# 1911 Encyclopædia Britannica/Cement

CEMENT (from Lat. caementum, rough pieces of stone, a shortened form of caedimentum, from caedere, to cut), apparently first used of a mixture of broken stone, tiles, &c., with some binding material, and hence of any material capable of adhering to, and uniting into a coherent mass, fragments of a substance not in itself adhesive. The term is often applied to adhesive mixtures employed to unite objects or parts of objects (see below), but in engineering, when used without qualification, it means Portland cement, its modifications and congeners; these are all hydraulic cements, i.e. when set they resist the action of water, and can, under favourable conditions, be allowed to set under water.

Hydraulic Cements.—It was well known to builders in the earliest historic times that certain limes would, when set, resist the action of water, i.e. were hydraulic; it was also known that this property could be conferred on ordinary lime by admixture of silicious materials such as pozzuolana or tufa. We have here the two classes into which hydraulic cements are divided.

When pure chalk or limestone is “burned,” i.e. heated in a kiln until its carbonic acid has been driven off, it yields pure lime. This slakes violently with water, giving slaked lime, which can be made into a smooth paste with water Pozzuolanic cement. and mixed with sand to form common mortar. The setting of the mortar is due to the drying of the lime (a purely physical phenomenon, no chemical action occurring between the lime and the sand). The function of the sand is simply that of a diluent to prevent undue shrinkage and cracking in drying. Subsequent hardening of the mortar is caused by the gradual absorption of carbonic acid from the air by the lime, a skin of carbonate of lime being formed; but the action is superficial. Mortar made from pure or “fat” lime cannot withstand the action of water, and is only used for work done above water-level. If, however, such “fat” lime is mixed in the presence of water, not with sand but with silica in an active form, i.e. amorphous and (generally) hydrated, or with a silicate containing silica in an active condition, it will unite with the silica and form a silicate of lime capable of resisting the action of water. The mixture of the lime and active silica or silicate is a pozzuolanic cement. The simplest of all pozzuolanic cements would be a mixture of pure lime and hydrated silica, but though the latter is prepared artificially for various purposes, it is too expensive to be used as a cement material. A similar obstacle lies in the way of using a certain native form of active silica, viz. kieselguhr, for it is too valuable as an absorbent of nitroglycerine, for the manufacture of dynamite, to be available for making pozzuolanic cement. There are, however, many silicious substances occurring abundantly in nature which can thus be used. They are mostly of volcanic origin, and include pumice, tufa, santorin earth, trass and pozzuolana itself. The following analyses show their general composition:—

 Neapolitan Pozzuo- lana (per cent) Roman Pozzuo- lana (per cent) Trass (per cent) Soluble silica (SiO2) 27.80 32.64 19.32 Insoluble silicious residue 35.38 25.94 50.40 Alumina (Al2O3) ${\displaystyle \scriptstyle {\left.{\begin{matrix}\ \\\ \end{matrix}}\right\}\,}}$ 19.80 22.74 ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \\\ \end{matrix}}\right.}}$ 13.86 Ferric oxide (Fe2O3) 3.10 Lime (CaO) 5.68 4.06 · · Magnesia (MgO) 0.35 1.37 0.13 Sulphuric anhydride (SO3) Trace Trace · · Combined water (H2O) ${\displaystyle \scriptstyle {\left.{\begin{matrix}\ \\\ \end{matrix}}\right\}\,}}$ 4.27 8.92 ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \\\ \end{matrix}}\right.}}$ 7.57 Carbonic anhydride (CO2) · · Moisture · · · · 5.04 Alkalis and loss 6.72 4.33 0.58 100.00 100.00 100.0

An artificial product which serves perfectly as a pozzuolana is granulated blast-furnace slag. The slag, which must contain a high percentage of lime, is granulated by being run while fused into abundance of water. This granulated slag differs from the same slag allowed to cool slowly, in that a portion of the energy which it possesses while fused is retained after it has solidified. It bears to ordinary slowly-cooled slag a similar relation to that borne by plastic sulphur to ordinary crystalline sulphur. This potential energy becomes kinetic when the slag is brought into contact with lime in the presence of water, and causes the formation of a true hydraulic silicate of lime. The following analysis shows the composition of a typical slag:—

 Per Cent. Insoluble residue 1.04 Silica (SiO2) 31.50 Alumina (Al2O3) 18.56 Manganous oxide (MnO) 0.44 Lime (CaO) 42.22 Magnesia (MgO) 3.18 Soda (Na2O) 0.70 Sulphuric anhydride (SO3) 0.45 Sulphur (S) 2.21 ——— 100.30 Deduct oxygen equivalent to sulphur 1.10 ——— 99.20

Granulated slag of this character is ground with slaked lime until both materials are in a state of fine division and intimately mixed. The usual proportions are three of slag to one of slaked lime by weight. The product termed slag cement sets slowly, but ultimately attains a strength scarcely inferior to that of Portland cement. Although it is cheap and suitable for many purposes, its use is not large and tends to decrease. Pozzuolanic cements are little used in England. Generally speaking, they are only of local importance, their cheapness depending largely on the nearness and abundance of some suitable volcanic deposit of the trass or tufa class. They are not usually manufactured by the careful grinding together of the pozzuolana and the lime, but are mixed roughly, a great excess of pozzuolana being employed. This excess does no harm, for that part which fails to unite with the lime serves as a diluent, much as does sand in mortar. In fact, ordinary pozzuolanic cement made on the spot where it is to be used may be regarded as a better kind of common mortar having hydraulic qualities. Good hydraulic mortars may be made from lime mixed with furnace ashes or burnt clay as the pozzuolanic constituent.

Cements of the Portland type differ in kind from those of the pozzuolanic class; they are not mechanical mixtures of lime and active silica ready to unite under suitable conditions, but consist of definite chemical compounds of lime andPortland Cement. silica and lime and alumina, which, when mixed with water, combine therewith, forming crystalline substances of great mechanical strength, and capable of adhering firmly to clean inert material, such as stone and sand. They are made by heating to a high temperature an intimate mixture of a calcareous substance and an argillaceous substance. The commonest of such substances in England are chalk and clay, but where local conditions demand it, limestone, marl, shale, slag or any similar material may be used, provided that the correct proportions of lime, silica and alumina are maintained. The earliest forms of cements of the Portland class were the hydraulic limes. These are still largely used, and are prepared by burning limestones containing clayey matter. Some of these naturally possess a composition differing but little from that of the mixture of raw materials artificially prepared for the manufacture of Portland cement itself. Although hydraulic limes have been in use from the most ancient times, their true nature and the reason of their resistance to water have only become known since 1791. Next in antiquity to hydraulic lime is Roman cement, prepared by heating an indurated marl occurring naturally in nodules. Its name must not be taken to imply that it was used by the ancients; in point of fact the manufacture of this substance dates back only to 1796.

With the growth of engineering in the early part of the 19th century arose a great demand for hydraulic cement. The supply of materials containing naturally suitable proportions of calcium carbonate and clay being limited, attempts were made to produce artificial mixtures which would serve a similar end. Among those who experimented in this direction was Joseph Aspdin, of Leeds, who added clay to finely ground limestone, calcined the mixture, and ground the product, which he called Portland cement. The only connexion between Portland cement and the place Portland is that the cement when set somewhat resembles Portland stone in colour. True, it is possible to manufacture Portland cement from Portland stone (after adding a suitable quantity of clay), but this is merely because Portland stone is substantially carbonate of lime; any other limestone would serve equally well. Although Portland cement is later in date than either Roman cement or hydraulic lime, yet on account of its greater industrial importance, and of the fact that, being an artificial product, it is of approximately uniform composition and properties, it may conveniently be treated of first. The greater part of the Portland cement made in England is manufactured on the Thames and Medway. The materials are chalk and Medway mud; in a few works the latter is replaced by gault.

The composition of typical samples of chalk and clay is shown in the following analyses:—

The composition of typical samples of chalk and clay is shown in the following analyses:—

 Chalk. Per cent. Silica (SiO2) 0.92 Alumina + ferric oxide (Al2O2 + Fe2O3) 0.24 Lime (CaO) 55.00 Magnesia (MgO) 0.36 Carbonic anhydride (CO2) 43.40 ——— 99.92
 Clay. Per cent. Insoluble silicious matter 26.67 Consisting of Silica (SiO2) 31.24 Quartz (SiO2) 19.33 Alumina (Al2O3) 16.60 Silica (SiO2) 5.19 Feldspar   7.34% Ferric Oxide (Fe2O3) 8.66 Alumina (Al2O3) 1.47 Lime (CaO) 0.25 Magnesia (MgO) 0.03 Magnesia (MgO) 1.91 Soda (Na20) 0.65 Soda (Na2O) 1.00 ——— Potash (K2O) 0.45 26.67 Sodium Chloride (NaCl) 1.86 Combined water, organic matter, and loss 11.36 ——— 100.00
 Fig. 1.
 Fig. 2.

Many different forms of kiln are used for burning Portland cement. Besides the chamber kilns which have been described, there are the old-fashioned bottle kilns, which are similar to the chamber kilns, but are bottle-shaped and openOther kilns. at the top; they do not dry the slurry for their next charge. Their use is becoming obsolete. There are also stage kilns of the Dietzsch type, which consist of two vertical shafts, one above the other, but not in the same vertical line, connected by a horizontal channel. At this middle portion and in the upper part of the lower shaft the burning proper proceeds; the upper shaft is full of unburnt raw material which is heated by the hot gases coming from the burning zone, and the lower shaft contains clinker already burned and hot enough to heat the incoming air which supplies that necessary for combustion at the clinkering zone. A pair of Dietzsch kilns, built back to back, are shown in fig. 2. There are other forms of shaft kiln, such as the Schneider, in which there is a burning zone, a heating and cooling zone as in the Dietzsch, but no horizontal stage, the whole shaft being in the same vertical plane. Another form is the Hoffmann or ring kiln, made up of a number of compartments arranged in a ring and connected with a central chimney; in these compartments rough brick-shaped masses of the raw materials are stacked, and between these bricks fuel is sprinkled. At a given moment one of these compartments is burning and at its full temperature; the air for combustion is drawn in through one or more compartments behind it which have just finished burning, and is thereby strongly heated; the products of combustion pass away through one or more compartments in front of it and heat their contents before they are subjected to actual combustion. It will be seen that the principle of the ring kiln is similar to that of the stage kiln. In each case the clinker which has just been burned and is fully hot serves to heat the air-supply to the compartment where combustion is actually proceeding; in like manner the raw materials about to be burned are well heated by the waste gases from the compartment in full activity before they themselves are burned. (It may be noted that here and generally in this article “burn” is used in the technical sense; it is technically correct to speak of cement clinker being “burned”, although it is not a fuel; in accurate terms it is the fuel which is burned, and it is the heat it generates which raises the clinker to a high temperature, i.e. technically “burns” it.) By this device a great part of the heat is regenerated and a saving of fuel is effected.

The methods of burning cement described above are obsolescent. They are being replaced by the rotatory process, so called because the cement is burned in rotating cylinders instead of in fixed kilns. These cylinders vary from 60 to 150 ft. in length, an ordinary length in modern practice being 100 Rotatory kilns. to 120 ft.; their diameter correspondingly varies from 6 ft. to 7 ft. 6 in. The cylinders are made of steel plate, lined with refractory bricks, are carried on rollers at a slight angle with the horizontal, and are rotated by power. At the upper end the raw material is fed in either as a dry powder or as a slurry; at the lower end is a powerful burner. In the early days of rotatory kilns producer gas was used as a fuel, but with little success; about 1895 petroleum was used in the United States with complete success, but at a relatively heavy cost. At the present time, finely powdered coal injected by a blast of air is almost universally employed, petroleum being used only where it is actually cheaper than coal. In the working of this type of kiln the rotation and slight inclination of the cylinder cause the raw material to descend towards the lower end. At the upper end the raw material is dried and heated moderately. As it descends it reaches a part of the kiln where the temperature is higher; here the carbonic acid of the carbonate of lime, and the combined water of the clay are driven off, and the resulting lime begins to act chemically on the dehydrated clay. The material is then in a partially burnt and slightly sintered state, but it is not fully clinkered and would not make Portland cement. The material continues to descend by the rotation of the kiln and reaches the lower end nearest the burner where the temperature is highest, and is there heated so highly that the union of the lime, silica and alumina is complete, and fully burnt clinker falls out of the kiln. It is extremely hot, and is cooled usually by being passed down one or more rotating cylinders, similar to the first, but smaller, and acting as coolers instead of kilns. On its way down the cylinders the clinker meets a current of cold air and is cooled, the air being correspondingly warmed and passing on to aid in the combustion of the fuel used in heating the kiln. This regenerative heating is similar in principle and effect to that obtained by means of the shaft and ring kilns described above. The output of these kilns varies from 200 to 400 tons per kiln per week according to their size and the nature of the raw materials burned, as against 30 tons per week for an ordinary chamber kiln. A large saving in labour is also secured. The rotatory system presents many advantages and is rapidly replacing the older methods of cement making. Fig. 3 represents diagrammatically a rotatory cement plant on the Hurry & Seaman system, which was one of the first to make cement by the rotatory process successfully on a large scale, using powdered coal as fuel. Rotatory kilns of various other makes are now in use, but the same principles are embodied, namely, the employment of a rotating inclined cylinder for burning the raw materials, a burner fed with powdered coal and a blast of air, and some device such as a cooling cylinder or cooling tower by which the clinker may be cooled and the air correspondingly heated on its way to the burner.

 Fig. 3.

Another method of making Portland cement which has been proposed and tried with some success consists in fusing the raw materials together in an apparatus of the type of a blast furnace. The high temperature necessary to fuse cement clinker makes this process difficult to accomplish commercially, but it has many inherent merits and may be the process of the future, displacing the rotatory method.

The composition of Portland cement varies within comparatively narrow limits, and for given raw materials the variations are tending Composition. to become smaller as regularity and skill in manufacture increase. The following analysis may be taken as typical of cements made from chalk and clay on the Thames and Medway:—

 Per cent. Silica (SiO2) 22.0 Insoluble residue 1.0 Alumina (Al2O3) 7.5 Ferric oxide (Fe2O3) 3.5 Lime (CaO) 62.0 Magnesia (MgO) 1.0 Sulphuric anhydride (SO3) 1.5 Carbonic anhydride (CO2) 0.5 Water (H2O) 0.5 Alkalis 0.5 ——— 100.0

There may be variations from this composition according to the nature of the raw materials employed. Thus the silica may range from 19 to 27%, the alumina and ferric oxide jointly from 7 to 14%, the lime from 60 to 67%. All such variations are permissible provided that the quantity of silica and alumina is sufficient to saturate the whole of the lime and to leave none of it in a “free” condition, likely to cause the cement to expand after setting. Other things being equal, the higher the percentage of lime within the limits indicated above the stronger is the cement, but such highly limed cement is less easy to burn than cement containing about 62% of lime; and unless the burning is thorough and the raw materials are intimately mixed, the cement is apt to be unsound. Although the ultimate composition of cement, that is, the percentage of each base and acid present, can be accurately determined by analysis, its proximate composition, i.e. the nature and amount of the compounds formed from these acids and bases, can only be ascertained indirectly and with difficulty. The foundations of our knowledge on this subject were laid by H. le Chatelier, whose work has since been supplemented by that of Spenser B. Newberry, W. B. Newberry and Clifford Richardson. As the outcome of these inquiries it has been established that tricalcium silicate 3CaO·SiO2 is the essential constituent of Portland cement. The constituent of next importance is an aluminate, but whether this is dicalcium aluminate, 2CaO·Al2O3, or tricalcium aluminate, 3CaO·Al2O3, is still in doubt. In the following description it is assumed to be the tricalcium aluminate. The remaining silicates and aluminates present, and ferric oxide and magnesia, if existing in the moderate quantities which are usual in Portland cement of good quality, are of minor importance and may be regarded as little more than impurities. The silicates and aluminates of which Portland cement is composed are believed to exist not as individual units but as solid solutions of each other, these solid solutions taking the form of minerals recognizable as individuals. The two principal minerals are termed alite and celite; according to the best opinion, alite consists of a solid solution of tricalcium aluminate in tricalcium silicate, and celite of a solid solution of dicalcium aluminate in dicalcium silicate. Celite is little affected by water, and has but small influence on the setting; alite is decomposed and hydrated, this action constituting the main part of the setting of Portland cement. Both the components of alite react, and for simplicity their reactions may be stated in separate equations, thus:—

(1) 2(3CaO·SiO2) + 9H2O = 2(CaO·SiO2)·5H2O + 4Ca(OH)2
Tricalcium silicate. Hydrated mono- Calcium
calcium silicate. hydroxide.

(2) 3CaO·Al2O3 + 12H2O = 3CaO·Al2O3·12H2O
Tricalcium aluminate. Hydrated tricalcium aluminate.

Since alite is a solid solution and, although an individual mineral, is not a chemical unit, the proportion of tricalcium silicate to tricalcium aluminate in a given specimen of alite will vary; but, whatever the proportions, each of these substances will react in its characteristic manner according to the equations given above.

The precise mechanism of the process of setting of Portland cement is not known with certainty, but it is probably analogous to that of the setting of plaster of Paris, consisting in the dissolution of the compounds produced by hydration while they are in a more soluble form, their transition to a less soluble form, the consequent supersaturation of the solution, and the deposition of the surplus of the dissolved substance in crystals which interlock and form a coherent mass. This theory being accepted, it is evident that a small quantity of water, by successive dissolution and deposition of a substance capable of existing in a more soluble and in a less soluble form, is able to bring about the crystallization of an indefinitely large quantity of material. It is not necessary that there should be present sufficient water to dissolve the whole of the reacting substance at any one time; it is sufficient if there is enough for hydration and a small surplus for the crystallization by successive stages as above described. It is generally admitted that the aluminate is the chief agent in the first setting of the cement, and that its ultimate hardening and attainment of strength are due to the tricalcium silicate.

As mentioned above, the constituents other than the tricalcium silicate and tricalcium aluminate of which alite is composed, are of minor importance. The function of the ferric oxide present in ordinary cement is little more than that of a flux to aid the union of silica, alumina and lime in the clinker; its role in the setting of the cement is altogether secondary. In fact, excellent Portland cement can be prepared from materials free from iron. Such cement, if free also from manganese, is white, and its manufacture has been proposed for exterior decorative use. Magnesia, if present in Portland cement in quantity not exceeding 5%, appears to be inert, but there is evidence that in larger proportion, e.g. 10-15%, it may hydrate and set after the general setting of the cement, and may give rise to disruptive strains causing the cement to “blow” and fail. In so-called natural cement which is comparatively lightly burnt, the magnesia appears to be inert, and as much as 20 to 30% may be present. Another constituent of Portland cement which influences its setting time is calcium sulphate, naturally formed from the sulphur in the raw materials or fuel, or intentionally added to the finished cement as gypsum or plaster of Paris. It has a remarkable retarding effect on the hydration of the calcium aluminate, and consequently on the setting of the cement; thus it is that a little gypsum is often added to convert a naturally quick-setting cement into one which sets slowly. It will be observed that in the hydration of tricalcium silicate, the main constituent of Portland cement, a large portion of the lime appears as calcium hydroxide, i.e. slaked lime. It is evident that this will form a pozzuolanic cement if a suitable silicious material such as trass is added to the cement. The ultimate product when set may be regarded as a mixed Portland and pozzuolanic cement. The use of trass in this manner as an adjunct to Portland cement has been advocated by W. Michaelis, and undoubtedly increases the strength of the material, but it has not become general.

The quality of Portland cement is ascertained by its analysis and by determining its specific gravity, fineness, mechanical strength and soundness. A good sample will usually have a composition within the limits cited above and approximating Testing. to the typical figures given above. It will be ground so finely that not more than 3% will be left on a sieve of 76 × 76 meshes per sq. in., the wires of the sieve being 0.005 in. in diameter. It will have, when freshly burned, a specific gravity not lower than 3.15, and briquettes made from it and kept in water will possess a tensile strength of 400-500 ℔ per sq. in. seven days after they are made, while briquettes made from a mixture of 3 parts by weight of sand and 1 of cement will give about 225 ℔ per sq. in. at twenty-eight days. Formerly the soundness of cement was determined by keeping thin pats of the cement in cold water for twenty-eight days, or in warm water (110°-120° F.) for twenty-four hours, and examining for cracks or other signs of expansion. Modern practice is to measure the expansion of a test piece of cement kept in water at a temperature of 212° F. The simplest and most generally used method is due to H. L. le Châtelier, and consists in measuring the increase in circumference of a cylinder of cement 30 mm. in diameter by means of a split ring encircling the cylinder, the motion of which is magnified by two light rods extending radially. Another quantitative test for soundness is that formulated by L. Deval, who has shown that briquettes of 3 of sand and 1 of cement kept in water for two days at 80° C. = 176° F. attain approximately the same strength as similar briquettes attain at seven days in water at the ordinary temperature. In like manner briquettes kept at 176° F. for seven days are approximately equal in strength to those kept at the ordinary temperature for twenty-eight days. A cement not perfectly sound will give low results in the hot test, and a cement of indifferent soundness will crack and go to pieces. The test is admittedly severe, but can be passed without difficulty by cement made with proper care and skill. There are many modifications and elaborations of all the tests which have been mentioned. Cement for all important work is submitted to a rigorous system of testing and analysis before it is accepted and used.

Hydraulic Lime is a cement of the Portland as distinct from the pozzuolanic class. The most typical hydraulic lime is that known as Chaux du Theil, made from a limestone found at Ardèche in France. This limestone consists of calcium carbonate most intimately intermixed with very finely divided silica. It contains but little alumina and oxide of iron, which are the constituents generally necessary to bring about the union of silica and lime to form a cement, but in spite of this the silica is so finely divided and so well distributed that it unites readily with the lime when the limestone is burned at a sufficiently high temperature. English hydraulic limes are of a different class; they contain a good deal of alumina and ferric oxide, and in composition resemble somewhat irregular Portland cement.

Analyses of the two classes of hydraulic lime are as follows:—

 Chaux de Theil. Per cent. Blue Lias. Per cent. Insoluble silicious matter 0.3 2.39 Silica (SiO2) 21.7 14.17 Alumina (Al2O3) 1.8 6.79 Ferric oxide (Fe2O3) 0.6 2.34 Lime (CaO) 74.0 63.43 Magnesia (MgO) 0.7 1.54 Sulphuric anhydride (SO3) 0.3 1.63 Carbonic anhydride (CO2) } 0.6 3.64 Water (H20) 2.69 Alkalis and loss · · 1.38 ——— ——— 100.0 100.00

Hydraulic lime contains a good deal of uncombined lime, and has to be slaked before it is used as a cement. In France this slaking is conducted systematically by the makers, the freshly burned lime being sprinkled with water and stored in large bins where slaking proceeds slowly and regularly until the whole of the surplus uncombined lime is slaked and rendered harmless, while the cementitious compounds, notably tricalcium silicate, remain untouched. In English practice hydraulic lime is slaked by the user. Seeing that regular and perfect slaking is more easily attained when working systematically on a large scale and by storing the material for a long period, the French method is the better and more rational. The product may then be regarded as a cement of the Portland class mixed with slaked lime. When gauged with water and made into a mortar it sets slowly, but ultimately becomes almost as strong as Portland cement. Its slow setting is an advantage for some purposes, e.g. for foundations and abutments where settlements may occur. The structure is free to take its permanent position before the lime sets, and cracks are thus avoided. A case in point is the employment of hydraulic lime in place of Portland cement as grouting outside the cast-iron tubes used for lining tunnels made by the shield system.

Roman Cement is another cement of the Portland class which came into use shortly before the manufacture of artificial Portland cement was attempted. It is still in use, though only for special purposes where a quick-setting material is required. It is made from septaria nodules which are dredged up on the Kent and Essex coasts and consist of about 60% of calcium carbonate mixed with clay, the mass being sufficiently indurated to remain coherent under water. The nodules are not prepared in any way, but simply burned at a moderate red heat.

The resulting cement varies somewhat in composition, but approximates to the following figures:—

 Per cent. Insoluble silicious matter 5.86 Silica (SiO2) 19.62 Alumina (Al203) 10.30 Ferric oxide (Fe2O3) 7.44 Manganese dioxide (MnO2) 1.57 Lime (CaO) 44.54 Magnesia (MgO) 2.92 Sulphuric anhydride (SO3) 2.61 Carbonic anhydride (CO2) 3.43 Water (H2O) 0.25 Alkalis and loss 1.46 ——— 100.00

The most characteristic constituent is the oxide of iron, which gives the cement a reddish colour, and the presence of manganese also differentiates Roman from Portland cement, which rarely contains appreciable quantities of that element. The high percentage of alumina causes the cement to be quick-setting, and it becomes hard in about five minutes. It resists the action of water, salt or fresh, very well, and is therefore useful in situations where the work is likely to be submerged immediately after it has been put in place.

The term Natural Cements is applied to cements made by burning mixtures of clay and carbonate of lime naturally occurring in approximately suitable proportions. They may be regarded as badly-mixed Portland cements, and need no special description. American “natural” cements are of a somewhat different class. They are usually made from a silicious limestone containing magnesia, and are comparatively lightly burned.

The following analysis is typical of a cement of this kind:—

 Per cent. Silica (SiO2) 24.30 Alumina (Al203) 7.22 Ferric oxide (Fe2O3) 5.06 Lime (CaO) 33.70 Magnesia (MgO) 20.94 Water, carbonic anhydride, and loss 8.78 ——— 100.00

These irregular cements of the Portland class are good building materials for ordinary purposes, but are not so suitable as good artificial Portland cement for heavy and important undertakings.

Good hydraulic cements are highly permanent materials provided certain conditions be observed. It might be supposed that hydraulic cements from their nature would be indifferent to the action of water, but this is only true if the structures of which they form part are sufficiently compact. In this case the action of the water is checked by the film of carbonate of lime which eventually forms oh the surface of calcareous cement. This, together with the compactness of the mortar, hinders the ingress and egress of water, and prevents the dissolution and ultimate destruction of the cement. But where the concrete or mortar is not well made and is porous, the continual passage of water through it will gradually break up and dissolve away the calcareous constituents of the cement until its strength is utterly destroyed. This destructive action is increased if the water contains sulphates or magnesium salts, both of which act chemically on the calcareous constituents of the cement. As sea-water contains both sulphates and magnesium salts, it is especially necessary in concrete for harbour work to take every care to produce an impervious structure. There are various minor external causes for the failure and ultimate destruction of cement mortar and concrete, but their discussion is a matter for the specialist. Failure from inherent vice in the cement has been already touched on; it can always be traced to want of skill and care in manufacture.

Calcium Sulphate Cements.—Under this term are comprehended all cements whose setting properties primarily depend on the hydration of calcium sulphate. They include plaster of Paris, Keene’s cement and many variants of these two types. The raw material is gypsum (q.v.). This may be almost chemically pure, when it is generally used for Keene’s cement; or it may contain smaller or greater quantities of impurities, in which case it is suitable for the preparation of cements of the plaster of Paris class. The mode of preparation is to calcine the gypsum at temperatures which depend on the class of cement to be produced. If plaster of Paris is to be made, calcination is carried out at about 204° C. (= 400° F.); at this temperature, gypsum, CaS04·2H20, loses three-quarters of its combined water and becomes 2CaSO4·H20. If a cement of the Keene’s cement class is to be prepared the temperature used is higher, e.g. 500° C. (= 932° F.), and the whole of the combined water of the gypsum is expelled, the anhydrous sulphate CaSO4 being obtained.

To produce plaster of Paris European practice consists in baking the mineral in ovens, and in America in heating it in kettles. Both processes are inferior in economy to calcination in rotatory kilns, a process which may be regarded as the method of Plaster of Paris; Keene’s cement. the present and the immediate future. Keene’s cement and its congeners are made in fixed kilns so constructed that only the gaseous products of combustion come into contact with the gypsum to be burnt, in order to avoid contamination with the ash of the fuel.

The setting of plaster of Paris depends on the fact that when 2CaSO4·H2O is treated with water it dissolves, forming a supersaturated solution of CaSO4·2H2O. The excess held temporarily in solution is then deposited in crystals of CaSO4·2H2O. In the light of this knowledge the mode of setting of plaster of Paris becomes clear. The plaster is mixed with a quantity of water sufficient to make it into a smooth paste; this quantity of water is quite insufficient to dissolve the whole of it, but it dissolves a small part, and gives a supersaturated solution of CaSO4·2H2O. In a few minutes the surplus hydrated calcium sulphate is deposited from the solution, and the water is capable again of dissolving 2CaSO4·H2O, which in turn is fully hydrated and deposited as CaSO4·2H2O. The process goes on until a relatively small quantity of water has by instalments dissolved and hydrated the 2CaSO4·H2O, and has deposited CaSO4·2H2O in felted crystals forming a solid mass well cemented together. The setting is rapid, occupying only a few minutes, and is accompanied by a considerable expansion of the mass. There is reason to suppose that the change described takes place in two stages, the gypsum first forming orthorhombic crystals and then crystallizing in the monosymmetric system. Gypsum thus crystallized is in its normal monosymmetric form, more stable under ordinary conditions than the orthorhombic form. Correlatively in its process of dehydration to form plaster of Paris, monosymmetric gypsum is converted into the orthorhombic form before it begins to be dehydrated.

The principles which govern the preparation and setting of the other class of calcium sulphate cements, that is, cements of the Keene class, are not fully understood, but there is a fair amount of knowledge on the subject, both empirical and scientific. The essential difference between the setting of Keene’s cement and that of plaster of Paris is that the former takes place much more slowly, occupying hours instead of minutes, and the considerable heating and expansion which characterize the setting of plaster of Paris are much less marked.

It is the practice in Great Britain to burn pure gypsum at a low temperature so as to convert it into the hydrate 2CaSO4·H2O, to soak the lumps in a solution of alum or of aluminium sulphate, and to recalcine them at about 500° C. On grinding they give Keene’s cement. Instead of alum various other salts, e.g. borax, may be used. The quantity of these materials is so small that analyses of Keene’s cement show it to be almost pure anhydrous calcium sulphate, and make it difficult to explain what, if any, influence these minute amounts of alum and the like can exert on the setting of the cement. It seems probable that the effect of the salts is inconsiderable, and that the governing condition is the temperature at which the cement has been burnt. The setting of Keene’s cement takes place by the same sort of process which has been described for the setting of plaster of Paris, the chief differences being that the substance dissolved is anhydrous calcium sulphate and that the operation takes a longer time.

All cements having calcium sulphate as their base are suitable only for indoor work because of the solubility of this substance. They form excellent decorative plasters on account of their clean white colour and the sharpness of castings made from them, this latter quantity being due to their expansion when setting.

See D. B. Butler, Portland Cement (London, 1905); E. C. Eckel, Cements, Limes and Plasters (New York, 1905); G. R. Redgrave and Charles Spackman, Calcareous Cements (London, 1905); F. H. Lewis, “Manufacture of Hydraulic Cements in the United States,” The Mineral Industry (New York, 1898); W. H. Stanger and Bertram Blount, “Cement Manufacture in Great Britain,” The Mineral Industry, New York, 1897 and 1905; Id. “The Testing of Hydraulic Cements,” Journ. Soc. Chem. Ind., 1894, 13, p. 455; Id., Proc. Inst. Civ. Eng., 1901; B. Blount, “Recent Progress in the Cement Industry,” Journ. Soc. Chem. Ind., 1906, 25, p. 1020; H. L. le Chatelier, Recherrhes experimenlales sur la constitution des mortiers hydrauliques; Desch, Concrete, No. 2, pp. 101-102; Davis, Journ. Soc. Chem. Ind., 1905, 26, p. 727.  (B. Bl.)

Adhesive Cements.—Mixtures of animal, vegetable and mineral substances are employed in great variety in the arts for making joints, mending broken china and other objects, &c. A strong cement for alabaster and marble, which sets in a day, may be prepared by mixing 12 parts of Portland cement, 8 of fine sand and 1 of infusorial earth, and making them into a thick paste with silicate of soda; the object to be cemented need not be heated. For stone, marble, and earthenware a strong cement, insoluble in water, can be made as follows:—skimmed-milk cheese is boiled in water till of a gluey consistency, washed, kneaded well in cold water, and incorporated with quicklime; the composition is warmed for use. A similar cement is a mixture of dried fresh curd with 110th of its weight of quicklime and a little camphor; it is made into a paste with water when employed. A cement for Derbyshire spar and china, &c., is composed of 7 parts of rosin and 1 of wax, with a little plaster of Paris; a small quantity only should be applied to the surfaces to be united, for, as a general rule, the thinner the stratum of a cement, the more powerful its action. Quicklime mixed with white of egg, hardened Canada balsam, and thick copal or mastic varnish are also useful for cementing broken china, which should be warmed before their application. For small articles, shellac dissolved in spirits of wine is a very convenient cement. Cements such as marine glue are solutions of shellac, india-rubber or asphaltum in benzene or naphtha. For use with wood which is exposed to moisture, as in the case of wooden cisterns, a mixture may be made of 4 parts of linseed oil boiled with litharge, and 8 parts of melted glue; other strong cements for the same purpose are prepared by softening gelatine in cold water and dissolving it by heat in linseed oil, or by mixing glue with one-fourth of its weight of turpentine, or with a little bichromate of potash. Mahogany cement, for filling up cracks in wood, consists of 4 parts of beeswax, 1 of Indian red and yellow-ochre to give colour. Cutler’s cement, used for fixing knife-blades in their hafts, is made of equal parts of brick-dust and melted rosin, or of 4 parts of rosin with 1 each of beeswax and brick-dust. For covering bottle-corks a mixture of pitch, brick-dust and rosin is employed. A cheap cement, sometimes employed to fix iron rails in stone-work, is melted brimstone, or brimstone and brick-dust. For pipe-joints, a mixture of iron turnings, sulphur and sal ammoniac, moistened with water, is employed. Japanese cement, for uniting surfaces of paper, is made by mixing rice-flour with water and boiling it. Jewellers’ or Armenian cement consists of isinglass with mastic and gum ammoniac dissolved in spirit. Gold and silver chasers keep their work firm by means of a cement of pitch and rosin, a little tallow, and brick-dust to thicken. Temporary cement for lathe-work, such as the polishing and grinding of jewelry and optical glasses, is compounded thus:—rosin, 4 oz.; whitening previously made red-hot, 4 oz.; wax, ¼ oz.