Steel and its heat treatment

Book i L

dpghtN?- -_

COPXRIGHT DEPOSCR

S 1 JlEjL

AND ITS HEAT TREATMENT

BY DENISON K. BULLENS

Consulting Metallurgist

SECOND EDITION, THOROUGHLY REVISED TOTAL ISSUE, FIVE THOUSAND

NEW YORK

CA v-

^ A

<°tft

BY

.DENISON K. BULLENS

All rights reserved

PRESS OF

ERAUNWORTH & CO.

ROOK MANUFACTURERS

BROOKLYN, N. Y.

IN MEMORY OF MY FATHER

Digitized by the Internet Archive in 2011 with funding from The Library of Congress

Preface To Second Edition

In the second edition the scope of this work has been broadened to include additional information of a practical nature to illustrate further the application of principles in everyday commercial practice, and to encourage a consideration of every element in the cycle of operations from the initial heating of the steel for forging to the cooling in the final heat-treatment process.

In the section on Heat, additional data are given to illustrate the difference between combustion and generation of heat and the application of heat to useful work; the difference between the mere indication of uniform temperature and uniformly heated product; the relation between temperature, time, mass, and surface in the determination of uniformly heated product; the influence of furnace design and operation on the quality and cost of finished product; the weakness of relying upon pyrometer readings without considerirg other equally important factors; and the factors governing the selection of furnaces and fuels and the use of both.

The section on Forging has been materially enlarged into a new chapter to illustrate the relation of forging to heat treatment, the effect of temperature, time and uniformity of heating upon the structure of the steel, together with original photomicrographs illustrating the variation in structure under distinctive conditions.

The influence of the " Human Element," while not involved in a technical consideration of scientific principles, is nevertheless an important factor in the practical side of the work, and it is considered more in detail for the reason that there has been shown nothing so far to prove that it is not essential. It is believed that the practice of this art of heat treatment is a specialty or trade in itself, requiring skilled men for its proper conduct, and that shop practice vi PREFACE

cannot keep pace with laboratory development until this point is recognized and sustained. To encourage favorable consideration of this view an effort is made to indicate some of the weaknesses of our so-called modern heating practice and to illustrate the relationship of furnace and operator to the quality and cost of finished product, with suggestions for the improvement of both.

The chapter on Annealing has been enlarged with considerable data of interest to the practical man, illustrating the effect of temperature, time, and saturation in both heating and cooling upon the grain size and uniformity of structure, and thus more fully emphasizing in a practical manner the basic principles of the effect of heat upon steel. Also there is shown a series of unusual photomicrographs which illustrate most graphically the theory of " diffusion and equalization " advanced in the previous edition.

The number of photomicrographs and other illustrations has been materially increased to supplement the written matter, and the whole is based substantially upon observations made under actual working conditions.

D. K. BULLENS.

Philadelphia,

Preface To First Edition

Modern Heat Treatment should be considered as an art or trade, since it certainly requires knowledge, skill and judgment for its proper performance. These, in turn, necessitate at least some knowledge of heat, of steel, and of the effect of heat upon steel. And all three factors are linked together by the " human element." The author has therefore endeavored to bring together the theoretical and practical sides of the general subject of steel and its heat treatment in such a manner as will, he hopes, be understandable by that " human element."

The theoretical or scientific side of heat treatment is not affected by a consideration of many elements which enter into the art or practical application of metallurgical principles, but unless performance in the shop compares favorably with the standard set in the laboratory, the value of the work is lessened. The technical principles have been dealt with heretofore more thoroughly than the application of these principles, and it is the purpose of this book more completely to consider this practical aspect in the hope that a better understanding of the factors governing commercial practice mry suggest means of improvement in shop methods.

It has been the author's attempt to make the chapters dealing with the heating problem illustrate the necessity for considering all of the elements entering into the production of a uniformly heated product; of heat application rather than details of construction; of the importance of the human element and the relation it bears to the ultimate results; of encouraging a broad and common-sense view of the principles governing this important operation and which generally have been overlooked; and finally, of viewing the viii PREFACE

heating problem as an engineering proposition, adapting each fuel to proper furnace design and operation to meet the requirements of the problem in hand, and by so doing aim for the adoption of the standard heating unit in terms of finished product — " the cost of a unit of quantity of given quality."

He has attempted to make as practical as possible those chapters relating to steel and the effect of heat upon steel. Theories have been advanced only so far as has been thought necessary for a clear understanding of principles. Wherever possible, illustrations in the form of photomicrographs and charts have been given. The data given under the various types of heat-treated steels have been checked as far as possible and every effort has been made to be correct.

To the many friends who have aided him in the preparation of this book the author would express his sincere appreciation. Effort has been made to give due credit for cuts and data at the proper place, and for such as may not have been made, acknowledgment is hereby rendered.

Denison K. Btjllens.

Philadelphia,

October 1, 1915.

NOTE TO SECOND IMPRESSION WITH ADDITIONS

To comply with requests of readers of the First Edition of this book for concise information upon the treatment of modern high-speed steels, it has been deemed advisable to add an Appendix on this subject.

D. K. B.

Contents

HAPTER PAGE

I. The Testing of Steel 1

II. Heat Generation 16

Chapter I

Growth of Heat Treatment.— Probably no one division in the metallurgy of steel has taken such wonderful strides in recent years as has the art of heat treatment. Twenty years ago the scientific knowledge and technical application of heat treatment were but very limited. Such as it was, it usually consisted in " heating to a red heat " for annealing, or perhaps the instructions called for " harden at a bright red and temper to a straw color." Then it was an art guarded with much secrecy and confined for the most part to makers of tools and a certain few specialties.

Practically all alloy steels require treatment of one sort or another. In the " natural " state very few steels present their full value, so that heat treatment is not only advisable but often mandatory.

Necessity for Heat Treatment.— Take for example the steels used in the automobile industry. The frame requires resistance to vibratory stresses occasioned by rough roads, as well as strength and toughness. Rear axles must have great tortional resistance; front axles " must withstand vibrations. The steering parts must be strong, tough and without brittleness; the springs must neither sag nor break. Crank-shafts must be able to resist impact, besides being stiff. Gears are subject to wear and must be capable of withstanding this action if a smoothly running transmission is to be had. And so each separate part might be named, all having a more or less severe duty to perform and requiring steels possessing various degrees of strength, toughness, resilience, endurance, shock-resisting and wearing qualities.

Testing. — These various combinations of static and dynamic strength are obtained by adjusting and correlating both the chemical composition and heat treatment of the steel. Certain chemical components intensify the static properties of the material; others may affect the dynamic qualities. Thus by coupling with a steel of suitable chemical combination the proper heat treatment, there arises a product with physical properties most adapted for the work in hand. Similarly, having once produced a suitable article, it then remains to duplicate it. To this end all rational heat treatment must be aligned and standardization of results be obtained. In order to accomplish these specific requirements, the influence of definite chemical composition and definite treatment must be known, as will be described in later chapters. The guide to this work is frequent and constant testing, and a definite knowledge of the various components should be possessed by every heat-treatment man. Thus we may say that the purpose of practical testing is (1) to supply information as to suitable material and its qualities for different purposes, both for the manufacturer of the material and for the designer or user, and (2) to test the specified uniform quality of the material.

Stresses and Strains. — Testing resolves itself into a determination of the strength of the material, which in turn is measured by the application of a force and its resultant effect. The force put upon a body is termed the stress, and the deformation resulting from that force is the strain. Upon the method of applying that force depends the nature of the test. Thus we may conveniently classify such stresses under the following headings:

A. Steady or constant loads- — static stresses;

B. Repeated static stresses and accelerated stresses — fatigue

stresses;

C. Suddenly applied loads — impact stresses;

D. Repeated impact or vibratory stresses — dynamic stresses;

E. Miscellaneous tests such as resistance to penetration, wear,

etc.

Tensile Strengths — The most common test for static strength, that is, the strength of the steel under constant load and without shock or vibration, is the tensile test. Thus we may call the tensile strength the absolute strength of the metal under tension, i.e., the force actually required to pull the metal asunder. A standard test piece is gripped between the upper and lower jaws of a testing machine and the total resistance to rupture is measured. Knowing the area of the cross-section of the test piece and the load required to break it, the strength per square unit may then be calculated. The tensile strength is usually given in pounds per square inch (American) , tons per square inch (British), or kilograms per square millimeter (metric system; 1 kg. per mm. 2 =1422.32 lbs. per square inch). The accuracy of the tensile test is dependent not only upon the conditions under which the test is made, such as the rate of pulling, alignment of test piece in the machine, etc., and which are more or less influenced by the human element, but also upon the metal itself. The higher the tensile strength and brittleness of the steel, the greater the possibility of error; differences of several thousand pounds per square inch are often encountered in the same piece of high-tensile, heat-treated steel, even in the absence of brittleness.

Test pieces are generally taken half way between the center and the outside of the piece, and longitudinally or " with the grain." Occasionally it is necessary to take tests transversely or " across the grain"; in this case the results will be lower than in the longitudinal test, the exact amount depending upon the composition and treatment of the steel.

The static strength increases in direct proportion to the carbon content of the steel. For the ordinary basic and acid open-hearth steels, without heat treatment, Campbell gives the following formulae by which the tensile strength of such steels may be roughly determined. These results apply for steel " in the natural."

Acid open-hearth steel :

Tensile strength = 40,000 + 1000C + 1000P+zMn. Basic open hearth steel:

Tensile strength = 41, 500+ 770C + 1000P+yMn.

In these formulae, C equals each one point (0.01 per cent.) of carbon as determined by combustion, P equals each 0.01 per cent, of phosphorus, Mn equals each 0.01 per cent, of manganese, and x and y are given in the table on page 4.

Elastic Limit (Tension). — The term " elastic limit " has probably been more ill-used than any other common technical testing name, with the possible exception of " hardness." Among its many definitions the two which stand out pre-eminently are (1) the least stress at which the material retains a permanent deformation or " set " after the removal of the stress; and (2) the least stress under

Percentage of Carbon.

On Acid Steel.

X

Lbs. per Sq. In.

On Basic Steel.

V

Lbs. per Sq. In.

0.05 ;

not

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

0.60

* Beginning only with 0.4 per cent, manganese, t Beginning only with 0.3 per cent, manganese.

which ductile material exhibits a marked yielding — sometimes denoted as the " yield point."

The determination of the true elastic limit should always be taken from a curve plotted, using an extensometer, from a series of careful observations, as otherwise sets caused by non-homogeneity and initial stress might be obtained which do not represent the plasticity of the material. This method of determining the elastic limit is but little used commercially, as the amount of labor involved is too great.

The yield point, or commercial elastic limit, is obtained by noting the stress at which the test piece first begins to " give " or elongate. This may be obtained by means of two prick-punch marks and observing the first signs of elongating by means of dividers held on these points; or by noting the drop of the weighing beam or halt in the load indicator (" jockey ") ; or by means of the general appearance of the test piece.

In its practical application the elastic limit may be called the working strength of the material, for in most cases the steel or machine part becomes useless when strained beyond its elastic limit. This is particularly true of automobile construction, in which the value of a car is dependent upon the correct adjustment and alignment of its several working parts, such as in transmissions and transmission suspensions. All tests given in this book, unless otherwise noted, refer to the commercial elastic limit or yield point.

The relation existing between the elastic limit and the tensile strength is too broad a subject for discussion here, as the varying chemical compositions and heat treatments exert such a tremendous influence; a study of the results given in following chapters will show a proportionality of 40 per cent, and upward.

Elongation. — The elongation is measured in per cent, of the original test section and is commonly the amount of stretch which will occur in the material when pulled apart by tension. It is usually measured in relation to an initial distance of 2 or 8 in., or 100 mm. when the metric system is used, but other specifications as used in Europe give a definite relation of original gaugelength to the thickness or diameter of the specimen.

Reduction (or Contraction) of Area. — The reduction of area refers to the area at the point of rupture, usually reported in per cent, reduction of the original area — that is, the original area of the test piece minus the area of the smallest cross-section after fracture; this divided by the original area is the percentage reduction of area. The reduction of area should be regarded as a measure of the saturation of the steel at the proper temperature.

Ductility. — The percentage elongation and percentage reduction of area are a measure of the " ductility " of the material, usually varying inversely with the tensile strength. The true measure of the ductility of the steel cannot be taken alone from either the elongation or reduction of area, as the results obtained in either case will depend in a large measure upon the size of the test piece, the method of testing, etc. Many engineers regard the reduction of area as the more reliable; this is offset by the fact that many steel specifications make no mention of the reduction of area, but particularly specify the percentage elongation. Ductility may also be defined as the amount of distortion of the material before final rupture.

Compressive Strength. — The compressive strength of material is its resistance to crushing. The test is generally carried out upon a small cylinder or 1-in. cube of the metal, using the same machine as for the tensile test. Care must be used to see that the line of strain passes exactly through the axis of the specimen, and that the plates above and below the piece have a greater resistance to penetration than the metal to be tested. The application of the term elastic limit is similar to that in the tensile test.

Torsional Strength. — As its name implies, the torsion test is used to determine the resistance to twisting. This test is very largely used to-day for automobile steel and is measured in inch-pounds with the amount of distortion given in degrees. The elastic limit is obtained as in a tension test, using either a tropometer or an autographic attachment. The usual comparison is by calculating the shearing stress in pounds per square inch.

Endurance. — The computation and understanding of such static stresses as have been previously outlined are comparatively simple. The requirement to be fulfilled in designing is that the working stress shall not exceed the elastic limit of the material, whether it be in tension, compression or torsion. Numerous every-day failures, however, which cannot be accounted for by the limited information given by such tests, have forced investigators to probe more deeply into the complicated kinematic forces which seem to have such a great influence upon the " life " of the metal. It is now a well-known fact that, if a stress is applied a great number of times, i.e., repeated, each application being made before the material has had time to recover from the preceding stress, the material will eventually break even though the stress is below the elastic limit of the material. These repeated stresses upon steel cause a gradual disturbance of the structure and its component particles, which greatly weakens the material, and is called fatigue. The resistance to fatigue and its numerical test value may be termed the endurance of the steel. The stresses embodied under the heading of fatigue may be broadly classed as repeated static stresses and acceleration stresses ranging from zero to maximum or from a negative maximum to a positive maximum — alternating — stresses.

Fatigue Stresses. — These stresses are produced in a machine part by an external force or forces of varying strength and direction acting upon the part. When the force is produced by a continuously varying acceleration or retardation of masses taking part in the movement of the machine part, they may be conveniently termed acceleration stresses. 1 Typical stresses of this category are the revolving shaft stress on a loaded wheel or machine axle, the piston pressure and the acceleration pressure of the movable parts in the piston rod and crank-shaft of high-speed steam and oil engines. These perpetual stresses or so-called fatigue stresses are the essential ones in the movable parts of most high-speed machines, and a knowledge of the capacity of the material to resist them should serve as a basis for the selection of the material and design.

Rotary Bending. — Such static endurance tests may be carried out in a machine of the rotary bending type, such as the Wohler or the White-Souther machines. From a study of a large number of

1 J. O. Roos af Hjelmsaeter, Int. Assoc. Test. Mat., 1912, Vol. II, No. 9.

experiments made on a rotary bending machine of the Wohler design, Foos concludes that such endurance tests are not suitable as specification tests, but are of great value in the selection of material and the heat treatment for various purposes.

On the other hand, the real value of the rotary bending test as a criterion of the brittleness-fatigue endurance has been of late greatly questioned. That the results usually obtained are largely indicative of the elastic limit alone is probably more in accord with our present-day knowledge. The results from a series of tests conducted by Foos upon a Wohler type rotary bending machine with steels of 0.11, 0.40 and 0.65 per cent, carbon, given in the following table, would tend to support the latter theory, as one would naturally expect from past experience that the 0.40 per cent, carbon steel would have a greater fatigue-resisting strength than the 0.65 carbon steel.

ROTARY BENDING TESTS, WOHLER MACHINE

Chemical.

Static Propert es.

Endurance Limit.

Fiber

Stress oj oi

Elongain Kg.

s a o

M

OS

O a

03 ox

el t-t

o

-a p.

m

o -a Ph u 3 ft

n

3 m o o

m

Strength Lbs. per Sq. In.

Limit Lbs. per Sq. In.

giving Fracture

after 1 Million Revolutions.

R

A

0.11

0.33

0.019

0.013

0.01

49,770

32,990

34.1

Si

A

0.40

0.51

0.027

0.011

0.20

82,760

50,770

23.8 s 2

O.T.

0.40

0.51

0.027

0.011

0.20

109,780

70,820

15.6

Ti

A

0.65

0.49

0.023 0.007

0.20

116,040

50,900

14.3

T 2

O.T.

0.65

0.49

0.023 0.007

0.20

151,020

94,560

11.0

Treatment. " A," heated at 1560° F. for 30 minutes, and air-cooled.

" O.T.," heated at 1560° F. for 30 minutes, quenched in mineral oil, and re-heated to 1025° F.

Suddenly Applied Loads. — Machine parts at one time or another may be exposed to abnormal working loads such as may result from a single accidental blow or a sudden retardation of masses in motion, and which in any case cannot be supposed to be frequently repeated. Such abnormal stresses are therefore in the nature of suddenly applied loads or impact stresses, and constitute a different group from those previously discussed under static stresses. It is evident that such stresses demand that the material be able to sustain a great work of deformation for a single or a few impacts without rupture — that is, the machine part shall sustain as little damage or deformation as possible. In general the " ductility " (elongation and reduction of area) has for a long time been used as a measure of the work of rupture. But, although such tests are of comparative value, they do not measure either the ductility under impact or the strength or resistance under impact.

Drop Test. — The drop test as commercially applied may be described as an aggravated bend test on a large scale. It is a relative or qualitative test only, usually made on a full-size forging, to determine roughly the homogeneity of the metal and its ductility under shock. We have arbitrarily separated this test from the " impact " tests, reserving the latter %s applied to the specific measurement of the impact strength upon a more or less standardized test bar. The most common application of the drop test is that of locomotive axles, in which it is required that the axle shall stand a specified number of blows at a given height without rupture and without exceeding, as a result of the first blow, a certain deflection.

Impact Strength. — The impact test is used to determine the ability of the metal to withstand a suddenly applied load in the nature of an impact or shock, thus detecting brittleness or lack of toughness. This function is called resilience by the foreign technical world, referring to fragility or the converse of brittleness, and is stated in terms of the specific work of rupture under impact. It should not be confused with the English word "resilience," which is interpreted in this country as " springiness." This fragility is not defined by the tensile test, although an experienced steel man, from an examination of the size and aspect of the grain and other conditions of the fracture of the test piece, can usually express an opinion as to the fragility, but he cannot assess any definite value. Although the drop test specifies the fragility in a qualitative manner, it does not measure the actual resistance to rupture, and is therefore but an imperfect test. In order to overcome such objections and to arrive at a definite value, machines have been devised which break a special notched bar by a blow — the force required to rupture the metal being measured in kilogram-meters or foot-pounds. Notched test bars are used in order to localize the deformations. The blow must be delivered with sufficient velocity to bring out the desired brittleness functions. This blow or impact may be obtained by a falling weight (the Fremont machine), by a falling pendulum (the Charpy principle), or by a revolving fly-wheel bearing a releasable knife (the Guillery machine), these three representing the most common types of impact machines in use abroad, as well as the Olsen pendulum type (using a test specimen in the form of a cantilever) in this country.

Impact Tests. — Fremont x recommends as the ideal conditions to be realized in the application of the impact test: (1) A minimum drop of four meters, or proportional to the impact speed; (2) the weight of the anvil block to be equal to at least forty times the weight of the tup; (3) sufficient ease and rapidity of adjustment of the machine.

On account of the many factors entering into the problem and the numerous designs of impact testing machines, the majority of the testing associations have abstained from prescribing any special type of apparatus for performing the test. The Copenhagen Congress (1909) of the International Society for Testing Materials has, however, recommended the use of a standard notched test bar of 30X30X160 mm., or a smaller bar of 10X10 mm. cross-section when the larger size is not available. On the contrary, there are many who believe that a smaller, rectangular test bar reveals more clearly the local defects which form the nucleus of future cracks, etc.

Use of Impact Tests. — The fragility test has a double purpose — to point out steel which is defective, either inherently or by incorrect heat treatment; and to act as a valuable aid for the adjustment of a proper heat treatment. It is evident that steels burdened with sulphur and phosphorus, or rotten with piping and segregation, will always remain brittle whatever one may do. But ordinary, sound stock, properly treated, is nearly always non-brittle. The degree of brittleness will of course vary according to the composition, treatment and use of the different steels. The effect of heat treatment upon the impact strength is very great, so that due care should be used in so adjusting the chemical composition and treatment of the steel as to give the best combination for the work in hand. The impact test, in conjunction with other tests, gives a quick method of determining a quality the importance of which is yearly becoming more prominent scientifically; commercially, however, the impact tests are so unreliable, or vary so greatly, that they can hardly be used with any degree of accuracy.

Fatigue Impacts. — An impact or shock has a considerably greater effect than a stress slowly applied, and if repeated a sufficient number of times will eventually result in the rupture of the specimen. When such frequently repeated stresses are comparatively small —

iCh. Fremont, Proc. Int. Assoc. Test. Mat., Vol. II, No. 9, 1912.

that is, are well below the elastic limit of the material — they may be termed fatigue impacts. Their measurement, as determined by the energy or amount of work they represent, is a principal component of the dynamic strength of the material. Apparatus for thus testing the material is developed upon the principle of alternating impacts. As practical examples of such stresses in commercial application there may be mentioned the stresses to which the axles of locomotives or railway cars are subjected every time the wheel passes over a rail joint, or the impact stresses sustained by different parts of a motor car when passing over bad roads, or in changing to different speeds, etc.

Alternating Impact Machines. — Various machines for establishing comparative numerical values for this dynamic strength have been designed, in the endeavor to produce an alternating flexure and at the same time deliver a blow or impact. This has been accomplished by applying blows to the upper part of a test piece with the aid of two pendulum balls which, are made to fall alternately from opposite directions from a certain height against the test piece. Other machines have been patterned along the lines of the Upton-Lewis machine, keeping the fiber stress well within the elastic limit.

Alternating Impact Test Results. — The study of a great number of alternating impact tests of a comparative nature, the views of other engineers, and the study of steel parts broken in service, would lead the author to the opinion that the dynamic strength (using the term in its broadest meaning) of straight carbon steels reaches a maximum at 0.25 to 0.35 per cent, carbon, with perhaps even narrower limits of 0.25 to 0.30 per cent, carbon. Further, the maximum endurance is obtained when the steels have been properly hardened from a temperature slightly over the upper critical range and toughened at a temperature of 1200° to 1250° F. The maximum combination of static working strength, ductility, resistance to shock and vibration, and endurance probably is obtained in straight carbon steels with 0.35 to 0.45 per cent, carbon when subjected to the above treatment.

Relation of Various Tests. — There does not seem to be any simple relation between the elastic limit at steady tensile stress and the limit of endurance in the rotary bending test, although some engineers consider it as a " reflection of the elastic limit." The rotary bending is undoubtedly less than the former, and according to the investigations of some engineers, the limit of endurance (rotary bending) will generally amount to about 50 to 80 per cent, of the elastic limit.