ASTM F1467-18

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: F1467 − 18
Standard Guide for
Use of an X-Ray Tester ('10 keV Photons) in Ionizing
Radiation Effects Testing of Semiconductor Devices and
Microcircuits
This standard is issued under the fixed designation F1467; 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.
1. Scope tions is used to provide an estimate of the ratio (number-of-
holes-cobalt-60)/(number-of-holes-X-ray). Several cases are
1.1 This guide covers recommended procedures for the use
defined where the differences in the effects caused by X-rays
of X-ray testers (that is, sources with a photon spectrum having
and cobalt-60 gammas are expected to be small. Other cases
≈10 keV mean photon energy and ≈50 keV maximum energy)
where the differences could potentially be as great as a factor
in testing semiconductor discrete devices and integrated cir-
of four are described.
cuits for effects from ionizing radiation.
1.8 It should be recognized that neither X-ray testers nor
1.2 The X-ray tester may be appropriate for investigating
cobalt-60 gamma sources will provide, in general, an accurate
the susceptibility of wafer level or delidded microelectronic
simulation of a specified system radiation environment. The
devices to ionizing radiation effects. It is not appropriate for
use of either test source will require extrapolation to the effects
investigating other radiation-induced effects such as single-
to be expected from the specified radiation environment. In this
event effects (SEE) or effects due to displacement damage.
guide, we discuss the differences between X-ray tester and
1.3 This guide focuses on radiation effects in metal oxide
cobalt-60 gamma effects. This discussion should be useful as
semiconductor (MOS) circuit elements, either designed (as in
background to the problem of extrapolation to effects expected
MOS transistors) or parasitic (as in parasitic MOS elements in
from a different radiation environment. However, the process
bipolar transistors).
of extrapolation to the expected real environment is treated
1.4 Information is given about appropriate comparison of elsewhere (1, 2).
ionizing radiation hardness results obtained with an X-ray
1.9 The time scale of an X-ray irradiation and measurement
tester to those results obtained with cobalt-60 gamma irradia-
may be much different than the irradiation time in the expected
tion. Several differences in radiation-induced effects caused by
device application. Information on time-dependent effects is
differences in the photon energies of the X-ray and cobalt-60
given.
gamma sources are evaluated. Quantitative estimates of the
1.10 Possible lateral spreading of the collimated X-ray
magnitude of these differences in effects, and other factors that
beam beyond the desired irradiated region on a wafer is also
should be considered in setting up test protocols, are presented.
discussed.
1.5 If a 10-keV X-ray tester is to be used for qualification
1.11 Information is given about recommended experimental
testing or lot acceptance testing, it is recommended that such
methodology, dosimetry, and data interpretation.
tests be supported by cross checking with cobalt-60 gamma
irradiations. 1.12 Radiation testing of semiconductor devices may pro-
duce severe degradation of the electrical parameters of irradi-
1.6 Comparisons of ionizing radiation hardness results ob-
ated devices and should therefore be considered a destructive
tained with an X-ray tester with results obtained with a
test.
LINAC, with protons, etc. are outside the scope of this guide.
1.13 The values stated in SI units are to be regarded as
1.7 Current understanding of the differences between the
standard. No other units of measurement are included in this
physical effects caused by X-ray and cobalt-60 gamma irradia-
standard.
1.14 This standard does not purport to address all of the
safety concerns, if any, associated with its use. It is the
This guide is under the jurisdiction of ASTM Committee E10 on Nuclear
Technology and Applications and is the direct responsibility of Subcommittee
E10.07 on Radiation Dosimetry for Radiation Effects on Materials and Devices.
Current edition approved March 1, 2018. Published April 2018. Originally
approved in 1993. Last previous edition approved in 2011 as F1467 – 11. DOI: The boldface numbers in parentheses refer to the list of references at the end of
10.1520/F1467-18. this guide.
Copyright © ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States
F1467 − 18
NOTE 2—The SI unit for absorbed dose is the gray (Gy), defined as one
responsibility of the user of this standard to establish appro-
J/kg. The commonly used unit, the rad (radiation absorbed dose), is
priate safety, health, and environmental practices and deter-
defined in terms of the SI units by 1 rad = 0.01 Gy. (For additional
mine the applicability of regulatory limitations prior to use.
information on calculation of absorbed dose see Practice E666.)
1.15 This international standard was developed in accor-
3.1.4 equilibrium absorbed dose, n—absorbed dose at some
dance with internationally recognized principles on standard-
incremental volume within the material in which the condition
ization established in the Decision on Principles for the
of electron equilibrium (the energies, number, and direction of
Development of International Standards, Guides and Recom-
charged particles induced by the radiation are constant
mendations issued by the World Trade Organization Technical
throughout the volume) exists (see Terminology E170).
Barriers to Trade (TBT) Committee.
3.1.4.1 Discussion—For practical purposes the equilibrium
absorbed dose is the absorbed dose value that exists in a
2. Referenced Documents
material at a distance in excess of a minimum distance from
2.1 ASTM Standards:
any interface with another material. This minimum distance
E170 Terminology Relating to Radiation Measurements and
being greater than the range of the maximum energy secondary
Dosimetry
electrons generated by the incident photons.
E666 Practice for Calculating Absorbed Dose From Gamma
3.1.5 ionizing radiation effects, n—the changes in the elec-
or X Radiation
trical parameters of a microelectronic device resulting from
E668 Practice for Application of Thermoluminescence-
radiation-induced trapped charge. These are also sometimes
Dosimetry (TLD) Systems for Determining Absorbed
referred to as ‘total dose effects.’
Dose in Radiation-Hardness Testing of Electronic Devices
E1249 Practice for Minimizing Dosimetry Errors in Radia- 3.1.6 time dependent effects, n—the change in electrical
tion Hardness Testing of Silicon Electronic Devices Using
parameters caused by the formation and annealing of radiation-
Co-60 Sources induced electrical charge during and after irradiation.
E1894 Guide for Selecting Dosimetry Systems for Applica-
4. Significance and Use
tion in Pulsed X-Ray Sources
2.2 International Commission on Radiation Units and Mea- 4.1 Electronic circuits used in many space, military and
surements Reports: nuclear power systems may be exposed to various levels of
ICRU Report 33—Quantities and Units for Use in Radiation ionizing radiation dose. It is essential for the design and
fabrication of such circuits that test methods be available that
Protection
can determine the vulnerability or hardness (measure of
2.3 United States Department of Defense Standards:
nonvulnerability) of components to be used in such systems.
MIL-STD-883, Method 1019, Ionizing Radiation (Total
Dose) Test Method
4.2 Manufacturers are currently selling semiconductor parts
with guaranteed hardness ratings, and the military specification
3. Terminology
system is being expanded to cover hardness specification for
3.1 Definitions:
parts. Therefore test methods and guides are required to
3.1.1 absorbed-dose enhancement, n—increase (or de-
standardize qualification testing.
crease) in the absorbed dose (as compared with the equilibrium
4.3 Use of low energy (≈10 keV) X-ray sources has been
absorbed dose) at a point in a material of interest; this can be
examined as an alternative to cobalt-60 for the ionizing
expected to occur near an interface with a material of higher or
radiation effects testing of microelectronic devices (3, 4, 5, 6).
lower atomic number.
The goal of this guide is to provide background information
3.1.2 average absorbed dose, n—mass weighted mean of
and guidance for such use where appropriate.
the absorbed dose over a region of interest.
NOTE 3—Cobalt-60—The most commonly used source of ionizing
3.1.3 average absorbed-dose enhancement factor, n—ratio
radiation for ionizing radiation (“total dose”) testing is cobalt-60. Gamma
of the average absorbed dose in a region of interest to the
rays with energies of 1.17 and 1.33 MeV are the primary ionizing radiation
equilibrium absorbed dose. emitted by cobalt-60. In exposures using cobalt-60 sources, test specimens
must be enclosed in a lead-aluminum container to minimize dose-
NOTE 1—For a description of the necessary conditions for measuring
enhancement effects caused by low-energy scattered radiation (unless it
equilibrium absorbed dose see the term ‘charged particle equilibrium’ in
has been demonstrated that these effects are negligible). For this lead-
Terminology E170 which provides definitions and descriptions of other
aluminum container, a minimum of 1.5 mm of lead surrounding an inner
applicable terms of this guide. In addition, definitions appropriate to the
shield of 0.7 to 1.0 mm of aluminum is required. (See 8.2.2.2 and Practice
subject of this guide may be found in ICRU Report 33.
E1249.)
4.4 The X-ray tester has proven to be a useful ionizing
For referenced ASTM standards, visit the ASTM website, www.astm.org, or
radiation effects testing tool because:
contact ASTM Customer Service at service@astm.org. For Annual Book of ASTM
4.4.1 It offers a relatively high dose rate, in comparison to
Standards volume information, refer to the standard’s Document Summary page on
most cobalt-60 sources, thus offering reduced testing time.
the ASTM website.
4.4.2 The radiation is of sufficiently low energy that it can
Available from International Commission on Radiation Units and Measure-
ments (ICRU), 7910 Woodmont Ave., Suite 400, Bethesda, MD 20841-3095,
be readily collimated. As a result, it is possible to irradiate a
http://www.icru.org.
single device on a wafer.
Available from Standardization Documents Order Desk, DODSSP, Bldg. 4,
4.4.3 Radiation safety issues are more easily managed with
Section D, 700 Robbins Ave., Philadelphia, PA 19111-5098, http://
dodssp.daps.dla.mil. an X-ray irradiator than with a cobalt-60 source. This is due
F1467 − 18
both to the relatively low energy of the photons and due to the and after the irradiation at which the measurement takes place
fact that the X-ray source can easily be turned off. (see X1.7 for further detail).
4.4.4 X-ray facilities are frequently less costly than compa-
NOTE 5—The dose rates used for X-ray testing are frequently much
rable cobalt-60 facilities.
higher than those used for cobalt-60 testing. For example, cobalt-60
testing is specified by Military Test Method 1019 to be in the range of 0.5
4.5 The principal radiation-induced effects discussed in this
to 3 Gy(Si)/s (50 to 300 rads/(Si)/s). For comparison, X-ray testing is
guide (energy deposition, absorbed-dose enhancement,
commonly carried out in the range of 2 to 30 Gy(Si)/s (200 to 3000
electron-hole recombination) (see Appendix X1) will remain
rads(Si)/s).
approximately the same when process changes are made to
5.4 Handling—As in any other type of testing, care must be
improve the performance of ionizing radiation hardness of a
taken in handling the parts. This especially applies to parts that
part that is being produced. This is the case as long as the
are susceptible to electrostatic discharge damage.
thicknesses and compositions of the device layers are substan-
tially unchanged. As a result of this insensitivity to process
6. Apparatus
variables, a 10-keV X-ray tester is expected to be an excellent
6.1 X-Ray Tester—A suitable X-ray tester (see Ref (3))
apparatus for process improvement and control.
consists of the following components:
4.6 Several published reports have indicated success in
6.1.1 Power Supply—The power supply typically supplies
intercomparing X-ray and cobalt-60 gamma irradiations using
10 to 100 mA at 25 to 60 keV (constant potential) to the X-ray
corrections for dose enhancement and for electron-hole recom-
tube.
bination. Other reports have indicated that the present under-
6.1.2 X-Ray Tube—In a typical commercial X-ray tube a
standing of the physical effects is not adequate to explain
partially focused beam of electrons strikes a water-cooled
experimental results. As a result, it is not fully certain that the
metal target. The target material most commonly used for
differences between the effects of X-ray and cobalt-60 gamma ionizing radiation effects testing is tungsten, though some work
irradiation are adequately understood at this time. (See 8.2.1
has been done using a copper target. X-ray tubes are limited by
and Appendix X2.) Because of this possible failure of under- the power they can dissipate. A maximum power of 3.5 kW is
standing of the photon energy dependence of radiation effects,
typical.
if a 10-keV X-ray tester is to be used for qualification testing 6.1.3 Collimator—A collimator is used to limit the region
or lot acceptance testing, it is recommended that such tests
on a wafer which is irradiated. A typical collimator is con-
should be supported by cross checking with cobalt-60 gamma structed of 0.0025 cm of tantalum.
irradiations. For additional information on such comparison,
6.1.4 Filter—A filter is used to remove the low-energy
see X2.2.4. photons produced by the X-ray tube. A typical filter is 0.0127
cm of aluminum.
4.7 Because of the limited penetration of 10-keV photons,
6.1.5 Dosimeter—A dosimetric system is required to mea-
ionizing radiation effects testing must normally be performed
sure the dose delivered by the X-ray tube (see Guide E1894).
on unpackaged devices (for example, at wafer level) or on
delidded devices.
NOTE 6—X-ray testers typically use a calibrated diode to measure the
dose delivered by the X-ray tube. These typically provide absorbed dose
in rads(Si).
5. Interferences
6.2 Spectrum—The ionizing radiation effects produced in
5.1 Absorbed-Dose Enhancement—Absorbed-dose en-
microelectronic devices exposed to X-ray irradiation are some-
hancement effects (see 8.2.1 and X1.3) can significantly
what dependent upon the incident X-ray spectrum. As a result,
complicate the determination of the absorbed dose in the region
appropriate steps shall be taken to maintain an appropriate and
of interest within the device under test. In the photon energy
reproducible X-ray spectrum.
range of the X-ray tester, these effects should be expected when
NOTE 7—The aim is to produce a spectrum whose effective energy is
there are regions of quite different atomic number within
peaked in the 5 to 15 keV photon energy region. This is accomplished in
hundreds of nanometres of the region of interest in the device
three ways. First, a large fraction of the energy output of the X-ray tube
under test.
is in the tungsten L emission lines. Second, some of the low-energy output
of the tube is absorbed by a filter prior to its incidence on the device under
NOTE 4—An example of a case where significant absorbed dose
test. Third, the high-energy output of the tube is only slightly absorbed in
enhancement effects should be expected is a device with a tantalum
the sensitive regions of device under test and thus has only a small effect
silicide metallization within 200 nm of the SiO gate oxide.
on the device. (See X1.2 for further detail.)
5.2 Electron-Hole Recombination—Once the absorbed dose
6.2.1 Control of Spectrum—The following steps shall be
in the sensitive region of the device under test is determined,
taken to insure adequate control of the X-ray spectrum:
interpretation of the effects of this dose can be complicated by
6.2.1.1 Anode Material—Unless otherwise specified, the
electron-hole recombination (see 8.2.1 and X1.5).
X-ray spectrum shall be produced by a tungsten target X-ray
5.3 Time-Dependent Effects—The charge in device oxides tube.
and at silicon-oxide interfaces produced by irradiation may 6.2.1.2 Anode Bias—Unless otherwise specified, the X-ray
change with time. Such changes take place both during and tube producing the X-ray spectrum shall be operated at a
after irradiation. Because of this, the results of electrical constant potential no lower than 40 kV nor higher than 60 kV.
measurements corresponding to a given absorbed dose can be 6.2.1.3 Spectrum Filtration—Unless otherwise specified,
highly dependent upon the dose rate and upon the time during the X-ray spectrum shall be filtered by 0.0127 cm of aluminum
F1467 − 18
prior to its incidence on the device under test. Further filtration 7.1.1.2 Dosimeter system to be used,
of the X-ray spectrum by additional intervening layers or by 7.1.1.3 Irradiation geometry to be used,
the device under test itself is to be minimized. 7.1.1.4 Devices to be tested, and
7.1.1.5 Parameters to be tested, including bias conditions
NOTE 8—Note that the X-ray spectrum is also filtered by the beryllium
and required accuracy.
window of the X-ray tube and by ;15 cm of air.
7.1.2 The test plan may also include a required sequence of
NOTE 9—For irradiation of Si to SiO based microelectronic devices
which are unpackaged, or packaged but delidded, filtration of the X-ray
actions for the test. A suggested sequence for the test is as
spectrum by the device under test is not expected to have a significant
follows:
effect (see X1.2 for further detail).
7.1.2.1 Prepare bias fixtures, test circuits, and test programs.
6.2.2 Determination of Spectrum—Generally, when using
7.1.2.2 Perform preliminary dosimetry if such measure-
the X-ray tester for ionizing radiation hardness testing, it is not
ments are not available.
necessary to have a detailed knowledge of the X-ray spectrum.
7.1.2.3 Make pre-irradiation electrical measurements.
Where it is necessary to know the spectrum, data exist in the
7.1.2.4 Bias the parts properly and irradiate them to the first
literature for some important cases. For unusual cases, experi-
radiation level.
mental and computational means exist to determine the spec-
7.1.2.5 Perform post-irradiation electrical measurements.
trum (see X1.2 for additional detail).
7.1.2.6 Irradiate the parts to the next level, if more than one
radiation level is required.
NOTE 10—If a thermoluminescent dosimeter (TLD) is used as a
dosimeter, it is necessary to know the spectrum. This is because the 7.1.2.7 Repeat 7.1.2.5 and 7.1.2.6 until all required levels
spectrum of the X-ray tester is substantially attenuated in passing through
have been achieved.
a TLD. For further information on the spectrum see X1.2. Given a
7.2 Device Bias:
spectrum, a dose versus depth correction can be made for the TLD (see,
for example, Ref (4)). 7.2.1 Ionizing radiation effects depend on the biases applied
to the device under test during and following irradiation (see
6.3 Dose Rate:
X1.4 and X1.5 for additional information).
6.3.1 Since ionizing radiation effects can depend strongly on
7.2.2 Biasing conditions for devices during irradiation shall
the dose rate of the irradiation, adequate steps shall be taken to
be maintained as specified in the test plan. In most cases, use
determine and control the dose rate (see 7.1 for additional
worst case bias conditions.
information).
7.2.3 If the time dependence of the behavior of the device
6.3.2 The dose rate shall be maintained at the value speci-
under test is to be studied, the biasing conditions on the device
fied in the test plan to a precision of 610 %.
following irradiation shall be maintained within 610 % of the
6.4 Device Preparation—The photons from the X-ray tester
bias conditions specified in the test plan.
have a limited range in materials as compared to photons from
7.2.4 If it is necessary to move the device from its location
a cobalt-60 gamma source (see X1.2 for further detail). As a
in the X-ray irradiation apparatus to a remote test fixture, the
result, microelectronic devices to be irradiated shall be tested
device shall be handled so as to minimize changes during the
either as regions on a wafer or as delidded packaged devices.
transfer.
6.5 Beam Collimation—X-ray testers may be used for irra-
7.2.4.1 If the device is packaged (and delidded), the contacts
diation of selected devices on a wafer. For this use, appropriate
on the device under test shall be shorted during transfer.
measures shall be taken to ensure that the X-ray beam is
7.2.4.2 If the device is either packaged or on a wafer, the
limited to the vicinity of the particular devices being irradiated.
device shall be handled so that electrical transients (for
See X1.6 for further detail.
example, from static discharge) do not alter the device char-
acteristics.
6.6 Test Instrumentation:
6.6.1 Various instruments for measuring device parameters
7.3 Temperature:
may be required. Depending on the device to be tested, these
7.3.1 Many device parameters are temperature sensitive. To
can range from simple current-voltage I-V measurement cir-
obtain accurate measures of the radiation-induced parameter
cuitry to complex integrated circuit (IC) test systems.
changes, the temperature must be controlled.
6.6.2 All instrumentation used for electrical measurements
7.3.2 In addition, time-dependent effects (see 5.3 and X1.7)
shall have the stability, accuracy, and resolution required for
can be thermally activated. Because of this, the temperatures at
accurate measurement of the electrical parameters as specified
which radiation measurements and storage take place can affect
in the test plan.
parameter values.
6.6.3 Cables connecting the device under test to the test
7.3.3 Devices under test (DUT) shall be irradiated at a
instrumentation shall be as short as possible. The cables shall
temperature measured at a point in the test chamber in close
have low capacitance, low leakage to ground, and low leakage
proximity to the DUT.
between wires.
7.3.4 All radiation exposures, measurements, and storage
shall be done at 24° 6 6°C unless another temperature range is
7. Procedure
agreed upon between the parties to the test.
7.1 Test Plan: 7.3.5 Temperature effects must also be considered in estab-
7.1.1 Parties to the test must agree upon the conditions of lishing the sequence of post-irradiation testing. Choose the
the test, as follows, and establish a test plan. sequence of parameter measurements to allow lowest power
7.1.1.1 Source and dose level to be used, dissipation measurements to be made first. Power dissipation
F1467 − 18
may increase with each subsequent measurement. When high absorbed dose in the device material of interest can be
power is to be dissipated in the test devices, pulsed measure- performed using Eq 1:
ments are required.
μ /ρ
~ !
en
a
D 5 D (1)
a b
7.4 Electrical Measurements: μ /ρ
~ !
en
b
7.4.1 The X-ray tester may be used to determine ionizing
where:
radiation effects on microelectronic devices for a broad range
D = equilibrium absorbed dose in the device material,
a
of applications including process control and research on
hardening technology (see Appendix X2 for further detail).
D = absorbed dose in the dosimeter,
b
7.4.2 A wide range of electrical measurements may be
(μ /ρ) = mass absorption coefficient for the device
en a
performed in conjunction with X-ray tester irradiations. These
material, and
may include current-voltage, subthreshold current-voltage, and
(μ /ρ) = mass absorption coefficient for the dosimeter.
en b
charge pumping measurements. These pre- and post-irradiation
NOTE 13—If, for example, the dose is measured in a PIN detector and
electrical measurements shall be performed as specified in the
the dose in an SiO region of the device is desired, the ratio (μ /ρ) /(μ /
2 en Si en
ρ) is, in the photon energy range of interest, approximately 1.8. Thus,
test plan. SiO2
in this case, D ≈ 1.8 D .
Si SiO2
7.4.3 Timing of Measurements:
7.5.1.4 A correction for absorbed-dose enhancement effects
7.4.3.1 Changes in electrical parameters caused by the
shall be considered. This correction is dependent upon the
growth and annealing of radiation-induced electrical charge
photon energy that strikes the device under test (see 8.2.1 and
within the device under test can be highly time dependent (see
X1.3).
5.1 for additional detail). As a result, particular care will be
given to the timing of the irradiation and electrical measure-
NOTE 14—A relatively simple case to analyze for dose enhancement is
ments as specified in the test plan.
one where the dose is desired for a thin (˜<50 nm) SiO layer bounded on
either side by thick (˜>200 nm) layers of silicon or aluminum (see, for
7.4.3.2 Long delays between the end of irradiation and the
example, Fig. X1.2 of X1.3). For this case, the dose-enhancement factor
start of electrical measurements are not recommended unless
is 1.6 to 1.8. That is, the dose in the thin SiO layer is approximately the
the purpose of the experiment is the study of time dependent
same as the dose in the adjacent silicon or aluminum. For a similar
effects (TDE). Unless otherwise specified, electrical measure-
problem, but with thicker SiO layers, the dose-enhancement factor is
˜<1.6 and˜>1 (see X1.3).
ments will be started within 20 min after the end of irradiation
or sooner.
7.5.2 Measurement of Dose Rate—Appropriate dosimetry
7.4.3.3 It is usually preferable to perform electrical testing
techniques shall be used to determine within 610 % the dose
on the device under test either during irradiation, immediately
rate of the irradiation of the device under test. Typically, the
following irradiation with the device left in place in the
dose rate will be the measured dose divided by the irradiation
irradiation fixture, or both.
time.
7.5 Dosimetry:
NOTE 15—Determination of the significance of the dose rate for
radiation effects can be quite complex (see 5.1, 8, and X1.7).
7.5.1 Measurement of Dose:
7.5.1.1 Appropriate dosimetry techniques shall be used to
8. Comparison with Cobalt-60 Gamma Results
determine within 610 % the dose applied to the device.
8.1 Physical Processes That Affect Radiation Effects:
7.5.1.2 The equilibrium absorbed dose shall be measured
with a dosimeter irradiated in the position of the device before,
8.1.1 When X-rays are used to test devices, the magnitude
or after, the irradiation of the device. of the irradiation-induced changes in electrical parameters may
be significantly different as compared to the changes resulting
NOTE 11—The dose from X-ray testers has most commonly been
from cobalt-60 gamma irradiation at the same exposure level
measured using a calibrated PIN diode detector (3). This method results in
(4).
a measured dose-rate in rad(Si)/s. Since there is some appreciable
attenuation of the X-ray beam on penetrating to and through the sensitive
8.1.2 The causes for these differences arise from the depen-
layer of the detector (even with a filtered spectrum as required by 6.2.1.3),
dence of radiation effects on the energy of the irradiating
a correction needs to be made to give the dose which would have been
photons. Two of the important mechanisms leading to these
deposited in a very thin layer of silicon. This correction is somewhat
differences are absorbed-dose enhancement (7) and electron-
spectrum dependent. At least one manufacturer provides detectors whose
hole recombination (8).
calibration includes this correction. During the calibration measurement
the front surface of the sensitive region of the PIN detector must be in the
8.1.3 In comparing radiation-induced effects caused by
same plane as the front surface of the device under test. Further, care must
X-rays and cobalt-60 gammas, the relative magnitude of
be taken that the entire front surface of the sensitive region of the PIN
absorbed-dose enhancement and electron-hole recombination
detector must be illuminated by the X-ray beam.
shall be assessed. The magnitude of such effects must be
NOTE 12—Other dosimetry methods that have been used include TLDs
assessed for the specific testing environment used.
(see Practice E668 and Ref (4)) and X-ray photographic film.
8.2 Use of Corrections for Physical Processes to Intercom-
7.5.1.3 This dosimeter absorbed dose shall be converted to
pare X-ray and Cobalt-60 Gamma Measurements:
the equilibrium absorbed dose in the material of interest within
the critical region within the device under test, for example the 8.2.1 Combined Effects of Absorbed-Dose Enhancement
SiO gate oxide of an MOS device. Conversion from the and Electron-Hole Recombination for Si-SiO Devices—In
2 2
measured absorbed dose in the dosimeter to the equilibrium order to compare the radiation effects caused by X-ray and
F1467 − 18
cobalt-60 gamma irradiations, it is necessary to make appro- (X-Ray)) must approach unity. Thus the differences between
priate allowance for the differences between these two sources. X-ray and cobalt-60 gamma irradiation are most serious for
In order to accomplish this, it has been suggested that it is relatively low doses. This caution is important to bear in mind
necessary and sufficient to correct for differences in absorbed- for doses approaching the failure dose for a device, where hole
dose enhancement and electron-hole recombination (9, 10, 11, trapping may be showing signs of saturation.
12, 13). A critical assessment of this body of work suggests that
8.2.1.6 Finally, the methodology of this section is appropri-
X-ray versus cobalt-60-gamma comparisons often can properly
ate for the calculation of effects within the gate or field oxide
be made in this fashion.
layers of individual transistors. To apply these methods to the
8.2.1.1 Although the methodology described in this section
radiation-induced failure of microcircuits, it is necessary to
is predominantly based on radiation-induced hole-trapping
apply them to the critical devices that result in the microcircuit
studies, the same approach can be applied to interface state
failure.
generation. (For additional discussion see X1.8.1.)
8.2.2 Corrections for Standard MOS Devices:
8.2.1.2 This section will present an estimate of the differ-
8.2.2.1 Table 1 presents estimates of the combined effects of
ences between X-ray and cobalt-60 gamma effects for several
absorbed-dose enhancement and electron-hole recombination
important cases. That is, an estimate will be presented of the
for several important cases for standard MOS technology. In
expected values of the ratio (Eq 2):
order to systematize these results, the problem has been split
into five cases of practical interest.
Number Holes ~Cobalt 2 60!
Relative 2 Effect 5 (2)
Number Holes X 2 Ray 8.2.2.2 The results of Table 1 have been calculated assum-
~ !
ing that the cobalt-60 gamma data are taken using a lead-
8.2.1.3 The combined effects of both absorbed-dose en-
walled test box (14, 15). The use of such a test box for
hancement and electron-hole recombination will be presented.
cobalt-60 gamma irradiations is recommended, and thus the
In calculating the ratio of Eq 2, it has been assumed that both
data of Table 1 should be regarded as representing the results
sources (X-ray and cobalt-60) produced the same dose (as
to be expected using best experimental practice (see Practice
measured by TLDs or silicon PIN detectors and corrected to
E1249).
dose in ‘bulk’ SiO ) with the same dose rate (in SiO ).
2 2
8.2.1.4 It should be noted that the material of this section
NOTE 16—The effects of using the lead-walled test box for cobalt-60
testing are especially important for cases where high atomic number
includes the combined effects of only dose enhancement and
materials are present. An example is the presence of a gold flashing on the
recombination. If other effects (for example, time dependent
interior surface of the lid. For additional details see Ref (14).
interface state growth or hole annealing effects) are important,
8.2.2.3 Note first, in Table 1, that there are cases where one
then those correction factors must be included also. Some of
would expect small differences between X-ray and cobalt-60
these other effects are discussed in X1.7.
gamma irradiation, and other cases where a factor of 1.5
8.2.1.5 Further, it is important to note that the values
differences are expected.
presented in this section (see Table 1) do not treat saturation
effects. That is, they are appropriate for cases where the effects 8.2.2.4 During cobalt-60 gamma exposures, if high atomic
are approximately linearly related to dose. Clearly, as one number elements are present, such as gold deposited on the
approaches the limiting case where hole trapping is completely inside of Kovar device lids, additional dose enhancement can
saturated, the ratio (Number Holes (cobalt-60))/(Number Holes occur. This may raise the numbers in Table 1 by 10 to 20 %
TABLE 1 Estimate of the Ratio of the Relative Effects of Cobalt-60 and X-Ray Irradiations for Silicon MOS Devices
(Using a Lead-Walled Test Box with Cobalt-60)
NOTE 1—These ratios of cobalt-60 to X-ray effects do not account for saturation. As radiation effects begin to saturate, cobalt-60 and X-ray effects
become more similar and, thus, the ratio of their effects approaches unity.
NOTE 2—The estimated values in this table are intended to give the reader a rough value of the experimental results that should be expected. The
number of significant digits used are not representative of what would be appropriate for reporting experimental results.
Number of Holes (cobalt-60)
Case Description of Case Comments
Number of Holes (X-ray)
I Gate (On):
oxide thickness = 25–50 nm ;0.9 Effects nearly cancel
oxide field ' 10 V/cm
II Gate (Off):
oxide thickness = 25–50 nm ; 1.2 Recombination dominates slightly
oxide field ' 10 V/cm
III Thick Gate (On):
oxide thickness = 100 nm ;0.9 Effects nearly cancel
oxide field ' 10 V/cm
IV Thick Gate (Off):
oxide thickness = 100 nm ;1.3 Recombination dominates slightly
oxide field ' 10 V/cm
V Field:
oxide thickness = 100–400 nm 1.3 to 1.5 Recombination dominates
oxide field ' 10 V/cm
F1467 − 18
(15, 16). (This estimate is for the case where a lead-walled test shows the variation of dose enhancement with the thickness of
box is used. The increase may be a factor of 1.5 to 1.7 in the the polysilicon layer separating the silicide layer and the gate
absence of this spectrum filtration.) oxide.
8.2.3 Example—The calculations for Case I are now treated
8.2.4.4 Electron-hole recombination corrections are ex-
in greater detail to clarify how to handle cases not treated
pected to be similar under fields of interest in devices with
explicitly in Table 1. The data sources and calculations leading
heavy-metal silicides as in more conventional devices (see
to the results shown in Table 1 are as follows:
X1.5). Thus, recombination corrections may be taken from, for
8.2.3.1 First, the X-ray absorbed-dose enhancement factor
example, Eq X1.1 and Eq X1.3 of X1.5.
can be obtained from the literature. See, for example, Fig.
8.2.5 Corrections for Silicon on Insulator (SOI) Devices:
X1.2b and Refs (11), and (17). Note, from Fig. X1.2b, that a
8.2.5.1 There is evidence that the back-gate threshold volt-
50-nm oxide corresponds to an enhancement factor of about
age in SOI devices can be particularly sensitive to photon
1.6.
energy. The top gates on SOI devices are expected to behave in
8.2.3.2 Second, the cobalt-60 gamma absorbed-dose en-
the same manner as for more conventional devices if back-gate
hancement factor was assumed to be 1.0 (no enhancement).
leakage is suppressed.
This is reasonable in the absence of high-Z material such as a
8.2.5.2 A comparison of X-ray and cobalt-60 gamma effects
gold-flashed lid. Estimates of the cobalt-60 gamma absorbed-
on SOI devices has been presented by Fleetwood et al (20).
dose enhancement factor in the presence of high-Z material can
This paper compared zone melt recrystallization (ZMR) de-
be found in Refs (14) and (15).
vices having 2 μm-thick buried oxides with separation by the
8.2.3.3 Third, the recombination correction factor can be
implantation of oxygen (SIMOX) devices having 0.4 μm-thick
obtained from Eq X1.1 and Eq X1.3 of X1.5. Consider the data
buried oxides.
of these equations for a field of 10 V/cm. Note that a
8.2.5.3 This work showed major differences for back-gate
comparison of the fraction of unrecombined holes for a
threshold-voltage shift with devices built with ZMR material.
cobalt-60 gamma source to the fraction of unrecombined holes
At zero back-gate bias, a given back-gate threshold-voltage
obtained using an X-ray tube shows a difference of about a
6 shift required three times the X-ray dose in comparison to the
factor of 1.4 (for example, at 10 V/cm the ratio is about
cobalt-60 gamma dose. This was the worst case of the
0.64/0.46 = 1.4).
experimental situations explored. These results are summa-
8.2.3.4 Using these numbers, the combined difference in
rized in Table 3 as Case VII.
effect is about (1.0/1.6) × (1.4) = 0.9. Such calculations are the
8.2.5.4 The differences were smaller for SIMOX devices. In
source of the numbers given in Table 1.
this case, X-ray exposures greater by a factor of approximately
8.2.3.5 Calculations similar to the ones just described can,
1.5 were required to give the same shift as was obtained with
of course, be carried out for values of oxide thickness and field
cobalt-60. It was inferred that this difference resulted from the
that are intermediate to the limiting cases used in Table 1.
smaller thickness of the buried oxide (0.4 μm) for the SIMOX
8.2.4 Corrections for Devices with Heavy-Metal Silicides:
devices.
8.2.4.1 Devices are now being manufactured with heavy
8.2.5.5 The differences between the results for ZMR and
metal metallization layers such as tungsten or tantalum silicide.
SIMOX devices was attributed to the field dependence of
8.2.4.2 The presence of such layers is expected to result in
electron-hole recombination in the buried oxide.
significant dose enhancement in adjacent SiO gate oxides for
X-ray irradiation (17, 18, 19). For example, Fleetwood et al 8.2.5.6 Note that the correlation factor for SOI or silicon on
(19) suggest dose-enhancement factors in excess of 2.5 for sapphire (SOS) devices can be strongly affected by the bias on
some cases. These results are summarized in Table 2 as Case the train of the top gate transistor during irradiation (20). In
VI. particular, it is expected that the two radiation sources should
8.2.4.3 Although the mechanisms for dose enhancement are agree more closely for the case in which the drain of the top
expected to be the same as for Si-SiO devices, the greater gate transistor is biased during irradiation (with zero back gate
magnitude of this effect in silicided devices require modifica- bias) because the field in the buried insulator is greater than for
tion of the method outlined in 8.2.1. Fleetwood et al (19) give zero drain bias and, hence, the differences in electron-hole
some suggestions on how to make such corrections. See, for recombination can be smaller.
example, X1.3 for suggested dose-enhancement factors (19). In 8.2.5.7 Additional data in Fleetwood et al (20) may be
particular, note Table X1.3 that shows the variation of dose helpful in comparing X-ray and cobalt-60 gamma results on
enhancement with gate oxide thickness, and Fig. X1.3 that SOI devices.
TABLE 2 Estimate of the Ratio of the Relative Effects of Cobalt-60 and X-Ray Irradiations for Cases
of Silicon MOS Devices with Heavy Metal Silicides
Number of Holes (cobalt-60)
Case Description of Case Comments
Number of Holes (X-ray)
VI Gate with Heavy Metal Silicide: 0.4 to 0.9 Substantial dose enhancement possible for X
gate oxide thickness = 25–50 nm, rays if heavy-metal layer is “near”—
gate oxide field ' 10 V/cm enhancement of factor of 2.5 possible
F1467 − 18
TABLE 3 Estimate of the Ratio of the Relative Effects of Cobalt-60 and X-Ray Irradiations for SOI Devices
Number of Holes (cobalt-60)
Case Description of Case Comments
Number of Holes (X-ray)
VII SOI Back Gate: 1.0 to 3.0 Substantial reduction of effect for X-rays for
buried oxide thickness = 0.4–2.0 μm small back-gate bias
8.2.6 Corrections for Recessed Field Oxides and Base 9.4 Device—State the manufacturer, device type number,
Oxides in Bipolar Devices—Titus and Platteter (21) have
package type, controlling specification, date code, other iden-
shown that X-ray and cobalt-60 gamma irradiations produce
tifying numbers given by the manufacturer, and any available
factor of two differences in radiation effects due to recessed
information on its specific construction;
field oxides. These differences have been attributed to differ-
9.5 Irradiation Geometry—State the position and orienta-
ences in electron-hole recombination in oxides with low fields
tion of source and device under test;
(see Fig. X1.5). That is, this is comparable to Case V (in Table
1) for standard MOS devices. Similar differences are expected
9.6 Electrical Bias—State the electrical bias conditions used
for the oxides that overlie the base-emitter junction of many
and provide a schematic for the bias circuit;
linear bipolar technologies. Such oxides often limit their total
9.7 Parameter Measurements—Provide a tabulation of test
dose response.
parameter measurement data, and
9. Report
9.8 Statistical Bias and Precision—State any experimental
9.1 As a minimum, report the following information (where
conditions that might lead to a bias or lack of precision in the
relevant):
measured results. State an estimate of the precision and bias for
9.2 Source—State the source type, target material, operating
the measured results.
voltage, fluence rate, and any information on a measured or
calculated energy spectrum. State the position, thickness, and
10. Keywords
composition of spectrum filtration materials, if any,
10.1 ionizing radiation effects; microcircuits; radiation
9.3 Dosimeter System—State the dosimeter type, calibration
hardness; semiconductor devices; X-ray testing
data, relevant environmental conditions during the irradiation,
dose enhancement and recombination corrections used;
APPENDIXES
(Nonmandatory Information)
X1. PHYSICAL PROCESSES THAT AFFECT RADIATION EFFECTS
X1.1 Introduction bias applied during irradiation. It will be shown that these
phenomena can, in some cases, lead to major changes in the
X1.1.1 This appendix will contain a discussion of four
correlation between incident radiation flux and the measured
classes of physical processes that are of concern to the user of
effect on the device.
an X-ray tester.
X1.1.2 First are the processes of attenuation and filtration of X1.1.4 The third class of physical processes is concerned
the incident spectrum before it strikes the region of interest with the possibility of improper localization of the incident
within the device-under-test. These are important because of X-ray beam caused by scattering or fluorescence. This is
their bearing on the question of whether the conversion from
believed to be a manageable problem.
the measured dose in a detector (PIN, TLD, etc.) to the required
X1.1.5 The fourth class of physical processes to be dis-
dose in the region of interest (such as the SiO gate oxide)
cussed includes phenomena that are less well understood than
within the device under test must be determined for each type
those treated in the first two classes. Included in this class are
of device. That is, will each type of device require determina-
interface state generation effects and annealing effects.
tion of a correction for spectrum absorption and filtration, or
are these corrections negligible? It will be shown that these are
X1.1.6 Radiation-test personnel must give consideration to
usually not major
...