of

THE

CHEMICAL NEWS

AND

JOURNAL OF PHYSICAL SCIENCE

Proprietor and Editor,

Sir Wm. Crookes, O.M., F.R.S..

୮

Established in the Year 1859.

(WITH WHICH IS INCORPORATED THE "CHEMICAL GAZETTE "). Published Weekly. Annual Subscription, free by post 1 Entered at the New York Post Office as Second Class Mail Matter. Transmissible through the Post-United Kingdom, at Newspaper rate; Canada and Newfoundland, at Magazine rate.

LEAD ASHES, SULPHATE OF LEAD, LEAD SLAGS, ANTIMONIAL LEAD, COPPER MATTE, TIN ASHES, &c., ORES, DROSS, or RESIDUES containing TIN, COPPER, LEAD, and ANTIMONY.

NEW TYPE
HIGH EFFICIENCY
OZONE GENERATOR
FOR RESEARCH.

Particulars from

OZONAIR LIMITED,

96, Victoria Street, London, S.W. 1.

SULPHUROUS

ACID and SULPHITES.

Liquid SO2 in Syphons, for Lectures, &c.

PHOSPHORIC

ACID and PHOSPHATES.

CARAMELS & COLORINGS

for all purposes.

Registered as PRICE 4d.
a Newspaper. LPOST FREE 4§d.

H. K. LEWIS & CO., Ltd.,

BOOKSELLERS,

SCIENTIFIC & TECHNICAL.

ORDERS by POST or TELEPHONE promptly executed.

Telegrams:

"PUBLICAVIT, EUSROAD," 136, Gower St.,

LONDON."

Telephone:
MUSEUM 1072.

Now Ready.

London, W.C.1.

TABLES FOR MIXED ACIDS. ABC FIVE-FIGURE LOGARITHMS and Tables for

CHEMISTS

By C. J. WOODWARD, B.Sc.

Logarithms of Numbers with Thumb Lateral Index - Instructions for using the Tables, with Examples and Exercises-Hydrometer Tables -Electrochemical Equivalents-Gas Correction or Nitrometer Tables from 680 to 799 mm. pressure and 7° to 23° C.-Thermometer Equivalents, -99° F. to +2990° F. in °C. and vice versa-Weight and Measure Conversion, Imperial to Metric and vice versa-Factors and Logarithms of Factors for use in Gravimetric Analysis. All tables involving atomic weights calculated from the International Atomic Weights of 1917.

LONDON:

E. & F. N. SPON, 57, Haymarket.

T. TYRER & CO., Ltd. stirilag Chemical Works,

STRATFORD, LONDON, E. CHEMICALS for Analysis, Research, and Technical purposes. TECHNICAL EXPERIMENTS carried out

on a MANUFACTURING SCALE. Inquiries solicited from INVENTORS ard FATENTEES (especially Foreign) under the New Act, 1906.

CHEMICAL APPARATUS
AND REAGENTS.

Pure Chemicals for Research Work.

A. BOAKE, ROBERTS, & CO. (LIMITED), JOHN J. CRIFFIN & SONS, LTD.,

Stratford, London, E.

KINGSWAY, LONDON, W.C

Feb., 1918

THE IDENTIFICATION AND ESTIMATION OF and NH3 fumes, even in traces. The standards, therefore,

to filter the water without adding any acid into another 100 cc. cylinder. The results are expressed as Pb in solution+Pb in suspension = total Pb.

The sample must be examined for colouring matter against distilled water, which can be done with one of the above portions. The colouring matter present may usually be estimated as compared with cc. lead solution in distilled water required to match it. The colour of many waters may be matched in this manner. The cc. lead solution used to produce equality of tint is deducted from the total cc. lead solution required in the Pb estima tion. This will only be necessary when the water colour is high and distilled water used for the standard. In some cases it is not possible to match the natural colour of the water by this means, and other substances such as caramel solution or peat extract used instead.

It is essential to test the glass bottle containing the sample for lead. This can only be done by emptying the contents and washing well with a bottle brush, and finally cleaning with a 1 per cent solution acetic acid and well washing with distilled water. Fill the bottle with 500 cc. distilled and 10 cc. strong acetic and allow to stand twenty-four hours and test for Pb. When the bottle cannot be emptied in this way the test may be made by cleaning the outside of bottle and immersing the base in a basin with the acidulated water and performing the test. The application of this test is imperative when the nature of the glass is unknown.

The Blank Experiment.-It is essential, more especially with new apparatus, to test the graduated cylinders, pipettes, and reagents for lead. This is best done by washing out all the cylinders, pipettes, and test glasses with HCl, tap and then distilled, and filling all with distilled and adding 1 cc. 20 per cent acetic and well mixing with glass rod mixers, and allowing to stand for an hour with the glass rods in the cylinders. Now add 1 cc. H2S and compare with cylinder full of distilled water only. The blank is considered satisfactory if no trace of coloration develops in two minutes.

be reduced to equality in colouring by lowering one cylinder or the other, and the respective values found. It is well to remember that during the final coloration measurement both columns ought to be as near as possible of the same height, otherwise low or high results will ensue. There ought not to be more than 10 cc. difference between the columns at most, and if more than this a little make-up water ought to be added. The point to remember is that the lead ought to be estimated at the actual concentration present in the water, and no dilution method ought to be used, but the estimation made with shorter, say 50 cc. cylinders, instead.

The Colour Measurement.-The tint estimations are made near a well-lighted window or by means of oil, gas, or electric light. A good white light is necessary. That from an ordinary cycle actylene lamp, or good oil lamp, are both satisfactory. Gas mantle jets also give good results. Metal filament and carbon film electric lamps require screening with oil paper, or sand blasted or opaque glass globes, otherwise reflection effects cause trouble. It is a great convenience to have a lamp which can be lowered, or moved towards the observer at will and where the candle-power can be increased if desired, or lessened. A good acetylene light or powerful oil lamp answers all requirements. Such a mode of illumination will be found most useful when the lead exceeds 1 part per 100,000, as the density of tint may require powerful illumination unless dilution is resorted to. Again, when traces of lead are only present the illumination requires to be subdued.

The test mixers or cylinders are covered with two or three folds of stout blotting paper, which are fixed with rubber bands, leaving the bottom uncovered. The cylinders must always be kept dry on outside, and the paper jacket also, otherwise reflection effects will obscure the reading. By using these paper jackets, brighter, stronger, and purer tints are obtained, which are essential for making a true colour comparison. Unless the cylinders are so screened, colour comparisons in the majority of waters become impossible, unless tintometers or colour meters are used. The paper jackets ought to be continued above the 100 cc. mark.

The glasses are placed on the white tile and during colour comparison are raised about one to two inches, with the light adjusted to fall on the plate, and the candle-power increased according as the depth of tint increases. Under these conditions, I part Pb per 4 million parts may be detected and estimated with a 7-inch or 100 cc. column. This is equal to 0025 part per 100,000 or 0.0175 grain Pb per gallon, which is the limit of the test. With 200 cc. or 14 inch columns I part Pb in 8 million parts may be detected, or 00125 part per 100,000, or 0.0067 grain per gallon, which is more sensitive than the chromate test.

The Process.-100 cc. of the sample which is suspected to contain Pb is placed in one of the cylinders and I cc. acid added, and well mixed and put aside for two minutes. I cc. H2S is added and again well mixed. This is compared with another 100 cc. of the sample or distilled water without any reagents being added. The least trace of coloration will indicate the presence of either copper or lead. The presence of copper is detected by testing another portion of the sample with ammonia, and if present the lead must be identified and estimated by the chromate method to be hereafter described. As the ferrocyanide test for copper also gives a reaction with zinc and iron and lead, brown, white, and blue tints respectively result, which, if all be present, will, more or less, make the test very indefinite. It is therefore better to use ammonia for identifying the copper. It is also well to remember that zinc in acetic acid solution is also precipitated by H2S, which causes a turbidity when present to the extent of I in 100,000 or more. The zinc, therefore, only becomes troublesome by obscuring the brightness of the lead sulphide coloration. Should this obscurity persist by the presence of zinc the lead will require to be esti-throught the tube. In these cases the paper jackets must mated by the chromate method. However, by adding only I cc. H2S or o 75 cc. no turbidity will result from zinc, even when present to the extent of 1 in 100.000.

The lead is estimated by comparing the intensity of coloration resulting from the addition of H2S to the leadcontaminated water, with a sample of the same water but free from lead, to which is added lead standard solution to produce equal intensity of coloration. The most convenient and reliable way to do this is to take 100 cc. of the sample, add acid and H2S. and compare with 100 cc. of the lead-free sample, to which is added 1 cc. acid, and lead solution till equally approximate tints are obtained. Then take a fresh 100 cc. and add to this the cc. lead solution found as above, add acid and H2S, and compare. A slight difference in tint should only result, which may

When the Pb present exceeds o 75 part per 100,000, on looking down the tube a semi black reflecting surface only is visible, which is almost opaque to ordinary daylight. At this and greater Pb concentrations vertical colour comparisons cannot be made with the full 100 cc. column. In such cases it is advisable to work with 50, 40, or 25 cc. columns, measuring on a vertical line of sight, or better still, to measure the tint on a horizontal line of sight right

be dispensed with. Even when the Pb is present at the rate of 5 parts per 100,000 the horizontal method may be used, but using carbon estimation tubes in place of the cylinders. By this means the estimations may be conducted without resorting to dilution of the sample.

Summary. Due to the colouring matter in the water sample, and also its saline constituents and other unknown factors, no estimation of Pb in water by the H2S process can be considered satisfactory unless a sample of the leadfree sample water is used as standard. As previously stated, the reason of this is that different waters with the same Pb contents give with H2S variable intensity of tint, amounting in some cases to 100 per cent. When distilled water is used for the standard, the lead is likely to be underestimated by 25 to 33 per cent at least, if not more.

THE SETTING OF CEMENT IN ITS RELATION, the trough with the fin dipping in the plastic cement. As TO ENGINEERING STRUCTURES.

By BERTRAM BLOUNT.

THE question of the setting of cements of all sorts has been a subject of investigation for many years, and except in the simple case of plaster of Paris remains unsettled. As there are many contributors to this discussion who have devoted great attention and much work to the endeavour to decide the true causes of setting, and as at this meeting they will formulate their views, I propose to deal chiefly with the practical aspect as it appears to the engineer.

Before doing this, however, I should like to say a word or two on the present state of the theory of the subject. In the first place the term cement is elastic, and it is wrong to suppose that colloids like casein should be excluded. Their behaviour may throw light on the mechanism of setting of more important cements. Next how is one to define setting? Generally it involves bydration, but is the induration due to loss of water, as in the case of clay, to be excluded? I take it that, broadly, setting does not connote hydration, and with this narrower definition we must be content. Taking the typical case of plaster of Paris, it has been established that a supersaturated solution is formed, that interlocking crystals are deposited from it, and that the released water passes on to perform the same function with adjacent particles. This naturally leads to a consideration of whether strontium sulphate and barium sulphate set in the same way. The low solubility of bo.h, especially the latter, would seem at first glance to negative the supposition, but as both are found crystalline in nature the question is not superflucus. Coming to industrial cements of the Portland cement class the question is much more difficult. I am not aware that any definite proof has been given that the constituents of Portlant cement form supersaturated solutions and set in the manner accepted as true for plaster of Paris. A fair indication of this is found in the fact that there are two schools of thought concerning setting, the one crystalline, the other colloidal. I profess myself strongly neutral, as evidence on each side appears inconclusive, but it may well be that in this discussion the whole matter will be cleared up by those who have had more recent oppor tunities of experiment than I.

Such theoretical considerations lie at the very root of the matter, and when they have been determined will be as complete a guide to practice as the principles of mechanics are to the design of a bridge. Until this end has been accomplished we must be content with humbler empirical methods.

As setting is a gradual process there is no such thing in reality as a setting point or setting time. But in practice it is convenient to fix an arbitrary condition which is called the "setting point," and to call the time occupied in reaching this condition the "setting time." With minor cements, such as those of the plaster of Paris class, it is not usual to define the condition or the time rigidly. It is generally sufficient to recognise that ordinary plaster of Paris "goes off" quickly and needs rapid handling, and that other calcium sulphate cements setting more slowly can be manipulated in a more leisurely way and require a considerable time to harden. With the more important cements, of which Portland cement is the chief instance, something less rough and ready is required. In consequence numerous methods have been devised and put into tentative use. One of the most mechanically attractive was that of my friend Prof. John Goodman, who in 1887, as far as I remember, constructed a machine consisting of a metal trough on which ran a loaded roller pulled by clock work and having a central fin. The trough was filled with cement and the ends of the roller rolled on the edges of

A contribution to a General Discussion on "The Setting of Cements and Plasters," held by the Faraday Society, January 14, 1918.

the cement set, and the roadbed became harder, the fin with its roller was gradually lifted upwards and eventually rode on the surface of the set cement. The rate and time of setting were automatically recorded on a diagram drawn by a pencil passing through the centre of the roller. Certainly a neat idea, but the difficulty common to all methods of determining setting time came in. There was no definite point of time when the fin on the roller ceased to make an impression, and thus the accuracy aimed at by this ingenious machine was greater than the nature of the material allowed. I hope that Prof. Goodman will correct me if I am wrong, as I write from memory.

In practice the only apparatus of any value for determining the setting time of cement is the Vicat needleor one of its modifications, such as the British standard needle formerly used-which is too well known to need description. By simple means of this kind cements can be classified quite as closely as their nature allows, having regard to the fact that the setting of cement is not like the setting point of a metal-and that too is a progessive phenomenon - but a gradual process.

Even with this modest standard of attainment the determination of the setting point of cement is beset with pitfalls. Cement is scarcely ever used neat, and yet its setting point is invariably determined neat, and I for one see no means of using any other method., The best that can be hoped for is that a cement with a given setting time will behave with some sort of relation thereto when it is made into mortar or concrete. But consider the difficulties here. An ordinary Portland cement will require about 22 per cent of water to make a plastic mass. When mixed with standard sand in the proportion of 1:3 it will need about 10 per cent of water, reckoned on the mixture, in order to make a plastic mass, i.e., roughly double the amount used when it is neat. The mechanism of setting will be very different under these different conditions, but there is more than that. The aggregate, sand, stones, and the like, will be machanically in the way of progressive hydration; that, however, is a trifle compared with the thermal effect. The setting of cement generates heat, which in large masses and with quick-setting cements may cause a considerable rise in temperature. Indeed, the method of measuring rise of temperature has been sug gested as a mode of determining the setting time-chiefly in Germany. Anything more lunatic is difficult to conceive. Where the cement is slow setting and the proportion of aggregate is considerable, say 6 or 8 to 1, as t generally is, the rise of temperature is not large unless the cement is so unsound that it contains what is conveniently termed "free lime." That idea may be dismissed with regard to cement made under modern conditions. What is of more moment is the influence of temperature on the rate of setting of cement. If the temperature is high the rate of setting is correspondingly high and may be too high for satisfactory construction.

Cements used in Madras at a tropical temperature and in winter in Canada, obviously are being employed under very different conditions, and it is the business of the cement manufacturer to provide a suitable material which shall act not as it does in a temperate climate, but in the place and under the conditions of actual use; he must be a little proleptic. I admit that the task is a hard one; proleptical tasks mostly are, but that the difficulties can be overcome is shown pretty clearly by results. When not only ordinary concrete but ferro-concrete has to be considered, the difficulties with regard to setting increase rapidly.

A 30-ton block may be made for harbour work and allowed to lie in the yard for six months, and if it does not set hard at once no one is much the worse, provided that the cement and aggregate are of good quality.

With ferro-concrete the matter is different. A reasonable time must be allowed for such things as piles, as they are not put immediately to use and may lie in the yard, but the ordinary ferro-concrete structure is mono