This international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles

for the
Development of International Standards, Guides and Recommendations issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.

Designation: E21 − 17ϵ1

Standard Test Methods for

Elevated Temperature Tension Tests of Metallic Materials1
This standard is issued under the fixed designation E21; the number immediately following the designation indicates the year of original
adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. A superscript
epsilon (´) indicates an editorial change since the last revision or reapproval.
This standard has been approved for use by agencies of the U.S. Department of Defense.

ε1 NOTE—Section 3 was editorially updated in April 2019.

1. Scope* E8/E8M Test Methods for Tension Testing of Metallic Ma-

1.1 These test methods cover procedure and equipment for terials
the determination of tensile strength, yield strength, elongation, E29 Practice for Using Significant Digits in Test Data to
and reduction of area of metallic materials at elevated tempera- Determine Conformance with Specifications
tures. E74 Practices for Calibration and Verification for Force-
Measuring Instruments
1.2 Determination of modulus of elasticity and proportional E83 Practice for Verification and Classification of Exten-
limit are not included. someter Systems
1.3 Tension tests under conditions of rapid heating or rapid E177 Practice for Use of the Terms Precision and Bias in
strain rates are not included. ASTM Test Methods
1.4 The values stated in inch-pound units are to be regarded E220 Test Method for Calibration of Thermocouples By
as standard. The values given in parentheses are mathematical Comparison Techniques
conversions to SI units that are provided for information only E633 Guide for Use of Thermocouples in Creep and Stress-
and are not considered standard. Rupture Testing to 1800°F (1000°C) in Air
E691 Practice for Conducting an Interlaboratory Study to
1.5 This standard does not purport to address all of the Determine the Precision of a Test Method
safety concerns, if any, associated with its use. It is the
responsibility of the user of this standard to establish appro- 3. Terminology
priate safety, health, and environmental practices and deter-
3.1 Definitions of terms relating to tension testing which
mine the applicability of regulatory limitations prior to use.
appear in Terminology E6, apply to this test method. These
1.6 This international standard was developed in accor-
terms include alignment, axial strain, bending strain, gauge
dance with internationally recognized principles on standard-
length, elongation, elongation after fracture, extensometer
ization established in the Decision on Principles for the
system, necking, reduction of area, tensile strength, yield
Development of International Standards, Guides and Recom-
strength. In addition, the definitions of the following terms
mendations issued by the World Trade Organization Technical
relating to tension testing are included.
Barriers to Trade (TBT) Committee.
3.2 Definitions:
2. Referenced Documents 3.2.1 reduced section, n—the central portion of the speci-
men that has a cross section smaller than the gripped ends.
2.1 ASTM Standards:2
3.2.1.1 Discussion—The cross section is uniform within
E4 Practices for Force Verification of Testing Machines
prescribed tolerances.
E6 Terminology Relating to Methods of Mechanical Testing
3.2.2 length of the reduced section—the distance between
the tangent points of the fillets that bound the reduced section.
1
These test methods are under the jurisdiction of ASTM Committee E28 on 3.2.3 adjusted length of the reduced section—the length of
Mechanical Testing and are the direct responsibility of Subcommittee E28.04 on the reduced section plus an amount calculated to compensate
Uniaxial Testing. for strain in the fillet region.
Current edition approved Dec. 1, 2017. Published January 2018. Originally
approved in 1933. Last previous edition approved in 2009 as E21 – 09. DOI:
10.1520/E0021-17E01. 4. Significance and Use
2
For referenced ASTM standards, visit the ASTM website, www.astm.org, or
4.1 The elevated-temperature tension test gives a useful
contact ASTM Customer Service at service@astm.org. For Annual Book of ASTM
Standards volume information, refer to the standard’s Document Summary page on estimate of the ability of metals to withstand the application of
the ASTM website. applied tensile forces. Using established and conventional

*A Summary of Changes section appears at the end of this standard

Copyright © ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States
Copyright by ASTM Int'l (all rights reserved); Wed Jun 5 23:35:02 EDT 2019
1
Downloaded/printed by
Choong Guowei (TUV SUD AG) pursuant to License Agreement. No further reproductions authorized.
 E21 − 17ϵ1
relationships it can be used to give some indication of probable sitate use of a material different from that used during the
behavior under other simple states of stress, such as elevated-temperature test. See Practice E1012 for recom-
compression, shear, etc. The ductility values give a compara- mended methods for determining specimen alignment.
tive measure of the capacity of different materials to deform 5.1.2.3 Gripping devices and pull rods may oxidize, warp,
locally without cracking and thus to accommodate a local and creep with repeated use at elevated temperatures. Increased
stress concentration or overstress; however, quantitative rela- bending stresses may result. Therefore, grips and pull rods
tionships between tensile ductility and the effect of stress should be periodically retested for axiality and reworked when
concentrations at elevated temperature are not universally necessary.
valid. A similar comparative relationship exists between tensile 5.1.3 The testing machine shall be equipped with a means of
ductility and strain-controlled, low-cycle fatigue life under measuring and controlling either the strain rate or the rate of
simple states of stress. The results of these tension tests can be crosshead motion or both to meet the requirements in 9.6.
considered as only a questionable comparative measure of the 5.1.4 For high-temperature testing of materials that are
strength and ductility for service times of many hours. readily attacked by their environment (such as oxidation of
Therefore, the principal usefulness of the elevated-temperature metal in air), the specimen may be enclosed in a capsule so that
tension test is to assure that the tested material is similar to it can be tested in a vacuum or inert gas atmosphere. When
reference material when other measures such as chemical such equipment is used, the necessary corrections must be
composition and microstructure also show the two materials made to determine the actual forces seen by the specimen. For
are similar. instance, compensation must be made for differences in pres-
sures inside and outside of the capsule and for any variation in
5. Apparatus the forces applied to the specimen due to sealing ring friction,
5.1 Testing Machine: bellows or other features.
5.1.1 The accuracy of the testing machine shall be within
5.2 Heating Apparatus:
the permissible variation specified in Practices E4.
5.2.1 The apparatus for and method of heating the speci-
5.1.2 Precaution should be taken to assure that the force on
mens should provide the temperature control necessary to
the specimens is applied as axially as possible. Perfect axial
satisfy the requirements specified in 9.4.
alignment is difficult to obtain especially when the pull rods
and extensometer rods pass through packing at the ends of the 5.2.2 Heating shall be by an electric resistance or radiation
furnace. However, the machine and grips should be capable of furnace with the specimen in air at atmospheric pressure unless
loading a precisely made specimen so that the maximum other media are specifically agreed upon in advance.
bending strain does not exceed 10 % of the axial strain, when NOTE 2—The media in which the specimens are tested may have a
the calculations are based on strain readings taken at zero force considerable effect on the results of tests. This is particularly true when the
and at the lowest force for which the machine is being properties are influenced by oxidation or corrosion during the test.
qualified. 5.3 Temperature-Measuring Apparatus:
NOTE 1—This requirement is intended to limit the maximum contribu- 5.3.1 The method of temperature measurement must be
tion of the testing apparatus to the bending which occurs during a test. It sufficiently sensitive and reliable to ensure that the temperature
is recognized that even with qualified apparatus different tests may have of the specimen is within the limits specified in 9.4.4.
quite different percent bending strain due to chance orientation of a 5.3.2 Temperature should be measured with thermocouples
loosely fitted specimen, lack of symmetry of that particular specimen,
lateral force from furnace packing, and thermocouple wire, etc. The scant in conjunction with the appropriate temperature indicating
evidence available at this time3 indicates that the effect of bending strain instrumentation.
on test results is not sufficient, except in special cases, to require the
measurement of this quantity on each specimen tested. NOTE 3—Such measurements are subject to two types of error.
Thermocouple calibration and instrument measuring errors initially intro-
5.1.2.1 In testing of brittle material even a bending strain of duce uncertainty as to the exact temperature. Secondly both thermo-
10 % may result in lower strength than would be obtained with couples and measuring instruments may be subject to variation with time.
improved axiality. In these cases, measurements of bending Common errors encountered in the use of thermocouples to measure
temperatures include: calibration error, drift in calibration due to contami-
strain on the specimen to be tested may be specifically nation or deterioration with use, lead-wire error, error arising from method
requested and the permissible magnitude limited to a smaller of attachment to the specimen, direct radiation of heat to the bead,
value. heat-conduction along thermocouple wires, etc.
5.1.2.2 In general, equipment is not available for determin- 5.3.3 Temperature measurements should be made with ther-
ing maximum bending strain at elevated temperatures. The mocouples of known calibration. Representative thermo-
testing apparatus may be qualified by measurements of axiality couples should be calibrated from each lot of wires used for
made at room temperature using the assembled machine, pull making base-metal thermocouples. Except for relatively low
rods, and grips used in high temperature testing. The specimen temperatures of exposure, base-metal thermocouples are sub-
form should be the same as that used during the elevated- ject to error upon reuse, unless the depth of immersion and
temperature tests and designed so that only elastic strains occur temperature gradients of the initial exposure are reproduced.
throughout the reduced section. This requirement may neces- Consequently base-metal thermocouples should be verified by
the use of representative thermocouples and actual thermo-
3
Subcommittee E28.10 on Effect of Elevated Temperature on Properties requests couples used to measure specimen temperatures should not be
factual information on the effect of nonaxiality of loading on test results. verified at elevated temperatures. Base-metal thermocouples

Copyright by ASTM Int'l (all rights reserved); Wed Jun 5 23:35:02 EDT 2019
2
Downloaded/printed by
Choong Guowei (TUV SUD AG) pursuant to License Agreement. No further reproductions authorized.
 E21 − 17ϵ1
also should not be reused without clipping back to remove wire holders or crossheads may be recorded and used to determine
exposed to the hot zone and rewelding. Any reuse of base- strains corresponding to the 0.2 % offset yield strength. The
metal thermocouples after relatively low-temperature use with- value so obtained is of inferior accuracy and must be clearly
out this precaution should be accompanied by recalibration marked as “approximate yield strength.” The observed exten-
data demonstrating that calibration was not unduly affected by sion should be adjusted by the procedure described in 9.6.3 and
the conditions of exposure. 10.1.3.
5.3.3.1 Noble metal thermocouples are also subject to errors 5.4.5 The extensometer system shall include a means of
due to contamination, etc., and should be periodically annealed determining strain rate.
and verified. Thermocouples should be kept clean prior to 5.5 Room-Temperature Control—Unless the extensometer is
exposure and during use at elevated temperatures. known to be insensitive to ambient temperature changes, the
5.3.3.2 Measurement of the emf drift in thermocouples range of ambient temperature should not exceed 10°F (6°C)
during use is difficult. When drift is a problem during tests, a while the extensometer is attached. The testing machine should
method should be devised to check the readings of the not be exposed to perceptibly varying drafts.
thermocouples on the specimen during the test. For reliable
calibration of thermocouples after use the temperature gradient 6. Sampling
of the testing furnace must be reproduced during the recalibra-
tion. 6.1 Unless otherwise specified the following sampling pro-
5.3.4 Temperature-measuring, controlling, and recording in- cedures shall be followed:
struments should be verified periodically against a secondary 6.1.1 Samples of the material to provide test specimens
standard, such as a precision potentiometer and if necessary shall be taken from such locations as to be representative of the
re-calibrated. Lead-wire error should be checked with the lead lot from which it was taken.
wires in place as they normally are used. 6.1.2 Samples shall be taken from material in the final
condition (temper). One test shall be made on each lot.
5.4 Extensometer System: 6.1.3 A lot shall consist of all material from the same heat,
5.4.1 Practice E83, is recommended as a guide for selecting nominal size, and condition (temper).
the required sensitivity and accuracy of extensometers. For
determination of offset yield strength at 0.1 % or greater, a 7. Test Specimens and Sample
Class B-2 extensometer may be used. The extensometer should
meet the requirements of Practice E83 and should, in addition, 7.1 The size and shape of the test specimens should be
be tested to assure its accuracy when used in conjunction with based primarily on the requirements necessary to obtain
a furnace at elevated temperature. One such test is to measure representative samples of the material being investigated.
at elevated temperature the stress and strain in the elastic range 7.2 Unless otherwise specified, test specimens shall be
of a metal of known modulus of elasticity. Combinations of oriented such that the axis of the specimen is parallel to the
stress and temperature which will result in creep of the direction of fabrication, and located as follows:
specimen during the extensometer system evaluation should be 7.2.1 At the center for products 11⁄2 in. (38 mm) or less in
avoided. thickness, diameter, or distance between flats.
7.2.2 Midway from the center to the surface for products
NOTE 4—If an extensometer of Class B-2 or better is attached to the
reduced section, the slope of the stress-strain curve will usually be within over 11⁄2 in. (38 mm) in thickness, diameter, or distance
10 % of the modulus of elasticity. between flats.
5.4.2 Non-axiality of loading is usually sufficient to cause 7.3 Specimen configurations described in Test Methods
significant errors at small strains when strain is measured on E8/E8M, are generally suitable for tests at elevated tempera-
only one side of the specimen.4 Therefore, the extensometer tures; however, tighter dimensional tolerances are recom-
should be attached to and indicate strain on opposite sides of mended in 7.6. The particular specimen used should be mainly
the specimen. The reported strain should be the average of the governed by the requirements specified in 7.1. When the
strains on the two sides, either a mechanical or electrical dimensions of the material permit, except for sheet and strip,
average internal to the instrument or a numerical average of the gauge length of the specimens should have a circular cross
two separate readings. section. The largest diameter specimen consistent with that
5.4.3 When feasible the extensometer should be attached described in 7.1 should be used, except that the diameter need
directly to the reduced section. When necessary, other arrange- not be greater than 0.500 in. (12.7 mm). The ratio of gauge
ments (discussed in 9.6.3) may be used by prior agreement of length to diameter should be 4, as for the standard specimens
the parties concerned. For example, special arrangements may described in Test Methods E8/E8M. If different ratios are used,
be necessary in testing brittle materials where failure is apt to the specifics should be reported in the results.
be initiated at an extensometer knife edge. NOTE 5—Specimen size in itself has little effect on tensile properties
5.4.4 To attach the extensometer to miniature specimens provided the material is not subject to appreciable surface corrosion, lack
may be impractical. In this case, separation of the specimen of soundness, or orientation effects. A small number of grains in the
specimen cross section, or preferred orientation of grains due to fabrica-
tion conditions, can have a pronounced effect on the test results. When
4
Tishler, D. N., and Wells, C. H., “An Improved High-Temperature Extensom- corrosion is a factor in testing, the results do become a function of
eter ,” Materials Research and Standards , American Society for Testing and specimen size. Likewise, surface preparation of specimens, if affecting
Materials, MTRSA, Vol 6, No. 1, January 1966, pp. 20–22. results, becomes more important as the specimen size is reduced.

Copyright by ASTM Int'l (all rights reserved); Wed Jun 5 23:35:02 EDT 2019
3
Downloaded/printed by
Choong Guowei (TUV SUD AG) pursuant to License Agreement. No further reproductions authorized.
 E21 − 17ϵ1
7.4 Specimens of circular cross section should have Threads of the specimen should be concentric with this axis
threaded, shouldered, or other suitable ends for gripping which within the same tolerance. Other means of gripping should
will meet the requirements of 5.1.2. have comparable tolerances.
NOTE 6—Satisfactory axial alignment may be obtained with precisely 7.7 For cast-to-size specimens it may not be possible to
machined threaded ends. But at temperatures where oxidation and creep adhere to the diameter, straightness, and concentricity limita-
are readily apparent, precisely fitted threads are difficult to maintain and to tions of 7.6, but every effort should be made to approach these
separate after test. Practical considerations require the use of relatively
loose-fitting threads. Other gripping methods have been successfully
as closely as possible. If the specimen does not meet the
used.5,6 requirements specified in 7.6, the test report should so state.
The magnitude of the deviations should be reported.
7.5 For rectangular specimens some modifications of the
standard specimens described in Test Methods E8/E8M are 8. Calibration and Standardization
usually necessary to permit application of the force to the
specimen in the furnace with the axiality specified in 5.1.2. If 8.1 The following devices should be calibrated against
the material available is sufficient, the use of elongated standards traced to the National Institute of Standards and
shoulder ends to permit gripping outside the furnace is the Technology. Applicable ASTM methods are listed beside the
easiest method. When the length of the specimen is necessarily device.
restricted, several methods of gripping may be used as follows: Force-measuring system E4 and E74
Extensometer E83
7.5.1 A device that applies the force through a cylindrical ThermocouplesA E220
pin in each of the enlarged ends of the specimen. The pin holes Potentiometers
should be accurately centered on extensions of the centerline of Micrometers

the gauge section. Grips of this type can provide good axiality A
Melting point methods are also recommended for thermocouple calibration.
of loading.5
7.5.2 High-temperature sheet grips similar to those illus- 8.1.1 Axiality of the loading apparatus should be measured
trated in Test Methods E8/E8M and described as self-adjusting as described in 5.1.2.
grips. These have proven satisfactory for testing sheet materi- 8.2 Calibrations should be as frequent as is necessary to
als that cannot be tested satisfactorily in the usual type of assure that the errors in all tests do not exceed the permissible
wedge grips. variations listed in this test method. The maximum period
7.5.3 Extension tabs may be welded or brazed to the between calibrations of the testing machine shall be one year.
specimen shoulders and extended to grips outside the furnace. Instruments in either constant or nearly constant use should be
When these are used, care must be exercised to maintain calibrated more frequently; those used only occasionally
coaxiality of the centerlines of the extensions and the gauge should be calibrated before each use.
length. Any brazing or welding should be done in a jig or
fixture to maintain accurate alignment of the parts. Any 9. Procedure
machining should be done after brazing or welding. 9.1 Measurement of Cross-Sectional Area—Determine the
7.5.4 Grips that conform to and apply force against the minimum cross-sectional area of the reduced section as speci-
fillets at the ends of the reduced section. fied in 7.2 of Test Methods E8/E8M. In addition measure the
7.6 The diameter (or width) at the ends of the reduced largest diameter (or width) in the reduced section and compare
section should not be less than the diameter (or width) at the with the minimum value to determine whether the require-
center of the reduced section. It may be desirable to have the ments of 7.6 are satisfied.
diameter (or width) of the reduced section slightly smaller at 9.2 Measurement of Original Length:
the center than at the ends. This difference should not exceed 9.2.1 Unless otherwise specified, base all values for elon-
0.5 % of the diameter (or width). When specimens of this form gation on a gauge length equal to four diameters in the case of
are used to test brittle materials, failure may regularly occur at round specimens and four times the width in the case of
the fillets. In these cases, the center of the reduced section may rectangular specimens, the gauge length being punched or
be made smaller by a gradual taper from the ends and the scribed on the reduced section.
exception to the requirements above noted in the report.
Specimen surfaces shall be smooth and free from undercuts NOTE 7—Elongation values of specimens with rectangular cross sec-
tions cannot be compared unless all dimensions including the thickness
and scratches. Cold work introduced through machining or are equal. Therefore, an elongation specification should include the
handling can produce high residual stresses or other undesired specimen cross-sectional dimensions as well as the gauge length. Using a
effects and should be minimized. The axis of the reduced gauge length equal to 4.5 times the square root of the cross-sectional area
section should be straight within 60.5 % of the diameter. compensates somewhat for variations in specimen thickness but even this
does not result in the same value of elongation when specimens of the
same material are machined to different thicknesses and tested.7
5
Schmieder, A. K.,“ Measuring the Apparatus Contributions to Bending in 9.2.2 When testing metals of limited ductility gauge marks
Tension Specimens ,” Elevated Temperature Testing Problem Areas, ASTM STP punched or scribed on the reduced section may be undesirable
488, American Society for Testing and Materials, 1971, pp. 15–42
6
Penny, R. K., Ellison, E. G., and Webster, G. A.,“Specimen Alignment and
7
Strain Measurement in Axial Creep Tests,” Materials Research and Standards, Stickley, G. W., and Brownhill, D. J., “Elongation and Yield Strength of
American Society for Testing and Materials, MTRSA, Vol 6, No. 2, February 1966, Aluminum Alloys as Related to Gage Length and Offset,” Proceedings, American
pp. 76–84. Society for Testing and Materials, ASTEA, Vol 65, 1965, pp 597–616.

Copyright by ASTM Int'l (all rights reserved); Wed Jun 5 23:35:02 EDT 2019
4
Downloaded/printed by
Choong Guowei (TUV SUD AG) pursuant to License Agreement. No further reproductions authorized.
 E21 − 17ϵ1
because fracture may occur at the stress concentrations so difference between the indicated temperature and the nominal
caused. Then, place gauge marks on the shoulders or measure test temperature to exceed the following limits:
the over-all length of the specimen. Also measure the adjusted Up to and including 1800°F (1000°C) 5°F (±3°C)
length of the reduced section to the nearest 0.01 in. (0.2 mm) Above 1800°F (1000°C) 10°F (±6°C)
as described in 9.2.3. If a gauge length, other than that specified When testing at temperatures of a few hundred degrees,
in 9.2.1 is employed to measure elongation, describe the gauge internal heating due to plastic working may raise the tempera-
length in the report. In the case of acceptance tests, any ture of the specimen above the limits specified. In these cases
deviation from 9.2.1 must be agreed upon before testing. include the temperature at maximum force and the reason for
NOTE 8—The availability of flexible ceramic fiber cords for mounting the overshoot in the report.
of high temperature extensometers with high purity ceramic rods with 9.4.5 The term “indicated temperature” means the tempera-
chisel or vee-chisel ends, provides a good measure of ductility without ture that is indicated by the temperature measuring device
excessive damage to the gauge section caused by other types of exten- using good quality pyrometric practice.
someters or traditional punch or scribe marks. Damage to the rods from
specimen failure may be minimized through the use of spring loaded NOTE 10—It is recognized that true temperature may vary more than the
attachment fixtures. Non contact extensometers may also be used for this indicated temperature. The permissible indicated temperature variations in
purpose. 9.4.4 are not to be construed as minimizing the importance of good
9.2.3 When the extensometer is to be attached to the pyrometric practice and precise temperature control. All laboratories
should keep both indicated and true temperature variations as small as
specimen shoulders, measure the adjusted length of the re- practicable. It is well recognized, in view of the extreme dependency of
duced section between points on the two fillets where the strength of materials on temperature, that close temperature control is
diameter (or width) is 1.05 times the diameter (or width) of the necessary. The limits prescribed represent ranges which are common
reduced section. The strain rate and offset yield calculations are practice.
based on this dimension (see 9.6.3, 10.1.2, and 10.3). 9.4.6 Temperature overshoots during heating shall not ex-
NOTE 9—In the yield region, stress is approximately proportional to ceed the above limits, unless agreed upon by the customer and
offset strain to a power which usually lies in the range from zero to 0.20. the supplier. The heating characteristics of the furnace and the
For specimens of circular cross section the above value of adjusted length temperature control system should be studied to determine the
of the reduced section was found by calculation to give an error in yield power input, temperature set point, proportioning control
stress of less than 1⁄2 % within this range of exponents and for fillet radii adjustment, and control-thermocouple placement necessary to
ranging from 1⁄2 to 1 times the diameter of the reduced section. The
method of calculation was similar to that used by Thomas and Carlson.8 limit transient temperature overshoots. It may be desirable to
stabilize the furnace at a temperature from 6 to 28°C (10 to 50
9.3 Cleaning Specimen—Wash carefully the reduced section
°F) below the nominal test temperature before making the final
and those parts of the specimen which contact the grips in clean
adjustments. If allowed, overshoots shall be reported, with
alcohol, acetone, or other suitable solvent that will not affect
details of magnitude and duration.
the metal being tested.
9.4.7 The time of holding at temperature prior to the start of
9.4 Temperature Control: the test should be governed by the time necessary to ensure that
9.4.1 Form the thermocouple bead in accordance with the specimen has reached equilibrium and that the temperature
Guide E633. can be maintained within the limits specified in 9.4.4. Unless
9.4.2 In attaching thermocouples to a specimen, the junction otherwise specified this time should not be less than 20
must be kept in intimate contact with the specimen and minutes. Report the time to attain test temperature and the time
shielded from radiation. Shielding may be omitted if, for a at temperature before testing.
particular furnace and test temperature, the difference in
indicated temperature from an unshielded bead and a bead 9.5 Connecting Specimen to the Machine—Take care not to
inserted in a hole in the specimen has been shown to be less introduce nonaxial forces while installing the specimen. For
than one half the variation listed in 9.4.4. The bead should be example, threaded connections should not be turned to the end
as small as possible and there should be no shorting of the of the threads or “bottomed.” If threads are loosely fitted,
circuit (such as could occur from twisted wires behind the lightly apply force to the specimen string and manually move
bead). Ceramic insulators should be used on the thermocouples it in the transverse direction until the force drops to its
in the hot zone. If some other electrical insulation material is minimum value before testing. If packing is used to seal the
used in the hot zone, it should be determined that the electrical furnace, it must not be so tight that the extensometer arms or
insulating properties are maintained at higher temperatures. pull rods are displaced or their movement restricted.
9.4.3 When the length of the reduced section is less than 2 9.6 Strain Measurement and Strain Rate:
in. (50 mm), attach at least two thermocouples to the specimen, 9.6.1 During yield strength determination, maintain the
one near each end of the reduced section. For reduced sections strain rate in the reduced parallel section of the test specimen
greater-than or equal to 2 in. (50 mm) add a third thermocouple at 0.005 6 0.002 mm/mm/min. After yield strength
near the center of the reduced section. determination, increase the rate of crosshead motion to 0.05 6
9.4.4 For the duration of the test, (defined as the time from 0.01 times the length of the reduced parallel section (A) of any
the application of force until fracture), do not permit the of the specimen configurations described in Test Methods
E8/E8M per minute.
8
Thomas, J. M., and Carlson, J. F., “Errors in Deformation Measurements for NOTE 11—The tensile properties of material tested at elevated tempera-
Elevated Temperature Tension Tests,” ASTM Bulletin, ASTM, May 1955, pp. 47–51. ture are, in general, affected by the rate of deformation. It is therefore

Copyright by ASTM Int'l (all rights reserved); Wed Jun 5 23:35:02 EDT 2019
5
Downloaded/printed by
Choong Guowei (TUV SUD AG) pursuant to License Agreement. No further reproductions authorized.
 E21 − 17ϵ1
important that this rate be controlled and reported. of coupling or crosshead separation. In order to adjust for the
9.6.1.1 In cases where it is desired to establish the entire extension which occurs outside the reduced section, two
engineering stress strain curve of a specimen, the speed of specimens must be tested, one with the standard reduced
testing may be maintained at 0.005 mm/mm/min throughout section and the other, a shortened specimen, with similar grip
the entire test ends and shoulders but without fillets and reduced section. The
latter need only be tested to the force level required to yield the
NOTE 12—Maintaining the strain rate in 9.6.1 may require the use of a former (see 10.1.3).
suitable extensometer system that has a calibrated range to record the
entire test and will not be damaged by the failure. Caution is advised since 9.6.7 To allow for the elastic strain in the machine parts,
control problems can occur if an extensometer slips or behaves in an pull rods and grips, set the rate of crosshead motion during
unexpected manner. yielding at or slightly higher than the upper limit of the
9.6.1.2 If it has been established that the crosshead speed recommended range, that is 0.007 times the adjusted length of
remains constant within the tolerance above, the extensometer the reduced section per minute. At strains exceeding that
and strain rate indicator may be used to set a strain rate of 0.05 corresponding to the yield strength of the material being tested,
6 0.01 mm/mm/min after yield strength determination. To apply the method of 9.6.2.
protect it from damage, the sensing element of the extensom- 9.7 Recording Maximum Force—If an autographic recorder
eter may be removed before maximum force is reached. of force and extension is used, continue the recording of force
NOTE 13—Even with constant crosshead speed, the strain rate in the
after the sensing element of the extensometer is removed. In
specimen may still vary. Before maximum force it will be less than the any case observe the maximum force and record manually.
nominal rate due to the elasticity of the machine and grips and the 9.8 Measurements of Specimen After Test:
progressive elongation of the specimen. After maximum force it will be
greater than the nominal rate due to nonuniform strain during necking. 9.8.1 For measuring elongation, fit the ends of the fractured
Available experimental evidence does not justify the added complexity of specimen together carefully and measure the distance between
maintaining a constant strain rate throughout the post-yield stages of a gauge marks or the over-all length to the nearest 0.01 in. (0.2
tension test. mm) at room temperature.
9.6.2 When yield strength determination is not required, an 9.8.2 If any part of the fracture surface extends beyond the
extensometer need not be used. The rate of crosshead motion middle half of the reduced section, the elongation value
shall be maintained at 0.05 6 0.01 times the adjusted length of obtained may not be representative of the material. In the case
the reduced parallel section, per minute throughout the test. of an acceptance test, if the elongation meets the minimum
9.6.3 When yield strength determinations are required, ob- requirements specified, no further testing is required; but if the
servations of force and extension during loading and through elongation is less than the specified minimum the test may be
yield are necessary. The following three means of making these discarded and a retest made.
observations are acceptable: 9.8.3 For measuring reduction of area of specimens of
9.6.3.1 For specimens of normal size and ductility, attach an circular cross section, fit the ends of the fractured specimen
extensometer to the reduced section. together carefully and measure the minimum diameter to the
9.6.3.2 When metals of limited ductility are tested attach the nearest 0.01 in. (0.2 mm) at room temperature. If the fracture
extensometer to the specimen shoulders. cross section is not round, make sufficient measurements to
9.6.3.3 For miniature specimens, measure coupling or establish the cross-sectional area at fracture. If the fracture
crosshead separation to determine an approximate yield occurs at a fillet or gauge mark the reduction of area may not
strength. be representative of the material. In the case of an acceptance
9.6.4 When the extensometer is attached to the reduced test, if the reduction of area meets the specified minimum, no
section, a strain rate indicator or controller shall be used to further testing is required, but if the reduction of area is less
maintain a rate of 0.005 6 0.002 mm/mm/min through the than the specified minimum the test may be discarded and a
yielding range except within the elastic range where smaller retest made.
strain rates are permissible. 10. Calculations
NOTE 14—With conventional testing machines, the strain rate often 10.1 Yield Strength:
cannot be controlled closely if the material yields in a relatively sudden 10.1.1 Unless otherwise specified determine the yield
manner. In such cases the speed control should be preset to the rate which
strength reported at an offset of 0.2 % as described in Test
experience shows will result in the specified strain rate at the force
corresponding to the yield stress. This will usually be at a strain rate which Methods E8/E8M.
will result from a rate of crosshead motion of 0.005 times the adjusted NOTE 15—The accurate measurement of proportional limit and offset
length of the reduced section, per minute. yield strength of 0.02 % or less is extremely difficult at elevated
9.6.5 When the extensometer is attached to the specimen temperatures. Even though the extensometer has the required accuracy
shoulders, use the adjusted length of the reduced section to and sensitivity during room-temperature calibration, this is not assurance
that the strain measurements during the elevated-temperature tension test
calculate the setting of the strain-rate indicator or pacer during will have equal accuracy. High temperature at the attachment points and
yielding. Otherwise the procedure is the same as that described extensometer rods passing through the furnace packing will probably
above. reduce the accuracy significantly. Therefore, the determination of propor-
9.6.6 When the stock size requires use of specimens less tional limit and offset yield strength of 0.02 % or less is not recommended.
than 0.25 in. (6.25 mm) in diameter, the approximate 0.2 10.1.2 If the extensometer must be attached to the specimen
percent offset yield strength may be determined from a record shoulders, base the offset extension (inches or millimeters) on

Copyright by ASTM Int'l (all rights reserved); Wed Jun 5 23:35:02 EDT 2019
6
Downloaded/printed by
Choong Guowei (TUV SUD AG) pursuant to License Agreement. No further reproductions authorized.
 E21 − 17ϵ1
the adjusted length of the reduced section, that is, 0.002 times 11. Report
the adjusted length of the reduced section for 0.2 percent offset 11.1 The report shall include the following:
yield strength and corresponding values for other yield 11.1.1 Description of material tested, including method of
strengths. manufacture, type and size of product, and other pertinent
10.1.3 If coupling or crosshead separation are recorded processing information, as well as heat treatment,
adjust the observed extension in the following two steps. First, microstructure, and chemical composition.
to compensate for machine elasticity, grip distortion and 11.1.2 Specimen dimensions, including cross-sectional
shoulder strain, reduce values for the standard specimen by the dimensions, fillet radius, length of reduced section, adjusted
values for the shortened specimen (9.6.6) at corresponding length of reduced section (if used), type of end connection, and
applied forces. Second, treat this adjusted strain datum by the whether machined, partially machined, or as cast.
method of 10.1.2. Report only yield strengths with offsets of 11.1.3 Temperature of test.
0.2 % or more and label these “approximate yield strength.” 11.1.4 Strain rate during yielding and strain rate after
10.2 Tensile Strength—Calculate the tensile strength by yielding.
dividing the maximum force, during a test carried to fracture, 11.1.5 Yield strength, if required, and the method used to
by the original minimum cross-sectional area of the reduced determine the yield strength.
section. 11.1.5.1 When one or more values of yield strength are
required, the amount of the offsets should be shown with the
10.3 Elongation: numerical values.
10.3.1 When the gauge length is marked on the reduced 11.1.5.2 If the extensometer was attached to the specimen
section having a nominally uniform cross-sectional area, the shoulders, this fact should be stated in a footnote to the values.
elongation is equal to the gauge length after fracture minus the 11.1.5.3 If an extensometer was not attached directly to the
original gauge length, the difference expressed as a percentage specimen, the value should be listed as “approximate yield
of the original gauge length. If the gauge length includes fillets, strength (offset = 0.2 %).’’
shoulders, threads, etc., the change in gauge length is expressed 11.1.6 Tensile strength.
as a percentage of the adjusted length of the reduced section. 11.1.7 Elongation and gauge length. If elongation was
10.3.2 A method that can sometimes be used when there is measured from gauge marks not on the reduced section this
autographic recording of strain up to the moment of fracture, is fact should be included in the designation of the quantity, for
to read the elongation as strain offset from the initial, linear, example“ elongation from shoulder measurements’’ or “elon-
loading line. This can be useful in the case of materials of very gation from over-all measurements.’’ If elongation was mea-
low ductility. Since these values are usually lower than those sured from the extensometer record at fracture instead of after
measured from the broken specimen, the method of measure- fracture, this should be noted.
ment should be stated with the results. 11.1.8 Reduction of area for specimens with circular cross
section.
10.4 Reduction of Area—Reduction of area is equal to the 11.1.9 Time to attain test temperature and time at tempera-
minimum cross-sectional area of the reduced section before ture before testing.
testing minus the minimum cross-sectional area of the reduced 11.1.10 Other special conditions, such as nonstandard atmo-
section after testing, the difference expressed as a percentage of sphere and heating methods, exceptions to required dimen-
the area before testing. Reduction of area is reported only for sional accuracy and axiality of loading, amount and duration of
specimens of circular cross section. temperature overshoot.
10.5 Rounding—Unless otherwise specified, for purposes of 11.1.11 Location and description of fracture. The descrip-
determining compliance with specified limits, observed or tion should include any defects, evidence of corrosion, and
calculated values shall be rounded as indicated below, in type of fracture (such as cup and cone, brittle, shear).
accordance with the Rounding Method of Practice E29 as 11.1.12 Identification of equipment used including make
follows: and capacity of testing machine, make and class of
extensometer, make and size of furnace, type of temperature
Rounded Unit for Observed or
Quantity Measured Calculated Value controller, and description of thermocouples including
Tensile or Yield Strength Nearest 500 psi (3.5 MPa) material, wire size, attachment, technique and shielding.
Elongation or Reduction of Area Nearest 0.5 % 11.1.13 Name of tester and date of test.

Copyright by ASTM Int'l (all rights reserved); Wed Jun 5 23:35:02 EDT 2019
7
Downloaded/printed by
Choong Guowei (TUV SUD AG) pursuant to License Agreement. No further reproductions authorized.
 E21 − 17ϵ1
12. Precision and Bias Coefficient of Variation, % Tensile Properties at 1100°F (593°C)
Elongation
9 Yield
12.1 Precision—An inter-laboratory test program gave the Tensile
Strength
gauge Reduction of
following values for coefficients of variation for the most Strength length = 4 Area
offset = 0.2 %
diameters
commonly measured tensile properties:10 CV %r 1.7 4.0 3.1 2.2
Coefficient of Variation, % Tensile Properties at 600°F (316°C) CV %R 4.4 4.4 11.1 4.5
Elongation CV %r = repeatability coefficent of variation in percent within a laboratory.
Yield CV %R = repeatability coefficient of variation in percent between laboratories.
Tensile gauge Reduction of
Strength
Strength
offset = 0.2 %
length = 4 Area 12.1.1 The values shown are the averages from tests on four
diameters
CV %r 1.0 3.0 3.8 4.6
frequently tested metals at two test temperatures, selected to
CV %R 1.4 5.1 8.2 4.9 include most of the normal range for each property listed
CV %r = repeatability coefficent of variation in percent within a laboratory. above. When these materials are compared, a large difference
CV %R = repeatability coefficient of variation in percent between laboratories.
in coefficient of variation is found. Therefore, the values above
should not be used to judge whether the difference between
duplicate tests of a specific material is larger than expected.
The values are provided to allow potential users of this test
method to assess, in general terms, its usefulness for a
proposed application.
9
Supporting data are available from ASTM Headquarters. Request RR:E28-
12.2 Bias—The procedures in Test Methods E21 for mea-
1015. suring tensile properties have no bias because these properties
10
For further information, see Practice E177 and Practice E691. can only be defined in terms of a test method.

SUMMARY OF CHANGES

Committee E28 has identified the location of selected changes to this standard since the last issue (E21–09)
that may impact the use of this standard.

(1) Section 3—revised to use terms in Terminology E6, add (3) 9.6—Clarification of strain rates.
“adjusted length of the reduced section.” (4) Editorial language clarifications were made throughout.
(2) Section 5—Mandatory language for strain rate.

ASTM International takes no position respecting the validity of any patent rights asserted in connection with any item mentioned
in this standard. Users of this standard are expressly advised that determination of the validity of any such patent rights, and the risk
of infringement of such rights, are entirely their own responsibility.

This standard is subject to revision at any time by the responsible technical committee and must be reviewed every five years and
if not revised, either reapproved or withdrawn. Your comments are invited either for revision of this standard or for additional standards
and should be addressed to ASTM International Headquarters. Your comments will receive careful consideration at a meeting of the
responsible technical committee, which you may attend. If you feel that your comments have not received a fair hearing you should
make your views known to the ASTM Committee on Standards, at the address shown below.

This standard is copyrighted by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959,
United States. Individual reprints (single or multiple copies) of this standard may be obtained by contacting ASTM at the above
address or at 610-832-9585 (phone), 610-832-9555 (fax), or service@astm.org (e-mail); or through the ASTM website
(www.astm.org). Permission rights to photocopy the standard may also be secured from the Copyright Clearance Center, 222
Rosewood Drive, Danvers, MA 01923, Tel: (978) 646-2600; http://www.copyright.com/

Copyright by ASTM Int'l (all rights reserved); Wed Jun 5 23:35:02 EDT 2019
8
Downloaded/printed by
Choong Guowei (TUV SUD AG) pursuant to License Agreement. No further reproductions authorized.