solar activity and ozone layer depletion - Inter Islamic Network on ...
solar activity and ozone layer depletion - Inter Islamic Network on ...
solar activity and ozone layer depletion - Inter Islamic Network on ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
SOLAR ACTIVITY AND OZONE LAYER DEPLETION<br />
COMPARISION OF PARAMETRIC<br />
EVALUATION OF TOTAL OZONE MAPPING<br />
SPECTROPHOTOMETER (TOMS) DATA<br />
NIMBUS -7 WITH THE DOBSON SPECTROPHOTOMETER<br />
OBSERVATIONS AT<br />
THE GEOPHYSICAL CENTER QUETTA, PAKISTAN<br />
M.Ayub Khan Yousuf Zai, Jawaid Quamar,M.Rashid Kamal Ansari,<br />
Jawaid<br />
Iqbal <str<strong>on</strong>g>and</str<strong>on</strong>g> Arif Hussian<br />
Institute of Space <str<strong>on</strong>g>and</str<strong>on</strong>g> Planetary Astrophysics <str<strong>on</strong>g>and</str<strong>on</strong>g> Department of Applied<br />
Physics, University of Karachi<br />
Abstract<br />
In this communicati<strong>on</strong> we present some approaches to relate the <str<strong>on</strong>g>depleti<strong>on</strong></str<strong>on</strong>g><br />
process with the <str<strong>on</strong>g>solar</str<strong>on</strong>g> <str<strong>on</strong>g>activity</str<strong>on</strong>g>. We propose to study the effectiveness of <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g><br />
<str<strong>on</strong>g>layer</str<strong>on</strong>g> structure. Developing a quantitative treatment for data covering a specific<br />
period, this study will report for Pakistan an explorative assessment of the link<br />
between <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> <str<strong>on</strong>g>layer</str<strong>on</strong>g> <str<strong>on</strong>g>depleti<strong>on</strong></str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>solar</str<strong>on</strong>g> <str<strong>on</strong>g>activity</str<strong>on</strong>g>. We have computed<br />
parameters, which signify the measurements drawn from the satellite, <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
ground based instruments. We will also try to look for underst<str<strong>on</strong>g>and</str<strong>on</strong>g>ing the<br />
interacti<strong>on</strong> of dynamics <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> photochemistry in the middle atmosphere (<br />
25-35 km) <str<strong>on</strong>g>and</str<strong>on</strong>g> the nature of stratospheric resp<strong>on</strong>ses to anomalous <str<strong>on</strong>g>solar</str<strong>on</strong>g> <str<strong>on</strong>g>activity</str<strong>on</strong>g>.<br />
We have compared both the data sets using different techniques <str<strong>on</strong>g>and</str<strong>on</strong>g> observed<br />
that the trend follows the same path. This could be helpful in obtaining the<br />
estimates of sunspots <str<strong>on</strong>g>and</str<strong>on</strong>g> establishing a correlati<strong>on</strong> with the <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> <str<strong>on</strong>g>layer</str<strong>on</strong>g><br />
reducti<strong>on</strong> then suitable predicti<strong>on</strong>s can be made for our government <str<strong>on</strong>g>and</str<strong>on</strong>g> private<br />
departments<br />
Introducti<strong>on</strong><br />
Solar studies are of importance in general astr<strong>on</strong>omy because the sun is the <strong>on</strong>ly<br />
star up<strong>on</strong> which surface detail an be resolved The investigators find the variety of<br />
<str<strong>on</strong>g>solar</str<strong>on</strong>g> structures <str<strong>on</strong>g>and</str<strong>on</strong>g> events to decide their physical causes. For our society, the<br />
study of the sun is more indirect but of great importance because <str<strong>on</strong>g>solar</str<strong>on</strong>g> radiati<strong>on</strong><br />
is essential to our biosphere. Cyclicity of <str<strong>on</strong>g>solar</str<strong>on</strong>g> <str<strong>on</strong>g>activity</str<strong>on</strong>g> (<str<strong>on</strong>g>solar</str<strong>on</strong>g> cycle) could be a<br />
factor supporting the emergence of OLD (Solar terrestrial relati<strong>on</strong>s analysis). It is<br />
suspected that Cl2 is produced during active period of the sun <str<strong>on</strong>g>and</str<strong>on</strong>g> this producti<strong>on</strong><br />
of Cl2 causes Oz<strong>on</strong>e <str<strong>on</strong>g>layer</str<strong>on</strong>g> <str<strong>on</strong>g>depleti<strong>on</strong></str<strong>on</strong>g>.<br />
The rates of <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> removal in the year 1992 are enhanced by volcanic erupti<strong>on</strong>s<br />
in 1991 <str<strong>on</strong>g>and</str<strong>on</strong>g> 1992. Because anthropogenically introduced chlorine(Cl) <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
bromine(Br) levels will remain high for so l<strong>on</strong>g, it is expected that there will be
more <str<strong>on</strong>g>and</str<strong>on</strong>g> more UV-B reaching the earth’s surface. Man-made causes of OLD<br />
are providing a threshold for OLD. We can not neglect the natural events which<br />
are occurring in the upper atmosphere <str<strong>on</strong>g>and</str<strong>on</strong>g> producing certain chemical<br />
c<strong>on</strong>stituents which react with O3 <str<strong>on</strong>g>and</str<strong>on</strong>g> dissociate it into nascent oxygen <str<strong>on</strong>g>and</str<strong>on</strong>g> a<br />
molecular oxygen an can be seen by Chapman reacti<strong>on</strong>s. Therefore, in view of<br />
this c<strong>on</strong>cepti<strong>on</strong> the M<strong>on</strong>treal Protocol is not to be all of the things.[1-8]. These<br />
investigati<strong>on</strong>s led to series of processes known as Chapman mechanism which<br />
explained the formati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> annihilati<strong>on</strong> of O3 <str<strong>on</strong>g>layer</str<strong>on</strong>g>. Using this mechanism,<br />
predicti<strong>on</strong>s were made which provided excellent agreement with observati<strong>on</strong>s<br />
available then [30-35]. The situati<strong>on</strong> can be explained by the following reacti<strong>on</strong>s:<br />
O2 + h ν → O + O<br />
(1)<br />
O + O2 + M → O3 + M<br />
(2)<br />
O3 + hν → O + O2<br />
(3)<br />
O + O3 → O2 + O2<br />
(4)<br />
M being a third body required carrying off the excess energy of the associati<strong>on</strong><br />
process. The rapid cycle composed of reacti<strong>on</strong> (1.2) <str<strong>on</strong>g>and</str<strong>on</strong>g> (1.3) does not, of<br />
course, destroy O3. Many years after the preceding initial work, it was suggested<br />
that catalytic processes can lead to the overall reacti<strong>on</strong> as follows:<br />
(5)<br />
O + O3 → 2 O2<br />
Oz<strong>on</strong>e c<strong>on</strong>tent in the stratosphere varies due to various natural <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
anthropogenic activities. Some prominent such activities creating OLD are<br />
described as follows<br />
(a) Technological<br />
CFCs produced in various technological processes gradually migrate upward into<br />
the stratosphere. At altitude above 30 km, intense UV radiati<strong>on</strong> breaks down the<br />
CFCs, causing them to release chlorine <str<strong>on</strong>g>and</str<strong>on</strong>g> destroying <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> as shown in the<br />
following chemical reacti<strong>on</strong>s:<br />
CFCl3 + h ν → CF Cl2 + Cl −<br />
(6)<br />
CF2Cl2 + h ν→ CF2 Cl + Cl<br />
(7)
Cl + O3 → Cl O + O2<br />
(8)<br />
During the past few years, satellite <str<strong>on</strong>g>and</str<strong>on</strong>g> ground-based observati<strong>on</strong>s have<br />
particularly c<strong>on</strong>tributed to our underst<str<strong>on</strong>g>and</str<strong>on</strong>g>ing of the general features of the<br />
vertical <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> distributi<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> latitude variati<strong>on</strong>s in the stratosphere <str<strong>on</strong>g>and</str<strong>on</strong>g> lower<br />
mesosphere. Oz<strong>on</strong>e photochemistry in the middle atmosphere (25-35 km)<br />
will c<strong>on</strong>tribute to link stratospheric resp<strong>on</strong>ses to anomalous <str<strong>on</strong>g>solar</str<strong>on</strong>g> <str<strong>on</strong>g>activity</str<strong>on</strong>g><br />
(b) Solar Cycle<br />
In the stratosphere sun drives the photochemistry as well as providing the<br />
heating that drives the <str<strong>on</strong>g>solar</str<strong>on</strong>g> circulati<strong>on</strong>. Because the atmosphere absorbs UV<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> shorter wavelength radiati<strong>on</strong>, accurate determinati<strong>on</strong> of its time variati<strong>on</strong><br />
should be measured from above the top of the atmosphere. Some informati<strong>on</strong><br />
about <str<strong>on</strong>g>solar</str<strong>on</strong>g> variability can be obtained by c<strong>on</strong>siderati<strong>on</strong> of l<strong>on</strong>ger wavelength<br />
features that do penetrate to the ground, as the same mechanisms that give rise<br />
to the l<strong>on</strong>ger-wavelength variati<strong>on</strong>s the most often is the radio radiati<strong>on</strong> with a<br />
10.7-micr<strong>on</strong> wavelength, albeit the other proxies such as sunspot numbers can<br />
also be used. Two of the most important periods for <str<strong>on</strong>g>solar</str<strong>on</strong>g> variati<strong>on</strong>s are 27 days<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> 11 years.[9-22]<br />
The <str<strong>on</strong>g>solar</str<strong>on</strong>g> cycle dependence of atmospheric <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> also has an altitude<br />
dependence which maximises at approximately 45 km. The temperature changes<br />
of the upper stratsphere are rather caused by the <str<strong>on</strong>g>solar</str<strong>on</strong>g> cycle variati<strong>on</strong>. Basis for<br />
the <str<strong>on</strong>g>solar</str<strong>on</strong>g> cycle theory was the noti<strong>on</strong> that reactive nitrogen compounds (NO <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
NO2), like chlorine can destroy stratospheric <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> quite effectively. These<br />
compounds are produced in abundance in the mesosphere <str<strong>on</strong>g>and</str<strong>on</strong>g> lower<br />
thermosphere (∼ 60-70 km) during period of high <str<strong>on</strong>g>solar</str<strong>on</strong>g> <str<strong>on</strong>g>activity</str<strong>on</strong>g>. The nitrogen<br />
compounds are rapidly destroyed in the mesosphere, but might be transported all<br />
to the stratosphere in the dark polar winter.<br />
(c) Solar Prot<strong>on</strong> Events<br />
Sometimes the sun emits large amount of prot<strong>on</strong>s which are channeled by the<br />
earth’s magnetic field to impact the earth at high altitudes. These are referred to<br />
as <str<strong>on</strong>g>solar</str<strong>on</strong>g> prot<strong>on</strong> events (SPEs). For prot<strong>on</strong>s of sufficient energy, they can reach<br />
the mesosphere <str<strong>on</strong>g>and</str<strong>on</strong>g> stratosphere <str<strong>on</strong>g>and</str<strong>on</strong>g> lead to reducti<strong>on</strong> of <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> by producing<br />
increased c<strong>on</strong>centrati<strong>on</strong>s of odd nitrogen <str<strong>on</strong>g>and</str<strong>on</strong>g> / or odd hydrogen compounds in<br />
the stratosphere. A principal advance in the determinati<strong>on</strong> of <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g><br />
distributi<strong>on</strong>s <str<strong>on</strong>g>and</str<strong>on</strong>g> variability came with the advent of satellite-based systems.<br />
The most important of these are given as follows:<br />
(1) Total Oz<strong>on</strong>e Mapping Spectrometer (TOMS) instruments for total <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g><br />
measurements<br />
(2) Solar Backscatter Ultraviolet (SBUV,I <str<strong>on</strong>g>and</str<strong>on</strong>g> II )<br />
(3) Stratosphere Aerosol <str<strong>on</strong>g>and</str<strong>on</strong>g> Gas Experiment (SAGE, I <str<strong>on</strong>g>and</str<strong>on</strong>g>II) instruments for<br />
vertical <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> distributi<strong>on</strong>s
(d) Investigating the correlati<strong>on</strong>s between the UV-B due to OLD <str<strong>on</strong>g>and</str<strong>on</strong>g> the<br />
<str<strong>on</strong>g>solar</str<strong>on</strong>g> variati<strong>on</strong>s using modeling strategies<br />
Oz<strong>on</strong>e Layer depths (Dobs<strong>on</strong> Units)<br />
420<br />
380<br />
340<br />
300<br />
260<br />
220<br />
sunspot<br />
y=279.266+0.051*x+eps<br />
180<br />
-20 20 60 100 140 180 220<br />
Sun spot numbers (Counts)<br />
Figure.1. Correlati<strong>on</strong> between <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> c<strong>on</strong>centrati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> the sunspot numbers<br />
Measurements from satellite can give near-global covering, however, it provides<br />
<strong>on</strong>ly averaged values of the measured quantity over large distances both in<br />
vertical <str<strong>on</strong>g>and</str<strong>on</strong>g> horiz<strong>on</strong>tal directi<strong>on</strong>s. So we have preferably utilised Dobs<strong>on</strong> data for<br />
our computati<strong>on</strong>s <str<strong>on</strong>g>and</str<strong>on</strong>g> compared these with the data. down loaded from satellite<br />
(TOMS, Nimbus-7 ). As far as the trend <str<strong>on</strong>g>and</str<strong>on</strong>g> cumulative behaviour of TOMS<br />
satellite data <str<strong>on</strong>g>and</str<strong>on</strong>g> Dobs<strong>on</strong> data measured with the help of ODD<br />
spectrophotometer that was installed at Geophysical Center Quetta for the<br />
durati<strong>on</strong> 1979-1993 are c<strong>on</strong>cerned we have tried to dem<strong>on</strong>strate through precise<br />
computati<strong>on</strong>s the comparis<strong>on</strong> between these two data sets.<br />
<str<strong>on</strong>g>Inter</str<strong>on</strong>g>nati<strong>on</strong>al Oz<strong>on</strong>e <str<strong>on</strong>g>Network</str<strong>on</strong>g> presently utilizes <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> depth detecting<br />
spectrophotometer (ODDS). On board the satellite, incident <str<strong>on</strong>g>solar</str<strong>on</strong>g> radiati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
backscattered UV sunlight are measured. Total <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> is derived from these<br />
measurements. To map total <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g>, total <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> mapping spectrometer (TOMS)<br />
instrument scan through the sub-satellite point in a directi<strong>on</strong> perpendicular to the<br />
orbit plane [18][23]. Details of these instruments could be found in every <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g><br />
detecti<strong>on</strong> techniques literature.<br />
Satellite-borne instruments (e.g. <strong>on</strong> Nimbus-7, Meteor-3, ADEOS <str<strong>on</strong>g>and</str<strong>on</strong>g> Earth’s<br />
probe) are used to measure decreases in the c<strong>on</strong>centrati<strong>on</strong> of stratospheric
Parameters<br />
/ Years<br />
<str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> in our atmosphere. Total Oz<strong>on</strong>e Mapping Spectrometer (TOMS) data<br />
(Nimbus-7) are available for the period from 1979-1993.<br />
A comparis<strong>on</strong> TOMS <str<strong>on</strong>g>and</str<strong>on</strong>g> Dobs<strong>on</strong> values using coefficient of variati<strong>on</strong> (CV) for<br />
each year of both the data sets, observing trends <str<strong>on</strong>g>and</str<strong>on</strong>g> calculating some other<br />
parameters is given in Table.1.<br />
TABLE.1: Comparis<strong>on</strong> of parametric computati<strong>on</strong> for TOMS satellite <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
Dobs<strong>on</strong> data during the sessi<strong>on</strong> 1979-1993 <str<strong>on</strong>g>and</str<strong>on</strong>g> Dobs<strong>on</strong> data<br />
during the sessi<strong>on</strong> 1979-1993<br />
Mean of TOMS<br />
/Dobs<strong>on</strong><br />
X<br />
TR – Mean<br />
TOMS /Dobs<strong>on</strong><br />
TR – X<br />
Sta. Dev.<br />
TOMS<br />
/Dobs<strong>on</strong><br />
σ<br />
SE mean<br />
TOMS<br />
/Dobs<strong>on</strong><br />
SE- X<br />
Coefficient of<br />
Variati<strong>on</strong><br />
TOMS<br />
/Dobs<strong>on</strong><br />
(CV)<br />
1979 285.76/ 298.25 284.18/296.32 23.01/21.79 1.30 / 6.29 0.08 / 0.07<br />
1980 277.50/ 294.00 276.37/293.23 18.41/11.27 0.98 / 3.25 0.07 / 0.04<br />
1981 280.89/291.50 279.62/ 290.45 16.85/13.32 0.88 / 3.84 0.06 / 0.05<br />
1982 287.07/301.25 285.28/299.13 24.85/20.46 1.31/5.91 0.08 / 0.07<br />
1983 272.55/283.92 271.64/281.34 18.45/8.93 0.97 / 2.56 0.07 / 0.03<br />
1984 275.88/ 287.67 275.83/285.41 21.81/17.07 1.14/4.93 0.08 / 0.06<br />
1985 262.09/ 277.42 262.15/276.21 12.71/ 6.68 0.67 / 1.93 0.05/ 0.02<br />
1986 272.01/ 292.58 271.43/ 290.65 18.52/13.81 0.97 / 3.98 0.07 / 0.05<br />
1987 270.81/ 286.92 271.12/ 285.22 23.12/20.64 1.21 / 5.96 0.08 / 0.07<br />
1988 264.50/281.83 264.31/280.34 19.70/ 7.93 1.03/2.28 0.07 / 0.03<br />
1989 279.62/297.75 279.13/280.34 15.43/10.57 0.81 / 3.05 0.05 / 0.04<br />
1990 274.46/282.83 273.87/280.32 16.73/27.07 0.83 / 7.80 0.06 / 0.09<br />
1991 280.25/297.83 280.23/295.11 20.52/15.39 1.08 / 4.44 0.07 / 0.05<br />
1992 272.70/278.33 271.45/ 277.76 23.27/18.49 1.22 / 5.33 0.08 / 0.07<br />
1993 262.54/264.92 262.77/ 263.22 18.10/5.82 1.61 / 1.68 0.06 / 0.02<br />
TABLE. 2: Trend equati<strong>on</strong>s<br />
1 TOMS-Nimbus – 7 Yt = 281.37 0 .85 t + ε<br />
2 Dobs<strong>on</strong> Instrument Yt = 297.23 1.179 t + ε<br />
This justifies our adopti<strong>on</strong> of Dobs<strong>on</strong> data for computati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> modeling. Data<br />
having less variati<strong>on</strong> is more stable <str<strong>on</strong>g>and</str<strong>on</strong>g> vice versa. Coefficient of variati<strong>on</strong> hence<br />
determines the stability or c<strong>on</strong>sistency of the data.
Oz<strong>on</strong>e depth in DU for Quetta Pakistan<br />
305<br />
300<br />
295<br />
290<br />
285<br />
280<br />
275<br />
270<br />
265<br />
260<br />
260<br />
0 2 4 6 8 10 12 14 16<br />
Time in years (1979-1993)<br />
290<br />
284<br />
278<br />
272<br />
266<br />
Oz<strong>on</strong>e depth in DU from TOMS Nimbus-7<br />
NEWVAR18 (L)<br />
NEWVAR19 (R)<br />
Figure. 2. Manifestati<strong>on</strong> of the trends of TOMS <str<strong>on</strong>g>and</str<strong>on</strong>g> Dobs<strong>on</strong> data sets<br />
Oz<strong>on</strong>e depth from Dobs<strong>on</strong><br />
305<br />
300<br />
295<br />
290<br />
285<br />
280<br />
275<br />
270<br />
265<br />
260<br />
260<br />
0 2 4 6 8 10 12 14 16<br />
Time in years (1979-1993)<br />
290<br />
284<br />
278<br />
272<br />
266<br />
Oz<strong>on</strong>e depth from TOMS Nimbus-7<br />
DOBSON (L)<br />
TOMS (R)<br />
Figure 3: Depicti<strong>on</strong> of the trends of TOMS <str<strong>on</strong>g>and</str<strong>on</strong>g> Dobs<strong>on</strong> data sets with 95 %<br />
c<strong>on</strong>fidence level.
Total Oz<strong>on</strong>e from TOMS Nimbus-7<br />
305<br />
300<br />
295<br />
290<br />
285<br />
280<br />
275<br />
270<br />
265<br />
260<br />
260<br />
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16<br />
Time in years (1979-1993)<br />
290<br />
284<br />
278<br />
272<br />
266<br />
Total Oz<strong>on</strong>e from Dobs<strong>on</strong><br />
DOBSON (L)<br />
TOMS (R)<br />
Figure. 4: Comparis<strong>on</strong> of ground-based data recorded at Quetta Pakistan<br />
<strong>on</strong> Dobs<strong>on</strong> Spectrophotometer with the<br />
TOMS Nimbus-7<br />
The trend equati<strong>on</strong>s are also evolved using least square methods as shown in<br />
table 2. These equati<strong>on</strong>s could be used for future predicti<strong>on</strong>s for both the data<br />
sets. Accuracy of the Dobs<strong>on</strong> Unit (DU) is based <strong>on</strong> the laboratory<br />
measurements of the <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> absorpti<strong>on</strong> coefficients as menti<strong>on</strong>ed above. An<br />
error in the absorpti<strong>on</strong> coefficients is due to the b<str<strong>on</strong>g>and</str<strong>on</strong>g> pass <str<strong>on</strong>g>and</str<strong>on</strong>g> the <str<strong>on</strong>g>solar</str<strong>on</strong>g> spectral<br />
shape. The error in the measurements of the absolute DU scale can be<br />
estimated as ranging from 3 % to 5 %.<br />
Figure.5.
Depth (DU)<br />
Oz<strong>on</strong>e depth (DU)<br />
340<br />
330<br />
320<br />
310<br />
300<br />
290<br />
280<br />
270<br />
260<br />
250<br />
Jan Apr M<strong>on</strong>ths<br />
Aug Dec<br />
1982.<br />
Figure 5: Secular trend by least square tech. for the year 1982<br />
306<br />
304<br />
302<br />
300<br />
298<br />
296<br />
294<br />
292<br />
290<br />
288<br />
1970 1975 1980 1985 1990 1994<br />
Time in year<br />
Figure 6: Secular trend by least square method for April<br />
The secular trends for both the data sets have manifest that the <str<strong>on</strong>g>depleti<strong>on</strong></str<strong>on</strong>g> at<br />
Pakistan coastal regi<strong>on</strong> has got some value albeit the rate is small. With the help<br />
of trend equati<strong>on</strong> we are in the positi<strong>on</strong> to predict future status of <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> <str<strong>on</strong>g>layer</str<strong>on</strong>g><br />
<str<strong>on</strong>g>depleti<strong>on</strong></str<strong>on</strong>g> not <strong>on</strong>ly for Pakistan but also for other countries of the regi<strong>on</strong>.<br />
Albeit the from satellite can give near-global covering, but can provide <strong>on</strong>ly<br />
averaged values of the measured quantity over large regi<strong>on</strong>s both in vertical It is<br />
rather difficult as well as expensive to properly h<str<strong>on</strong>g>and</str<strong>on</strong>g>le the working of satellite
instruments because of the variati<strong>on</strong>s in the speed <str<strong>on</strong>g>and</str<strong>on</strong>g> orbit of the satellites.<br />
Also, in case of failure of any of the comp<strong>on</strong>ents, necessary repairs take time<br />
that influences the accuracy of data. Whereas, the ground based radars can<br />
provide data with high vertical resoluti<strong>on</strong> by measuring small differences in the<br />
time delays of the return pulses above the radar sites.<br />
REFERENCES<br />
(1) Andrews G D (2000), “An Introducti<strong>on</strong> to Atmospheric Physics”, Cambridge<br />
University Press, L<strong>on</strong>d<strong>on</strong><br />
(2) Johnst<strong>on</strong>, H S (1971), “Reducti<strong>on</strong> of stratospheric <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> by oxide catalyst<br />
supers<strong>on</strong>ic transport”, Science, 173: 517-522<br />
(3) Mészàros E (1995), “Global <str<strong>on</strong>g>and</str<strong>on</strong>g> Regi<strong>on</strong>al Changes in Atmospheric<br />
Compositi<strong>on</strong>”, CRC Press Inc., 2000 Corporate Blvd., Florida<br />
(4) Box G E P <str<strong>on</strong>g>and</str<strong>on</strong>g> Jenkins G M (1976),” Time series analysis, forecasting <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
c<strong>on</strong>trol”, rev. ed. Holden-Day, San Francisco<br />
(5) Bojkov R, Bishop L, Hill W J, <str<strong>on</strong>g>and</str<strong>on</strong>g> Tiao G C (1990), “A statistical trend<br />
analysis of revised Dobs<strong>on</strong> total <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> data over the northern hemisphere”,<br />
J. Geophs. Res. 95: 9785-9808<br />
(6) Stolarski R S, Bloomfield P, Herman J R, (1991), “Total <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> trends<br />
deduced from Nimbus 7 TOMS data”, Geophys. Res. Lett. 18: 1015-1018<br />
[7]. Gupta S (1999), “Research Methodology <str<strong>on</strong>g>and</str<strong>on</strong>g> Statistical Techniques”, Deep<br />
& Deep Publicati<strong>on</strong>s, New Delhi.<br />
[8] Granger C W J, <str<strong>on</strong>g>and</str<strong>on</strong>g> Newbold P (1986), Forecasting Ec<strong>on</strong>omic Time Series,<br />
2nd.Edn., Academic Press, New York<br />
[9] Priestley M B(1988), N<strong>on</strong>-linear <str<strong>on</strong>g>and</str<strong>on</strong>g> N<strong>on</strong>-stati<strong>on</strong>ary Time series Analysis,<br />
Academic Press, L<strong>on</strong>d<strong>on</strong><br />
[10] Newbold P (1988), Some recent developments in Time Series analysis,<br />
I,II,<str<strong>on</strong>g>and</str<strong>on</strong>g> III. <str<strong>on</strong>g>Inter</str<strong>on</strong>g>. Statist. Rev., 56, 17-29<br />
[11 ] M.Ayub Khan Yousuf Zai & Jawaid Quamar (1998), Sing. J. Phys., Vol. 14,<br />
# 1, pp 69-79<br />
(12) M.Ayub Khan Yousuf Zai & Jawaid Quamar (2001), “ Study the<br />
phenomen<strong>on</strong> of <str<strong>on</strong>g>oz<strong>on</strong>e</str<strong>on</strong>g> <str<strong>on</strong>g>layer</str<strong>on</strong>g> <str<strong>on</strong>g>depleti<strong>on</strong></str<strong>on</strong>g> as a physical process”, Indian. J.<br />
Phys., 75B (4), 307-314.
[13] M.Ayub Khan Yousuf Zai, (2004),”A Quantitative Study of effects of Oz<strong>on</strong>e<br />
Layer Depleti<strong>on</strong> <strong>on</strong> Marine Organisms with reference to coastal regi<strong>on</strong>s of<br />
Pakistan,” A Ph.D. thesis.<br />
[14] Anders<strong>on</strong> T W (1971), The Statistical Analysis of Time Series, Wiley, New<br />
York<br />
[15] Haggan V <str<strong>on</strong>g>and</str<strong>on</strong>g> Ozaki T (1979), Amplitude-dependent AR model fitting for<br />
n<strong>on</strong>-linear r<str<strong>on</strong>g>and</str<strong>on</strong>g>om vibrati<strong>on</strong>s, Paper presentati<strong>on</strong> at the <str<strong>on</strong>g>Inter</str<strong>on</strong>g>nati<strong>on</strong>al Time<br />
Series meeting, University of Nottingham, UK, March 1979.<br />
[16] Clevel<str<strong>on</strong>g>and</str<strong>on</strong>g> W S <str<strong>on</strong>g>and</str<strong>on</strong>g> McGill (1987), Graphical percepti<strong>on</strong>: the visual decoding<br />
of quantitative informati<strong>on</strong> <strong>on</strong> graphical displays of data, J. Roy. Stat. Soc.,<br />
A 150, 300<br />
[17] Basu S K, Sinha B K, & Mukherjee S P (2000), Fr<strong>on</strong>tiers in Probability <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
Statistics, Narosa Publishing House, New Delhi<br />
[18] Subba Rao T (1998), “ Applicati<strong>on</strong>s of Time Series Ananlysis in Astr<strong>on</strong>omy<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> Meteorology”, Chapman & Hall, L<strong>on</strong>d<strong>on</strong><br />
[19] Rowl<str<strong>on</strong>g>and</str<strong>on</strong>g> F S (1991), Ann. Rev. Phys. Chem. 42, 731<br />
[20] Bojkov R D <str<strong>on</strong>g>and</str<strong>on</strong>g> Fioletov (1995), Geophys., Res. Lett., 22, 2729<br />
[21] Henriksen K <str<strong>on</strong>g>and</str<strong>on</strong>g> Roldugin V (1995), Geophys., Res., Lett.. 22, 3219<br />
[22] Henriksen K And Larsen H H S (1992), J.Atmos. Terr. Phys., 55, 1113<br />
[23] Herman J R <str<strong>on</strong>g>and</str<strong>on</strong>g> Larko D (1993), J. Geophys. Res., 98, 12783<br />
[24] Hofmann D J, Oltmans J S, <str<strong>on</strong>g>and</str<strong>on</strong>g> Harris J M (1994), Geophys. Res. Lett., 21,<br />
1779<br />
[25] Hofmann D J, Oltmans J S, <str<strong>on</strong>g>and</str<strong>on</strong>g> Koenig G L (1996), Geophys. Res.Lett., 23,<br />
1533<br />
[26] McPeter R D <str<strong>on</strong>g>and</str<strong>on</strong>g> Hll<str<strong>on</strong>g>and</str<strong>on</strong>g>sworth (1996), Geophys. Res. Lett., 23, 3699<br />
[27] Rol<str<strong>on</strong>g>and</str<strong>on</strong>g>o R G (1994), J. Geophys. Res. Vol. 99, # D6. 12,937-12,951<br />
[28] Portmann R W, Solom<strong>on</strong> S, <str<strong>on</strong>g>and</str<strong>on</strong>g> Rol<str<strong>on</strong>g>and</str<strong>on</strong>g>o R G (1996), J. Geophys. Res.<br />
Vol. 101, # D17, 22,991-23,000<br />
[29] Rol<str<strong>on</strong>g>and</str<strong>on</strong>g>o R G, Frode S, <str<strong>on</strong>g>and</str<strong>on</strong>g> Jeffery T (1992), J. Geophys. Res. Vol. 97, #<br />
D12, 12967-12,991<br />
[30] Rol<str<strong>on</strong>g>and</str<strong>on</strong>g>o R G (1994), Physics World, April, 1994, 49-55<br />
[31] Tatsuo Shimazaki (1989), Geophysics <str<strong>on</strong>g>and</str<strong>on</strong>g> Astrophysics M<strong>on</strong>ograph,<br />
ed., Terra Scientific. Tokyo
[32] Attilio B <str<strong>on</strong>g>and</str<strong>on</strong>g> Shar<strong>on</strong> B (1997), Encychlopedia of Energy <str<strong>on</strong>g>and</str<strong>on</strong>g> Envir<strong>on</strong>mental,<br />
p 198-1225, Wiley, New York<br />
[33] Solom<strong>on</strong> S, <str<strong>on</strong>g>and</str<strong>on</strong>g> Daniel L A (1992), Nature, 357, 33-37<br />
[34] Prather M J, <str<strong>on</strong>g>and</str<strong>on</strong>g> Wats<strong>on</strong> R T (1990), Nature, 344, 729-734<br />
[35] Kallenrode B M (2004), “ Space Physics”, Springer Verlag, Berlin, New York<br />
[36] Zeilik M (1994), Astr<strong>on</strong>omy, the evolving universe, John Wiley, New York<br />
[37] Richard S L (1990), Dynamics in Atmospheric Physics, Cambridge<br />
University Press, New York<br />
[38] Kaler J B (1994), Astr<strong>on</strong>omy, HarperCollins College Publishers, New York<br />
[39] Zirin H (1989), Astrophysics of the Sun, Cambridge University Press,<br />
New York<br />
[40] Unno W & Yuasa M (1992), Astrophys.<str<strong>on</strong>g>and</str<strong>on</strong>g> Space Sci., 189, 271-278<br />
[41] Seinfeld H J & P<str<strong>on</strong>g>and</str<strong>on</strong>g>is N S (1998), Atmospheric Chemistry <str<strong>on</strong>g>and</str<strong>on</strong>g> physics,<br />
from air polluti<strong>on</strong> to climate, JohnWiley, New York<br />
[42] R<str<strong>on</strong>g>and</str<strong>on</strong>g>el W J & Gille J C (1993), Nature, 365, 533-535
SPACE RADIATION EFFECTS ON SPACECRAFT<br />
MICROELECTRONICS<br />
Dr. Shazia Maqbool <str<strong>on</strong>g>and</str<strong>on</strong>g> Ayesha Javed<br />
Satellite Research <str<strong>on</strong>g>and</str<strong>on</strong>g> Development Center (SRDC) Lahore<br />
SUPARCO, Pakistan<br />
ABSTRACT<br />
Srdclhr@ nexlinx.net.pk<br />
One of the primary design c<strong>on</strong>siderati<strong>on</strong>s for a space missi<strong>on</strong> is the<br />
survivability of its electr<strong>on</strong>ic systems in an i<strong>on</strong>izing radiati<strong>on</strong> envir<strong>on</strong>ment. The<br />
increasingly complex missi<strong>on</strong> requirements of modern spacecraft for military,<br />
commercial <str<strong>on</strong>g>and</str<strong>on</strong>g> scientific applicati<strong>on</strong>s have m<str<strong>on</strong>g>and</str<strong>on</strong>g>ated the need for highly<br />
capable microelectr<strong>on</strong>ics. This trend has been accompanied by the need to<br />
provide affordable, lightweight, low volume <str<strong>on</strong>g>and</str<strong>on</strong>g> low power systems. Both these<br />
trends have resulted in the need to adopt commercial microelectr<strong>on</strong>ics for<br />
space missi<strong>on</strong>s as well as to examine emerging technologies. By their very<br />
nature, commercial technologies are c<strong>on</strong>stantly changing.<br />
The effects of radiati<strong>on</strong>s <strong>on</strong> commercial device technologies have widely been<br />
discussed in literature. This paper summarizes these discussi<strong>on</strong>s to outline<br />
trends in different radiati<strong>on</strong> effects with changing device technologies. A review<br />
<strong>on</strong> radiati<strong>on</strong> resp<strong>on</strong>se of very large scale integrati<strong>on</strong> (VLSI) devices will also be<br />
presented.<br />
I. INTRODUCTION<br />
In the past, rad-hard technology derived from military programmes has<br />
dominated the space industry <str<strong>on</strong>g>and</str<strong>on</strong>g> military applicati<strong>on</strong>s have had significant<br />
influence <strong>on</strong> the directi<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> capability of the electr<strong>on</strong>ic industry. However,<br />
c<strong>on</strong>tinued reducti<strong>on</strong>s in military budgets, coupled with dramatic growth of the<br />
c<strong>on</strong>sumer electr<strong>on</strong>ics industry, have reduced this influence. Market drivers for<br />
c<strong>on</strong>sumer electr<strong>on</strong>ics have taken their products <str<strong>on</strong>g>and</str<strong>on</strong>g> supporting technologies<br />
far bey<strong>on</strong>d the capability of specialized military technology. C<strong>on</strong>sequently, radhard<br />
technology lags behind the commercial development by about two<br />
generati<strong>on</strong>s. This situati<strong>on</strong> has led space industry to commercial<br />
microelectr<strong>on</strong>ics. However, the use of commercial off-the-shelf (COTS)<br />
technology brings new issues <strong>on</strong> the reliability in the space envir<strong>on</strong>ment.<br />
One of the most significant trends in commercial microelectr<strong>on</strong>ics is the move<br />
towards smaller dimensi<strong>on</strong>s – i.e. “scaling”. Scaling has some benefits for<br />
radiati<strong>on</strong> effects resilience: As gate oxide thick-nesses decrease, hole trapping<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> interface trap build-up are becoming less significant, leading to a trend<br />
toward improved total-i<strong>on</strong>ising dose (TID) performance [13, 4]. Similarly, the<br />
increasing use of buried epitaxial substrates over the last 20 years [14, 30] has<br />
helped decrease single-event latch-up (SEL) sensitivity – both by limiting the<br />
charge-collecti<strong>on</strong> volume, <str<strong>on</strong>g>and</str<strong>on</strong>g> by decreasing the substrate series resistance
[35, 29, 3]. In additi<strong>on</strong>, the c<strong>on</strong>comitant trend towards reduced supply voltage<br />
levels suggests that device threshold voltages should so<strong>on</strong> fall below that<br />
required sustaining latch-up. Indeed, if c<strong>on</strong>tinued scaling forces the industry to<br />
adopt silic<strong>on</strong>-<strong>on</strong>-insulator (SOI) technology, latch-up should so<strong>on</strong> be eliminated<br />
[13, 30].<br />
However, whilst some radiati<strong>on</strong> effects are becoming less of an issue, new<br />
threats have emerged: Another significant trend in COTS technology is the<br />
move towards the use of “smart” logic, in device architectures, e.g. advanced<br />
memories, complex micro-c<strong>on</strong>trollers <str<strong>on</strong>g>and</str<strong>on</strong>g> field-programmable gate arrays<br />
(FPGAs), etc. Here the performance of the logic device requires the inclusi<strong>on</strong><br />
of complex c<strong>on</strong>trol circuitry internal to the die. Whilst transparent to the user,<br />
such circuitry may offer a significant target to i<strong>on</strong>izing particles <str<strong>on</strong>g>and</str<strong>on</strong>g> thus be<br />
pr<strong>on</strong>e to single-event effects (SEEs), such as a single-event upset (SEU) or<br />
single-event transient (SET). In such circuitry, these can manifest themselves<br />
as single event functi<strong>on</strong>al interrupts (SEFIs), whereby the device exhibits an<br />
unexpected change in its observable output state [18].<br />
II. SINGLE EVENT EFFECTS (SEE)<br />
SEEs are the result of free-charge generati<strong>on</strong> in semic<strong>on</strong>ductor materials<br />
caused by a single i<strong>on</strong>izing particle strike. Different phenomena may arise<br />
depending <strong>on</strong> the locati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> the instant of the particle strike.<br />
II.I. Single Event Upset (SEU)<br />
Single-Event Upset (SEU) is a n<strong>on</strong>-destructive single-event-effect. These are<br />
soft errors associated with changes in the state of memory elements – i.e.<br />
unexpected bit-flips due to particle strikes. Re-writing the data will restore the<br />
memory states.<br />
As technologies moved towards smaller feature sizes <str<strong>on</strong>g>and</str<strong>on</strong>g> thus had less charge<br />
stored <strong>on</strong> their circuit nodes, it was expected that the critical charge (<str<strong>on</strong>g>and</str<strong>on</strong>g> thus<br />
the linear energy transfer (LET) threshold) for SEU would decrease [1]. Indeed<br />
this trend was observed in the 1980’s in the now older generati<strong>on</strong> of devices.<br />
However, results <strong>on</strong> more recent technologies, with minimum feature sizes<br />
ranging from 3μm to 0.35μm do not follow this trend; in fact, the threshold LET<br />
has remains approximately c<strong>on</strong>stant. This levelling off at a minimum LET has<br />
been postulated to be the result of manufacturers taking into account the need<br />
to maintain a low upset rate from naturally-occurring radio<str<strong>on</strong>g>activity</str<strong>on</strong>g> (e.g. alpha<br />
particles) or atmospheric neutr<strong>on</strong>s during the design <str<strong>on</strong>g>and</str<strong>on</strong>g> manufacture of their<br />
parts [13, 4]. Recent SEE testing of complementary metal oxide semic<strong>on</strong>ductor<br />
(CMOS) technologies (0.15 μm, 0.13 μm, <str<strong>on</strong>g>and</str<strong>on</strong>g> 90 nm) c<strong>on</strong>forms with these<br />
predicti<strong>on</strong>s <str<strong>on</strong>g>and</str<strong>on</strong>g> it is found that the decreased size of the shrinking latches<br />
combined with robust design efforts has stabilized the critical neutr<strong>on</strong> cross<br />
secti<strong>on</strong>s of state of the art static latches. Indeed, the mean time between failure<br />
(MTBF) calculated for a milli<strong>on</strong> logic gate FPGA fabricated with a 90-nm<br />
technology is better than that of the MTBF of a similar complexity array<br />
fabricated in 150- or 130-nm processes.
Figure 1: Relati<strong>on</strong>ship between Feature Size <str<strong>on</strong>g>and</str<strong>on</strong>g> Critical Charge, [31]<br />
II.II. Single Event Transients (SET)<br />
This is are also a class of n<strong>on</strong>-destructive soft-error that can cause changes of<br />
logical state in combinati<strong>on</strong>al logic, or may be propagated in sequential logic<br />
through “glitches” <strong>on</strong> clock or set/ reset lines, etc.<br />
So far, it has been observed that radiati<strong>on</strong> resp<strong>on</strong>se of logic devices is<br />
dominated by errors in registers <str<strong>on</strong>g>and</str<strong>on</strong>g> memory cells – i.e. SEUs [14]. However,<br />
as devices are further scaled down to smaller feature sizes <str<strong>on</strong>g>and</str<strong>on</strong>g> faster speeds,<br />
soft errors in combinati<strong>on</strong>al logic, SETs, are expected to become more<br />
probable. In c<strong>on</strong>trast to SEUs, which do not show frequency dependence,<br />
SETs depend significantly <strong>on</strong> the operating speed of the devices in questi<strong>on</strong> [4,<br />
14]. According to models developed by Shivakumar et al., soft errors in<br />
combinati<strong>on</strong>al logic are likely to become comparable to SEUs in unprotected<br />
memory elements by the year 2011 [38].<br />
II.III. Single Event Latch-Up (SEL)<br />
Single Event Latch-Up (SEL) is a potentially destructive single event effect. It<br />
can occur due to the presence of parasitic bipolar transistors inherent in CMOS<br />
structures, <str<strong>on</strong>g>and</str<strong>on</strong>g> may lead to an effective short circuit through the device. There<br />
are two types of SEL: In traditi<strong>on</strong>al or destructive SEL, the device current<br />
c<strong>on</strong>sumpti<strong>on</strong> increases immediately above the maximum specified for the<br />
device, causing permanent damage to the device through the burn-out of b<strong>on</strong>dwires,<br />
etc. The other type is known as a micro-latch-up, <str<strong>on</strong>g>and</str<strong>on</strong>g> during such an<br />
event, the device current increases above the normal operating voltage but not<br />
above the maximum specified. Micro-latch-up halts device operati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> a<br />
power reset is required to recover the device. The device may still operate but<br />
show signs of permanent damage, such as exhibiting a step-change in its<br />
current c<strong>on</strong>sumpti<strong>on</strong>. Alternatively the damage may not be immediately<br />
apparent. Sometimes, a series of micro-latch-up events in quick successi<strong>on</strong><br />
can lead to destructive SEL.<br />
II.IV. Single Event Functi<strong>on</strong>al <str<strong>on</strong>g>Inter</str<strong>on</strong>g>rupt (SEFI)<br />
One of the most significant trends in COTS technology is the move towards<br />
device architectures with in-built “smart logic” – that is for devices which have<br />
internal operati<strong>on</strong>s transparent to the user. Smart logic requires the inclusi<strong>on</strong> of<br />
more c<strong>on</strong>trol circuitry internal to the die. For example, flash memory devices
c<strong>on</strong>tain a state machine that generates the signals needed to perform<br />
read/write operati<strong>on</strong>s. Many dynamic r<str<strong>on</strong>g>and</str<strong>on</strong>g>om-access memory (DRAM)<br />
devices, microprocessors <str<strong>on</strong>g>and</str<strong>on</strong>g> field-programmable gate-arrays (FPGAs) have a<br />
built-in self-test capability, which again increases device complexity. Many<br />
DRAMs include redundant memory rows <str<strong>on</strong>g>and</str<strong>on</strong>g> columns intended to increase<br />
device reliability. When the DRAM is powered up, weak or bad rows or<br />
columns are replaced with their redundant counterparts <str<strong>on</strong>g>and</str<strong>on</strong>g> a redundancylatch<br />
has to be used to hold the device c<strong>on</strong>figurati<strong>on</strong>. Many synchr<strong>on</strong>ous-<br />
DRAM (SDRAM) devices also have internal state machines that implement<br />
pipelines, programmable refresh modes <str<strong>on</strong>g>and</str<strong>on</strong>g> various power states. In most of<br />
these cases, this c<strong>on</strong>trol logic is not directly accessible to the user.<br />
Experience from space flight <str<strong>on</strong>g>and</str<strong>on</strong>g> ground-based i<strong>on</strong>-beam testing of many<br />
COTS devices reveals that this internal c<strong>on</strong>trol circuitry is susceptible to<br />
radiati<strong>on</strong> effects, often manifesting itself as the class of single event effects<br />
(SEEs) called single event functi<strong>on</strong>al interrupts (SEFIs), whereby an internal<br />
upset in this logic causes an unexpected change in the observable state of the<br />
device. SEFI was first menti<strong>on</strong>ed in 1996 [5, 18], although SEFI-like signatures<br />
had been reported earlier [17, 42, 19]. Early reports were c<strong>on</strong>fined to<br />
microprocessor SEFIs, however, emerging data h<str<strong>on</strong>g>and</str<strong>on</strong>g>ling devices, such as<br />
advanced memories <str<strong>on</strong>g>and</str<strong>on</strong>g> FPGAs, have also been found to be susceptible.<br />
III. MULTIPLE BIT UPSET (MBU)<br />
Sometimes, more than <strong>on</strong>e memory cells can be corrupted by the charge<br />
depositi<strong>on</strong> originating from the same particle. Such events are characterized as<br />
single-event multiple-bit upsets. These errors are mainly associated with<br />
memory devices, although any register is a potential target. MBUs, which have<br />
been observed during ground-based i<strong>on</strong>-beam testing <str<strong>on</strong>g>and</str<strong>on</strong>g> in during space<br />
flight, may be attributed to three mechanisms:<br />
• A single particle strike affects more than <strong>on</strong>e bit in <strong>on</strong>e or more words;<br />
• Coincidental, independent SEUs (from different particle strikes) appear<br />
as an MBU;<br />
• A single SEFI event causes MBUs (termed as block SEFIs).<br />
The sec<strong>on</strong>d mechanism is not in fact a true MBU, but is instead an artifact of<br />
the diagnostic algorithm <str<strong>on</strong>g>and</str<strong>on</strong>g> the underlying (usually high) SEU rate. Both<br />
SRAMs <str<strong>on</strong>g>and</str<strong>on</strong>g> DRAMs are pr<strong>on</strong>e to MBUs, with DRAMs, generally having a<br />
higher probability of getting such errors. MBU rates are expected to increase<br />
with the device scaling. However, in some cases, less dense devices have<br />
been found to show higher MBU rates than more dense devices [39]. On the<br />
other h<str<strong>on</strong>g>and</str<strong>on</strong>g>, reports of block SEFI events are increasing. Floating gate<br />
memories, which do not exhibit traditi<strong>on</strong>al MBUs, are susceptible to this class<br />
of MBUs.<br />
IV. TOTAL IONISING DOSE<br />
Total i<strong>on</strong>izing dose in semic<strong>on</strong>ductor devices depend <strong>on</strong> the creati<strong>on</strong> of<br />
electr<strong>on</strong>-hole pairs within dielectric <str<strong>on</strong>g>layer</str<strong>on</strong>g>s <str<strong>on</strong>g>and</str<strong>on</strong>g> subsequent generati<strong>on</strong> of traps<br />
at or near the interface with the semic<strong>on</strong>ductor. This can produce a variety of<br />
device effects such as flat-b<str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> threshold voltage shifts, surface leakage<br />
currents, <str<strong>on</strong>g>and</str<strong>on</strong>g> speed degradati<strong>on</strong> of the devices. Many current COTS parts are<br />
very susceptible to total dose damage, <str<strong>on</strong>g>and</str<strong>on</strong>g> may fail at total-dose levels of 5<br />
krad (Si) or even less. Total dose is therefore an issue in space flight where
l<strong>on</strong>g missi<strong>on</strong> lifetimes, <str<strong>on</strong>g>and</str<strong>on</strong>g>/or exposure to the Van Allen radiati<strong>on</strong> belts means<br />
that these dose levels can be easily exceeded.<br />
As commercial devices are scaled to smaller feature sizes, so gate oxide<br />
thicknesses are also decreasing. The threshold voltage shift that results from<br />
hole traps in gate oxides decreases as the square of oxide thickness,<br />
assuming that the hole trapping efficiency is unchanged as the oxide thickness<br />
is reduced [23]. It has been shown that hole trapping will reduce further<br />
because of tunneling as oxides are scaled below 10nm [34]. Thus, for the<br />
highly scaled devices, hole trapping <str<strong>on</strong>g>and</str<strong>on</strong>g> interface traps are not likely to remain<br />
major c<strong>on</strong>cerns for devices operating in a radiati<strong>on</strong> envir<strong>on</strong>ment. However, it<br />
should be noted that even in highly scaled devices there would be isolati<strong>on</strong><br />
dielectrics, which are still relatively thick. Thus this is isolati<strong>on</strong>/field oxide that<br />
dominates total dose resp<strong>on</strong>se of commercial devices [39, 13, 4,].<br />
V. TRENDS IN DEVICES<br />
This secti<strong>on</strong> presents a review of radiati<strong>on</strong> issues in SOTA <str<strong>on</strong>g>and</str<strong>on</strong>g> emerging<br />
microelectr<strong>on</strong>ics.<br />
V.I. MEMORIES<br />
Space radiati<strong>on</strong> effects <strong>on</strong> different types of memory devices are discussed<br />
below.<br />
V.I.I. DRAMs<br />
DRAM technology has grown rapidly in recent years. The first DRAM,<br />
introduced in the early 1970s, c<strong>on</strong>tained <strong>on</strong>ly 1,024 bits, but modern DRAMs<br />
are available c<strong>on</strong>taining 256 megabits or more. Large DRAMs incorporate a<br />
c<strong>on</strong>troller to provide parallelism in operati<strong>on</strong>s <str<strong>on</strong>g>and</str<strong>on</strong>g> hence an increased memory<br />
throughput. Their high density <str<strong>on</strong>g>and</str<strong>on</strong>g> superior performance characteristics make<br />
them an attractive c<str<strong>on</strong>g>and</str<strong>on</strong>g>idate for <strong>on</strong>-board data recording to fulfill increasingly<br />
higher missi<strong>on</strong> data storage requirements. In additi<strong>on</strong>, recent technology has<br />
permitted the stacking of dies into single units providing even denser<br />
packaging of DRAMs into a single device. However, the use of DRAMs for<br />
space applicati<strong>on</strong>s does not come without certain penalties. DRAMs have<br />
historically been c<strong>on</strong>sidered as devices very sensitive to SEUs, since the circuit<br />
errors were observed <strong>on</strong> these comp<strong>on</strong>ents back in the seventies. Their high<br />
sensitivity to i<strong>on</strong>ising particle strikes comes from their characteristic of<br />
passively storing bit informati<strong>on</strong> as charge stored in a circuit node. Although<br />
the amount of charge stored decreases steadily from <strong>on</strong>e technology<br />
generati<strong>on</strong> to the next, unexpectedly, the error rate has been found to<br />
decrease or stay c<strong>on</strong>stant [24, 7, 14]. However, the observati<strong>on</strong> of n<strong>on</strong>traditi<strong>on</strong>al<br />
SEEs, e.g. SEFIs, raises new c<strong>on</strong>cerns over the reliability of these<br />
devices in radiati<strong>on</strong> envir<strong>on</strong>ment. Both the Cassini <str<strong>on</strong>g>and</str<strong>on</strong>g> Hubble Space<br />
Telescope (HST) missi<strong>on</strong>s have used solid state data recorders (SSDRs)<br />
made from DRAM [21]. Observed SEE results include the st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard single bit<br />
upset (SEU) errors, stuck bits, MBUs, <str<strong>on</strong>g>and</str<strong>on</strong>g> column or row errors (where a single<br />
i<strong>on</strong> strike induces a partial or full address column or row to be in error in a<br />
single device) [21, 39]. Label termed column <str<strong>on</strong>g>and</str<strong>on</strong>g> row errors as block SEFIs<br />
[20]. Ground-based i<strong>on</strong>-beam testing of DRAMs has also shown st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard
SEFIs (e.g with the device entering test or st<str<strong>on</strong>g>and</str<strong>on</strong>g>by modes) in additi<strong>on</strong> to<br />
above-menti<strong>on</strong>ed errors [20, 18].<br />
It should be noted that advanced DRAM architectures are also emerging. An<br />
example of an advanced DRAM is Synchr<strong>on</strong>ous DRAM (SDRAM) that can<br />
provide even more storage capacity <strong>on</strong> a chip al<strong>on</strong>g with an increase in access<br />
speed. These devices have underlying characteristics of stuck bits <str<strong>on</strong>g>and</str<strong>on</strong>g> MBUs.<br />
Increased level of complexity in these devices leads to higher susceptibility of<br />
SEFIs [33].<br />
V.I.II. SRAMs<br />
These devices are susceptible to SEUs, generally with quite low thresholds<br />
(often below 1 MeVcm2/mg) [2]. They are also subject to MBUs. However, their<br />
MBU rates are significantly smaller than their SEU rates. There is no evidence<br />
from in-orbit observati<strong>on</strong>s that the denser devices are necessarily more pr<strong>on</strong>e<br />
to MBUs than the older devices [39, 41] although ground based tests suggest<br />
this. Stuck bits are sometimes observed, but <strong>on</strong>ly during irradiati<strong>on</strong> with<br />
particles of high LET (i.e. heavy-i<strong>on</strong>s) [7]. Newer generati<strong>on</strong>s of SRAMs, using<br />
a 6T cell design, are expected to have an improved SEU <str<strong>on</strong>g>and</str<strong>on</strong>g> total dose<br />
resp<strong>on</strong>se [22, 32, 7].<br />
V.I.III. Flash Memories<br />
Flash memory technology comes with advantages in terms of density <str<strong>on</strong>g>and</str<strong>on</strong>g> n<strong>on</strong>volatility.<br />
It provides fast read access, comparable to that of DRAMs. However,<br />
erasing <str<strong>on</strong>g>and</str<strong>on</strong>g> writing new data in flash memories is a more complex operati<strong>on</strong>,<br />
requiring applicati<strong>on</strong> of high voltage. Newer flash devices use complex internal<br />
architectures to improve write <str<strong>on</strong>g>and</str<strong>on</strong>g> erase times, as well as to make device<br />
operati<strong>on</strong> more transparent to user.<br />
There are two types of flash architecture: NOR <str<strong>on</strong>g>and</str<strong>on</strong>g> NAND. Whilst both<br />
architectures have the same basic storage element, it is the interc<strong>on</strong>necti<strong>on</strong> of<br />
these memory cells that distinguishes structure. In the NOR flash memory<br />
array, the bit line logic goes to “0” if any of the memory cell transistors is <strong>on</strong>. In<br />
the NAND flash memory array, all memory cell transistors are required to be <strong>on</strong><br />
for bit line to go to “0” state. The NAND architecture is more sensitive to<br />
radiati<strong>on</strong> [25].<br />
Radiati<strong>on</strong> testing has been performed <strong>on</strong> Intel, Samsung, Toshiba, AeroFlex,<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> SanDisk Flash memory devices [36, 28, 27]. The observed SEEs are<br />
dominated by errors in their complex internal architectures rather than in the<br />
n<strong>on</strong>-volatile storage elements. Test results have shown that flash devices are<br />
immune to SEUs when powered off. The probability of getting an upset is<br />
particularly high during write mode. Static <str<strong>on</strong>g>and</str<strong>on</strong>g> dynamic read modes have also<br />
dem<strong>on</strong>strated complex c<strong>on</strong>trol-related upsets, where the functi<strong>on</strong>al<br />
c<strong>on</strong>sequences of these errors can be multiple. Sometimes, a steep increase in<br />
the current c<strong>on</strong>sumpti<strong>on</strong> is observed during or after irradiati<strong>on</strong>.<br />
V.I.IV. EEPROMs<br />
Electrically Erasable Programmable Read-Only Memory (EEPROM) provides<br />
n<strong>on</strong>-volatile storage like Flash memories. The principal difference between the<br />
two technologies is that an EEPROM requires data to be written or erased <strong>on</strong>e
yte at a time whereas flash memory allows data to be written or erased in<br />
blocks. This makes flash memory faster. Radiati<strong>on</strong> test results <strong>on</strong> EEPROMs<br />
are presented in [18, 20]. Similar to flash memories, the devices are found to<br />
be more susceptible during write mode. SEFI has been observed in these<br />
memories as well.<br />
V.I.V. Mitigati<strong>on</strong>s for Memory Devices<br />
There are three approaches to improve survivability of microelectr<strong>on</strong>ics in<br />
space envir<strong>on</strong>ment.<br />
• Firstly, to invent better semic<strong>on</strong>ductor process to make the memory<br />
cell less susceptible to direct radiati<strong>on</strong>. Researchers found that performance<br />
under radiati<strong>on</strong> can be improved by repositi<strong>on</strong>ing various comp<strong>on</strong>ent juncti<strong>on</strong>s<br />
or by using silic<strong>on</strong>-<strong>on</strong>-sapphire (SOS) or SOI technologies. All these methods<br />
require redesign of the memory chips <str<strong>on</strong>g>and</str<strong>on</strong>g> call for new process technologies<br />
that are not widely used for memory chips <str<strong>on</strong>g>and</str<strong>on</strong>g> therefore, present a risk in<br />
producti<strong>on</strong> cost <str<strong>on</strong>g>and</str<strong>on</strong>g> in scalability when memory technology changes.<br />
• The other way is to use COTS memory parts. It has been found that<br />
every memory chip exhibits slightly different characteristic under radiati<strong>on</strong><br />
envir<strong>on</strong>ment. Even the same batch of chips can react quite differently am<strong>on</strong>g<br />
each other. Some chip works well in the energy envir<strong>on</strong>ment while others fail<br />
miserably. Therefore, the most suitable way is to test-<str<strong>on</strong>g>and</str<strong>on</strong>g>-select. NASA<br />
decided to use COTS parts even in the <str<strong>on</strong>g>Inter</str<strong>on</strong>g>nati<strong>on</strong>al Space Stati<strong>on</strong>. This is<br />
obviously for cost reas<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> to allow upgrade path as memory technology<br />
advances during their course [43].<br />
• Implementati<strong>on</strong> of EDAC (Error Detecti<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> Correcti<strong>on</strong> Codes)<br />
techniques in both hardware <str<strong>on</strong>g>and</str<strong>on</strong>g> software can prove helpful in fighting against<br />
the hazardous effects of radiati<strong>on</strong>s.<br />
V.II. MICROPROCESSORS<br />
The most complex type of digital logic device used in <strong>on</strong>-board is the<br />
microprocessor. Using commercial processors for space missi<strong>on</strong>s brings<br />
potential advantages in terms of cost <str<strong>on</strong>g>and</str<strong>on</strong>g> design time reducti<strong>on</strong>. These devices<br />
are accompanied by COTS software applicati<strong>on</strong>s, development tools, <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
operating systems. Processors such as the 80Cx86 family <str<strong>on</strong>g>and</str<strong>on</strong>g> the Power PC<br />
are being flown in space despite being susceptible to i<strong>on</strong>izing radiati<strong>on</strong>. These<br />
processors are not radiati<strong>on</strong> hardened <str<strong>on</strong>g>and</str<strong>on</strong>g> have been found to be susceptible<br />
to SEUs.<br />
Microprocessors work by c<strong>on</strong>tinuously moving data between registers; the<br />
memory unit with arithmetic <str<strong>on</strong>g>and</str<strong>on</strong>g> logic unit operati<strong>on</strong>s <str<strong>on</strong>g>and</str<strong>on</strong>g> flag registers updates<br />
etc. It is unlikely that all secti<strong>on</strong>s will be in use during the processing of a<br />
program. The applicati<strong>on</strong> software, which is being executed, determines how<br />
many secti<strong>on</strong>s are being used at any given time. Moreover, each secti<strong>on</strong> might<br />
have a different sensitivity to SEU. Therefore, the SEU-induced effects in<br />
microprocessors are str<strong>on</strong>gly applicati<strong>on</strong> dependent. Many commercial<br />
processors have been tested to evaluate their suitability for space applicati<strong>on</strong>s.<br />
These include the 80x86 architectures, SPARC, Motorla 68020, Pentium,<br />
Alpha, <str<strong>on</strong>g>and</str<strong>on</strong>g> PowerPC’s [6, 26, 20, 11, 37, 12]. All devices experienced some<br />
type of SEFI or device lockups at fairly low threshold LETs <str<strong>on</strong>g>and</str<strong>on</strong>g> authors have<br />
suggested that reliable operati<strong>on</strong> in the radiati<strong>on</strong> envir<strong>on</strong>ment will require at<br />
least a detecti<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> reset capability.
One result, which was certainly unexpected, is that the threshold LET has<br />
essentially not changed during last many years [13]. Table1 compares<br />
threshold LET of different microprocessors.<br />
Mitigati<strong>on</strong>:<br />
Implementati<strong>on</strong> of either offline or <strong>on</strong>line fault detecti<strong>on</strong> techniques like<br />
watchdog timers, Locksteps <str<strong>on</strong>g>and</str<strong>on</strong>g> built in testing in microprocessors can<br />
help against the radiati<strong>on</strong> caused damages.<br />
Device Manufacturer Feature Size<br />
(approx)<br />
Threshold LET<br />
(MeV-cm 2 /mg)<br />
Z-80 Zilog 3μm 1.5-2.5<br />
8086 Intel 1.5μm 1.5-2.5<br />
80386 Intel 0.8μm 2-3<br />
68020 Motorola 0.8μm 1.5-2.5<br />
LS64811 LSI 1.2μm 2-2.5<br />
90C601 MHS 1.2μm 2-2.5<br />
80386 Intel 0.6μm 2-3<br />
PC603e Motorola 0.4μm 1.7-3<br />
Pentium Intel 0.35μm 2-3<br />
Power PC750 Motorola 0.25μm 2-2.5<br />
PowerPC7455 Motorola 0.18μm 1<br />
Table1: Upset Threshold of Microprocessors [13]<br />
V.III. FIELD PROGRAMMABLE GATE ARRAYS (FPGAs)<br />
This technology offers number of advantages including a highly compact<br />
soluti<strong>on</strong>, high integrity, flexibility, reduced cost, faster <str<strong>on</strong>g>and</str<strong>on</strong>g> cheaper prototyping,<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> reduced lead-time before flight. The capacity <str<strong>on</strong>g>and</str<strong>on</strong>g> performance of FPGAs<br />
suitable for space flight have been increasing steadily for more than a decade.<br />
The applicati<strong>on</strong> of FPGAs has moved from simple glue logic to complete<br />
platforms that combine several real-time system functi<strong>on</strong>s <strong>on</strong> a single chip [9].<br />
The choice of type of storage for the c<strong>on</strong>figurati<strong>on</strong> drives the radiati<strong>on</strong><br />
performance of the device [15]. FPGAs can be split into two categories: namely<br />
re-programmable <str<strong>on</strong>g>and</str<strong>on</strong>g> <strong>on</strong>e time programmable (OTP). Reprogrammable<br />
technology offers volatile SRAM or n<strong>on</strong>-volatile EEPROM/Flash cells to hold<br />
the device c<strong>on</strong>figurati<strong>on</strong>. SRAM-based technology has been evolving at a<br />
faster pace than OTP technology, <str<strong>on</strong>g>and</str<strong>on</strong>g> now features a milli<strong>on</strong> system gates or<br />
more <strong>on</strong> a single chip, <str<strong>on</strong>g>and</str<strong>on</strong>g> hence has become an attractive choice for high<br />
performance applicati<strong>on</strong>s. However, <strong>on</strong>e must be aware not <strong>on</strong>ly of radiati<strong>on</strong>
issues, but also of c<strong>on</strong>cerns such as the total loss of c<strong>on</strong>figurati<strong>on</strong> from single<br />
i<strong>on</strong> hit. OTP technology <strong>on</strong> the other h<str<strong>on</strong>g>and</str<strong>on</strong>g> offers radiati<strong>on</strong> performance<br />
comparable to applicati<strong>on</strong> specific integrated circuit (ASICs). The most<br />
comm<strong>on</strong>ly used OTP FPGAs use antifuses for storing its c<strong>on</strong>figurati<strong>on</strong>, either<br />
using oxide-nitride-oxide (ONO) or metal-to-metal (M2M) antifuse structures<br />
[16].<br />
FPGAs are available at different reliability levels, ranging from inexpensive<br />
COTS devices to high reliability, rad-hard devices, as well as different device<br />
speeds, capabilities, <str<strong>on</strong>g>and</str<strong>on</strong>g> c<strong>on</strong>figurati<strong>on</strong>s. Both Actel <str<strong>on</strong>g>and</str<strong>on</strong>g> Xilinx have been<br />
offering radiati<strong>on</strong> tolerant devices for space applicati<strong>on</strong>s. The Actel OTP<br />
devices have been the primary FPGA technology used in space flight systems<br />
to date. Actel’s ACT1, ACT2 <str<strong>on</strong>g>and</str<strong>on</strong>g> ACT3 have been very popular am<strong>on</strong>g space<br />
designers [15]. Commercial SX <str<strong>on</strong>g>and</str<strong>on</strong>g> SX-A series have also been reported as<br />
being capable of tolerating total doses as high as 100 krads(Si) or more. These<br />
devices are latch-up immune as well. The flip-flop element of the basic device<br />
is c<strong>on</strong>sidered as being radiati<strong>on</strong> “tolerant”; enhanced radiati<strong>on</strong> hardened levels<br />
can be achieved via incorporati<strong>on</strong> of software or hardware implemented<br />
mitigati<strong>on</strong> strategies [16]. Although ONO antifuses are susceptible to SEGR,<br />
the threshold is found to be quite high [15]. The probability of an SEGR event is<br />
low <str<strong>on</strong>g>and</str<strong>on</strong>g> <strong>on</strong>e rupture does not cause permanent damage to the device.<br />
Typically it would take at least 10 ruptures to cause a failure. Actel introduced<br />
products using a M2M antifuse. It is reported that this antifuse technology is<br />
immune to SEGR effect to a currently tested LET threshold of 80 MeV-cm 2 /mg.<br />
Actel Corporati<strong>on</strong> offers two classes of rad-hard FPGAs, RH1020 <str<strong>on</strong>g>and</str<strong>on</strong>g> RH1280<br />
with equivalent gate densities of 2,000 <str<strong>on</strong>g>and</str<strong>on</strong>g> 8,000 respectively. These products<br />
are latchup immune <str<strong>on</strong>g>and</str<strong>on</strong>g> can withst<str<strong>on</strong>g>and</str<strong>on</strong>g> total dose in excess of 300 krads (Si).<br />
Its SEU performance is predicted to be 10-6 upsets per bit-day in a 90% worstcase<br />
geosynchr<strong>on</strong>ous earth orbit. Actel proposes mitigati<strong>on</strong> techniques to<br />
improve SEU performance of these devices by incorporating TMR or by<br />
avoiding flip-flops in the sequential units.Adaptati<strong>on</strong> of <strong>on</strong>e of these techniques<br />
is recommended for critical design secti<strong>on</strong>s.<br />
Another rad-hard FPGA RHAX250-S device technology offers a gate count of<br />
250, 000 with even better radiati<strong>on</strong> performance. In additi<strong>on</strong> to these rad-hard<br />
products, Actel offers radiati<strong>on</strong> tolerant devices such as RTAX_S. This device<br />
technology has TID resp<strong>on</strong>se comparable to RH1020 <str<strong>on</strong>g>and</str<strong>on</strong>g> RH1280 devices,<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> latchup <str<strong>on</strong>g>and</str<strong>on</strong>g> SEU immunity to 104 <str<strong>on</strong>g>and</str<strong>on</strong>g> 60 MeV-cm2/mg respectively.<br />
Xilinx offers the XQR series for space applicati<strong>on</strong>s. This series is a subset of<br />
their commercial XC40000XL family. Although these devices have<br />
dem<strong>on</strong>strated acceptable total dose <str<strong>on</strong>g>and</str<strong>on</strong>g> latch-up performance for many space<br />
applicati<strong>on</strong>s, they are susceptible to single event effects including SEFIs [8, 9].<br />
Work is in progress at Xilinx for the development of the single event immune<br />
rec<strong>on</strong>figurable FPGA (SIRF).<br />
SEUs in reprogrammable FPGAs can be grouped into three categories [9]:<br />
C<strong>on</strong>figurati<strong>on</strong> upsets are defined as the upsets in the c<strong>on</strong>figurati<strong>on</strong> memory of<br />
the device <str<strong>on</strong>g>and</str<strong>on</strong>g> can be detected by read-back. The likelihood of failure will
depend up<strong>on</strong> the upset locati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> the specific design utilisati<strong>on</strong> of the device<br />
resources. A c<strong>on</strong>figurati<strong>on</strong> upset can sometimes lead to high current states.<br />
For example, a SEU may cause two output drivers to be c<strong>on</strong>nected together.<br />
User logic upsets represent upsets in the circuit elements that are not directly<br />
accessible by read-back. The effect of these is dependent up<strong>on</strong> the particular<br />
logical design implemented by the user.<br />
System upsets are those upsets that occur in the c<strong>on</strong>trol circuitry of the FPGA.<br />
As an example, it is possible for a single upset in the c<strong>on</strong>figurati<strong>on</strong> c<strong>on</strong>trol<br />
circuitry to change many c<strong>on</strong>figurati<strong>on</strong> bits simultaneously .This kind of<br />
anomaly can <strong>on</strong>ly be detected by noting the malfuncti<strong>on</strong> of the device.<br />
V.IV. GaAs DEVICES<br />
GaAs devices are used in high speed digital <str<strong>on</strong>g>and</str<strong>on</strong>g> microwave applicati<strong>on</strong>s. Since<br />
no dielectric <str<strong>on</strong>g>layer</str<strong>on</strong>g>s are used in GaAs devices, they exhibit extreme radiati<strong>on</strong><br />
tolerance - up to several hundred Mrad (GaAs) (two orders of magnitude higher<br />
than for equivalent Si-based technologies) [44].<br />
In general GaAs technology is very susceptible to SEU but hardening with low<br />
temperature buffer <str<strong>on</strong>g>layer</str<strong>on</strong>g>s offers orders of magnitude improvement.<br />
V.V. OPTO-ELECTRONIC DEVICES<br />
Broadly opto-electr<strong>on</strong>ic devices include light emitting diodes (LEDs), laser<br />
diodes, optical detectors, <str<strong>on</strong>g>and</str<strong>on</strong>g> optical couplers. These products are also not<br />
c<strong>on</strong>sidered radiati<strong>on</strong> hard without special processing. For example opto<br />
couplers can be sensitive to degradati<strong>on</strong> of current transfer ratio (CTR) at a<br />
dose of 2-3 krad (Si). Some types of opto-emitters are severely degraded by<br />
the prot<strong>on</strong> bombardment at a TID of 1-2 krads. Displacement damage dose<br />
(DDD) is most pr<strong>on</strong>ounced in optocouplers <str<strong>on</strong>g>and</str<strong>on</strong>g> lasers. High b<str<strong>on</strong>g>and</str<strong>on</strong>g>width<br />
optocouplers are extremely sensitive to SEUs. Permanent damage can be<br />
caused by displacement damage, for example loss in LED power (though<br />
devices with a hardened LED will ultimately suffer from photo-resp<strong>on</strong>se<br />
degradati<strong>on</strong>) [45] <str<strong>on</strong>g>and</str<strong>on</strong>g> from i<strong>on</strong>izing damage in the photo detector.<br />
Optocouplers can be of linear or digital type <str<strong>on</strong>g>and</str<strong>on</strong>g> many different designs exist,<br />
each with a different radiati<strong>on</strong> resp<strong>on</strong>se. Optocoupler failures occurred <strong>on</strong> the<br />
Topex-Poseid<strong>on</strong> missi<strong>on</strong> [45] at TID levels of 20-30 krad (Si). For systems that<br />
use lens coupling, radiati<strong>on</strong> darkening of the lens, due to TID, can sometimes<br />
be a problem (particularly for graded index, GRIN, lenses).<br />
VI. CONCLUSIONS<br />
Adopti<strong>on</strong> of COTS devices <str<strong>on</strong>g>and</str<strong>on</strong>g> st<str<strong>on</strong>g>and</str<strong>on</strong>g>ards has become a prevailing practice for<br />
space applicati<strong>on</strong>s. C<strong>on</strong>tinuously changing COTS technologies impose a threat<br />
to their reliable use in space radiati<strong>on</strong> envir<strong>on</strong>ment.<br />
A general trend towards an improved TID <str<strong>on</strong>g>and</str<strong>on</strong>g> SEL resp<strong>on</strong>se has been<br />
observed. SEFIs <str<strong>on</strong>g>and</str<strong>on</strong>g> SETs are recognized as two potentially dominating<br />
radiati<strong>on</strong> hazards from current knowledge of device technology <str<strong>on</strong>g>and</str<strong>on</strong>g> its<br />
radiati<strong>on</strong> performance. SETs are predicted to be of serious c<strong>on</strong>cern in near
future. SEFIs, <strong>on</strong> the other h<str<strong>on</strong>g>and</str<strong>on</strong>g>, have been reported with present technology<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> likely to get worse as device technology moves towards smarter <str<strong>on</strong>g>and</str<strong>on</strong>g> more<br />
transparent architectures.<br />
Over the past many years, COTS technology has been flown in space systems<br />
with a reas<strong>on</strong>able expectati<strong>on</strong> of success. This has primarily been achieved by<br />
a thorough underst<str<strong>on</strong>g>and</str<strong>on</strong>g>ing of radiati<strong>on</strong> resp<strong>on</strong>se of these devices <str<strong>on</strong>g>and</str<strong>on</strong>g> then,<br />
applicati<strong>on</strong> of appropriate mitigati<strong>on</strong> techniques <str<strong>on</strong>g>and</str<strong>on</strong>g> sound engineering<br />
practices.<br />
VII. ACKNOWLEDGMENT<br />
This work is attributed to the unmatched guidance of Dr. Riaz M Suddle <str<strong>on</strong>g>and</str<strong>on</strong>g> to<br />
SUPARCO for sp<strong>on</strong>soring it.<br />
VIII. REFERENCES<br />
[1] Akers, L.D., “Microprocessor Technology <str<strong>on</strong>g>and</str<strong>on</strong>g> Single Event<br />
Suscptibility”, Proc. 10th Annual AIAA/USU C<strong>on</strong>ference <strong>on</strong> Small<br />
Satellites, Logan, Utah, 1996<br />
[2] J.R.Coss et al., “Device SEE Susceptibility Update: 1996-1998”, IEEE<br />
NSREC Data Workshop Record, pp. 60-81, July, 1999<br />
[3] P.E. Dodd, M.R Shaneyfelt, E. Fuller, J.C Pickel, F.W. Sextan <str<strong>on</strong>g>and</str<strong>on</strong>g> P.S.<br />
Winokur “Impact of Substrate thickness <strong>on</strong> Single-Event Effects in<br />
Integrated Circuits”, IEEE Trans. Nucl. Sci., Volume-48, Issue-6, pp.<br />
1865-1871, December, 2001<br />
[4] Paul V. Dressendorfer “Basic Mechanisms for the New Millennium”,<br />
IEEE NSREC Short Course, July, 1998<br />
[5] EIA/JEDEC St<str<strong>on</strong>g>and</str<strong>on</strong>g>ard, “Test Procedures for the Measurement of Single<br />
Event Effects in Semic<strong>on</strong>ductor Devices from Heavy I<strong>on</strong> Irradiati<strong>on</strong>”,<br />
Electr<strong>on</strong>ic Industries Associati<strong>on</strong>, Engineering Department, Arlingt<strong>on</strong>,<br />
1996<br />
[6] Estreme, F. et al., “SEU <str<strong>on</strong>g>and</str<strong>on</strong>g> Latchup Results for SPARC Processors”,<br />
IEEE Radiati<strong>on</strong> Effects Data Workshop Record, pp. 13-19, July, 1993<br />
[7] Faccio, F., “COTS for the LHC Radiati<strong>on</strong> Envir<strong>on</strong>ment: The Rules of<br />
the game”, 6th Workshop <strong>on</strong> Electr<strong>on</strong>ics for LHC Experiments,<br />
Krakow, Pol<str<strong>on</strong>g>and</str<strong>on</strong>g>, 2000<br />
[8] Fuller, E. et al., “Radiati<strong>on</strong> Characterizati<strong>on</strong>, <str<strong>on</strong>g>and</str<strong>on</strong>g> SEU Mitigati<strong>on</strong>, of the<br />
Virtex FPGA for Space-Based Rec<strong>on</strong>figurable Computing”, Presented<br />
at IEEE NSREC C<strong>on</strong>ference, October, 2000<br />
http://klabs.org/richc<strong>on</strong>tent/Papers/Xilinx/Xilinx_NSREC2000.pdf<br />
accessed <strong>on</strong> 5th March 2004<br />
[9] Gaisler, J., “Suitability of Reprogrammable FPGAs in Space<br />
Applicati<strong>on</strong>s”, ESA Technical Report, 2002<br />
http://www.estec.esa.nl/microelectr<strong>on</strong>ics/asic/fpga_002_01-0-4.pdf<br />
accessed <strong>on</strong> 5th March 2004<br />
[10] Hodgart, M.S. <str<strong>on</strong>g>and</str<strong>on</strong>g> Tiggeler, H., “Introducti<strong>on</strong> to error Correcting Codes<br />
in Satellite Applicati<strong>on</strong>s”, Technical Report, Surrey Space Centre, 1999<br />
[11] Howard, J.W. et al., “Total Dose <str<strong>on</strong>g>and</str<strong>on</strong>g> Single Event Effects Testing of<br />
the Intel Pentium III (P3) <str<strong>on</strong>g>and</str<strong>on</strong>g> AMD K7 Microprocessors”, IEEE<br />
Radiati<strong>on</strong> Effects Data Workshop Record, pp. 38-47, July, 2001
[12] Irom, F. et al., “Single-Event Upset in Commercial Silic<strong>on</strong>-On-Insulator<br />
PowerPC Microprocessors”, IEEE Trans. <strong>on</strong> Nucl. Sci., Vol. 49, Issue-<br />
6, pp. 3148-3155, December, 2002<br />
[13] A.H. Johnst<strong>on</strong>, “Radiati<strong>on</strong> Effects in Advanced Microelectr<strong>on</strong>is<br />
Technologies”, IEEE Trans. Nucl. Sci., Volume-45 , Issue-3 , pp. 1339-<br />
1354, June, 1998<br />
[14] Johnst<strong>on</strong>, A.H., “Scaling <str<strong>on</strong>g>and</str<strong>on</strong>g> Technology Issues for Soft Error Rates”,<br />
4th Annual Research C<strong>on</strong>ference <strong>on</strong> Reliability, Stanford University,<br />
October, 2000<br />
http://nepp.nasa.gov/docuploads/40D7D6C9-D5AA-40FC-<br />
829DC2F6A71B02E9/Scal-00.pdf accessed <strong>on</strong> 5th March 2004<br />
[15] Katz, et al., “Radiati<strong>on</strong> effects <strong>on</strong> Current Field Programmable<br />
Technologies”, Presented at IEEE NSREC C<strong>on</strong>ference, July, 1997<br />
http://klabs.org/richc<strong>on</strong>tent/Papers/nsrec97fpgapaper.PDF accessed<br />
<strong>on</strong> 5th March 2004<br />
[16] Katz, R. et al., “Logic design Pathology <str<strong>on</strong>g>and</str<strong>on</strong>g> Space Flight Electr<strong>on</strong>ics”,<br />
MAPLD <str<strong>on</strong>g>Inter</str<strong>on</strong>g>nati<strong>on</strong>al C<strong>on</strong>ference Proceedings, 1999<br />
http://klabs.org/richc<strong>on</strong>tent/Papers/Mapld99/A4_Katz_P.pdf accessed<br />
<strong>on</strong> 5th March 2004<br />
[17] Koga, R. et al., “Techniques of Microprocessor Testing <str<strong>on</strong>g>and</str<strong>on</strong>g> SEU rate<br />
Predicti<strong>on</strong>”, IEEE Trans. <strong>on</strong> Nucl. Sci., Vol.-32, Issue- 6, pp. 4219-<br />
4224, December, 1985<br />
[18] Koga, R. et al., “Single Event Functi<strong>on</strong>al <str<strong>on</strong>g>Inter</str<strong>on</strong>g>rupt (SEFI) Sensitivity in<br />
Microcircuits”, Proc. 4th European C<strong>on</strong>ference <strong>on</strong> Radiati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> its<br />
Effects <strong>on</strong> Comp<strong>on</strong>ents <str<strong>on</strong>g>and</str<strong>on</strong>g> Systems (RADECS), pp. 311-318, 1997<br />
[19] Label, K. et al., “SEU Tests of 80386 Based Flight-Computer/Data<br />
H<str<strong>on</strong>g>and</str<strong>on</strong>g>ling System <str<strong>on</strong>g>and</str<strong>on</strong>g> of Discrete PROM <str<strong>on</strong>g>and</str<strong>on</strong>g> EEPROM Devices <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
SEL tests of Discrete 80386, 80387,PROM, EEPROM, <str<strong>on</strong>g>and</str<strong>on</strong>g> ASICs”,<br />
IEEE Radiati<strong>on</strong> Effects Data Workshop Record, pp. 1-11, 1992<br />
[20] Label, K. et al., “Radiati<strong>on</strong> Effect Characterizati<strong>on</strong><str<strong>on</strong>g>and</str<strong>on</strong>g> Test Methods of<br />
Single Chip <str<strong>on</strong>g>and</str<strong>on</strong>g> Multi Chip Stacked 16Mbit DRAMs”, IEEE Trans. <strong>on</strong><br />
Nucl. Sci. Vol 43, Issue-6, pp. 2974- 2981, December, 1996<br />
[21] Label, K., “Single Event Effect Test Results for C<str<strong>on</strong>g>and</str<strong>on</strong>g>idate Spacecraft<br />
Eectr<strong>on</strong>ics”, IEEE Radiati<strong>on</strong> Effects Data Workshop, pp. 14-21, July,<br />
1997<br />
[21] Label, K. et al., “Anatomy of an In-Flight Anomaly: Investigati<strong>on</strong> of<br />
Prot<strong>on</strong>-Induced SEE Test Results for the Stacked IBM DRAMs”, IEEE<br />
Trans. <strong>on</strong> Nucl. Sci., Vol 45, Issue-6, pp. 2898-2903, December, 1998<br />
[22] A.J.Lelis et al., “Radiati<strong>on</strong> Resp<strong>on</strong>se of Advanced Commercial<br />
SRAMs”, IEEE Trans. Nucl. Science, Vol.43, Issue-6, pp. 3103-3108,<br />
December, 1996<br />
[23] T.P. Ma <str<strong>on</strong>g>and</str<strong>on</strong>g> P.V. Dressendorfer, “I<strong>on</strong>izing radiati<strong>on</strong> effects in MOS<br />
devices <str<strong>on</strong>g>and</str<strong>on</strong>g> circuits”, Wiley <str<strong>on</strong>g>Inter</str<strong>on</strong>g>-science, New York, 1989<br />
[24] Massengill, L.W, “Cosmic <str<strong>on</strong>g>and</str<strong>on</strong>g> Terrestrial Single Event Radiati<strong>on</strong><br />
Effects in Dynamic R<str<strong>on</strong>g>and</str<strong>on</strong>g>om Access Memories”, IEEE Trans. <strong>on</strong> Nucl.<br />
Sci, Vol. 42, Issue-2, pp. 576-593, April, 1996<br />
[25] Miyahira, T. <str<strong>on</strong>g>and</str<strong>on</strong>g> Swift, G., “Evaluati<strong>on</strong> of Radiati<strong>on</strong> Effect in Flash<br />
Memories”, JPL Publicati<strong>on</strong>, 2003<br />
[26] Moran, A. et al., “Single Event Effect testing of the Intel 80386 Family<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> the 80486 Microprocessor”, Proc.3rd European C<strong>on</strong>ference <strong>on</strong><br />
Radiati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> its Effects <strong>on</strong> Comp<strong>on</strong>ents <str<strong>on</strong>g>and</str<strong>on</strong>g> Systems (RADECS), pp.<br />
263-269, September, 1995<br />
[27] Nguyen, D.N <str<strong>on</strong>g>and</str<strong>on</strong>g> Scheick, L.Z., “SEE <str<strong>on</strong>g>and</str<strong>on</strong>g> TID of Emerging N<strong>on</strong>-<br />
Volatile Memories”, JPL Publicati<strong>on</strong>, 2002
[28] Nguyen, D.N. et al., “Radiati<strong>on</strong> Effects <strong>on</strong> Advanced Flash Memories”,<br />
IEEE Trans. <strong>on</strong> Nucl. Sci., Vol 46, Issue-6, pp. 1744 – 1750,<br />
December, 1999<br />
[29] A. Ochoa, Jr., <str<strong>on</strong>g>and</str<strong>on</strong>g> P.V. Dressendorfer, “A Discussi<strong>on</strong> of the Role<br />
Distributed Effects in Latchup,” IEEE Trans. Nucl. Sci., vol. NS-28, pp.<br />
4292–4294, Dec. 1981<br />
[30] Oldham, T.R., “How device scaling affects single event effects<br />
sensitivity”, IEEE NSREC Short Course, July, 2003<br />
[31] Petersen, E.L et al., “Calculati<strong>on</strong>s of Cosmic Rays Induced Soft Upset<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> Scaling in VLSI Devices”, IEEE Trans. Nucl. Sci., Vol-29, Issue-6,<br />
pp. 2055-2063, December, 1982<br />
[32] C.Poivey et al., “Radiati<strong>on</strong> Characterisati<strong>on</strong> of Commercially Available<br />
1Mbit/4Mbit SRAMs for Space Applicati<strong>on</strong>s”, NSREC Data Workshop,<br />
1998<br />
[33] Rodgers, J et al., “Advanced Memories for Space Applicati<strong>on</strong>s”, IEEE<br />
Microelectr<strong>on</strong>ics Reliability & Qualificati<strong>on</strong> Workshop, December, 2001<br />
http://parts.jpl.nasa.gov/mrqw/mrqw_presentati<strong>on</strong>s/S1_rodgers.pdf<br />
accessed <strong>on</strong> 5th March 2004<br />
[34] N.S. Saks, M.G. Anc<strong>on</strong>a <str<strong>on</strong>g>and</str<strong>on</strong>g> J.A. Modola, “Radiati<strong>on</strong> Effects in MOS<br />
Capacitors with Very Thin Oxides at 80K”, IEEE Trans. Nucl. Sci., Vol.<br />
31, pp. 1249, 1984<br />
[35] J.E. Schroeder, A. Ochoa, Jr., <str<strong>on</strong>g>and</str<strong>on</strong>g> P.V. Dressendorfer, “Latchup<br />
eliminati<strong>on</strong> in bulk CMOS LSI circuits,” IEEE Trans. Nucl. Sci., Vol-27,<br />
pp. 1735–1738, December, 1980<br />
[36] Schwartz, H.R. et al., “Single Event Upset in Flash Memories”, IEEE<br />
Trans. <strong>on</strong> Nucl.Sci., Vol-44, Issue-6, pp 2315-2324, December, 1997<br />
[37] Seifert, N. et al., “Impact of Scaling <strong>on</strong> Soft-Error Rates in Commercial<br />
Microprocessors”, IEEE Tans. <strong>on</strong> Nucl. Sci., Vol. 49, Issue-6, pp. 3100-<br />
3106, December, 2002<br />
[38] Shivakumar, P. et. al, “Modeling the Effect of Technology Trends <strong>on</strong><br />
Soft Error Rate of Combinati<strong>on</strong>al Logic”, Proc. <str<strong>on</strong>g>Inter</str<strong>on</strong>g>nati<strong>on</strong>al<br />
C<strong>on</strong>ference <strong>on</strong> Dependable Systems <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Network</str<strong>on</strong>g>s (DSN), pp. 389-<br />
398, June, 2002<br />
[39] Swift, G.M., <str<strong>on</strong>g>and</str<strong>on</strong>g> Guertin, S.M., “In-flight Observati<strong>on</strong>s of Multiple Bit<br />
Upset in DRAMs”, IEEE Trans. <strong>on</strong> Nucl. Sci., Vol. 47, Issue-6, pp.<br />
2386-2391, December, 2001<br />
[40] Underwood, C.I., “Single Event Effects in Commercial Memory Devices<br />
in the Space Radiati<strong>on</strong> Envir<strong>on</strong>ment”, PhD Thesis, University of<br />
Surrey, 1996<br />
[41] Underwood, C.I., “Observati<strong>on</strong>s <strong>on</strong> the reliability of COTS-devicebased<br />
SSDRs operating in low earth orbit”, Fifth European c<strong>on</strong>ference<br />
<strong>on</strong> Radiati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> its Effects <strong>on</strong> Comp<strong>on</strong>ents <str<strong>on</strong>g>and</str<strong>on</strong>g> Systems (RADECS),<br />
pp. 387-393, September, 1999<br />
[42] Velazco, R. et al., “SEU Testing 32-bit Microprocessors”, IEEE<br />
Radiati<strong>on</strong> Effects Data Workshop Record, pp. 16-20, July, 1992<br />
[43] Products guide www.Simmtester.com. Radiati<strong>on</strong> hardened memory<br />
devices<br />
[44] C Claeys <str<strong>on</strong>g>and</str<strong>on</strong>g> E Simoen, Literature study <strong>on</strong> radiati<strong>on</strong> effects in<br />
advanced semic<strong>on</strong>ductor devices, ESTEC c<strong>on</strong>tract report P35284-IM-<br />
RP-0013, March 30 2000<br />
[45] A H Johnst<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> B G Rax, "Prot<strong>on</strong> damage in linear <str<strong>on</strong>g>and</str<strong>on</strong>g> digital<br />
optocouplers," IEEE Trans Nucl Sci, vol 47, pp 675 -681, 2000.<br />
[46] http://klabs.org/richc<strong>on</strong>tent/MAPLDC<strong>on</strong>98/Papers/c4_miyahira.pdf<br />
accessed <strong>on</strong> 5th March 2004<br />
[47] http://parts.jpl.nasa.gov/docs/PID16621.pdf<br />
accessed <strong>on</strong> 5th March,<br />
2004