphate retards the decomposition. There must be a misprint somewhere. A retarding of the decomposition would be more in line with the observation by Schmid that sodium sulphate retards the precipitation of alumina by sodium carbonate.

With what they call a sulpho-acetate, AL(SO)(CH,CO), made from 200

'Mitt. techn. Gewerbe-Museums in Wien, 2Sektion für Färberei, 3, 59, 60 (1886). 'Jour. Soc. Chem. Ind., 14, 654 (1895).

Al(SO),.18H,0 + 227.6 g Pb(CH,CO2)2 per litre, a precipitate formed on heating to 90° and a jelly on heating to 100°. The solution cleared up on cooling. There was no precipitate on diluting sixty-fold. On adding sodium bicarbonate to this so-called sulpho-acetate, Liechti and Suida obtained solutions of what they called basic sulphoacetates, from Al2(SO)(CH,CO2),(OH) down to Al,(SO1)(CH ̧CO2)(OH),.

was

There

is no real reason for assigning these formulas to these solutions. In the first place there is no obvious reason why the sodium bicarbonate should take acetate rather than sulphate out of the aluminum salt. It would be distinctly more reasonable to consider the basic compounds as varying from Al(SO ̧)(CH,CO2)(OH), to AL(CH,CO,), (OH),. Probably the reason for not doing this was that these solutions did not behave like basic acetate solutions obtained by adding sodium bicarbonate to aluminum acetate; but that does not prove anything because these latter solutions do not contain sodium sulphate. In the second place there is no reason to suppose that Liechti and Suida ever had a sulpho-acetate solution. The so-called sulpho-acetate solution undoubtedly merely a mixture of 1, Al2(SO), and /, Al,(CH,CO,). Similarly there is no satisfactory evidence of the existence of any of the alleged basic compounds. They may perfectly well have been aluminum sulphate or aluminum acetate or a varying mixture of both with peptized alumina. Liechti and Suida undoubtedly thought they were dealing with true basic salts because of their getting an apparent solution; but it is quite evident, even from the abstract, that they knew nothing about colloidal solutions. The formulas which are written therefore indicate the relative amounts of sodium bicarbonate which have been added to the original solution and nothing more than that. This point seems to have been overlooked by all the people who have discussed this work.

3

2

The so-called basic sulpho-acetates all formed jellies on heating and all formed precipitates on dilution, the temperature at which the first turbidity occurred being lower the more basic the solution. Less water had to be added to cause precipitation as the solution became more basic. With straight aluminum acetate, made from 200 g Al(SO,),.18H2O + 341.4 g Ph(CH,CO), per litre, there was no turbidity on heating and no precipitation on cooling. This result does not contradict that of Gay-Lussac because he was work

ing with a more dilute solution. All the socalled basic aluminum acetates formed precipitates or jellies on heating which did not dissolve when the solutions were cooled. The one with the alleged formula Al2(CH,O2) (OH), precipitated on dilution; but the more basic ones did not. This may only be a question of a time factor. If not, sodium acetate must peptize peptize alumina strongly.

To a

By adding sodium carbonate to aluminum chloride, solutions were obtained having the analytical compositions, Al,Cl,(OH), ALCI (OH)2, Al2Cl, (OH)3, A12Cl2(OH),. None of these solutions become turbid either on heating or on diluting. Leichti and Suida were not able, however, to make these alleged solutions synthetically. "The solubility of aluminum hydrate in aluminum chloride was tested, and the following remarkable results were obtained. solution of AlCl a quantity of aluminum hydrate was added sufficient to form the basic salt Al,Cl,(OH)2. The alumina dissolved only on heating, and the solution remained clear on cooling. To this clear solution a further quantity of aluminum hydrate was added, sufficient to form the compound Al,Cl,(OH),. It was, however, found that no more alumina could be made to dissolve, the precipitate even increasing, and on filtering it was found that the solution contained equal molecules of normal AI,CI and of HCI. The nascent ALCI,(OH), had apparently decomposed according to the following formula: 7 AICI,(OH), + 2 H20 =

5 Al(OH) + 2 Al2CI + HCl. In the same way it was proved that, on adding aluminum hydrate to Al2(SO), solution sufficient to produce the basic no alumina at compound Al(SO,),(OĦ), all dissolved, the filtrate only containing Al,(SO1). Here, too, we must suppose that the nascent basic compound decomposes as follows:

AL(SO), (OH) = Al(OH) + AL(SO1),.'

We know that hydrous ferric oxide, if present in excess, will remove from suspension all the chromic oxide' peptized by caustic alkali; but nobody has studied the removal from suspension of a substance by an excess of itself, unless perhaps we have such a case in the Bayer process for the purification of alumina. Since Liechti and Suida were obsessed with the idea of basic compounds, they made no experiments to see whether addition of sodium chloride would cause more alumina to go into apNagel: Jour. Phys. Chem., 19, 331 (1915).

parent solution. According to Schmid' sodium sulphate retards the precipitation of alumina.

A whole series of so-called basic nitrate solutions were made up by adding sodium bicarbonate to aluminum nitrate solutions. None of these solutions became turbid either on heating or on dilution. The sum total of all these experiments is that basic aluminum acetate solutions hydrolyze very readily because the resulting acetic acid is a weak acid; that basic aluminum sulphate solutions become turbid because of the coagulating action of the sulphates, this more than counterbalancing the effect due to the strength of sulphuric acid; and that basic aluminum chlorides and nitrates do not become markedly turbid because the acids are strong ones having low coagulating powers.

Although Liechti and Suida got no cloudiness with their aluminum sulphate solution on heating, that may have been a question of concentration or of time, for Naumann' reports that "when a solution of potash alum is heated to the boiling point of water, a white precipitate is formed, which, after washing with water, is an amorphous powder, with an admixture of glittering lamina, and dissolves with difficulty in strong hydrochloric acid, but easily in potash. The precipitate contains 31.2-32.6 per cent. of alumina, about 11 per cent. of potash, 30-40 per cent. of sulphuric acid, and water; and is therefore a more or less basic compound of alumina, potash, and sulphuric acid, with water. It was found that with pure alum solutions the decomposition was most rapid at first, gradually becoming less for equal intervals of time, so that a state of equilibrium in the liquid was reached only after a very long time. Dilution of the solutions favoured decomposition. Free sulphur c acid, added to alum solutions, prevented the decomposition, partially or entirely, according to the amount added. Neutral potassium sulphate, on the contrary, expedited the decomposition."

There is a good deal of uncertainty as to what compounds are formed when salts of with aluminum are boiled or are treated soda. Cajar states that the action of soda on cold aluminum sulphate solution gives Covers well; a white precipitate which

1 Jour. Soc. Chem. Ind., 14, 654 (1895). 2 Ber. deutsch chem. Ges., 8, 1639 (1875); Jour. Chem. Soc., 29, 682 (1876).

3 Zeit. angew. Chem., 27, 793 (1911).

whereas a more transparent precipitate is obtained when the soda is added to a hot sulphate solution. Underwood' states that alumina precipitated cold with sodium carbonate contains basic sulphate and dries soft and powdery; but that it dries horny when precipitated hot. Jennison says that the hydrate of aluminum, Al(OH), "is produced when caustic soda or potash, ammonium hydrate or carbonated alkalies is added to a solution of an aluminum salt; it is soluble in caustic alkalies, and therefore is usually obtained by means of the carbonates. When produced from cold, dilute solutions, it is of a transparent, gelatinous nature; but, on heating, becomes opaque and more contracted in bulk. With carbonated alkalies it is must denser and is very lumpy, owing to the reaction being incomplete. Aluminum hydrate, when dried, forms a hard, white, horny substance, which has the composition of Al(OH), and only on ignition is the whole of the water driven off, leaving Al,O,."

According to Knecht, Rawson and Loewenthal a crystallised basic sulphate, Al,(SO)(OH)2, has been placed on the market by Messrs. Peter Spence and Co., Ltd. Schlumberger claims that there is one basic sulphate of aluminum. On adding 4.5 mols of caustic potash to a solution containing one mol of aluminum sulphate, the supernatant liquid was still distinctly acid. On adding 5 mols KOH the solution was neutral and contained no aluminum salt. On adding 6 mols KOH the supernatant liquid was alkaline and contained some alumina, presumably as potassium aluminate. The precipitate must therefore analyse for (Al,O,),SO,xH2O where x is not less than six. Actually it came out seven, so Schlumberger writes the formula of the basic salt (Al,O H ),.H2SO.. Of course, it will be noticed that he analysed a precipitate obtained only under one set of conditions. While it may be that he was dealing with a definite compound, there is nothing in his experiments to prove it. He should have analysed the precipitate after adding three or four mols of caustic potash and have shown that it had exactly this

1Underwood and Sullivan: “Printing Inks," 81 (1915).

2" The Manufacture of Lake Pigments from Artificial Colours," 51 (1900). 3Bull. Soc. chim. Paris, [3] 13, 41 (1895). 4" A Manual of Dyeing," 225 (1910).

same composition. Then his results could have been accepted as proving something. This is the more necessary because many basic aluminum sulphates are to be found in the books. 1 Böttinger contributes a new one of his own by evaporating repeatedly to dryness a mixture of aluminum sulphate and sodium chloride at 130°-140°. This one analyses for Al,O,.SO,.6H,O or Al2O H H2SO,.2H,O if we adopt Schlumberger's way of writing the formula.

ADSORPTION OF ALUMINUM SULPHATE BY WOOL.

Since the different salts of aluminum hydrolyze of themselves under suitable conditions, they should hydrolyze even more readily in presence of a textile fibre which adsorbs the alumina and they do as a matter of fact. In so far as the acid radical is also adsorbed, there may be minor complications and the matter may seem less simple than it really is.

With

When wool is treated with aluminum sulphate, Al(SO,),.18H,O, less than about five per cent. on the wool, the bath is exhausted completely, all the alumina and all the sulphuric acid being taken up.' higher concentrations, more and more aluminum sulphate is left in the bath. There has been some discussion whether the wool takes up aluminum sulphate or alumina and sulphuric acid in the cases when the bath is exhausted completely. When wool mordanted in this way is boiled with water, the wash water is always acid, sulphuric acid being removed slowly and alumina left behind. This has been considered by von Georgievics as a proof that the sulphuric acid is free. While this may be true, it does not follow, because the adsorbed aluminum sulphate might be hydrolyzing. It is much wiser to admit frankly that we have no way at present of deciding this point. In case the acid is not combined in definite proportions, the next problem is whether it is adsorbed by the alumina, by the wool, or by the two in some unspecified ratio.