Practical cement testing

Preface

Although during the past year or two, the additions to the literature on cement and concrete have been especially notable, it nevertheless has appeared to the author that a complete description of the methods of handling practical tests of cement has not yet been brought before the public. The methods usually given in text books assume too much knowledge for the use of the student or beginner, while to the practical operator the directions are too general to be of much value. The excellent standard methods of testing recently issued by committees of several technical societies have already done much to promote uniformity in testing, but are of little real assistance either to the novice or the expert in enabling him to increase the accuracy or to simplify the routine of his work. This volume therefore has been designed primarily for the use of the student, the novice, and the practical operator in conducting actual routine tests of cement to determine its suitability for purposes of construction, but it is hoped that both the expert and the engineer who directs this work may also find something of interest in its pages.

The general scope of the book covers a description of the properties of cement, the objects of the various tests, the methods of conducting them, the common influences and errors that are most likely to affect the determinations, and the practical interpretation of the results which are finally obtained. No attempt has been made to consider the practical use of cement and concrete except in so far as the conditions of actual work regulate the use of the various tests, while the data given are also applicable more to the conduct of tests than to the final use of the material. In other words, the scope of the book is intended to cover only the methods and the application of the tests of cement commonly employed in routine work, and not to consider theoretical properties, investigations of a research character, nor the use of cement.

Chapters I., II. and III. are of an introductory nature, and are included only for the information of the student, anil for the jy PREFACE.

logical development of the subject. The constitution of cement is considered very briefly, while the chapter on manufacture is purely descriptive and makes no attempt at being technical.

The body of the book, Chapters IV. to XII., is devoted to the conduct of routine tests of cement. The chapter on Chemical Analysis was prepared in collaboration with Mr. Charles S. Reeve, Chemist of the Philadelphia Laboratories.

Chapter XIII. deals with simple tests by which the character of a cement may be ascertained with practically no apparatus, and which might be of considerable service to the small consumer or to the expert when it is impracticable to obtain apparatus to make tests in the orthodox manner.

The practical operation of a cement laboratory is considered in Chapter XIV., which also describes the organization, the labor required, and the cost of testing.

To avoid an endless number of qualified statements, in describing the different properties and tests, the body of the book is devoted entirely to Portland cement, while in Chapter XV. is given a description of the properties and tests of natural, improved, slag, sand, and other varieties of cement.

Chapter XV^I. is devoted to the subject of cement specifications, while the different Appendices give the standard specifications and methods of testing issued by various scientific bodies.

The methods of testing used in the author's laboratory are given at some length throughout the course of the book, and it is hoped that the frequent allusions to them will be pardoned on account of the long and varied experience of the laboratory, and on account of the large amount of testing performed, which has required both development of system and simplification of method.

The author wishes to express his appreciation of the kindness of Mr. George S. Webster, Chief Engineer of the City of Philadelphia, in permitting him to describe the methods employed, and to use data obtained in the Philadelphia Laboratories, and also to acknowledge the interest that has enabled the Laboratories to attain their present stage of development.

Acknowledgements for the use of cuts are also due to The Allis-Chalmers Co., J. W. Bramwell, The Bonnot Co., The Brad-

PREFACE. V

ley Pulverizer Co., The Cement Age, F. L. Smidth & Co., The Geo. V. Cresson Co., F. D. Cummer & Son Co., The Fairbanks Co., The Kent Mill Co., Lathbury & Spackman, Tinius Olsen & Co., Thos. Prosser & Son, Riehle Bros. Testing Machine Co., Frederick W. Taylor, Henry Troemner and the International Instrument Co.

W. PURVES TAYLOR.

January i, 1906.

Table Of Contents

Page CHAPTER I.— CLASSIFICATION AND STATISTICS 1

Definitions. — Classification. — Distinguishing characteristics of the different varieties. — Historical data. — Statistics of the Portland cement industry.

Chapter Ii.— Composition And Constitution 7

The components of Portland cement. — Silica. — Lime. — Alumina. — Iron oxide. — Magnesia. — Sulphuric acid. — Sulphur. — Alkalies. — Carbonic acid. — The constitution of Portland cement. — Various theories.

Chapter Iii.— Manufacture 15

Raw materials. — Calculation of mixtures. — Processes of manufacture. — Dry process with rotary kilns. — Dry process with stationary kilns. — Wet process with rotary kilns. — Wet process with stationary kilns. — Portland cement from slag. — Essentials to good manufacture.

Chapter Iv.— Inspection And Sampling 34

Reception and storage of shipments. — Shipments in bags and barrels. — Time to hold shipments for test. — Field inspection. — Color of cement. — Weight of packages. — Methods of sampling. — Treatment of samples.

CHAPTER v.— THE TESTING OF CEMENT 41

Classification of tests. — Qualities necessary for a good cement to possess. — The common tests of cement. — The development of testing.— Methods of testing'. — Uniformity in testing. — Requisites for good testing.

Chapter Vi.— Specific Gravity 4g

Definition, — Objects of test. — Underburning. — Adulteration. — Effect of seasoning. — Effect of composition. — Effect of fineness. — Effect of humidity. — The test of apparent density. — Forms of apparatus. — Methods of operation. — Preliminary treatment of sample. — Sources of error. — Interpretation of results.

Chapter Vii.— Fineness C3

Necessity for fine grinding.— Objects of test. — Various methods of determining fineness. — Sieves. — Wire cloth. - Specifications for wire cloth. — Mechanical sifting. — Size and shape of sieves. — Treatment and size of sample. — Accuracy of test. — Methods of operation. — Sources of error. — Interpretation of results.

viij CONTENTS.

Page

Chapter Viii.— Time Of Setting 80

Definitions. — Objects of test. — Theory of setting. — Effect of composition. — Effect of seasoning. — Effect of mixing water.— Effect of fineness. — Effect of environment. — Rise of temperature during setting. — Normal consistency. — — Methods of determining consistency. — Forms of apparatus. — Methods of operation. — Sources of error. — Accuracy of test. — Interpretation of results.

Chapter Ix.— Tensile Strength 101

Definition. — Objects of test. — Reasons for use of tensile test. — Neat versus sand tests. — Effect of composition. — Effect of seasoning. — Effect of fineness. — Effect of environment. — Making briquettes. — Amount of mixing water. — Temperature of mixing water. — ^Purity of mixing water. — Sand. — Form of briquette. — Moulds. — Method of mixing and moulding. — Hand methods. — Mechanical methods. — Storage of briquettes.— Environment during setting. — Appliances for storing. — Marking briquettes. — Various types of testing machines. — Adaptability of the different types. — Form of clip. — Rate of stress. — Wet briquettes.— Number of specimens to test. — Accuracy of test. — The strength of Portland cement. — Interpretation of results.

Chapter X.— Soundness 156

Definition. — Causes of unsoundness. — Effect of seasoning. — Effect of fineness. — Methods of determining soundness. — Measurements of expansion. — Normal lests. — Accelerated tests. — Methods of conducting accelerated tests. — Forms of apparatus. — Value of accelerated tests.— Interpretation of results.

Chapter Xi.— Chemical Analysis 184

The components of cement. — Significance of analyses. — Methods of analysis. — General method. — Alternative and additional methods. — Rapid methods for control work. — The detection of adulteration. — Microscopical tests. — Equipment for chemical testing. — Value as a routine test.

Chapter Xii.— Special Tests 212

Compression tests. — Transverse tests. — Tests of adhesion. —Shearing tests. — Abrasion tests. — Porosity. — Permeability. — Yield tests of mortar. — Tests of sand. — Tests of stone.

Chapter Xiii.— Approximate Tests 226

The use of approximate tests. — Tests of fineness. — Tests of setting. — Tests of strength, by the tensile test. — Tests of strength by the transverse test. — Tests of soundness. — Tests of purity. — Apparatus. — Interpretation of results.

Chapter Xiv.— Practical Operation 239

Equipment of a laboratory. — Amount of labor required. — Cost of operation. — Organization of the laboratory. — Recording systems.

CONTENTS. ix

Page CHAPTER XV.— OTHER VARIETIES OF CEMENT 252

Natural cements, their manufacture, properties, tests and the interpretation of results. — Improved cements. — Pozzuolana cements. — Slag cements. — Sand cements. — Mixed cements.

Chapter Xvi.— Specifications 275

Discussion of specifications. — Standard specifications. — A form of specification for Portland and natural cements. — The general interpretation of specifications.

APPENDIX A. — Standard methods of testing, proposed by the Committee of the American Society of Civil Engineers 287

APPENDIX B.— Standard method for the chemical analysis of cement, proposed by the Committee of the New York Section of the Society for Chemical Industry 296

APPENDIX C. — Standard specifications for Portland and natural cement, adopted by the American Society for Testing Materials 299

APPENDIX D. — Specifications for Portland, natural and puzzolan

cement, adopted by the Corps of Engineers, U. S. Army.... 303

APPENDIX E. — British standard specifications for Portland cement,

issued by the Engineering Standards Committee 308

APPENDIX F. — Specifications for Portland cement, adopted provisionally by the Canadian Society of Civil Engineers 312

APPENDIX G. — A list of reference books on cement and concrete. . 313

INDEX 31G

Chapter I. Classification And Statistics

Definitions. — Hydraulic cement may be somewhat broadly defined as a material which, when pulverized and mixed with water into a paste, acquires the property of setting and hardening under water. In engineering construction, four classes of cement are generally recognized — (i) Portland cement, (2) natural cement, (3) Pozzuolana cement, (4) mixed or blended cement.

Portland cement is the product resulting from the process of grinding an intimate mixture of calcareous and argillaceous materials, calcining the mixture to incipient fusion, and grinding the resulting clinker to a fine powder. It must contain no materials added subsequent to calcination other than small amounts of certain substances used to regulate its setting properties. The German Association of Portland Cement ^lanufacturers has adopted the following definition of Portland cement :

"A hydraulic cementing material having a specific gravity of not less than 3.10 in the calcined condition, and containing not less than 1.7 parts bv weight of lime to one part each of silica, alumina, and ferric oxide, the material being prci)are{l by intimately grinding the ravv' ingredients, calcining them to not less than clinkering temperature, and then reducing this clinker to a proper fineness."

English and American societies* also have adopted standard definitions of Portland cement similar in tenor, although not as explicit, except that they limit the amount of substances that may be added after calcination to two or three i)er cent.

Natural cement is the product resulting from the burning and subsequent pulverization of an argillaceous limestone or other suitable rock in its natural condition, the heat of burning being insufificient to cause vitrification.

♦See Appendices.

This class of cement is also commonly known as ''Rosendale, ' it being; so named from the district in the eastern part of New York State which is the greatest producer of natural cement. This term, however, when applied to all naturals is a misnomer. In England the name "Roman" is applied to certain grades of this material.

Pozzuolana cement is obtained by grinding together an intimate mixture of slaked lime and blast-furnace slag or volcanic scoria. The cement is not burned, the hydraulic ingredients being present only as a mechanical mixture. This material must not be confused with slag Portland, which is a regular Portland cement in v/hich the slag furnishes the silicious ingredients, thus replacing the clay or shale in the mix.

Mixed cements cover a large variety of products made by combining the other forms of cement or by mixing them with an inert material. The so-called "second-grade" Portlands generally belong in this class, since they consist usually of Portland cement mixed either with natural cement or with raw rock, cinder or slag. Sometimes, however, these cements are merely made of inferior clinker, in which case the>- are to be classed with Portlands. 'Improved cements" are naturals containing from 10 to 30 per cent, of Portland clinker. Sand cements are made by finely grinding a mixture of Portland cement and sand, usually in equal proportions. These varieties of mixed cements are those most commonly encountered, although many other forms are to be found on the market.

Distinguishing Characteristics. — The distinguishing characteristics of Portland cement are — in manufacture, the use of an artificial mixture, the grinding before burning, and the calcination to incipient fusion — and in use, its heavier weight, its slower set, and its greater strength.

Natural cement is distinguished in manufacture by its production from a single variety of material, unground, and burned at a low heat, and in use by its lighter weight, quicker set, and lower strength in the earher stages of hardening.

In what follows in this book, the discussion will be limited to Portland cement (except for Chapter XV.) , so that whenever the unqualified term cement is employed^ Portland cement alone is to be understood.

CLASSIFICATION AND STATISTICS. 3

Historical.* — Although Smeaton, when building the Ecldystone lighthouse in 1756, discovered that the addition of clay to lime would render it capable of setting under water, no real Portland cement was produced until 1824, when Joseph Aspdin, a bricklayer, of Leeds, England, took out a patent for producing a cement by calcining a mixture of lime and clay. He gave to it the name ''Portland" on account of its resemblance when hardened to the famous oolitic limestone, used extensively for building, found in the ''Isle of Portland," a peninsula on the southern coast of England. The first works for producing this material were estabhshed at Wakefield by Aspdin in 1825, while the construction of the Thames tunnel in 1828 was the first important piece of engineering work to use it in any quantity.

This early cement, however, was very different from that of modern days, chiefly in that the burning was never carried up to the point of vitrification, so that the elements could never have been properly combined. It was not until about 1845 that the manufacture began to be placed upon a scientific basis, and Portland cement, as we now^ know it, to be produced. In Germany, the first works were established near Stettin in 1852, while 1875 niarked the beginning of the industry in the United States.

Statistics of Industry. — The development of the Portland cement industry, particularly in the past decade, has been of such remarkable proportions, that a brief summary of statistics regarding it cannot fail to show the importance of the scientific study of a material so widely employed. The following data are taken largely from the report of the Geological Survcyf for 1903:

Table I. shows the domestic production of Portland, natural and slag cements. It will be observed that the production of natural cement has remained practically constant for the past fifteen years, while the Portland production, particularly since 1895, is remarkable in its rapid increase. The chief reason that Portland cement production, prior to 1895, advanced so slowly, was the deep-seated prejudice of our engineers against

♦For more complete data on the history of cement consult Redprrar*'!. -mVcareous CTements," and f!ummings' "Anicrlcnn Cenicnts •'

f'The production of Comont in lOm.' h- 1.. L. Kimhnll-ex nuM from Mineral Resources of the United States" for lOO^-issued by Division of Mtnlng and Mineral Recources, U. S. Greoloplcal Survey.

4 PRACTICAL CEMfiXT TESTING.

the (loniostic and in favor of the forcii^n material. Even as late as the oml of the past decade, it was not uncommon for the specifications of imi)ortant work to call for foreign cement, or at •least for a cement "that shall be equal to the best German brands." The excellence of the domestic product and its

TABLE I.— Total Production of Natural, Portland and Slag Cement in the

United States. 1818-1903.

(From Mineral Resources of the United States, 1903.)

Year.

I88I i^S3 l88q

i8qo

i8qi

i8q2 i8q4 i8q6 iSqg

I QOO

lOOI.

iqo3

Pozzuulana

Natural.

Portland.

or Slag.

300,000

1,000.000

4,250,000

11,000,000

16,420,000

22,000,000

82,000

2.030,000

42,000

2,440,000

60,000

3,165,000

85,000

4,190,000

90,000

4,000,000

100,000

4,100,000

150,000

4,186,152

150,000

6,692,744

250,000

6,253,295

250,000

6,531,876

300,000

7,082,204

335'000

7.451^535

454.813

8,211,181

547.440

7,411,815

590,652

7,563.488

7,741.077

990.324

7,970,450

1.543.023

12,265

8,311,688

2.677,775

48,329

8,418,924

3,692,284

150,895

9,868,179

5,652,266

335,000

8.383.519

8,482,020

446,609

7.084,823

12.711,225

272,689

8,044.305

17,230,644

478,555

7.030,271

22,342,973

525,896

209,132,526

79,608,196

2,270,238 equality to the best foreign material is now, however, generally conceded.

Table II. gives the distribution of production by States, Pennsylvania leading both in the number of mills and in quantity of output.

The distribution by sections (Table III.) divides the production into the "Lehigh district" of Pennsvlvania and Ne\v Jersey, the New York district, both of which operate on limestone, and

CLASSIFICATION AND STATISTICS.

TABLE II. — Production of Portland Cement in the United States in 1903,

by States. (From Mineral Resources of the United States, 1903.)

State.

Alabama

Arkansas

California

Colorado

Georgia

Illinois

Indiana

Kansas

'Michigan

Missouri

New Jersey

New York

Ohio

Pennsylvania . . .

South Dakota. . .

Texas

Utah

Virginia

West Virginia. . .

Tumber of Work.s.

Quantity Barrels.

a I

631,151 258,773

5 3

d I

1,257,500

3 12 n

I

1,955.183 825,257

2,693,381

1,602,946 729 519

9.754.313

Value.

$1,019,352 436,535

1,914.500

538,13^

690,105

(a) Includes product of Utah and South Dakota.

(b) Includes product of Texas, (r) Includes product of Arkansas.

{d) Includes product of Alabama, Georgia and West Virginia.

TABLE III. — Showing the Development of the Portland Cement Industry in

the United States. (From Mineral Resources of the United States, 1903.)

Section.

New York

Lehigh and Northampton counties, Pa., and Warren County, N. J.

Ohio

Michigan

All other sections

Total

New York

Lehigh and Northaiiipton counties, Pa., and Warren County, N. J.

Ohio

Michigan

All other sections

Total (>s

Nurnbor of works.

QiMutity.

Barrels.

65,000

Per cent.

Number of works.

Quantity.

Harrols.

465,832

Per cent.

5.5

5 2

47 5""

60.0 6.5

14.1

100.0

6 6 >5

8,482,020

72.6

•6.3

7.8

335.500 lOO.O

0.8

>3

M

«.7

as.6

17,230,04.1

UXJ.O

PRACTICAL CEMF.XT TESTING.

the Ohio-Michigan district, which generally employs marl. The Lehigh district produces over half the Portland cement made in the country. The recent development of the industry in Michigan is also a feature of the table.

The diagram, Fig. i, shows a comparison of the consumption

30,000,000

28,000,000

26,000,000

24,000,000

22,000.000

20,000,000

18,000,000

16,000,000

CDc0Q0C0a0C0COGDQ000CDaD00C0COaQcDGnCX>OlO>

Years, YiG. I. — Diagram Showing the Quantity of Cement Consumed in the United States During the Twenty Years, 1883-1903.

of Portland cement in the United States with the entire amount of hydraulic cement imported in the last few years, which illustrates not only the growth of the industry, but also the marked increase in the ratio of domestic production to total consumption.

r"

—

\l

N

- k

^1.^.

Y leV/"^

¥h

N

1^ r l//lv rt'l

Y M i

0^ yu

^i-^i Jciil !.,j>,j«t;i

£j i#A

X

N i ^Toj3l^. ; yf i ^yC^'—s^zSACin^ \

Chapter II. Composition And Constitution

Composition. — The essential components of Portland cement are silica, alumina and lime. Other ingredients always occurring in appreciable quantities are iron, magnesia, alkalies, sulphuric and carbonic acids, and water.

Le Chatelier has stated* the following to be the limits of these ingredients in good commercial cements :

Silica 21 to 24 per cent.

Alumina 6 " 8

Iron oxide 2 " 4

Magnesia 0.5 '' 2

Sulphuric acid 0.5 '* 1.5

Carbonic acid and water . . . i *' 3

Bleininger, in a chaptert on the Nature of Portland Cement, sets the following limits :

Silica 19 to 26 per cent.

Alumina 4 " 11

Iron oxide 2 '' 5

Magnesia o " 5

Sulphuric acid o " 2.5

Alkalies o " 3

The percentages given b\ Le Chatelier should be considered as average rather than limiting values, since many good cements have an average composition largely out of these limits ; Hleininger's limits, on the other hand, are somewhat extreme.

Table IV. gives the results of chemical analyses of many standard brands and serves to show the composition of conunercial cements. In addition to these important ingredients, the following usually occur in very small (juantities: Titanium, phosphoric acid, sulphur, manganese, silicious sand, coal ami ash from the kilns, flint and iron from the mills, etc. Tlie sum total of these elements, however, rarely e(|uals (Mie per cent.

f'Manufac-ture cf Portland Comonts." by A. V. HleJniMKor. l-ourlh s.<rl.>s. Rullolin :{, Ohio Slato C.ooloKical Survey.

PRACTICAL CEMEXT TESTIXC.

uonniSi no'ssoq

'OS

NMrOMN--HN

'O '•IT

M « N N N

"~ LT-. o X i .i g i

^i§ i

III i I ^ I

C = C

'^7 5 a ^

COMPOSITION AND CONSTITUTION. g

Silica (SiOg). — Silica constitutes from 19 to 24 per cent, of a Portland cement, and combined with lime to form calcium silicate furnishes the active factor in its hardening. It should exist entirely in a combined state and not as quartz sand, in which condition it does not combine with the lime, and is inert in the subsequent reactions.

Lime (CaO). — A well proportioned cement will contain from 59 to 67 per cent, of lime, depending upon the amounts of silica and alumina present. Provided all of it exists in a combined state, the greater its amount, the higher will be the strength of the product. If, however, more lime is present than will combine with the silica and alumina, an unsound cement will result, since the excess of lime in slaking will expand and so cause disintegration. The recent demand for high strength values at early periods has caused many manufacturers to produce very high limed cements, often passing the limit of safety unless exceptional care has been taken to so finel} grind the raw materials and to burn so perfectly that the theoretical limit of combination can be approximated. A high proportion of lime also requires greater heat in burning, and forms a slow-setting cement.

*"Too low a proportion of lime, on the other hand, produces a fusible clinker, liable to overburning. This is especially the case with aluminous materials. If hard-burned, such mixtures give a fused clinker liable to fall to dust on cooling, hard to grind, and yielding slow-setting cement of poor hardening properties. If light burned, an over-clayed mixture yields soft brownish clinker, grinding to a brownish, quick-setting cement of inferior strength."

Alumina (ALO.j). — From 5 to 10 per cent, of alumina i> usually present in Portland cement, mostly combined as calcium aluminates, and since the setting ])roperties are due to these aluminates, the greater the proportion of this ingredient, the (|uicker will be the setting of the product. The ultimate tensile strength of high alumina cements is also apt to be inferior. The presence of alumina tends to facilitate the burning, since its compounds are much more fusible than those of silica, and it present in

*S. B. Newberry in "Coiicreto. I'lain ;iml Itfiiifoicod." by F. W. Tiiyli^r ami 6. E. Thompson.

excess make a mixture difficult to burn uniformly and hard to grind.

Iron Oxide (FCoOg). — According to Le Chatelier, ferric oxide and calcium carbonate on burning yield products which slake with water and possess no hydraulic properties. Schott and Richardson, however, have prepared cements in which all of the alumina was replaced by iron, and therefore conclude that the function of these two ingredients is practically similar. The investigations of S. B. and W. B. Newberry also appear to show that the action of ferric oxide and alumina in promoting the combination of silica and lime is practically identical. The amount of iron usually present in Portland cements is less than 4 per cent.; and it exerts but little influence on the physical properties of the material. The dark gray color of cement is due to the presence of iron compounds, cement prepared from silica, alumina and lime only being colorless.

Magnesia (MgO). — The role of magnesia in Portland cement has not yet been definitely determined. Certain investigators, including Le Chatelier, claim that magnesia may replace lime, and form silicates and aluminates of magnesia whose characteristics are similar to those of calcium. Others, including Erdmenger, Richardson and Newberry, consider that magnesia remains free in cement and is present only as an adulterant.

Investigation of two or three failures of important engineering works placed in sea water developed the fact that the concrete was high in magnesia, but it has been claimed that its presence was due to a deposition of magnesian salts from the water, and that the magnesia originally present in the cement was not responsible for the failures. A committee of the Association of German Portland Cement Manufacturers in 1895 presented a report stating that their investigations had shown that a content of magnesia up to 8 per cent, was harmless. Dyckerhofif, however, presented a minority report, stating that his experiments had shown that more than 4 per cent, of magnesia, whether added to a normal mixture or substituted for an equivalent portion of lime, caused a steady falling off in the strength of the cement, although actual cracking was observed only with 8 or more per cent. S. B. Newberry has stated that he has made a cement containing 9 per cent, magnesia which

COMPOSITION AND CONSTITUTION. u

stood the boiling test, but that one containing 15 per cent, failed after several hours' boiling.

American Portland cements contain on an average from 2 to 4 per cent, magnesia, the latter value being the limit placed by the specifications"^' of the American Society for Testing Materials.

Sulphuric Acid (SO3). — Portland cement contains usually from 1.25 to 1.75 per cent, of this ingredient, a large proportion of which is due to the admixture of calcium sulphatet with the finished cement for the purpose of regulating the setting properties. Such additions, however, should not exceed 2 or 3 per cent., which is equivalent to about i or 1.5 per cent, of the anhydrous sulphuric acid, since an abnormal amount is injurious to both the strength and soundness of the product, especially if placed in sea water. The standard French specifications limit sulphuric acid to i per cent., while American practice is to permit from 1.5 to 2 per cent. The specifications of the American Society for Testing Materials place the limit at 1.75 per cent.

Sulphur (S). — A small amount of sulphur is sometimes introduced from the coal in burning, and occasionally from the reduction of sulphates in the raw materials. These sulphides cause the cement in hardening to become mottled with dark blue spots, and if present in appreciable quantities may cause disintegration, due to the expansion in oxidizing on exposure to the air.

Alkalies (K2O and NaoO). — These appear to have but little influence on either the burning or quality of the cement. It has been held that alkalies acted more or less as a flux in facilitating the combination of lime with the silica and alumina, but later experiments have apparently disproved this theory. Excess of alkaHes, under certain conditions, is said to be responsible for unsoundness, although this fact has never been (lefinitel\' proven. The percentage of alkalies is usually from 0.5 to 2 per cent.

Carbonic Acid (CO.). — Provided the temperature o{ burning has been sufficiently high, the i)resence of (his elenuMit is due solely to absorption from the air, and since lime wIumi insntTi clently combined is very active in this absorption, an abtK^rniMl percentage of carl^onic acid must show either nn(ler])urning or an

*See Appendix C. tSfie paRes 20 and 8,^.

excess of lime. Normal cements contain from 0.5 to 1.5 per cent, of carbonic acid.

Constitution. — Although Portland cement has been extensively manufactured for well over hah" a century, it is only in comparatively recent years that a scientific study of the constitution of cement has been undertaken, and even at present the exact state of combination m which the different components exist has not been definitely ascertained. For many years it was the practice to proportion raw materials and to study the finished product by means of the "hydraulic modulus," or the ratio of the weight of lime to that of silica, alumina and iron, which varied between 1.8 and 2.2, but it soon became understood that the lime combined in different ratios with the other components, and the inadequacy of the formula was thus recog^iized.

Le Chatelier in 1887 was the first to attempt scientifically to explain the constitution of Portland cement.* Following petrographic methods, and examining with a microscope thin sections of clinker and hardened cement, he succeeded in separating them into two chief constituents, which later were named by Tornebohm alit and celit. two accessory constituents, belit and felit, and in addition an amorphous isotropic mass of matter. He considered aht to be composed of tri-calcium silicate, 3 CaO SiO.. which is the active element in the hardening ot cement, and which hydrates as follows :

3 CaO SiO, - Aq = CaO SiO,2.5 H.O - 2 Ca rOHV

Cement also contains a tri-calcic aluminate, somewhat unstable, but setting rapidly in water, hydrating according to the formula :

3 CaO ALO, - Aq = 3 CaO Al,0,i2 H,0

Portland cement, according to Le Chatelier. therefore consists of tricalcium silicate mixed v.ith calcium aluminate and ferrate and also containing mono and dicalcium silicates. He expresses the hydration of cement by the following two reactions :