UNIVERSITY OF CALIFORNIA AT LOS ANGELES 5 ( I WORKS OF PROFESSOR F. P. SPALDING rrm.isHKn r.v JOHN WILEY & SONS A Text=Book on Roads and Pavements. The aim of this work is to give a brief dis- cussion, from an engineering standpoint, of the principles involved in highway work, and to outline the more important systems of construc- tion, with a view to forming a text which may serve as a basis for a systematic study of the subject. I2mo, cloth, $2.00. Hydraulic Cement Its Properties, Test- ing, and Use. This work embodies the results of a careful study of the nature and properties of hydraulic cement, and the various methods which have been proposed, or are in use, for testing cement. I2mo, cloth, $2.00. HYDRAULIC CEMENT. 775 PROPERTIES, TESTING, AND USE. 0. w. VOL Ho. FREDERICK P. SPALDING, Assistant Professor of Civil Engineering at Cornell University Member c/ the American Society of Civil Engineer*. FIRST EDITION. FIRST THOUSAND. NEW YORK: JOHN WILEY & SONS. LONDON: CHAPMAN & HALL LIMITED 1898. Copyright, 1897, BV F. P. SPALDING. TT 37 1 S7U PREFACE. THE following pages contain the results of a care- ful study of the nature and properties of hydraulic cement, and the various methods which have been pro- posed, or are in use, for testing cement. The subject is not so simple as it might seem to the casual observer, but abounds in contradictions in the results of experiments and conflicts of opinion between authorities, which at times are quite bewildering. The views of the author, as derived from his own observation of the behavior of cement in use or in the laboratory, have been stated without reserve, and free use has been made of the results of available European investigations. The recommendations of the recent commissions appointed in Europe for the study of the methods of testing materials are fully given in so far as they relate to cement. The various tests applied for determining the quality of cement are discussed, and an effort is made to point out the limitaticvis within which they may be accepted as reliable indications of value. A chapter is given upon the use of cement in mor- tar and concrete, and a number of sample specifica- tions are appended for the purpose of showing the present practice of leading American engineers. F. P. S. ITHACA, N. Y., March, 1897. iii 182027 CONTENTS. CHAPTER I. 'HYDRAULIC LIME. PAGE Art. i. Definition I " 2. Chemical Ingredients 3 " 3. Hydraulic Index 7 " 4. Classification of Limes 8 " 5. Common Lime 9 " 6. Hydraulic Lime 13 " 7. Manufacture of Hydraulic Lime 16 " 8. Grappicrs 19 " 9. Puzzolana 21 CHAPTER II. CLASSIFICATION AND CONSTITUTION OF CEMENT. Art. 10. Classification of Cement 23 1 n. Manufacture of Cement 26 " 12. Portland Cement 30 ' 13. Natural Cement 37 " 14. Slag-cement 42 1 15. Mixed Cement 46 ' 16. Grappiers Cement 47 ' 17. Sand-cement 48 CHAPTER III. THE SETTING AND HARDENING OF CEMENT. Art. 18. The Setting of Cement 49 " 19. The Hardening of Cement 50 VI CONTENTS. PAGE Art. 20. Chemical Theory 51 " 21 Influence of Calcium Sulphate 55 " 22. Influence of Calcium Chloride 57 ' ' 23. Effect of Sand 60 " 24. Water used in Gauging 61 " 25. Effect of Environment 64 " 26. Effect of Temperature 64 " 27. Effect of Age upon Cement 67 " 28. Effect of Fineness , 68 CHAPTER IV. THE SOUNDNESS OF CEMENT. Art. 29. Permanence of Volume 70 " 30. Free Lime 72 " 31. Magnesia 74 " 32. Aluminate of Lime 77 " 33. Sulphur Compounds 78 " 34. Exterior Agencies So " 35. Effect of Sea-water 82 CHAPTER V. METHODS Or TESTING CEMENT. Art. 36 Object of Tests 85 " 37. Apparent Density 88 " 38. Specific Gravity 96 " 39. Tests for Fineness 103 " 40. Rate of Setting 105 " 41. Change of Temperature during Setting in CHAPTER VI. TESTS FOR THE STRENGTH OF MORTAR. Art. 42. Methods Employed 113 " 43. Form of Briquette 117 " 44. Quantity of Water used in Gauging 120 " 45. Methods of Making Briquettes 123 CONTENTS. Vli PAGE Art. 46. Mechanical Appliances for Making Briquettes 129 " 47. Tensile Tests 139 " 48. Compression Tests 146 " 49. Transverse Tests 151 " 50. Tests of Sand-mortar 153 " 51. Interpretation of Results 156 CHAPTER VII. TESTS FOR SOUNDNESS. Art. 52. Ordinary Tests 161 " 53. Measurement of Expansion 166 " 54. Accelerated Tests 171 " 55. Kiln Test 173 " 56. Steam and Hot-water Test 176 " 57. Boiling Test 180 " 58. Pressure Test 183 " 59. Chloride of Calcium Test 184 " 60. Value of Accelerated Tests 185 " 61. Air-slaking , 193 CHAPTER VIII. SPECIAL TESTS. Art. 62. Test of Adhesive Strength 195 " 63. Chemical Analysis 199 " 64. Tests for Homogeneity 207 " 65. Abrasive Tests 213 " 66. Tests for Porosity 215 " 67. Tests for Permeability 217 " 68. Frost Tests 221 " 69. Tests for Yield of Mortar 222 ' 70. Tests of Sand 223 CHAPTER IX. CEMENT-MORTAL AND CONCRETE. Art. 71. Sand for Mortar 224 " 72. Proportioning Mortar 227 Vlii CONTENTS. PAGE Art. 73. Gauging Mortar 229 " 74. Preparation of Concrete 231 " 75. Yield of Mortar and Concrete 235 " 76. Mixtures of Lime and Cement 237 " 77. The Freezing of Mortar 238 " 78. Porosity and Permeability of Mortar 241 " 79. Expansion and Contraction of Mortar 244 " 80. Effect of Retempering Mortar. 245 APPENDIX. SPECIFIC A TIONS FOR THE RECEPTION OF CEMENT. A. Specifications of the Pennsylvania Railroad; Joseph T. Richards, Engineer Maintenance of Way 249 B. Specifications of the Canadian Pacific Railway; Mr. P. Alex. Peterson, Chief Engineer 252 C. Extracts from Specifications for Furnishing Portland Cement for Wallabout Improvement, Brooklyn; Mr. W. E. Belknap, Engineer .253 D. Specifications for Municipal Work in St. Louis; Mr. M. L. Holman, Water Commissioner 259 E. Specifications for Municipal Work in Philadelphia; Mr. G. S. Webster, Chief Engineer, Bureau of Surveys 261 HYDRAULIC CEMENT. CHAPTER I. HYDRAULIC LIME. ART. 1. DEFINITION. LlME is the name commonly applied to the product obtained by the calcination of limestone. The lime- stones employed differ greatly in composition, and the properties of the limes obtained from them vary with the nature and proportions of the substances combined in them. When the limestone is composed of nearly pure carbonate of lime, the clinker resulting from the burn- ing, known as quicklime, possesses the property of breaking up, or slaking, upon being treated with a sufficient quantity of water. The slaking of lime is due to its rapid hydration when in contact with water, and the process is accompanied by a considerable in- crease in the volume of the mass of lime and by a rise in temperature. If the quantity of water be only sufficient to cause the hydration of the lime, the quick- 2 HYDRAULIC CEMENT. lime is reduced to a dry powder; while if the water be in excess it becomes a paste. The slaked lime thus formed possesses the further property, when mixed to a paste with water and allowed to stand in the air, of hardening and firmly adhering to any surface with which it may be in contact. This hardening of common limes will take -place only when exposed to the air and allowed to become dry. When lime is nearly pure and its activity is very great it is known as fat lime. If the lime have mixed or in combination with it considerable percentages of impurities of an inert character, which act as an adulteration to lessen the activity of the lime, causing a partial loss of the prop- erty of slaking and also diminishing its power of hard- ening, it is known as meagre lime. When the impurities in the lime are composed mainly of silica and alumina, they may, while lessening or destroying its property of slaking, impart to it the power of hardening under water and of setting without reference to the presence of air. When the proportions of the hydraulic ingredients are such that the material possesses the property of hardening in water, without having entirely lost that of slaking, the material is known as hydraulic lime. If the acquirement of hydraulic properties has been accompanied by an entire loss of the property of slak- ing, the product is hydraulic cement. Hydraulic limes and cements may be made either by burning limestones containing the proper propor- tions of hydraulic ingredients, in which case they are known as natural limes or cements, or by the admix- HYDRAULIC LIME. 3 ture of material containing such ingredients to the limestone before burning, or to the lime afterward, in which case they are known as artificial limes or cements. The materials used in the manufacture of lime and cement vary widely in different localities, and the resulting products differ greatly in their properties, being affected both by the composition of the raw materials and by the manipulation given them during the process of manufacture. Because of this variation in the character of the material, it is extremely diffi- cult to formulate any general laws governing its prop- erties, or to devise any system of testing which shall give an accurate determination of value. Hydraulic limes are used quite extensively in Europe, but are not made to any extent in this coun- try. The American cement industry is, however, a very important one. Natural cements are manufac- tured in immense quantities, and the production of Portland cement is rapidly growing to large propor- tions, although considerable of high-grade foreign cement is still imported. ART. 2. CHEMICAL INGREDIENTS. The most important ingredients of hydraulic lime- stones, in addition to the carbonate of lime, are usually alumina, silica, oxide of iron, and magnesia. They commonly contain also small quantities of sulphuric acid, phosphoric acid, oxide of manganese, potash, soda, bituminous or carbonaceous matter, organic sub- stances and water. The lime after burning may also 4 HYDRAULIC CEMENT. contain small particles of cinders from the combus- tibles employed. The volatile substances are without effect upon the lime because they disappear in burning. A small percentage of carbonic acid and water which escapes being driven off in the burning or is afterward absorbed from the air will appear in the lime. If this quantity of carbonic acid be large, it indicates either that the burning has been incomplete or that the lime has become carbonated by subsequent exposure. The energy of the lime is thus diminished, the portion of lime in combination with the carbonic acid being rendered inert. Silica is undoubtedly the most important element in rendering lime hydraulic and is always present. When in the form of silicious sand not attacked by acids it is unaffected by the burning and remains as inert material in the product. It is only that portion of the silica which is present in a condition to be re- duced in the burning and combine with the lime which is of value in imparting hydraulic properties to the lime. Alumina is an important element in hydraulic lime, although not, as seems to be the case with silica, an essential to its hydraulicity. Vicat found that alumina without silica would not render lime hydraulic. When, however, it is combined with silica it becomes one of the active elements in the setting and harden- ing of hydraulic limes and cements, provided it be not present in too large proportions. When in excess of about one part alumina to two parts silica, it is claimed that the surplus re-mains inert in the lime and detracts HYDRAULIC LIME. 5 from its energy. This, however, very rarely happens in practice. Oxide of Iron is commonly thought to be without influence upon the hydraulicity of lime, although there is some question concerning it. Like alumina, it confers no hydraulic properties when alone, and it is probably always inert. Magnesia seems to act quite differently as the con- ditions under which it is present vary, and its real action in most cases is in considerable doubt. Vicat made hydraulic limes by burning carbonate of mag- nesia with fat lime, thus showing that the magnesia alone might impart hydraulic properties to the lime. It has also been shown by Mr. H. St. Claire Deville that the oxide of magnesium by itself is sometimes an hydraulic material and sets under water. Other ex- periments have seemed to indicate that magnesia might act like alumina, and replace that element in forming an hydraulic material with silica and lime. These results may possibly all be due to the hydraulic properties of the magnesium oxide and independent of any action of the magnesia upon the lime. The silicates and aluminates of magnesia also have the property of hardening underwater like the similar salts of lime, and in some cements a portion of the lime which would otherwise be required seems to be replaced by magnesia. The action in these instances is not definitely known, and there is a difference of opinion amongst authorities concerning it, some think- ing that the magnesia acts like lime, others that it is inert and does not contribute to the energy of the material. In view of the activity of the magnesian 6 HYDRAULIC CEMENT. salts when alone, it seems reasonable to suppose that they have a similar effect in the magnesian cements. The activity of magnesia in hydraulic material depends largely upon the temperature at which it is burned, and in the experiments which have demon- strated its hydraulic properties it has been found that too high temperature destroys these properties. This possibly accounts for many of the contradictory results obtained by different investigators with these salts. This also is true in a less degree with the lime salts, which may sometimes be rendered inert by too high temperature. SulpJiuric Acid occurs in some limestones as sulphate of lime, and sulphur also sometimes occurs as a sul- phide, usually of iron. During the burning the sul- phide may become transformed into sulphate, and either or both forms may result in the final product. Experience seems to indicate that the sulphate is, in general, a deleterious substance, likely to affect the durability of mortar made from lime in which it occurs. The sulphide, however, is supposed to in- crease the hydraulic properties of the material, although its action is doubtful. In Europe it is com- mon to limit the percentage of sulphur or sulphuric acid which may be present in the lime. Phosphoric Acid occurs in very small quantities in material of this character, and is thought to have no action other than that of combining with a small quantity of lime, which is thus rendered inert. Oxide of Manganese is comparatively unimportant, as the quantity present is always too small to have any considerable effect upon the properties of the HYDRAULIC LIME. 7 lime. It has been supposed by some authorities to have strong hydraulic properties. The Alkalies contained in the limestone act as a flux during the burning, causing the chemical reactions to take place more readily and completely. The small amount which ordinarily ocurs in lime and cement is unimportant in its effect, the alkali being gradually dissolved out of the mortar. ART. 3. HYDRAULIC INDEX. The hydraulic activity of a lime or cement, that is, its ability to harden under water, depends primarily upon the relative proportions of the hydraulic in- gredients and of lime. Silica and alumina are consid- ered to be the effective hydraulic ingredients, and it is common to designate the ratio of the sum of the weights of silica and alumina to that of lime in the material its hydraulic index. The hydraulic index gives, therefore, within certain limits, a measure of the hydraulicity of the various classes of limes. It is to be remembered, however, that there are other factors to be considered in judg- ing of the action of a lime than this simple proportion. The other ingredients may by their combinations withdraw portions of the active elements so as to modify the effective ratio between them, while the activity of the lime depends largely upon the state of combination in which the active elements exist. This is not shown by analysis, and may. be greatly modified by the manipulation given the material during manu- facture. HYDRAULIC CEMENT. ART. 4. CLASSIFICATION OF LIMES. Limes may be classified according to their hydraulic indices, or according to the rate of hardening under water. Limes with hydraulic index less than 10/100 pos- sess little if any hydraulic properties, and are known as common limes, which may be either fat limes or meagre limes according to the proportion of inert material contained by them. When the hydraulic index is- between 10/100 and 20/100 the lime is known as feebly hydraulic and may require from 12 to 20 days to harden under water. Hydraulic lime proper includes that of index varying from about 20/100 to 40/100. These may harden in 2 to 8 or 10 days, the more rapid ones being sometimes classed as eminently hydraulic. When the hydraulic index is between about 40/100 and 60/100 the lime is of the class known as limiting lime. These limes have not the characteristics of hydraulic lime, but form the boundary between the limes proper and the cements, losing the property of slaking, and as a general thing not possessing a suffi- cient quantity of the hydraulic ingredients to make a safe natural cement. When burned at a low tem- perature it may give hydraulic lime of rather poor quality, and when properly treated and burned at a high temperature it makes good Portland cement. From material with an index of 65/100 to i T ? ff i5 are obtained most of the common natural cements. With hydraulic index of i^ to 3 T Q yo> the material yielded HYDRAULIC LIME. 9 by burning is known as meagre cement, usually a weak material of little value as a cement. The material of hydraulic index S^A m ^y be puzzolana, which has not the properties of cement, but when mixed with fat lime renders the lime hydraulic. The above divisions are all more or less arbitrary, and there are no sharp lines between the classes, which merge into and overlap each other. The character of the material is also affected by other factors than the hydraulic index, and classification by this method is by no means invariable, material of one class frequently behaving like that of another class. It is difficult to determine at exactly what point lime begins to be hydraulic, but when it requires a month or more to harden under water it is usually considered as being non-hydraulic. In determining the rate of hardening of limes there are so many external circumstances which may affect the result that there is always chance for error, which causes classification by that method to be somewhat uncertain. Variation in temperature has always an important effect upon the rate of hardening. ART. 5. COMMON LIME. Common lime is such as does not possess hydraulic properties. It is divided into fat or rich lime and meagre lime, according to the quantity of impurities of an inert character it may contain. When made into a paste and left in air it slowly hardens. The process of hardening consists in the gradual formation IO HYDRAULIC CEMENT. of carbonate of lime through the absorption of car- bonic acid from the air, accompanied by the crystalli- zation of the mass of hydrated lime as it gradually dries out. In common lime the final hardening takes place very slowly, working inward from the surface, as it is dependent upon contact of the mortar with the air. When the lime is nearly pure the resulting car- bonate is likely to be somewhat soluble, and conse- quently to be injured by exposure. Nearly all limes, however, contain small percentages of silica and alumina, and these ingredients, even when in quanti- ties too small to render the lime hydraulic, impart a certain power to set, causing the hardening to take place with greater rapidity and without the same dependence upon contact with air. It also renders the material less soluble and more durable in exposed situations. Limes containing but a small amount of impurities consist mainly of calcium oxide, which is very caustic and becomes hydrated very rapidly when brought into contact with water. This hydration, or slaking, pro- duces a rise in temperature and an increase in volume, which vary in amount according to the purity of the lime, the volume being doubled or tripled for good fat lime. The common method of slaking consists in covering the quicklime with water, using two or three times the volume of the lime. This method is known as drown- ing. The lime is usually spread out in a layer, per- haps 6 or 8 inches thick, in a mixing-box, the water poured over it and allowed to stand. Sufficient time must be allowed for all of the lumps to be reduced. HYDRAULIC LIME. II When the lime contains much foreign matter, the operation frequently requires several days. Too great a quantity of water is to be avoided, the amount necessary being such as will reduce the lime after slaking to a thick pasty condition. All the water should be added at once, as the addition of water after the hydration is in progress causes a lowering of tem- perature and checks the slaking. For the same reason, the lime should be covered after adding water, and not stirred or disturbed until the slaking is com- pleted. The covering is commonly effected by spread- ing a layer of sand over the lime; the sand being afterward used to mix with it in making the mortar. A second method of slaking is sometimes employed having for its object the reduction of the slaked lime to powder and known as slaking by immersion. Slaking by immersion is accomplished in two ways. By the first method, the lime is suspended in water in baskets for a brief period to permit the absorption of the necessary water, after which it is removed and covered until the slaking takes place and the lime falls to powder. By the second method, sprinkling is substituted for immersion proper, the lime being placed in heaps and sprinkled with the necessary quantity of water, then covered with sand and allowed to stand. The difficulty in using these methods is to get just the right quantity of water to make the slaking com- plete. Lime so slaked may be barreled and shipped in form of powder. Spontaneous Slaking is also sometimes resorted to; 12 HYDRAULIC CEMENT. it consists in exposing the lime to the air until slaking is effected by absorption of moisture. Lime slaked by immersion swells less than when slaked by drowning and requires less water to form into a paste. Slaked lime, either as powder or paste, may be kept indefinitely if protected from the air. Lime is commonly sold as quicklime, and should be in lumps and not air-slaked. When it is old and has been exposed to the air, it is likely to have absorbed both moisture and carbonic acid, thus becoming less active, the portion combined with the carbonic acid being inert. A simple test of the quality of quick- lime is to immerse a lump for a minute, then place in a dish and observe whether it swells, cracks, and dis- integrates with a rise of temperature. Slaking some days in advance of use is desirable in order to insure the complete reduction of the lime, and it is quite common to slake lime much longer before it is to be used. The swelling of lime has been found to be increased by slaking with steam instead of water, and slaking has been found by M. Candlot to be accelerated by the addition to the water of a small quantity (2% to 6#) of chloride of calcium or chloride of magnesium. Common lime is ordinarily used in construction as a mortar, mixed with sand. The quantity of lime in the mortar should be just sufficient to fill the voids in the sand, without leaving any part formed entirely of lime. Mortar of rich lime shrinks on hardening, while masses composed entirely of lime on the interior are likely to remain soft, so that an excess of lime may be an element of weakness. If too little lime be used, HYDRAULIC LIME. 13 the mortar may be porous and weak. The propor- tions ordinarily required are between i part lime to 2 parts sand, and I part lime to 3 parts sand. Mortar of common lime should not usually be em- ployed in heavy masonry or in damp situations. Where the mass of masonry is large, the lime-mortar will become hardened only with great difficulty and after a long time. The penetration of the final in- duration due to the absorption of carbonic acid is very slow. The observations of M. Vicat showed that car- bonization extended only a few millimeters the first year and afterward more slowly. The induration of the lime along the surfaces of contact with a harder material, when used in masonry, is usually more rapid than in the interior of the mass of lime itself, and hence the strength of adhesion to stone or brick is often greater than that of cohesion between the par- ticles of mortar. Meagre limes are similar in action to fat limes, but less energetic. They swell feebly in slaking and with slight change in temperature. The mortar hardens like that of fat lime, but cracks less and contracts less. Meagre limes proper, those containing so much inert matter as to materially reduce the energy, are not commonly employed in construction. ART. 6. HYDRAULIC LIME. Hydraulic lime is obtained by burning limestone containing silica and alumina in sufficient quantity to impart the ability to harden under water. The hy- draulic elements are present in such quantities that 14 HYDRAULIC CEMENT. they combine with a portion of the lime, forming sili- cates and aluminates of lime, leaving the remainder as free lime in an uncombined state. When treated with water the free lime is slaked, the action being much less energetic than that of fat lime and varying in intensity with the quantity of hydraulic ingredients. The quantity of free lime in the material is depend- ent upon the degree of burning, as well as upon the amount of lime contained by the stone. If the stone be underburned, the combination of the hydraulic ele- ments with the lime is not complete, and more of the lime remains in a free state. For this reason a stone of high hydraulic index may give a lime when under- burned, but become unslakable when burned at a high temperature, as in the case of the limiting limes. The best limes are usually those which can be burned at a high temperature to complete the chemical combina- tions. t is necessary that sufficient free lime be present to cause the lime to slake properly, but it is desirable also that the quantity of uncombined lime be as small as possible, as the setting properties are due to the silicates and aluminates, while the hydrated lime remains inert durinp 1 the initial hardening of the mortar. According to Professor Le Chatelier, limestone for hydraulic lime should contain but little alumina, as the aluminates are hydrated during the slaking of the lime and become inert, while the silicates are not affected, the heat of the slaking preventing their hydration. HYDRAULIC LIME. 1 5 The following is given as an average analysis of the best French hydraulic lime: Silica 22 Alumina 2. Oxide of iron I. Lime 63. Magnesia. ....*.... 1.5 Sulphuric acid 0.5 Water. . 10.0 100 It is important that the slaking be very thorough, as the presence of unhydrated free lime in the mortar while hardening is an element of danger to the work. Any lime becoming hydrated after the setting of the mortar may, by its swelling, cause distortion and per- haps disintegration of the mortar. Hydraulic lime is used in the same manner as fat lime, being mixed with sand to a paste. When in the air hydraulic lime acts like common lime, dries", har- dens, and slowly absorbs carbonic acid. It contracts and cracks when without sand, but much less than fat lime. In water, or in damp situations, the action of the two are altogether different. The hydraulic lime then hardens more or less rapidly. In running water a small amount of the lime is at first dissolved, but this is soon arrested as the hardening progresses. Hydraulic lime is commonly slaked at the manufac- tory and shipped in form of powder. It may be kept, without injury, in this form by covering and protect- ing from the air. l6 HYDRAULIC CEMENT. ART. 7. MANUFACTURE OF HYDRAULIC LIME. The manufacture of hydraulic lime as commonly carried on in Europe consists, after the quarrying of the rock, of burning, slaking, and bolting the material. As already stated, it is usually slaked at the works and sold in the form of powder. The varieties of furnaces used in burning are quite numerous, but may be divided into those in which the stone is burned in contact with the fuel, and those in which the fire is outside the chamber in which the burning takes place. Furnaces of the first class have been more generally employed, and are claimed (Candlot, Ciment et Chaux hydrauliques, p. 7) to be preferable from the point of view of the uniformity of product. Continuous fur- naces are commonly used. The furnace is filled by placing alternate layers of combustible and limestone. When full it is lighted at the bottom, and as the mass settles, new layers of material are added at the top, while the burned lime is drawn out at the bottom, the furnace being kept in continuous operation. The rapidity of burning is controlled by dampers and the movable cover at the top. In furnaces of the second class, the flame and gases of combustion are passed through the stone, which is not in direct contact with the fuel. The regulation of a furnace to secure the proper degree of burning is a matter requiring skill and ex- perience, and demanding the close attention of the attendant. The lime must be completely burned, but not overburned. The character of the limestone HYDRAULIC LIME. I/ determines the amount of burning necessary. A lime of low hydraulic index may be burned at a much higher temperature than one of high index. When the limestone is irregular in character it will not burn evenly, but the parts of high index will be vitrified before that of low index is properly burned. The burning is accelerated when the stone is moist, and stone fresh from the quarry is preferred to that which has been exposed until the hygrometric water has been evaporated. The chemical phenomena of the burning of lime are approximately as follows: The hygrometric water is first driven off. The carbonate of lime is next de- composed. Then the clay is dehydrated and decom- posed, and the combination of the silica and alumina with the lime takes place. The temperatures required to effect these changes depend upon the composition. The decomposition of the carbonate of lime takes place at a lower temperature in presence of silica, and hence silicious limestones are burned at a comparatively low temperature maintained for a considerable time, while argillaceous limestones require a higher tem- perature maintained for a shorter period. According to M. Bonnami, the carbonate of lime is commonly decomposed at about 440 C., while the clay is dehy- drated and decomposed at above 700 C., and the silicates and aluminates of lime are then formed. If a temperature intermediate between these points be maintained for a considerable time, the result will be the decomposition of the carbonate of lime, leaving caustic lime mixed with insoluble clay. The slaking of hydraulic lime is commonly accom- 18 HYDRAULIC CEMENT. plished by sprinkling, as mentioned in Art. 5. The lime after coming from the furnace is spread in layers 4 to 8 inches deep in the slaking-chambers. It is then sprinkled so that all of the quicklime is well moistened; from "]% to io# of water is commonly required. When the lime is wet it is thrown into large heaps and left for a sufficient time for the slaking to be completed. The time required varies with the hydraulic index and the degree of burning. Limes of high index may require 15 or 20 days. When the index and degree of burning increase together, the lime soon becomes unslakable. In slaking, the object is to obtain the complete hydration of the uncombined lime without causing any change in the silicates. To accomplish this the slaking must take place at a high temperature. The heat of slaking volatilizes the surplus moisture and prevents the hydration of the silicates. If the temperature be too low, the lime may partially harden during slaking and the portion of silicate which is hydrated becomes inert. If the lime be imperfectly slaked, the free lime left in the material may cause injury to the mortar after hardening. Lime so affected will set more quickly than it would if sound, but afterward is likely to swell and crack. Steam is sometimes used for slaking instead of water, and is said by M. Le Chatelier to act more rapidly upon the lime while producing no effect upon the silicates. It is claimed that the aluminates are hydrated dur- ing the slaking of the lime, being readily acted upon by steam, and hence are undesirable in hydraulic lime. HYDRAULIC LIME. 19 After the lime has been reduced to powder by slak- ing it is forced through sieves which permit the passage of all pulverized particles, but hold those of appreciable size, including the underburned rock and the overburned parts which refuse to slake. The lime resulting from the first bolting is known as the flour of lime. ART. 8. GRAPPIERS. The residue left after the sifting of hydraulic lime is known as grappiers. It differs very much in its composition in various instances, depending upon the limestone used and the manipulation in manufacture. It includes the under- burned portions of the limestone and the overburned particles which will not slake. When the burning is thoroughly done and the limestone used regular in composition, the proportion of unburned particles is small. The larger part of the residue is then com- posed of hard material more rich than the other por- tions of the lime in silica and alumina, obtained from the clay which is disseminated through the limestone or formed by the combination of the cinders of com- bustion with the lime. This is what is properly meant by the term grappiers. M. Bonnami found in his investigations that the larger part of the grappiers are from the surfaces of the limestone which are in contact with the combusti- ble during the burning, and due to the silica and alumina of the cinders. These cinders are usually quite aluminous. 20 HYDRAULIC CEMENT. The grappiers are ground and sifted and either added to the lime or used separately as cement. The addition of ground grappiers to hydraulic lime has the effect of raising the hydraulic index of the lime and increasing its activity, and offers a means of controlling to a certain extent the properties of the lime. In this case the mixture between the lime and the grappiers must be very intimate in order to obtain a homogeneous material. It is also very important that the lime in the grappiers be entirely slaked, to prevent the introduction of free lime into the product. To secure this the grappiers after grinding are exposed to the air for a considerable time before using, thus permitting any unslaked particles to become air-slaked. The following analyses of lime and grappiers from the great works at Teil are given by Prof. Durand- Claye (Chimie appliquee a 1'art de 1'ingenieur). ANALYSES OF LIME AND GRAPPIERS FROM TEIL. Slaked Lime. Merchantable Lime. Grappiers. Rejected Material. Silica Per Cent. Per Cent. 2*1 qc Per Cent. 31 8s Per Cent. Alumina and iron oxide. 2-75 3.10 63 35 4.25 55 6 8.20 45 2 5 i 15 I 20 o 85 6 Q5 8 50 7. 10 2 60 The first column gives the ordinary slaked lime. Merchantable lime has a portion of powdered grappiers added to augment its hydraulic properties. The third column gives an analysis of powdered grappiers, which is sold as a slow-setting cement. The rejected ira- HYDRAULIC LIME. 21 terial is a calcareous sand which has puzzolanic proper- ties. It is employed with cement in making water- pipe, brick, etc. ART. 9. PUZZOLANA. The term puzzolana is commonly applied to a class of materials which, when made into a mortar with fat lime or feebly hydraulic lime, impart to the lime hydraulic properties and cause the mortar to set under water. Puzzolana (or pozzuolana) proper is a material of volcanic origin, deriving its name from Pozzuoli, a city of Italy near the foot of Mount Vesuvius, where its properties were first discovered. It was extensively used by the Romans in their hydraulic constructions, being pulverized and mixed with slaked lime and a small amount of sand for the formation of hydraulic mortar. The puzzolana is a silicate of alumina in which the silica exists in a state easily attacked by caustic alka- lies, and hence readily combines with the lime in the morta- The class of puzzolanas also includes several other materials of somewhat similar character. Trass is the name given to a volcanic material found in Germany and Holland, much resembling puzzolana and used in the same manner. Arcncs is a sand found in France and applied to the same purpose. It is quartzose in character and mixed with clay in considerable proportions; from 1/4 to 3/4 of the total volume. It may be made into a paste, 22 HYDRAULIC CEMENT. with water, which will harden on drying out, and is sometimes used for common mortar without lime. Psammitcs is a sandstone consisting of grains of quartz, schist, feldspar, and mica, agglutinized with a variable cement. It is slaty in character and may be worked into a paste with water. Puzzolana may be made artificially by burning clay, and natural ones may frequently be improved by burn- ing, which has the effect of dehydrating the silicate of alumina of which they are mainly composed and leav- ing it in condition to combine readily with the lime. Berthier gives the following analyses of average samples of puzzolanas: Trass. Puzzolana. Silica 0.570 0.445 Alumina 0.120 0.150 Lime 0.026 0.088 Magnesia o.oio 0.047 Iron oxide 0.050 o. 123 Potash 0.070 0.014 Soda o.oio 0.040 Water, etc 0.444 0.096 CHAPTER II. CLASSIFICATION AND CONSTITUTION OF CEMENT. ART. 10. CLASSIFICATION OF CEMENT. HYDRAULIC cements may be classified according to the method of manufacture under five general head- ings: Portland cements, natural cements, slag-cem- ents, mixed cements, and grappiers cements. The term Portland Cement is commonly used to designate hydraulic cement formed by burning to the point of vitrifaction a mixture of limestone and clay in proper proportions and reducing the resulting mass to powder by grinding. The cement so classified is of lower hydraulic index than the other cements, and is consequently burned at a higher temperature. Portland cement is usually made artificially by a mix- ture of limestone and clay or of nearly pure lime- stone with stone of high index, and in all cases the material must be very uniformly incorporated into the mixture. The high temperature employed in burning and the necessity of reducing the raw material, whether natural rock or artificial mixture, to powder before burning, for the purpose of homogenizing it, may be considered the distinctive characteristics of this class. The conference for the unification of methods for 23 24 HYDRAULIC CEMENT. testing materials at Munich * propose the following additional definition of Portland cements: " They contain a minimum of 1.7 parts of lime per unit of hydraulic substances. The addition of 2% by weight of foreign matter may be tolerated in the manufacture of Portland cement, with the view of augmenting cer- tain important qualities, without the necessity of changing the name." Natural Cements are those which are made by burn- ing limestones less rich in lime than those giving hydraulic limes or Portland cement. These are burned like the hydraulic limes without pulverization of the raw material, and require a much lower tem- perature in burning than does Portland cement. This class includes a number of sub-classes varying widely in composition and varlue. In Europe they are commonly divided into quick-setting natural cements, frequently called Roman Cement, and semi- slow-setting cements, known sometimes as natural Portland cement. In the United States there is much greater variety in the materials coming within this classification and much confusion in their nomencla- ture. They are most commonly designated by a name derived from the locality in which they are obtained, and this seems the most feasible and satis- factory method. Thus, Rosendale cements are those from the region of the lower Hudson, Lehigh cements are from Southeastern Pennsylvania, Louisville cem- ents from the Ohio valley, Potomac and James River * M6moires de la Soci6te des Ingenicurs Civils, 1891. vol. I. p. 112. CLASSIFICATION AND CONSTITUTION OF CEMENT. 2$ cements from the corresponding valleys, while Utica, Akron, and Milwaukee are names indicating the loca- tion of manufactories of particular brands. The term American Cement has sometimes been applied to include all natural cements made in the United States. This, however, often leads to con- fusion because of the fact that other than natural cements are now made quite extensively in this coun- try, and the term American Portland Cement is also in common use The term Rosendale Cement has frequently been given a general meaning and used as synonymous with natural and American to include all natural cements. It is, however, more properly restricted to the cements of the district in which it was first applied, and there seems to be no good reason for extending it to include other and totally different material. Slag Cement or, as a more general term, Puzzolana Cement, is the product obtained by an intimate mix- ture of slaked lime with finely pulverized puzzolanic material, commonly blast-furnace slag. In this ma- terial the hydraulic ingredients are not burned with the lime, but are present in the cement in a mechanical mixture only. Graf piers Cements are obtained by grinding the par- ticles which are not pulverized in slaking hydraulic lime. Mixed Cement is the name given in Europe to an extremely variable class of products obtained by mix- ing different grades of cement together, or by mixing cement with other material for the purpose of impart- ing desired properties. 26 HYDRAULIC CEMENT. ART. 11. MANUFACTURE OF CEMENT. The manufacture of hydraulic cement as commonly practised consists of four operations, viz., the prepara- tion of the raw material, the burning, the grinding, and the bolting. The methods of preparing the raw material differ according to the nature of the material and the method to be used in burning. For natural cements it is usually only necessary to select the proper portions of the rock and break it into fragments of suitable size for introduction into the furnace. The production of good cement requires the use of homogeneous material, and care must be used to prevent the introduction of variable rock into a single burning. For Portland cement there are three general methods of preparing the material, in all of which it is essential to good results that the various ingredients be very carefully proportioned and that they be formed into a very uniform and homogeneous mixture in order to facilitate the chemical changes in all parts of the material during the burning. The first, known as the wet method, consists in working the raw materials into an intimate mixture by reducing to a paste with water, then drying into bricks which may be stacked in the furnace for burning. In the wet process proper a large excess of water is employed and afterward drawn off. In the semi- wet process, now more com- monly employed, only enough water is used to reduce the mixture to a plastic condition. The second method, called the dry method, con- sists in grinding the materials together dry or with a CLASSIFICATION AND CONSTITUTION OF CEMENT. 2? very small quantity of water, ana making bricks of the powder by subjecting it to pressure in a brick-machine. The bricks in all cases require thorough drying before being placed in the furnace. The third method is to grind the dry materials into powder and burn in a rotary furnace without forming them into bricks, or to mix to a plastic condition and dry in small lumps on the circumference of a drying- cylinder for burning in the rotary furnace. The exact proportioning of the ingredients and the intimacy of their incorporation into the mixture have very important bearing upon the value of the cement. The materials used in manufacturing cement differ greatly in different localities, and the method employed depends somewhat upon the character of the raw materials. For natural cements a limestone of high hydraulic index is usually employed, but differing much in composition, some having a high percentage of alumina, others of magnesia, and still others of both. For Portland cements the most common materials are a fat or slightly hydraulic limestone with clay or shale, made into bricks by the semi-wet process. Sometimes a hydraulic limestone of high index (such as is used for natural cements) is mixed with a fat limestone, commonly by a dry method. These materials are also sometimes used by the method of double calcination, that is, the fat limestone is first burned in the ordinary manner, the resulting quicklime is slaked and bolted, after which the slaked lime is mixed and ground with the argillaceous limestone, the object being to get a very perfect distribution of the lime through the mixture. 28 HYDRAULIC CEMENT. The' method of double calcination is also sometimes employed with rock containing naturally about the right proportions of hydraulic ingredients for making Portland cement. The composition of the rock being usually somewhat irregular, it is lightly burned and reduced to powder, in order to secure greater uni- formity, and then formed into bricks and burned in the usual manner. The furnaces used for burning cement are of several kinds. The ordinary vertical lime-kiln, as mentioned in the previous chapter, is very commonly used, the kiln being used intermittently, and requiring usually 5 to 10 days to burn a charge. The Hoffman continuous kiln is a series of cham- bers arranged in a circle, one chamber being fired at a time and the products of combustion passing through the chambers containing unburned material, the firing progressing from chamber to chamber continuously around the kiln. The Dietzsch kiln is of the same character as the Hoffman, but has the fire outside the chamber contain- ing the material to be burned. The Ransome kiln consists of a rotary cylinder lined with fire-brick. The axis of the cylinder is inclined at an angle of 10 or 12 degrees with the horizontal, and the heat is applied as a gas- or oil-flame introduced through the axis of the cylinder at its lower end. The slurry is introduced in the form of powder at the upper end of the cylinder, and is slowly carried to the lower end by the corrugations of the inner surface of the cylinder, thus being gradually subjected to the heat, which reaches a maximum at the lower end, CLASSIFICATION AND CONSTITUTION OF CEMENT. 2Q where the clinker falls out either into cooling cylin- ders or upon floors from which it is removed to storage-rooms. The nature of the burning is indicated by the color of the clinker, which for Portland cement is dark green or black when well burned ; underburned material being of a lighter color and weight, while the over- burned material may powder upon cooling, forming what is known as heavy powder. This heavy powder is inert and does not set as a cement, but ordinarily possesses puzzolanic properties and becomes active when mixed with lime. The degree of burning required, as already indi- cated, varies with the nature of the material used, the heat required being greater as the hydraulic index of the material becomes less. The heavy powder mentioned above is formed at lower temperatures as the proportion of lime becomes less, and each grade of material has a certain range of temperature within which it should be burned, below which it will be underburned and above which it will be rendered inert. In underburned cement the chemical changes are incomplete; a part of the lime may be left as caustic lime uncombined with the clay. This is shown by its light weight. In the burning, as the dehydration of the materials and the decomposition of the carbonate of lime is first effected, the limestone loses in weight without loss of volume, and thus suffers a loss in apparent density. As the subsequent combination of the lime with the clay occurs, a contraction in volume takes place and the density becomes greater. 30 HYDRAULIC CEMENT, After the burning is completed, the clinker is broken in a crusher to small fragments and then ground to powder. The cement is then bolted through sieves to separate the coarse particles, which are afterward returned to the grinders. The grinding is a very im- portant matter, as only the extremely fine impalpable powder is of real value in the cement, and to secure good results the method of grinding must be such as to produce a large proportion of such material. After the bolting of the cement it is usually carried to chambers and spread out for aeration, but in some cases is packed directly in barrels or bags for shipment. The necessity for aeration depends upon the ac- curacy of the composition and the completeness of the chemical combinations. Its object is to eliminate the quicklime which may be present by allowing it to become air-slaked. The method to be used in manufacturing cement must in each instance be modified to suit the material to be operated upon. The rock of a single quarry usually varies so much as to require different treat- ment in its various parts; or if mixtures are to be made, constant watchfulness is required in regulating the proportions in order to obtain a product of uni- form quality. ART. 12. PORTLAND CEMENT. The term Portland Cement is usually limited to material containing a high percentage of lime and burned at a high temperature. It is usually low in alumina and magnesia. In order to make a good CLASSIFICATION AND CONSTITUTION OF CEMENT. 31 cement of this character it is necessary that the in- gredients be very accurately proportioned and that the material be very homogeneous. This requires ordinarily the pulverization of the raw materials and their uniform incorporation into the mixture in a finely divided state. The action of Portland cement seems to depend upon the formation, during the burning, of certain silicates and aluminates of lime which constitute the active elements of the cement, the other ingredients being considered in the light of impurities. The ideal cement would be that in which the proportion of lime is just sufficient to combine with all the silica and alumina in the formation of active material. If there be a surplus of clay beyond this point, it forms inert material. Any surplus of lime remains in the cement as free lime, and constitutes one of the chief dangers in the use of cement, as, although it may not prevent the proper action of the cement when used, it may cause the mortar to afterward swell and become cracked and distorted as the lime slakes. As perfect homogeneity is not attainable in practice, it is always necessary that the clay be somewhat in excess in order that free lime be not formed. The amount of excess of clay necessary evidently depends upon the thoroughness of the process used in manu- facture and the evenness which may be reached in the mixture of raw materials. The hydraulic index of Portland cement varies from about 42/100 to 60/100. The value of the index is affected by the relative proportions of silica, alumina, and iron oxide contained by the cement as the equiva- 32 HYDRAULIC CEMENT. lent weights of these oxides differ. The normal com- position of Portland cement is usually within the fol- lowing limits: ' Silica 20 to 25 per cent. Alumina... 5 " 9 " Iron oxide 2 " 5 " Lime 57 " 65 " Magnesia 0.5 " 2 " ! Sulphuric acid 0.25" 1.50" Table I, taken from Candlot,* gives analyses of a number of representative European Portland cements, while Table II, collected from various sources, gives analyses of a few of the leading brands both American and foreign sold in this country. A large number of analyses of European Portland cements are given by Professor Tetmajerf which show for the most part about the same range of variation as those already given. Professor Le Chatelier has made a very careful study of the constitution of Portland cement by analyzing sections of clinker under the microscope, as .well as by studying synthetically the various compounds of the principal ingredients. He concludes $ that the tri- calcic silicate, SiO B Ca 3 , is the only one that is really hydraulic and is the active element in cement. In Portland cement he finds it to be the principal com- * Ciment et chaux hydrauliques (Paris, i8gi). Method en und Resultate der Priifung der Hydraulischen Bindemittel (Zurich, 1893). J Annales des Mines, September, 1893. CLASSIFICATION AND CONSTITUTION OF CEMENT. 33 ponent, occurring in cubical crystals. It is formed by combination of silica and lime in presence of fusible compounds formed by the alumina and iron. TABLE I. COMPOSITION OF PORTLAND CEMENTS. a i c* 'G SJ I 2 15 25-85 IO.OO 4-85 54-20 1.65 1. 00 2-45 .... 16 29.10 12.50 4.65 48.60 1.70 1.90 i-55 17 3.40 24 65 "35 5-25 50.45 1-15 1-25 2.50 .... 18 0.50 20.00 8.40 5-70 52.05 c-95 2.80 9.60 .... 19 2.60 27.10 4.10 3-75 48.70 0.65 0.95 12.15 .... 20 .... 28.06 6.65 3-30 47-79 i. 08 1.66 10.44 21 .... 27-54 9-25 3-8o 54.58 0.50 0.64 3.69 22 20.54 8.72 3-23 5I-85 2.31 3-93 9.28 .... 23 .... 21.04 12.72 4.04 5I-38 1.20 5-5i 3-65 24 21.29 9-3 6 3-5i 5-74 4.24 6.01 3-22 .... 25 23 35 9.69 2.96 54-71 i.oS 1.95 4-3 1 .... 26 20.60 9.92 2.56 52.42 3-91 6.06 3-30 .... 27 28.36 12.12 1-57 47.28 2.24 2.OO 4-52 .... 28 .... 25.64 8.76 2.15 44-87 1-93 5-U 10. 16 .... 29 25.70 9.26 3-38 50.86 1-54 1.51 6.67 .... 30 .... 29.74 11.92 3-68 42.98 2.0S 2.34 6.20 .... 31 .... 22.14 8.74 3-69 58.41 2. 02 2.90 2.12 32 23-35 8.20 3-74 57-94 1.63 2.98 2.S2 are given a somewhat lighter burning. They are, however, more heavily burned than the Roman cem- ents. Limestones in nature are not so homogeneous as the artificial mixtures used in making Portland 4O HYDRAULIC CEMENT. cement, and the proportion of lime cannot be so great as in the more homogeneous mixtures without danger of producing an objectionable quantity of free lime in the cement. The use of material of this character, therefore, requires much care in order to produce good results. As the hydraulic index becomes greater the homogeneity becomes less important, as free lime becomes less likely to occur and less dangerous, and irregularities only have the effect of increasing the quantity of inert matter, which causes mortar made from the cement to gain strength much more slowly than with Portland cement of low index. It is to be observed that the material spoken of as inert, and which delays the gain in strength in the early period of hardening, may not be altogether inert, and may contribute to the final strength of the cement, as it is of a puzzolanic character and perhaps ultimately com- bines with the hydrated lime in the mortar. These cements occupy an intermediate position between the artificial Portland cements and the Roman cements, and may approach either in composition. In fact, the same raw material may frequently produce either if burned lightly giving the quick-setting Roman cement, or burned more heavily a slow-setting natural Portland. Heavy burning increases the amount of silica combined with lime at the expense of the aluminates, thus relaxing the rapidity of set and in- creasing the early strength of the mortar. The Llagnesian Natural Cements are those in which a portion of the lime of the Roman cement is replaced by magnesia. Very little is known as to the action of the magnesia in these cements. It seems probable CLASSIFICATION AND CONSTITUTION OF CEMENT. 41 that the magnesia replaces lime or combines with it in the formation of double silicates and aluminates, and that it bears some part in the setting and harden- ing of the mortar. That certain magnesian salts possess hydraulic properties is well known, their action according to M. Fremy being probably much slower than the corresponding lime-salts. The action of cements of this class is somewhat similar to that of Roman cements: they gain strength very slowly, but may be either quick or slow setting. The composition of the magnesian cements varies from that of the Roman cements to one in which the pro- portion of magnesia is as large as that of lime. As the proportion of magnesia to lime increases, the hydraulic index, considering magnesia as lime, fre- quently decreases and becomes less than would be admissible in Roman cement. Magnesian cements are but little used in Europe, but in the United States they form the largest part of the natural cements in use, and many of them have been found by experience to be very useful and reli- able materials. The Rosendale cements are of this character. The rock from which these cements are made differs greatly in character in the same locality, and in the different strata of the same quarry. In some of the works the nature of the product is regu- lated by mixing in proper proportions the clinker obtained by burning the rock from different strata. Each portion of rock must be burned in such degree as is suited to its composition, and hence as the ma- terial is not pulverized before burning it must be burned separately and mixed afterward. To produce 42 HYDRAULIC CEMENT. uniformly good cement, therefore, requires close and careful attention; and for this reason there is often considerable difference in the quality of cement made by works in the same locality and from very similar material. Cement of high index, unlike Portland cement, is usually materially changed by age. When these cements are kept exposed to the air for a considerable length of time, they gradually become slower-setting and perhaps eventually lose the power of setting altogether", sometimes becoming puzzolana, which again becomes active cement by reburning. ART. 14. SLAG-CEMENTS. Slag-cement is formed by the admixture of slaked lime with ground blast-furnace slag. The slag has approximately the composition of an hydraulic cement, but lacks a proper proportion of lime to render it active as a cement. These cements are sometimes called puzzolana cements, the slag used being of the same nature as the puzzolana commonly employed in making lime hydraulic. The method employed in forming slag-cement is to cool the slag suddenly by plunging it into a current of water as it emerges from the furnace. This makes the slag granular, and causes it to retain the heat of crys- tallization, thus rendering its elements more ready to enter into combination in presence of water. Experience in Europe shows that the slag must be basic in order to be of use in making cement. Pro- CLASSIFICATION AND CONSTITUTION OF CEMENT. 43 fessor Tetmajer* arrives at the conclusion that slags in which the ratio -TTT is unity are not suitable, and that silica above this proportion the value of the product in- creases with this ratio. He also finds that the best results are obtained from slags giving a ratio Al,0 3 T- 3 = 45/100 to 50/100. M. Prost f states that a considerable amount of sulphur may be unobjectionable in slag-cements, and mentions a case where good results had been obtained with sulphurous slag, the only effect being discolora- tion attributed to sulphide of iron. He also concludes that a slag is most advantageous for this purpose which is most rich in lime and alumina. It is very important in slag-cements that the slag be ground very fine, and be very intimately mixed with the lime. The lime is slaked and bolted, arid then ground mechanically with the slag powder so as to insure thorough incorporation into the mixture. In consequence of the necessity of attaining extreme pulverization of the slag, it is necessary to first dry it. The water which serves to make it granular remains to some extent between the grains and makes bad lumps at time of grinding. It has been attempted to substitute quicklime for slaked lime and use this water for slaking, but unsuccessfully, the slag combining to some extent with the lime and thus weakening the * Annales de les Construction, Juillet* 1886. f Annales des Mines, 1889. vol. II. p. 158. 44 HYDRAULIC CEMENT. cement, while particles of quicklime being left in the cement cause swelling of the mortar after setting. The drying is done in a furnace at a dull red heat. The powdered slag is bolted through a fine sieve about 10,000 meshes per square inch before mixing with the lime. The lime may advantageously be. kept for some time after slaking before being used, as this insures the complete reduction of the quicklime, but the slag seems to deteriorate when kept long after grinding. Fat lime is commonly employed for this purpose, but there seems to be an advantage in using meagre lime on account of the mortar being less likely to crack when used in the air. M. Prost found that there was no advantage to the strength of the cement in using hydraulic lime. Various additions of puzzolanic or other material are also sometimes resorted to for the purpose of preventing the cracking in air when fat lime is used.. This also increases the activity of the cement. The composition of slag-cement usually differs from that of Portland in having a less quantity of lime, more silica and alumina, and more alumina in proportion to the silica. Table IV gives the composition of a number of samples of the leading European slag-cements, taken from Candlot and Tetmajer. Slag-cement is usually slow-setting, but the activity varies greatly with the circumstances of use. The rapidity of action is greater as the proportions of lime and alumina increase. Slag-cement acts better under water than in the air. CLASSIFICATION AND CONSTITUTION OF CEMENT. 4$ TABLE IV. COMPOSITION OF SLAG-CEMENTS. Silica. Alumina. Iron Oxide. Lime. Mag- nesia. Sul- phuric Acid. Loss on Ignition. Not Deter- mined. I 24.80 I9-I3 2.6 7 36.60 6.76 .IO 7-50 0-44 2 24.60 13.46 0.8 4 5O.22 2.6 5 70. 5-40 0.13 3 24.90 I3-46 2.8 3 50.40 I. 2O .IO 6-45 4 24.30 13-85 I-I5 49 50 2.16 .86 6.90 O.28 5 27-45 14.65 r -75 46.20 1.86 72 7.00 o-37 6 25.20 15.23 0.77 50.00 1-35 .72 6.50 0.23 7 20.40 18.59 o 41 50.07 0.50 .08 8.30 8 18.30 18.07 0-34 53-16 0.64 . 18 8.07 9 22.35 12.83 0.64 55.61 2.17 27 4.01 10 27-35 9-13 1.50 50.28 5.72 .40 2.59 ii 20.35 14-05 o.3J 50.26 2.68 39 6.99 12 18.69 9.20 2.14 46-36 4.92 -25 12.19 13 19.87 14.84 0.80 48.54 2-44 .00 8.41 M i8.ii 15-54 0.92 54-72 0-54 37 8.64 15 20.94 14-85 1.03 48.18 3.58 .69 7.22 16 19.24 I7.I5 1.07 54-21 0.81 39 6-39 It is essentially a hydraulic material, and it is especially important that it be kept damp during the early period of hardening, in order that the water necessary to its proper hardening may not evaporate. M. Frost states that slag-cement is sensitive to the action of frost, and should not be used in freezing weather; while Mr. Redgrave declares that it resists frost better than Portland showing a difference of experience in France and England. Mr. Redgrave also says that it may be kept a long time without injury, and if kept free from moisture' that it undergoes no change whatever; while M. Bon- naini states that exposed to air in powder it rapidly loses its hydraulic properties, probably through car- bonization. 46 HYDRAULIC CEMENT. ART. 15. MIXED CEMENTS. The term Mixed Cement is sometimes used to in- clude a considerable number of cements which are formed by a mixture of various products occurring at works where -other cement is made. These mixtures may be made either for the purpose of cheapening the product or of imparting to it certain desired qualities. They consist of admixtures of different grades of cement, of the overburned or underburned portions of clinker, or of foreign material added to the cement. Slag-cements and certain natural cements which, like some of the Rosendales, are made by mixing different grades of clinker are sometimes included under this head, but are not what is usually meant by the term. Mixed cements differ so widely in character that no general discussion of their attributes is possible. Their values depend upon the care used in selecting, pro- portioning, and incorporating the ingredients, and each works has its own method of manufacture. In some cases, light-burned Roman cement is made slow-set- ting by the admixture of grappiers obtained in the slaking of hydraulic lime, with sometimes an addition of Portland cement. The overburned clinker from the manufacture of Portland cement is also sometimes utilized by being mixed with natural cement, a cer- tain amount of Portland cement being added to bring up the initial strength and reduce the rapidity of set. Cement of. this kind is usually sold under the desig- nation of Portland or natural cement, and not accord- CLASSIFICATION AND CONSTITUTION OF CEMENT. 47 ing to its real character. Some of them when carefully and regularly made give good results in practice. ART. 16. GRAPPIERS CEMENT. Grappiers cements are made by grinding to powder the grappiers left from the slaking and bolting of hydraulic lime. Very great care is necessary in elimi- nating all of the free lime from the grappiers by thorough slaking, the operation of slaking and bolting being repeated several times. The grappiers includes the underburned stone, and overburned material formed in contact with the fuel, as well as a certain amount of hard-burned material of too high hydraulic index to slake, and similar in composition and action to Portland cement. This latter is the effective por- tion of the cement, and it predominates in grappiers of good quality. These cements are usually of low index and very slow-setting. They are liable to contain free lime unless carefully handled and usually require exposure to the air after grinding to permit them to become air-slaked. At Teil the grappiers are passed through coarse grinders, which serve to remove all the soft parts. It is then bolted, allowed to air-slake for a month, then bolted again. Finally the parts resisting slaking are ground, steam being present to slake the particles of free lime, after which it is air-slaked before packing for shipment 48 HYDRAULIC CEMENT. ART. 17. SAND-CEMENT. Sand-cement is the name given to material formed by grinding together Portland cement ana sand to an extremely fine powder and a very intimate mixture. It is claimed that a very considerable amount of sand may thus be mixed with the cement without materially reducing its strength, and that the sand-cement so made may still be mixed with the usual proportions of ordinary sand and give good results in use. It is said that the additional grinding given the cement in pulverizing the sand reduces the cement to impalpable powder, thus increasing its power of " tak- ing sand." Experiments also seem to indicate that if sand be powdered separately, a certain amount may be mixed with cement without serious injury to mor- tar made from the cement. CHAPTER III. THE SETTING AND HARDENING OF CEMENT. ART. 18. THE SETTING OF CEMENT. WHEN cement-powder is mixed with water to a plastic condition and allowed to stand, it gradually combines into a solid mass, taking the water into combination, and soon becomes firm and hard. This process of combination amongst the particles of the cement is known as the setting of the cement. Cements of different character differ very widely in their rate and manner of setting. Some occupy but a few minutes in the operation, while others require several hours. Some begin setting immediately and take considerable time to complete the set, while others stand for a considerable time with no apparent action and then set very quickly. The points where the set is said to begin and end are necessarily arbitrarily fixed, and are differently determined usually by trying when the mortar will sustain a needle carrying a specified weight. The beginning of set is usually supposed to be when the stiffening of the mass first becomes perceptible, and the end of set is when the cohesion extends through the mass sufficiently to offer such resistance to any 49 50 HYDRAULIC CEMENT. change of form as to cause rupture before any per- ceptible deformation can take place. It is sometimes stated that the chemical change in- volved in setting is an instantaneous occurrence at about the time we call the beginning of set, and that the gradual hardening then begins and is a continuous process until the maximum strength is reached. How- ever this may be, with some cements a quite notice- able change suddenly shows itself at about this time in the disappearance of water from the surface of the mortar and the sudden stiffening of the mass. ART. 19. THE HARDENING OF CEMENT MORTAR. After the completion of the setting of the cement the mortar continues to increase in cohesive strength over a considerable period of time, and this subsequent development of strength is called the hardening of the cement. The process of hardening appears to be quite dis- tinct from, and independent of, that of setting. A slow-setting cement is apt, after the first day or two, to gain strength more rapidly than a quick-setting one; but it does not necessarily do so. The ultimate strength of the cement is also quite independent of the rate of setting. A cement imperfectly burned may set more quickly and gain less ultimate strength than the same cement properly burned, but of two cements of different composition the quicker-setting may be the stronger. There is as wide variation in the rate of hardening of different cements as in the rate of setting: some THE SETTING AND HARDENING OF CEMENT. 51 gain strength rapidly and attain their ultimate strength in a few days, while others harden more slowly at first and continue to gain in strength for several years. The rate of early hardening gives but little indication of the ultimate action of the cement, as the final strength of the mortar may be the same, however rapidly the strength is attained. Portland cement usually hardens more rapidly and gains its maximum strength more quickly than natural cement, and also, as a rule, the Portland cement attains greater final strength when used in the same manner. Of two cements of the same class, however, it is not safe to infer that that which most rapidly gains strength will prove the stronger and more per- manent material ; in fact, where an abnormally high strength is shown in a few days the presumption as to final strength is against the cement giving such result, and in favor of one hardening at a more moderate rate. The rate at which cement should harden for a given use depends, of course, upon the necessity of devel- oping early strength in the work. For many pur- poses, such as most subaqueous construction, high early strength is quite desirable if not necessary ; but for most engineering work a very rapid hardening does not seem necessary, and better results may often be obtained by the use of a material of more gradual action. ART. 20. CHEMICAL THEORY. Very little is definitely known concerning the chem- ical reactions which take place in the process of setting 52 HYDRAULIC CEMENT. and hardening of cement-mortars. Many theories have been proposed to account for the phenomena by different observers, based mainly upon the study of the properties of various compounds of lime, silica, and alumina formed synthetically. Chemical analysis shows the proportions of the various elementary sub- stances of which the cement is composed, but not their state of combination ; and the action of a cement may be greatly modified by altering the condition in which the ingredients exist, through changing the manipulation in manufacture, without altering their relative quantities. M. Fremy considered Portland cement to be very complex in composition, and ascribed the setting to the action of lime upon certain puzzolanic compounds, composed of double silicates of lime and alumina, the calcination of the clay giving rise to a porous material which absorbs the lime by capillary affinity. M. Landrin concluded that a substance correspond- ing to the formula 3SiO 2 ,5CaO is found in both Port- land cements and puzzolana, and he considered this to be the active element in the hardening of cement, although he states that aluminate of lime contributes to the setting and accelerates that action. Prof. Le Chatelier, from his study of Portland cements, explains the phenomena of setting by show- ing that certain salts, including the aluminate and silicate of lime which form the active elements of Portland cement, while soluble in an anhydrous state, form insoluble salts when hydrated. When they come into contact with water in mixing mortar the anhy- drous sale enters into solution, then, becoming hy- THE SETTING AND HARDENING OF CEMENT. 53 drated, the hydrate is precipitated from the saturated solution in a crystalline form. Those salts which are thus capable of being dissolved in an anhydrous state and then becoming hydrated arrive at stability in two ways by decomposition and by combination. The tricalcic silicate, which is the essential element of Portland cement, is decomposed in presence of water to a hydrated monocalcic silicate and a hydrate; thus Si0 6 Ca 2 + Aq = SiO,,CaO,2.5H,O + 2CaO,H,O. The monocalcic silicate crystallizes in the form of needle-like crystals and the hydrate in hexagonal lamina visible to the eye. The tricalcic aluminate is hydrated by simple combination with the water. Al 2 O 6 Ca 3 -f Aq = Al a O 3 .3CaO, I2H 2 O. The double silicate of alumina and iron, 2SiO.,, Al a O s ,3CaO, is thought to be quite inert in Portland cement, and to merely serve the purpose of assisting the combination of the silica and lime by acting as a flux during burning. It seems, however, to be an active element in slag-cement, forming by combina- tion with lime in presence of water the same com- pounds that are produced in the setting of Portland cement. The difference in its action is explained by the fact that in the slow cooling of Portland cement the salt exists in crystalline form, while through the sudden cooling of the slag it is made vitreous, and is then in condition to be attacked by the lime. The first setting of Portland cement is attributed to 54 HYDRAULIC CEMENT. the hydrating of the aluminates and ferrites of lime, while the subsequent hardening is due to the slower progress of the hydration of the tricalcic silicate. The rapidity of set is therefore dependent upon the rela- tive proportions of aluminates and silicates. When the burning is done at a low temperature, therefore, the aluminates, which are first formed, will cause a rapid set, while as the degree of burning becomes greater the aluminates give place to silicates which cause the setting to become slower and the subse- quent gain in strength greater. The aluminates are thought to add but little to the final strength of the mortar, as they are not permanent compounds, but are acted upon by water and various salts with which they are likely to come in contact in the work. Cements of low hydraulic index harden more rapidly and gain their full strength earlier than those of high index. They are more nearly of the composition which, according to the theory, should give the highest proportion of active ingredients, while those of high index have a surplus of silica and alumina, forming inert material. It is perhaps questionable whether in all cases this so-called inert material is in reality inert in the final hardening of the cement. Sometimes those cements which from this cause harden very slowly continue to gain in strength over a long period, and ultimately surpass those which gain strength more rapidly in the beginning; and it is quite possible that this overclayed portion, which is of puzzolanic char- acter, may bear an important part in the final harden- ing. Most slow-setting cements have a period during THE SETTING AND HARDENING OF CEMENT. 55 which they lose strength after hardening for several months, probably due to the decomposition of salts formed by the parts of too low hydraulic index during the burning. This loss of strength is usually tem- porary when the cement is of normal composition; but if it be overlimed, the loss of strength may continue, to the final destruction of the mortar. The experiments of M. Candlot indicate that the presence of carbonic acid is essential to the hardening of hydraulic cement-mortar. He found that if the mortar were placed in distilled water, frequently renewed, it became gradually decomposed, and finally lost all coherence; but the presence of carbonic acid, as is common in all natural waters, prevented this action and caused proper hardening to take place. ART. 21. INFLUENCE OF CALCIUM SULPHATE. The action of sulphate of lime to slacken the rate of setting of Portland cement is well known. In Ger- many it has been common to utilize it, for the purpose of regulating the rate of set, by adding powdered gypsum to the cement. M. Candlot * has made a careful study of the influ- ence of sulphate of lime upon the action of Portland cement. He found that the increase in time required to set varied with the quantity of sulphate added ; an addition to a quick cement of from I to 4 per cent being sufficient to change the time of set from a few minutes to several hours. Cement which has been * Ciment et Chaux Hydrauliques (Paris, 1891). $6 HYDRAULIC CEMENT. made slow-setting by the addition of calcium sulphate becomes again quick-setting with age, more or less rapidly, as it is or is not exposed to the air. In some cases the cement by exposure soon becomes quick- setting, and then by longer exposure again becomes slow. When cement treated with the sulphate of lime has regained its quick action by exposure, it may again be made slow by the addition of a small quantity of lime. Fresh cement with sulphate of lime added, and setting slowly in consequence, will set rapidly if the mortar be mixed with a solution of the carbonate of soda. " Cement having sulphate of lime added set more rapidly when mixed with sea-water than with fresh water, and that which had been exposed and regained its former activity set more rapidly when mixed wet than when mixed stiff. " The addition of a small quantity of calcium sul- phate to Portland cement augments the strength. When the mortar is kept in sea-water and the propor- tion of sulphate exceeds I or 2 per cent, the mortar cracks and perhaps disintegrates. When the cement containing the sulphate was kept in sacks during several weeks it showed less strength during the early period of hardening." M. Candlot concludes from his experiments that the explanation of the action of sulphate of lime lies in the fact that in the presence of water the sulphate combines with the aluminate of lirae forming the compouHd, Al !1 O 3 ,3CaO,2.5(SO. 1 CaO), which crystal- lizes with a large quantity of water. The action depends upon the fact that the aluminate is insoluble THE SETTING AND HARDENING OF CEMENT. 57 in lime-water, and, as most of the quick-setting cements contain a certain quantity of free lime, when the cement is gauged the lime at once enters into solution and prevents the action of the aluminate until the sulphate is combined with it. When the cement becomes old, the free lime becomes carbonized, and fails to prevent the immediate solution of the alumi- nate. Aluminous cements burned at low temperatures often contain considerable aluminate of lime, and these may bear an addition of 5 to 10 per cent of sulphate without loss of strength. The proportion of sulphate must always be limited to what may be neutralized by the aluminates. ART. 22. INFLUENCE OF CALCIUM CHLORIDE. M. Candlot has also made a careful study of the effect upon the setting and hardening of cement-mor- tar of chloride of calcium, either dissolved in the water with which the cement is mixed, or that in which the mortar is immersed. He found that Portland cement gauged with water containing a few grammes per litre of chloride of calcium sets more slowly than if gauged with pure water; while if the solution of chloride be concentrated, 100 to 400 grammes per litre, the set- ting is very rapid. " The influence of calcium chloride in weak solution upon the set of Portland cement may be attributed to the salts which determine the set entering into solu- tion more slowly in that solution than in pure water. Hydrate of lime agitated with a large excess of water 8 HYDRAULIC CEMENT. is dissolved much less in the chloride solution than in pure water, while with the aluminate of lime this result is much more marked. " If cement of Vassy, quick-setting, be gauged with a solution of chloride of calcium, 20 to 40 grammes per litre, the setting is about the same as with pure water. If the cement be diluted to a fluid with the same solution it will not set or harden. Portland cement treated in the same way hardens very slowly, but acquires a hardness comparable to that given by fresh water. " Feeble solutions of chloride of calcium have no appreciable effect upon cements exempt from alumina, like certain grappiers cements composed almost ex- clusively of silicate of lime. " From this the conclusions are drawn: " I. That in Portland cement the aluminate exists in feeble proportions; that it acts in an energetic manner upon the set, but very little upon the hardening, which is caused by the silicate of lime. "2. That in the Vassy cement the aluminate of lime is the essential element, and determines both the set- ting and the hardening; the role of the silicate being unimportant, especially during the early period of hardening. "3. That in the phenomena of setting the relative quantities of the elements present determine the action. The solution of chloride in presence of a large quantity of the aluminate of lime perhaps does not hinder the hydration and crystallization; but if, on the contrary, a small quantity of aluminate be mixed in an THE SETTING AND HARDENING OF CEMENT. 59 excess of chloride solution, the action of that prepon- derates, and the aluminate will not enter into solution. ' A weak solution of CaCl has the property of pro- voking the rapid hydration of lime. A cement con- taining an excess of free lime, gauged with pure water, swells and disintegrates under the slow expansive action of the free lime. The same cement gauged with a solution of CaCI, 30 to 60 grammes per litre, does not swell, because the lime is slaked before the set. 44 As already stated, when Portland cement is mixed with a solution of 100 to 400 grammes per litre CaCl it sets very quickly. This set is accompanied by a strong rise in temperature. This only occurs with a fresh cement. With an old cement the setting re- mains slow, no rise in temperature is produced, and the mortar swells and disintegrates. " Mortar of cement gauged with a concentrated solution of CaCl is disintegrated if placed in water immediately after setting, but 15 or 20 hours after- ward it may perhaps be submerged without loss of strength. " The action of a concentrated solution CaCl upon Portland cement is due to the fact that aluminate of lime is attacked very energetically by that solution. While it is very slightly soluble in a feeble solution, it is dissolved in large quantities in a concentrated solution. " When a fresh cement is agitated with a concen- trated solution of CaCl it dissolves not only the alu- minate, but the oxide of lime. The lighter the cement is burned, the more it will dissolve. When an old 60 HYDRAULIC CEMENT. cement is agitated with the concentrated solution of CaCl the aluminate dissolves but very little. ART. 23. EFFECT OF SAND. Cement is ordinarily employed in mortar formed by mixing it with sand, and the action of the mortar is necessarily largely affected by the nature and quantity of sand used. When the cement is finely ground and the sand of good quality, a mortar composed of equal parts of each, as a general thing, finally attains a strength as high as, or higher than, the neat cement. Cements of different characters, however, vary considerably in their power to " take sand " without loss of strength; some of the weaker ones may not be able to take more than half their weight of standard sand, while others can be mixed with considerably more than their own weight without loss of strength at the end of one year after mixing. All have a certain limit within which they may be made stronger by an admixture of good sand than they would be if mixed neat. Cement mixed with sand always hardens more slowly than' neat cement, and requires a much longer time to attain its maximum strength. As the proportion of sand to cement is increased both the rate of hardening and final strength are diminished. Within certain limits, however, the strength of mortar increases over a longer period of time as the proportion of sand becomes greater, and as the time of observation is extended the loss of strength due to larger propor- tions of sand becomes less. Thus a good Portland THE SETTING AND HARDENING OF CEMENT. 6l cement in a mortar containing I part sand to I of cement at the end of a year may be expected to be stronger than mortar of neat cement. At the end of three years the I to I mortar should be much stronger, while a 2 to I may be as strong as the neat mortar. At the end of four or five years the 2 to I mortar may be on even terms with the i to I, while a 3 to I mortar may have steadily gained to perhaps three fourths the strength of the others, where it probably stops. Beyond the limit where the quantity of cement is sufficient to fill all the interstices in the sand the ultimate strength diminishes very rapidly as the pro- portion of sand increases. Clean and sharp sand usually gives a higher strength in mortar than that containing an admixture of clay or earth, or that composed of rounded grains. Coarse sand also usually gives greater, strength than that which is very fine. It is often difficult, however, to judge of the quality of sand without experimenting with it. In some cases a small amount of fine clay does not appear to injure the strength of the mortar, while a judicious mixture in the sand of grains of various sizes may be of benefit, through reducing the volume of interstices. A mortar composed of sand and cement usually possesses greater ability to adhere to other surfaces when coarse sand is used than if the sand be fine. ART. 24. WATER USED IN GAUGING. The quantity of water used in mixing the mortar is one of the most important conditions; the less the 62 HYDRAULIC CEMENT. quantity, provided there be sufficient to thoroughly dampen the mass of cement, the quicker will be the set. With some Portland cements, changing the quantity of water used in mixing neat cement from 20 per cent to 25 per cent of the weight of cement doubles or even triples the time required for the mortar to set. In other cases the effect is compara- tively slight. When the quantity of water used in mixing is suffi- cient to reduce the mortar to a soft condition the hardening as well as the setting becomes more slow, and the strength during the early period is less than if a less quantity be used. This difference disappears to some extent with time, and the mortar mixed wet may eventually gain nearly as much strength as though mixed with less water. When the quantity of water employed is not suffi- cient to reduce the mass to a plastic condition, the mortar will not be so thoroughly compacted, and will not reach the same strength as when made plastic, unless pressure be applied to it. But if just sufficient water be used to thoroughly dampen the mortar, and pressure be applied to expel the air and close the voids, the early strength will be greater than when more water is used. This difference, like the former one, disappears to a certain extent with time, but the final strength is usually greater with the less quantity of water. According to Prof. Le Chatelier, the solidity of the crystalline mass varies with the form, dimensions, and mode of aggregation of the crystals. In general, the strength of a single crystal is greater than its adher- THE SETTING AND HARDENING OF CEMENT. 63 ence to neighboring crystals. Long needle-like crystals give greatest strength, and crystals have this character more as the solution is more strongly supersaturated. The nature of the water used in mixing may also affect the rate of setting. When sea-water is used the setting is usually slower than with fresh water, the chloride and sulphate of magnesia being the principal retarding elements. Cements with a high hydraulic index show a less difference between fresh and sea water than those of low index, and well-burned cem- ents less than imperfectly burned ones. The experi- ments of M. Candlot indicate that this is due to the action of the salts mentioned above upon the aluminate of calcium, and that those cements containing the highest percentage of aluminate are affected the most by being mixed with sea-water. Water containing sulphate of lime in solution retards setting (see Art. 21). Mortar kept immersed in sea-water usually hardens more rapidly than that kept in fresh water. This difference is commonly much more noticeable with neat cement than with mortar containing considerable proportions of sand. The strength gained in sea- water, however, although gained much more quickly, is generally less in final amount than that in fresh water. There is, however, a very great difference between various cements in this particular. Cements with a low hydraulic index show the great- est difference between sea and fresh water. Those containing small quantities of free lime give much greater early strength in sea than in fresh water, but are also sooner disintegrated by the action of sea-water. HYDRAULIC CEMENT. ART. 25. EFFECT OF ENVIRONMENT. Cement-mortar kept under water ordinarily hardens more rapidly in the early period than that exposed to the air, but usually that kept in air ultimately reaches greater strength. The highest strength is commonly produced by keeping the cement for a time in water, and later placing in dry air. Nearly any cement- mortar will harden more rapidly and attain greater strength if kept moist during the operation of setting and the first period of hardening than if it be exposed at that time to dry air. A sudden drying out about the time of completing the set usually causes a con- siderable loss of strength in mortar, and frequently the mortar so treated is filled with drying cracks. This result is usually more marked when the cement is mixed with a large quantity of water to a soft con- dition. The nature of the water in which the mortar is allowed to harden is of more importance to its strength than that of the water used in gauging. When the mortar is to be kept in air, the nature of the water used in mixing becomes more important, although probably the variations in ordinary natural water are rarely sufficient to produce any appreciable difference in the strength of the mortar. Mortars gauged with sea-water harden best in air. ART. 26. EFFECT OF TEMPERATURE. The temperature of the water used in mixing has an important bearing upon the time required for setting; THE SETTING AND HARDENING OF CEMENT. 65 the higher the temperature, within certain limits, the more rapid the set. Many cements which require several hours to set when mixed with water at a tem- perature of 40 Fahr. will set in a few minutes if the temperature of the water be increased to 80 Fahr. Below a certain inferior limit, ordinarily from 30 to 40 Fahr., the mortar sets with extreme slowness or not at all; while at a certain upper limit, in some cements between 100 and 140 Fahr., a change is suddenly made from a very rapid to a very slow rate, which then gradually decreases as the temperature in- creases, until practically the mortar will not set. The temperature of the cement, and that of the air in which the mortar is placed during setting, influence the rate of setting in about the same manner as that of the water. In case the air in which the mortar is placed be dry, the setting will usually be somewhat more rapid than if it be moist; and if it be too dry, the rapid evaporation of the water from the surface of the mortar may cause drying cracks in the mortar. Quick-setting cements usually show a rise in tem- perature during setting, due to the rapidity of the action which takes place. It has been suggested that the time occupied by the setting would be better shown by observing the period of advanced tempera- ture, than by noting the stiffening of the mortar, as is common. Most slow-setting cements, however, do not show sufficient change of temperature, if any takes place, to be appreciable; and the rise in tem- perature, where it does take place, may not always be the result of the process of setting. If the air at the time of mixing mortar be suffi- 66 HYDRAULIC CEMENT. ciently cold to freeze the mortar before it can set, it will not set while frozen; but most cements will do so after thawing out, and but few of them will be injured by such freezing in so far as their ultimate strength is concerned. Recent experiments* have, however, seemed to show that mortar may set while frozen if it remains in that condition fora sufficient length of time. The temperature of the water with which cement- mortar is mixed has a quite appreciable effect both upon its rate of hardening and its ultimate strength, and the temperature of the air at the time of mixing has a similar effect. The lower the temperature at which the mixing is done, the slower the hardening and the greater the final strength. This difference is not sufficient to be important at ordinary air tempera- tures in so far as the use of mortar is concerned, but is quite appreciable in making comparative tests. The temperature of the air or water in which the mortar is immersed during the time of hardening has a very appreciable effect upon the rate of hardening of many cements. This effect differs very radically for different material ; with some the process is greatly accelerated by keeping them hot as compared with what would be the result in cold air or water; others are not appreciably affected, while still others seem to be retarded in their hardening by the application of heat. This variation is to be found among cements of the same class, and is seemingly independent of their value. Cements of low hydraulic index usually show the greatest gain in rate of hardening under the action of heat. * Paper by Cecil B. Smith; Trans. Canadian Soc. C. E. THE SETTING AND HARDENING OF CEMENT. 67 ART. 27. EFFECT OF AGE UPON CEMENT. The effect upon a cement of retaining it a long time before using depends upon the nature of the cement and the method of keeping. When the cement is enclosed so as to prevent the access of air, as in barrels, it may usually be preserved for a considerable time without experiencing any alteration, provided it is kept dry. When exposed to the air the cement commonly undergoes more or less alteration. Portland cements of good quality are usually but slightly affected, as they are composed for the most part of stable com- pounds. The change which occurs is limited usually to making the cement slower-setting. Where the cement is originally slow-setting this effect may be very slight, and the cement may perhaps be retained for a long period, two or three years at least, without appreciable change in its properties. A hard-burned cement, originally quick-setting, usually becomes slower-setting with age, but commonly without injury as to its ability to harden and its ultimate strength. Light-burned cements, particularly Roman cements, are affected in much greater degree by age. These cements not only become slower-setting when exposed to the air, but commonly also they gradually lose the power of hardening and become finally inert in many cases becoming puzzolanic material, the activity of which may be restored by the addition of slaked lime. The changes which occur in cements kept in dry air are attributed to the action of carbonic acid upon the 68 HYDRAULIC CEMENT. free lime which they may contain, and perhaps also upon the less stable compounds, as the aluminates of lime, which contribute to the rapid set and are found most plentifully in the light-burned material ART. 28. EFFECT OF FINENESS. The degree of fineness to which the cement is ground is always very important in its effect upon the strength of mortar made from the cement. The valuable part of the cement is practically only that portion which is ground extremely fine to an impalpable powder. The coarse parts are not altogether inert, but are more or less active, depending upon the size of the grains of which they are composed. As the clinker obtained in burning cement is very dense and hard, it is ground with difficulty, and the coarser particles are apt to be of the best burned, and therefore most valuable, part of the material. Coarse grinding is likely therefore to leave in a useless condition much of what should be the most active portion of the cement. The rate of setting is accelerated as the fineness to which the cement is ground becomes greater. In a finely ground cement the amount of active material is greater than in one coarsely ground, and the chemical reactions which take place in setting are facilitated by fine subdivision of the particles. When the cement is gauged with sea-water the rate of setting is less in- fluenced by the fineness. The hardening of cement mixed neat is not greatly affected by fineness; that finely ground usually har dens more rapidly, but attains less final strength than. THE SETTING AND HARDENING OF CEMENT. 69 when more coarsely ground. The hardening of the coarsely ground cements is more gradual and regular, and the ultimate strength greater. Cement, when used, is commonly mixed with sand, and the attainment of strength in sand-mortar rather than mortar of neat cement is therefore of impor- tance. The finer ground the cement, the greater its resistance when mixed with sand, both in the earlier and later stages, and also the sooner will it reach its ultimate strength. The effect of fine grinding is much greater when the proportion of sand to cement is large, as the power of the cement to take sand without diminution of strength is thereby greatly in- creased. The coarser particles of the cement may be considered as practically inert material, which acts rather as sand than as cement in the mortar, and the power of the cement to harden and develop strength in sand-mortar is thus dependent upon the amount of fine material contained in it. The adhesive strength of cement increases very rapidly with the fineness, at least in the early period of hardening. The difference between coarse and fine grinding is greater in the early period of hardening than later. The fine cement hardens much more rapidly, but the coarse cement, especially in rich mortar, often reaches nearly the same ultimate strength. The attainment of extreme fineness may not therefore always be eco- nomical when the extra cost is considered. CHAPTER IV. THE SOUNDNESS OF CEMENT. ART. 29. PERMANENCE OF VOLUME. THE permanence of any structure erected by the use of cement is dependent upon the power of the cement, after the setting and hardening processes are complete, to retain its strength and form unimpaired for an indefinite period. Experiment has shown that mortars made from cement of good quality frequently continue to gain in strength and hardness through a period of several years, or at least that there is no material diminution of strength with time; and that changes of temperature, or in the degree of moisture surrounding it, produce no injurious effects upon the material. This durability in use is commonly known as \.\\Q permanence of volume or soundness of the cement. Heat has the same effect in causing expansion or contraction of cement-mortar that it has upon other materials. The coefficient of expansion for neat Portland cement, according to a series of experiments at " 1'ecole des Ponts et Chaussees," is about the same as that of iron. For sand-mortar the coefficient is somewhat less. When mortar which has been immersed in water is 70 THE SOUNDNESS OF CEMENT. ?I transferred to dry air a slight contraction may take place in volume, together with an increase in strength, while a transference the other way may produce the opposite result; but no distortion of form or disinte- gration of the mortar will take place in either case if the cement be of good quality. Sometimes cement when made into mortar sets and hardens properly, and later, when exposed to the action of the atmosphere or water, becomes distorted and cracked, or even entirely disintegrated. If the composition deviates but slightly from the normal this process of disintegration may not show itself for a con- siderable time, and proceeds very slowly. It thus becomes an element of considerable danger, as it is liable to escape detection in testing the cement. The unsoundness of cement may be occasioned either by defective composition, causing the mortar to yield to the action of expansions proceeding from within itself, or by exterior agencies which act upon ingredients of the cement susceptible to their influ- ence, or are permitted to act by the method of mak- ing and using the mortar. Very little is definitely known concerning these various destructive agencies, and there is considerable doubt concerning the causes which operate in many instances. The expansive action is commonly attributed to free lime or magnesia. The exterior agencies are the action of frost, of dry air and heat, and of sea-water. Most cements probably contain small amounts of the expansive elements, which when in small quantity act with extreme slowness, and perhaps produce no visible effect for several months after the use of the 72 HYDRAULIC CEMENT. mortar; then there occurs a decrease of strength, which probably disappears with time. Cements of lew index, which gain strength very rapidly in the earlv period, are quite apt to act in this manner, and occasionally, as already noted, the cement may not possess sufficient strength to resist, and the expansive action continues to ultimate distintegration. The term Permanence of Volume, if limited to the power of the material to resist actual change of form or dimension in the body of mortar, is not necessarily synonymous with soundness, if by soundness we desig- nate its power to resist disintegration over a long period. Most unsound cements fail by swelling and cracking under the action of expansives. In some cements, however, the failure occurs by a gradual softening of the mass of mortar, without appreciable change of form or dimension, the process being very slow, sometimes not noticeable for several months after the mortar is mixed. ART. 30. FREE LIME. The presence of small quantities of free lime in the cement is doubtless one of the most frequent causes of disintegration in cement-mortar. The lime being distributed through the cement in small particles is hydrated very slowly after the setting of the cement, causing, through its swelling during slaking, a strong expansive force on the interior of the mortar, and pro- ducing an increase in volume, loss of strength, and perhaps final disintegration. THE SOUNDNESS OF CEMENT. 73 The effect of the lime depends upon its physical condition, and is affected by the degree of burning. Lime burned at a high heat slakes much more slowly, and is therefore more likely to be injurious than when burned at a low temperature. Prof. Le Chatelier found that where the addition of quicklime formed by burning the carbonate produced no result, that obtained from the. nitrate caused swelling and cracking in the mortar. The presence of free lime in hard-burned cements of low index is therefore of special importance, and must be carefully guarded against by securing accurate composition and com- plete reactions in burning. The fineness of the cement also modifies the action of the free lime, as finely divided material will slake quicker than coarse grains, and the lime is more apt to become hydrated before setting is completed ; or if the cement be exposed to the air before use, the lime in a fine state will sooner become air-slaked. If free lime be present in such condition that it becomes slaked before the initial set of the cement, it causes no injury. If it becomes slaked during the setting or first period of hardening, the strength of the mortar may be reduced, being rendered less compact and more porous. In case this action be not sufficient to cause disintegration the loss of strength may to a great extent subsequently disappear. When the slaking of the free lime does not take place until a longer period of time has elapsed, the danger in the use of the cement is more serious. When the expansive action becomes sufficient to over- come the tenacity of the mortar, disintegration ensues. 74 HYDRAULIC CEMENT. If the expansive action be not sufficient to overcome the tenacity of the mortar, an increase in volume and loss of strength in the mortar may take place, the extent of which is dependent upon the relative inten- sities of the expansive and resisting forces. These effects may afterward gradually disappear, but they probably have the effect of making the mortar more easily attacked by external agencies. Cements of low hydraulic index, which gain strength rapidly in the early period of hardening, are particularly liable to contain appreciable quantities of free lime, which is frequently shown by a loss of strength when tests are extended over a considerable period of time. Mortar kept under water is acted upon much more rapidly than that exposed to dry air. ART. 31. MAGNESIA. Free magnesia in cement acts very much like free lime. The action of magnesia, however, is much slower than that of lime, and for this reason it is a more serious defect, because less likely to be detected in the tests applied before using. Prof. Le Chatelier mixed 5 per cent of lime and of magnesia with two samples of Portland cement, and observed the time required for the swelling to begin, resulting as follows: Swelling at o C. Swelling in Water at 100 C. Commenced. Ended. Commenced. Ended. 5*lime S% magnesia. in 3 hours in 6 monhts 36 hours immediately 6 hours in hour 40 hours THE SOUNDNESS OF CEMENT. 75 In hard-burned cements of low hydraulic index any magnesia which may be present is likely to be in a free state, and hence the percentage of magnesia in such material should be low, and specifications fre- quently limit the amount of magnesia that may be allowed in Portland cement. When mortar fails from this cause the expansive action may not be shown for several months after the mortar has set, and then in a comparatively short time the swelling, cracking, and disintegration take place. Dr. Erdmenger made a large number of experiments upon the effect of adding small quantities of magnesia to Portland cement, and found * that all expanded in water and contracted in air; those containing consid- erable percentages disintegrated, beginning after about 90 weeks. The fineness of the grains of magnesia, like those of lime, is important as affecting the intensity of the effect. Prof. Le Chatelier found that where magnesia coarsely ground produced swelling and cracking, the same quantity of finely ground magnesia produced swelling, but without distorting or cracking the mortar. The extent of the change in volume varies with the quantity of magnesia, increasing rapidly as the quantity is increased. The expansion is reduced as the mortar is mixed more wet, being less as the porosity of the mortar is greater. In light-burned cements the danger of free mag- nesia is greatly lessened, and many good cements of * Journal Society of Chemical Industry, vol. xil. p. 927. 76 HYDRAULIC CEMENT. this class contain large percentages of magnesia. An overdose may, however, be an element of danger in .any case, and several failures of mortar in Europe have been attributed to magnesia in both hard-burned and light-burned cements. M. Durand-Claye gives an instance * where the cement in a work swelled and cracked, and an exam- ination showed that it contained 16 to 28 per cent of magnesia. Experiments were made by mixing mag- nesia with good cement, and swelling resulted. The rock which served for the manufacture of the cement contained a large proportion of magnesia, which was probably present in an uncombined state in the cement. The time in which the swelling occurred was found to depend upon the amount of water available. When mixed with the normal quantity and left in dry air no expansion took place. It is therefore only dangerous in water. Analyses of this cement are given as follows: Silicious Silica. Alumina. Iron. Lime. Magnesia. Sulphur. Loss on Sand. % % % % % % Ignition. 14.80 8.OO 4.60 47.30 24.30 0.60 0.40 18.30 2.95 3.60 44.80 28.15 2.30 1.90 0.35 20.70 3.35 3.65 43.30 26.70 0.15 i. So This is not similar to the magnesian cements com- monly used in the United States on account of the extremely low hydraulic index. If it be assumed that in the normal cements of this class the magnesia acts like lime in combining with silica and alumina, the presence of free magnesia might be accounted for by * Annales des Fonts et Chaussees, 1886, vol. I. p. 845. THE SOUNDNESS OF CEMENT. 77 the lack of a sufficient quantity of these hydraulic elements. The silicates and aluminates of magnesia are known to possess, like those of lime, the property of harden- ing under water. Their action, however, is said to be much slower than that of the lime salts, and it has been suggested that the presence of the magnesian salts might be sometimes injurious on account of their hydrating after the hardening of the lime salts in the cement. It has not been shown, however, that any swelling takes place in the setting of these salts, and the effect may be to contribute to the final strength. Whatever the nature of the process, it is certain that some good magnesian cements continue to increase in strength over a long period, the proportionate increase in the later period being much greater than for Port- land cement of even very moderate action. ART. 32. ALUMINATE OF LIME. The exact role of aluminate of lime is in many cases a matter of considerable doubt. M. Bonnami considers the basic aluminates to act as expansives in hydrating like lime, being decomposed in presence of water. The action of water upon these aluminates is very rapid, heat being given off, with the result of greatly accelerating the set of cements containing them in appreciable quantities. The disintegration of mortar has in certain in- stances been attributed to the hydration of aluminates subsequent to the set. It seems probable, however, that where the aluminate is present in small quan- ?8 HYDRAULIC CEMENT. titles that its hydration usually takes place before the setting of the cement, and in fact is the first cause of the setting action. Aluminate of lime is freely acted upon by calcium sulphate, as stated in Art. 21, forming the sulpho- aluminate (3CaO.Al 2 O 3 -f 2.sSO 4 Ca -f- 6oH 2 O), which crystallizes with great expansive force. When a cement containing aluminate is exposed after setting to water containing calcium sulphate the combination of the sulphate with the aluminate may take place, with the result of causing swelling and, if sufficient in quan- tity, disintegration of the mortar. Aluminates should therefore be avoided in cements to be used for mortar to be exposed to the action of sea-water, as the sulphate of magnesia acts upon the lime of the cement, forming the sulphate of lime, which then combines with the aluminate of lime, producing the expansive action. For this reason cements high in alumina are not considered desirable for marine work. ART. 33. SULPHUR COMPOUNDS. The action of sulphur in cement is extremely vari- able, depending upon the state in which it may exist and the nature of the cement. The effect of adding sulphate of lime for the purpose of rendering the setting slower has already been discussed (Art. 21). This action depends upon the presence of aluminate of lime in sufficient quantity to take all of the sulphur into combination. When the sulphate is added in ex- cess, or to a cement without the aluminate, it remains THE SOUNDNESS OF CEMENT. 79 soluble in the mortar and is gradually dissolved out, having only the tendency to make the mortar porous. It is pointed out by M. Candlot that if the aluminate be prevented from acting by adding slaked lime to a cement with which sulphate of lime has been mixed, the combination of the sulphate and aluminate may take place after setting, causing the destruction of the mortar through its expansive action. The effect of the existence of sulphate in the material before burning the cement may be quite different from that of adding it afterward. In Roman cement the two seem to give analogous results; but in the heavily burned cements it may be in a state not readily soluble, and hence slow in acting upon the aluminate, thus causing the expansion to be delayed until after the set. Mr. Spachman, who experimented upon the pro- duction of Portland cement from alkali waste, con- cludes * that the danger in using material containing too large proportions of sulphate of lime is due to the likelihood of forming calcium sulphide, CaS, during the burning, which afterward forms, with the iron oxide of the cement, the sulphide of iron, FeS. This upon exposure is oxidized to a sulphate of iron, changing the color of the cement to a brown, and causing it to lose much of its activity, in some instances scarcely setting at all. Mr. Spachman gives the limit of about 5# of SO 4 Ca as what may be safely used. Prof. Tetmajer states that calcium sulphate in Port- land cement sometimes acts as an expansive, through * Journal Society of Chemical Industry, vol. xi. p. 497. 80 HYDRAULIC CEMENT. the fact that it is readily oxidizable and expands in oxidation. In slag-cements the presence of calcium sulphide is thought to be less injurious. According to M. Prost, it gives a green color to the cement when kept in water, but without injury to its strength. In the air it may cause the mortar to crack. ART. 34. EXTERIOR AGENCIES. The principal exterior agencies which operate to cause the destruction of mortar are changes in tem- perature or in humidity, and the nature of the water with which it may be in contact. Exterior mechani- cal agencies, such as the shocks of waves or of ice and sand produced by a current, have an abrasive action and may overtax the strength of mortar in the early period of hardening, but they do not cause disintegra tion through injury to the cement. The effect of frost is to set up a mechanical action through the freezing of water in the pores of the mor- tar and resistance to it probably depends mainly upon the strength of the mortar and its ability to resist this expansion. The nature of the water to which the mortar is ex- posed is important because of the possible chemical action of salts which it may hold in solution. This is shown by the disintegration of mortar in sea-water or in sewer water which is quite sound when subjected to fresh water. According to M. Candlot, " all hydraulic materials are alterable by pure water. A mortar traversed by THE SOUNDNESS OF CEMENT. 8 1 pure water finally loses all coherence, the elements constituting the agglomerant being little by little decomposed. But natural water contains always car- bonic acid, which intervenes in the majority of cases to arrest decomposition and close the pores of the mortar traversed by the water. When the cement does not give up its lime too readily the lime is transformed into carbonate, which forms a deposit in the voids of the mortar. If, on the contrary, the dissolving of the lime in the water which traverses the mortar is abun- dant, there is produced a large quantity of carbonate without cohesion, which is carried off by the water." In regard to the effect of temperature Prof. Le Chatelier says: " At elevated temperatures certain solid hydrates lose their water and are reduced to powder, like the crystalline carbonate of soda, and cause the disintegration of the mortar. This is the case with certain aluminates of lime, especially the alumino-sulphate, but precise experiments are still necessary upon this subject. ' This dehydration occurs in dry air. This explains the well-known fact that certain cements stay for months in water and attain high strength, but the same exposed to dry hot air disintegrate into a sandy mass." M. Candlot says that " aluminous cements are subject to alteration in surroundings exposed to alter- nate dryness and humidity, and also when exposed to a high temperature." It should be remarked, how- ever, that this probably depends upon the alumina b~ing present as basic aluminate of lime, and that cements with a high proportion of alumina, such as 82 HYDRAULIC CEMENT. certain Roman cements, containing considerable sul- phate, commonly give good results when used in situations exposed to changes in humidity. The Louisville cements are a prominent example of this. When cement-mortar during the early period of hardening is exposed to very dry air, the hardening may be interfered with by the lack of moisture neces- sary to admit of the completion of the hydration and crystallization of the cement, thus causing a lack of cohesive strength, and perhaps ultimate destruction of the mortar. Different cements vary greatly in the extent to which they are influenced by this cause; slow-setting Portland cements being ordinarily least and the slag-cements most affected. ART. 35. EFFECT OF SEA-WATER. The destructive effect of sea-water upon hydraulic mortars which are sound in fresh water is probably due to the action of magnesian salts upon the lime of the cement, thus forming sulphate and chloride of calcium. The action of these salts upon the harden- ing of cement-mortars has already been discussed. When mortar in sea-water fails by swelling, the failure is usually attributed either to too large a pro- portion of free lime or magnesia, or to aluminate of lime in the cement. When the cement contains free lime, the expansive action is greatly intensified in sea- water as compared with that in fresh water. This may be explained by the presence of calcium chloride, which increases the rapidity of slaking of quicklime, causing the expansion to be shown sooner, and to act THE SOUNDNESS OF CEMENT. 83 more violently than in fresh water. When the cement contains considerable aluminate of lime the calcium sulphate may act upon it, as indicated in Art. 32, causing the formation of the sulpho-aluminate and the corresponding expansive effect. When mortar made from cement of good quality is exposed to the action of sea-water its durability depends largely upon the permeability of the mortar. The lime salts formed by the action of sea-water are readily soluble, and if the mass is freely permeated by water those salts may be washed out, leaving the mortar more open to the action of the disintegrating agencies. Thus mortar of any Portland cement may be injuriously affected by sea- water if used in such manner as to permit the continuous action of the mag- nesian salts through the mass. The ultimate hardening of mortar in sea-water, as in fresh water, seems to depend upon the action of carbonic acid in forming a protection to prevent the operation of the elements of disintegration. When the mortar resists the penetration of the water so as to prevent its renewal in the interior of the mass, the outside soon becomes protected by the action of the carbonic acid, and effectually prevents further action of the magnesian salts. M. Durand-Claye examined the mortar from a sea- wall where parts of it were disintegrated, and found a large proportion of magnesia, although it was not con- tained in the original rnortar or in the portions of the wall which were still sound. The percentage of sul- phuric acid was also increased in the disintegrated portions, seeming to show that the magnesia had been 84 HYDRAULIC CEMENT. precipitated from the sulphate of the sea-water, and the resulting sulphate of lime had for the most part washed out. Where the water against the wall is under pressure from one side, or where tidal flow keeps the work sub- merged only a part of the time, the action of the sea- water is more strongly felt than in work always en- tirely covered. M. Alexander submitted blocks of cement-mortar to the filtration of both fresh and sea water.* Those in fresh water were unaffected, but those in sea-water were disintegrated in six months. Analysis showed that those in fresh water suffered a slight loss of lime and sulphuric acid, while those in sea- water were much changed by loss of lime and gain in magnesia and sul- phuric acid. M. Alexandre also found that " argillaceous or soft calcareous sand is attacked by sea-water, and mortar containing them may be decomposed although the cement is good." * Annales des Fonts et ChaussSes, 1890, vol. i. p. 408. CHAPTER V. METHODS OF TESTING CEMENT. ART. 36. OBJECT OF TESTS. TESTS of cement may have for their object either the examination of the quality of the material in order to determine its fitness for use, or investigation of the properties of the cement for the purpose of in- creasing knowledge of its behavior under the varying contingencies of use. Where experiments are made with this latter object, the tests to be applied and methods of operation must of course be dependent upon the special point to be investigated. In many instances it may be possible to combine to a certain extent the two objects. This is particu- larly the case where a permanent laboratory is estab- lished to regulate the reception of material for exten- sive works, as in the case of the laboratories connected with the Government experiment stations in Europe. In some of these stations careful examination of every sample of cement in a number of particulars is made, with the result of accumulating a mass of valuable in- formation regarding the characteristics of all the differ- ent kinds of cement. Systematic series of tests of this character possess much greater value as a means of 85 6 HYDRAULIC CEMENT. deducing the laws governing the action of the mortar than special examinations upon particular points, which often fail to take into account the variable nature of the material, and the necessity for exact knowledge of the nature of the cement upon which the tests are made. The French " Commission upon the Method of Testing the Materials of Construction" recommend that in the permanent laboratories cement should be systematically tested in the following particulars: Chemical analyses; fineness; specific gravity; appar- ent density; homogeneity; time of setting; tensile, compressive, flexural, and adhesive strength; perma- nence of volume; porosity; permeability; resistance to decomposition by sea- water; and yield of mortar. Tests of cement, as commonly made for its recep- tion upon engineering work, have for their object only the determination of the quality of the material and its fitness for the use. Tests for this purpose must be made according to some recognized standard, and cannot closely approximate the conditions of use with- out impairing their value as means of judging the quality of the cement. What it is necessary to know about the cement is that it will set and harden into a solid mass, which will firmly adhere to any surface with which it may be in contact, and that it will en- dure through a long time without change of form or loss of solidity. As ordinary tests must be made in a short time, but a few days at most being usually allowed for determin- ing the quality of the material, the problem to be met in testing is to apply such tests as will enable a pre- METHODS OF TESTING CEMENT. 87 diction to be made, from its behavior under them in a short time, as to what it will do in a long time under the circumstances of its use. The difficulty of this with a material varying so greatly in character and in its behavior under various conditions is evident. Having a particular brand of cement whose character- istics are known, it may readily be determined whether a given sample is of normal quality, and something may be predicted of its future from its behavior under short-time tests. Very little, however, can be done .in the way of generalization, and for a new and un- known material it is only possible to state a somewhat indefinite probability as to final results. Tests may be imposed which in nearly all cases will secure good material, but often at the expense of rejecting equally good or better material. This, how- ever, will be unavoidable until the characteristics of the various brands of cement are more fully known, and the tests to which each should be subjected better understood. The tests which are usually imposed to determine the quality of hydraulic cement are those of weight, fineness, time of setting, tensile strength, and sound- ness. Chemical analysis is sometimes made, and spe- cific-gravity test is substituted for that of weight, or both are frequently omitted. Compression-tests are also sometimes added. The greatest weight is usually given to the test of tensile strength, and much greater value is commonly placed upon the results of that test than they deserve. It is much the simplest and best means of making a test for strength, and is very desirable as showing the 88 HYDRAULIC CEMENT. proper hardening of the mortar, but cements cannot be graded in value by the strength attained in a short time. Cement giving high early strength is to be relied upon only in so far as it has been shown by ex- perience capable of subsequently maintaining such strength. The attempt to produce cement which will develop great strength on short-time tests is liable to result in lowering the hydraulic index, or the addition of calcium sulphate, and sometimes in the presence of free lime, giving a material more likely to be unsound than one of more moderate strength. The test for soundness or permanence of volume is an important one, as giving an indication of the prob- able durability of the material; but in this, as in the other tests, a knowledge of the usual action of the material will contribute greatly to the proper inter- pretation of the test. The test for fineness is also important as bearing upon the power of the cement to take sand. It was recommended by the committee of the American Society of Civil Engineers upon a uniform system of testing, that tests for quality be limited to the above three most important tests fineness, tensile strength, and soundness; and this recommendation is now commonly followed in the United States, although the test for soundness as usually made is of little value. ART. 37. APPARENT DENSITY. The apparent density of cement is measured by determining the weight of a given volume of the material. This test is made as a means of showing METHODS OF TESTING CEMENT. 89 whether the process of manufacture has been well con- ducted. If the cement be not thoroughly burned or if it lack homogeneity so that in portions of it the combinations are not complete, the weight is less than when the material is homogeneous and well burned. Variations of composition also affect the weight, so that there may be considerable variations in the weight of various cements of good quality, equally well burned. The apparent density is affected by the fineness to which the cement is ground ; the coarser the particles of the cement, the greater its weight per unit volume. The weight test when employed should therefore be combined with one for fineness to prevent the attain- ment of heavy weight by coarse grinding. The test for apparent density is not usually em- ployed for the reception of material, as it is somewhat indefinite in result. It is, however, sometimes in- cluded in specifications in England, and is used in many European laboratories where a careful study is made of the properties of cement. As the cement powder may be packed in the meas- ure so as to give very different weights for the same volume, it is necessary to use a uniform method of filling the measure in determining the weight. The common method of conducting the test is to pass the powder through a sieve and allow it to fall through a funnel or down an inclined plane through a given height into a measure, which when full is struck and weighed. The height of fall and the size of the meas- ure both affect the result, the cement packing closer in a large than in a small measure. (JO HYDRAULIC CEMENT. In Europe several appliances are used for testing apparent density. Tetmajer's Apparatus. The apparatus of Prof. Tetmajer is used in a number of the leading labora- tories. It is shown in Fig. i, and consists essentially FIG. j. of a cylindrical measure (M) of I liter capacity and 10 centimeters high, provided with ears which catch upon a frame formed of two levers (Z,). The frame is raised and dropped at each turn of the hand-wheel by the cam (O), thus giving a succession of jars to the meas- ure. Above the measure a sieve (7?) is oscillated upon a system of levers which are hinged to the base, and moved by the rod (V), giving two oscillations at each turn of the hand-wheel. The number of revolutions is recorded by the revolution-counter (T). In the operation of the apparatus the cement is filled into the sieve and shaken through by the oscil- lations produced by turning the hand-wheel. It is caught in the measure and jarred down by the raising and dropping of the frame. About 500 revolutions METHODS OF TESTING CEMENT. 9! are necessary to secure the best results in compacting the powder in the measure. The compactness is found to vary with the rapidity of motion, a moderate speed of about 200 revolutions per minute giving a maximum effect, and being considered most desirable. Inclined-plane Apparatus. The inclined-plane ap- paratus for apparent density has been used in a num- ber of forms, one of which, employed in France and recommended by the " Commission des Methodes d'Essai des Materiaux de Construction," is repre- sented in Fig. 2. FIG. 2 The inclined plane is formed of sheet zinc 30 cm. long and inclined at 45 with the horizontal. It is 10 cm. wide for the upper two thirds of its length, and through the lower third diminishes gradually to 5 cm. at the lower end. The zinc is turned up at the sides to form a channel in which the material may slide. 92 HYDRAULIC CEMENT. At the point where the larger plane begins to narrow a second sheet of zinc, 20 cm. long and 10 cm. wide, is set at right angles to the first, leaving an opening of 1 cm. The measure 10 cm. high and of I litre capacity is placed with its top 5 cm. below the lower edge of the plane. The cement is poured in small quantities on the summit of the secondary plane so slowly as not to clog the opening between the planes, until the meas- ure is full when it is struck and weighed. A single inclined plane of somewhat greater length (50 cm.) is sometimes used, the cement being sifted upon the upper end and allowed to slide directly into the measure. It is said, however, to give less uniform results than the double plane unless handled with extreme care. German Funnel Apparatus. This apparatus was recommended by the German conference upon methods of testing materials, as was also the Tetmajer apparatus. The funnel is formed of a hollow cone with its axis vertical, as shown in Fig. 3. The height of the cone is 1 8 cm., its upper base is 20 cm. and lower base 2 cm. in diameter, terminated at the lower end by a second cone 5 cm. high, with a lower base 1.6 cm. in diameter. The funnel is supported upon a tripod, with its lower end 20 cm. above the table and 10 cm. above the top of the liter measure. To facilitate the flow of cement into the measure a rod 7/10 cm. in diameter is rotated in the axis of the funnel. This rod is guided by two cross-rods supported upon the interior surface of the funnel. METHODS OF TESTING CEMENT. 93 In the operation of this apparatus sufficient cement to fill the measure is placed in the funnel, and the rod is then rotated, about 45 revolutions per minute, by gear or by hand until the material has passed through FIG. 3. and filled the measure, which is then struck and weighed. Sieve and Funnel Apparatus. This apparatus as used in France is shown in Fig. 4. It consists of a funnel with a sieve fitting into the upper part of it and the measure below. The cement is put into the sieve, and gradually worked through by the use of a spatula. It then slides down the funnel into the measure until that is filled. 94 HYDRAULIC CEMENT. The "Commission des Methodes d'Essai des Mate- riaux de Construction" made a careful comparison of results obtained by the various methods. They found the German funnel apparatus quite precise in its FIG. 4. results with certain materials, but that with some cement it always became clogged by the packing of the material in the funnel. The Tetmajer apparatus is capable of great precision, but is somewhat compli- cated, and requires careful manipulation to secure always the same rate of filling the measure and the same amount of compacting. The inclined plane and the sieve and funnel apparatus are found to give good results, and are recommended by the commission for METHODS OF TESTING CEMENT. 95 use in France. The two latter give nearly identical results, the German apparatus somewhat higher and the Tetmajer apparatus much larger results than the others. The following method for apparent density was recommended in the preliminary report of the com- mittee of the American Society of Civil Engineers upon cement testing, but has never come into com- mon use: " Procure a cylinder of a height of 6 inches, having an interior area of 2 square inches. Sifting the cem- ent to be measured so that it may not be compact, weigh carefully 5 ounces if of Portland cement and 4 ounces if of natural cement, and pour the same into the cylinder, which should stand upright with its lower end resting upon a close-fitting and suitable base; then, without shock or sudden impact, lower a close-fitting piston, moving without friction, slowly down the cylinder on to the cement; said piston and its attachments to weigh 50 pounds. After resting thereon one minute, remove the same and ascertain the bulk of cement thus compressed." In making tests for apparent density it is advisable to sift the cement, and use only that portion which passes the finest sieve, thus making the result to a certain extent independent of the fineness of grinding. To accomplish this the sieve used should be as fine as possible in order to eliminate all but the impalpable powder. In Europe a sieve of 5000 meshes per square centimeter is employed for this purpose, correspond- ing to the No. 1 80 sieve, 32,400 meshes per square inch. 96 HYDRAULIC CEMENT. The ordinary weight of Portland cement varies from 70 to 100 Ibs. per cubic foot, depending largely upon the method of making the tests. Natural cement is usually somewhat lighter. ART. 38. SPECIFIC GRAVITY. The determination of specific gravity is often sub- stituted for that of apparent density, and is a much better guide to a knowledge of the actual density of the material, as it is not subject to the fluctuations due to fineness or method of determination which characterize the weight tests. The differences of specific gravity to be determined are, however, very small, and great care is necessary in the manipulation of the test in order to obtain reliable results. The test for specific gravity is commonly made by immersing a known weight of the cement in a liquid which will not act upon it, and obtaining its volume through noting the volume of liquid displaced. In making the test by this method it is necessary that all the air-bubbles contained in the cement powder be eliminated, and that the volume obtained be that of the cement particles only. Schumann Volumenometer. Several forms of appa- ratus have been used for this purpose. Of these the Schumann volumenometer, shown in Fig. 5, is perhaps the most common. It consists of a graduated tube, the bottom of which is ground to fit closely into the top of a flask. In the use of the apparatus the tube is placed upon the flask and filled with benzine to the zero-point on METHODS OF TESTING CEMENT. 97 the scale. 100 grammes of cement are then weighed and carefully poured into the top of the tube so as to sift gradually through the liquid, thus allowing the air to escape. The elevation of the surface of the liquid in the tube FIG. 5. gives the volume of the cement. The scale as ordi- narily made has a range of 40 cubic centimeters, and is graduated to i/io centimeter. This volumenometer gives very satisfactory results when carefully used, but much care is required to fully eliminate the air and prevent the powder from adhering to the surface of the tube. It is well in operating in this manner to 98 HYDRAULIC CEMENT. introduce the liquid through a long funnel, thus keep- ing the inner surface of the tube dry. In some instances greater accuracy may perhaps be obtained by filling the apparatus to the zero-point with liquid and weighing the whole, then pouring in enough cement through a funnel to raise the surface of the liquid through a definite volume, and determin- ing the weight of cement by reweighing the whole. This eliminates the error due to loss of cement powder in introducing it into the flask. Candlof s Volumenometer. M. Candlot has modified the Schumann volumenometer by closing the upper end of the tube with a glass bulb. In using this form of the apparatus the tube is turned bottom upward and the bulb filled with benzine. The flask is then placed upon the tube, the apparatus is inverted, and a reading taken of the height reached upon the tube. A weighed quantity of cement is then placed in the tube, which is turned upside down and separated from the flask for the purpose. The flask being replaced and the whole again inverted, the volume is shown by the use of the liquid in the tube. In order to elimi- nate the air-bubbles the apparatus is shaken before taking the second reading so as to thoroughly mix the liquid and the cement. This method is much more rapid than that with the Schumann volumenometer, but in practice is hardly so satisfactory in result. Le Chatelier 's Volumenometer. This apparatus is shown in Fig. 6. It consists of a flask the top of which is drawn out into a tube i cm. in diameter and 20 cm. long. Above the middle the tube enlarges -METHODS OF TESTING CEMENT. 99 into a bulb for a short space, and then again continues with uniform diameter to the top. The flask to a point marked on the tube just below the bulb has a capacity of 100 cubic centimeters. From this point to a mark above the bulb the capacity is 20 cubic FIG. 6. centimeters. This latter mark is very carefully deter- mined, and the upper part of the tube is graduated to 1/10 cubic centimeter. In using this apparatus, the flask is filled with liquid to the mark below the bulb, and the cement is then slowly introduced through a funnel, and settles through the liquid into the flask, the air being eliminated by IOO HYDRAULIC CEMENT. its long passage through the liquid. Cement is added until the surface of the liquid rises to the 20- cubic-centimeter mark. The weight of this volume of cement is obtained by weighing the apparatus before and after the cement is introduced. Or, this volu- menojneter may be used in the same manner as that of Dr. Schumann, the bulb serving to prevent the cement sticking to the sides of the tube. Mann Gravimeter. This apparatus consists of a flask which when filled to a certain mark upon its neck contains an accurately known quantity of liquid. A graduated tube with a stop-cock at its lower end con- tains when full the same quantity as the flask. A weighed quantity of cement is placed in the flask and the tube is filled with the liquid to be used. The liquid is then allowed to run from the tube into the flask, reliance being placed upon shaking the flask to eliminate air-bubbles, until the flask is filled to the volume mark. The volume of liquid remaining in the graduated tube is then equal to the volume of cement powder. The complete eradication of the air is a matter of difficulty, and if the operation is conducted altogether in the air the change of temperature may be sufficient to affect the result. This method in practice is not likely to give very concordant results. Erdmenger* s Volumenometer is a modification of the Mann apparatus, looking to the maintenance of a con- stant temperature during the test. This arrangement is shown in Fig. 7. It consists of a graduated tube of 50 cubic centimeters capacity, partially enclosed in a larger vessel which acts as a cooler. The upper end of the tube is closed with a ground-glass stopper, the METHODS OF TESTING CEMENT. 101 lower end with a stop-cock. Near its lower end the tube has a horizontal branch, also closed by a cock, and connecting it with a jar of I liter capacity having two openings at the top, one being connected with the measuring tube, the other with a rubber pressure-ball. The cooler has two openings at the top, one of which serves to permit the passage of air, the other to admit a small thermometer. To fill the cooler with water, DDfl FIG. 7. or empty it, an opening is placed at the bottom which may be controlled by a stop-cock. A narrow-necked flask of the capacity of 50 cubic centimeters is used for the measurement of the volume of the cement. In making the test the apparatus is kept at a tem- perature of about 60 Fahr. by filling the cooler with water at that temperature and standing the double- necked jar containing the liquid to be used in the test and the measuring-flask in vessels of water of the same temperature. The measuring-flask is then filled to 102 HYDRAULIC CEMENT. the 5o-cubic-centimeter mark with the fluid, and the test is made as with the Mann apparatus; or the mix- ing of the liquid with the cement may be done by putting the cement in a small funnel, and washing it through with the liquid, thus eliminating the air. Any of these volumenometers may give good results when carefully handled, the Schumann and Le Chate- lier forms being easiest to handle satisfactorily. Benzine or turpentine is usually employed as the fluid, and it is important that the liquid should be maintained at a nearly uniform temperature during the test. For this reason it is common to immerse the apparatus in cool water. Where benzine is used the temperature should not rise above 60 Fahr. It is desirable also to sift the cement to be tested through a fine sieve, on account of the better elimination of air-bubbles possible with the fine material. In order to make the determination of specific gravity of value, it must be reliable to two decimal places. Portland cement varies from about 3.00 to 3. 18, and is usually above 3.05. Natural cements vary from 2.75 to 3.05. An inferior limit is sometimes fixed in specifications usually for Portland cement about 3.00 or 3.05, and for Roman cement about 2.80. The presence of the volatile elements due to incom- plete burning, or of adulterations added after the burning, tends to lower the specific gravity. The quantity of adulteration, however, needs to be con- siderable before it becomes appreciable in the results of this test. The specific gravity, unlike the apparent density, is not affected by the fineness of grinding; and it has METHODS OF TESTING CEMENT. 103 been suggested by examination of the results of cer- tain experiments that a comparison of the two tests for the same material may sometimes give a better determination of the actual fineness than can be obtained by the use of sieves. This fineness is shown practically by the greater ability of the fine cement to " take sand " without losing in early strength. ART. 39. TESTS FOR FINENESS. The fineness to which a cement is ground is usually considered a matter of importance, as upon it depends very greatly the early adhesive strength of the mortar and the ability of the cement to take sand. A test for fineness is nearly always included in specifications for cement, and the test is particularly necessary where the tensile strength is tested for neat cement only. In such case the attainment of a proper strength neat, together with a fair degree of fineness, practically insures that the cement will give good results when used with sand. The fineness which should be required is largely a matter of relative economy; the finer the cement, the larger the quantity of sand that may legitimately be used with it. The coarse parts of the cement are to be considered as inert material, or practically as a certain amount of sand already mixed with the cement. It is a question therefore of relative costs of different degrees of fineness. There is, however, some dispute as to the value of fineness. Some European authorities question the wisdom of a fineness test. It is well known that the IO4 HYDRAULIC CEMENT. effect of fineness in the strength of sand-mortar dis- appears to some extent with time, but the impalpable powder seems to be the really valuable part of the cement, and if this be omitted the cement loses its value. Prof. Le Chatelier in his microscopic examination of mortar found that, after setting, in the more fine particles no trace is left of the grains of the cement. With the larger ones the central part of the grain remains unaltered. It seems that the grains which are completely attacked are limited to o. I millimeter in diameter, but. further study is needed upon this point. Coarse grinding also, as has been elsewhere noted, increases the intensity of the action of expansives which may be contained in the cement, causing a coarse-ground cement to expand and crack, when per- haps if finely ground it would be unaffected. The test for fineness simply consists in sifting the cement through a sieve or a set of sieves, and observ- ing the amount retained by each sieve. The committee of the American Society of Civil Engineers upon " standard tests" recommend the use of sieves of 2500, 54/6, and 10,000 meshes per square inch. Specifications usually, however, require only a single sieve generally that of 2500 meshes, but some- times that of 10,000 meshes. A more general use of the finer sieve would undoubtedly be advantageous, as it is now generally admitted that all material coarser than that dimension is practically inert, and a real measure of useful fineness is not given by the 21500- mesh sieve A common requirement is that not more METHODS OF TESTING CEMENT. 10$ than io# by weight of the cement be retained upon a sieve of 2500 meshes, or that not more than 20$ be retained upon that of 10,000 meshes, or both. Most of the cements commonly in use in the United States easily comply with these requirements. Very many of them do not give a residue on the coarse sieve of more than i# to 3$, or on the fine one of more than 8# to io#. Some cements, however, seem to be bolted with special reference to passing the test of the 25OO-mesh sieve, and are very coarse when tested with a finer one. The sizes of wire of which the sieves are made is of course important as regulating the sizes of the open- ings, and should always be stated ; the common stand- ard is that the diameter of wire should be about 1/3 of the spacing between centres.. The sieves of the American Society of Civil Engineers mentioned above are of Nos. 35, 37, and 40 wire-gauge. It is not usually practicable to get sieves with per- fect regularity either of spacing or diameter of wires. A sufficiently near approximation for practical work may be obtained by using care in selecting the sieve, but the gauge frequently offered for this use differs very considerably in the sizes of openings for the same number per inch, and sometimes the openings are quite irregular in size in different parts of the same sieve. ART. 40. RATE OF SETTING. The rate of setting of cement is tested for the pur- pose of determining if it be suitable for a given use, 106 HYDRAULIC CEMENT. and not as a measure of the quality of the material. For most purposes, where immediate setting is not required to prevent disturbance of the mortar before hardening, the moderately slow-setting cements are found most convenient, as they need not be handled so quickly, and may be mixed in somewhat larger uantities. Testing for time of setting consists in arbitrarily fixing two points in the process of consolidation, which are called the beginning and the end of setting. These points are differently determined in the vari- ous methods of testing, and are not marked by any distinguishing phenomena which admit of definite determination. The method recommended by the committee of the American Society of Civil Engineers is that proposed by General Gillmore, and consists in mixing cakes of neat cement, about 2 or 3 inches in diameter and 1/2 inch thick, to a stiff plastic consistency, observing the time when they will _bear_. a needle 1/12 inch in diameter sustaining a weight of 1/4 pound, and noting this as the beginning of setting; then continuing the observations with a needle 1/24 inch in diameter carrying a weight of one pound until the material is sufficiently firm to bear this, when it may be called fully set. The committee call those cements which set in 1/2 hour or less, quick-setting; those requiring x. more time, slow-setting. The time of setting is often roughly determined in practice by making small cakes of mortar and observ- ing when they will resist penetration under a light METHODS OF TESTING CEMENT. IO7 pressure of the thumb-nail. This is a standard test in Germany. For ordinary practical purposes these methods are sufficiently accurate, as all that is necessary is to know whether the cement sets quickly or slowly, but for experimental and comparative purposes more elaborate methods are valuable. The beginning of setting is the point of most value, as the cement in practice should be used before that point is reached, in order that it may not be disturbed after the stiffening has begun. In Germany and France the Vicat needle is com- monly employed for accurate determinations. This arrangement is shown in Fig. 8. By this method a FIG. 8. briquette of neat cement is made in a cylindrical brass or rubber mould 10 centimeters in diameter and 4 centimeters high, placed upon a plate of glass or metal, the cement being mixed to a plastic consist- 108 HYDRAULIC CEMENT. ency as determined by the consistency test. The apparatus is so arranged that a weight of 300 grammes may be brought either upon a needle of I square mil- limeter section, or upon a cylindrical plunger I centi- meter in diameter, and allowed to settle into the cement, the depth of penetration being shown by a scale along which the weight slides. As soon as the mould is filled with the mortar it is placed in the apparatus, and the plunger, sustaining the 300 grammes, is brought to the surface of the briquette and allowed to sink into it. If the plunger penetrates to a point 6 millimeters from the bottom the mortar is of proper consistency for the test. The needle is then substituted for the plunger, and the time when the needle first refuses to sink entirely through the mortar is observed and noted as the beginning of set- ting; the time when the needle first rests upon the briquette without penetrating it is considered the end of setting. The accurate determination by this method of the points where the set is said to begin and end is a matter of some difficulty, as the lack of perfect homo- geneity causes the needle to sink more deeply in some parts than in others, and the cement sets more rapidly at the circumference than in the interior of the mass. However, these defects are not very serious when due care is exercised in mixing the mortar, and the pene- tration is not taken too near the edges. The time of completion of set is much less well defined than that of beginning of set, as there is usually a considerable period during which a very slight penetration takes place, decreasing insensibly to final disappearance. METHODS OF TESTING CEMENT. 109 The point to be used for completion of the set is that at which the penetration becomes very small, so that the curve of penetrations becomes practically horizon- tal. Such a point is usually fairly well defined. Various modifications of this method have been pro- posed for the purpose of securing greater uniformity in result, but they have not come into general use. M. Bonnami has proposed to modify the test by vary- ing the weight upon the needle instead of the depth of penetration. He measures the time at which various weights will cause the penetration of the needle to mid-depth in the mortar, beginning with 50 grammes which marks the beginning of setting, and increasing to 3000 grammes, which gives the end of setting, thus obtaining a curve of times in terms of weights sus- tained. The penetration to mid-depth is selected as the point of maximum variation. The time of setting is usually tested upon paste of neat cement, on account of the difficulty of obtaining a satisfactory test with sand-mortar. The Vicat needle is quite useless when sand is employed because of the interference of the grains of sand with the descent of the needle. Rough tests of sand-mortar by the ordinary methods may readily be made with sufficient accuracy for practical purposes, and are very desirable as showing more nearly what may be expected of the mortar when used. The rate of setting of neat mortar gives but little indication of what the action may be with sand. For different cements a mortar of 3 parts sand to I of cement may require from about i to 8 or 10 times as long as neat paste when the same sand and method of mixing are employed. I 10 HYDRAULIC CEMENT. Several propositions have been made with reference to a standard test for the purpose of comparing the rate of setting of sand-mortars. These usually have been to substitute cylinders of larger diameter for the Vicat needle, in order to reduce the effect of the sand grains, and then to use correspondingly heavy weights to produce penetration. The usual method is to determine the weight necessary to indent the surface of the mortar. Thus in one apparatus a cylinder I cm. in diameter is employed: when the mortar will just bear a weight of 400 grammes it is considered as beginning to set; when it will sustain 10,000 grammes the setting is complete. M. Feret has also proposed to make standard tests by using fine sand composed of grains which pass the sieve of 75 meshes and are held by one of 180 meshes per linear inch, the test being made as for neat cement with the Vicat needle. This serves as a com- parison of the effects of sand upon different cements. In making tests for rate of setting, the temperature of the ingredients of the mortar before gauging, that of the atmosphere in which it is gauged, and of the air or water in which it is placed during setting have a very large influence upon the results. A temperature of 60 to 65 Fahr. is usually accepted as standard, although the air in the laboratory may have a some- what higher temperature perhaps 65 to 70. The amount of water used in gauging and the methods of mixing the mortar and filling the mould are important. For the purpose of regulating these, the consistency of the mortar is prescribed. This is probably the best means of arriving at uniform results, METHODS OF TESTING CEMENT. in but it should be pointed out that the same consistency may be arrived at in two ways by using a small quantity of water and working thoroughly, or by using a larger quantity and working less. The quantity of water which will bring the mortar to proper consistency after three or four minutes of vigorous working is the most desirable, but there is opportunity for consider- able variation in the results of the test as carried out by different operators. A number of automatic appliances have been devised for the purpose of making these tests without constant attendance. That of Professor Fuertes, in the labora- tory of the College of Civil Engineering, Cornell Uni- versity, consists of a trough i inches wide by 2 inches deep, in which the mortar is placed. A car is drawn along a track over the trough by means of a screw rotated by a clock, thus giving a motion vary- ing uniformly with time. From the car is suspended a Vicat needle, which is dropped and raised at regular intervals by the action of a small stream of water. A pencil attached to the shaft of the needle draws the curve of penetrations upon a board at the side of the trough. ART. 41. CHANGE OF TEMPERATURE DURING SETTING. Observations of the change of temperature during setting are commonly taken in many of the European laboratories. It has been thought by some observers that the points of beginning and end of setting might be more accurately marked by observing the change of temperature than by the needle test. The opera- tion of setting is a chemical action which takes place 112 HYDRAULIC CEMENT. with the disengagement of heat, but in many cases the rise in temperature is so slight and indefinitely marked that it would be difficult to use it in this manner. The amount of the variation in tempera- ture varies somewhat with the activity of the cement, increasing rapidly as the cement sets more quickly. The total rise of a quick-setting cement may reach 15 or 20, while in a very slow one it may be quite im- perceptible. The change in temperature also varies with the nature of the cement, and attempts have been made to connect it with the soundness of the material, par- ticularly the presence of free lime. This, however, does not seem to be supported by facts, or at least the indications are very indefinite. Expansives which are slow in action, and therefore dangerous in the cement, are not likely to cause increase in temperature during setting. The test for change in temperature is ordinarily made by placing the mortar in a cylindrical mould, like that used with the Vicat needle, fitted with a cover through which is an opening to permit the intro- duction of the thermometer. This cover prevents the mortar coming into contact with the air or water in which it may be placed, thus neutralizing the effect of the internal change of temperature. Some very slow-setting cements show a fall in tem- perature if left exposed to the air while setting, prob- ably due to surface evaporation. This test does not seem of importance as a measure of the quality of cement, but it is worthy of attention in a systematic study of the properties of the material, and may be capable of giving interesting results. CHAPTER VI. TESTS OF THE STRENGTH OF MORTAR. ART. 42. METHODS EMPLOYED. THE strength of mortar is frequently tested in three ways: the tensile test is the one more commonly em- ployed, but compressive and transverse tests are also often used. The test for tensile strength is made by making briquettes of the mortar in moulds having a definite section at the middle, in the United States usually one inch square, and enlarging at the ends to fit in clips by which they may be placed in the testing- machine and pulled apart by direct tension. This test is in common use, because it can be more readily and uniformly applied than the others, and seems, when coupled with other tests, to give a satisfactory indication of the value of the material. The compressive test consists in crushing small blocks of the mortar between the jaws of the testing- machine and weighing the force required. This test is more difficult in manipulation to secure uniform results, and also requires much heavier appliances, on account of the high resistance offered by the material to crushing. "3 114 HYDRAULIC CEMENT. The transverse test is made by moulding the mortar into bars; the bar is afterward placed horizontally upon supports near its ends, and broken by a load brought upon its middle, causing it to break by bend- ing. The proper conduct of any test for strength is a matter requiring care and experience. There are many points connected with the circumstances and manipulation of the work which have an important bearing upon the result. These are: the form of the briquette; the method of mixing and moulding; the amount of water used in tempering the mortar; the surroundings in which the mortar is kept during the hardening; the rate and manner of applying the stress; the temperature at which all of the operations are performed. In order to secure uniform results it is essential that the tests be standardized in all of these particulars. Much has been accomplished in this direction during recent years, but there is still great disparity in the results of different operators, un- doubtedly due mainly to differences in making the briquettes. Every laboratory seems to have to a certain extent its own practice, which makes its work incomparable with that of any other laboratory. Even where pre- sumably the same methods are used it is very difficult to frame rules that all will understand alike, while in all cases the personal equation of the operator is an important matter in hand-work. The committee upon standard tests of the American Society of Civil Engineers in their report call attention to this matter in the following words: TESTS OF THE STRENGTH OF MORTAR. 115 " The testing of cement is not so simple a process as it is sometimes thought to be. No small degree of experience is necessary before one can manipulate the materials so as to obtain even approximately accurate results. " The first tests of inexperienced though intelligent and careful persons are usually very contradictory and inaccurate, and no amount of experience can eliminate the variations introduced by the personal equation of the most conscientious observers. Many things, apparently of minor importance, exert such a marked influence upon the results, that it is only by the greatest care in every particular, aided by experi- ence and intelligence, that trustworthy tests can be made." Experience, since the report of the committee was made, has shown that the difficulties in the way of uniformity in such tests are much greater than was then imagined. The variations in the results of the tensile test between the work of different experienced operators working by the same method and upon the same ma- terial are frequently very large, and often make all the difference between the acceptance and rejection of the cement. Differences of 40$ to 60% with neat cement are not uncommon, while for sand-mortar they are much greater. An investigation of this matter by Prof. J. M. Porter, of Lafayette College, is interesting in this connection. He divided a sample of cement into a number of parts, sending each to a different laboratory with the request that tests be made of it in I to 3 Il6 HYDRAULIC CEMENT. mortar, according to the rules recommended by the committee of the American Society of Civil Engineers. The resulting average strengths of each of the nine laboratories were as follows, in pounds per square inch: 75, 102, 114, 133 and 140, 153, 163, 176, 225, 247. These results (see Engineering News, March 5, 1896) show that the lowest strength was but 30$ of the highest, while the remainder were quite evenly distributed between the two extremes. Each result was the average of five briquettes, which agree fairly well among themselves. If the results of experienced men in the permanent laboratories vary so much, what is to be expected of tests made by less experienced men for the reception of material upon temporary work, and how can a specification be framed which shall fairly determine the value of the material? Evidently, to secure proper results with hand-work, the inspector must first be calibrated, and the specifications drawn in accordance with the practice of the laboratory. It is at least very desirable that some means be devised by which the work of these tests may be made automatically, and the personal factor eliminated in so far as possible. In standard tests it is customary to adopt a nearly constant temperature of 60 to 65 Fahr. for the air in the laboratory in which the briquettes are prepared and the tests made, and about the same or slightly less for the water used in tempering and that in which the mortar is immersed during hardening. TESTS OF THE STRENGTH OF MORTAR. ART. 43. FORM OF BRIQUETTE. Briquettes of mortar for tests of strength are com- monly formed in moulds of metal of the form to be used in the tests. As the size and shape of the speci- mens have an important effect upon the result, it is necessary to adopt standard dimensions in order to obtain uniform results. For compressive tests a parallelepiped, usually a cube, is employed. In the United States a cube whose edges are each two inches in length is com- monly used, although sometimes an inch cube is used. In Europe generally the standard specimen is a cube with edges seven centimeters in length. The French " Commission upon Methods of Testing Materials," however, rejected the rectangular section for compression specimens on the ground that it is difficult to so fill the corners of the moulds as to make homogeneous briquettes, and they recommend the cylindrical form as preferable. They also recommend the use of half-briquettes obtained by the tension test as blocks for the crushing test. For transverse tests bars of rectangular section are used. Different experimenters have used quite differ- ent dimensions, and there is no size which may reason- ably be called a standard. Those who have proposed the adoption of this test in place of that for tension in the acceptance of material have usually advocated a test-piece of a section one inch square and from eight to twelve inches long, although sometimes the section is made two inches square. HYDRAULIC CEMENT. For tensile tests many forms of briquettes have been tried, but at present there are but: two in com- mon use: the one recommended by the committee of the American Society of Civil Engineers (shown in Fig. 9), which was derived from that used by Mr. Grant in England, is now the standard in the United States, and commonly used in England; the other is the form adopted by the Association of German FIG. 9. FIG 10. Cement Makers, and is the standard in Germany, and generally employed in France. This form is shown in Fig. 10, which gives the dimensions in millimeters. The middle section is 22.5 mm. wide by 22.2 mm. thick, giving a cross-section of 5 square centimeters. Comparative tests of briquettes of the two forms indicate that the English form gives higher results than the German, the difference being commonly for neat briquettes as much as 30$ to 40$ of the smaller. This may perhaps be accounted for by the fact, stated by Mr. Faija, that a sudden change of cross-section is always an element of weakness, and while the English TESTS OF THE STRENGTH OF MORTAR. I 19 form diminishes gradually from the ends to the middle, in the German form the area is suddenly decreased by a circular notch at the middle. It may also be noted that the surface upon which the clip catches the briquette when being tested is in the German form inclined at a greater angle to the centre line of the briquette, and consequently the adjustment in the clips to produce axial stress is less perfect. In England and the United States the standard area for the middle is one square inch; in Germany and France a little smaller, being, as already noted, five square centimeters. The use of these small sec- tions is advantageous, as it admits of lighter apparatus in making the test, and because greater uniformity is easily attainable in making the briquette. The work also is facilitated by the fact that less mortar is required for each specimen than with larger sections, so that more briauettes may be prepared from each mixing. The size of the breaking section has an important effect upon the strength, the smaller sections giving much higher strength per unit of area than the larger ones. Thus for neat cement a change from a section i inch square to one 2 inches square has been found to lessen the tensile strength per square inch to about one half that of the smaller section. M. Durand- Claye has shown that the strength varies more nearly with the perimeter of the section than with its area, and that the interior may be removed without loss of strength. M. Alexandre made a number of experiments upon I2O HYDRAULIC CEMENT. the relative strengths of briquettes of different sizes.* He found that large briquettes gave much less strength per unit area than small ones, but for sand-mortars the effect diminished as the proportion of sand in- creased, and the difference also became less as the age of the mortar increased, seeming to indicate that the effect may be partly due to the more perfect hardening of the small specimens ART. 44. QUANTITY OF WATER USED IN GAUGING. The determination of the proper consistency for mortar is very important in its effect upon the results of tests of strength. The effects upon setting and hardening of varying the quantity of water used in mixing have already been discussed, Art. 24. In making standard tests it is common to regulate the quantity of water by trying to bring the mortar to a normal consistency which shall be uniform for all tests. Different cements require very different quan- tities of water to reach the same consistency, and in the use of sand-mortar the nature and condition of the sand may also cause considerable variation. It should also be noted that the consistency of mortar does not depend altogether upon the quantity of water used, but may be varied by the manner and extent of working the mortar during the gaugings. In French and German practice it is required that the mortar be vigorously worked for five minutes. Under the different systems of making briquettes there are two consistencies employed as standards * Annales des Fonts et Chaussees, 1890, vol. n. p. 277. TESTS OF THE STRENGTH OF MORTAR. 12! the plastic and the dry. In using each of these meth- ods the quantity of water used and the consistency reached vary greatly in different places, as the man doing the work may interpret the terms used in describ- ing the desired state of the mortar. The rules recommended by the committee of the American Society of Civil Engineers describes the condition of the mortar to be used as " stiff and plas- tic," thus leaving much to individual judgment. In Europe two methods are employed for determin- ing the proper consistency of plastic mortar. The Tetmajer method consists in determining the normal consistency by the method already described for find- ing the time of setting by the Vicat needle (Art. 40). This method was recommended by the conference of German, Austrian, Swiss, and Russian engineers at Dresden, and also by the French commission. The Boulogne method is commonly used in France, and also approved by the French commission. It is as follows: The mortar is to be vigorously worked for five minutes to bring it to the required consistency. " i. The consistency of the mortar should not change sensibly if the mixing be continued three minutes after the expiration of the required five minutes. "2. If a small quantity of the mortar be taken up on the trowel and allowed to fall upon the mixing slab from a height of 50 centimeters it should be detached from the trowel without leaving any small particles adhering, and after falling should approximately retain its form without cracking. 122 HYDRAULIC CEMENT. " 3. A small quantity taken in the hand and patted into a round form, until water flushes to the surface, should not stick to the hand, and, when allowed to fall from a height of one-half meter, the ball should retain its rounded form without showing any cracks." To meet these requirements leaves but a narrow limit within which the consistency may vary. If a slightly too small quantity of water be used, the mor- tar would crack upon falling. If the quantity be very slightly too great, the mortar continues to soften upon further working, becomes sticky, and loses its form upon falling. Mortar of the dry consistency is used when the briquette is to be made by beating the mortar into the moulds. For this purpose the mortar is required to have the consistency of damp earth. In the event of moulding the briquettes by machinery the quantity of water may perhaps be controlled by the amount of compacting employed and the tendency to force the water out in moulding. The quantity of water required, as already noted, varies with the fineness, age, and other conditions of the cement, as well as with its nature. To bring mortar to a plastic condition the quantity of water required is approximately as follows: " For briquettes of neat cement: Portland cement about 25$, natural cement about 30$. ' For briquettes of I part cement, I part sand: about 15$ of total weight of sand and cement. " For briquettes of I part cement, 3 parts sand: about \2% of total weight of sand and cement," TESTS OF THE STRENGTH OF MORTAR. 123 In any particular instance the proper amount can only be determined by trial. ART. 45. METHODS OF MAKING BRIQUETTES. The wide differences commonly found in the results of tensile tests made by different men are, without doubt, mainly due to differences in making the briquettes. Probably if the other strength tests were as commonly used as that of tension the same want of uniformity would be observable in them. These differences occur, not only in the work of novices, but in that of skilled operators, who, while able to main- tain practical uniformity in their own work, disagree in results with each other when experimenting upon the same material and apparently using the same methods. The extent of this difficulty has already been alluded to in Art. 42. In order to secure uniform results it is essential that a uniform procedure be adopted as to all the operations of forming the briquette. The points of importance are the quantity of water used in temper- ing which has been discussed in the last article, the method of gauging and amount of working to which the mortar is subjected in bringing it to a proper con- sistency, and the method of forming the briquette and amount of fopce used in placing the mortar in the mould. In making briquettes by hand two general methods are employed, corresponding to different consistencies of the mortar. The plastic method is most commonly employed, being used in England, France, and the 124 HYDRAULIC CEMENT. United States, while the dry method is standard in Germany. The rules recommended by the committee of the American Society of Civil Engineers give the follow- ing method of making the briquettes: 4 ' The proportions of cement, sand, and water should be carefully determined by weight, the sand and cement mixed dry, and the water added all at once. The mixing must be rapid and thorough, and the mortar, which should be stiff and plastic, should be firmly pressed into the moulds with a trowel, without ramming, and struck off level; the mould in each in- stance while being charged and manipulated to be laid directly on glass, slate, or some other non-absorbing material. 44 The moulding must be completed before incipient setting begins. As soon as the briquettes are hard enough to bear it, they should be taken from the moulds and be kept covered with a damp cloth until they are immersed. For the sake of uniformity, the briquettes, both of neat cement and those containing sand, should be immersed in water at the end of 24 hours, except in the case of one-day tests." The report of the French commission upon standard tests recommends the following method: " The moulds are placed upon a plate of marble or polished metal, which has been well cleaned and rubbed with an oiled cloth. Six moulds are filled from each gauging if the cement be slow-setting and four if it be quick-setting. Sufficient material is at once placed in each mould to more than fill it. The mortar is pressed into the mould with the fingers so TESTS OF THE STRENGTH OF MORTAR. 12$ as to leave no voids and the side of the mould tapped several times with the trowel to assist in disengaging the bubbles of air. The excess of mortar is then removed by sliding a knife blade over the top of the mould so as to produce no compression upon the mortar. " The briquettes are removed from the moulds when sufficiently firm, and are allowed to remain for 24 hours upon the plate in a moist atmosphere, protected from currents of air or the direct rays of the sun, and at a nearly constant temperature of 15 to 18 C. They are then placed in the surroundings in which they are to be kept until the time for breaking. With quick-setting cements the delay is reduced from 24 hours to i hour for neat cement and 3 hours for sand- mortar. "It is recommended to weigh the briquettes after removing them from the moulds to make sure of the regularity of their formation." By the German method the mortar is mixed with less water than in the above, and the mould is filled and heaped with it. It is then rammed into place and pounded until the water flushes to the surface, after which the briquette is struck off level, and when hard enough is taken from the mould and treated as in the other case. Following are the specifications adopted by the Association of German Cement Makers:* " On a metal or thick glass plate five sheets of blot- ting-paper soaked in water are laid, and on these are placed five moulds wetted with water. 250 grammes * Engineering News, Nov. 13, 1886. 126 HYDRAULIC CEMENT. of cement and 750 grammes of standard sand are weighed and thoroughly mixed dry in a vessel; then 100 cubic centimeters of fresh water are added, and the whole mass mixed for five minutes. With the mortar so obtained the moulds are at once filled, with one filling, so high as to be rounded on top, the mortar being well pressed in. By means of an iron trowel, 5 to 8 centimeters wide, 35 centimeters long, and weighing about 250 grammes, the projecting mortar is pounded, first gently and from the sides, then harder into the moulds, until the mortar grows elastic and water flushes to the surface. A pounding of at least one minute is necessary. An additional filling and pounding in of the mortar is not admissible, since the test-pieces of the same cement should have the same density at the different testing stations. The mass is now cut off with a knife and the surface smoothed. The mould is carefully taken off, and the test-piece placed in a box lined with zinc, which is to be provided with a cover to prevent a non-uni- form drying of the test-pieces at different tempera- tures." For making the test-pieces of neat cement : ' ' The inside of the moulds are slightly oiled, and the same are placed on a metal or glass plate without blotting- paper. 1000 grammes of cement are weighed out, 200 grammes of water added, and the whole mass thoroughly mixed for five minutes. The forms are well filled, and then proceed as for hand-work with sand-mortar. " The mould can only be taken off after the cement has sufficiently hardened. TESTS OF THE STRENGTH OF MORTAR. 1 27 " The quantity of water for finely-ground or quick- setting cements must be increased." A method somewhat in use in France for sand- mortar is that proposed by M. Candlot, and recom- mended by the Commission upon Methods of Testing Materials, for use with sand-mortars. It is as follows: " Sufficient mortar is gauged at once to make six briquettes, requiring 250 grammes of cement and 750 grammes of normal sand. The weight of water neces- sary exceeds by 30 grammes the amount necessary to bring the cement alone to normal consistency. The mortar is prepared in the ordinary manner. In forming the briquette the mould is placed upon a metal plate, and a guide fitted above it having the same section as the mould and a height of 125 millim- eters. " 1 80 grammes of mortar are introduced and roughly distributed in the mould and guide with a rod. By means of a metallic pestle weighing one kilogramme, and with a base of the form of the briquette but of slightly less dimensions, the mortar is pounded softly at first, then stronger and stronger until a little water escapes under the bottom of the mould. " The pestle and guide are then removed and the mortar cut off level with the top of the mould." It is claimed that by this method very uniform results have been obtained. There are two points to be especially noted in mak- ing briquettes by hand: first, the mortar must be very thoroughly worked in gauging; both the German and French rules require that it shall be briskly mixed for at least five minutes, only sufficient mortar being pre- 128 HVDRAULIC CEMENT. pared at once for five or six briquettes; second, the air- bubbles must be well worked out of the mortar in filling the moulds. The neglect of these precautions causes much of the irregularity which commonly exists in the work of inexperienced operators. It is perhaps easier to secure uniform results with the dry than with the plastic method. The greater density of the hammered briquette also gives it higher strength. The plastic method, however, accords more nearly with the conditions of the use of the material in practice. Even with the, most experienced operators there exist differences in the amount of working, the pres- sure given in forming the briquette, and the quantity of water used, which cause wide variations in result. In order to secure good results in tests of strength it is necessary that the briquettes should be kept in a moist condition during setting and the first period of hardening. For this purpose it is customary in the United States to cover the briquettes with wet cloths after moulding and until submerging them. In Europe they are commonly placed in zinc boxes dur- ing this period. In the laboratory of the City of Philadelphia Mr. Richard L. Humphreys uses a soapstone closet for this purpose. He describes * the arrangement as follows: " This closet, which is made of soapstone i inches thick, is supported on a wooden frame, and is 3 feet high and 18 inches wide. Along the front is a strip of soapstone 3 inches wide, forming a basin of Proceedings Engineers' Club of Philadelphia, November, 1896. TESTS OF THE STRENGTH OF MORTAR. 1 29 the bottom of the closet, in which water is placed for keeping the air moist. The doors are made of wood, covered with planished sheet copper, and are rabbeted to fit tightly. There are two sets of shelves, the lower being a wooden rack, and the upper is formed of strips of glass 33 inches long, 3 inches wide. When closed the closet is perfectly tight, the water in the bottom keeping the air moist, preventing the briquettes from drying out, and thus checking the process of setting. " Briquettes which have been removed from the moulds are placed on edge on the glass shelves, while the moulds containing the briquettes too soft to be removed are placed on the rack." ART. 46. MECHANICAL APPLIANCES FOR MAKING BRIQUETTES. In order to reduce the effect of the personality of the operator in making tests of the strength of cem- ents, various appliances for gauging and moulding briquettes by machinery have been proposed and tried. Greater uniformity in these tests is highly desirable, and it seems possible to reach it only by the applica- tion of automatic appliances in making the briquettes. No entirely satisfactory system of automatic testing has as yet been devised. In Europe machines are quite commonly employed for moulding briquettes, but the mixing is done by hand in the ordinary manner. In Germany and Switzerland all blocks for compres- sive tests are made by machine, as are many of the 130 HYDRAULIC CEMENT. tensile specimens. The machine used for this purpose is either the Bohme hammer or the Tetmajer appa- ratus. The Bohmt hammer, designed by Dr. Bohm6 of the Charlottenburg Experiment Station, is shown in Fig. II. It consists of an arrangement by which the mor- FlG. II. tar is compacted in the mould by a succession of blows struck by a hammer of the weight of two kilogrammes upon a plunger sliding in a guide-mold placed over the mold in which the briquette is to be formed. The machine is arranged to lock after striking 150 blows. A high degree of density is thus produced in the briquette, and the air is thoroughly expelled. More regular results are thus obtained, depending much less upon the personality of the operator than by the hand method. The arrangement of this apparatus, however, is such that its operation must be extremely slow, in order to give time for the hammer TESTS OF THE STRENGTH OF MORTAR. 13! to strike a full blow without being caught on the next stroke, and the time required to make a briquette is too great to admit of its general use. The rules of the Association of German Cement Makers specify this machine, and are as follows:* " In order to obtain concordant values in compression tests machine-making is necessary. 400 grammes of neat cement and 1200 grammes of dry standard sand are thoroughly mixed dry in a vessel, 160 cubic centi- meters of water are added thereto, and then the mortar is thoroughly mixed for five minutes. Of this mortar 860 grammes are placed in the cubic mould, provided with guide-mould, and. the mould is then screwed on the bed-plate under the pounding-machine. The iron follower is placed in the form, and by means of Bohme's trip-hammer one hundred and fifty blows are struck by a hammer weighing 2 kilogrammes. After removing the guide-mould and follower the test- piece is smoothed off, taken with the mould from the bed-plate, and then treated as for hand-work." The Tetmajer apparatus is similar in character to the Bohme hammer. It consists of an iron rod carry- ing a weight upon its lower end, which is raised through a given height and dropped upon the mortar in the mould. The ram in this machine weighs 3 kilograms. This machine is used in the Zurich laboratory for both tensile and compressive specimens, and Prof. Tetmajer regulates the number of blows by requiring a certain amount of work to be done upon a unit volume of mortar, 0.3 kilogrammeter of * Engineering News i Nov. 13, 1886. 132 HYDRAULIC CEMENT. work per gramme of dry material of which the mortar is composed. This apparatus is subject to the same limitations in practice as the Bohme hammer, in being very slow in use, and somewhat expensive in first cost of apparatus. Canadian Method. In Canada, and to some extent in England, the method has been adopted of gauging the mortar quite soft, using a high percentage of water, with machine-mixing, and then moulding the briquettes under light pressure 20 Ibs. per square inch on the surface of the briquette. This gives much lower results for strength than the ordinary methods, which is unimportant provided the results are con- cordant, and specifications are made to agree with the method. A number of tests made by Mr. Cecil B. Smith at the McGill University seem to show that the method is capable of yielding uniform results in so far as the variations of the individual tests from the mean of a single series is concerned. The method appears to be defective, however, in not affording a satisfactory method of determining the proper consistency of the mortar, upon which largely depends the comparability of the results of different observers. The Jamieson briquette machine, shown in Fig. 12, was designed by Prof. Jamieson, of the Iowa State University, for the purpose of making briquettes under diect pressure, and is intended to secure rapid manip- ulation. The following description is from Engineering News of February 7, 1891 : " The principle of the machine is very simple. A vertical cylinder (#) whose cross- TESTS OF THE STRENGTH OF MORTAR. 133 section conforms to the outline of the standard bri- quette receives a charge of the mixed cement from a FIG. 12. hopper at the side. In this cylinder a close-fitting piston is worked by a hand-lever (&), so that the 134 HYDRAULIC CEMENT. charge of cement in the cylinder may be subjected to a pressure of about 175 Ibs. per square inch. The bottom of the cylinder is raised above the bed-plate a distance equal to the thickness of the briquette (i inch), and in this space works a triangular block (/) having two holes of the same cross-section as the briquette. This block can be oscillated to bring either of these holes beneath the cylinder. When one hole is in this position, however, the other is clear of the cylinder at one side, and the briquette of cement which has been pressed into it can be lifted by a plunger, worked by a lever (;;/) and guided by the pins (kk). Owing to the pressure used the briquettes are hard enough to handle as soon as lifted from the mould, and they are at once removed and placed on a glass slab." " In practical working it has been found possible to make briquettes as rapidly as 600 per hour." This machine may be found useful in laboratories where large numbers of briquettes are needed for the purpose of comparing cements under various condi- tions, as it greatly lessens the manual labor of form- ing the briquette. It is to be observed, however, that the time occupied in making briquettes is largely used in mixing the mortar, and that rapid moulding involves equally rapid mixing. Ihe macnine, as designed, provides no means of regulating the amount of pressure applied in forming the briquette, which may vary with different operators, and is thus likely to produce variations in result. This may be com- paratively unimportant for neat briquettes mixed dry, but has a large influence upon the strength of sand- mortar. It may be suggested also that to make TESTS OF THE STRENGTH OF MORTAR. 13$ by this method briquettes, which are firm enough to handle immediately, the cement must be mixed very dry too dry for the best results; with sand-mortars it will be very difficult to produce solid cakes by this method. The experience of the author in experimenting with various appliances for moulding briquettes shows that quite uniform results may be obtained from briquettes moulded under a single application of a steady pressure. This method may be applied much more rapidly than the hammer method, with about as good results, and render the results of different operators much more concordant than can be obtained by hand-work, although the variations from the mean in the work of a single observer may be as great or greater than in the hand-work. Briquettes of neat cement, machine-mixed, and moulded under a pressure of about 500 Ibs., upon the surface of the briquette, give good results when the averages of different men are compared, and small variations in pressure are not important in the results. For sand-mortar, one part cement to three parts sand, a pressure of IOOO to 1500 Ibs. is desirable to sufficiently compact the mortar to form homogeneous briquettes and give uniformity in the results. To obtain the best results the mortar should be gauged to such a consistency that the water begins to ooze out under the pressure, the cakes being reduced to a semi-plastic condition, and becoming too soft to handle before setting. An apparatus for making briquettes by this method 136 HYDRAULIC CEMENT. is easily arranged. To do good work it should not aim at too great rapidity, but sufficient time should be given to each briquette to permit the weight to produce its full effect in the compression of the mortar. The gradual application of the load in a testing-ma- chine has been found preferable to the sudden descent of a lever as usually employed. Possibly a screw and hand-wheel giving a gradually increasing pressure may prove the best method of applying the pressure. The time occupied need not be long, and briquettes may be made much more rapidly than in hand-work. A means of regulating the pressure so that it may always be the same is an essential to good work. It is quite as important to eliminate the personal element from the gauging as from the moulding of briquettes. For this purpose several appliances have been tried. Thejigmixcr is an apparatus in which the materials are placed in cups with covers clamped on, and shaken rapidly up and down. It has been tried in a number of places, but has usually been found quite unsatisfac- tory in practice. It is difficult to make a satisfactory mixture by this method, and the result depends very much upon the rapidity of operation. The Faija mixer was designed and first used by Mr. Faija in England. It is shown in Fig. 13, as made by Riehle Bros. Testing Machine Company of Phila- delphia, and consists of a cylindrical pan in which a mixer, formed of four curved blades, revolves both on its own axis and about that of the pan. This arrange- ment gives fairly good results in use. Fig. 14 shows an arrangement used by the author TESTS OF THE STRENGTH OF MORTAR. 137 in the laboratory of the College of Civil Engineering at Cornell University. In this apparatus the cylinder containing the ma- terials is closed, thus avoiding the dust, which is very disagreeable with the open pan. The cover revolves about the axis of the cylinder, and the gear is placed FIG. 13. FIG. 14. outside to remove any danger of it becoming clogged. The mixer is formed of vertical rods held by a hori- zontal arm, and revolves about the axis of the cylinder and also about the middle point of the arm. This arrangement seems to give a somewhat more thorough working of the mortar than the same number of turns of the Faija mixer, but if the briquettes be moulded under heavy pressure the two give about the same results. By the use of some apparatus of this kind the mor- tar may be much more expeditiously mixed than by hand, and with greater uniformity. 138 HYDRAULIC CEMENT. To devise any system of making briquettes by hand which will secure uniformity in the work of different men seems hopeless. If such uniformity is to be secured, it must be by the use of automatic appliances. This involves not only the use of the same apparatus, but its use in the same manner in the different labora- tories. This would be very difficult of attainment; but if various mechanical appliances should come into use the relations between them would soon become known, and statements of the means employed in form- ing briquettes would make the results to a certain extent, at least, comparable with each other. For the purpo'se of securing uniformity in the results of pressure-made briquettes the following pro- cedure has been followed, with good results, in the laboratory of the College of Civil Engineering at Cor- nell University: In gauging the mortar two pounds of the dry materials are employed in one mixing. For sand-mortar, the dry materials are put in the mixer and given 50 turns; the water is then added and the mixing completed by an additional 50 turns, the mixer being operated at the rate of about 100 turns per minute. With neat cement, the cement is put in the mixer, the water added, and the mixing completed with 50 turns of the handle. After the completion of the mixing the mould and guide are filled with the mortar, and a pressure of 1000 Ibs. put upon the piston in the pressure-machine. The piston is then raised, the guide removed, and the surplus of mortar sliced off; after which the briquette is removed from the mould by being pressed out upon the surface upon which it is to remain until set. The necessary quan- TESTS OF THE STRENGTH OF MORTAR. 139 tity of water is first determined by trial, and is such that the pressure used will cause the water to slightly ooze out beneath the mould. ART. 47. TENSILE TESTS. The test for tensile strength is commonly made by placing the briquette in a pair of clips which catch its ends and are attached to a machine by which the load necessary to break the briquette may be weighed. In order to secure uniform results it is necessary that the stress shall be so applied as to bring the tension axially upon the small section of the briquette, and also that the rate of application of the load shall be always the same. There are various types of testing-machines in use, and it seems unnecessary to enter into any detailed description of them here. One of the simplest is the old machine in the College of Civil Engineering at Cornell University, which consists of a lever, suspended from a wooden framework, carrying near one end the clips to hold the specimen, and at the other a bucket into which a stream of water flows until the briquette breaks, when the water is shut off automatically. The weight of water multiplied by the ratio of lever- arms gives the stress on the briquette. All of the points of suspension are on knife-edges, and an adjust- able weight upon one end serves to balance the lever. The Michaelis Machine, shown in Fig. 15, is similar in character to the above, but is arranged with a double lever, and uses shot in the bucket instead of water. This machine is the one commonly employed in Europe. FIG. 16. TESTS OF THE STRENGTH OF MORTAR 14! The Fairbanks Machine is practically tne same as that of Michaelis. In the Fairbanks machine the shot is weighed and the stress determined by the same beam used in breaking the specimen, the bucket being hung at the other end of the beam from that used in breaking the specimen, and the weight found by means of a sliding weight. The Richie Mac/tine, shown in Fig. 16, is an ordi- nary lever machine in which the load is brought upon the specimen by means of the lower hand-wheel, while the weight is moved along the scale-beam by the upper hand-wheel. In testing a briquette both wheels must be operated simultaneously and the scale-beam be kept balanced. The Olsen Machine, shown in Fig. 17, is similar in character to the above, but differs somewhat in detail. This machine has been modified by Prof. Porter in one constructed for Lafayette College,* by adding a second lever below the clips, through which the stress is applied by means of water flowing into a bucket attached to its end. The stress is measured by the weight sliding upon the upper scale-beam as in the ordinary machine, but the weight is moved auto- matically by means of an electric contact at the end of the beam. Various other more complicated types of machines are sometimes employed, most of them hydraulic in principle. Nearly any of the machines in common use may give good results in practice. In selecting a machine, * Engineering News, March 5, 1896. 142 HYDRAULIC CEMENT. however, those are to be preferred in which the load may be automatically applied at a uniform rate. The FIG. 17. attainment of a constant rate of application with a hand-machine is a matter of considerable difficulty. In order that the stress upon the briquette shall be axial, care must be exercised in properly centring the TESTS OF THE STRENGTH OF MORTAR. 143 briquette in the clips, and the form of the clip must be such that it shall not clamp or bind upon the head of the briquette, but may be free to adjust itself to an even bearing. The surface of contact between the clip and briquette must be large enough to prevent the material of the briquette being crushed at the point of contact, and yet as small as possible to admit of its more free self-adjustment. The suspension of the clips, as is usual, by conical bearings permits them to turn so as to always transmit the stress in a right line between bearings. Fig. 1 8 shows the form of clip which is used with FIG. 18. the standard German briquette. As here given it is approved by the " Commission des Mthodes d'Essai des Materiaux de Construction," as giving very satis- factory results. With the briquette used in the United States and 144 HYDRAULIC CEMENT. England several forms have been used for clips. Fig. 19 shows the form adopted by the Committee of the American Society of Civil Engineers. This form does not offer sufficient bearing-surface for good results, as the briquette is likely to break on account of the crushing of the surface of the briquette at the point of contact. The form shown in Fig. 20 has been much used, but FIG. 19. FIG. 20. has the disadvantage of clamping the head of the briquette too closely, and unless great care is used may cause the briquette to break by twisting, thus giving irregular results. When the break occurs through the crushing of the material the fracture usually extends from one of the points of contact irregularly to the small section, but a break due to twisting may often be a centre break, and the irregu- larity show only in the results. The form of clip shown in Fig. 21 is to be preferred to those given above, as affording a sufficient bearing- TESTS OF THE STRENGTH OF MORTAR. 145 surface without clamping, and thus tending to diminish the irregularities in the results of tests. In order to prevent crushing at the points of contact FIG. 21. and to permit more free adjustment Mr. W. R. Cock has proposed* a rubber bearing, as shown in Fig 22. The use of this clip undoubtedly increases to some extent the proportion of centre breaks, and perhaps slightly raises the breaking strength. A rubber bear- ing-surface may also be secured on the clips of form shown in Fig. 21, by stretching a rubber band around the jaw of the clip so as to cover the surface of con- tact. This may somewhat increase the regularity of the tests, but when carefully used the ordinary clip is capable of giving very satisfactory results. Various appliances have been proposed for the accu- rate centring of briquettes, or to prevent the more free adjustment of the clips to the direction of stress, by * Engineering News, Dec. 20, 1890. 146 HYDRAULIC CEMENT. using a template for the exact placing of the briquette, or by hinging the clips at the upper corners. These arrangements do not, however, seem to be necessary. FIG. 22. The method of keeping the briquettes between the time of moulding and breaking is of course important in its effect upon the resulting strength. The tem- perature should, for standard tests, always be kept as nearly uniform as possible between 60 and 70 Fahr. When the mortar has been conserved under water it should be tested immediately on being taken from the water. The standard rate of applying the load is ordinarily about 400 Ibs. per minute. ART. 48. COMPRESSIVE TESTS. The compressive strength of cement-mortar is much greater than its tensile strength, and as it does not TESTS OF THE STRENGTH OF MORTAR. 147 seem to give a better indication of value, while more difficult of satisfactory determination, and also requires heavier apparatus, it is not usually employed as a test of quality in the acceptance of material. The com- pressive test is, however, valuable for purposes of comparison, and is desirable as an addition to the showing made by the tensile test. In the European experiment stations it is customary to test all cements under compression as well as tension. For this test, as for tension, it is essential that a standard method be followed if comparable results are to be obtained. The size and shape of the specimen are of the greatest importance. The effect of com- pression is ordinarily to cause the material to spread laterally by pressing out the sides, failure usually occurring by shearing along surfaces inclined at about 30 with the vertical, leaving pyramidal or conical blocks at the middle. When the specimen is small in height the resistance is greater per unit area than if it be higher, and for blocks similar in form the resistance increases with the size. Cubes are commonly employed for this purpose, but cylinders are sometimes preferred, on account of the greater ease with which homogeneous specimens may be prepared, and because of the liability of the corners of cubes to crack off under comparatively light pressure. The common piece in Europe is a cube with edges 7 centimeters long, moulded under the hammer, and in testing it is required that the pressure be always exerted on two faces of the cube, that is, on the faces which are against the surface of the mould, in forming the block. 148 HYDRAULIC CEMENT. The German conference at Berlin, however, recom- mended the use of blocks 5 square centimeters in area, the same as the tensile specimens. The French " Commission on Methods of Testing Materials" made the following recommendations: " For tests of compressive strength the half- briquettes separated by tension are to be used. Each half-briquette to be crushed separately, and the sum of the two results taken as the strength of the specimen. " In the absence of half-briquettes cylinders 45 millimeters in diameter and 22 millimeters high may be used, made and conserved like the tensile briquettes. " Those specimens which show visible irregularities or distortions are smoothed by lightly rubbing in the hand upon a stone surface." " The testing apparatus should be so arranged that the stress may be continuously applied at such a rate as to crush the half-briquette in one or two minutes." " For tests having for their object the comparison of mortar with other materials, it is provisionally recommended to employ the cube, with faces of 50 square centimeters area, placed upon one side. These tests will thus conform in a general way to the rules adopted for the other materials." In testing cement the compressive specimens may be readily obtained with smooth faces, which may be placed in the jaws of the testing-machine directly in contact with tne compression-blocks. This is the usual practice in Europe. It is common, however, in testing other materials, stone or brick, to set the specimen in the machine TESTS OF THE STRENGTH OF MORTAR. 140 with a thin layer of plaster of Paris between the plate of the machine and the surface of the specimen. A small pressure being placed upon the specimen before the plaster of Paris sets, the bearing becomes even and the load uniformly distributed over the surface. In many instances this seems to conduce to greater uniformity in results with cement, but it is not com- mon practice and is hardly necessary. For making compressive tests of cement, any of the ordinary lever or hydraulic machines in common use, with a capacity of 40,000 to 50,000 Ibs., is usually sufficient. It is desirable that the load be applied as uniformly as possible, as the result will be more or less affected by unsteadiness or shocks. A very handy and efficient machine for this purpose in use at the College of Civil Engineering, Cornell University, is shown in Fig. 23. It is a hydraulic machine, constructed by J. Amsler, Lafon & Son, Shiffhausen, Switzerland. The machine is a hydraulic press, in which fluid pressure is reduced through a system of pistons, so that it can be measured by a mercury manometer. A is the compression-piston, B and C are the pistons which reduce the pressure exist- ing under piston A. The mercury manometer consists of a glass tube connected at the bottom with the space under the piston C. The test-piece E is placed between the compression- plates, of which one rests upon a spherical surface in the piston A, so that it can adjust itself; the other hangs on the end of the hand-screw, and may be placed at any height. The cylinder K is filled with castor-oil. The rod 1 50 HYDRAULIC CEMENT. L may be driven into the cylinder by the crank and screw, producing a pressure on the oil between pistons A and B. The piston A presses upward on the sped- FIG. 23. men, while piston B presses downward upon the larger piston C, causing pressure on the fluid beneath. Under piston C is a layer of machine-oil, which serves the purpose of making it tight and lubricating it. The mercury fills the bottom of cylinder M (under TESTS OF THE STRENGTH OF MORTAR. 15 1 the machine-oil) and the pipes connecting this cylin- der with the manometer, which stands at the side of the machine and reads to 32,000 kilogrammes. The pistons are fitted accurately to the cylinders, and are not packed. When the machine is operated by turning the crank which drives the rod L the pis- tons C and B are caused to oscillate about their axes, thus equalizing the wear and eliminating the friction. This machine is easily operated by hand, and the speed may be regulated by watching the manometer while turning the crank. It has been proposed to employ punching tests instead of crushing the entire specimen. This method has been employed for a number of years in the labo- ratory at Teil, France. A punch five square centi- meters in section (circular) is employed, and it is claimed that the results are more regular than those obtained by crushing the entire specimen, while requir- ing less force in the testing-machine. The " Com- mission des Methodes d'Essai des Materiaux," how- ever, concluded that it presented no advantage over the ordinary test. ART. 49. TRANSVERSE TESTS. Tests of the strength of mortar under transverse loading are seldom employed as a measure of the quality of the material, but are frequently made with a view to determining the action of the material in service. Propositions have often been made to substi- tute the transverse for the tensile test in the reception of material. These suggestions have usually been 152 HYDRAULIC CEMENT. based upon the simplicity of the test and of the apparatus with which it may be carried out. All that is necessary, after the bar is prepared, is the arrange- ment of a couple of knife-edges upon which the ends of the bar may be rested, and a third knife-edge to carry the weight brought upon the middle of the bar. The ordinary test by tension is, however, quite simple, and there seems to be little if any advantage in mak- ing a change, although the transverse test offers an equally effective means of determining quality. Much less is known as to what the transverse strength should be, and its use in specifications would need to be pre- ceded by experiments to obtain proper values for the loads to be required, while the errors due to differences in making briquettes would exist the same in the one case as in the other. Prof. Durand-Claye made a large number of tests to compare tensile and transverse strength for both neat cement and mortars in small test-pieces. He used bars 2 centimeters square and 12 centimeters long, and tension-pieces of the ordinary 5-square-centimeter sec- tion, and found the results quite regular. The modulus of rupture per square centimeter, computed by the 3 PI ordinary formula (R = ,, where P = load, / = length, and a* = area), were found to average a little less than twice the unit stress for tension. The use of this formula in this instance is of course inexact, as it assumes the material to have the same coefficient of elasticity for tension as for compression, and to be strained only to the elastic limit. As all the speci- TESTS OF THE STRENGTH OF MORTAR 153 mens are of the same size, however, this is immaterial for purposes of comparison. In making the transverse test the most common method has been to provide supports for the ends of the bar and hang weights directly upon its middle. Care should be exercised to guard against the crushing of the material under the knife-edges; a good plan is to use small plates of iron between the knife-edge and the surface of the briquette to distribute the load. Where a tension-machine is in use, a transverse attachment is readily added, by which both tests may be made by the same machine. Fig. 24 shows such FIG. 24. an attachment as it is commonly used in Europe in connection with the Michaelis machine; the rod K is attached to the end of the lower lever in the machine (A y Fig. 15), thus receiving a less effect from the ap- plication of the load than the tension specimen. ART. 50. TESTS OF SAND-MORTAR. Tests of the strength of sand-mortar, although com- monly recommended, are not very generally employed 154 HYDRAULIC CEMENT. in ordinary specifications, reliance being usually placed upon the neat test, coupled with that for fineness, to indicate how the cement will act when mixed with sand. A tensile test with sand is, however, of the greatest value when properly conducted, because it accords more nearly with the conditions under which the cement is to be used. Tests with sand may be made either as a means of judging the value of the cement, or to determine the probable strength of mortar under the circumstances which may obtain in special work. When the tests are intended as a determination of value, it is essential that they be made according to some standard forming ready means of comparison with other material. This requires that the specimens be made and tested according to standard methods, and that the sand used be of standard quality. In the German specifications the standard sand is described as follows: " In order to obtain concordant results in the tests, sand of uniform size of grain and uniform quality must be used. This standard sand is obtained by washing and drying the purest quartz sand obtainable, sifting the same through a sieve of 60 meshes per square centimeter, thereby separating the coarsest particles, and by removing from the sand so obtained, by means of a sieve of 120 meshes per square centimeter, the finest particles. The diameters for the wires of the sieve shall be 0.38 millimeter and 0.32 millimeter respectively." The German Confer- ence upon Uniform Tests specified Freienwalde sand as the standard. The committee of the American Society of Civil TESTS OF THE STRENGTH OF MORTAR. 155 Engineers recommend an artificial sand made by crushing quartz. Their report is as follows: " The question of a standard sand is of great im- portance, for it has been found that sand looking alike and sifted through the same sieve gives results vary- ing within rather wide limits. " The material that seems likely to give the best results is the crushed quartz used in the manufacture of sand-paper. It is a commercial product, made in large quantities and of standard grades, and can be furnished of a fairly uniform quality. It is clean and sharp, and although the present price is somewhat excessive (3 cents per pound), it is believed that it can be furnished in quantity for about $5.00 for 300 Ibs. As it would be used for tests only, for purposes of comparison with the local sands and with tests of different cements, not much of it would be required. " The price of German standard sand is about $1.25 for 112 Ibs. ; but the article being washed, river-sand is probably inferior to crushed quartz. Crushed granite could be furnished at a somewhat less rate than crushed quartz, and crushed trap for about the same as granite; but no satisfactory estimate has been obtained for either of these. " The use of crushed quartz is recommended by your committee, the degree of fineness to be such that it will all pass a No. 20 sieve and be caught on a No. 30 sieve." This sand can now be obtained of dealers in testing apparatus, but much of what is furnished for the pur- pose requires resifting to bring it to standard size. The use of this sand is probably advantageous as con- I $6 HYDRAULIC CEMENT. ducing to uniformity, but it gives less strength in mortar than good natural sand. In France both artificial sand (crushed quartz) and natural sand are used to some extent in tests. The Commission on Methods of Testing recommend the use of natural sand. They divide the sand into these sizes: i. Sand which passes openings of i mm. and is retained by those of 0.5 mm. 2. Sand which passes openings of 1.5 mm. and is retained by those of i.o mm. 3. Sand which passes openings of 2.0 mm. and is retained by those of 1.5 mm. The name simple standard sand is given to No. 2, and the name compound standard sand to a mixture of equal parts of the three sizes. The former is to be used in making mortar for standard tests where the dry consistency is employed, and the briquettes made by pounding. The latter is required in gauging when the plastic consistency is used. Tests of sand-mortar for the purpose of comparing various sands with the standard sand, or of estimating the efficiency of mortar under varying circumstances of use, are often of much value as a guide to the proper use of the material. In such work the method of testing necessarily depends upon the point to be investigated. ART. 51. INTERPRETATION OF RESULTS. The test for strength is regarded as a measure of the value of a cement, as showing the possession of the active elements. There are, however, different TESTS OF THE STRENGTH OF MORTAR. 1 57 elements which act at different rates, and it is unwise to classify cements according to strength alone. The tensile strength developed by cement in a test extend- ing over a short period of time is not necessarily an indication of the strength that may be attained by it during a longer period, unless the normal action of the particular material be known. That which is strongest at first may not continue to be the strongest. The development of good strength soon after the use of the mortar is a desirable attribute in most engineering work, and the probability of the material being good is greater where it shows a fair early strength, and therefore it is usually wise to specify that the cement shall develop a fairly good strength on a short-time test, but there is no object in requir- ing extremely high values. Mr. Faija recommends that the gain in strength between the 7 and 28 day periods be considered, rather than the absolute early strength, in determining the probable subsequent gain in strength. This is doubtless a better guide than the usual one, but it is not usually practicable to require tests extending over a period of 28 days, and in many instances it would be misleading, if comparison of different cements were attempted. Prof. Unwin gives a formula for the strength at any period, y = a -f- b(x i)", in which y is the strength at x weeks after mixing; a, the strength at end of one week; n, a constant for the particular material, to be determined by observations extending over consider- able time; b, a constant to be determined from the strengths given by the sample at I week and 4 weeks 158 HYDRAULIC CEMENT. after mixing. Prof. Unwin gives the value n = 1/3 for Portland cement in general, and shows that the formula gives values well in accord with the results of tests in many instances. The formula depends upon the assumption that for any two Portland cements the gains in strength at end of any period are to each other as the gains between the 7 and 28 day test, a propo- sition quite wide of the mark in many instances. The strengths commonly required by specifications, in the United States, are based upon the recommen- dations of the committee of the American Society of Civil Engineers, which gives the following average values: ' American Natural Cement, Neat. I day, I hour or until set in air, the rest of the 24 hours in water, from 40 to 80 Ibs. i week, i day in air, 6 days in water, from 60 to 100 Ibs. I month, i day in air, 27 days in water, from 100 to 150 Ibs. i year, i day in air, remainder in water, from 300 to 400 Ibs. American and Foreign Portland Cements, Neat. day, i hour or until set in air, the rest of the 24 hours in water, from 100 to 140 Ibs. i week, i day in air, 6 days in water, from 250 to 550 Ibs. i month, i day in air, 27 days in water, from 350 to 700 Ibs. i year, i day in air, remainder in water, from 450 to 800 Ibs. TESTS OF THE STRENGTH OF MORTAR. 159 American Natural Cements, I part Cement to I part of Sand. i week, i day in air, 6 days in water, from 30 to 50 Ibs. i month, I day in air, 27 days in water, from 50 to 80 Ibs. i year, i day in air, remainder in water, from 200 to 300 Ibs. American and Foreign Portland Cements, I part of Cement to 3 parts of Sand. i week, i day in air, 6 days in water, from 80 to 125 Ibs. i month, i day in air, 27 days in water, from 100 to 250 Ibs. i year, i day in air, remainder in water, from 200 to 350 Ibs." At least the minimum value here given for i and 7 days are usually required in ordinary specifications, but higher values are often employed. Many of the better cements commonly give results above the maxi- mum values stated above; this depends, however, upon how they are tested. The values to be required in specifications need to be modified, especially with the natural cements, to accord with the particular kind of cement to be used, and also with the practice of the laboratory. Where large quantities of cement are regularly em- ployed, and the same men continuously make the tests, it is a comparatively simple matter to conform the specifications to the work of the laboratory so as to l6o HYDRAULIC CEMENT. get reliable indications of the value of the material. A very large portion of the testing for reception of material must, however, be done upon detached works, where temporary laboratories are to be used, and in- spectors employed for the occasion. In these cases it is a difficult matter to adopt a specification which shall give good results, unless the operator can himself first be calibrated. The results of tests in the permanent laboratories usually give higher strength for the same material than would be obtained on an ordinary outside test, espe- cially by a comparatively inexperienced man. It is not to be inferred, however, that the highest results are necessarily the outcome of the greatest skill. As a rule, the most expert and reliable operators get only moderate strength for the best material. Lack of skill in conducting tests nearly always tells against the material tested, and good material may often be rejected because of inexperience in the in- spector; but, on the other hand, it is a frequent trick of contractors having inferior material rejected on an ordinary test to send it to one of the laboratories known to obtain abnormally high strengths, and thus get results which seem to show error on the original test. In specifications it is usually desirable to require tests showing a fair degree of strength rather than very high values. The latter are if anything less likely to give good material and unnecessarily limit competi- tion. CHAPTER VII. TESTS FOR SOUNDNESS. ART. 52 ORDINARY TESTS. SOUNDNESS is the most important quality of a cement, as it means the power of the cement to resist the disintegrating influences of the atmosphere or water in which it may be placed. Unsoundness in cement may vary greatly in degree, and show itself quite differently in different material. Cement in which the unsoundness is very pronounced is apt to become distorted and cracked after a few days, when small cakes are placed in water. Those in which the disintegrating action is slower may not show any visi- ble change of form, but after weeks or months gradually lose coherence, and soften until entirely disintegrated. The method in common use for testing unsoundness is to make small cakes or pats of neat cement, usually about 3 or 4 inches in diameter and 1/2 inch thick, upon a plate of glass, and keep them in air or water for a few days, carefully watching them to see if they show any signs of distortion or surface cracks, which may indicate a tendency to disintegration. The German standard specifications require that the 161 162 HYDRAULIC CEMENT. cakes for this test shall be 1.5 centimeters thick at the centre and have thin edges. These cakes are placed in water 24 hours after they are made, or at least not until they are firmly set, and observations are con- tinued over a period of 28 days, when, if no cracks or distortions appear, the cement is considered sound. " The cakes, especially those of slow-setting cement, must be protected against draughts and sunshine until their final setting. This is best accomplished by keep- ing them in a covered box lined with zinc or under wet cloths. In this manner the formation of heat cracks is avoided, which are generally formed in the centre of the cake, and may be taken by an inexperi- enced person for cracks formed by blowing." The Committee of the American Society of Civil Engineers recommend the following: "Make two cakes of neat cement two or three inches in diameter, about one half inch thick, with thin edges. Note the time in minutes that these cakes, when mixed with water to the consistency of a stiff plastic mortar, take to set hard enough to stand the wire test recommended by General Gillmore, i/12-inch diam- eter wire loaded with one fourth of a pound, and i/24-inch loaded with one pound. " One of these cakes, when hard enough, should be put in water and examined from day to day to see if it becomes contorted, or if cracks show themselves at the edges, such contortions or cracks indicating that the cement is unfit for use at that time. In some cases the tendency to crack, if due to free lime, will disappear with age. The remaining cake should be kept in the air and its color observed, which for a TESTS FOR SOUNDNESS. 163 good cement should be uniform throughout, yellowish blotches indicating a poor quality the Portland cements being of a bluish gray and the natural cements being dark or light, according to the character of the rock of which they are made. The color of the cements when left in the air indicates the quality much better than when they are put in water." The color test above given is not of much value, as unsound cement is very commonly of good color. The time during which these observations shall con- tinue is not specified in these rules, but in practice they are not usually carried over more than two days to a week before acceptance of the material. The French Commission upon Methods of Testing Materials recommend both a pat test and test of the amount of swelling which takes place in the mortar, as follows: " Cold Tests. (a) For this test the cement paste is formed into a pat about 10 centimeters in diameter and 2 centimeters thick, made thin at the edges. " Immediately after being made the specimens in- tended for tests in water are immersed in the same conditions as the briquettes used for tensile tests. " The specimens intended for use in the air are at once exposed to the conditions indicated for bri- quettes." " The condition of the specimens is observed at the same periods of time employed in making tensile tests (7 days, 28 days, 3 months, 6 months, I year, etc.). " (b) To measure the increase in volume of the mortar of neat cement after prolonged immersion in cold water, a bar of cement is employed 80 centi- 164 HYDRAULIC CEMENT. meters in length and 12 millimeters square, placed vertically in a glass tube 25 millimeters in diameter, which is then filled with water. " " The elongation is measured by the motion upon an arc of a needle moved by a rod resting upon the upper extremity of the bar of mortar." (See Fig. 25.) The method of testing by measuring the variation in length is also used to some extent in Germany. Methods of conducting this test are described in Art. 53- It is important in testing soundness in this manner that the tests should be continued over as long a period as possible, and many cases of unsoundness are not discovered with a 28-day test. Instances have been observed in which mortar in the form of 2-inch cubes has completely disintegrated within two years, where incipient checking was not observable for three months in a small cake test. The most common and dangerous cases of unsoundness are probably discov- ered by the ordinary tests. It may be observed, how- ever, that the fact that disintegration of mortar is not oftener observed in large constructions is probably due more to the general good quality of the cement sup- plied by the best makers, and to the frequent stability of work regardless of the nature of the mortar, than to the efficiency of the test for soundness. The quantity of water to be used in mixing mortar for tests of soundness is about the same as that used for tests of strength, although a variation in the quan- tity, within small limits, does not seem to materially influence the results. Care must be taken that the cakes be kept moist during setting and previous 10 TESTS FOR SOUNDNESS. 165 immersion, in order that they may be free from dry- ing cracks. On this account the French commission consider it preferable to immerse the specimens im- mediately after mixing, without waiting for the mor- tar to set. Some natural cements do not stand imme- diate immersion although apparently quite sound in water if first allowed to set in air. When mortar is to be used in sea-water, the pats of cement are placed in water of the same character, and as nearly as possible corresponding to the conditions of practice. Practically, however, the action of sea-water is so slow that the test is comparatively useless. M. Alex- andre found * that the first indication of disintegration may not be shown for several months or perhaps years. He also found that the action in the labora- tory did not always accord with that in the work. This was probably due to failures occurring because of the method of employing the mortar, when the cement was not defective. Tests are sometimes made of mortar to be used in sea-water by causing the water to filter through the block of mortar under pressure. This test is made in France by employing the standard cubes used in compression, which are arranged and submitted to filtration as in the tests for permeability (see Art. 67). These blocks are afterward crushed, and their strengths compared with those kept under normal conditions. *Annales des Fonts et Chaussees, Sept. 1890, p. 131. l66 flYDRAULIC CEMENT. ART. 53. MEASUREMENT OF EXPANSION. Unsoundness in cement is doubtless for the most part due to the presence of expansive elements, the action of which subsequent to the setting of the cement produces internal forces, tending to disrupt the mass of mortar and usually causing an increase in its volume. The various tests for soundness have for their object the determination of the presence of these expansives, which may be indicated either by the distortion and cracking of the mortar when present in sufficient quantity, or by a simple increase in volume without visible distortion when present in less quantity, or when more finely divided and uni- formly distributed through the mass. As a more efficient and accurate method of deter- mining the presence of expansives than is afforded by the observation of thin pats of mortar, various appli- ances have been devised for the purpose of measuring directly the increase of volume of a block of mortar. The Long-bar Apparatus. The first apparatus for the purpose of measuring change of dimensions was devised by MM. Durand-Claye and Debray. It has been in use for a number of years in France, and was recommended for use in cold tests by the French Commission on Methods of Testing Materials (see Art. 52). This apparatus is shown in Fig. 25. The test is made upon bars 80 centimeters long and 12 millim- eters square in section. The moulds in which these bars are formed consist of iron rods considerably longer than the bars to be formed, and of section 30 by 12 TESTS FOR SOUNDNESS. I6 7 millimeters, laid flat upon a table, held apart at the ends by blocks of the same section as the cement bar. and prevented from spreading by clamps at the ends. In making the test, the bar of cement is placed in a vertical glass tube 80 centimeters long and 23 nui FIG. 25. limeters in diameter, closed at the bottom and rilled with the water, to the action of which the mortar is to be exposed. To the top of the glass tube is fixed a ring carrying on one side an arc and on the other a rod, to which is hinged a needle, which travels upon the arc, being actuated by a rod resting upon the top of the cement bar. The extension of the cement bar is thus multiplied by 10 in the reading on the arc, which may be graduated, and the successive positions i68 HYDRAULIC CEMENT. of the needle read from the scale, or it may be covered with blank paper, and have the positions of the needle- point marked and afterward measured. This method requires much care in manipulation, both in making the bars and in handling them in the test. Bauschinger 's Caliper Apparatus. This apparatus, designed by Prof. Bauschinger and used in the German FIG. 26. ard Swiss laboratories, is an arrangement for measur- ing the change of length of a short bar. It is shown in Fig. 26. TESTS FOR SOUNDNESS. 169 The bars used in the test are about 100 millimeters in length and 5 square centimeters in section. They are moulded with square cavities in the ends of the bars, in which are set small plates containing centres, to form bearings for the points of the microm- eter-screw. The change in length may be measured to 1/200 millimeter. The measuring apparatus is suspended by a rod from a bar pivoted upon the top of the standard and balanced by counterweights so as to hang level. It consists of the caliper bar A, one end of which carries a micrometer-screw, reading to 1/200 millimeter, and the other end, a vertical bar to the bottom of which is hinged the needle D. The lower end of the needle has a point which is pressed by a spring against the end of the specimen in taking a measurement. The pressure of the micrometer against the specimen is regulated by bringing the point of the needle to zero on the scale F. A small bar of metal encased in wood is used as a standard in calibrating the instru- ment, the length of the standard being known very accurately at a definite temperature. Le Chatelier's Apparatus. The apparatus of Prof. Le Chatelier is designed to measure the increase in circumference of a cylindrical block of mortar. This method is recommended by the French commission for use in making hot tests, and is said to have the advantage of being much more easily manipulated than the other methods. It has also been suggested that the increase in length for long bars may not be an accurate indication of the actual swelling, as there would be a tendency for the expansion to take lines 170 HYDRAULIC CEMENT. of least resistance, and therefore the transverse swell- ing may be more than the longitudinal. This apparatus is shown in Fig. 27. It is described by Prof. Le Chatelier as follows:* "A much more simple and yet sufficiently precise measurement of the expansion can be made by letting the cement harden in cylindrical moulds of a diameter equal to their FIG. 27. height (for example 30 mm.), constructed of metal 0.5 mm. thick, slit along generatrix and provided on each side of the slit with two long needles (150 mm. for example), which serve to magnify any widening of 'the slit. The widening is equal to the enlargement, not of the diameter, but of the circumference of the cylinder of cement. Very slow-setting cements or limes, the water of which would evaporate or drain away in air, it is indispensable to immerse as soon as moulded, without waiting for them to set. The im- mersion in water of a porous mass filled with air may sometimes, by reason of capillary phenomena, give rise to a certain expansion, and even to more or less disintegration, if the hardness be insufficient. During the moulding and until the setting has taken place the mould should be kept firm by means of a suitable holder, which is removed after setting and before the measurements are begun." * Trans. Am. Inst. Mining Eng. TESTS FOR SOUNDNESS. 17 1 " For products of good quality, the distance be- tween the points of the needles does not attain I mm. in 28 days from the time of the end of setting. This test for invariability of volume, when made cold, has but little interest, since it detects only exceptionably bad products." ART. 54. ACCELERATED TESTS. The fact that many cases of unsoundness in cement are not shown by the ordinary tests when extended over a short period of time has long been recognized, and many efforts have been made to find some means of determining with accuracy and within a reasonable time whether the material be reliable. The difficulty of this with a material of so variable a nature, and in which failure may be due to so many and so diverse causes, is self-evident. Each test must be directed to the determination of the presence of some particular cause of unsoundness, and all of them seem, when indiscriminately applied to all cements, to meet material which will not pass them, although of good quality, and to which they are evidently inapplicable. Nearly all of these tests are directed to the detec- tion of the so-called expansives, and most of them attempt to accelerate the chemical action, which causes the swelling and disintegration, by the action of heat. Some of the heat tests have proven fairly successful in use, although none have been exten- sively employed as tests for the reception of material. Hot tests were first suggested by Dr. Michaelis, who proposed the use of heat to advance the hardening of 172 HYDRAULIC CEMENT. cement, with a view to determining, from the strength gained in a short time in hot water, that which would result in a longer period under normal conditions. From the first experiments in this direction it ap- peared that the results obtained in from I to 7 days in hot water might bear a definite relation to those obtained in much longer periods at ordinary tempera- tures. Later experiments showed, however, that while this might be true of a limited class of materials, as the composition is varied the effect of heat upon the strength also varies in a marked degree, some cement even showing a loss of strength in hot as com- pared with cold water. In a series of tests made by M. Deval it was found that the addition of a small percentage of quicklime to a good Portland cement caused the cement to attain less strength when kept in hot than in cold water, and Prof. Le Chatelier proposed to utilize this discovery for detecting the presence of free lime. He suggested the use of briquettes of I to 3 mortar, and the comparison of the strengths of briquettes conserved for 3 and 7 days in water at 80 C with those kept for 7 and 28 days in water at ordinary temperatures; the cements of good quality to show a strength at least equal in hot to that attained in cold water. It is to be observed in considering this test that there are some good cements which give less strength hot than cold when of normal quality. M. Alexandre found that cements rich in aluminates behave in this manner. There are also certain cements which give high tests in hot water notwithstanding the presence TESTS FOR SOUNDNESS. 1/3 of appreciable quantities of expansives. These are apparently siiicious cements of low hydraulic index, in which the free lime while rendering the cement unsound does not cause it to lose strength in hot water during the short period of the test. There are several methods of testing the soundness of cement by the aid of heat, which have come more or less into use, and have in many instances given satisfactory results. These all aim at the detection of the presence of expansives through accelerating their action by heat, and then observing the deforma- tions or measuring the expansion as in the correspond- ing cold tests. Descriptions of these tests are given in the following articles. ART. 55. KILN TEST. This test was originated by Dr. Bohme, and consists in exposing small cakes of the cement to heat in a drying oven for a definite period, and observing whether it cracks. The specifications of the Association of German Cement Makers recommend this test as a means of forming an opinion quickly, but make the ordinary 28- day test decisive as to those cements which fail to pass the kiln test. In these specifications the kiln test is described as follows: " For making the heat test, a stiff paste of neat cement and water is made, and from this cakes 8 cm. to 10 cm. in diameter and i cm. thick are formed on a smooth impermeable plate covered with blotting- paper. Two of these cakes, which are to be protected 174 HYDRAULIC CEMENT. against drying, in order to prevent drying cracks, are placed, after the lapse of twenty-four hours, or at least only after they have set, with their smooth sur- faces on a metal plate and exposed for at least one hour to a temperature of from 1 10 C. to 120 C., until no more water escapes. For this purpose the drying closets in use in chemical laboratories may be utilized. If after this treatment the cakes show no edge cracks, the cement is to be considered in general of constant volume. If such cracks do appear, the cement is not to be condemned ; but the results of the decisive test with the cakes hardening on glass plates under water must be waited for. It must, however, be noticed that the heat test does not admit of a final conclusion as to the constancy of volume of those cements which contain more than 3$ of calcium sul- phate (gypsum) or other sulphur combinations." This test is considered by some authorities to be of value for cements to be used in the air. It differs very radically, however, from the way the material is used in practice, as it effects the complete drying out of the mortar. In many instances also it is very diffi- cult to interpret, in consequence of the loss of cohe- sive strength due to drying, where no distortions appear. The effect of the withdrawal of the water necessary to the proper hardening of the mortar may vary as the rapidity of action of the material varies. The kiln test has sometimes been modified by using a moist atmosphere in place of dry air. A pan of water is placed in the oven under the specimens; the evaporation serving to keep the air saturated with moisture. Prof. Tetmajer used this method in a TESTS FOR SOUNDNESS. I?5 series of comparative tests and found it to give results similar to those of his boiling test, but somewhat less effective. His method was as follows: " The speci- mens are placed on a support in the oven, on the bottom of which are several millimeters of water. The heat is gradually applied so as to evaporate all the water in three to six hours, first that which is on the bottom of the oven, then that which has been absorbed by the mortar. Until the water is entirely evaporated the temperature remains at about 95 C. The heating is continued a half-hour after the dis- engagement of the vapor ceases, in such manner as to raise the temperature in the oven to 12.0 C. Under these conditions the interior of the briquette will reach but little above 100 C." " It should be remarked that by this method it is difficult to render the results comparable. It is not possible to make the duration of heat exactly the same for all the specimens, and after the evaporation of the water the heat in the bottom is much greater than at the top of the oven." Flame Test. A dry heat test has been proposed, and is sometimes made in Europe, by making a ball of the cement paste about two inches in diameter and placing it on a gauze in the flame of a gas-jet. The heat is gradually applied, so that at the end of an hour it reaches a temperature of about 90 C. The heat is then increased until the lower part of the ball becomes red-hot, after which it is cooled and examined for cracks. The results of this test are much like those of the dry-kiln test, and are usually difficult to interpret satisfactorily. 176 HYDRAULIC CEMENT ART. 56. STEAM AND HOT-WATER TEST. This test consists in subjecting cakes of cement, prepared in the ordinary manner, to the action of steam for three or four hours, then immersing in hot water for the remainder of twenty-four hours, and examining for cracks and distortions. Mr. Faija, by whom this test was devised, uses it in his specifications for cement in England. He describes his method of conducting the test as follows: " Briefly, it is a vessel containing water, the water being maintained at an even temperature of about 1 10 to 115 Fahr. ; there is a cover to the vessel, so that above the water there is a moist atmosphere which has a temperature of about 100 Fahr. The manner of carrying out the test is by making a pat, in the manner already described, on a small piece of glass; immediately the pat is gauged, it is placed on a rack in the upper part of the vessel, and is there acted upon by the warm vapor rising from the hot water; when the pat is set quite hard, it is taken off the rack and put bodily into the water, which, as has been already stated, is maintained at a temperature of 100 to 115 Fahr., and in the course of twenty-four hours it is taken out and examined, and if found then to be quite hard and firmly attached to the glass, the cement may be at once pronounced sound and perfectly safe to use; if, however, the pat has come off the glass and shows cracks or friability on the edges, or is much curved on the under side, it may at once be TESTS FOR SOUNDNESS. 177 decided that the cement in its present condition is not fit for use." Mr. Faija prefers the temperature given above, but other experimenters have seemed to get better re- sults using a higher one. Prof. Tetmajer obtained fairly good results with a temperature just below the boiling-point about 200 Fahr. He subjected the cakes to the action of steam for four hours, and hot water twenty hours, placing the cakes in the steam as soon as mixed. Mr. Maclay modified this method of testing, and introduced it into the specifications of the Depart- ment of Docks, New York City. Four pats or cakes of cement prepared in the usual manner were used by Mr. Maclay for his tests, the conduct of which he describes as follows:* " One of these pats is placed in a steam-bath, temperature 195 to 200 Fahr., as soon as it is made. The second pat is placed in the same steam-bath as soon as it is set hard, and can bear the I -pound wire. The third pat is placed in the steam-bath after double the interval has elapsed that it took the pats to set hard, counting from the time of gauging. The fourth pat is placed in the steam-bath at the end of twenty-four hours. " The first four pats are each kept in the steam- bath three hours, then immersed in water of a tem- perature of about 200 Fahr. for twenty-one hours each, when they are taken out and examined. To pass this test perfectly, all four pats, after being twenty- one hours in hot water, should upon examination show * Transactions, American Society of Civil Engineers, vol. xxvn. p. 412. 178 HYDRAULIC CEMENT. no swelling, cracks, nor distortions, and should adhere to the glass plates. The latter requirement, while it obtains with some cements nearly free from uncom- bined lime, is not insisted upon, the cracking, swelling, and distortion of the pats being much the more im- portant features of this test. " In hot-water tests, where the cement is very objectionable from excess of free lime, improper burn- ing, or other causes, the trouble generally shows itself in the cracking or distortion of all four pats. Where the cement is not so bad the cracking and swelling take place in the first three pats only, and when the cement is still less objectionable only the first two pats crack or swell. The cracking or swelling of No. I pat alone can generally be disregarded." " In every case of failure and rejection the cement should have been allowed to set hard in a normal temperature before subjecting it to a steam-bath." It should be noted that the effect of exposing the cement to steam before setting seems to differ with different material, depending perhaps upon the rela- tive effects the heat may exercise upon the rate of setting and upon the action of the expansives. Where the rapidity of setting is greatly increased by the heat, the severity of the test may be augmented by placing in the steam at once; but where the rate of setting is less affected, the heat may cause the action of the expansives to take place before the set, thus lessening the severity of the test. In most cases the result of the test is the same either way, but it seems fairer to permit the cement to set before submitting it to the test. TESTS FOR SOUNDNESS. 179 Mr. Maclay, however, in his specifications does not accept the results of the steam and hot-water test as conclusive in case of failure, but only considers it as cause for suspicion of the cement failing to pass it, and adds a further test for the purpose apparently of giving the material one more chance. This test con- sists in testing the strength of briquettes conserved in hot water and comparing them with those kept cold. " These briquettes are prepared and treated as fol- lows: When making the briquettes for the ordinary cold-water tests, four additional sets of five each of neat cement, and four additional sets of mortar, one part cement and two parts sand, are prepared, and allowed to set twenty-one hours in moist air of about 60 Fahr. They are placed for three hours in a steam-bath about 195 Fahr., then immersed in water maintained at 200 Fahr., after which they are broken when two, three, four, and seven days old respec- tively, and the breakings compared with the normal breakings of briquettes seven and twenty-eight days old kept in cold water." " The writer finds, in a general way, that the aver- ages of the breakings of hot-water briquettes of pure cement, four days old, are nearly as high as the nor- mal seven-day breakings cold, and the hot water seven- day breakings of the pure cement are nearly as high as the normal twenty-eight-day breakings cold, where the cement is of good quality. Where the cement is poor, and the pats show cracking and distortion, there is generally a remarkable falling off in the strength of the hot-water briquettes from the above comparison. ISO HYDRAULIC CEMENT. and one system can therefore be used as a check on the other." This is practically the same test mentioned in Art. 54, as proposed by M. Le Chatelier. Its use in this manner is recommended by M. Candlot: " The cements which contain free lime show less resistance in hot water than in cold. The cements of good quality in hot water show resistances equal to or greater than those in cold. Cements properly proportioned and homogeneous, but not completely burned, give, in this test, satisfactory results." As already pointed out, however, the relative strengths hot and cold do not depend altogether upon the presence of expansives, and it is questionable if this method is as accurate as that it is designed to check. Some unsound materials certainly give high results in the measurement of the strength of briquettes conserved hot, while there seems to be no authentic instance of any unsound cement being accepted on the steam and hot-water test, although good cement may perhaps be condemned. ART. 57. BOILING TEST. The boiling test, which is very similar in effect to that by steam and hot water, was first suggested by Prof. Tetmajer of Zurich. It consists in placing the mortar to be tested in cold water, and then gradually raising the temperature of the water to boiling. Prof. Tetmajer's method is to place the pats in cold water, immediately after gauging, raise the temperature to boiling in about an hour, continue boiling for three TESTS FOR SOUNDNESS. l8l hours., and then examine the pats for checking and softening. This method seems rather more severe in its effects upon the mortar than the other hot tests, although in general differing but little from the steam and hot- water test when a boiling temperature is employed in that test; the action, however, seems to be more energetic, and less time is required to arrive at the same results. It seems desirable in using the boiling test to permit the cement to set before subjecting it to the lest, as giving a more reliable indication of value. The results of the test is in most cases practically the same whether the cement has previously set or not. When cement is subjected to the boiling test before setting takes place it is necessary to exercise much care in the manipulation of the test to avoid any disturbance of the mortar through the motion of the water when heated. The results of the test also depend somewhat upon the rate of setting of the material, and upon the influence of heat upon the rate of setting. With quick-setting cements this action is unimportant, but with those very slow the heat may cause the action of the expansives to take place in advance of the setting, or the cement may remain soft until late in the test, and appear to fail, in consequence of dis- turbance due to the ebullition of the water. Prof. Tetmajer recommends for this, and in fact for all pat tests, that the cakes shall not be made with thin edges. His method is to roll a ball of the mor- tar, and then flatten the ball to the required thickness. The consistency of the mortar is determined by the 182 HYDRAULIC CEMENT. requirement that it shall not crack in flattening or run at the edges. For tests in boiling water this seems desirable, but for pats to be used in the ordinary cold tests the thin edges are of advantage in expediting the results where unsoundness exists in the mortar. The boiling test is frequently used in connection with apparatus for measuring expansion, in place of observing the distortions or cracks. The Bauschinger caliper apparatus is sometimes employed in this way, the bars being subjected to the boiling test, and the increase in length noted. The Le Chatelier apparatus (Fig. 27) is also usually employed in this manner. The French commission upon methods of testing materials recommend the use of the Le Chatelier apparatus for this purpose, in addition to the cold test given in Art. 52, as follows: " Hot Tests. For these tests cylindrical test-pieces are employed, 3 centimeters in diameter and 3 centi- meters high, made in metal moulds 1/2 millimeter thick, cut on a generatrix, and carrying, one on each side of the slit, two needles, 1 5 centimeters long. The increase of the distance between the ends of the needles gives a measure of the swelling. " The molds as soon as they are filled are im- mersed in cold water. After an interval of not more than twenty-four hours beyond the completion of the set the temperature of the water is gradually raised to 100 C. in from a quarter to half an hour. The temperature is maintained at 100 for six hours, and then it is allowed to cool before taking the final measurements." TESTS FOR SOUNDNESS. 183 " This method of testing is not applicable to quick- setting cements." " The standard test for deformation is to be made upon neat cement of standard consistency." The boiling test is more simple than the steam and hot-water test, and requires very little in the way of apparatus. It may readily be made anywhere without difficulty. In a laboratory where apparatus may be kept in continuous operation the steam and hot- water test offers some advantages in ease of operation, and in permitting continuous study of material under examination beyond the limits of the regular test. ART. 58. PRESSURE TEST. Dr. Erdmenger has devised a high-pressure steam test for soundness. In this test the pats are allowed to harden for three days, &nd are then exposed to steam for six hours at a pressure of from 3 to 20 atmospheres. The originator has used this test for a number of years, and claims that it gives very satis- factory results, and that when properly carried out it enables a complete and rapid judgment to be formed on a cement containing magnesia.* As with all accelerated tests, there has been much dispute as to the reliability of this one, some authori- ties claiming that many of the best Portland cements fail under it, others considering it nearly infallible. It is a test requiring more extensive appliances for its execution than the other hot tests, and has not been so largely used as the others. * Journal Society of Chemkal Industry, vol. XII. p. 927. 184 HYDRAULIC CEMENT. The test is executed both upon the neat cement and sand-mortar, the severity of the test being greater upon the neat cements. Dr. Erdmenger claims that the best cements show no defect under this test at high pressures (forty atmospheres); that others may show defects at high pressures, although safe in prac- tice (especially in the neat test), but they are not first quality; while cement which cannot stand the pressure test of about twelve atmospheres in the sand tests should be rejected as faulty. ART. 59. CHLORIDE-OF-CALCIUM TEST. The action of chloride of calcium in the water used in gauging or conserving the briquettes is similar to that of heat. The results of the experiments of M. Candlot have already been discussed (Art. 22). In presence of calcium chloride the slaking of lime is greatly accelerated, the effect increasing as the solu- tion is more concentrated. It also makes the setting slower, especially for cement rich in aluminate of lime. If cement containing free lime be gauged with a concentrated solution of calcium chloride, it may therefore cause the lime to become slaked before setting, so that no subsequent swelling will take place. If the solution be less concentrated, the lime may not be completely slaked before setting unless the quantity be very small, while the slaking is hastened and swelling occurs after the setting if the free lime be in sufficient quantity. If a chloride-of-calcium solution is used to keep the cement after setting, the expansive action of the free TESTS FOR SOUNDNESS. I5 lime is accelerated, and a very small quantity may produce swelling. This test, as devised by M. Candlot, is to mix the cement paste with a feeble solution of calcium chloride, which causes the slaking before setting of so small a percentage of lime as may not be objectionable in the cement, but does not eliminate a larger quantity. Then immerse the mortar in the same solution, and thus augment the swelling if the free lime be present in serious quantity. The method of conducting the test as proposed is to " mix the paste for the cakes with a solution of 40 grammes of chloride of calcium per liter of water, allow them to set, immerse in the same solution for twenty- four hours; then examine for checking and softening as in the other tests." The chloride of calcium is said to exert no action upon free magnesia. Upon free lime it seems, how- ever, to be quite efficient, and it may often be useful in pointing out doubtful material. The action becomes less energetic as the tempera- ture is increased, and it should not be used in con- nection with the hot tests. ART. 60. VALUE OF THE ACCELERATED TESTS. The reliability of the various accelerated tests for determining the soundness of cement in use is a matter concerning which there is much dispute amongst authorities on the subject. These tests have not as yet come into general use, and considerable opposi- 1 86 HYDRAULIC CEMENT. tion has been developed to them, although in certain instances they are employed. The question of adopting the accelerated tests was discussed by the Association of German Cement Manufacturers in 1891, and it was voted to adhere to the standard test already existing. It was also stated in the report that experiment had not proved the in- accuracy of the standard test (the twenty-eight-day cold-pat test), and while the accelerated tests may be useful to the manufacturers in determining the char- acter of their cements, they should not be used by the consumer, with a view to forming a sound opinion on the constancy of volume. The results of a number of experiments are also given in this discussion, showing that a number of cements which had not withstood the kiln test were sound when used under water at normal temperatures, or if placed in air after being kept moist for several days. It may be remarked also that most of the dis- cussion seemed to refer mainly to the dry-heat test. Some of the leading German cement experts, how- ever, are strong advocates of the use of heat tests. Dr. Michaelis expressed his approval of them, and stated that he had experimented with and used them satisfactorily for a number of years at the Charlotten- burg experiment station. Dr. Erdmenger also declares that experience has shown the high-pressure steam test to give an accurate determination of the per- manence of volume of the material. He states that most of the best German cements have stood this test up to a pressure of forty atmospheres for sand-mortars and several for neat cement, the best of them show- TESTS FOR SOUNDNESS. l8/ ing no defect whatever; and he thinks there is no ground for the statement that many good cements will not pass the test. Prof. Bauschinger found that cements which had given good results in the ordinary 28-day test, and also upon the cold-pat test for a year, failed when formed into prisms of I to 3 mortar for testing in the Bauschinger caliper apparatus. Expansion was de- tected in six months by measurement, and afterward became visible to the naked eye. The German Conference upon Methods of Testing* recommend the continuance of the present practice (the kiln test for quick determinations and the 28-day cold-pat test as decisive), but they state that " the boiling test may undoubtedly be considered as the most conclusive and rapid test for the determination of constancy of volume of Portland cement, of slag- cement, and of trass," and refer a particular test to the sub-committee for further examination and report. The French Commission upon Methods of Testing f recommend the use of the boiling test as the best method of arriving at a quick determination of the. permanence of volume for Portland cement, the amount of the increase in volume to be measured directly, instead of simply observing the effect upon a small pat. * Resolutions of the Conventions held at Munich, Dresden, Berlin, and Vienna, for the purpose of adopting uniform methods for testing materials, by J. Bauschinger; translated by O. M. Carter and E. A. Gieseler (Washington, 1896). f Commission des Methodes d'Essai des Materiaux de Construc- tion; Rapport (Paris, 1894 and 1895). 188 HYDRAULIC CEMENT. In England the steam and hot-water test has been introduced by Mr. Faija, using a low temperature; while in the United States the same test is employed in the specifications of the New York Department of Docks, using temperature near boiling. In some in- stances the boiling test has also been used in the United States by manufacturers in the study of their product, with very good results. The advisability of adopting some form of acceler- ated test in ordinary specifications is still an open question, needing for its determination more accurate knowledge than is now available of the behavior of the various kinds of cement when subjected to such tests. Numerous experiments have been made for the purpose of deciding the matter, the results of which differ so widely from each other as to involve the question in great confusion. Most of these experi- ments have been made by studying the action of heat upon the various ingredients to which unsoundness is usually attributed, and arguing from the results whether the hot test gives an accurate indication of the presence of these ingredients in the cement. The most common method has been to mix a small quan- tity of quicklime with a good Portland cement, and then observe the action of the test upon the mixture. This involves the assumption that certain percentages of free lime are sufficient to render the cement un- sound. The results of tests made in this manner are also subject to much variation, due to the nature of the cement, its rapidity of action, and the quantity of free lime which may have been originally present in it. TESTS FOR SOUNDNESS. 189 These experiments have been of much value in showing the effect of the accelerated tests upon various substances, and in discovering the reasons for many of the apparently contradictory results with them. The question of the ultimate adoption of tests of this character must, however, be determined by a comparison of the results obtained by their use upon the material as found in market with the action of the same material in practical use. To this end experi- ments are desirable which shall systematically compare the results of accelerated tests upon ordinary cements with the results of tests under normal conditions ex- tended over long periods of time. An extended series of experiments of this character has already been car- ried out by Prof. Tetmajer at Zurich,* and a smaller series upon material in use in the United States at the laboratory of the College of Civil Engineering at Cornell University. A careful study of the available results of experi- ment seems to justify the following statements: 1. Small percentages of uncombined lime or mag- nesia in the cement are commonly detected by the use of the heat tests, and the same ingredients in suffi- cient quantity render the cement unsound in ordinary use. 2. Cement liable to change of volume when em- ployed under normal conditions is almost invariably detected when submitted to the hot-water test. There seems to be no well-authenticated instance of failure to condemn really defective material. * Methoden und Resultate der Priifung hydraulischen Binde- mittel (Zurich, 1893). 190 HYDRAULIC CEMENT. 3. Nearly, if not quite, all of the best brands of Portland cement, and many of natural cement, as found in market, readily meet the requirements of these tests, which therefore do not impose so severe limitations upon the choice of cement as is commonly supposed. With natural cement the results of these tests vary somewhat with the character of the cement, and the same tests do not seem to be universally applicable. This, however, is a matter which can only be determined by careful experiment upon each of the various classes of natural cement. Many of them bear the severe tests fully as well as the Port- land. 4. While these tests rarely, if ever, fail to detect an unsound cement, and most good cements readily pass them, there are occasional instances of cements con- demned by the heat tests, which are not unsound when kept at normal temperatures in fresh water. These cements for the most part seem to fail if kept in dry air or are subjected to the action of sea-water. Apparently, a degree of unsoundness may exist which is sufficient to cause the change of volume to take place in a short time in hot water or in a longer time in dry air, while in cold water the action of the ex- pansives takes place without injury to the mortar. This is shown by the fact that pats of mortar which had failed in the boiling test at the time of mixing, after being kept several months in cold water, and then subjected to the boiling test, were found to stand the test perfectly, showing that the action of the expansives must meanwhile have taken place. In TESTS FOR SOUNDNESS. IQI several instances pats of cement acting in this manner were found to blow in dry air. The unsoundness of cement condemned by the heat tests, when the mortar is to be kept submerged in fresh water, is therefore, in many instances at least, questionable. In the large series of tests made by Prof. Tetmajer, out of 139 samples of Portland cement 17 failed in the boiling test, all of which also failed in a long time in dry air, while only 2 were defective in long time in fresh cold water. In the tests made by the author, the percentage of failures to total number of samples is less than at Zurich, 5 in 53, and 3 of the 5 samples which were rejected by the hot test were later disintegrated when kept in cold fresh water. 5. The fineness of the cement has an important bearing upon its behavior under the heat tests. All ordinary cement probably contains some small propor- tion of expansives, which in the finely ground material may become hydrated without injury to the mortar; but if the fine material be sifted out, and only the coarse particles employed in the test, the slower action in the larger particles may cause distortion and crack- ing to take place. Whether the same difference usually follows in cold water has not been satisfactorily determined. 6. Portland cement which contains free lime in sufficient quantity to cause it to swell in the hot test may frequently be made to pass that test by adding a small quantity of sulphate of lime. The action of this salt upon the setting of cement has already been discussed (Art. 21). The rules of the Association of IQ2 HYDRAULIC CEMENT. German Cement Makers permit the addition of a small proportion of sulphate of lime for the purpose of regulating the rate of setting, and conclude that no injury is thereby caused to the cement. The action of the sulphate of lime to correct the expansive action of free lime is but imperfectly known. The fact that it causes cement containing free lime to pass the hot tests is well known, but whether the corrective influence extends to the action of the expansives when the material is used under normal conditions has not as yet been satisfactorily determined. The strongest objection that has been urged against the use of hot tests is that they fail to detect free lime in the presence of the calcium sulphate. The justice of this objection can only be decided by ex- periments extending over considerable periods of time to determine whether the material so passed is sound under normal conditions. Doubtless for use in sea- water it would be necessary to limit the quantity of sulphuric acid. The whole subject of hot tests must still be regarded as in the experimental stage, and further experiment is necessary to determine more fully the connection between the results of tests and the action of the cement in use. Under present conditions it may be said that the presumption is fairly against the soundness of a cement failing to pass the hot-water test and in favor of that which succeeds in passing it; but variations in the cement and in the conditions under which it is employed may affect the results, and they cannot be TESTS FOR SOUNDNESS. 193 relied upon with certainty for all material. Upon ordinary work to be kept in fresh water there is probably no considerable danger of unsoundness in the use of any good brand of cement, although failure sometimes occurs; but for the more important works, and particularly those to be subjected to the action of sea-water, it is reasonable to apply such tests as are likely to insure good material, even at the risk of ex- cluding other good material. ART. 61. AIR-SLAKING. Sometimes fresh cement, when first opened after being shipped, will, if tested at once, show an abnor- mally rapid rate of setting, and subsequently harden very slowly, so that on short-time tests very low ten- sile strength may be given. If, however, this cement be exposed to the air for a few days, it may resume its natural rate of setting, and attain proper strength upon the tests. In some laboratories it is customary to thus expose cement to the air a short time before testing, and this process is termed air-slaking. The propriety of air-slaking in testing cement is questioned by some engineers, upon the ground that the cement to be used in the work will not be treated in the same manner. In England it is customary to give such exposure to all cement to be used upon im- portant work for at least ten days, but in the United States the cement is commonly used just as received from the manufacturer. The general practice seems to favor air-slaking in testing, and probably a cement capable of regaining 194 HYDRAULIC CEMENT. its normal condition in a few days' exposure will not endanger the work, even if used at once, but it would doubtless be better in using such cement to air-slake the whole before using. It may be remarked, how- ever, that air-slaking does not ordinarily seem neces- sary. The cement commonly placed upon the market by the best makers does not need it. While it may be allowable to give the material the benefit of the operation, probably few instances of rejection would occur on account of its omission. In many instances the effect of air-slaking a cement requiring it disappears with time; that is, the strength of the mortar after three or six months may be as great for that mixed before as for that mixed after air- slaking, although the difference of strength on a test extending over a few days is very considerable. This would indicate the necessity of air-slaking the whole of the material if early strength is to be developed in the work, although the ultimate strength might per- haps be satisfactory either way. If the cement blows or shows unsoundness on the first test, it should not be used without exposure, as it would indicate a degree of unsoundness likely to be serious, even though this also disappear in the second test. The question is simply as to the quantity of the expansives which may be present without danger to the work in which it is used. CHAPTER VIII. SPECIAL TESTS. ART. 62. TESTS OF ADHESIVE STRENGTH. THE ability of cement-mortar to firmly adhere to any surface with which it may be placed in contact is one of its most valuable properties, and quite as impor- tant as the development of cohesive strength. Tests for adhesive strength are not commonly employed as a measure of the quality of the material, because of the uncertain character of the test and the difficulty of so conducting it as to make it a reliable indication of value. The adhesive properties of the cement are to a certain extent called into play in the tensile tests of sand-mortar, and may be inferred from a compari- son of neat and sand tests. Adhesive strength is developed more slowly than that of cohesion. The difference between the two, which is usually considerable during the early period of hardening, is gradually lessened with time. This is illustrated in the greater time required for sand- mortar than for neat cement to harden. Experiments u-pon the adhesion of mortars to various substances are sometimes made, both for the purpose of comparing various cements or methods of use, and to study the relative effects of various kinds 195 196 HYDRAULIC CEMENT of surfaces. Such experiments are quite desirable with a view to the extension of knowledge of this very important quality. The common method of making this test is to pre- pare briquettes, of which one half are of neat cement or sand-mortar of the ordinary form for tensile speci- mens, and the other half a block of stone, glass, or other material to be used, of the same section as the mold at its middle, and arranged to be held by a special clip in the testing-machine at the other. In Germany and Switzerland the apparatus shown in Fig. 28 is employed. The test-piece is shown at a; FIG. 28. it has a section of 10 square centimeters twice that of the standard tensile specimen. The mortar end is enlarged to a wedge shape to catch the upper clip of the testing-machine, while the other end is formed of a parallelepiped of marble or ground glass for standard tests, with a cylindrical groove cut in its side, which fits into a special clamp (shown at c]. This clamp is held by screws in the lower clip of the testing-machine, as shown at d. For forming the briquettes molds SPECIAL TESTS. 197 are used, in the bottom of which the blocks are placed and the mortar filled in on top. In some laboratories blocks of marble have been used for standard tests in comparing different ma- terials; but the large amount of labor involved in the preparation of the blocks, and the difficulty of getting always the same surface, has been a bar to the exten- sion of this method. Ground glass has been more commonly employed, the same blocks being repeatedly used. Dr. Michaelis has also used for this purpose standard blocks of cement-mortar, of the same form, which are easily prepared, and more uniform in material and surface. The German Conference on Methods of Testing, before mentioned, did not define a standard test, but referred the matter to a sub-committee, with the recommendation that the German apparatus just described be utilized. The French commission recommend the use of a briquette of double T form, suggested by M. Candlot. It is shown in Fig. 29. A mold is employed, made to the form of half the briquette, which is set down over the block to be used in forming the specimen. The recommendations of the commission are as follows : " A. To compare the adhesive strengths of cements, there is submitted to the tension test a briquette of the double T form, shown above, each of the ma- terials to be studied constituting half of one of the specimens. " B. Standard tests, intended to compare the ad- I 9 8 HYDRAULIC CEMENT. hesive strengths of various cements to the same ma- terial. " (a) The standard adhesion-blocks are prepared of mortar composed of one part, by weight, of artificial Portland cement, passed thiough the sieve of 900 meshes per square centimeter, and two parts of stand- ard sand No. 3 (that passing a sieve of i mm. open- FIG. 29. ings and retained upon one of 1/2 mm. openings). The mortar is gauged with 9 per cent water and strongly compressed in the moulds, in the bottom of which is placed a movable metallic block. The ad- hesion-blocks are immersed in fresh water after twenty-four hours, and kept thus until used, or at least for twenty-eight days. When they are to be utilized, they are dried and the surface is smoothed with emery paper. " (b) The standard plastic mortar (see Art. 44) is employed for these tests, introduced by the pressure SPECIAL TESTS. 199 of the trowel into the moulds, placed in such manner that the standard adhesion-block forms the base. The briquette is removed from the mould when com- pletely set." " (c) The briquettes are tested in the same manner and at the same periods prescribed for tensile tests." " C. Tests intended to compare the adhesive strength of a cement to various materials; ' ' (a) For these tests the same methods are adopted, with the difference that the standard blocks are re- placed by blocks formed of the various materials to be used. If the material can be moulded the block may be made in the same way as the standard block. If the material is solid, a plate several millimeters thick may be formed having one face dressed. This is placed in the bottom of the mould and the block filled out with neat cement." ART. 63. CHEMICAL ANALYSIS. Chemical analysis is of very great value in the study of the properties of various cements, and is com- monly employed by manufacturers in regulating the quality of their products. It is not commonly used for the purpose of determining the quality of a cement, and is not of much value as a test for the reception of material. The quality of the cement depends not only upon the ingredients being properly proportioned, but also upon the state of combination of the ingredients, and this in turn depends upon the manipulation given the material in manufacture. Analysis shows the propor- 2OO HYDRAULIC CEMENT. tions of the various ingredients, but does not show their state of combination. The results of an analysis may show that the composition is such that a good cement may be made from the ingredients, but other tests are necessary to show whether it has been made. Chemical analysis .may, however, prove a cement to be bad through its containing objectionable propor- tions of ingredients known to be injurious. Thus the allowable percentage of sulphur compounds is some- times prescribed for cement to be used in maritime work; in fresh water they may be unobjectionable. The expansive elements, free lime or free magnesia, cannot be detected by analysis, as their presence is not necessarily dependent upon the proportions of total lime or magnesia present. The French specifications, devised by M. Guillain, for Portland cement reject those containing more than 1$ of sulphuric acid, or sulphides in appreciable quan- tity; while those containing more than 4$ of oxide of iron, or which give less than 44/100 for the ratio of the sum of the weights of silica and alumina to the weight of the lime are to be regarded with suspicion. A large percentage of volatile elements in a cement indicate -either insufficient burning or deterioration with age through exposure to the air. M. Candlot states that a chemical analysis may be useful in showing the adulteration of cement, some- times practised in Europe. Upon sifting the cement and separately analyzing the coarse and fine portions an unadulterated cement should show practically identical results for the two analyses. He also states that blast-furnace slag, which is a common adultera- SPECIAL TESTS. 2OI tion in Portland cement, may sometimes be discovered by the odor of sulphuretted hydrogen upon treating it with hydrochloric acid. Method of Analysis. The following method of making an analysis of cement is given by M. Durand- Claye:* " Weigh two grammes of the specimen in powder, place in a porcelain evaporating-dish with a little water and hydrochloric acid in excess. Evaporate dry in a sand-bath at a moderate temperature. The mass should be stirred several times with a glass rod after the silica has jellied. When the drying is com- plete, place the dish in a furnace and heat quickly, but not to redness, to remove the last trace of acid and humidity. After cooling, treat the new material with a slight excess of hydrochloric acid, and evaporate a second time to dryness. Then treat again with acidulated water. The liquor should appear yellow, and the residue white, not reddish. If it has that color, let the dish and its contents heat some time in a sand-bath until all the red color disappears. " Silica. Place the contents of the dish upon a fil- ter. The white powder obtained is the silica of the specimen. Dry, calcine, and weigh in the ordinary manner. "Silicious Sand. The cement sometimes contains grains of inert silicious sand, which is found with the silica. It is recognized by the sound of the grains on the bottom of the dish when stirred with the glass rod. To determine the silicious sand, dissolve the cement * Chimie appliquee a 1'art de 1'ingenieur (Paris, 1885). 2O2 HYDRAULIC CEMENT. in hydrochloric acid, and let the liquor digest during one day. Then, upon decanting, the silicious grains rest upon the bottom; wash, collect, and weigh. "Alumina and Iron. The filtrate, after removing the silica, is treated with a few drops of nitric acid, and brought to ebullition for a moment. The pro- toxide of iron, if any, is transformed to peroxide. Then, by saturating with a slight excess of ammonia, and boiling, a precipitate is formed, which is collected on a filter placed in a funnel above a flask. It is washed carefully with warm water. When well drained, it is placed on a sand-bath, dried, burned, and weighed. The weight found, diminished by that of the cinders of the filter, is that of the peroxide of iron and alumina soluble in acids. The ammonia precipitates the two bases from the solution, while it is without effect upon the salts of lime, and does not affect the salts of magnesia previously acidulated. When the limestone contains much magnesia, like dolomite, some of that base may be precipitated by the ammonia. In this case it is prudent to redissolve the precipitate in hydrochloric acid, after having separated by filtra- tion, and to reproduce it a second time with ammonia. " Lime. Add oxalate of ammonia to the last fil- trate, and the lime only is precipitated, as magnesia is not separated in the water containing ammonia salts. In order that the oxalate of lime may be well precipi- tated, it is necessary to employ a large excess of the reagent. The precipitate is collected better when the temperature is higher; it is therefore well to heat the liquor, and dissolve the oxalate of ammonia in boiling water. In order to obtain a complete precipitation SPECIAL TESTS. 2O3 the liquor must be alkaline. The precipitate should not be filtered immediately, as the liquor continues working. The flask should be put in a furnace and boiled for half an hour, or still better be left in a sand-bath half a day. The boiling in the furnace of a grainy precipitate like oxalate of lime gives rise frequently to shocks, which perhaps throws the material out. When the precipitate is well formed and the liquor clear, place upon a filter in a funnel over a flask and wash carefully with hot water; then, placing the funnel in a sand-bath, dry the filter and precipitate before calcining. A simple scorching is not sufficient. The oxalate at the temperature of burning is decomposed more or less completely: a part becomes caustic lime; some is transformed to carbonate, or it remains as oxalate. It is therefore necessary to transform it to a definite compound. " When the means are at hand for a sufficiently energetic burning to heat the platinum cherry-red, put the precipitate, with the filter, in a platinum crucible furnished with a cover, and calcine. Five minutes are sufficient with an annealing bellows. The oxalate is transformed to caustic lime, which is imme- diately weighed in the crucible when cold. The weight, less cinders of filter, is the lime contained in the sample. It is well to assure that the calcination is complete; for this purpose, after weighing, pour a little water in the crucible, then acid; if any carbon- ate is present, bubbles of gas will be given off. " When an apparatus for heating so energetically is lacking, the oxalate is sometimes changed to car- bonate. Place the precipitate in a platinum dish and 204 HYDRAULIC CEMENT. burn in a muffle, or to a heat which makes the dish red. When it is cold, pour a solution of carbonate of ammonia; then evaporate in a sand-bath, dry over a furnace, and weigh. It is well to renew the treat- ment of carbonate of ammonia until two weights are obtained which are in accord. The weight of lime is 0.56 that of the carbonate. " Magnesia. The filtrate contains the magnesia, with the ammonia salts which are introduced in the course of the analysis. Precipitate with phosphate of soda. Let it stand cold half a day before filtering, until finally assured that the ammonia-magnesian phosphate which is formed is entirely crystallized, or at least that not more than insensible traces remain in solution. When placed on the filter some of the crystals adhere to the flask, and may not be detached by washing. These are dissolved in acid and then again precipitated with ammonia in non-adherent granular crystals, which readily wash upon the filter. The washing from the flask to the filter should be done with cold water charged with ammonia. Pure water, especially if warm, dissolves an appreciable quantity of the precipitate. After drying, burn with the filter and weigh. 40/111 of the weight, less the cinders of the filter, gives the weight of magnesia. " Volatiles. Weigh a new specimen in a crucible and calcine quickly for five or ten minutes. The crucible when cold is weighed with its contents, which are composed of lime and the other fixed elements. The difference in weight is the loss upon ignition. " Sulphuric Acid. If it is desired only to obtain the sulphate of lime, take a specimen (5 grammes for SPECIAL TESTS. 2O$ example) and digest cold in a solution of carbonate of ammonia for twenty-four hours. Agitate the mix- ture often, especially in the beginning. The sulphate of lime is decomposed, carbonate of lime and sul- phate of ammonia being produced. Filter, saturate the liquor with hydrochloric acid, and boil to remove all carbonic acid. Then pour in chloride of barium, and the sulphuric acid is precipitated as sulphate of barium, which is collected on a filter, dried, and weighed. The filtration should not take place imme- diately: it should stand until the sulphate of barium gathers and the liquor clears. " Sulphides. If the total sulphur is to be obtained, it is to be brought into contact with an oxidant, which transforms the sulphide into sulphate. A little chlo- rate of potassium is added, and then it is attacked with precaution by hydrochloric acid. Evaporate dry, and digest the residue, after cooling, with carbonate of ammonia. The analysis is then like the preceding. If the two methods be simultaneously applied, and by the second a result be found larger than by the first, the difference represents the sulphuric acid formed from the sulphides. " Alkali. Take a special specimen and separate the silica, alumina, and iron as already indicated. Pour the filtrate so obtained in a porcelain dish and evaporate dry, then calcine in a muffle to drive off the ammonia salts and decompose the oxalates. The residue is composed of alkaline salts, and of magnesia, free or carbonated. The latter are insoluble, and by treating with water and filtering, a liquor is obtained containing only the alkaline salts. The liquor is 2O6 HYDRAULIC CEMENT. evaporated to dry ness in a platinum dish in presence of an excess of sulphuric acid, which displaces the other acids. The residue is alkaline sulphate. Heat to redness to drive off all excess of acid, and weigh. Treat with water acidulated with hydrochloric acid, and precipitate the sulphuric acid with chloride oi barium. The weight of sulphate of barium gives that of sulphuric acid, which subtracted from that of the alkaline sulphates gives fixed alkali. " When the specimen contains considerable sul- phuric acid this method may give erroneous results. The sulphate of magnesia is only partially decomposed at the moment of calcination of the ammonia salts, and a part of that salt is found with the alkaline sul- phates. To remove that cause of error, after having eliminated the ammonia salts and treated the residue with water, pour on a little baryta water, which pre- cipitates the magnesia and gives at the same time compounds insoluble in sulphuric or carbonic acid. The filtrate contains then only alkaline chlorides, free alkali, and baryta. Saturate with sulphuric acid and filter, then evaporate dry, and calcine, and obtain the residue of alkaline sulphates as in the other case. " Separation of Iron and Alumina. After having burned and weighed these two bases, digest and heat with hydrochloric acid until dissolved. Mix with citric or tartaric acid until ammonia in excess does not produce a precipitate ; then add the sulphohydrate of ammonia. The alumina is not precipitated, while the iron passes to the state of insoluble sulphide. Boil and collect on a filter, dry, burn, and weigh. The sulphide decomposes and becomes peroxide. SPECIAL TESTS. 2O/ The alumina is the difference between the total weight and that of the peroxide of iron. " Or, after dissolving in hydrochloric acid, evaporate the larger part of the hydrochloric liquid so as to have a slight excess of acid; then pour a solution of pure potash in excess, and boil until the precipitate obtained presents a dark reddish brown color. Filter and wash the precipitate of peroxide of iron. An appre- ciable quantity of alumina will be precipitated with the iron; to remove this, redissolve in hydrochloric acid and repeat the operation a second or third time. The absence of alumina is shown by the ammonia salt giving no precipitate with the filtrate. " This operation requires great care and consider- able time." ART. G4. TESTS FOR HOMOGENEITY. In Europe various tests have been proposed for the purpose of detecting the adulteration of Portland cement. These tests are not usually of a nature in- tended for use in specifications for the reception of material, but may sometimes be of use in studying the characteristics of various brands of cement or in classi- fying a product of doubtful character. The materials sometimes met with as adulterations include powdered limestone or shale, blast-furnace slag, hydraulic lime, trass, etc. The nature of the tests depends upon that of the adulteration to be discovered. The specific-gravity test is sometimes utilized for this purpose, the foreign matter being lighter than cement. The differences, however, are so small that the test at best is a rather uncertain one. 208 HYDRAULIC CEMENT. The determination of the loss upon ignition may show the presence of foreign matter containing vola- tile elements, while chemical analysis, separating the fine and coarse parts of the cement and comparing the results, is sometimes resorted to, as stated in Art. 63. Microscopic Test. The use of the microscope for the purpose of determining the character of the sub- stances present in cement has often been proposed. Prof. Le Chatelier made a careful study of Portland cement by examining sections of unground clinker by polarjscopic analysis. He thus demonstrated the possibilities of this method in the scientific investiga- tion of the nature of the material. This method is not, however, applicable as a test for homogeneity. Attempts have also been made to determine the character of the cement by studying the grains of which the powder is composed under the microscope. For Portland cement, it has been ob- served that the active portion of the cement is com- posed of grains of angular form and metallic lustre, and that the parts of earthy appearance are probably inert. It has also been found that the color of the grains seems to bear some relation to their value in the cement. Further study may reveal more positive indications of value based upon the microscopic ap- pearance, but it seems unlikely that this method will be applied to the determination of value in practice. For tests of homogeneity, the employment of an ordinary magnifying-glass may perhaps give useful results. M. Feret, who made a study of the matter and presented a report to the French Commission upon Methods of Testing Materials, recommends that the SPECIAL TESTS. 2OQ material to be examined be sifted through the sieve of 4900 meshes per square centimeter and the portion retained by that sieve be used in the examination, on account of the difficulty of observing the grains when mixed with the impalpable powder. It is also sug- gested that two glasses be used : the first of a power of about three diameters to examine the general appearance and uniformity of the material; the second of about eight diameters to study in detail the various grains. The material is placed upon a black surface to be examined. The points to be noticed are the form, color, and transparence of the grain and the appearance of its fracture. It is also desirable to test the hardness with a steel point and the solubility in a drop of water or acid. It is to be borne in mind that the character of the material in the coarse particles is not necessarily the same as in the finer portions, or at least the proportions of the various ingredients may not be the same. M. Feret found that pure cement, thoroughly well burned, gave " grains all of the same appearance, black, opaque, angular, hard, and with a rough frac- ture. Mixed with a drop of water they show at first no change, but after several minutes a sort of halo appears, produced by the beginning of crystallization of the soluble compounds. These grains color im- mediately in hydrochloric acid to a yellow, but are completely dissolved with difficulty. " Rock less burned gives grains equally opaque, but of color less deep, varying to brown, gray, or green. The underburned rock gives gray or yellow grains, which crush easily under the point of a knife, 210 HYDRAULIC CKMENT. and are attacked by acids with the disengagement of carbonic acid." " When, instead of material prepared in the labo- ratory and thoroughly homogeneous, the residue ob- tained from market grades of cement is used, the appearance is very different even if the material be of best quality. The color is less deep and the material appears very heterogeneous. Such cements contain always, with many of these black, brown, or green grains, a large number of more vitreous material, green, yellow, and white, without containing foreign matter." " When the cement is less well burned the color of the large grains becomes more clear, and the glass shows an increasing number of gray friable grains, the black grains decreasing in number." " It is not uncommon to find in the cement clinker pieces of bright colors, blue, green, violet, red, and white; these materials are usually soft and porous." " Carbon in black and brilliant grains, with con- choidal fracture, is found in all the specimens, and easily known." " The debris of flint from the millstones is not easily distinguished from certain underburned cement grains; they are the gray morsels, hard and opaque. They differ from the cement grains in not being attacked by water or acids." " The appearance of grains of slag vary according to the nature of the slag. Commonly, the grains are compact, of a bluish gray color, and smooth clean fracture. Sometimes the grains are vitreous and black. The granular slags employed in making slag- SPECIAL TESTS. 211 cements are soft, and leave few grains. Their debris has the appearance of colorless glass, or tint of yellow or green, and fractures easily under the point of a knife." " Grappiers give round grains, usually more clear in color than the grains of cement." " Gypsum added to cement clinker before grinding may show large white crystalline grains easily seen. They may be identified by their hardness and solu- bility. Plaster is difficult to recognize, because it grinds fine and leaves no crystals." When plaster has been added to cement, it may sometimes be detected by separate analyses of the fine and coarse parts, as to sulphuric acid. The fine parts containing the plaster give the higher results. Le Cliatclicr 's Test. This method of detecting adulteration, proposed by Prof. Le Chatelier, consists in forming a liquid, by the mixture of methylene iodide and benzine, of density slightly below that of cement and above that of slag, and separating by its means the cement from the adulteration. It is thus described by Prof. Le Chatelier: " The first operation is the preparation of a liquid of proper density for the separation, 2.95 for exam- ple. In this liquid the cement sinks to the bottom and the slag floats on the surface. To prepare the liquid add to the methylene iodide of density 3.1 a small quantity of benzine, stopping the moment a crystal of aragonitc of density 2.94 just remains at the surface. It is well not to make the mixture directly, but to make two mixtures the one a little lighter, the other a little heavier, than the density sought; 212 HYDRAULIC CEMENT. thus obtaining a more progressive variation, and more easily regulated." " The apparatus consists of a glass tube (Fig. 30), 10 millimeters in diameter, 70 millimeters high, FIG. 30. widened into a funnel at top and terminated at bottom in a point, with an orifice I millimeter in diameter. This orifice is closed on the interior a little above the bottom, by a small emery stopper, fastened to a long glass rod, which issues from the top of the tube. " To make an experiment, the stopper is wet with water (grease is dissolved by the liquid), to prevent leakage. Two grammes of cement are introduced, then five cubic centimeters of liquid (density 2.95). It is then agitated in a lively manner with a small SPECIAL TESTS. 21$ platinum thread bent around into a hook, in order to drive out the bubbles of air and thoroughly mix the cement in suspension in the liquid. Finally, it is allowed to settle for an hour; after that time there is formed two layers the cement at the bottom and the slag at the top. The stopper of emery is lightly raised by the rod to which it is attached, letting out the cement and part of the liquid, and is then replaced. The cement is caught upon a filter, through which the liquid passes into a flask, when it is ready for another operation. " The slag and remainder of the liquid is received upon another filter. Finally, the tube and filters are washed with benzine and dried; the cement and slag are weighed, and analyzed chemically if thought proper." The cost is said to be the principal objection to this method, although but a small quantity, a part of a cubic centimeter of iodide of methylene, is required for each operation. ART. 65 ABRASIVE TESTS. Cement to be used for sidewalks, floors, or artificial stone is sometimes submitted to tests for resistance to abrasion. This test is frequently employed in Ger- many, the apparatus designed by Prof. Bauschinger being used. This apparatus consists of a cast-iron disk 122 centimeters in diameter and 3 centimeters thick, mounted to rotate horizontally at about 20 revolutions per minute. Specimens 6x6, 10 X 10, or 12 X 12 centimeters in section are employed. 2I 4 HYDRAULIC CEMENT. They are held upon the disk with a pressure of 30 kilogrammes, and 20 grammes of standard sand are FIG. 31. added for each 10 turns. 200 turns are given, and the loss in weight of the specimen is determined. Fig. 31 shows a similar apparatus made by the SPECIAL TESTS. 21$ Riehle Bros. Testing-machine Co., and used for brick and stone tests. Abrasion tests, when employed, are usually made for both neat cement and sand-mortar. The test of mortar as used in practice is evidently the more important. Resistance to abrasion varies with the character of sand used, and for sand-mortar de- pends upon the adhesion of the cement to the sand and the hardness of the grains of sand. Sand-mor- tars with moderate proportions of sand give better resistance to abrasion than neat cement. ART. 66. TESTS FOR POROSITY. Tests for the porosity of mortar are interesting in studying the properties of the materials and methods of gauging them. The porosity may often be a matter of importance as affecting the durability of mortar subjected to the action of disintegrating agencies. These tests are not employed for the reception of material. The porosity depends to a much greater extent upon the quantity of water used in gauging, and the degree of compression used in forming the briquette, than upon the character of the cement or sand employed, although probably the fineness of these materials has some influence. The test for porosity consists in determining the ratio of the volume of voids to the total volume of the mortar. The difficulty of determining when the voids are completely filled with liquid, or when the mortar is quite dry, makes the process a somewhat uncertain one, and requires that a definite procedure be followed in order to arrive at concordant and comparable results. 2l6 HYDRAULIC CEMENT. The method usually followed is to measure the total apparent volume and the volume of material: the volume of' voids is then the difference of the two; this divided by total volume is the percentage of porosity. To measure the total apparent volume, the simplest method is to make the block of such form that it may be directly measured. When this method is not employed the total volume may be obtained by weighing the block in a saturated condition in water and in the air; the difference between the two weights is the weight of water displaced, from which the volume may be found. In order to obtain the same state of saturation the weight in water should be taken immediately before that in air. Grease is also some- times applied to the surface of the block to prevent change during the weighing. To obtain the volume of solid material in the block, the difference between the weight of the block when dry, in air, and of the saturated block in water is obtained. This difference is the weight of water dis- placed by solid material. To secure good results the entire dryness in the first instance and the complete saturation in the second is essential. The block may be placed in warm dry air for a period sufficient to permit the weight to become constant. For this pur- pose it is necessary that the temperature be the same in all cases, as the amount of hygrometric water given off depends upon the temperature of drying; 100 to I IO Fahr. has been sometimes employed, and is recommended by the French Commission on Standard Tests. After the dry weight is obtained, considerable difficulty may be experienced in getting complete SPECIAL TESTS. 2 1/ saturation. If the block be simply immersed, air will be retained in the voids, and a long period required to obtain a constant weight. Boiling is sometimes resorted to, but has the disadvantage of perhaps caus- ing change of volume in the mortar. The best method of expediting the test is to exhaust the air by placing the specimen in water under the receiver of an air- pump. Prof. Tetmajer recommends* that a temperature of 1 10 C. be used in drying, and that the volume of solid matter be obtained by weighing dry in air and satu- rated in paraffine, and determining the volume of paraffine displaced. The block is then put in water, and the total volume obtained by measuring directly the volume of water displaced. For this purpose an apparatus is employed which is very similar to the Schuman volumenometer, Fig. 5, but with a remov- able cover to the dish to admit the specimen. ART. 67. TESTS FOR PERMEABILITY. The permeability of mortar is quite distinct from its porosity, and the more porous is not necessarily the more permeable. Tests for permeability, like those for porosity, are interesting and important in the studies of the properties of mortar, but are not suitable for use in specifications for the reception of material. The common test for permeability is made by forcing water through a cake of mortar under pressure. This * Methoden und Resultate der Priifung der Hydraulischen Bindemittei (Zurich, 1893). 2l8 HYDRAULIC CEMENT. may be accomplished either by subjecting the mortar directly to the pressure of a considerable head of water, or by subjecting the block of mortar to a small pressure from a column of water above, while the air is exhausted, forming a partial vacuum below. The latter method has been usually preferred in Europe, although the former has been made standard in France. The apparatus shown in Fig. 32 has been frequently employed in these tests. It consists of a heavy cylin- drical glass jar (a) with a ground upper edge, upon which is accurately fitted the second cylinder (*/). The specimen (c), having a section of 20 square centi- meters and a thickness of I centimeter, is fastened and made water-tight in the cylinder (d) by means of a rubber packing-ring. A ground-glass stopper covers the cylinder (b) and carries the graduated tube (e), which has a capacity of 200 cubic centimeters and is graduated to 1/2 centimeter. The space under the specimen is connected with an air-pump and a mer- cury manometer, a stop-cock being placed in the tube connecting with the air-pump. In using the appa- ratus, after the specimen has been placed in the cylin- der and the cover clamped down, the air is exhausted from a, the stop-cock closed, and the graduated tube filled with water to the zero mark. The quantity of water percolating through the specimen may then be read from the scale for any desired unit of time. The arrangement shown in Fig. 33 is also used to some extent in Europe. It consists of a hollow cylindrical block, 1 10 millimeters in diameter and 200 millimeters high, of the mortar to be tested, into the top of which a glass tube is set with neat cement. In SPECIAL TESTS. 2I 9 making the test a rubber tube is connected with the glass tube and with a vessel placed at such an eleva- FIG. 32. FIG. 33 tion as will give the pressure desired, which varies according to the mortar to be tested. A cubical block, arranged as shown in Fig. 34, is 220 HYDRAULIC CEMENT. also frequently employed. For this purpose the standard block as used for compressive tests may be employed. This form is recommended by the French FIG. 34. Commission upon Methods of Testing Materials, as follows: " The permeability of cements and mortars may be expressed by the number of liters of water which traverse a cube with faces of 50 square centimeters during a given time. " The water is supplied by a glass tube 35 mil- limeters in diameter and no millimeters high, sealed vertically by the aid of neat cement upon the upper face of the block. The upper end of the tube is con- nected by a rubber tube with a reservoir raised to an elevation corresponding to the head desired. The heads adopted, according to the permeability of the mortar, are o. 10 meter, I meter, and 10 meters. " Before the test is made, the blocks should be im- mersed for 48 hours, with such precautions as are necessary to secure as complete saturation as possible. SPECIAL TESTS. 221 After beginning the experiment the block is kept completely immersed." " The rate of filtration is determined after 24 hours, 7 days, 28 days, 3 months." ' ' The determinations are made for three blocks, the mean results being taken for the two blocks most con- cordant." " The standard test for permeability is to be made upon standard plastic mortar 28 days old, kept under water. ' ' " For tests upon mortars of different ages and com- positions, it is recommended to employ i to 2 and I to 5 mortars made plastic, and 7 days, 28 days, and 3 months old." The permeability of mortars to gases is a matter upon which experiments are very desirable, but con- cerning which comparatively little is known. ART. 68. FROST TESTS. Tests are frequently made upon cement-mortars to determine the effect of freezing before setting or while the mortar is still comparatively fresh, for the purpose of investigating the safety of using the mortar in freez- ing weather, or the best method of so using it. These tests are usually made by exposing briquettes to freezing temperatures, either by taking advantage of natural low temperatures, or by using a freezing mixture, and comparing the results of tensile tests upon briquettes so treated with those kept under nor- mal conditions. Tests of this kind may be of much value in showing 222 HYDRAULIC CEMENT. the relative properties of various materials, and often give very interesting results. They are to be used with caution, however, in determining from them the probable effect of freezing upon work in which the mortar may be used. Frequently the results of briquette tests, and the action of the material in large masses in construction are not concordant. Injury to work may perhaps result not only from the injurious effect of frost upon the strength of the material, but also from expansive action upon the mass of mortar, after setting, while still too weak to offer effectual resistance to distortion. ART. 69. TEST FOR YIELD OF MORTAR. In some laboratories it is the custom to make tests of the yield of mortar obtained from given weights of the materials employed. This may sometimes be of importance as affecting the economy of use of various materials, while a study of the differences obtained with different material is interesting. In conducting such tests it is evidently necessary to adopt a standard method of gauging the ingredients, the volume of the resulting product being much affected by variations in manipulation. The method employed is usually to measure directly the volume of paste obtained by mixing to standard consistency a unit weight of neat cement or of cement and sand in proper proportions. For this purpose the paste is put in a graduated glass cylinder, care being taken to eliminate all of the air-bubbles. Sometimes the test is made by making blocks of the SPECIAL TESTS. 22$ paste, which are allowed to set, their volumes being subsequently obtained by greasing their surfaces and taking the difference of weight in air and water. ART. 70. TESTS OF SAND. Tests of the sand to be used in> mortar are of much value in determining the relative value of different sands, as well as in studying the effect of variations in the nature, form, or size of grains of which it may be composed. Complete tests should include an examination of the nature of the sand, its fineness as shown by the amount retained by various sieves; the form of grain, which may be examined under a glass; its specific gravity or the weight of a unit volume; its minera- logical character. Tests of the tensile strength of mortar made from the sand to be tested as compared with similar tests made upon standard sand are of most importance as indicating the value of the sand to combine with cement in forming mortar. CHAPTER IX. CEMENT-MORTAR AND CONCRETE. ART. 71. SAND FOR MORTAR. As hydraulic cement is commonly mixed with cer- tain proportions of sand, when used in construction, the nature and quantity of sand used, and the method of manipulating the materials in forming the mortar, have nearly as important an effect upon the final strength of the work as the quality of cement itself. In testing cement, a standard sand is usually em- ployed. This sand may be obtained quite uniform in quality. In the execution of work, however, local sand must generally be employed ; this varies widely in character, and should always be carefully considered upon any important work, where the development of strength and lasting qualities in the mortar is of im- portance. The importance of the quality of sand in mortar is not commonly appreciated, and but little attention is usually given to securing good sand even when the cement is subjected to rigid requirements. Mr. Newman, in his book on concrete, in speaking of the materials used in concrete, says: " Considering the very varied character of sand and gravel, it seems that more attention should be given to the particu- 224 CEMENT-MORTAR AND CONCRETE. 22$ larization of the sand and gravel, remembering the locality of the work in each case, and the geological features of the district from which, for reasons of economy, the sand or gravel must be obtained. " The value of it from an engineering point of view may be very different, even in a small area; and to be most particular as to the character and quality of Portland cement, and apparently regardless of that of the sand and gravel, although the latter form 85$ to 93$ of the volume of concrete at the time of mixing, is hardly capable of vindication, especially as Portland- cement concrete should be a monolithic mass, and the effect of sand is to retard induration and decrease strength." The chemical nature of the sand does not appear to have any important bearing upon its usefulness in mortar. Silicious sand is sometimes thought to exer- cise a slow puzzolanic action, and perhaps aid some- what in the final hardening of the mortar. It is usually the best sand for the purpose. Calcareous sands are good, if not friable or composed of soft par- ticles. Argillaceous sand is usually less desirable, and has been found in some instances apparently to cause ultimate disintegration in sea-water,* although a small admixture of clay in the sand may not be objection- able, and has been shown in some instances not to decrease the strength when present to an extent not exceeding 10$ of the sand. A sand for use in mortar should be clean, and as free from loam, mud, or organic matter as possible. *Annales des Fonts et Chaussees, 1890, vol. II. p. 277. 226 HYDRAULIC CEMENT. In general, the presence of any foreign matter is to be avoided. The sand should also be as sharp as possible ; if it be composed of angular grains, it will compact much closer and make a stronger mortar when used with the same proportion of cement than if it be composed of rounded grains. Coarse sand is usually preferable to that which is very fine, provided it be fine enough to give a smooth mortar, as it gives better strength. The coarse sand presents less surface to be coated with cement, and the interstices are more easily filled. Fine cand requires more water in mixing in order to arrive at the same consistency, and thus gives usually a more porous mortar. Fine sand may, however, be desirable when an impervious mortar is the object. The use of a mixture of grains of various sizes is usually desirable, as giving less voids to be filled by the cement ; and it is frequently found that when the cement is not in considerable excess the strength obtained by such a mixture is much greater than is given by either the large or small grains alone. This is doubtless due to the voids in the sand being more completely filled by the cement. Large grains of uniform size seem desirable where a meagre mortar is to be employed, the quantity of cement being insuffi- cient to fill the voids and only used to coat the grains and cement them together. Fine sand is objectionable in mortar exposed to the action of sea-water on account of the increased poros- ity. In using quick-setting cement the drvness of thp CEMENT-MORTAR AND CONCRETE. 22? sand is a matter of importance; if the sand be damp, when the mixture of sand and cement is made, suffi- cient moisture may be given off to induce hydration previous to the addition of the water. With slow cements this is of less consequence. M.- Candlot found * cement left in contact with sand which was slightly damp, not sufficiently so to cause the cement to set, was greatly modified in action, probably through the hydration of the aluminate of lime. Cement left ten minutes in contact with sand con- taining 3$ of moisture, and then sifted out, had its time of setting increased from a few minutes to several hours. When the sand was very wet, the action was more serious and a loss of strength resulted. ART. 72. PROPORTIONING MORTAR. The relative proportions of the ingredients to be used in mortar are usually stated as a ratio of parts of cement to sand, and the quantities are determined either by weight or volume. Cement should always be measured by weight on account of the variation in volume caused in packing, and the difficulty of meas- uring by volume always in the same way, while the sand may conveniently be measured by volume. The proportions of sand and cement to be used in any instance depends upon the nature of the work ind the necessity for the development of strength or imperviousness in the mortar. The volume of in- * Ciment et Chaux hydrauliques (Paris, 1891). 228 HYDRAULIC CEMENT. terstices to be filled varies with the forms and sizes of the grains of sand, and the quantity of cement neces- sary to reach the same strength with different sands varies considerably. The volume of a given weight of sand is also greater when damp than when dry, and the same proportion to a given volume of sand gives a richer mortar when the sand is measured in a damp condition than when measured dry. The proportions most commonly used in ordinary work are, for natural cements, one part cement to one or sometimes two parts sand, and for Portland cement one part cement to three parts sand. If the propor- tions for the mixture were regulated by the value of the sand the interests of economy might frequently require changes in proportions, and would usually demand the use of the best sand obtainable. Good sand in a I to 3 mixture frequently gives greater strength than a poorer one mixed I to 2, and either mortar may give equally good results in practice. The cement in mortar must, for the best results, both coat the grains of sand so as to cause them to adhere to each other, and fill the voids between them. Mortar to be exposed to the action of water, particu- larly sea-water, should always contain a surplus of cement over what is necessary to fill the voids in the sand. M. Candlot gives as a minimum for mortar to be used in sea-water a proportion of 600 kilogrammes of cement to a cubic metre of sand, which is to be in- creased when the cement is not finely ground. Fine sand is to be avoided, if possible; when used, the proportion of cement to be increased. CEMENT-MORTAR AND CONCRETE. 22Q The complete filling of the voids in the sand so as to exclude the water from the interior of the mass and prevent the action of the magnesian salts upon the cement is in such work a matter of first importance. When the mortar is to be subjected to the action of fresh water its permeability may not be a matter of so great consequence, and in many instances less rich mortar may frequently be used to advantage, provided sufficient strength be obtained for the given purpose. The carbonizing of the lime forms a protection to the mass of mortar, which is not subject, as in sea-water, to the action of the magnesian salts. ART. 73. GAUGING MORTAR. In mixing cement-mortar, the cement and sand are first thoroughly mixed dry, the water then added, and the whole worked to a uniformly plastic condition. The value of the mortar will depend upon the thoroughness of the operation; the cement must be uniformly distributed through the sand during the dry mixing, while thoroughly working the mass after the addition of the water will greatly increase its strength. In mixing by hand, by the ordinary method, a plat- form or box is used ; the sand and cement are placed upon the platform in layers, with a layer of sand at bottom, and then turned and mixed with shovels until properly distributed through the mass. The material is then formed into a ring, or a mound with a crater at the centre, and all the water necessary added at once, after which the material is thrown up from the 230 HYDRAULIC CEMENT. sides until the water is all taken up, and then worked into a plastic condition. In order to secure proper manipulation of the ma- terials on the part of the workmen, it is quite common to require that the whole mass shall be turned over a certain number of times with the shovel, both dry and wet. The mixing should be quickly and energetically done, only such quantity being mixed at once as can be used before the initial set of the mortar takes place. The cement should not be left in contact with the sand for any considerable time before being used, or a considerable quantity should not be mixed dry and left to stand until wanted, as the moisture commonly in the sand will to some extent act upon the cement. Upon large works mechanical mixers are frequently employed, with the advantage of greatly lessening the labor of manipulating the material, and also of insur- ing thorough mixing. The quantity of water to be used in gauging mortar can be determined only by experiment in each instance. It depends upon the nature of the cement and sand, and upon the proportion of sand to cement. The water may be considered as made up of two parts that necessary to gauge the neat cement to a paste, and that required to wet the surfaces of the sand. The first varies directly with the quantity of cement, the second with that of the sand. Fine sand requires more water than coarse sand to reach the same con- sistency, and mortar of fine sand should be made a little more wet than when of coarse sand, to give the CEMENT-MORTAR AND CONCRETE. 23! best results in practice. The quantity of water also varies with the dryness of the sand and its porosity. The amount of water to be used in mixing mortar for ordinary masonry is such that the mortar when properly mixed shall have a stiff plastic condition. It should not be a soft, semi-fluid mass. The proper consistency is described by M. Candlot as such that if a ball of mortar be formed in the hand and allowed to fall through a small height, it should neither lose its form nor crack; the ball should not be wet enough to stick to the hand. The best results are usually obtained by mixing with as little water as will admit of proper manipulation in the work, and wetting the surfaces with which it is to be in contact. In all cases the proper quantity of water should first be determined by experiment, and afterward, in preparing the mortar for use in work, the required quantity should each time be added by measurement. The addition of water little by little, or from a hose, should never be allowed. ART. 74. PREPARATION OF CONCRETE. Concrete is any mixture of mortar with coarse material, usually gravel or broken stone, the office of the mortar being to bind together the pieces of the aggregate and fill the spaces between them. In engineering work the mortar for concrete is commonly formed from hydraulic cement, the term beton being also frequently used to designate hydraulic concrete. In preparing concrete by hand the mortar is mixed in the usual manner; then the stone is spread over the 232 HYDRAULIC CEMENT. top of the layer of mortar and thoroughly mixed with it by turning with shovels. The stone should be sprinkled before being mixed with the mortar, suffi- ciently to wet its surfaces and prevent the absorption of the water from the mortar, thus promoting the adherence of the mortar to the aggregate. Mortar for concrete should never, as is frequently done, be reduced to a fluid state; not only is the resulting strength of the mortar reduced by so doing, but it cannot be properly mixed with the aggregate to form a homogeneous mass, as the cement washes out of the mixture. The mortar should, however, be suf- ficiently soft to mix readily with the aggregate to a cohesive mass, which may be placed and compacted without difficulty in the work. The consistency of concrete must in each instance depend upon its nature and method of use; the greatest strength may usually be attained by mixing somewhat dry and heavily ramming, the mortar being of such consistency that the concrete becomes somewhat jelly-like, water being brought to the surface in ramming. An extreme either of dryness or wetness may be injurious. Mechanical mixers are frequently employed for pre- paring concrete, and are very useful in saving labor where considerable quantities are used. There are a number of forms which have proven effective in use, but it seems unnecessary to enter into a discussion of them here. The aggregate used for concrete should be as hard and durable as possible, and that of angular form is preferable to rounded. Angular forms give a greater surface for the adherence of the mortar in proportion CEMENT-MORTAR AND CONCRETE. 233 to volume while leaving a less volume of interstices to be filled with mortar. The materials should be uni- form in quality. Where gravel is used which varies in quality, it should be blended by mixing in order to obtain uniform strength in the concrete. Porous aggregates are to be avoided, as they are likely to absorb the cement- When the materials are absorb- ent, they should be saturated in sprinkling before using, in order to avoid withdrawing water from the mortar before setting takes place. The quantity of sand used in concrete should be such as is necessary to fill the voids in the aggregate, while the quantity of cement depends upon the strength necessary in the work under consideration. When the concrete is required to be water-tight, the amount of cement paste must be sufficient to fill the interstices in the mixture of stone and sand. The quantity of sand necessary to fill the interstices in the stone may be determined by filling a measure with stone as closely as possible, and then measuring the quantity of water which can be poured into the meas- ure; this gives the volume of sand required. If the proper quantity of damp sand be added to the stone in the measure by shaking it down so as to fill the voids, the volume of water which can then be put in the measure is the volume of cement paste necessary to fill the voids in the aggregate. The strength of concrete usually varies nearly in proportion to the amount of cement used in forming it. When a strong concrete is desired, it should be obtained by increasing the richness of the mortar in cement, not by increasing the proportion of mortar to 234 HYDRAULIC CEMENT. large material above the point at which the sand fills the interstices in that material. If the proportion of sand be less than this, the resulting concrete will be porous and not thoroughly solidified ; if it be greater, the excess of sand may be an element of weakness in the concrete. In the use of concrete in considerable masses the main body of the work is sometimes formed of very weak concrete, with a facing of stronger water-tight concrete to protect it. This weak concrete is fre- quently formed by omitting the sand altogether and simply coating the stone lightly with neat cement, causing the stones to adhere to each other, thus form- ing a mass sufficiently firm for foundations in many locations when protected by a covering of richer con- crete. The voids in a mass of ordinary broken stone vary from about 4/10 to 5/io of the volume, depending upon its uniformity in size. Where there is consid- erable variation in size, the voids may be somewhat less. When the interstices are to be filled, it is desir- able that the aggregate contain material of various sizes, to reduce the volume of interstices. For this reason small gravel is sometimes mixed with broken stone in the preparation of concrete. The proportions in common use for concrete of Portland cement vary from I part cement, 2 parts sand, and 5 parts broken stone to I part cement, 4 parts sand, and 8 or 10 parts broken stone or gravel. Usually the mortar is made richer when natural cement is used. The proportions, of course, vary with the character of the materials to be used as well as that of the /ork to be done, and can only be properly CEMENT-MORTAR AND CONCRETE. 235 determined by the exercise of good judgment, in the light of experience. ART. 75. YIELD OF MORTAR AND CONCRETE. The volume of mortar formed by mixing given quantities of cement and sand depends upon the den- sities of the materials and the volume of interstices in the sand. It is affected also by the method of pre- paring the mortar, the uniformity of the mixing, and the degree of compactness given. The net volume of materials entering into the com- position of mortar or concrete is readily found from their weights and densities, but it represents only approximately the resulting volume. An accurate knowledge of the yield of any particular mixture is to be obtained only by experiment upon the materials to be employed. Portland cement is usually sold in barrels containing about 375 Ibs. Natural cements are lighter, and are put up in barrels of 260 to 320 Ibs. Barrels of Rosen- dale cements usually contain 300 Ibs. The amount of neat cement paste made by a given weight of cement powder varies with the specific gravity of the cement and the amount of water neces- sary in gauging. The lighter cements require more water and yield less paste for a given volume of cement than the heavier ones. To form a cubic foot of plastic paste requires usually from 75 to 90 Ibs. of natural-cemerfit powder; about 80 to 85 Ibs. of Rosen- dale cement being required, while about 95 to loolbs. of Portland cement are necessary. 236 HYDRAULIC CEMENT. In mixing sand-mortar, where the cement and sand are proportioned by volume measured loose, the quantities required to form a cubic yard of mortar are approximately as follows: Uatural Cement. Portland Cement. Sand. Pounds. Pounds. Cu. Yd. to i mortar. . . . 1050 to 1250 1350 to 1530 .65 to .70 to 2 . . . 640 to 720 810 to 920 .80 to .85 to 3 " ... 500 to 575 620 to 690 93 to .96 to 4 " 400 to 460 500 to 575 I.OO to 5 " ... 32010 375 400 to 460 I.OO For concrete, when the aggregate is broken stone of uniform size, it is necessary, in order to fill the in- terstices with mortar, that the volume of mortar be 50$ to 60% that of the aggregate. For such concrete a mixture of about .9 cubic yard of broken stone with .50 to .55 cubic yard of the mortar as given above yields about one cubic yard of concrete. This gives the proportions sometimes employed for strong con- crete: I part cement, 2 parts sand, and 4 parts broken stone; or I part cement, 3 parts sand, and 5 parts broken stone. Where the stone is more irregular in size, or if gravel of smaller size be added, a smaller proportion of mortar may give good results. Thus, .9 cubic yard of broken stone with .4 cubic yard of gravel and .3 cubic yard of mortar has been found to yield i cubic yard of good concrete. This, using I to 2 mortar, gives the proportion I part cement, 2 parts sand, 3 parts gravel, and 7 parts broken stone. When the amount of mortar used is less than that indicated above, and the interstices in the aggregate CEMENT-MORTAR AND CONCRETE. 237 are not filled, the yield of concrete is about equal to the volume of aggregate employed. ART. 76. MIXTURES OF LIME AND CEMENT. Slaked lime is sometimes mixed with hydraulic cement for the purpose of decreasing the cost of con- struction. Experiments seem to indicate that a very considerable percentage of lime may frequently be added without material loss of strength in the mortar. With Portland cement the addition of lime weakens the mortar somewhat, the decrease in strength aug- menting rapidly as the proportion of lime increases. With some American natural cements it has been found that a certain amount (sometimes 30$ to 40^) of lime may be added without sensibly decreasing the strength of the mortar or impairing its hydraulic prop- erties. This may be due to puzzolanic action on the part of the cement, which is of high hydraulic index. When mortar is not to be used under water, and only moderate strength is necessary, it may often be economical to form the mortar by the admixture of lime, better results being obtained than by using a higher proportion of sand. The mortar thus formed is less porous than that made with a larger proportion of sand, and it is also more plastic and easier to work. In making the mixture the lime is ordinarily slaked in the usual manner and used in the form of paste, although it may be slaked to powder and mixed dry with the cement. By the first method the thorough slaking of the lime is insured. The admixture of lime causes the cement to become 238 HYDRAULIC CEMENT. slower-setting, the quick-setting cement being affected more strongly than the less active ones. In France a small proportion of Portland cement is sometimes added to hydraulic lime for the purpose of accelerating the setting and increasing the strength of the lime. Mixtures of natural and Portland cement have fre- quently been used in the United States for the pur- pose of modifying the action of the quick-setting material or of cheapening construction. They seem generally to give results compounded of those which would be obtained by using them singly. The results of all these mixtures will be found to vary with the particular cement employed, and the effect can only be known by trial in each instance. In all cases, to get good results, the mixtures must be very intimate. ART. 77. THE FREEZING OF MORTAR. Mortar of good Portland, or of many kinds of nat- ural cement, is not injured by freezing, when frozen before it is set. The cement sets with extreme slow- ness, if at all, while frozen, but after thawing it sets and hardens properly. Mortar frozen for short periods a few days does not set while frozen, but the experiments of Mr. Cecil B. Smith at McGill Uni- versity seem to show that if kept frozen for a sufficient period it may finally set while frozen. The hardening of cement which has been frozen is much more slow than that unfrozen, but it may ulti- mately gain the same strength. CEMENT-MORTAR AND CONCRETE. 239 Masonry constructed during freezing weather is fre- quently injured by freezing, notwithstanding the fact that the cement itself shows no loss of strength due to freezing. The effect of frost coming upon the work before it is fully hardened is frequently to distort or cause unequal settlement in it, and sometimes repeated freezing and thawing gradually causes the mortar to be thrown out of place or perhaps to become cracked and disintegrated on the outside. The construction of cement masonry during freezing weather is there- fore usually more or less hazardous, unless some means be adopted of preventing the freezing action. Many instances may, however, be cited where ex- treme cold has not injured work constructed, without such precaution, with Portland-cement mortar, and it is claimed by many engineers that Portland may be used with impunity in freezing weather, but usually it is not placed in work while a freezing temperature prevails. It is commonly agreed that most natural cements should not be used when a very low tempera- ture is likely to reach the work in advance of it having attained good strength, and instances are numerous of work having been injured by changing temperature of winter weather, although it may not have frozen for considerable time after setting. Salt is quite commonly used in cold weather to prevent the freezing of mortar while it is soft. A strong solution, frequently a saturated one, is em- ployed. The salt, by preventing the freezing of the water, prevents any distorting or disrupting action upon the work due to the change in volume of the mortar. The use of salt considerably decreases the 240 HYDRAULIC CEMENT. activity of the cement, and mortar may stand in a soft condition at freezing temperatures and finally set when the temperature becomes sufficient to induce action. The loss in early strength of cement-mortar which has been mixed with salt water on exposure to low temperature before setting is usually greater than that of mortar without the salt and exposed at the same temperature. The effect of salt upon the strength of various kinds of cement is quite different. In nearly all, the strength of mortar kept in air is increased by its use. When the mortar is kept under water, most cements have an access of early strength from the use of salt, which is lost later, the final strength being somewhat reduced. This is true of most Portland cements. Some natural cements suffer a material loss of strength when mixed with salt water, while others are entirely ruined by a low temperature with or without the use of salt. Care should always be taken to determine the action of salt and cold upon the particular cement before using it in this manner. It is advisable in using salt to protect the mortar from contact with water immediately after setting, as sometimes salt mortar which has been exposed to low temperature may lose its cohesion if submerged soon after setting. Soda has sometimes been employed to prevent the freezing of mortar, but its use has not become exten- sive, and has usually proven unsatisfactory. Hot water should not be used in mixing mortar in freezing weather. It not only decreases the strength of the mortar, but renders it more liable to injury from CEMENT-MORTAR AND CONCRETE. 24! frost. Heating the stones or bricks in the construc- tion of masonry in freezing weather may be beneficial, as serving to accelerate the setting and keep the mor- tar from freezing while soft. The injury done to mortar by low temperatures is probably not usually due to freezing before setting, but to alternate thawing and freezing while work is still fresh, before hardening is sufficiently advanced to render the mortar capable of adequately resisting the expansive forces. The effect of frost upon mortar which has set is similar to that upon stone or brick, and is due to the increase in volume of water freezing in its pores. Its effect therefore depends both upon the porosity of the mortar and upon the strength it possesses to resist disruption. The more rapid acqui- sition of strength by Portland cements may give them the advantage they possess in this regard. Prof. Le Chatelier, from his experiments upon the matter, concludes as follows: " This disintegration, like that of frozen stone, is more easily accomplished when the mortar offers small mechanical resistance, when the total volume of voids is large, and when the dimensions of each separate void are small. When the voids are sufficiently large, the ice breaks with a pres- sure less than that which will rupture the mortar. For this reason mortars of large sand are less affected, the voids being larger and less numerous." ART. 78. POROSITY AND PERMEABILITY OF MORTAR. The porosity of cement-mortar depends rather upon the manipulation of the materials in gauging than 242 HYDRAULIC CEMENT. upon the quality of the cement. When the quantity of cement is insufficient to fill the voids in the sand, spaces are left which permit the absorption of water without increasing the volume of the mortar. In gauging mortar air-bubbles attach themselves to the wet sand, the number of which is greater as the mortar is mixed more wet. Working the mortar tends to eliminate them. Voids in the mortar are also caused by the evaporation of surplus water used in mixing. Porosity is greater as the quantity of water used in gauging is increased and as the sand used is finer. The permeability of cement-mortars varies with the quality of the cement and the circumstances of its use. Mortar of neat Portland cement may be made prac- tically impermeable under a considerable head of water; that composed of cement and sand seems always more or less permeable, but when properly proportioned and mixed eventually permits very little water to pass. The permeability of mortar decreases rapidly with its age: for the first few days or weeks after mixing, water passes quite freely through it, but as the har- dening process approaches completion its power of resistance is in this particular greatly augmented. If blocks of mortar be submitted to the continuous filtration of water, the permeability diminishes very rapidly, and after a few months all mortars, except those of very coarse sand and feeble proportion of cement, become practically impermeable. Both the porosity and permeability are less for mortar rich in cement than for that in which the pro- CEMENT-MORTAR AND CONCRETE. 243 portion of cement is small. Mortar mixed dry is penetrated more readily than that mixed to a plastic or semi-wet condition. With the lapse of time, how- ever, the mortar mixed dry, if constantly exposed to water, approaches the others in resistance to permea- tion. The thoroughness of mixing and degree of compacting employed are more important factors than the absolute quantity of water used in mixing Fine sand, according to the experiments of M. Alex- andre, renders the mortar more porous and less per- meable than coarse sand. When the sand is of varying sizes both the porosity and permeability may be low. In any case, to attain a reasonable resistance to penetration, it is necessary that the interstices in the sand be entirely filled with cement. Cleanliness of the sand, its freedom from all foreign material, is of first importance in the preparation of impermeable mortar. Masonry of ordinary brick or stone can only be made impervious by the application of a coating of some kind to its face. A plastering of neat cement or rich mortar may sometimes be used for this purpose, and coatings of asphalt or coal-tar have sometimes been successfully employed. In concrete work where imperviousness is essential, it may be advisable, as with masonry, to coat the face of the concrete. In order that concrete may be reasonably water-tight, it is necessary that the quantity of cement-mortar used in preparing it be sufficient to fill the voids in the large material employed, as well as that the voids in the sand be completely filled with cement paste in making the mortar. HYDRAULIC CEMENT. ART. 79. EXPANSION AND CONTRACTION OF MORTAR. In the use of large masses of masonry or concrete the change that is liable to occur in the volume of mortar may frequently become of importance, and it may be necessary to make provision by which changes in dimension an take place without in : ury to the work. The coefficient of expansion for neat cement under the action of heat is, as previously stated, about the same as for iron, although it may vary in individual instances. For mortars containing sand, the coefficient is less than for neat cement. Cements differ considerably in their behavior during the continuance of the hardening process, as to the change that takes place in the volume of the mortar. Unsound cement is apt to swell and become distorted at the commencement of the process of disintegration, and of course any considerable change of this nature indicates the probable destruction of the mortar. Perfectly sound cement, although not altered in form, is usually changed somewhat in dimensions during hardening: if the mortar be conserved in dry air, a slight shrinkage takes place ; if under water, the mor- tar swells a little. Prof. Swain, in a series of experiments at the Massachusetts Institute of Technology for a committee of the American Society of Civil Engineers, found that for small blocks of mortar the change was the same in all directions; that for neat cements the CEMENT-MORTAR AND CONCRETE. 245 linear contraction in air varied from 0. 14$ to 0.32$ for the first twelve weeks after mixing, and the linear ex- pansion in water varied from 0.04$ to 0.25$. When sand was used the change was less, giving a contrac- tion in air from o.o8# to 0.17$, and an expansion in water of from 0.00$ to 0.08$. The rapidity of the change varies somewhat with the activity of the cement; the conclusion being that a quick-setting cement changes more in volume than a slow-setting one. Further experiment is desirable, that the action of the various classes of cement may be better under- stood. ART. 80. EFFECT OF RETEMPERING MORTAR. Masons frequently mix water in considerable quan- tities, and, if the mass becomes stiffened before being used by the setting of the cement, add more water and work again to a soft or plastic condition. After the second tempering the cement is much less active than at first, and remains a longer time in a workable condition. This practice is not usually approved by engineers, and is not permitted in good engineering construction, although there is some dispute as to its injurious effect. M. Alexandre, from an extensive series of experiments,* concludes that no injury is usually done to mortar by retempering, provided sufficient water be added to make the mortar plastic at the second work- * Annales des Ponts et Chaussdes, 1888, vol. i. p. 375. 246 HYDRAULIC CEMENT. ing. The hardening of mortar so treated is very slow at first, but it may subsequently (the tests extend over three years) gain as much strength as when gauged immediately upon mixing. It is also frequently claimed that the adhesive prop- erties of mortar are improved by giving it the " second set." The common practice of masons who set fire- place tiling and similar work is based upon this idea. Further experiment to determine this point would be interesting. The results of experiments other than those already quoted have seemed to show that in some instances injury is done to mortar by retempering, some cements even refusing to set the second time. Until more is known of the action of the material when subjected to this treatment, it seems advisable to mix only such quantity at once as may be used before the initial set of the cement, and to reject any material that may have set before being placed in the work. APPENDIX. SPECIFICATIONS FOR THE RECEPTION OF CEMENT. IN the discussion which has been given of the various tests applied to cement, the requirements of specifications have been considered, but it is thought desirable to append a few actual specifications to show the requirements employed in practice. Some engineers who use considerable quantities of cement of a few brands, or of a single brand, employ no specifications, but depend upon the reliability of the brand, or perhaps upon occasional examinations to show that the material is up to standard. This is frequently the practice upon ordinary railroad work, and in some instances, ivhere the use is continuous and private contracts may be made for the material, is quite satisfactory. Some specifications (quite frequently employed) make no mention of any specific requirements, but call for cement which shall be satisfactory to the en- gineer in charge of the work, and lodge in him full power to accept or reject the material. The effect of such a specification depends upon the circumstances 248 APPENDIX. under which it is employed. When used as a speci- fication for work, in open competition, it is very apt to prove unsatisfactory; but when it represents the established practice of an office dealing with particular brands of cement, it may not be objectionable. There seems, generally, to be no good reason for stating specifications in an indefinite manner. The conditions to be imposed may as easily be plainly stated, and thus leave no doubt as to the requirements. In some instances, however, engineers consider a specification of this kind advantageous where the cement to be used is limited to a few accepted brands (particularly with natural cements), the tests to be im- posed varying with the brand and being only for the purpose of showing that the various lots of cement are normal in action. In such a case the tests to be im- posed are usually an understood, though not an ex- pressed, part of the specifications. The specifications which follow, A, B, C, D, E, are thought to represent fairly well the range of require- ments in the best American practice Specifications "A" represent about the common requirements for ordinary masonry, but are more complete and explicit in statement than is common. The tensile test requirements vary considerably in various specifications, being frequently much higher where permanent laboratories are maintained, as in specifications " D " and " E." Specifications " B " have much the same require- ments, but apply the Faija warm-water test for sound- ness. Specifications " C " are intended for cement to be SPECIFICATIONS FOR THE RECEPTION OF CEMENT. 249 used for heavy work in sea-water. Soundness is assured by applying the hot test, combined with a limit of sulphuric acid and magnesia. The requirements are high, but may be met without difficulty by most of the leading brands of Portland cement, and are con- sidered necessary because of the importance of the work and its exposure to the action of sea-water. It is quite common to place considerable reliance upon the brand in selecting cement, and in many specifications the approval of the brand is an essential condition. This, seemingly, places an arbitrary power in the hands of the engineer, but may often increase the likelihood of getting good material, while perhaps saving the trouble and loss of time incident to testing material obviously unfit for use. A, PENNSYLVANIA RAILROAD COMPANY. P., W. & B. R. R. N. C. Ry. W. J. & S. R. R. STANDARD SPECIFICATIONS FOR CEMENT. Cement will be tested by the Railroad Company, upon its receipt, in accordance with the following specifications, and such tests are to be final as deter- mining the matter of its acceptance or rejection. The cement for testing shall be selected by taking, from each of six well-distributed barrels in each car- load received, sufficient cement to make five to ten briquettes; these six portions, after being thrown together and thoroughly mixed, will be assumed to represent the average of the whole car-load. 250 APPENDIX; Fineness. Not more than io# of any cement shall fail to pass through a No. 50 sieve (2500 meshes per square inch), wire to be No. 35, Stubb's wire- gauge. Cracking. Neat cement, mixed to the consistency of stiff plastic mortar, and made in the shape of flat cakes, two or three inches in diameter and one-half inch thick, with thin edges, when hard enough shall be immersed in water for at least two days. If they crack along the edges or become contorted, the cement is unfit for use. Tensile Strength. The test for tensile strength shall be made with briquettes of standard form, as recommended by the American Society of Civil En- gineers, moulds for which will be furnished from the office of the Engineer Maintenance of Way. They must have an average tensile strength not less than that given in the table below: One Day. One Week. Four Weeks. American Natural Cement: Neat 7O qc 50 }O 60 American and Foreign Portland Cement: Neat . . Proportion of Water. The proportion of water used in making briquettes varies with the fineness, age, and other conditions of the cement and the tem- perature of the air, but is approximately as follows: SPECIFICATION^ FOR THE RECEPTION OF CEMENT. 2$ I Neat Cement. Portland, 2(5# to 30$. Natural, 20# to 30$. I part cement, I part sand, about 15$ total weight of cement and sand. i part cement, 2 parts sand, about 12$ total weight of cement and sand. Mixing. The cement and sand, in proper propor- tions, shall be mixed dry, and all the water specified added at one time, the mixing to be as rapid as possi- ble to secure a thorough mixture of the materials, and the mortar, when stiff and plastic, to be firmly pressed to make it solid in the moulds without ramming, and struck off level. Moulding. The moulds to rest directly on glass, slate, or other non-absorbent material. As soon as hard enough, briquettes are to be taken from moulds and kept covered with a damp cloth until immersed. In the one-day test briquettes shall remain on the slab for one hour after being removed from mould and twenty-three hours in water. In one week or more test, briquettes shall remain in air one day after being removed from moulds and balance of time in water. Briquettes are to be broken immediately after being taken from the water. Stress to be applied at a uni- form rate of four hundred pounds per minute, starting each time at zero. No record to be taken of briquettes breaking at other than fti^'sjiiallest section. Sand. The sand used in test shall be clean, sharp, and dry, and be such as shall pass a No. 20 sieve (400 meshes per square inch), wire to be No. 28, Stubb's wire-gauge, and to be caught in a No. 30 sieve (900 252 APPENDIX. meshes per square inch), wire to be No. 31, Stubb's wire-gauge. Water. Ordinary fresh, clean water, having a tem- perature between 60 and 70 Fahr., shall be used for the mixture and immersion of all samples. Proportions. The proportions of cement and sand and water shall in all cases be carefully determined by weight. In preparing briquettes for test, sufficient material is to be taken to make one briquette at a time, and enough of water being added to make a stiff plastic paste as above stated. The temperature of the testing-room not to be below 45 Fahr. By order of the General Manager. JOSEPH T. RICHARDS, Engineer Maintenance of Way, OFFICE OF THE ENGINEER OF MAINTENANCE OF WAY PHILADELPHIA, January 10, 1897. B. SPECIFICATIONS OF THE CANADIAN PACIFIC RAILWAY COMPANY. Mr. P. ALEX. PETERSON, Chief Engineer. (Extract from General Specifications for Rubble Masonry.) " 5. The cement used in making concrete or mor- tar shall be freshly ground Portland cement of approved brand, or such other cement as the Engineer may approve. It shall weigh not less than 1 10 Ibs. to the struck bushel; and not less than 90$ of it shall pass through a sieve containing 2500 meshes per square SPECIFICATIONS FOR THE RECEPTION OF CEMENT. 2$3 inch. The tensile strength of the neat cement after being kept in water at a temperature of about 60 Fahr. for seven days shall not be less than 350 Ibs. to the square inch. Also, when mixed with one part of cement to three parts of sand by measure, it shall stand 170 Ibs. to the square inch at the end of 28 days. " In preparing sand for this test, sand shall be re- jected which passes through a sieve made of No. 31 wire with 900 meshes to the square inch, or which will not pass through a sieve made of No. 28 wire with 400 meshes to the inch. " A pat made and submitted to moist heat and warm water at a temperature of about 100 Fahr. shall show no sign of blowing in 24 hours. " Cement shall be tested by the Engineer on deliv- ery ; and it shall be kept in a dry place, and in as good order as when delivered, until it is used." C. EXTRACTS FROM SPECIFICATIONS FOR FURNISH- ING PORTLAND CEMENT FOR WALLABOUT IMPROVEMENT, BROOKLYN, N. Y. W. E. BELKNAP, Engineer. 11 9. All the cement to be furnished under this con- tract must be of the class of such material known as high-grade ' Portland ' cement free from lumps, dry and finely ground, and unless as otherwise specified, must be of one or more of the following brands, known as ' DyckerhofT, ' ' Alsen's White Label,' and ' Stettiner Star Brand.' Cement of other brands 2$4 APPENDIX. may be furnished, provided the Contractor submits proof satisfactory to the Engineer that it has been used in making large masses of concrete, which have been exposed to the action of sea-water for at least two years previous to the date of this contract, and that such concrete now shows no signs of deterioration which might be imputed to defective qualities in the cement. " 10. All the cement shall be composed of lime, silica, and alumina in their proper forms and propor- tions, be as free as possible from all other substances, and contain no adulterant in injurious proportions. The ratio of the weight of silica and alumina to the weight of the lime in the cement shall not be less than 45/100. The cement shall not contain more than 3/ of magnesia nor more than i# of sulphuric acid. " ii. The cement shall not have a lower specific gravity than 3.10. " 12. All the cement shall be of a fineness so that 99^ by weight shall pass through a No. 50 sieve (2500 meshes per square inch) of No. 35 wire; 90$ shall pass through a No. 100 sieve (10,000 meshes per square inch) of No. 40 wire, and 70% shall pass through a No. 200 sieve (40,000 meshes per square inch) of No. 45 wire. All wire numbered as per ' Stubb's ' gauge. "13. The cement must not take its ' initial ' set in less than 30 minutes after mixing. It shall take its ' hard ' set in not less than 3 hours and in not more than 8 hours. " The cement will be said to have attained its ' initial ' and its ' hard ' set when it bears without indentation respectively a wire of 1/12 inch diameter SPECIFICATIONS FOR THE RECEPTION OF CEMENT. loaded to weigh 1/4 pound, and a wire of 1/24 inch diameter loaded to weigh I pound, it having been previously mixed neat with about 25$ of its weight of water, and worked for from one to three minutes into a stiff plastic paste. " 14. All the cement shall be capable of developing a tensile strength under various conditions as follows: Age. Tensile Strength in Pounds per Square Inch. Mixed neat with about 25$ 01 water by weight and worked to stiff plastic paste. 24 hours 7 days, 28 " , in water after hard set I in air, 6 in water, 70 I " " 27 " 150 400 600 Mixed with 3 parts sand by weight, and about I2#of combined weight 7 days, I in air, 6 in water, 70 150 of sand and cement of 28 " T " " 27 " 240 water to stiff plastic paste. ' To determine the tensile strength, 4 briquettes of the cement under each of the above conditions will be broken in a ' Riehle ' or Fairbanks or other testing- machine satisfactory to the engineer. " The sand to be used in making briquettes will be clean, dry crushed quartz, trap-rock or granite, pass- ing a No. 20 sieve of No. 28 wire, and caught on a No. 40 sieve of No. 31 wire ' Stubb's ' gauge. The briquettes will be of the form recommended by the American Society of Civil Engineers. "15. All cement must be sound in every respect, 256 APPENDIX. and show no indications of distortion, change of volume, or blowing when subjected in the form of pats to exposure in air and fresh and sea water of tempera- ture from 60 to 212, as follows: " The pats will be made of neat, unsifted cement, mixed with fresh water to the same consistency as before stated for briquettes, and will be about 3 inches in diameter, having a thickness at the centre of about 1/2 inch, tapering to about 1/8 inch at the edges. They will be moulded on plates of glass and kept thereon during examination. " (a) One or more of these pats will, when set ' hard,' be placed in fresh water of temperature between 60 and 70 for from I to 28 days. " (b) One or more of these pats will be allowed to set in moist air at a temperature of about 200 for about 3 hours. It will then be placed and kept in boiling water for a period of from 6 to 24 hours. " (c) One or more of these pats will be allowed t& set in moist air at a temperature of about 100 for 3 hours; it will then be placed and kept in water of temperature of 110 to 115 for a period of from 24 to 48 hours. " (d~) One or more of these pats may be subjected to any or all of the above indicated tests ( (a), (b), and (c) ), using sea-water instead of fresh water. " (e) One pat will be kept in the air for 28 days and its color observed, which shall be uniform throughout, of a bluish gray, and free from yellow blotches. " A failure to pass test (b) will not necessarily cause the rejection of the cement, provided it passes the other tests for soundness as noted in (a), (c), (d), SPECIFICATIONS FOR THE RECEPTION OF CEMENT. 257 and (