preprint t r 'r. , , a j6s76 7 · · 2013-08-31preprint, , t ~ 9 ' r ~ 'r ... 2....

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


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



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


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


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


Figure 5. Same as Fig. 4, except p-channels

24

10470

-Ew

1000

100


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


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


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


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


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-


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


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


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


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


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


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...........


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


1012 10'3


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


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


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


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


1000

100

10

1

.1

.011012 1013 1014


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


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


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

preprint t r 'r. , , a j6s76 7 · · 2013-08-31preprint, , t ~ 9 ' r ~ 'r ... 2....

Documents