preprint t r 'r. , , a j6s76 7 · · 2013-08-31preprint, , t ~ 9 ' r ~ 'r ... 2....
TRANSCRIPT
X-711-71-410PREPRINT
, , t 9 , , A~ r ' ~ 'r. I 7RADIATION DAMAGE IN MOS
INTEGRATED CIRCUITSPART I'
-. I
VITALY DANCHENKO. . .- . . .~~~~~~~~
I I
SEPTEMBER 1971
Reproduced by
NATIONAL TECHNICALINFORMATION SERVICE
Springfield, Va. 22151
I.- - GODDARD SPACE FLIGHT CENTERGREENBELT, MARYLAND
N.
1N72-.12767
Unclas09762
(N AS A TM-X-65764) RADIATIONINTEGRATED cIRCUITS, pART 1
(NASA) Sep. 1971 64 P
DAM AGE I1V. Danch(
CS
Ner
MOSnkoL 09C
G3/ 2 6
(CAl O RY)* I- NASA dOR VMX OR AD)ER)
.4
N
I
J6S76
I *
, . . I -
I I
-I
https://ntrs.nasa.gov/search.jsp?R=19720005118 2018-06-23T11:47:10+00:00Z
X-711-71-410
RADIATION DAMAGE IN MOS INTEGRATED CIRCUITS, Part I
Vitaly Danchenko
September 1971
Goddard Space Flight CenterGreenbelt, Maryland
RADIATION DAMAGE IN MOS INTEGRATED CIRCUITS, Part I
by
Vitaly Danchenko
ABSTRACT
Complementary and p-channel MOS integrated circuits made by four
commercial manufacturers were investigated for sensitivity to radiation
environment. The circuits were irradiated with 1.5 MeV electrons. The
results are given for electrons and for the Co-60 gamma radiation equiv-
alent. The data are presented in terms of shifts in the threshold poten-
tials and changes intransconductances and leakages. Gate biases of -10V,
+10V and zero volts were applied to individual MOS units during irradiation.
It was found that, in most of circuits of complementary MOS tech-
nologies, noticable changes due to radiation appear first as increased
leakage in n-channel MOSFETs somewhat before a total integrated dose
of 1012 electrons/cm 2 (-3 x 104 rads/cm2 , Co-60) is reached. The in-
ability of p-channel MOSFETs to turn on sets in at about 1013 electrons/
cm 2 . Of the circuits tested, anRCA A-series circuit was the most radia-
tion resistant sample.
1ii pp,;kl-"
u' " I '~
' ):f ' 4lT: Tl
CONTENTS
Page
I. INTRODUCTION .................. 1.. 1
II. EXPERIMENTAL APPROACH AND METHOD ................ 3
A. Approach ......................... i............. 3
B. Method ........................................ 4
III. RESULTS ........... ' ................ .. ...... 6
A. Solid State Scientific, complementary, SCL 5102 Quad 2input NAND ................................... 6
B. RCA COS/MOS (complementary symmetry) circuits .......... 8
1. Low threshold integrated circuits, CD 4001 Quad 2input NOR .................. 8................. 8
2. COS/MOS of earlier vintage, CD 4007 Dual ComplementaryPair Plus Inverter .............................. 10
3. RCA CD 4001A, new A-series ...................... 11
C. Motorola complementary circuits, 2501 Quad 2 input NOR ..... 12
D. American-Microelectronics, Inc., (AMI) p-channel integratedcircuit, MX03C10 channel switch ................... 13
IV. CONCLUSION ...................................... 13
V. APPENDIX. Radiation damage in MOS devices ................ 15
1DW~~~~~ '~,j ,~,~-,~~,
LIST OF ILLUSTRATIONS
Figure Page
1 Schematic and photograph of an irradiated IC made by Solid
State Scientific, Inc ............................... 20
2 Shift in threshold potentials in n-channel MOS devices of Solid
State Scientific IC irradiated with various gate biases ....... 21
3 Same as Fig. 2, except p-channel devices ................ 22
4 Degradation in transconductances, gi's, in n-channels of Solid
State Scientific, Inc., as a function of total accumulated radiation
dose ........................................ 23
5 Same as Fig. 4, except p-channels ................... . 24
6 Leakage in n-channels of Solid State Scientific ICs as a function of
total accumulated radiation dose . . . . . . . . . . . . . . . . . . .. 25
7 Recovery of threshold potentials in n-channel and p-channel MOS
devices of Solid State Scientific ICs at room temperature ..... 26
8 Diagram and photograph of an irradiated RCA IC........... 27
9 Shifts in the threshold potentials in n-channel MOS devices of RCA
IC bombarded with various gate biases ................. 28
vii -i .: ..e~ .. ,: --:~k l,.1%; '~i'''
Figure Page
10 Same as Fig. 9, except p-channels ..................... 29
11 Degradation of the transconductances in n-channel and p-channel
MOS devices of RCA ............................. 30
12 Leakage in n-channels of RCA IC as a function of total accumulated
radiation dose .................................. 31
13 Same as Fig. 12, except total leakage in all 8 n-channels ...... 32
14 Same as Figs. 9 and 10, except irradiated with 5V gate biases... 33
15 Same as Fig. 11, except irradiated with 5V gate biases ...... 34
16 Same as Fig. 12, except irradiated with 5V gate biases ...... 35
17 Same as Fig. 13, except irradiated with 5V gate biases ...... 36
18 Schematic and photograph of irradiated RCA CD 4007 IC ...... 37
19 Shifts in threshold potentials of n- and p-channel MOD devices of
RCA IC, irradiated with various gate biases .......... 38
20 Degradation of the transconductances in n- and p-channels of
RCA IC as a function of total accumulated radiation dose ...... 39
viii
Figure Page
21 Leakage in n-channels of RCA IC, irradiated with various gate
biases ........................................ 40
22 Recovery of threshold potentials in n- and p-channel MOS of RCA
IC at room temperature ............................ 41
23 Shifts in the threshold potentials of n- and p-channel MOS devices
of RCA new A-series, irradiated with various gate biases ...... 42
24 Degradation in transconductances in n- and p-channels of RCA
new A-series, as a function of total accumulated radiation dose.. 43
25 Leakage in n-channels of RCA new A-series ............... 44
26 Diagram and photograph of irradiated Motorola IC .......... 45
27 Shifts in the threshold potentials of n- and p-channel MOS devices
of Motorola IC, irradiated with various gate biases .......... 46
28 Degradation in transconductances in n- and p-channels of Motorola
IC .......................................... 47
29 Leakages in n-channels of Motorola IC, irradiated with various
gate biases .................................... 48
30 Schematic and photograph of an irradiated AMI p-channel IC... 49
ix
Figure Page
31 Shifts in the threshold potentials of p-channel MOS devices of
AMI p-channel IC, irradiated with various gate biases ....... 50
32 Degradation in transconductances in p-channels of AMI p-channel
IC ...................................... .... 51
33 Radiation damage in a single MOS device unit ............. 52
34 Shift of the I-V characteristic of n-channel MOS device (GME) with
various radiation total accumulated doses ............... 53
35 Same as Fig. 34, except p-channel MOS device ............ 54
x
RADIATION DAMAGE IN MOS INTEGRATED CIRCUITS, Part I
I. INTRODUCTION
There are many advantages of Metal-Oxide-Semiconductor (MOS) integrated
circuits (ICs) over bipolar ICs in the design of large-scale integrated logic and
memory circuitry for spacecraft. Among these are extreme compactness and
low power consumption. Such circuits are being orbited to an ever increasing
extent. They have been flown or are planned to be flown on IMP-I, H and J,
ATS, NIMBUS, Pioneer, Helios and many other missions.
MOS ICs are found to be more sensitive to space radiation than their bipolar
counterparts. Consequently, we find ourselves under considerable demand from
flight project people at Goddard and other NASA Centers, as well as from
principal investigators for flight experiments outside NASA, who use or plan
to use MOS type ICs, to investigate the instability of these circuits in the space
radiation environment. Inquiries also come from foreign countries such as
France, Germany and Denmark, to whom Goddard extends assistance in space-
craft electronics. In particular, many questions are being asked about comple-
mentary MOS integrated circuits. To meet these demands we have initiated an
extensive program to systematically measure the sensitivity to radiation of
MOS circuits. It is our intent that reports, similar to this, will be issued from
time to time and made available to those who are involved in planning and design
of space-circuitry.
1
Experience has shown that radiation effects in MOS IC s not only differ
greatly from one manufacturer to another, but also change within a given company
due to a continual effort in improving stability and production of MOS devices.
These changes, directed primarily to facilitate production or lower threshold
potential, may or may not improve radiation hardening. There is some direct
effort on a laboratory scale to harden MOS ICs against radiation, but circuits
available for space flight use at present have not been manufactured to be radia-
tion resistant.
This report describes radiation effects in MOS integrated circuits most of
which have been used by GSFC in space missions. Of the four MOS technologies
investigated here three are of complementary and one is of strictly p-channel
type. In the choice of ICs for the investigation and the presentation here, no
preference whatsoever was given to any particular technology other than the
requirement of GSFC for the data and availability of suitable samples for ir-
radiation. The data are presented in terms of the shifts in the threshold poten-
tials of individual MOS units, changes in the transconductances and leakages in
n-channel units with various biases applied during irradiation. Other effects,
such as breakdowns in the gate insulation and leakages in the drain-to-source
junctions were also observed, but discounted here. Since some amount of han-
dling of the samples was necessary, it was not directly evident that these effects
were induced by radiation or caused by handling.
2
For those who are not familiar with a general aspect of radiation damage
in individual MOS devices and its effect on the performance of an IC a short
discussion of it is given in the Appendix.
II. EXPERIMENTAL APPROACH AND METHOD
A. Approach
In the investigation of the sensitivity of various MOS integrated circuits to
radiation environment full current-voltage characteristics of individual MOS
units in an IC were taken, rather than merely measure the degradation of the
output of the circuit as a whole. In choosing this method we have the following
advantages: 1) we are able to characterize individual MOS units in a circuit in
a radiation environment separately, and thereby to determine whether the n-
channel or the p-channel devices are responsible at the onset of the degradation
of the circuit. 2) The degradation in the transconductances and the shifts in the
threshold potentials may be measured separately in n- and p-channel devices.
3) It gives us some insight into the physical mechanism of radiation damage if
the differences in the technologies of various MOS circuits are known. 4) Anneal-
ing of radiation damage, if any, can be observed at room temperature separately
in n- and p-channel devices.
For practical purposes the degradation in the performance of an IC as a
whole can be seen and predicted from the observation in the degradation of
individual MOS units. The crucial points to observe in individual MOS units as
3
a function of total accumulated radiation dose are when the VGT of n-channel
units approach zero gate bias, whereby the IC will become leaky, and when the
VGT of p-channel approach the total bias of VDD , whereby the functionality of
an IC will be impaired.
Since the sensitivity to radiation environment is dependent strictly on the
processing technology of the gate oxide, the results of radiation damage in one
MOS IC are valid and can be extended to any other IC having the same gate oxide
technology.
B. Method
In selecting particular MOS ICs for the characterization of the radiation
damage in a certain series of integrated circuits having the same gate oxide
technology no preferences were exercised other than the simplicity of a circuit
and the accessibility to the individual gates with direct voltage biases with
respect to the substrates.
Since a shift in a threshold potential depends very strongly on the amount
of gate bias applied with respect to the substrate during irradiation, several of
the n- and p-channel gates on the same chip were biased with +10 and -10 volts
continuously during irradiation, respectively. Others gates were shorted to the
substrate and the rest were operated with a square wave of 10 volts (at 100 kHz)
to simulate the actual operation of an IC. For example, if an IC contained four
identical sections,then two of the sections were operated with a square wave, in
4
the third section n-channels were biased and p-channels were shorted, and in
the fourth section the biasing configuration was the reverse to that of the third.
In this way, since it is done on the same chip of silicon, a very good comparison
of the sensitivities to radiation under various biasing conditions are obtained.
Experience has shown that if the two identical type devices are physically lo-
cated next to each other on a chip of an IC, then they will have to be biased the
same way. Otherwise, if one is biased with 10 volts and the other is shorted,
during irradiation, then the shorted devices exhibits source to drain leakage,
which usually disappears after a couple of days at room temperature. This is
believed to be due to the adsorption of gas ions on the surface of the chip. Since
some of the MOS units remain shorted or biased for relatively long periods of
time in space environment this effect is important, and should be investigated
further.
Integrated circuits were irradiated with 1.5 MeV electrons from Van de
Graaff accelerator at Goddard and the electron fluxes were measured with cali-
brated Faraday cup. Irradiations were performed in steps until a final total
accumulated dose of 1014 e/cm2 was reached. At this total dose the performance
of most of the MOS ICs was degraded beyond their usefulness. The radiation
effects due to electrons may be converted to the effects with Co-60 gamma
source by using the factor 2.8 x 10-8 rads/electron/cm2. This factor was
taken from the work by W. L. Brown, J. D. Gabbe and W. Rosenzweig and
5
checked out by an experiment to be within 10% correct. The conversion to Co-60
gamma ray sensitivity is indicated on all the data graphs at the upper horizontal
axis.
After each radiation step a full I-V characteristic of each MOS device in an
IC was taken in a semilog representation, where the logarithm was that of the
drain current. From these curves the changes in the threshold potential, trans-
conductance and leakage were measured. The leakage measured in most cases
was that of the individual devices, but in some cases a total leakage of all the
n-channels in parallel was measured. Since most of the MOS ICs investigated
had diode protection, it was not possible to measure directly the shift of
the threshold potentials in n-channel devices below about -0.7 V. In that case
the shifts in V T were extrapolated by the use of the shapes of the curves of
the devices which, for example, were shorted during irradiation and therefore
were completely represented above the zero gate voltage line. These curves
were matched to the high current portion of the curves of the measured device.
These extrapolation are considered to be valid since it was observed that the
production of the radiation-induced interface states is only a weak function of
the applied bias during irradiation.
III. RESULTS
A. Solid State Scientific, Complementary, SCL-5102 Quad 2 Input NAND.
These integrated circuits were chosen for the tests because of their exten-
sive use by various flight projects and because a ready access to the individual
6
gates for biasing purpose and testing. The ICs were obtained from the company
in early spring of 1971 and represented the latest MOS technology at that time.
At the time of the preparation of this report the company was in the process of
changing and cleaning up their oxide gate technology, so that the radiation effects
in the new technology will most probably be reported in the next issue of this
report.
A schematic and a photograph of the 5102 irradiated are shown in Fig. 1.
Figures 2 and 3 show the shifts in the threshold potentials of n- and 1-channels
respectively, irradiated with +10V, +10V square wave 50% duty cycle and zero
volt, gate biases for n-channels, and -10V, -10V square wave of 50% duty cycle
and zero volt, gate biases for p-channels. A considerable stability in radiation
environment is exhibited by the MOS units irradiated with zero gate biases of
both n- and p-channel type. A shift of n-channels irradiated with zero gate bias
upward indicates that an electron trapping in the oxide predominates as compared
to the hole trapping.' 2 The degradation in the transconductances is shown in
Figs. 4 and 5 for n- and p-channels, respectively. Although, the initial trans-
conductance of n-channels is high for this technology as compared to other manu-
facturers, it's degradation is more severe as compared to others. It also indi-
cates a slight bias dependence, particularly in p-channel units. The leakages
begin already at about 5 x 1011 e/cm2 for the devices irradiated with +10V fully
on biases, as shown in Fig. 6, and increase quite rapidly with dose. As shown
in Fig. 3, the threshold potentials of p-channels irradiated with -10V fully on
7
gate bias approach the threshold potential of 10V at about 5 x 1012 e/cm2 . Thus,
it would appear from these data that the ICs will become too leaky before the
loss of functionality (due to the p-channels not being able to turn on) will set in.
The data in Fig. 7, however, show that the n-channels of this technology exhibit
a considerable recovery of the threshold potentials at room temperature as com-
pared with the p-channels. In space environment, therefore, at a small enough
dose rate the n-channels may not show any leakage at all before any loss of the
functionality will be evident.
B. RCA COS/MOS (complementary symmetry MOS) Circuits
1. Low threshold integrated circuits, CD 4001 Quad 2 Input NOR
These ICs were obtained from RCA at the end of 1970 and represent the
technology just before the switch over to the so-called "A-series" technology.
The ICs were designed to operate also at 5 volts input levels and at higher
speeds. Consequently, they were irradiated both with 10V and 5V gate biases
for comparison purposes.
Figure 8 shows the photograph and a schematic of an irradiated IC and Figs.
9 and 10 show the shifts in the threshold potentials of n- and p-channel MOS
units, respectively. The n-channel units, irradiated with +10V and zero gate
biases, have developed leakages in the gates and their measurements had to be
terminated soon after a few irradiation doses. Whether these gate-to-substrate
as well as some of drain-to-source leakages were caused by radiation could
8
not have been ascertained and, therefore, their analysis is not included in this
report. The worst case for p-channels happens already at a dose of 1012 e/cm2
when the threshold potential approaches the -10V value for the fully on during
irradiation. This is a factor of 5 sooner than in Solid State Scientific technology.
The leakages, as shown in Fig. 12, begin to show up at-about the same level as
in SSS technology. The initial transconductances of n-channels are much lower
than in p-channel, but the overall degradation in the transconductances is much
less than in SSS technology (Figs. 11, 4 and 5). Fig. 13 shows the total leakage
of the n-channel part of the IC and is about the same dependence on the dose as
the worst case in the individual leakages (Fig. 12).
The results for 5V irradiation biases are shown in Figs. 14 through 17. As
expected the overall shifts in the threshold potentials (Fig. 14) are smaller than
in the case of 10V bombardment biases (Figs. 9 and 10). The leakages in the
n-channels also develop later in dose (Fig. 16) than in the case of 10OV bombard-
ment bias. The degradation in the transconductances is about the same in both
cases since it is a weak function of the gate bias.
In the context of this comparison the important question, which is being
asked by many circuit designers, is: From the radiation damage point of view
which of the two gate biases is better to use, high or low? These data show
clearly that although the shifts in the threshold potentials for the lower gate
biases are smaller, the 5V value of the threshold potential is reached already
at a dose of about 4 x 1011 e/cm2 in the case of 5V operation mode (Fig. 14),
9
whereas the 10V value of the threshold potential is reached at a dose of about
10' 2 for the case of 10OV operation. So that the loss of functionality will happen
sooner in the case of 5V operation as compared to 10V operation. The leakage,
however, will happen later in the case of 5V operation.
2. COS/MOS Pair plus Inverter of the Earlier Vintage, CD 4007 Dual
Complementary
The data on CD 4007 presented here were obtained on an IC made by RCA
sometime in the middle of 1969. The integrated circuit is shown in Fig. 18 and
the results in Figs. 19 through 22. The sensitivity to radiation of these ICs is
not much different from the sensitivity of ICs made one and a half years later
(CD 4001 irradiated at 10V bias). The peculiar feature of these devices is that
even p-channel devices show some recovery in the threshold potentials when the
irradiation is discontinued for two or three days. This is indicated by the breaks
in the curves in Fig. 19. But as soon as the irradiation is resumed the shape of
the curve is regained. This suggests that the processes responsible for the
recovery are of low activation energy and, therefore, are easily counteracted by
additional irradiation. The reason for the recovery of a p-channel with fully on
gate bias during irradiation after about 6 x 1012 e/cm 2 is not known. The initial
values of the transconductances were much lower in n-channels than in p-channels
and the degradation was observed to be small, as shown in Fig. 20.
10
The recovery of the threshold potentials at room temperature is shown
in Fig. 22. Here, as in the case of Solid State Scientific, the n-channels exhibit
faster recovery than the p-channels.
3. RCA CD 4001, New "A Series"
In the beginning of 1971 RCA has introduced its new "A-series." The letter
A appears after the number of an IC, for example, CD 4001A. The company
claims that it has cleaned up its gate oxide technology, with the result that
the ICs are more reliable and less radiation sensitive. The results of the tests
of two such ICs are shown in Figs. 23, 24 and 25. As can be seen from these
results a considerable improvement in radiation hardness has been achieved.
A significant difference in radiation response is observed in the case of p-channel
devices. Aside from increased radiation hardness, a reversal in the gate bias
dependence is observed (Fig. 23). In the previous gate oxide technology a p-
channel with gate bias applied during irradiation suffers a greater shift in the
threshold potential. In this technology the shift is greater in the case of zero
gate bias. In p-channels irradiated with the gate bias the response is about the
same for gates with fully on bias or 50% duty cycle. This response of the new
RCA devices is very similar to the responses of p-channels of Motorola and
AMI to be discussed below.
Fig. 24 shows the degradation in the transconductances of n- and p-channels.
The degradation is considerably greater in this technology than in the previous
11
one (compare Fig. 11). Also, it exhibits a much greater gate bias dependence
in p-channels. In fact, it is very similar to the gate bias dependence in the
shifts of the threshold potentials, which indicates that the greater shift of the
threshold potentials in zero gate bias devices is most probably due to the greater
production of ionized interface states.
The leakages in n-channels of this IC appear only in devices with fully on
10V gate bias at about 2 x 1012 e/cm2 (Fig. 25). The threshold potentials
of the devices with 50% duty cycle and zero gate bias never reach the zero gate
bias line of Fig. 23, as radiation dose increases. This is due to the balance
between electron and hole trapping and, as radiation dose increases, the electron
trapping predominates.
From these data it can be predicted that if the ICs from this new A-series
is constantly operational in space environment, it will pass a radiation dose of
1014 e/cm2 before a leakage or a loss of functionality will be observed. To
substantiate these results, however, more tests of the A-series need to be
performed.
C. Motorola Complementary Circuits, 2501 Quad 2 Input NOR
The Motorola Complementary 2501 - Quad 2 input NOR circuit that was ir-
iradiated is shown in Fig. 26. Fig. 27 shows the results of the shifts in the threshold
potentials of n- and p-channels units with the various gate biases applied during
irradiation. The response in p-channels is very similar to the response of RCA's
12
new A-series. The radiation hardness of the n-channels, however, is very poor
with the leakage starting already at 2 x 10l1 e/cm2 (Fig. 29). The degradation
in the transconductance is shown in Fig. 28 and here again the degradation in
n-channels is very rapid and dose dependent. The production of radiation-induced
interface states in p-channels with zero gate bias during irradiation is most
probably responsible for the larger shift in the threshold potentials in those
devices (Fig. 27 and 28).
D. American-Microelectronics Inc., (AMI) p-Channel Integrated Circuit,
MX03C 10 Channel Switch
For comparison purposes one entirely p-channel IC made by AMI was
irradiated. A schematic and photograph of AMI MX03C - p-channel, 10 channel
switch is shown in Fig. 30. The shifts in the threshold potentials bombarded
with various gate biases are shown in Fig. 31. Here, as in Motorola and RCA's
A-series, the reversal of bias dependence is observed: a device with zero gate
bombardment bias exhibits a larger shift than the ones with applied gate bias.
It is expected that for this type of a circuit in the worst case the loss of func-
tionality will happen at about 5 x 1013 e/cm2 . The degradation in the trans-
conductances is shown in Fig. 22 and is comparable to the degradations in RCA's
A-series and Motorola p-channels.
IV. CONCLUSIONS
Selective biasing during irradiation of individual MOS units in an integrated
circuit has proven to be a good method to characterize the radiation sensitivity
13
of individual MOS devices as well as an IC as a whole. In general, the results
have shown that in most of the complementary circuits tested, the radiation
effects will appear first as a leakage in n-channel devices somewhat before the
integrated dose of 1012 e/cm2 is reached. The loss of functionality will appear
soon after the total accumulated dose of 1012 e/cm2 , when the value of the
threshold potential will reach the biasing voltage.
In the complementary type ICs the most radiation resistant integrated
circuits were found to be those of RCA within the new A-series. However, to
substantiate this more tests are necessary with other types and batches of these
ICs. As compared to complementary, the p-channel IC made by AMI is the most
radiation resistant from those tested simply because it does not contain the n-
channels, which are prone to leakage first in most of the complementaries tested.
14
V. APPENDIX. Radiation Damage in MOS Devices
Radiation damage in the MOS devices is evidenced by two major effects:
a shift in the threshold potential and degradation in the transconductance.
A. Shift in the Threshold Potential
As an energetic particle (or a shower of secondary electrons resulting
from a primary energetic particle) passes through the gate insulator, it ionizes
the insulator, producing an electron-hole pair in the oxide. Some of the electrons
excited in this way migrate to the semiconductor and/or to the metal of the gate
contact, leaving behind a net positive charge in the insulator. This process is
schematically shown in Fig. 32. The accumulation of positive charges in the
gate oxide results in a shift of the current-voltage, I-V, characteristic along the
gate potential, which continues to increase as long as the positive charge in the
gate oxide continues to build-up. Since the excess build-up of charges in the gate
oxide is that of positive charge, the shift in I-V characteristic in both n- and
p-channel devices is toward more negative values of the gate potential. Examples
of the shifts in I-V characteristics are shown in Figs. 34 and 35 for n- and p-
channel MOS devices (GME), respectively, having the usual thermally grown
gate oxides. The numbers at each curve are the total accumulated dose of 1.5
MeV electrons per cm2 at the MOS devices unshielded except for the covers of
the flat packs. Strictly speaking, in this type of effect, the operation of the device
has not been really damaged in the full meaning of the word; the operation of the
15
device has been merely shifted to another region of the gate potential, if we
neglect the accompanying changes in the transconductance. This shift, which
usually can be reversed by thermal annealing4 ' 5 , is the first effect that one ob-
serves during irradiation and is the most harmful effect in the operation of and
MOS circuit as a whole.
Threshold potential of an MOS device, VGT , is usually defined as the gate
potential at which 10 .La of drain current flows and is the definition adhered
to in this report. In Figs. 34 and 35 the values of the threshold potentials are
the gate voltages at which I-V characteristics cross the 10 1Pa vertical line.
B, Degradation of the Transconductance
Transconductance, gi, is usually given in the specifications for single MOS
devices, and is defined as the change in the drain current per unit change in the
gate potential. In the data of this report it is measured in the vicinity of the
threshold potential. In generals transconductance characterizes the degree of
"cleanliness" of the channel region from the interface states and other charge
traps and gives a measure of "channel resistivity."
Irradiation causes a decrease in the transconductance in both n- and p-
channel MOS devices. This degradation is due to the production of interface
states by radiation. Interface states act as momentary charge traps, which
in the course of channel current flow trap charge carriers momentarily and
then reemit them again into the conduction band. This trapping and reemitting
16
of the charge carriers decreases their mobility and causes the resistivity of
the channel to increase. The decrease in the transconductance may be observed
in Fig. 34, for example, as the slopes of the I-V characteristics increases with
increasing accumulated radiation dose. In most of the MOS complementary inte-
grated circuits measured in this report the increase in the slopes is greater in
n-channels than in p-channels. This is probably due to the greater mobility of
electrons and the smaller physical size of the n-channel units as compared to
the p-channel units.
VI. EFFECTS OF THE SHIFT IN THE THRESHOLD POTENTIALS AND
DEGRADATION IN THE TRANSCONDUCTANCE ON THE OPERATION
OF MOS IC
In a complimentary IC the effects of the shifts in the threshold potentials
of both n- and p-channel devices and their degradation of the transconductances
have to be considered.
As the threshold potential of the n-channel device decreases due to radia-
tion (Fig. 34), currents increase because the n-channel device tends to conduct
and switch sooner. The voltage drop across the n-channel MOS device will
cause the output not to have a full power supply swing. Also, noise immunity
degrades rapidly as the threshold potential moves near the ground potential.
Finally, as the radiation dose continues to increase, the dynamic power consump-
tion becomes too excessive because the n-channels stay turned on even with
zero gate bias.
17
In p-channel devices radiation causes an increase in the threshold potential
(Fig. 35). When this occurs, switching slows down since it becomes harder to
turn the device on. Delays in the output signals increase since the switching
becomes more sluggish. This will occur until the device will no longer conduct.
At this point no static current exists, the circuit is unable to switch from one
state to another, and no logic function can be performed.
Whatever happens first, either the effects with n-channel or p-channel MOS
units in an IC, depends both on the charge trapping capabilities of the respective
gate oxides and the values of the gate voltages applied to the devices during
irradiation. Both shifts, occuring at the same time, multiply the effects and the
degradation becomes very complicated.
18
REFERENCES
1. W. L. Brown, J. D. Gabbe, and W. Rosenzweig, "Results of the Telstar
Radiation Experiments," The Bell Syst. Techn. J. p. 1505, July 1963,
2. V. Danchenko and S. S. Brashears, "Investigation of Radiation Damage in
MOSFETs Using Bias-Temperature Treatments," IEEE Transactions on
Nuclear Science, NS-15, 182 (1968).
3. V. Danchenko, J. D. Walker and S. S. Brashears, "Radiation-Induced Electron
Trapping in M.I.S. Transistors," Proceedings of Grenoble Conference on
Properties and Use of M.I.S. Structures, Grenoble, France, June 1969.
4. V. Danchenko, U. D. Desai and S. S. Brashears, "Characteristics of Thermal
Annealing of Radiation Damage in MOSFETs," J. Appl. Phys., 39, 2416 (1968).
5. V. Danchenko, U. D. Desai, J. M. Killiany and S. S. Brashears, "Effects of
Electric Fields on Annealing of Radiation Damage in MOSFETs," IEEE
Transactions on Electron Devices, ED-15, 751 (1968).
19
1 14 13
2 t·: W e | | 0 ; 1 2
-t:: :i.]: 'i~t'9 'V~t, j4.at :. _ _ ;....¼..f..... ,- _ ; _ ,__
3 ,' . I ( ' ;' . '
Z3 -4
B3 % ,fB4
)
GND
SOLID STATE SCIENTIFIC SCL -5102 - COMPLEMENTARYPOSITIVE NAND
Figure 1. Schematic and photograph of an irradiated IC made bySolid State Scientific, Inc
20
CONVERTED DOSE, RAD - Si (CO - 60)105 106
1012 1013INTEGRATED DOSE/cm 2 , 1.5 MeV ELECTRONS
Figure 2. Shift in threshold potentials in n-channel MOS devices of SolidState Scientific IC irradiated with various gate biases
21
6
54
3
2
1
0
-1
-2
-3-4
-5
-6-7
z
I
n
1-J
0_
0
o9'T'I'-
-81014
CONVERTED DOSE, RAD - Si (CO - 60)105 106
1012 1013
INTEGRATED DOSE/cm 2 , 1.5 MeV ELECTRONS
Figure 3. Same as Fig. 2, except p-channel devices
22
Hn
0-
U]a-
0
-_J0I,U0
r
0-1
-2
-3-4-5
-6-7
-8
-9-10
-11
-12
-13-14
1014
104
70
60
50
I 40
E 30to
20
10
0
104
CONVERTED DOSE, RAD - Si (CO-60)105 106
1012 10'3
INTEGRATED DOSE/cm2 , 1.5 MeV ELECTRONS
Figure 4. Degradation in transconductances, gm's, in n-channels of Solid State Scientific, Inc.,-as a function of total accumulated radiation dose
23
1014
CONVERTED DOSE, RAD - Si (CO - 60)
105
1012 1013
106
1014
INTEGRATED DOSE/cm2 , 1.5 MeV ELECTRONS
Figure 5. Same as Fig. 4, except p-channels
24
10470
-Ew
1000
100
CONVERTED DOSE, RAD - Si (CO - 60)
104 105 106
101
.1
,1, , , ,I , ,,,,i , ,I ,I I Iill I I.01 o102 1013 1014
INTEGRATED DOSE/cm 2 , 1.5 MeV ELECTRONSFigure 6. Leakage in n-channels of Solid State Scientific ICs as a
function of total accumulated radiation dose
25
zLUCrQ
z
I I II11 I I III I II I t I rlI I I I I III I I II I 1 I I I 111li
n-CHAN N ELS
SOLID STATE SCIENTIFIC, INC.SCL 5102
p-CHANNEL
I I i I 1 I I I I I I I I III I I I i i i,.1 -...
1 10 100 1,000TIME (HOURS) AT ROOM TEMPERATURE
Figure 7. Recovery of threshold potentials in. n-channel and p-channel MOS devicesof Solid State Scientific ICs at room temperature
10,000
26
+3
+2+1
O--1 -2 -
-3 -
-4 -
-5 -
-6 -7-8 -
-9
-10-11 .
-0
0-j
zI`J
0
Ld
I
I I I II
13
14
1
2
-L~I~ M I"II i
3 4 5
4 7 11VSS
RCA CD 4007 - COMPLEMENTARYDUAL COMPLEMENTARY PAIR PLUS INVERTER
Figure 8. Diagram and photograph of anirradiated RCA IC
27
1;, t,# ,
II
I -
J. >
.·
_L~ f. .*..
9
8
7
6
Vm� -- I
12 11 10
CONVERTED DOSE, RAD - Si (CO - 60)105 106
1012 1013
INTEGRATED DOSE/cm2 , 1.5 MeV ELECTRONSFigure 9. Shifts in the threshold potentials in n-channel MOS devices of RCA
IC bombarded with various gate biases
28
+4+3+2
o +1
O0
> -1
z -3
E-
o -4"-7O
LL
I -
-10
104
1014
CONVERTED DOSE, RAD -Si (CO - 60)
105 106
1012 10'3
INTEGRATED DOSE/cm 2 , 1.5 MeV ELECTRONS
Figure 10. Same as Fig. 9, except p-channels
29
0-1
a_ -2
o -3''-4
> -5<I -6z -7
o-80 -9-J0 -10w3 -11
I -12-13-14
104
1014
CONVERTED DOSE, RAD - Si (CO - 60)105 106
1012
INTEGRATED DOSE/cm2 ,10141013
1.5 MeV ELECTRONS
Figure 11. Degradation of the transconductances in n-channel and p-channelMOS devices of RCA
30
10470
60
50
20
10
0
I l II I1 I I I I1 I I I I I I I1 I
p-CHANNELS
INITIAL gm's
n-CHANNELS
RCA 4001
I I I I I I Ii 1 I 1 I I 1 1 1 I i I I II 1
40
E 30wo
CONVERTED DOSE, RAD - Si (CO - 60)
R(ICA 4 010
_ 1LL'L...L_ J L.
1011 1012 1013 1014
INTEGRATED DOSE/cm 2, 1.5 MeV ELECTRONS
Figure 12. Leakage in n-channels of RCA IC as a function oftotal accumulated radiation dose
31
1000
100
-<I:
LI_
10
1
.1
.01 L-
CONVERTED DOSE, RAD - Si (CO - 60)
1000 - 1r1111
100
10
1
.1
.011011
INTEGRATED1012
DOSE/cm2 ,
1013
1.5 MeV
1014
ELECTRONS
Figure 13. Same as Fig. 12, except total leakage in all 8n-channels
32
c
C-)
z
z
CONVERTED DOSE, RAD - Si (CO - 60)
1012 10'3INTEGRATED DOSE/cm 2 , 1.5 MeV ELECTRONS
Figure 14. Same as Figs. 9 and 10, except irradiated with 5V gate biases
33
-j0
-r
zUiF--0
CLL
0
-o00I
H -8-9-10
1014
CONVERTED DOSE, RAD - Si (CO - 60)
10570
60 -
50
40:L
E:o
10 4
I ' 'T I I I I I I I I I I106
I I T I rI I I I
I. n-CHANNELS
T
20 kRCA CD 4001
10o
I I L I Il L lliL
1012
I I I I I I I I I
1013
INTEGRATED DOSE/cm2, 1.5 MeV ELECTRONS
Figure 15. Same as Fig. 11, except irradiated with 5V gate biases
34
01014
-- - - - -I I- wrs2tRh
1e~i~ tIp- OHA NN E Ls
e m~~~~~~~~~~~~~~~~~~~
304 rc
CONVERTED DOSE, RAD - Si (CO - 60)
104 105 106
.01o 1 I I
INTEGRATED1012 1013 10 T4
DOSE/cm 2 , 1.5 MeV ELECTRONS
Figure 16. Same as Fig. 12, excepted irradiated with5V gate biases
35
1000
100
10
1
a:
zUi.J
0
z
.1
CONVERTED DOSE, RAD - Si (CO - 60)104 105 106
.01 -- Iw
1011INTEGRATED
1012
DOSE1013 1014
/cm2 , 1.5 MeV ELECTRONS
Figure 17. Same as Fig. 13, except irradiated with 5V gate biases
36
1000
100
10
1
Z-a-
LU
zUr
M
.1
0
11 10 912
13
14 _____ 6
14
1 2 3 414 2 11
7 4 9
RCA CD 4007- COMPLEMENTARYDUAL COMPLEMENTARY PAIR PLUS INVERTER
Figure 18. Schematic and photograph of irradiatedRCA CD 4007 IC
37
4CONVERTED DOSE, RAD - Si (CO - 60)
105
·° -J INITIAL VGT'So -1
-2 n-CHANNELS
-4F --
° -6 p-CHANNELS
3 -7o RCA CD 4007I: - 8
In BOMBARDMENT BIAS:-9 - O -OV
-l-o x L -l, 10 V, .1 MHz-11 / - 10 V, FULLY ON-121
-120loll 1012 1013 1014
INTEGRATED DOSE/cm2 , 1.5 MeV ELECTRONS
Figure 19. Shifts in threshold potentials of n- and p-channel MOD devices ofRCA IC, irradiated with various gate biases
38
CONVERTED DOSE, RAD - Si (CO - 60)105
I I I I I I I I
1061 1 I I I I I II
p-CHANNELS
BOMBARDMENT BIAS:
o - OVX - LFJ, 10V, .1 MHzA - 10OV, FULLY ON
RCA CD 4007n-CHANNELS
I I I I II III I I I I I I II
1012
I I I I I I I I
1013 1014
INTEGRATED DOSE/cm2 , 1.5 MeV ELECTRONS
Figure 20. 'Degradation of the transconductances in n- and p-channels of RCA IC as a functionof total accumulated radiation dose
39
104I I 1 I II70
60 H
)
40-:::_EU0 30 H
20
10 F
0
5(
, =E...........
CONVERTED DOSE, RAD - Si (CO - 60)
1000
100
-- 10zLL
z 1
.1
.01i J l1011
INTEGRATED
1012 1013 1014
DOSE/cm 2 , 1.5 MeV ELECTRONS
Figure 21. Leakage in n-channels of RCA IC, irradiatedwith various gate biases
40
- V Wu 1000TIME (DAYS) AT ROOM TEMPERATUREFigure 22. Recovery of threshold potentials in n- and p-channel MOS of RCAIC at room temperature
41
-9-10-11-12
. fof
CONVERTED DOSE, RAD - Si (CO - 60)
1012 10'3
INTEGRATED DOSE/cm 2 , 1.5 MeV ELECTRONS
Figure 23. Shifts in the threshold potentials of n- and p-channel MOS devices of RCA newnew A-series, irradiated with various gate biases
42
o-j
0
0
--
LLI
0
'-I
1014
CONVERTED DOSE, RAD-Si (CO-60)105 106
1012 1013
INTEGRATED DOSE/cm 2 , 1.5 MeV ELECTRONS
Figure 24: Degradation in transconductances in n- and p-channels of RCAnew A-series, as a function of total accumulated radiation dose
43
70
60
50
I. 40:
E 30
20
10
104
CONVERTED DOSE, RAD-Si (CO-60)104 105 106
INTEGRATED1012
DOSE/cm 2 ,101 3
1.5 MeV1014
ELECTRONS
Figure 25. Leakage in n-channels of RCA new A-series
44
1000
100o
10n<i
z
z
0
I II11 I I I 111 1 III I e1 I
RCA CD 4001An-CHANNELBOMBARDMENT
BIAS:+10V FULLY ON
_ J i Il l{ I ,_ 1 , l 1 I ilt
0.1
0.01
1I
12 11 10 9
1313_4 ;,.a. f 1 ... .
1 8xJ " ~,
'1 - . . 6 . .
3 4 5VDD
1 P P 8GF,>t .10
6 P P 13
Motorola IC4 7
VSS
CONVERTED DOSE, RAD - Si (CO - 60)
105 106
1012 1013INTEGRATED DOSE/cm2 , 1.5 MeV ELECTRONS
Figure 27. Shifts in the threshold potentials of n- and p-channel MOS devices of MotorolaIC, irradiated with various gate biases
46
(n
2
_J
zLU0a_
0-j0
0LU
o104
10 4
CONVERTED DOSE, RAD - Si (CO - 60)
105104 106
1012 1013INTEGRATED DOSE/cm 2 ,1 .5 MeV ELECTRONS
Figure 28. Degradation in tansconductances in n- and p-channels of Motorola IC
47
70
Eto
01014
CONVERTED DOSE, RAD - Si (CO - 60)
1000
100
10
1
.1
.011012 1013 1014
INTEGRATED DOSE/cm2 , 1.5 MeV ELECTRONS
Figure 29. Leakages in n-channels of Motorola IC, irradiatedwith various gate biases
48
It
zL
0
M
20 19 18 17 16 2222 15
21
1
2
3
4
14
13
22
12
11
5 6 7,7 8 9 10
AMI S - 1304 OR MXO3C - p-CHANNEL, 10 CHANNEL SWITCH
Figure 30. Schematic and photograph of an irradiated AMI p-channel IC
49
CONVERTED DOSE, RAD - Si (CO - 60)105104 106
1012 1013INTEGRATED DOSE/cm 2 , 1.5 MeV ELECTRONS
Figure 31. Shifts in the threshold potenials of p-channel MOS devices of AMI p-channel IC,irradiated with various gate biases
50
(-,J0
c
-jO
0
-J0
U)LUT~
0-1-2-3
-5-6-7
-8-9
-10-11-12-13-14
1014
CONVERTED DOSE, RAD - Si (CO - 60)104 105 106
70 [ Il II [ I
60
50 INITIAL gm's
20 -CL
E 30 2
p-CHANNEL
AMI S-1304 OR MX03C
10 1
1 112 1013 10 1 4
INTEGRATED DOSE/cm 2 , 1.5 MeV ELECTRONS
Figure 32. Degradation in transconductonces in p-channels of AMI p-channel IC
51
IONIZING RADIATION
ELECTRON RECOMBINATION ELECTRONSTRAPPING LEAVING OXIDE
Figure 33. Radiation damage in a single MOS device unit
52
+4+3
+2Q)2~~~~~ + 1 L INITIAL
122-0O~~~~~ O F 1.2xO 10 2e/c m2+2 /- i4.1x1012 e/cm2
a> 1 6.9x10 lOe /c-2i-3
2
Fz 3.2x10 0e/cm-4
0-5-
(_ 0-60 ~10
-810
DRAIN CURRENT I D, aFigure 34. Shift of the I'V characteristic of channel MOS device
radiatin total accumulated doses (GM) wih various
53
1000
-1
-2 - INITIAL
0 -5 1.2102ecm2
X -..2nxlo e/cm
-1-0
-14
.01 .1 1 10 100 1000
DRAIN CURRENT, IDS A
Figure 35. Same as Fig. 34, except p-channel MOS device
54
LIST OF FIGURES
Figure 1. Schematic and photograph of an irradiated IC made by Solid State
Scientific, Inc.
Figure 2. Shift in threshold potentials in n-channel MOS devices of Solid State
Scientific IC irradiated with various gate biases.
Figure 3. Same as Figure 2, except p-channel devices.
Figure 4. Degradation in transconductances, gi's, in n-channeles of Solid State
Scientific, Inc., as a function of total accumulated radiation dose.
Figure 5. Same as Figure 4, except p-channels.
Figure 6. Leakage in n-channels of Solid State Scientific ICs as a function of
total accumulated radiation dose.
Figure 7. Recovery of threshold potentials in n-channel and p-channel MOS
devices of Solid State Scientific ICS at room temperature.
Figure 8. Diagram and photograph of an irradiated RCA IC.
Figure 9. Shifts in the threshold potentials in n-channel MOS devices of RCA IC
bombarded with various gate biases.
Figure 10. Same as Figure 9, except p-channels.
Figure 11. Degradation of the transconductances in n-channel and p-channel
MOS devices of RCA.
Figure 12. Leakage in n-channels of RCA IC as a function of total accumulated
radiation dose.
Figure 13. Same as Figure 12, except total leakage in all 8 n-channels.
55
Figure 14. Same as Figures 9 and 10, except irradiated with 5V gate biases.
Figure 15. Same as Figure 11, except irradiated with 5V gate biases.
Figure 16. Same as Figure 12, except irradiated with 5V gate biases.
Figure 17. Same as 13, except irradiated with 5V gate biases.
Figure 18. Schematic and photograph of irradiated RCA CD 4007 IC.
Figure 19. Shifts in threshold potentials of n- and p-channel MOS devices of
RCA IC, irradiated with various gate biases.
Figure 20. Degradation of the transconductances in n- and p-channels of RCA IC
as a function of total accumulated radiation dose.
Figure 21. Leakage in n-channels of RCA IC, irradiated with various gate biases.
Figure 22. Recovery of threshold potentials in n- and p-channel MOS of RCA IC
at room temperature.
Figure 23. Shifts in the threshold potentials of n- and p-channel MOS devices of
RCA new A-series, irradiated with various gate biases.
Figure 24. Degradation in transconductances in n- and p-channels of RCA new
A-series, as a function of total accumulated radiation dose.
Figure 25. Leakage in n-channels of RCA new A-series.
Figure 26. Diagram and photograph of irradiated Motorola IC.
Figure 27. Shifts in the threshold potentials of n- and p-channel MOS devices
of Motorola IC, irradiated with various gate biases.
Figure 28. Degradation in transconductances in n- and p-channels of Motorola IC.
Figure 29. Leakages in n-channels of Motorola IC, irradiated with various gate
biases.
56
Figure 30. Schematic and photograph of an irradiated AMI p-channel IC.
Figure 31. Shifts in the threshold potentials of p-channel MOS devices of AMI
p-channel IC, irradiated with various gate biases.
Figure 32. Degradation in transconductances in p-channels of AMI p-channel IC.
Figure 33. Radiation damage in a single MOS device unit.
Figure 34. Shift of the I-V characteristic of n-channel MOS device (GME) with
various radiation total accumulated doses.
Figure 35. Same as Figure 34, except p-channel MOS device.
57