distributed by: national technical information service u ... · national technical information...

AD-777 186

A STUDY OF ELECTRONICS RADIATION HARDNESS ASSURANCE TECHNIQUES. VOLUME II, PART 2, ELECTRICAL SCREENING FOR IONIZING RADIATION RATE AND TOTAL DOSE EFFECTS

I. Arimura, et al

Boeing Company

Prepa red for:

Air Force Weapons Laboratory

January 1974

DISTRIBUTED BY:

National Technical Information Service U. S. DEPARTMENT OF COMMERCE 5285 Port Royal Road, Springfield Va. 22151

5 *

.

*^SBJ ' ■■ . m

jCKse "■' tjr

KTI3 X -■•» I DIC F. . i' -:t:jfl f]

j utm*;: "2D Ü

i CA 'OK

lir 1 :■■■ * .

?:■ *ii> »MR.« "U irr K3ES

^ AIR FORCE WEAPONS LABORATORY

Air Force Systems Command Kirtland Air Force Base

New Mexico 87117

*l When US Government drawings, specifications, or other data are used for

any purpose other than a definitely related Government procurement operation, the Government thereby incurs no responsibility nor any obligation whatsoever, and the fact that the Government may have formulated, furnished, or In any way supplied the said drawings, specifications, or other data, is not to be regarded by implication or otherwise, as in any manner licensing the holder or any other person or corporation, or conveying any rights or permission to manufacture, use, or sell any patented invention that ay in any way b<» related thereto.

DO NOT RETURN THIS COPY. RETAIN OR DESTROY.

/>-

'SuMtäU« . I

UNCLASSIFIED Srtunty Classification Kb-TillZL

DOCUMENT CONTROL DATA R&D (Security classification of title, body ol abstract and indexing annottt'ivn must be entered when the overall report la classified)

i ORIGINATING ACTivtiy (Corporate author)

The Boeing Company •'Seattle, Washington 98124

Zm. REPORT VC C U «* I T V CLASSIFICATION

UNCLASSIFIED 2b. GROUP

REPORT TtTLE

A STUDY OF ELECTRONICS RADIATION HARDNESS ASSURANCE TECHNIQUES

Volume II, Electrical Screening for Ionizinq Radiation Rate and Total Dose Effects, Part 2

». DESCRIPTIVE NOTES (Type ot report end inclusive dale«;

31 July 1970 through 16 July 1973 5 AUTHORS) ;fj:r«f name, middle initial, teat name)

I. Arimura, et al.

a. REPORT DATE

January 1974 na. CONTRACT oa GRANT NO

F2960 .-71-C-0Ö01 b. PROJt.C T NO

8809, 133B, WDNE r- Task 11, DE, 01 d.

Ja. TOTAL NO OF PACES

3INATOR'S *.EFO^T NUMBER«51

7h HO IF RfFJ

44

AFWL-TR-73-m, Vol. II, Pt 2

96. OTHER REPORT NO^'Sl (Any other number» that may be astlQred this report)

10 DISTHlBUriOU STATEMENT

Approved for public release; distribution unlimited. NATIONAL TECHNICAL INFORMATION SERVICE U S Department of Commerce

Springfield VA 22151

11 SUPPLF.MEM . a P- "OTeS 12 SPONSORINS MILITÄR.' ACT'VFTr

Air Force Weapons Laboratory (ELP) Kirtland Air Force Base, NM 87117

\Distribution Limitation Statement A) ill ABSTRACT

Thii program determined physical failure modes of a range of discrete and integrated semiconductor devices exposed to ionizing rate, neutron, and total dose environments. rrorn physical reasoning possible electrical parameter measurements were determined which had some probability of correlating with the radiation sensitivity and failure thresholcs. It was determined that base transit time normal i^4-ion for nejtrcn degradation was genera'My effective for low-power transistors, but ineffective for power devices. Other AC and a few DC measurements also were found to be effective potential screens for neutron degradation. No particular advantage was noted for using electrical storage time constant s compared to electrical storage time for screening primary photocurrents of low-o^wer discrete devices. In some cases, the integrated circuits were obtained with nonstanda,~d metallization, "special lead," topologies to enable electrical measurements to b2 made at internal circuit nodes. In I'r's Volume (Volume II), the utility of these measurements as correlation parameters K.-s compared to that obtained from measurements made using unmodified circuits. Excellent correlation was obtained between tie neutron degradation of the logic circuits and the emitter-base turn-on voltage obtained from the leasurements made usinc special leads. Electrical screens for total dose hardness assurance were found to be relatively ineffective even with parameter correlation factors of 0.7 to 0.8. Similar results were obtained from an evaluation of the low dose screening concept since relative device sensitivity varied with absorbed dose. A mathematical expression was developed for the neutron induced reduction of the energy required for second

\J \J I NOV «s l*t / O I UNCLASSIFIED Security Classification

UNPASSIF1EÜ Sacurlty Classification

K«V KOKOI

Hardness Assurance Transient Radiation Effects Integrated Circuits Radiation Quality Assurance Irradiate-Anneal

ABSTRACT (Cont'd)

breakdown in ehe investigated power transistor. Second breakoown of dielectrically isolated inte grated circuit (logic) transistors was investigated and determined not to be a problem in neutron or ionizing r?te environments. Mean time between failure was not significantly affected (1) by any of the annealing schedules, or (2) by a nondestructive

jj second breakdown test used in the program, or (3) by exposure to an ionizino rate environment of lO- to 10io rad (Si)/s. Additional Volumes (Volu es I and III) contain a description of the program, summary and conclusions, and homogeneity and irradi anneal studies.

ate-

/ */ UNCLASSIFIED Security Classification

AFWL-TR-73-134, Vol. II, Pt 2

A STUDY OF ELECTRONICS RADIATION HARNESSS

ASSURANCE TECHNIQUES

Volume II, Part 2

Electrical Screening for Ionizing Radiation Rate and

Total Dose Effects

I. Arimura, et al.

The Boeing Company Seattle, Washington 98124

Final Report for Period 31 July 1970 through 16 July 1973

Approved for public release; distribution unlimited.

This work was sponsored in part by the Defense Nuclear Agency under Nuclear Weapons Effects Research Subtask TD-072, Work Unit No. 1, Electrical Parameter Statistical Correlation to Radiation Effects.

I JU

KOREWORD

This report was prepared by Tne Boeing Company, Seattle, Washington, under Contract F29601-71-C-0001. The research was performed under Program Elements 62601F, T.213F, and 627Ü7H; Projects 8809, 133B, and WDNE; Tasks DE, 1«, and 01.

Inclusive dates of research were 31 July 1970 through 16 July 1973. The report was submitted 29 October 1973 by the Air Force Weapons Laboratory Project Officer, Mr. John L. Mull is (ELP).

Principal authors and contributors were Mr. Allan H. Johnston, Dr. L. L. Sivo, and Mr. D. W. Egelkrout. Technical direction and coordination of the program were performed by Dr. R. S. Caldwell and Mr. C. Rosenberg.

Capt J. L. Guidry, Capt G. B. Crocker and Capt P. J. Vail at the Air Force Weapons Laboratory also made significant contributions to the overall planning and execution of the program.

Mr. L. D. Milliman, Mr. K. D. Friddell, Mr. J. Thomas, and Dr. R. C. Kennedy assisted with th? large tasks of Data Storage, Retrieval, and Analysis while Messrs. E. T. Ball, E. A. Gallaway, J. F. Foster, D. L. Cushing, S. D. Brooks, and S. Cady performed the Experimental Measurements. Documentation was very ably handled by Mrs. E. L. Taylor, Mr. C. R. Brittain, and Mr. R. Kent. Mr. D. K. Myers at Fairchild Semiconductor assisted with the low power device selection and Mr. K. Martin at Texas Instruments and Mr. R. Jarl ac RCA provided timely support of the Irradiate-and-Anneal Program. Mr. J. Snyder and his staff at Sandia Laboratories were very helpful in providing the necessary reactor tests.

This technical report has been reviewed and is approved.

JOHN L. MULL IS Pro;^ct Officer

GORDON G. WEPFER / \) /JOHN P. PORTASIK Lt Colonel, USAF i/ Colonel, USAF Chief, Phenomenology and Technology Chief, Electronics Division

Branch

ii

PREFACE

This report describes the results of a comprehensive study which

was designed to determine improved techniques for providing radiation

hardness assurance on modern electronic systems. The two basic goals

considered were (J) to determine from physical reasoning and large scale

testing the effectiveness of established electrical screening parameters

and the existence of additional ones which might be correlated with radia-

tion responses and (2) to establish a statistical comparison between the

various hardness assurance techniques including electrical screening, lot

sampling and irradiate-and-anneal. Tor reasons of physical convenience,

the report is divided into three volumes:

Volume I - Background, Approach, and Summary of Results

Volume II - Electrical Screening, parts 1, 2, and 3

Volume III - Lot Sampling and Irradiate-and-Anneal

This Volume (Volume T.I) contains a detailed description of the results

obtained from the. electrical screening portion of the program. The elec-

trical screening approach examined correlations between certain initial

electrical parameters and the radiation sensitivities of the devices.

The correlation parameters were selected on the basis of physical reasoning

and the radiation sensitivies were defined differently for the various

radiation environments. Neutron hardness assurance is treated first and

the various classes of devices such as low-power transistors, high-power

transistors, JFETs and ICs are discussed separately. Ionizing radiation

rate hardness assurance is treated second with subdivision determined

again by the various classes of devices. MT3F results are also discussed

for parts subjected to ionizing rate tests. Total dose hardness assurance

is discussed third for the low-power transistors and for the op amp

separately. Low dose screening is included in this section although it

differs slightly from the "normal" techniques of electrical screening.

Finally, second breakdown hardness assurance is discussed in its entirety.

IV ,-

CONTENTS

Section l3&*

III IuNIZING RADIATION RATE HARDNESS ASSURANCE 1

1. INTRODUCTION 1

2. LOW-POWER TRANSISTORS 1

3. POWER TRANSISTORS 5

4. JUNCTION FIELD EFFECT TRANSISTOR b

5. INTEGRATED CTRCUITS

2. ELECTRICAL SCREENING - LOW-POWER TRANSISTORS

8

6. MTBF RESULTS FOR PARTS SUBJECTED TO 21

IONIZING RATE TESTS

IV IONIZING RADIATION TOTAL DOSE HARDNESS ASSURANCE 76

1. INTRODUCTION 76

76

3. LOW DOSE SCREENING - LOW-POWER 81 TRANSISTORS

4. ELECTRICAL SCREENING - yA744 83

OPERATIONAL AMPLIFIER

5. LOW DOSE SCREENING - yA744 34 OPERATIONAL AMPLIFIER

References 117

ILLUSTRATIONS

FiSure Page

49 Histogram of Primary Photocurrents at 5.3 x 108 rad(Si)/s for 2N696 23


51 Histogram of Primary Photocurrents at 6.0 x 109 rad(Si)/s for 2N2905A 25



54 Dose Rate Dependence of Mean Primary Photocurrent for 2N696 and 2N2905A 28

55 Dose Rate Dependence of Mean Primary Photocurrent for 2N2222 and 2N3960 ' 29

56 Scatter Diagram of I [5,3 x 108 rad(Si)/s] versus tSE (50 mA/10mA) for 2N696 30

Q

57 Scatter Diagram of I [5.3 x 10 rad(Si)/s] versus Tg for 2N696 31

58 Histogram of Ip at 3.0 x 10 rad(Si)/s for TA8007 Showing Devices with "Anomalous" Photocurrents 32

59 Scatter Diagram of Base Doping Concentration. Ng0, versus Primary Photocurrent, I_p, at 3.0 x 107 Rad(Si)/s) for TA8007 33

60 Scatter Diagram of rß versus Ipp at 3.0 x 10 rad(Si)/s for TA8007 34

61 Histogram of Threshold Rate for Turn-On (ISp = 2A) for RCA TA8007 in Shorted-base Configuration 35

62 Histogram of Threshold Rate for Turn-On (Isp = 1A) for Solitron BR200A in Shorted- base Configuration 36

63 Scatter Diagram of hpE (3V/1A) versus Turn-On Threshold Rate for RCA TA8007 37

VI

Figure

64

67

68

69

ILLUSTRATIONS (Cont'd)

Scatter Diagram of rß versus Turn-On Threshold Rate for RCA TA80Ü7

65 Scatter Diagram of hFE (3V/1A) versus Turn-On Threshold Rate for Solitron BR200A

66 Scatter Diagram of rß versus Turn-On Threshold Rate for Solitron BR200A

Superlinearity of Primary Photocurrent, IpD

as a Function of Dose Rate for the Dual JFET

Correlation Between CGSg (V=1V) and I for the Dual JFET '" PP

Correlation Between ig and I for the Dual JFET

70 A Sample Histogram of the Radiation Response Threshold Data [TI Inverter 1-State Response]

71 Histogram of TI Inverter i-State Response Thresholds Showing Truncation With Electrical Storage Time

72 Truncation of TI Buffer Radiation Responses with Offset Voltage

73 Histogram of Radiation Response Threshold Data (Motorola Puffer 1-State Response)

74 Method Used to Detect Open Internal Resistors in the Inverter Circuits

75 Truncation of the Word Switch Secondary Photocurrent Thresholds with h„„ and rD

76 Experimental Method Used in LINAC Tests of the Motorola Sense Amplifier

77 Illustration of the Relative Radiation Sensitivities of 2N709, 2N930 and 2N2905A

78 Histogram of AIß Illustrating the Variation in the Radiation Sensitivities Among the 2N709 Transistors of Different Wafers (Dose = 1.25 x 106 rads; I = 3 uA)

79 Histograms of Ig/Iß Illustrating the Variation in the Radiation Sensitivities Among the 2N709 Transistors of Different Wafers (Dose= 1.25 x 106 rads; Ic - 3 uA)

Page

38

39

40

41

42

43

44

45

46

47

48

49

50

85

86

87

VI1


Figure Page

80 Histogram of hpE/hFE0 Hlustrating the Variation

in the Radiation Sensitivities Among the 2N709 f

Transistors of Different Wafers (Dose ■ 1.25 x 106 rads; I_ = 100 uA) 88

b

81 Histogram of A(l/hpE) = Alg/I^ Illustrating the Variation in the Radiation Sensitivities Among the 2N709 Transistors of Different Wafers (Dose = 1.25 x 106 rads; IE = 100 uA) ,89

82 Histogram of Alß Illustrating the Variation in t,.e Radiation Sensitivities Among the 2N930 Transistors of Different Wifers (Dose = 3.0 x 105 rads; IE = 1 uA) 90

83 Histogram of Iß/Ig Illustrating the Variation in the Radiation Sensitivities Among the 2N930 Transistors of Different Wafers (Dose= 3.0 x 105 rads; I£ = 1 pA) 91

84 Histogram of hpE/hpEQ Illustrating the Variation in the Radiation Sensitivities Among the 2N930 Transistors of Different Waf-j.r!= (Dose » 3.0 x 105 rads; I_= 1 mA) 92

b

85 Histogram of A(l/hpE) = Alg/Ic Illustrating the » Variation in the Radiation Sensitivities Among the 2N930 Transistors of Different Wafer« (Dose = 3.0 x 105 rads; I-, = 1mA) 93 * i t

86 Histogram of Alg Illustrating the Variation in »» the Radiation Sensitivities Among the 2N2905A Transistors of Different Wafers (Dose = 5.6 x 10 rads; I„ = 3 uA) 94

b

87 Histogram cf Iß/^ß Illustrating the Variation in the Radiation Sensitivities Among the 2N2905A Transistors of Different Wafers (Dose = 5.6 x 106 rads; I_ = 3 uA) 95

b

88 Histogram of hpE/hEEQ Illustrating the Variation in the Radiation Sensitivities Among the 2N2905A Transistors of Different Wafers (Dose = 5.6 x 106 rads; I = 3 mA) 96

b

89 Histogram of A(l/hFE) = AIB/lc Illustrating the Variation in the Radiation Sensitivities Among the 2N2905A Transistors of Different Wafers (Dose = 5.6 i 106 rads; I,. = 3 mA) 97 p

b

VI11

s

FiS.u.re Page

90 Histogram of Initial Gain, hFE, Marked to Illustrate J f the Extent of the hFEo Versus hFE Correlation

(Marked Devices Came From the Lower Tail of the hFE (Dose) Histogram). [2N2905A, hFE (10mA) <_ 64 at 1.3 x ?06 rad] The Pronounced Structure in the Histogram is Due co the Different Wafers 98

91 Histogram of Initial Gain, hFE0, Marked to Illustrate the Extent of the hEEo Versus hEE Correlation [Marked Devices Came From the Lower Tail of the hpE (Dose) Histogram], [2N9 30, hFE (1mA) < 64 at 3.0 x 105 rad] 99

92 Histogram of Initial Gain, hpEo, Marked to Illustrate the Extent of the hpEQ Versus hpE Correlation [Marked Devices Came From the Lower Tail of the hpE (Dose) Histogram]. [2N709, hFE (1mA) < 19 at 1.3 x 106 rad] " 100

93 Histogram of Low Dose A Iß, Marked to Illustrate the Limitations of the Low Dose Screening. Marked Devices Came From the Upper Tail of the High Dose Alg Histogram (2N930, High Dose = 3 x 105 rads) 101

94 Histogram of Low Dose Alß, Marked to Illustrate the Limitations of the Low Dose Screening. Marked Devices Came From the Upper Tail of the High Dose Alg Histogram (2N2905A, High Dose = 5.6 x 106

rad) 102

95 Histogram of Low Dose Alg Marked to Illustrate the Limitations of the Low Dose Screening. Marked Devices Came From the Upper Tail of the Hip l Dose AIB Histogram (2N709, High Dose = 1.3 x 106 rad) 103

96 Histogram of the Initial Bias Currents Illustrating the Variation Among the Op Amps 104

97 Histogram of Alg Illustrating the Variation in the Radiation Sensitivities Among the Op Ams (Dose = 5.6 x 106 rads) 105

98 Histogram of IB (5.6 x 10 rad) Illustrating the Variation Among the Op Amps 106

IX

TABLES

Table

51 Summary of Low Power Transistors Tested for Transient Ionization Effects

Page

51

52 Summary of Photocurrent D2ca for Low-Power Transistors

5?

53 Comparison of Electrical Storage Time and Storage Time Constant, as Screening Parameters for Primary Photocurrent

54 Summary of Rank Correlation Coefficients for Various Screening Parameters versus I for Low-Power Transistors PP

53

54

55 Relative Efficacies of Various Screening Parameters for Primary Photocurrents

56 Summary of MLR Predictions of Ipp for 2N29C5A and 2N2222 Low-Power Transistors Using Total Sample

57 Summary of Rank Correlation Coefficients for I

sp

58 Summary of Rank Correlation Coefficients for Primary Photocurrent - Dual JFET

59

60

61

Summary of MLR for I - Dual JFET PP

Summary of TTL Transient Ionization Data

Some Rank Correlations for the y Response of the TI Inverter

Some Rank Correlations for the y Response of the TI Buffer

54

55

56

57

58

59

60

61

63 Some Renk Correlations for the Ionization Response Threshold of the TI A-O-I Gate

64 MLR Results for TI A-O-I Gate Transient Response Threshold

65 Some Rank Correlations for Ionizing Rate Response of the Motorola Inverter

66 Effects of Electrical Screens on the Regression Results for the Motorola Inverter 1-State y Threshold

62

63

64

65

TABLE (Cont'd)

"V

Table

67 Some Rank Correlations for the y Threshold of the Motorola Buffer

68 Regression Results for the Motorola Buffer Y Threshold

69 Summary of Ionizing Rate Data (Non-TTL Integrated Circuits)

70 Some Rank Correlation Factors for the Ionization Response of the Word Switch

71 Multiple Linear Regression Results for the I Threshold of tbe TI Word Switch sp

7? Some Rank Correlation Coefficients for the Ionizing Rate Response of the Motorola Sense Amp

"3 An Example of MLR Predictions for the Ionizing Rate Response of the Motorola Sense Amp

74 Some Rank Correlation Coefficients for the Ionization Response of the yA744 Op Amp

75 MLR Results for the Ionization Response of the yA744 Op Amp

76 Failure Rates for Parts Subject to Ionizing Rate Tests

77 Rank Correlation Coefficients for Total Dose Damage Prediction (overview)

78 Summary of Data to Illustrate the Radiation Sensitivity of 2N709

79 Summary of Data «-c Illustrate the Radiation Sensitivity of 2N930

80 Summary of Data to Illustrate the Radiation Sensitivity of 2N2905A

81 Results of Rank Correlations Between Radiation Sensitivity (Low Dose) and Radiation Sensitivity (High Dose)

82 Summary of Data Illustrating the Procedure and the Results of the Low Dose Screening (2N930)

66

67

68

69

70

71

72

73

74

75

107

108

109

110

111

112

XI

TJ

Table

83

84

85

86

TABLES (Cont'd)

Summary of Data Illustrating the Procedure and the Results of the Low Dose Screening (2N2905A)

Summary of Data Illustrating the Procedure and the Results'of the Low Dose Screening (2N709)

Rank Correlation Coefficients for Damage Prediction at 5.6 x 10 rads - uA744

Summary of Data Illustrating the Procedure and the Results of the Low Dose Screening

for the yA744 Op Amp

Page

113

114

115

116

*

XI1

/ /

ABBREVIATIONS AND SYMBOLS

A Ampere, area, surface area of second breakdown region

A_ Base area

A„ Emitter area E

A,,™ Effective emitter area fc.fr

OL Open-loop gain

A Cross Sectional area of a second breakdown site SB

AC Alternating current

AH1 Effective emitter area determined from gain-bandwidth product, emitter base capacitance and current gain

AH3 Effective emitter area determined from gain-bandwidth product, current gain, and base doping concentration

AH4 Effective emitter area determined from gain-bandwidth product, breakdown voltage, and current gain

A-O-I And-or-invert

BOT Breakout transistor

BV Breakdown voltage

BV Breakdown voltage of collector-base junction with CBU . . .

emitter open

BV _ The emitter-base breakdown voltage of a transistor with the collector open |

BV Breakdown voltage of gate-channel in JFET. Source and drain shorted

BVCBO Base to collector breakdown voltage, emitter open

BVCEO Open base collector to emitter breakdown voltage, measured at a collector current of 10 milliamperes

BVEBO Open collector, emitter to base breakdown voltage, measured at a base current of 10 milliamperes

C Damage factor, average specific heat, capacitor

xm

-*- . .*- \r

ABBREVIATIONS AND SYMBOLS (Continued)

'CB

'GSS

IB

"OB

SBL

S C-B

CB - X, Y

C-E

CH

CMRR

COV

COVAR

CRO

CSBLXYZ

CSBXYZ

D

DB

DE

DC

D.I.

Collector-base junction capacitance

Gate-channel junction capacitance of JFET

Emitter-base junction capacitance

Collector-base junction capacitance

Damage factor of wafer transistors

Linear second breakdown neutron degradation constant

Neutron damage factor

Collector to base

Neutron damage factor for current gain measured at a collector current of X amperes and collector to emitter voltage of Y volts

Collector to emitter

Channel

Common-mode rejection ratio

Coefficient of variation

standard deviation Covariance =

mean

Cathode ray oscilloscope

Linear second breakdown neutron degradation constant determined from neutron levels ö , $ , and $

x y z

CPT, determined from neutron levels <j> , d> , and <|> SB x y z

Diffusion constant

Diffusion constant for minority carriers in base region

Diffusion constant for minority carriers in emitter region

Direct current

Dielectric isolation

XIV I

/ S


DUT Device under test

E energy or electric field strength, second breakdown energy measured at I_ ■ 5A

E-5MSIA Second breakdown energy for a 5 millisecond wide, 1 ampere pulse of emitter current

E-5MSIB0 Second breakdown energy for a 5 millisecond wide voltage pulse equal to the collector-emitter breakdown voltage, base open

ECR Energy required to produce second breakdown: measured at device terminals

JSB

i

E„ Energy required to produce second breakdown: delivered to a localized site of second breakdown

F F value

FSC Fairchild Semiconductor Corporation

FM1 Transistor figure of merit determined from gain bandwidth product and base doping concentration

FM3 Transistor figure of merit determined from gain bandwidth product and collector to base breakdown voltage, emitter open

FM4 Transistor figure of merit determined from gain bandwidth product and collector to base breakdown voltage, emitter open

FT A constant proportional to the gain-bandwidth product measured at an emitter current of 1 ampere

HA Hardness assurance

HFE - X, Y Common emitter DC current gain measured at a collector current of X amperes and a collector to emitter voltage of Y volts

I Current

IR Base current (bipolar transistors) or input bias current (op amps, sense amps)

I Initial (pre-radiation) base current or bias current ö g

IR The surface component of the base current

XV

EBO

EBO

«e


I-.. The forward current into the base of a saturated transistor

I_2 The current flowing out of the base of a saturated transistor when it is switched out of saturation

I Input bias current BIAS

I _ Base current of saturated transistor

I_ Collector current

I Collector-base junction leakage current

I Power supply current in positive supply lead

I , , Power jupply current of a digital circuit in a 0-state

I n Pre-exponential saturation current

I_„ Collector current of saturated transistor

I, Drain current in JFET

I _s Source to drain saturation current at zero gate bias

I„ Emitter current E

I„Dr. Emitter-base junction leakage current, collector open

I Initial (pre-radiation) emitter-base junction leakage current, collector open

I The surface component of the leakage current across the reverse biased emitter-base junction

XVI

I Power supply current in negative supply lead EE |

I_, Forward current of gate-channel diode in JFET r

ITXT/T>\ Reverse input current measurement for TTL circuits

IN(R; i

I . . Current out of TTL input with input voltage low

IT n. Current into TTL input with input voltage high IN(,1)

I._ Input offset current OS

I Primary photocurrent PP »

j

/"

/

TJ

:0L

h ISK

I sp

IC

ICBO

ICEO

IEBO

I'V

JE

J.I.

h h

ABBREVIATIONS AND SYMBOLS (Continued

Output leakage current of an integrated circuit

Reverse current of gate-channel diode in JFET

The output sink current of a TTL device with lower than normal supply voltage

Secondary photocurrent

Integrated Circuit

Open emitter, collector to base leakage current measured at 50 volts

Open base, collector to emitter leakage current measured at 30 volts

Open collector, emitter to base leakage current at 5 volts

Portion of second breakdown power not directly dissipated in a second breakdown site

Emitter current density

Junction isolation

Damage constant, an abbreviation for Kilo-ohm used in circuit schematics

Thermal conductivity at second breakdown site averaged over temperature

Carrier removal damage factor

Total delay time normalized damage constant

Damage constant due to lifetime degradation in emitter- base depletion region

Diffusion length, length of JFET channel in direction paralleled to current flow

Diffusion length of minority carriers in collector region

Diffusion length of minority carriers in emitter region

xvn

^


LINAC

LSI

M

«B

MB -3.0

MeV

MLR

MLRP

MLR-R

MLR-r

MLR--r

MOT

MSI

MTBF

N

BO

N o

NXYZ

N(CH)

Linear accelerator

Large scale integration

Mass of material at a localized second breakdown site

Temperature sensitivity of the common emitter DC current gain

Temperature sensitivity of the common emitter current gain measured at a collector current of 3 amperes and a collector to emitter voltage of 3 volts

Mega electron volts

Multiple linear regression

Multiple linear regression prediction

Coefficient of multiple linear correlation

Multiple linear regression partial correlation coefficient

Multiple linear regression coefficient

Motorola

Medium scale integration

Mean time between failure

Value of exponent in the equation relating second breakdown energy and neutron fluence. N is determined from, carrier concentration 4 fluence levels

Base doping concentration

Base doping concentration adjacent to emitter region

Minority carrier density in the emitter region

-3 Channel doping concentration for JFET in cm

Value of exponent in equation relating second breakdown energy and neutron fluence. NXYZ is determincu from fluence levels of 4 , 4 and 4

x y z

Channel dopi.ig as calculated from BV measurements

XV1U

-fc-

*v


Noise (1/f) Surface component of the 1/f noise

P Power

P Power of a square pulse required to produce second breakdown

Q A symbol for transistors used in schematic diagrams

R Resistance

*B

h

External resistance from the base of a transistor in an electrical circuit

R^ Maximum percentage range in the data

R^ Radius of emitter region

Load resistor

R^ Maximum percentage range in the prediction

RBC The reciprocal of base impurity concentration

RBV A constant inversely proportional to the 2.8 power of the open collector-emitter, to base breakdown voltage measured at a base current of 10 milliamperes

RCA Radio Corporation of America

RCC The reciprocal of the collector impurity concentration

RCV A constant inversely proportional to the 2.8 power of the open emitter, collector to base breakdown voltage measured at a base current of 10 milliamperes

RMS Root mean square

S Switch

Sat. Saturation

SB Second breakdown

SCR Silicon controlled rectifier

SLR Single linear regression

SSI Small-scale integration

XIX

T Absolute temperature (°K), Temperature

T Ambient temperature

T ,, The turn-off time of an integrated circuit off

t The temperature at a site of second breakdown at the onset of second breakdown

TI Texas Instruments

TTL Transistor-transistor logic

TUT Transistor under test

V Volt

Vnr, Base to emitter voltage BE

V„„^,, Emitter to base forward voltage BEON

Vr Collector voltage

V_„ Collector to base voltage LB

V The positive supply voltage of an integrated circuit CC

V Collector to emitter voltage » CE

V„_/e.T> Collector to emitter voltage for saturated transistor CE(SAT) I'

Vnc Drain voltage

V_D The voltage from the emitter to base of a transistor EB

V__ The negative supply voltage of an integrated circuit EE

VT T/D\

T^ie reverse input voltage of a TTL circuit

V The 1-state output voltage of a digital circuit OH

V The 0-state output voltage of a digital circuit UL

XX

i

• V Gate to source voltage GS

V. Input voltage

I t

.'..>,-,<..■■■- ■ ■■■'• ■■■

OS

VP

v SB

V-I

VBEON'-X

VCES-X

VSB-5A

VSB-5MSIBO

WB

wc

WD

WK2,3,4)

WB

a

c

cm

fT

F a

gm

gmo


The offset voltage of an integrated circuit

Pinch off voltage for JFET

Voltage across a device prior to second breakdown

Voltage-current

Emitter to base forward voltage for an emitter current of X amperes

Collector to emitter saturation voltage at a collector current of X amperes

Voltage across a device prior to second breakdown at an emitter current of 5 amperes

Voltage across a device prior to second breakdown at a delay time of 5 milliseconds and x.xth zero base current

Base width

Collector depletion width of a transistor

Depletion region width of emitter-base junction

Wafer number 1 (2,3,4)

The inverse square root of the common emitter gain- bandwidth product

Channel width in direction perpendicular to thickness and length in JFET

Channel thickness in JFET

Capacitance

Centimeter

Common-emitter gain-bandwidth product

Common base cutoff frequency

Post-irradiation transconductance

Pre-irradiation transconductance

XXI

-fc_

'FE i

hFE «hFEI

FEO

hFE<J>

i n

k

keV

mA

mV

mW

n

P

ns

P o

pF

q

r

rank-r

rB

rBl


Common-emitter DC gain

Common-emitter DC current gain for inverted transistor

Pre-irradiation giin

Gain after neutron irradi?tion

Equivalent input noise current

Boltzmann constant

3 1CT electron-volts

in"3 10 amperes

lCf3 volts

-3 10 watts

Value of exponent in equation relating second breakdown energy and neutron fluence

Exponent of voltage dependence of emitter-base junction capacitance

Electron concentration in intrinsic silicon

Minority carrier density in a p-type semiconductor

-9 10 seconds

Equilibrium hole concentration

-12 10 farads

Electron charge

rads

Rank correlation coefficient

Transverse base resistance

Transverse base resistant for inverted transistor

Collector body resistance

r Resistance of collector region of saturated transistor cs c

r,. v Pre-irradiation value of r,, . d(on)o d(on)

rd(on) The on resistance of a JFET

rad(Si) A deposited dose of 100 ergs per gram of silicon

s Second

t Time

t„ Base transit time a

t. Total delay time

tpn Propagation delay time of a logic circuit

t „ Turn-Off time of the word switch

tp.„~ The delay time of an integrated circuit when the output goes from low to high, measured at 50% of the equilibrium value

t „„ Reverse recovery time of gate-channel diode in JFET

t„¥ Electrical storage time

t„ Saturation time of the word switch

t-statistic Measure of significance for multiple linear regression analyse:

w„ Width of emitter stripe

x Modulation length m °

s s AI (burn-in)The change in I during burn-in

C„(T) The change in I„ AI_(T) The change in I when measured at different temperatures

AI (burn-in)The change in I__,, during burn-in EBO EBG s s

AI__rt The change in I™,, when measured at different temperatures EBO hüU

An Excess electron concentration

<j> Base-emitter junction contact potential Bb

XX111 >

I

J

ABBREVIATIONS VW SYMBOLS (Continued)

a

B

Y

e

e

HS

n

Mp

yA

pF

Us

'B

'SE

Ohm

Grounded emitter DC current'gain (i.e. h™)

Forced beta of saturated transistor (E I /I )

Initial (pre-radiation) grounded emitter DC current gain

Ionizing dose rate

Permittivity of dielectric

Permittivity of vacuum

The electronic charge divided by the product of Boltzmann's constant and the absolute temperature

Hot spot thermal resistance

Dielectric constant

Mobility

Carrier mobility in base region

Carrier mobility in collector region

Electron mobility

Hole mobility

Micro amperes

1C"6 farad

10 second

Resistivity of base region

Resistivity of collector region

Emitter sheet resistivity

Electrical conductivity or standard deviation

Excess minority carrier lifecime

Minority carrier lifetime in base region

»

!

♦ 1

XXIV .-#'

/

/


no

po

SB

SBO

rC

h

*4

*j

0-state

1-state

INI

2N1

3N1

4N1

Minority carrier lifetime in collector region

Minority carrier lifetime in highly p-type

Minority carrier lifetime in highly n-type

Electrical, storage time constant

Storage time constant for gate-channel diode in JFET

Second breakdown delay time

Delay time to second breakdown measured in previous tests

Hot spot thermal time constant

feutron Fluence

Critical dose or threshold dose for failure

First neutron fluence

Second neutron fluenca

Third neutron fluence

Fourth neutron fluence

The gain-bandwidth product of a transistor (f )x 2r

The low output voltage state of a digital circuit

The high output voltage state of a digital circuit

First neutron irradiation

Second neutron irradiation

Third neutron ir*«»»!'"tion

Fourth neutron irradiation

xxv/xxvi

, ■■.,.,.-:..-. -■

SECTION III

IONIZING RADIATION RATE EFFECTS HARDNESS ASSURANCE

1. INTRODUCTION

This section discusses hardness assurance techniques for ionizing

rate effects in discrete transistors, junction field-effect transistors

(JFET), and integrated circuits. For discrete transistors, hardness

assurance techniques were investigated for both primary and secondary

photocurrent. The transistor types studied include high- and low-power

transistors with a wide range of device geometries and electrical prop-

erties. The radiation rate depencence of primary photocurrent was also

considered. For the 1FET, gate primary photocurrent is the fundamental

radiation parameter used in the correlation and hardness assurance studies.

The integrated circuits included in this study are all dielectrically-

isolated, radiation-hardener, devices, and include both digital and linear

circuit types„ The circuit output response to ionizing radiation was used

as the parameter of interest for hardness assurance and correlation studies.

For the digital circuits, the radiation level was varied to find the par-

ticular radiation threshold level at which a given output response occurred,

A detailed description of the various device types and the schematic

diagrams for the integrated circuits are contained in Paragraphs 2-a and

2-b, Section IV, Volume 1. A description of the radiation test conditions

and techniques is included in Paragraph 4-c, Section IV, Volume 1.

2. LOW POWER TRANSISTORS

a. Primary Photocurrent

Steady-state primary photocurrents were measured for five types

of low-power transistors and two types of IC breakout transistors using

the Boeing Radiation Effects Laboratory 10-MeV Electron Linear Accelerator

as an ionization source. Experimental details of the measurements were

described previously in Paragraph 4-b, Section IV, Volume 1. Table 51 is

a summary of the parts tested for primary and secondary photocurrents.

The electrical screening parameters for primary photocurrents were in-

vestigated only for the first seven part types shown in Table 51.

^.'^■C/''-:VH >.? ifc^^^..,/, ;; ■-:. ,Uv/-.-j'^/rfifcV-M..V

3

f

The primary photocurrent', I , measured for five types of low-

power transistors are summarized in the sample histograms shown on Figures

49 through 53. The mean, standard deviation, maximum, minimum and range

of photocurrents are summarized in Table 52. The range of responses for

these part types (given by the maximum I divided by minimum I ) varied

from -1.8 to ~3.5 (excluding one 2N3960 which exhibited a breakdown or

anomalous 1 a factor of 10 greater than the other devices of that tvpe). PP

However, the data at the very high rates are of questionabe accuracy,

particularly for the rates in the high 10 rad(Si)/s range. Accurate

dosimeL.i.*y was difficult at these rates since the electron beam tpot was

quite small and the position of the beam varied from shot-to-shot.

I , based on the simple theory given in Paragraph 3-a, Section

V, Volume 1 for collector-bast primary photocurrent, is expected to vary

linearly with ionizing radiation rate. The rate dependence of I is an

important quantity to establish since extrapolation of data to other rates

is generally required for analyses of system performance. The rate de-

pendence of I was established for four of the part types by measuring

I at two rates. Table 52 includes a summary of linearity of the rate PP dependence of I ; the ratio of the photocurrent per unit rate at the

PP higher and lower rates indicates the rate dependence is nearly linear for

two of the part types with the 2N3960 and 2N2905A of questionable linearity.

The linear dependence of I with rate is shown in Figures 54 and 55 where PP

the mean value of I is plotted against, dose rate for four of the part >. PP

types. The bars associated with each data point are plus and minus one I

standard deviation for the distribution of photocurrents for the ~150 de-

vices of each type.

The only well-known electrical technique for screening transistors

against excessive photocurrents is the method of measuring electrical

storage time, t , as ueveloped by Carr (Ref. 17). The technique was

originally developed to evaluate different device types and wuvks fairly

well for this purpose when restricted (1) to silicon planar diffused and

mesa npn transistors with power ratings less than 800 mW, (2) to dose 6 8

rates between 10 and 10 rad(Si)/s, and (3) to transistors with electrical — 8 —5

storage times between 10 and 10 seconds. Although t can be usnd oh

f /

/

acceptably when comparing different types of devices, it has not been

demonstrated to be useful as a screening parameter for devices of a given

type.

However, as discussed in the theory of primary photocurrents

(Paragraph 3-b, Section V, Volume 1), the main contribution to I origi-

nates almost entirely in the collector regions for non-Au-doped planar

epitaxial devices. For these cases, the collector lifetime, i , and dif-

fusion length are the major parameters controlling 1 . The electrical

screening and correlation parameter then is the electrical storage time

constant, i , rather than t since,

h S 1/2 Tc

where T is obtained from

t = t In -—jr —-— (12) SE S V^E + XB2

The results of the correlations between I and t__ and T are pp SE S

shown in Table 53. Electrical storage time constant, r , appears to

offer practically no improvement as a screening parameter over electrical

storage time, t . This is partly the result of 'the poor quality of the

"fit" of the data to Equation (12) which produced "noisy" values of T

and thus poor rank correlations.

These results are further verified by examining the scatter £

diagrams (shown in Figures 56 and 57) of I versus t and T for one PP bh b

example in Table 53. It can be seen that, while a trend is obvious,

these plots are truly scattered and neither of the parameters are partic- ■i

ularly useful as a screen for I PP

The efficacy of other screens for I were investigated. Ac- PP 1 cording to the theory presented in Paragraph 3-b, Section V, Volume 1,

collector-base junction capacitance, C , and possibly base transit time,

t , should be the next best correlation parameters after r . Neither of B b

these parameters were particularly successful as I screens. The largest

\ ae

correlation coefficient for C was +0.538 for the 2N2905 which is opposite OB

in sign to that anticipated. At least t , however poorly, correlated with B

the I in the proper direction. PP

Tables 54 and 55 briefly summarize the order of efficacy of

potential screening parameters for the various device types. In general,

tor or ic are the best screening parameter for primary photocurrent. CTn ab o " IB as a potential screening parameter was unexpected since it was assumed

that the junction areas were nearly constant for a given device type.

However, C is dependent upon the base doping concentration near the e-b IB

junction and, thus could reflect a large gradient in base imparity doping

concentration. The electric field produced by this gradient w>uld enhance

the carrier flow to the c-b junction and the subsequent photocurrent.

The correlation with t is reduced by the presence of an internal electric D ■ ■

field.

Since no single screening parameter appeared to be successful as

a primary photocurrent screen, multiple linear regression (MLR) analyses

were performed using the parameters predicted to be effective from theo-

retical reasons as well as those the data indicated were correlated to I

These results (summarised in Table 56) were, in general, disappointing

and while effective in reducing the prediction errors, were not suffi-

ciently effective to implement as a screening technique.

PP

b. Secondary Photocurrents

Secondary photocurrents, I , were measured in a grounded base sp

configuration. The main "turn-on" mechanism for this case is the trans-

verse voltage generated by the IR drop of the primary photocurrent and

transverse base resistance. This model (presented in Paragraph 3-b,

Section V, Volume 1) predicts that the following parameters are important

for screening 1 : sp

1. I (and parameters for which I is dependent), PP PP

2. rD, transverse base resistance, B

3. h „, common-emitter DC current gain,

4. N , base region impurity doping concentration, and BO

5. t , base transit time. B

y

JL

In general, the correlations for I were much better than for sp

I . A summary of the correlation coefficient for potential I screening PP sp

parameters are shown in Table 57.

3. POWER TRANSISTORS

a. Primary Photocurrent

Primary photocurrents for power devices, though considerably

larger as a result of larger device geometries, do not differ in principle

from those of low-power devices. However, power devices do exhibit greater

tendencies for secondary effects such as the onset of "anomalous" photo-

currents due to transverse effects enhanced by their larger photocurrents

and geometries.

An inspection of a histogram of I for the RCA TA8007 in Figure PP

58 shows seven devices which exhibited these anomalous I s even at rates 7 PP

as low as 3.0 x 10 rad(Si)/s. Two are shown on the histogram and the

other five had such large I s that they are shown as outside the range

of the histogram. These devices all had low values of base doping con-

centration N . The mean values of the doping level for all devices was 16 -3

~5.6 x 10 cm , while the devices which exhibited anomalous turn-on all

had values ~2 x 10 cm This result is apparent in scatter diagrams

of I versus N„„ and r (Figures 59 and 60). pp BO B

b. Secondary Photocurrent, I , and Turn-on Threshold sp

The radiation rate threshold for I "turn-on" varied nearly a sp i

factor of 25 for the TA8007 ar.n nearly a factor of A for the BR200A. f'

These ranges can be seen in the histograms shown on Figures 61 and 62. i

Based on a simple analytical description for I , the key parameters for sp

screening I are: (1) h , (2) r , and (3) I . . sp FE B pp

The scatter diagrams in Figures 63 through 66 show the strong

correlation between turn-on threshold and both h and r . Thus, where bt, B

transient photocurrents constitute the major failure threat, screens on

high h„„ and rD values would significantly truncate those failures which FE B occur at the lower rates.

4. DUAL JUNCTION FIELD EFFECT TRANSISTOR

Primary photocurrent was measured at three dose rates for the dual

JFET. The experimental arrangement was similar to that used for the bi-

polar transistors with the collector-base junction being replaced by the

channel-gate junction. Source and drain were shorted together in this

measurement and 30V was applied across the gate-channel junction.

Figure 67 shows the response (I ) as a function of dose rate, y,

for tha sample. In this plot the "A" and "B" halves are both shown.

We can see that the primary photocurrent is super-linear over the

range of y which was used. Therefore, photocurrent is following a

relationship

I = Ay" (13) PP

where n > 1. For the mean of our sample n was found to be ~1.2.

This super-linearity is expected due to two-dimensional mechanisms

and the fact that the reverse bias was relatively near the breakdown

voltage of the gate-channel junction (BV ).

a. Correlation Parameters

Parameters which are expected to correlate with I are: the PP

reverse recovery time of the gate-channel diode, t , the storage time

constant, T , defined as

(14)

and the width of the depletion region which is inversely proportional to ■

the depletion capacitance. It was expected that the lifetime measurements

would correlate pcsitively with I and that the capacitance would cor- PP

relate negativelv with I PP

Table 58 shows a summary of rank correlation coefficients for

the various parameters of interest. The parts have been divided into

two groups labeled u and ß; group a contains serial numbers 1-35, and *

;

I

■i

tfmta* ik.

group ß contains serial numbers 36-60. The reason for this grouping is

that the two groups seem to come from different wafers or diffusions.

Histograms of many of the electrical parameters appear as bimodal distri-

butions with the groups appearing as two separate distributions. Hence,

when a rank correlation is performed between two parameters which are

each bimodally distributted, the correlation may be artifically high be-

cause of the separation of the two groups. This is shown by Figure 68 O

which shows a scatter diagram C„„_ (IV) versus I [y = 1.2 x 10 rad(Si)/sl baa pp

One can see from this plot and Table 58 that the rank correlation of the

total (0.738) is high and that the correlation with capacitance within

each group is significantly lower.

It is apparent from Table 58 that the best correlations were ob-

tained with the parameters which are a function of lifetime. The storage

time constant, i„, in general correlates as well or better than the reverse

recovery time, and it appears that this parameter is the most likely can-

didate for a screening parameter for primary photocurrent. Note that there

is a marked difference in the correlations between I and T_ observed in pp S

the two subgroups. The a subgroup data correlates much better than the

ß subgroup. This is, perhaps, more apparent in Figure 69 which shows

a scatter diagram of the "A" chips.

b. Multiple Linear Regression Analyses

Multiple linear regression techniques were used to predict pri-

mary photocurrent using the correlation parameters discussed in the pre-

ceding section. Since the JFETs were apparently grouped into two sub-

groups, which implied that the parts came from either two wafers or

diffusions, this allowed a measure of the predictive ability of the re-

gression coefficients of the MLR parameters. For this to be the case,

the independent variables used in the regression must be variables which

physically describe the dependent variable. Were this not the case, the

MLR coefficients would fit the data for which the coefficients were gen-

erated, but would not accurately fit another set of parts with slightly

different physical properties.

MLR's were therefore run using t , T and C and functions rrGS s GSb

of these parameters. Table 59 summarizes the results of these analyses.

We can see from this table that MLR techniques do not work well for the

prediction of the response of devices with electrical parameters much

different than those of the devices from which the MLR coefficients were

generated unless the independent variables are theoretically plausible.

The first two sets of MLR's shown used independent variables related to

I but with a meaningless functional relationship. Hence, the prediction pp 6 ——— * ' * errors are large. The third run, however, used x and 1/C which are

o Go S expected to be related to I and, as expected, the regression works well.

PP As Figure 69 showed in the previous material (Paragraph 4-b) the groups

(a and 3) show quite different degrees of correlation with T . The ad-

dition of the capacitance term, however, allowed a much better estimate

of I than obtained with T alone. PP s

c. Conclusion

As would be expected lifetime measurements give the best cor-

relation and prediction for I in the dual JFET. Therefore, an MLR PP

technique or an electrical screen using t or several values of t rrGo rrub

to give a fitted value of T should provide the best HA screen for de-

vices of this type.

5. INTEGRATED CIRCUITS

The transient ionization response of the integrated circuits was

measured with direct exposure to 10 MeV electrons from a linear accel- j *■„••

eratcr. The pulse width used was sufficiently wide to assure time equili- |,*1

brium, and pulse widths ranged from approximately 100 ns for the TTL

devices to 3 us for the op amp. The decision to use a wide radiation

pulse was reached because of the difficulty of measuring the response of

fast circuits to narrow pulses and because of the more universal appli-

cability of hardness assurance data obtained with wide radiation pulses,

.£., if t ;e circuit is hard to a pulse width sufficient for time equili-

brium, it is generally hard for any pulse width. Extreme care was taken

to assure repeatability in the transient ionization measurements. Thres- i

hold response data was in general repeatable to better than ±5% relative

-=V

accuracy. Because this type of experimentation forces continual adjust-

ments of the radiation intensity, it is very time consuming, and to

minimize the expenditure of effort data was not taken for all sections of

multiple devices. Additional general information on the experimental

procedures is given in Paragraph 4-c, Section IV, Volume 1.

a. TIL Circuits

(1) Failure Criteria

The TTL circuits were loaded with a resistor-diode network

that simulated maximum fanout for both logic states. The loading condi-

tions for the 1-state are particularly important when measuring the icni-

zation response because this causes the pull-up transistor tu be in the

active region which minimizes the output transient. In general, the

failure criterion used for either logic state for these circuits was a

100 mV transient voltage shift at the output terminal, i.e., the failure

criterion was equivalent to a 100 mV reduction in the allowable noise

margin. Trr'.s seems low, but it is the typical fraction of the normal 400

mV noise margin which is relinquished by the system designers for transient

radiation. Most of the available noise margin is usually required for

normal electrical design purposes. Since these circuits fail in either

state because of secondary photocurrent in internal transistors, the

radiation response is strongly nonlinear with dose rate, and moderate dif-

ferences in response criteria will cause only slight differences in the

radiation level for transient failure. One practical difficulty which

arose during the 1-state tests was the stability of the power supply volt-

age. The 10 uf capacitor placed close to the device to bypass the power

supply leads still allowed V to change by as much as several hundred

millivolts. This in turn affected the 1-state output voltage. To solve

this problem, it was necessary to place a large (4000 uf) capacitor a

few feet away from the experiment, out of the radiation beam. This held

the supply voltage to within 20 mV during the transient pulse.

The transient voltage shift was used as a failure criterion

instead of the absolute DC output voltage because: (1) it was experi-

mentally more convenient to measure a consistent pulse amplitude, (2)

failure mechanisms involving secondary photocurrent are likely to be more

i i iW "

consistent from unit to unit, and (3) circuits which have DC log'.c levels

close to the worst-case values for these parameters will fail with tran-

sient output voltages of a few hundred millivolts.

(2) Failure Mechanisms

The expected failure mechanism for the 1-state response of

the TTL devices is turn-on of the output transistor. Obvious correlation

parameters for secondary photocurrent are the h and r_ of the output tt B transistor and the value of the external base resistance. Storage time,

capacitance and stored charge were expected to correlate with primary

photocurrent.

The expected failure mechanism for the 0-state is turn-on

of the pull-up transistor. One obvious correlation parameter is the value

of the external base resistance. For all of the TI circuits except the

inverter, this resistor was significantly lower than for the corresponding

circuits made by Motorola (lKi; versus 4KQ for the standard gate). Be-

cause the secondary photocurrent of the pull-up transistor may cause

transient increase in the ^n-p(QAT) °^ t*ie outPut transistor, the dynamic

resistance of the circuit in the 0-state is also important.

(3) Radiation Data

Transient radiation data for the five TTL device types were ,

taken for approximately 140 units of each type. The sample histogram of

the data in Figure 70 shows the transient threshold data for the TI

inverter (1-state). The data for all of the device types is summarized jv

in Table 60, which lists the mean, standard deviation, ratio of maximum

to minimum value, and worst or most radiation sensitive value. This is

simply a tabular presentation of the radiation data which allows the

reader to make a comparison of the data for different device types and

different test conditions. Since the data for the TI buffer has a maxi- i

mum to minimum ratio of only 2, one can anticipate more difficulty in

obtaining high rank correlation coefficients for this device.

The limited spread in the radiation data for most of these

devices was a major limitation in investigating the effectiveness of

various electrical parameters in screening the radiation response distri- *

butions. The parts had already been through a stringent selection procedure

10

because of the electrical specifications and processing controls which

were included in the parts specifications. This is probably the reason

that the AC and switching measurements were relatively ineffective. Any

device with extreme AC parameters had already been eliminated. Another

aspect is that the basic resolution and accuracy of these AC measurements

are significantly lower than those of the DC parameters and hence, these

measurements would be expected to be less effective in correlations with

tightly grouped radiation data. An alternative approach would have been

to deliberately allow some fraction of the devices in the test population

to te out of tolerance. This would have broadened the range of the elec-

trical parameters, and made it easier to assess correlation effectiveness.

However, this approach would also raise a question about the applicability

to practical groups of devices which do undergo a stringent initial elec-

trical screen.

In the following material, it is apparent that the results

are best for devices which have a wide spread in radiation parameters.

The relative accuracy of the radiation dosimetry and the electrical meas-

urements are less important for devices with wide spreads in radiation be-

havior. In most applications, a range of 3 to 4 ii. the spread of radiation

behavior would be acceptable, and there would be no need to apply tech-

niques to further truncate the distribution. The main goal of hardness

assurance is to eliminate devices from one end of the distribution, and

it is difficult to accomplish this goal if the sample of parts used do

not have sufficient variability.

(4) TI Inverter

Some interesting results were obtained for the TI inverter

1-state response. High rank correlations were obtained for h„„, as shown

in Table 61. However, when h „ was appliel as a screening parameter,

eliminating devices with highest h„„ values, "wo devices stubbornly per-

sisted in the lower, more radiation sensitive, side of the histogram, and

it was apparent that, in spite of the high correlation coefficient, h rt

was not working well as a single correlation parameter. The electrical

parameters of the two devices were examined in more detail, and it was

found that both devices had abnormal values of V , one being very high,

11

the other very low - in fact, the low device was just beneath the 2.4V

specification limit. The reason for the high V reading was an open in- Un

ternal resistor. This same problem occurred for one of the Motorola cir-

cuits, and is further discussed in Paragraph 5-b of this Section. When

V was included as a screening parameter, the results shown in Figure OH

71 were obtained. These results suggest that abnormal device behavior,

in spite of electrical test limits, can still affect radiation behavior,

and support the concept of monitoring a wide number of device parameters

to achieve hardness assurance. The significance of V as a parameter for On

electrical evaluation of TTL circuits is discussed in detail in Paragraph

5-b of this Section. These results show that radiation hardness can be

improved by monitoring appropriate data, adding additional tests, and

imposing tighter limits for some parameters. The existing limits in the

Honeywell specification were not sufficient for this pur-pose.

A further interesting result was obtained for the TI in-

verter. Returning to Figure 71, one device is conspicuously located on

the leading edge of the histogram even after the h_ and V , screens were rh On

applied. This device had the highest value of propagation delay time, and

this value was significantly greater than that of any of the other devices.

This is one of the few examples of successful application of switching

parameters for the TTL ICs. Conversations with Honeywell at the beginning

of this program yielded the information that both manufacturers were having

difficulty meeting the switching time specifications for these parts, and

this had a major effect on device yields. Certainly, as discussed in

Paragraph 3-b, Section IV, Volume 1, storage time is an expected corre-

lation parameter for primary photocurrent, and the fact that measurements

affected by storage time were of limited success in this program is prob-

ably due to the stringent selection imposed by the vendors in meeting the

Honeywell specifications. For more loosely processed parts, theory and

the results for the TI inverter support the inclusion of switching times

in any hardness assurance program.

12

(5) TI Buffer

The TI buffer was the only digital circuit which used pri-

mary photocurrent compensation. The LINAC test results revealed that

this compensation was not completely effective. For a 100 mV transient

output response, the uncompensated Motorola buffer was actually harder.

However, for higher transient voltage shifts, the compensation worked

reasonably well; the dependence of output voltage on dose rate was much

less sharp for the TI buffer than for any of the other TTL circuits. For

this reason, the responses at a fixed rate could be used to coupare dif-

ferent units of this circuit.

The range of responses for the TI buffer was only a factor

of 2, and because of the limited range and the compensation, it was dif-

ficult to find good correlation parameters for this circuit. Nevertheless,

for higher output responses, a single correlation parameter, V , was

found which was reasonably successful. Figure 72 shows a histogram of

the output responses, with the high and low offset voltage units elimi-

nated from the distribution. This parameter, which is expected to cor-

relate with lifetime, was successful in eliminating the worst units from

the already narrow histogram. Considering the fact that this is a com-

pensated circuit, the results are quite promising. The rank correlation

factors of several parameters of interest are summarized in Table 62.

(6) TI A-O-I Gate

The TI A-O-I gate was selected because it contains two

phase splitter transistors which are connected in parallel. The increased %

photocurrent in the common collector resistor of the phase splitter tran-

sistors makes these circuits more sensitive to ionizing radiation than

Rank correlation coefficients for these devices are sum-

marized in Table 63. There were no parameters with high rank correlation

v

the simple gates, which have only a single phase splitter transistor.

No abnormal V values were observed for any of the A-O-I gates.

I 1

coefficients for this device type. However, the results of an MLR run,

shown in Table 64, show a relatively low rms error of 16%, and a maximum

ervor of only 45% when 5 parameters are used to obtain MLR coefficients.

13

-faA^b^

As with all the MLR runs for the'ICs, the first half of the devices were

used to obtain the MLR coefficients and the radiation response of its

entire population was predicted using these coefficients. The rms and

maximum errors apply to the prediction of all units.

(7) Motorola Inverter

The Motorola inverter population also had two devices with

abnormal V behavior, and the radiation thresholds of these units were OH

also abnormally low. One of these devices had extremely high leakage

current, and although it passed the vendor's electrical tests, would not

have passed an additional i00% test by the user because the I . measure-

ment was "out of spec". However, even though the other device met all

specifications, curve tracer tests revealed that the cause of the high

V value for this unit was that R , the resistor from the base of the OH j output transistor, was open! This greatly increased the sensitivity of

the device to transient radiation. A more detailed discussion of the

cause and significance of abnormal V values, along with suggestions for OH

eliminating this problem with additional electrical tests, is included

in Paragraph 5-b of this Section.

Rank correlation coefficients between several electrical

parameters of the Motorola inverter and the ionization threshold are

listed in Table 65. Once the two abnormal devices were identified, the

range of the ionization response data was reduced an order of magnitude.

The results of two MLR computer runs are shown in Table 66. The first

run includes all devices; the second run was generated with the two ab-

normal devices removed. The first MLR run had an rms error in excess of

100%, and the maximum error is even higher. However, when the two de-

vices were removed, large reductions in the rms and maximum errors were

obtained. This example illustrates the significant effect of a small

number of abnormal devices or bad data points on the results of an MLR

calcula'-ion. High quality data must be used in order to obtain reasonable

results with multiple linear regression.

14

■ ' - ""- tfMMi

(8) Motorola Buffer

Results for the Motorola buffer were better for the 0-state

than the 1-stace. Unfortunately, the 1-state is the more sensitive of

the two states. Rank correlation coefficients of various electrical para-

meters vs. ionization response are shown in Table 67. However, an ex-

amination of the 1-state data, presented in Figure 73, shows that the

data is already well truncated on the lower, more radiation sensitive

side. Except for one device, the minimum values are only a factor of 2

below the mean value of the distribution. With this type of histogram,

low rank correlations are expected because of the large peak on the low

side. Slight errors or "noise" in either the radiation or electrical data

will cause large rank differences in such a skewed distribution. When

this is considered, the resui-- for the Motorola buffer are net unreason-

able. Regression technicues were also applied for the Motorola buffer,

and these results are presented in Table 58. Again, the 0-state results

are better than the results for the 1-state, but the wider range of the

1-state data at the high end would be expected to affect the MLR results

for this case.

b. The Effect of Resistor Reliability on Radiation Response

(1) Open Resistor Problem

One extremely soft device was found iu each group of in-

verters from the two manufacturers. The only signifxeant difference in

the electrical behavior of these two devices was a slightly higher value

of Vn„. Additional measurements, which were made on these devices with

a curve tracer, proved conclusively that the high V readings were caused OH

by an open R„ resistor from the base of the output transistor to ground

(see Figure 74). This also explains the extreme sensitivity of these

particular devices to ionizing radiation.

After identifying this problem, it is disquieting to realize

that these parts still meet all electrical specifications, in spite of

the open resistor. Similar devices could easily be installed in "radiation-

hard" systems, and would lower the transient failure level of such systems

by more than one order of magnitude. In the following section, we will

15

i I il toi» i ^rmmäm

discuss the means of detecting this problem and implementing tighter

electrical specifications to screen this fault at the vendor or at an

incoming inspection level.

(2) Method of Detecting Open Base-Emitter Resistors

The V readings for the two faulty devices were approxi-

mately 300 mV higher than the average value of 2.7V which occurs when

V = 4.5V and V = 0.8V for the inverters. There are several factors L/L< i-iN

which can cause V to be high, but the distribution of V readings

for normal devices is tightly grouped around 2.7V with a total spread of

less than 100 mV. High V readings can be caused by the following

conditions' (1) high leakaje in the output transistor, (2) an open R, ,

(3) an open R-, or (4) a short between base and emitter of Q. or Q, (see

Figure 74 for a circuit schematic). Condition (3), R„ open, can be veri-

field by examining the V-I characteristics of the 1-state output as the

input voltage is changed from 0.4 to 1.2 volts. In normal devices, the

output voltage will change approximately 300 mV as the input voltage is

changed from 0.4 to 0.8V. This is caused by the phase splitter transistor

turning on, because of the inverted saturation of 0.. In order for the

phase splitter to turn on uith 0.8V applied to the input, R must be

present. If R is open, the phase splitter and the output transistor will

turn on simultaneously at input voltages above IV but there will be no

change in 1-state output voltage until the input level exceeds IV. The

output V-I characteristics of a normal device and a device with R, open

are shown in Figure 74.

Another resistor which could be open, and still allow the

circuit to pass electrical specifications is R,. This would greatly in-

crease the sensitivity of the circuit in the 0-state, because Q, effec-

tively has an open base if R, is open. This particular problem was not

observed on any of the devices in this program, but is nevertheless an

important consideration given the incidence of open R~ resistors en-

countered in these devices. The other resistors (R.. , R9 and R,) are all

necessary for proper circuit operation, ~nd the circuit would not pass

electrical or functional tests unless these resistors are connected

properly within the circuit.

16

■*($■.->-"\,'V ^-W *■.'.>»*■***

mmä idri^

(3) Electrical Screening Methods

One screening method for the presence of R is simply to

put a maximum as well as a minimum value on V . The maximum value could On

be assigned as 2.85V with V\„ = 4.5V and V_„ = 0.8V, which would eliminate CC IN

this particular problem for devices with the pull-down resistor connected

to ground. The limit would have to be adjusted upwards for devices which

use a pull-down resistor across the base-emitter junction of Q, , which is

typical of ail the TI devices except the inverter. For circuits with

different resistor values, such as the buffers, slightly different values

of V may be required. OH

For complex MSI circuit« . it may be impossible to verify

proper resistor connections for points buried within the circuit from

measurements with the external leads. There are two approaches to this

problem. One is to improve resistor reliability so that resistor contact

failures will have a very low probability of occurring. Based on our re-

sults, the present technology has not solved this problem. An alternative

approach is to measure critical internal V values with extra pads during

the wafer probing, or the resistor could be measured directly. This

approach assumes that the resistor reliability problem is not affected

by the die bond, packaging and burn-in. This is probably a poor assumption,

because of the high temperatures which occur during the die bonding and

electrical stress caused by bum-in procedures.

(4) Conclusions and Recommendations

We have discovered circuits with open internal resistors

which are much more sensitive to ionizing radiation than normal circuits.

Users of hardened circuits shoula be aware that this problem does not

occur with sufficient regularity to be discovered with the typical sample

testing that is done to evaluate hardened parts, but does occur often.

enough to seriously effect system hardness. Techniques for screening

such devices with tighter V_„ measurement limits have been suggested, Uh

and it is recommended that these methods be applied to eliminate these

bad devices. Additional effort should also be made to solve the problem

at the manufacturing l,?vel.

17

-«««■ -*- ' * m

c. Other Integrated Circuits

This section presents the ionization results of the TI word

switch, Motorola sense amplifier and Fairchild pA744 operational amplifier.

The ionization test results are summarized in Table 69, which shows the

mean value and range of the ionization response data for each of these

devices.

(1) TI Word Switch

The primary photocurrent and the radiation threshold for a

secondary photocurrent of 200 mA were used as radiation response criteria

for the word switch. The ionization response data were very tightly

grouped, with a ratio of maximum to minimum value of about 1.5 for both

the I s and the response thresholds. Considering the estimated relative PP

inaccuracy of the dosimetry, which could be as high as 5%, correlation

factors would be expected to be somewhat worse for the word switch than

for devices with higher spreads in radiation data. A table of rank cor-

relation coefficients for I and the I threshold is shown in Table 70. PP sp

Although the distributions of I and I were narrow, the PP sp

increased flexibility of measurements with the word switch resulted in

good screening parameters for I . Figure 75 shows a histogram of the sp initial data, and also shows the improvement after using h and the ex-

ternal base-emitter resistance value as screening parameters. Storage

time also worked as a screening parameter for this circuit. Since the

initial histogram was already very tight, with a steep lower side, it is

encouraging that the application of these relatively simple parameters

(see Paragraph 3-b, Section V, Volume 1 for a theoretical discussion)

further narrows the distribution.

A multiple linear regression run was also made !.o see how

well the word switch turn-on threshold could be oreuicted. The results,

shown in Table 71, have an rms prediction error of 3.9% with a maximum

prediction error of 11.5%. These were the best MLR result« obtained for

any of the integrated circuits, and the success of the rt;.. , ts for the

word switch is probably due to the increased flexibility of measurements

which are possible when the output transistor is accessible.

18

■-w- I '-"'- • A.

(2) Motorola Sense Amplifier

Radiation testing of the Motorola sense amplifier was com-

plicated by the wide variations in offset voltage between units, which

was further aggravated by the fact that the offset voltage specification

was relaxed in order to obtain the devices on a reasonable schedule. The

first stage of a plated-wire sense amp always operates in a linear, non-

saturating mode, and the offset voltage problem war resolved by operating

the devices at a fixed output voltage of 1.5 volts. This is the center

of the operating characteristics of the typical logic gate which would be

driven by the sense amp. The radiation failure criterion was that a 3 mV

differential input, signal from the pulse generator would no longer drive

the output beyond ehe TTL noise margin limits of 0.4 or 2.4 volts during

the radiation pulse. The 1.5 volt operating level was achieved by using

an operational amplifier to provide the necessary input biasing voltage.

The op amp was located in the data room in order to eliminate its radia-

tion response, and a large frequency compensation capacitor was used to

prevent the op amp from responding to the radiation transient. A diagram

of this experimental arrangement is shown in Figure 76.

The sense amplifier is a complex circuit, and based on the

range of offset voltages encountered for the 4-input channels, the inter-

nal transistors were not as well matched as the transistors in modern

junction-isolated circuits. The internal mismatches in V are important BE

in establishing bias currents in the various stages, and the mismatches

in V make it unlikely that external measurements will correlate with BL

transient radiation behavior. The types of measurements which can be

made are also severely restricted by the limited number of pins and fixed

resistor values. An examination of the LINAC data showed that, although

there was often a correlation between the data for different channels on

the same device, there were also cases where large differences occurred

between the channels on the same device. Rank correlations of several

electrical parameters with the ionization response of the sense amplifiers

are shown in Table 72. All of these correlation coefficients are very

low, and are an indication of the difficulty of establishing good cor-

relation parameters for this device. The MLR approach was also unsuc-

cessful, as can be seen from the results listed in Table 73.

19

(3) Fairchild Operational Amplifier - viA744

As mentioned previously in (Paragraph 2-b, Section IV,

Volume 1), the Fairchild uA744 is a radiation hardened, dielectrically

isolated operational amplifier which is not gold doped. Therefore, this

device exhibits long radiation storage time ones, saturation occurs, and

also is more sensitive to ionizing radiation than some of the newer

hardened op amps which are gold doped. Three different ionization response

measurements were made on this circuit, (1) the output voltage response

at 4.3 x 10 rad(Si)/s. (2) the radiation level required to just saturate o

the circuit, and (3) the radiation recovery (storage) time at 9 x 10

rad(Si)/s. For these tests, the circuit was connected as an inverting

voltage amplifier with a gain of 10. Power supply voltages were ±12 volts.

The rank correlations of several electrical parameters with

the radiation responses are summarized in Table 74. As expected, the

electrical saturation recovery time was one of the best correlation para-

meters (see Paragraph 3-b, Section V, Volume 1). However, because of the

complex behavior of this device, no single parameter worked very well in

screening the more sensitive devices. For the high transient response at

the edge of this saturation threshold, the electrical saturation recover/

was a good correlation parameter. A multiple linear regression run was

made to compare this approach with the single parameter approach. These

results are listed in Table 75, and are not particularly good. The five

units with the highest low-level response were not screened by either the

single parameter screen or the MLR approach.

These results indicate that complex linear circuits will be

difficult to handle when measurements are restricted to those possible

with external leads. It would be interesting to examine the feasibility

of breakout transistor measurements as screening parameters for the ioniza-

tion response of linear circuits. This would certainly increase the

flexibility of measurements, although both pnp and npn transistor types

would have to be made available.

20

d. Summary and Conclusions

The most significant result of the ionization response study on

ICs was the discovery that open internal resistors occurred for devices

from two manufacturers, which increased the radiation sensitivity of

these devices by approximately one order of magnitude. Electrical screen-

ing procedures were developed which will screen such devices with tighter

limits on one electrical parameter for TTL small scale integrated (SSI)

circuits.

The normal ionizing rate response data of most of the devices

were tightly grouped. Because of this narrow range of data, the uncer-

tainty in dosimetry and the small errors in electrical measurements tended

to obscure fundamenLal correlations. The electrical measurement problem

was further complicated by the fact that the ICs were procured to a rigid

set of specifications which screened out devices with extreme electrical

behavior and narrowed the distribution of electrical parameters. This

magnified the effect of the errors and resolution limits in the electrical

measurements.

In spite of the narrow distributions, correlation and screening

parameters were found which further truncated the already narrow distri-

bution. In general, the results were best for circuits which allowed a

reasonable inference of internal transistor parameters from external

measurements. For very complex circuits, with limited access to internal

transistor parameters, such as the sense amp and op amp, results were

considerably worse.

6. MTBF RESULTS FOR PARTS SUBJECTED TO IONIZING RATE TESTS

Although the general applicability of a gamma-rate screen was not a

part of this program, it was felt that MTBF testing of parts subjected

to high dose rate environments would yield results of immediate appli-

cability to some military systems. To this end a group of inverters and

buffers were submitted to life testing after exposure to the ionizing

rate tests. Neither catastrophic nor drift failures were observed in

the exposed group of 49 inverters although the control group (99 parts)

showed three failures, 2 catastrophic and 1 drift. For the buffers, 2

drift and 1 catastrophic failure were observed out of the exposed group of

21

1

48, although the control group of 94'parts showed no failures of either

kind. Converting the catastrophic failure numbers to failure rates at a

60% confidence level results in Table 76.

22

:y

• 1 fck AM^riMM mma

O vO z CM

o

< x.

CO

•a ca n

00 o

in

O CJ

CO «-.? r-« —. r- -T

o »p r-i O J->

^ o o

sC 00 J, o oo

r. u-1 «Z-! en ■H C>

o 0 u-i u-> ir> T "~l o o SS r-i

o o o rj r- -T X> -j

c. X> -I 1*1 „-1 C O o o f 1 ': ' -.

-J f. u"i r-i <r o c o o c .

-. rl -' tfi a o o o o o a o O

OOOOOO OO

o o X1 o c1 -t * —■

f*w o O » o w r-i O —« '-4 OOOOOOOO^OO

c a> M M 3 O O 4-> o

& tfl e

•H M P*

o

o o o o o o o :

..- r o o r .T~ N B T) Ji ^ 5* O" 0* CP ■

u o 4-1 CO

■H 33

tO vf iD —4 (

**1 -T O fi '

3 O O X >■£> ü o O O O C

o o c o o r* ;.• c o o —

2 a.

23

24

CM CM (M CM 55 CM

U O

5? O ti!

E n CO u

53

o

m CO

n l j O

p- «j -^ co r- i p* sT -i O ^ t

•-tc*»irtu"»p*',",("",>o I O ■* *•» O O «"* ^ ""* t

O O1» iflijl »vtiflMNOV'J I

'-J © 000^,^,o^'H-' I 0A'»OO-H-lOPÔ'-,r--,a3—' I

.H O O —tQO-*0-<-«0-*-<-^ I

i 000OOO"Q0-*O -'-«-N P-m»-<»T'-<OO'--'DP"1r>00,^-tn I P)M 81 «1 PI -7 n O iT, ^ i Q, « H o rt | O'-« o o■ *r> g o o ö o. ,o o g -* -♦ «■*« t

H o> H O) ri " M PI 'i I« >£■ o ii s o o> H I ©o —t o o o o o o a o o o o --* o —' , t OiH^plNiÔ'iriPv^nn^« o nfin | OOOOÔCOOQ:3 00000-*0-+ I ft i/i P« vO >1 n •> ■• .() ^ * ^ O ^ ^ M in Cfl B 00 I O O "■ **' '■* O "* O C —* O P J ^ -I —' -^ CTi "■* 1 ""> i OOOOOOOOOOO OOOOOOOO <H I I o

I I ! I

ft Ci —i -A .-J - I - t -J H H H r< >-l H H H >H r-l H >4 rl -H rt H >-( H H H r4 H H H -4 M"H H H; <-l H | OOOOOOOO OOOOOOOO OOOOOOOOOQOOOOOOOOOOOOO 1 l I T I i i i.;ii T i i. T i l T i i l i i i t t t T T I i I i i i 1 i t i i I I

bl fl ti) t'- U! b' b4 tU W W W [il :■■' .J U J) U ffl b. L-3 u) b: tu bJ bJ W *JJ U) j! u] u.; U- UJ U4 bJ uJ W tj] u { lA ui -< CO "'1 •-] <7, u-, fi O f- ^t —< 00 u-\ -i (^ î !1 o r^ -.t -H er) in PJ (7^ o ^ 'j P- •' H ti i/i n 7> Oi Pi I r"'00-l'IPIl*,'TiO''1'CNN*^(J'0-<Wfl^'J-1iCiii^l'00<0\0H'-llN^f1-3in 1

WChHHrHH-<r^ HH-(rH^HHrlrlNt'iriPJf4(S(SNNriPIN^N^P^nfl(nnnP1 I

a c OOO0OOOOO I

OOOOOOOO PJp^.-(«-<*Hu->iAvT»*P'»u"i-T»i>Xini©ma0inoOOOOOOOOGOO •

OOOOOOOO-» -»rj<*>( O^WPlm * t> O O ' I O O O O O I O

o

8 ä

s?

5?

-J r-i T> "( (— -1 r> <t Wl J~\ Ju T o •ti ^

O o ■ 1 O CO f\ rr> -J

CO Si u o M a cc

o o r- r] ■ i ro T- in "t

m r> in ■-1 ■P n —' ri

O-l M in O

Ö o

rt rs 00 m ii ■T to -T -t -< O —i O '■!> o O u -■ '-J o

< O CT\ CN 53 CM

M O

<n

CO

^"s •H

v^ 13 ta n

o\ o rH

X

O *

4J « 0}

c

u 3 CJ o o

—< O —t o o ■—. o r o - • n —. <J f*> ?■ rt 'T O <T>

-t o •-' o c <*) i i r-> < 1 f-> o O O ■P ..-, 1)

-r> -•< ,c M O ( i -4 —■ >J3 .C lA ■ 3

O O O o o •-> O o <—> o o o o o o o o o

I (N M I'Jfl'l NM M IN NM tN riMriMTJCN r» 'I fl -J H i-l ^1 ^rt-t'J-tH JWH-» —' rH —' -Hj O O O O Q O O O OOOOOOOOOO Q O O O O © O O O O O O O ■ O O O O OOOOI

1 I 1 I I I ' I I I I I t I 1 t I 1 ' I I I I 1 I i I I I I i I I I 1 I t I I j l'l l'- Li J1 t: W W t; U U) W W U] Ul li] CO lu uJ JJ U7 US ul :>J U IÜ u: W „U UJ t-1 [xl LI WPU1 Ltl Id UJ Ul I r. T f I is ^ ■ i M m r.| c^ -r> <*» O r- -J —< uO . -i el IT. -X. —' m ty. ri .-; -, .t r- •-< »A </; >--i vD O r^ f- — .7 I fi s I-J -7 to 'i if o^ M >o c -1 w H iî o> n * o ^ f~ o o o H -t N 'i N * f" " -l -t -t «n «A -X- >£i

PM

4-1 o

2 00 o

-1 -4 vt IA LA vß «o >o T~* r 3- tf> (J. —' •

OOOOOOOOOOCOiAfi-5 HHrn-t t,i*ya^'X>fAOcT1'~>—tr-i^'ü'.COOOO oooo

O O CJ O C O O O O O '*** CM' f| Oi 00 rt <& -1 ''~C^ÔcO^DÂ"''-^rt-HOOOOOOC, OOOO

.OOOOO^-T-Oi tHHOOOO OOOO

3 00

25

•* ,< — f<', »» I

« ♦ I < <N I > (VI I O it O II

• IM I > C ;l •j I I Cui

• * » t- • I I/» IM I

I

2 UJ I •a f- • UJ "^ f X O I

♦ I -" I'

I I

I <^l I O I

^ 1 I

C IM

° !

9-

r» r- fr tf\ •* rf\ <c o s Ift ■» •*

1ft -"• o Ki\ !**■ *f

>J- »A 4-4

m <M >r

D ~« m i*- n m *M <>j

*•» r^ *-

>J O p <■

lS

rv

O I

IM .0 0- o i N v»- O tr *— | ' ; i

•-I p-< o t-< j 1 j fr- •J? *t *o u^ ' ;

,-n O 0 _! , t «-• -4 r*4 o ^: 1

1 1

n C> -i •c J- i : ; -0 o V* I.TN 1 1 —4 •■* o O i 1 *> -O 1- m v-4 17- i i * IM 0> *J ir> r-t ■ j

3 ~ o o i_> —4 ' i » r m in vt a JLJ 1 i i r T--» l/> .f *j ^-4 i

^ --• "5 o CJ -■4 1 NJ <n r- ^1 r*- f- j r- c n ■t ■f r-4 1 b o o a CJ r-4 1

; I p -1 -J <<\ o> J) f- 1 t f r* C* fi r" .--* cc 1 I o o o o O * 1 o |

f> .» i ü» m » O J r* -J <*i ~J- .o u"\ ^1 IM -n ^-4 rr X ; •-H r-" 3 O "5 -1 "3 r>4 '*> o IV !f» 3 O D n -4 o • T *- ,^- /J .rt o fi ^ •* -* IM -*> in X> r. .-- X t- ! »-« ^ J O 3 O _J c c> o o 0 o « \J o >n VI in —i >c •o 3 40 ■n "•'! ■4* O O ;

< <1 o -. J 1*1 —4 4-4 rv J -'* l^ 1- f- en f- i 4 -4 —J •-4 y O 'J o O O :J c c; o CJ O

t- D -J ". X -T ., ■ \j _^ —* -^ -- ," ■ .' N .-< ( J —4 "> *\| -4 *-. r-.i ^i ■r\ <c\ *- T. 4" r- »-. .•'■ N-

-4 J ~> '- o 'J o C3 o 3 O c 'J. w- "3 .> O ifi n *> * o r~ in 3- ?'< J* -J ,-« rj •0 o (\J f ,^ r', **. •*) C5 -4 n —4 r> -5 M r> ri ^t —4 ^ •N. ■n o .T o r- n o J> o

o <^ to z CM

o

^* «

o O i-l

X

o vO

c

M Jl O o 4-1

o .G

ET « B •H

(X,

>4-|

o

% u 60 0 4J W

•i-l SB

I I I I I

i- I X 7T. I at OJ 1 Q. J I

«M (\j fNj ^ r\i f\J (M rs! r\j PSJ f\i r^ r^ <vj (\j (\j »-I ^< ^« --< ^-4 —4 -^ rH -■ i »■ < . , --• —« *-4 ^^ .4 .-4 .-. r-: —• ~ -1 ~* 0 o o o o o o r> O O O O O O o O O C> O O O O O "i c> o o o o 3 O O C O ") o - > "5 o

1 I I I I I I I I I I I I I I I I I I I I I I I I ' I I > I I I I I I I I I ' UJ UJ uj lii UJ u1 'i-l J« ui LiJ ji ui 'U UJ Ui ui u' '-'J LLI a' LiJ u.l UJ MJ ui A! ..'. a> u; Lil üi UJ at JJ IU J u u ^J ?M (.'^ if- '4 T >r "* ^ ro c> -r <".. .7 -^ % ^" r: : 'vt- "•■■ '" vi ^ o ^' -- -g a. %* i.r- in *-' x" ' i *" ^ > * ^5 rf\ «-4 r*- rg .0 j . ^ H o ^i r- 'i T ^ '- C' --4 ~-t ,\> ^j :--i *j tu- .n r .0 ^ 1^- a" 0*' ^ ' r> ,--, -. ^.

••• H N N PI l« >J idftvUNMStOff t <-4 »-4 * 4 ■ (\. \| '^

O <J O O O O O O O PI O1 H -4 O tr. -t a*\ -4 i, X ^4 u-'! P4-i c

C C? D '.1 a O O C O r-4 r-4 .\ (>', ^ *\.i -t t| ^ Vi T'l lw f\J *- .-,

^ "i (ft < *o o o ■

-■4 »\, -4 n o o 3

tN m a) 3 00

•H Ü4

O O O T O O O —1 O fs4 (n u-i L^ o "j fi o ^ 0' -0 in -t *\t -« "M st -4 N TM fij -4

Ai ,4 -l O •> "3

26

( I I I II I I I I

M I — I*- I • ao I t • I < 4 I > ~> II O il O I

1 • **> ll

> O I - ' I O •- O in ii i- • i tn N- i

i »

»M I O 1 I I

2 OJ i < US I UJ f- I r o i

> i

IM * O I

2 I I < -J t -. O I a o i lii o I

i i ,i i i i i

o I - CO F^ 1 •P tr -f 1 * r>j ft 1 £ J tf *NJ /-* I »/• *. o- a 1 o C rC f. i •* «r. c\i rn i to m O f^ 4 t> . o o o ^ ■X "j 1

• r ** -r f^ rv* **> ! ■ ► 0* m G ^f r> ^' ) o c- in m a fvi •~* ■r f*. «■, m rsi f* ** ! ir- r- ^ ^- CO rNi *f *

1 «s r- <r •4D r- H r-< sO n CO rr\ •^■, fj ri •* (^ 1 ♦3 J5 a- •-* r- i^ h" r- »\ *■- r»t ^ h- «r>4 Q m

j o r. : PÜ rî sO j 1^ r n-' f. j. ■i~ J- X C- si>

■ e,' ■s r^ m f*-; (-•J fM •" «~. •-t

i f\j 4> «rt o M t\i r* ir ^4 o (^ vl* I (SI +« "■î 'g ■J -i> X f o m J> sC

n« «i en rr, CJ •'■^ r. si- .■*! m .-4

wA (?\ —t a %r 1—4 -n cc .•»■■ O -^ i;

O f - •'^ D .^« .£ »o -n a. si U u"i m fvl f^ m iî *"vj <">* - n (<» .•n «-4

a C sO a.» r- a. C <» •C ^ r- s."

1 0 f u O o j> r- x *r\ !>•> «M li-^ ! (N, M ra r*'. Wi <>i .->*■ r* o ^ ci -H 1 f\ I f -0 *M o rü ^^ ^>- *"j *-t sC U1

\ C- I T\ o- O r> ■n .-i «\ ^-< ^J r j I

j (V CM *M "1 î "J ■J f r^' ■*"■. en —t I -* m r^ >Ti î -a 4 ^C J r^» r—

1 c «\ u^ OJ -» <r r- *> X. —t -^ c

IN *J. ^Ni ~vj f.J n. —* ■« ~t IW, ■^ •-4 ! -t p f <^t *-* "J ^ o o cr —4 ^-4

r- us •O r- 'X' J- ■Ü c r* O 0" ■n i (N ( \1 r>j •■\j '•4 fj -•i *M _< •VJ —• -^ I o- *• o r**, ti («■ -f- T .— ^' *-- ■—

o tf •;J O r ■o s^ r: 'C ^-i t" r1'.' IN V ^ lJ *■ i r-j —* i J .-4 Ni —4 «—■

i -*fc >i .r» 0 •Ö <V k ■J' •* -.. >s-

if ^. Jv *- -* *i »J U"- r* J^ ._. IN iki «*sj f^j r-i . si "4 . r-t -4 1 --*

i -r I sO '■ J ■r« -J- <"J 3 00 —- .? u*. ,- >•♦ * si .n «-. •—< o • j"l ■A *; 1-4 ">

i -s; IN (Ni M ^, '■; •—I « -' . < —t .—i -' -V X r-i r^ r»i f*'> s. J 'C O 4; r>. ,."

i •<S tv\ o". ro .r i7 T- rfl ■£ sj -". .-( »-I

w » \J <NJ \i f\i ^-4 o • -^ —- IM —• _4 -\ U 1*- u" CJ r- -I -i. " V ^. ■v -4 «. IN . J f-1 .-» <r X» «^ /I ^t 3 Z- f- ^. i (SJ (N f\i 'V tM »-* t? " .4 - i —* O — 1

! rr i. sj c I— r- r> r> / -- '7 r* '\J <"* 1 J ■n O ^-, .*■ r~- .^ c O .■; " J

(SI \i \j f--J i\J .4 O -* r- O --* O —-* j -<« ̂ '7 a- i.) U^ "^ —* vT r ■• —4 iTi c

IN 1 M "J o -n -sT *»■ . rg •-- '^ v: ^•* r* ( J l\J f\J ■-sj r-» o - --4 J '~J .j —4

US a. - 1

IT. >- >sl -*-, i , is f' j- —4 '*"> O

a- -i « .-< a ni ^ n ' h ■—« O r** .T _■

e-' iM >Sl r\j '^ fV, ^^ c ■» -H ' ) r'.* O —4

m ns ifi C _> r- r-. r- ,) C.'1 c C\J i) ^T

o O O o ^> f- ^ .n c f"l l-'l ■.* iT. T,

i- to rs: rg r-i Cvj 'si o n> ^-< 'J '5 O .J r- *-< H •s? OJ o si/ a.' f»j wr-< " 1 TsJ --4 r-

r4 ■o O C' o f^- > r ,-~i c l\, !/» IT '-* »•> (N (N •J "M r-* ^4 o ^ O • -< o O O 0 «N |T> .n on cr ft! a-' »-J f\ n '■J '1 O J->

fsl ■O i> 0- o O n 'D ro in O H sj sf •0

f-J <N —* ~* -H î «~* :J ~> o ^-* '"J ;) T) :i tf r> -J ISJ ''j -^ .o -> ^- o- -o T: 7"j -~t c M J o- c- ^T« 00 U"1 0 NJ 0- ^,' —4 sT >t •0 Aj ■'S! .-4 - -4 o Ü j—* -'. ' > ' '1 *~1 r.1

"■ '3 Jj 0"' "H .-1 (NJ «1 *.) ro T. ^: ,f r> - -. ^ ^sj

T T -c c- -r 'a" »i .7« r.i • vt *": ■0 -M .»■

r-i w~* ,-t —4 O O —i ■3 o * 1 o ■ 7 ' 1 O 0

i s

-< O ^1 (SI

^ 1 -H o cr -vi ir If -I I«- 1 il

*N (si IN (N (v) rj -J ('j 's] fsj -J g I'J IS' >J r- 0 a o o o o o o o o -"> noo o ~

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 u.i './j m m uj i.u uj jj L;I in u. IIJ u.1 11. a. ui i* in u' aj u t.' n .i t p. i - .. r? ir - • _ . 1 f- -, *- •. - - n .]

'"J vf tn ;- oD o —t -s .A so -1 7- .-< -si + _n ^- -r.' "S «^ r

\J -g N (sj fM ,-vj r-j M -g ~j v f. 3 O "3 O '; 'l o O "3 n i 3 I I I I I1 I I I I I I

\i IN (\, rsj ".' f., -M -M ~.j 1 o ':> o o o • j -> • -s -5

1 I ' I 1 1 1 1 I

—« " -r o r-

(sl CN <N (N .-n 1^ m -t ^f .f j J * m fli r ;r. us ur. o

O O O .J O O t- -3 M3 I

O O O O C) '.' : J ! . O O -•

J^J■'C4JSOJ•00'', (\-H.-I-C .1 • ,\i ..• 1

.(1 j1 0 -o (^- rj> ir Lei -i» LTV a i T'I -\i -J "' .-1 z

;■: ' J \i r, ;M .v, o fM

^ o o '' > .: -J >'J ^

rj n n « -j ^ o :> T „'> r'i H H :s ^j -£'

OS O

z CM

U o

<4-l

CO s^ ■v a)

C OJ u u 3 CJ o u o x:

CO B •rl u Pi

M 60 0 AJ to

•H

iro iri

HI l-i 3 60

•H

27

< in o «N Z {M

S

z «M

H O

«4-1

4J c e u 3 O o 4J o IX

J^ n «9 6

•rt H

IX

'i

u c

c <u

0) V) o o

u-i

4) M 3 00

(V«) INlHHfDOlOM iWMISd

28

Jk

(v«) iNiwanooioHi AHVWIHJ

29

I -f IT

I

C± >- * c r w »V & I

o u-i s-*

-id CO

4-1

a to 3 *»» 0) ^~. c

•H 01 10 > ^

j

•a en « K ^-N

•H CO CO o \/ t-H •a

Ctf X h O co • o m .H

H X <:

fl ■t • a m P. i—J

M a * a.

p § e «0 o ed BC •H PJ Q vO

o> >j l-i vO pc (U z ■dj ■w ts p*l 4-1

S td SJ o o cu CO «4-1

vO IT,

01 M 3 M

(s) 3 '3HII 33TO0IS TVDIHI33T3

30

Jk

.. g

-I T

3 i 

, o

—: s

(s) si 'INVISNOD awn aovnois ivomoaia

O r-^

CO to

CO

OO

P. P.

H

s O

PL.

PI

U-I

O

60

0)

u CO

in

<U U 0 oo •rl b<

31

1

x *

» u oo ■5- t» . O O . _0 . iy?n . o © . C r- < r err t o o . a A - UN» O O - ■4 •* M X- *- - oo-

O O ~« r ^ « f- r* «n o o -* o -o o F*- .Q »** OO ^ «0 •* »- * O O ^ 3D J. £?• ■piAO oo -■ ■e -^ * ,n i/-, t» OO o

CD «T1 -* <"t sj C if i a» u £> o Q m Ifl o •-* V if. lA t) o O o o >n * a: o f # f <r o o o o

flD *, -4 f» O O O O

*- i, .f o o o o o

w o o o **■ OJ —< CT" i/S <M ■* -J o o o O K. 5 * o »« Si rn * o o o o ■$ $■ r\( rf\ >1 ^ K) ^ o o o o i«-. qo 'J O

.^ -- "N4 r

O t» O < rf> # ^ I -* *4<>i t o c> O ( •n («j X - o o-

> o -*

i (Ni O O ö o

U oj Uf o o o *■* « o OftiO

iA -'S O O O tf* o — .-# O O o rt c -* o o o o o o oo o

UJUi Ui o o O O -j o O rf% o

* -« tr OO -* o -■ -* oo o — o -• o o o o o o o oo

U LU iJ Oo o Or* O O tf\ O

oo o «■• o »•* O o o o o o oo o

Ui oi Ui o o o o o o o ff- o

to

Ü

> a) Q

00 c

o

1^ o o 00 < H u o

W

f\i O ?

a.

rj (•* m tst «\j r*j o o o O O O o

a» LU li f» O »M

U' Ui u

■ôcooooo; i i

J uu U' I i -^ a« *

JOOOÔOOOO^ I I I 1 I I I t I I

ui a' UJ <-' O) U.» U.' UJ U' u ^ M tr K c, * N o ofi o

O o o i T •

uu U' ui

o o

c o

OO OO

0009

o o <

o o <

t r- a-

I

i O t*

1 ^^ ,«« ry r\j r»j r»* ^ f*i f»j (\' ru <M

> o ^

f» O f

X O r

OO (

O O (

»OM-r-oOOCJOOOC-'Ot

) O O O O O O O O O © Q O <

000006» i i

\ii05TO^»^(l)» o<ttc*o-*»r»oooo<ô

n OS* 0 0 -*

UJ o 0 0 or t4k~r4 < 000

000 -*• 000 »/• Ui 0 000

*« O (A > Ui U. 111 UJ

OOO .-> O *M O

0 O «N O •jr • • •

-* -« ■-«

■»

O -j -*0 -A -J O -0 + 0 OO^ u. OOO

•a n h

01 1^ 4-1

O c »H 0)

h X n

3 O Ü • O m 4J

0 4J £ «d PH

O- » ft u M 3

0

O 13 e

ä h

0

5 60 O

1-1 J5 (0 4J

1-1 •H Ä »

00 if)

9) U 0 00

£ S8S fa

32

itf ^

a I

5Z

a « act

(V) I •lN3HUnOOIOHc[ «JVWiad

33

r-^ *J o o

00 <

**" 3 H

*> 0) u o M o

V «4-t r\j >

r^ •» ~* o CO UJ a *<. O r. /-% 4, •^ * •t C/j

["! V*' CO o "0

~'e ■H cd 4J OS uu O at

% *" H f- • U o o i pa G

i) t-l

1 -z V X c

! "HT o o

■ *.s u •

1 -H nn is ti 4J

a a

a) a

z CO

«s «9 W u

i go IM C SI : °5 n h

1 "H u 1 "*P- e 3 1 O § U i ° P O I 01) «J

2 OT $ «

cd •H

o St

P PM

U 5* 1 >o 0) M i "" U cd 1 u, u <3 ! 2 cd •H

u M i * w p-l

• ON in

t) s u 1 3

00

fn 1

rita_4 ▲Jft ■MM

♦ 1 :

♦ ' !

1 ! r k

| j

i! ; ■

I I I

of 1 i '■

. O 1 . 1

u.

i j

* 1 -

i

I

1

1

;

'

t

' 1 1

1

i 1

1

K S ■ j i ! UJ ©

+ 1 ' ' i

o j i ; «f • ' j ! i | ;

* - |

i 1 i

4 ! i 1 j

it | i 4

i ^ ; V i | O 1 •

OP f\* ' ! 1 i *-4

-■ 1 1 r^i ih

1 ! 1 J i

! j —t

• © * j i 1 f* */» © 1 1

"* i i ] | I

1 «4 2 •i. I i 1 I I ( i ^ J» t j ! 1 1 t "Ü

1 t 4T

| i 1 -> i 1 1

1 l **

■ * i [ 1 1 i "" 1 i ' i 1 | 1 - I J j 1 ! ' wM

1 «. : 1 , ; —1 u> * 1

1 •N -*

y * i [ i •* j i 1 .-« (St

1 1 [ f\ *1 M 4 O • i 1 f\t

> © 1

i 1 ! o jh

♦ I

1 ! j .

rg

1 , i \ ** ;o i ! 1 * cC • X * i 1- 1

i 1 i t

1 »A i <M K *-* * ^ O

OD i i 1 i 1 01

IM ♦ ; 1 ! i ! *■

U. 1 j i i i O - I f 1 i ) !

r ♦ i »- 1 1 i 1 o • 1 1 l i ; -J* i 1 t i tt. I E ! i i !

* i 1 i i t

j i ' ; j i 1

i 1 !

o 1 o o i I | § i O % o

o O i U <

O ! o . ; ! v ! O s o © < > o O 1 «0 ■ 1 <M ö an -0 f ; fs» ; • [ • • • I • • «

(V -4 ! " " * ■4 o ° ' 1 > o

1 =

o o 00

* u o

«4-1

'S u

o H

X o

a. ft

go 3 to u

«

o

§

n V 4J 4J « u

o vO

V (J 3 60

I» •

(v) i 'iNauanooioHd Äwwm

34

o

I I I I

'I — I *f I ~ «f 'l • f«J i

a. • !i

> -o I O I U |!

> O || UJ I o -u i

U o

O c\. I

i/llf l

i» i O l

i* «3 *n I UJ r'l 1 3" - 1

• 1 <X> ,1

1 1 1 1

r- i O 1

Z i < M 1 t—i C 1 3 !\ | Jl X 1 x: • 1

-0 1 1 1 1 1 1 1 1 it

y- ►- I I I I I I

i- I ct 7 I .11 ULI I a o I

I I I

• I O I

«* «r

6- c* 1 i t- «■ 1 j

0 i j « r*^ o r-4 i

r~ f ./\ i t 1NJ t ^. * ; 1 m f -? ! i o t-4 s • i •s\ p o CM '?' o« F fi n O i a o r-4 --) •-< , «» *' >r o Ci o- i

a & '^ .J- C\i O ,

o .-• •-< <■* --. ,

ro >J- r. >-<"l »^ O •-• , a? tf «\J vT fi <M 0" i

o p wA r-t —8 -H o I *^ fi CT I"! •u r. r-. CD If -.» fO .vi —* X O C -4 •-4 •-H f— Cj , r- 0 x> ft o h A j ! r- r-4 m '_ > .-* 'JO o P ■~4 »t -* —* O m r» r- aj r-. vN 0 I^I

r- fi 0s '."> O & r*w £*»

o o o —< «-< g O •H <\l «-« •n O sO m CTJ »—■ T^

r- #c r-- o en a: j~* «SJ 'si '\j

o o o o o O o .-4 -* ,H o o r> n 'M >0 n ** X 0 a> .0 ri r- T) to J: «i ^J tr .-( r\j

o c o O o o o r-4 o ^-i »-* 0 *r <XJ O :o -( 0 1 i— f- *c ^n f\J -Q Lf\ f- *J f+-\ w-t o o x\ O O ~> o o ^"5 'O .-) o r~-i c «J o L* c .-* f\l '^ M ■i\ "". nsJ J". c \JC\ ^J J rj j- ^ <■ r> ~* .." -J J"' \f1

o O O O u —. c: *.j o o -j o o .-• y- ri , 4 ■7 C*' \ f* *- * -■-, . t l'1

**-■■

o ^ «■■4 • H , , r>j O v? *-^ ^« r<> -A N- .n r ' r-A

w o o •"■ ~- o *--i ^ c !> ^J ', ~- "' : O J3 01 c- r- ■*o ^ r\. T; '•J r- .r« •*> LTI it) "',4. ^ N O

C -i r> o -> •\! ■^ ■^ —* —• rvt •VI r* »T «-.' -■< *NJ T "

-- -n h- K f^- r- »>. t- ^ r^- -d o P*- K r- r^ P*" ^- l

oooor^ooocooooooooo

ui a) JJ i_i u' UJ u1 u! ii." .u in .J uj fj JJ a. uu '.■; ' -J O :N cr o r^ -,■, v* r *~ .~ -4 -*> -- n _*> r- r,. tVj ^-f ^ w'\ -3 vj O c1 i' j T r '- <:■ 1 *"> ;'■ '- '

ii -t1 "t~ .1', -0 ■" :<" n; ■ o o o -■> o :5 r- o

T: -rj rr T- ,-;: i- ^1 '0 er cr O C O O Q O C* O O O

-H »•) cr -T) r- o r\ m vj —4 ,-■ -1 ,~,

(JJ u .:: u; 11 ^ ■ 0 ;- '-1 '■* c i-

-1 ^' "*- jr. r- ■*?

—t r> ,-U vM ' ■j n -n «A

o o ' ■ ^> o o o o o o o ro m: eo •> o n o

o o c a3ocoaoo-iJ^'v«t-K I X c

it«, h- r- '"> r^

i -J o o -j •:

obooooooooo.-j«tmo>o-< -«NHH-«

N N ß o i« ^ ir\ o (v .i I cj o o c; ■-» o o o

P. m

G 0

es 5 00

■rl U H O

§§ M 00 00 <i O H *J (0 <

• H «O

1) Vi

tu

VJ

5 i

ii i

.1

5 ;! • ■£> I

of • 'I < •»• II > fi i C |l

° i!

J >o « UJ '| Q "J l

O- I • -* l

Q O ,1 t- • I i/l <M |

l I I

K5 'I i i

2" c; I < f- I U- fsi t n "o i

• i 1A |

I

O I 2 ,1 4 Ui I — => I u f- I

SE • I ■o ;i

K I Of Z I UJ UJ I O. U I

"3 7.

a

OOOOOOOOOOOOoCJOOOoOOOOSOOoO^oOOOOOOOOOO

UJ UJ U_ III UJ ii.' "I UJ UJ Ul III 111 UJ l:| Ul UJ Ul UJ L'J Ul UJ UJ UJ UJ UJ U U' flu 'JJ UJ JJ UJ UJ UJ U-l UJ UJ O.I Ul ii o p- r<- in r*- o r-; r\i JD 9' CM ir. '• <v m *r> *M ir\ on —i ,j- -r , a -)■ f- .-• ? r- ^ T *I r- •-« .r — t T* ,v* -I f* M ^ f> N £ ir- ^ »■■ H 0 O J ^ -»»^ M ^t O tf\ ^ (^ -o \ M1 .- ~* »~i (V| *j o** -A ri *t

N H H ai 4 CM ^ (O •>-* t~* <M f\j <*iCO m *t *t m m «0 JO «0 P"" r- 0^- .V CT* U* (7> ~* •"' ■-I »-1 ~* *•* -I <-< •■< .-I I I I I I

O o O o O O O O o o O .<<( O O o o m m m nv r- o r- o O n o■ > otioonoooooO^.O'txi'jHHinNi^.Tji: .r *\;

1 o -j o o o o c o n> o a ■ o D o p o a o o c o a

O000OonOOOOiHC»Hiv«iMMKlrtO»«»»»l,l''JO'536aoO353O

o

o§ M 4J »^ «

n C 3 0 00 1 -H

6'S 9 O H O

H « O ca

«n «

«f - «1 5

J-l ■O O f-l J2 O OT

u

o «W «si o P2

§ c K 0 CO M O « *i -rl m tH

■H O 35 OT

• VO

«1 U 3 to

36

o s 3 06

O

>

«S 4) U

<2

.-I O x:

CO

8

c 2 H

3 E

g $

K C K

a s

e a

§S

3 o

§ M M « n u 4)

y

en

**;

[8/(TB)PBJ] aiVH aioRsamu. Ho-tram

37

n

m e

8 is

a g

S G **';.; »,1. :* 4

3 3 ♦-+-■

a o

w

H CA

W -

8 * S w

§ 1

o s

o

2 o ■ 8

? S 3 H

2

8 « Q

U

O

8 a

5J

[•/(T«)f»*l aiva aioHSMHi No-romi

38

-f- --

I o

JG B

& £

fl ß 3

< H

.-1 3

1 «

M

u

> < •s.

I >

8 V»»

> e « *<

«H o

J3

O o

8g M * s?g

< as •H

M H V O W OT y at M

U o OT <H

• m vO

V M 3 60

[8/(TS)/P«i] aiVH Q10HS3HHX No-Nanx

39

■S <

5 8 ft S

' O

3 =

O o

2S 4. * - u .

NJ -- . - , -

^-*

.J..-

a o

i M

H

S3

§ § i o

g o

«I

I O

a «

5 I

fi 9 H

3 C I

is I! SS • O M w «i

a e

5J

s s e

[B/(fB)p«j)ajya aiOHsaam NO-NHM

40

1 10 — a

10 DOSE RATE, f [RAD(Si)/SEC]

Figure 67. Superlinearity of Primary Photocurrent, Ipp as a Function of Dose Rate for the Dual JFET

41

■4*d *+m+ämtih «Mi

- Y

a I

-I

o*

+ *- — «• - ♦ — -• V — - * - • ♦ — •

-

- « - * —

:

-"-♦••

-

! ! !

1 i

i

j

[ + • -

.. rg

i ,

j !

i

! i

M j j ' »

!

1

-4

0

~*

! 1 ! !

j ■ i ; i '

i ! 1

j *0 ft I

! | i a

w

S X

;

.

!

to «•1

1 i

j 1 <

■

1

w

1

| :! i i ' !

1 1

i !

a 0

I 1

i

!

1 l

j ! i t

[ |

1

i

j i

1 ~+ i

1 ! i

' >

i I i

i i

i i

i j

I i •-* «H • ■ ' i

1 } i |

1

i ; ;

•**

—* 1

- r -

i ! 1

j {

{ •

! 1

4 1 1 i

1

I

! i

i : i

i -

r-«

i 1 i

1 1

i

* | i

i

! 1 1 1

r* j

1 , - ♦ _«. - . fl -. ..

6 e i-l cd

& ai X

o «t-l

p. p.

la r\

W CO o CO

Crt ü

o u "• w - < (A B - O a)

*-> HI

§ •H

0)

o O

00

0)

3 00 •rl I*

(v) ddi

42

—h J i - m

—» 3

- I

_* « c

IS t © I

3 O

<u x:

u o

U-l

CO

I Or

o

! ow

i 8 T ~ &

w

ID CQ

O •H

0) f-i H o u

(V) 6/(IS) P«J ,01 x L IV INaHMOOIOHd AHVWIM

43

*M^k4rifti^ MHH

o

> f* o

O X. (0

u

^.1 r\j

01 (0 c o a (0

m o

K © j- ■ w n\ -l • ~i © o - ^ v o < -* * ^ •

w* IA■ rt **■ «■> - -« t* O * -i r ■M rH O O O < O ® « «J * '. n x i/-. «r t r -* O O O o '

* m \r -o ~« —« •*> - o •-• o o o o o '

• — r.i o ~ ■a- c —• y- r- ■> o* ■o P-- o t ■i <_■ O u -*.-« « HI r 7 o N ) m U PI ^ ^< « ■«"* t> o O J , .J f\ J N „,

v ^ "1 N il d' j 4 r\ -;->

T f, c\i m r>

> rf\ f\l -J r ■ r~ j O Ci o o ;> J J H\ 'O « l\

o ■rl 4-i

CO ■H <~» T3 01

J

* O J3 i; JE» c * iT

c o 01 ex

.C 0)

O 41 4J

H «0 5S 4-i

60 I O tH 4J 0) U •H 0) 33 4J

u 0) 01

rH > a c g M

• ■ I- -, -1 ov i

H ft| Hi Kl * ^ UMf J Jl (^ M)' i i* o

C' O o O O O O O p* f- -« —* «J 1' n s* vl tf ■ "' 'I' 'r> ' J ^ J- 4 -<f- J f nN rjOOO^OO

c> <"> o C* O O C O O O »j 's) r. J> ii i/\ 3 I- -1 r- .1 j, 4 •-' rvi ' ' O O O O O OOf''.'l)

O

0)

3

•rH

lo-ij«

44

ttt ' i A^

o o X o

60 c •H s o X. CO

m X) tH o .c w 0)

ITl u • f—i x. t-l H

CO 0) 53 C H O M W (X. CO M X

s H 4J •r-l

i5 o o ^ c H a o H CO -rl <; u w N 00 cfl

in M 0 0 z S5 o o M •H P

BC H

OJ

CN 60

O •H tu

siiNii AU naadiN

45

L l«MM

0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2x10'

lONIZATION RESPONSE THRESHOLD [RAD(.Si)/SEC]

Figure 72. Truncation of TI Buffer Radiation Responses with Offset Voltage

46

• - "Ihn i MMM

I • -

*«» n CT x, ■

t —« O O •-* O - < O o ' * «i 4: >M <" O N ?■ ■/"' 1 o *-. r-- > j oj r ' ■.

f\ JJ N »- .f, 0 'Owe ' O . '.> O I

»N 1; ,0 (."• .• ~« f d- j- ■ ■ , '-■ r«J I HCI'JOO'J'j:1",'. ,'I:'J 1

u-t LO »-< r~ j- CJ in .-. x *- -N — — -* " *A . ! I o rsj r- o **"' w —■ J " <■-. -.. 1 .j --, o> 1

,4 ~« -> r o "> "' O -■-.;' r r -1 o , 1 O PJ. IT O "" T. '•■' -t u .• r< .-. - j .) «J ri »• r» j •*■ O f>J **- -J- —< -' '• ■' "i - J -- -' "J ■*" ■.' ZJ ■*• *f O •-•.-* T> <"> o '< ^ <""• ° ">'"-""> ~"> O C o n ^ ■ |

nj a K r> r- *r -t v ■'. f> r v 1. r er -r> v T ** T ~ ? 7 ? 1* c P■ (T- •*- «■ C 0- ff ft- ft- f OOOO^O ■-> -> ' * O >j '. J ■" r- •*"' O '.' ."" <"• O O r, <-. O O O O . ■ ■".* O I" O O O

i.11 LJ ;i. : J !i" u u .J u ui .. ■ '■ ii L,I Vi ■■ 1 ■ j IJJ J ; a- ,f ■,: a' 1., a: .*( 1. ■ a,' u 1 u >:J U1

f .1 / '5jv'f f^ * ' ,1 rj - > "i" T1 **" if ."* *" '.T —i T~ -i": n *) f o- -Vj **■ .' ^ r-i %: -K. *" ... .«• ■■ "> n -* "; "> r j -i *" r - r* »- • *.. J -J -J ■■. ■ f •. ./> ,. ■• T, V C T> -■* —• >.

r^. -> 0 -T >- «-* >'• p\, f ^ X' ^> c r- x a y . • -1 -. - 1 -• •■ < .■ . < •—* ~* •■• »- *-t i'-j rg ra % >j t t t 1

i c> kj .• --'- o r> rv -j J n 1 „..-,..- _, ^ -f. 3 -. j ■- ► ^ c-1 o r- -j o c o o o o :* o

■1 O :.i n O Ci r, r> ~. *- -* - j - . .: ri 7 T, ^ A f *i ftj. (T* r. rv ., (_;% ,4 *• o o -_■> o CD O r-'

O -■* "■' O O •*! C> O r> '.»

C a f/S Of

I

a)

3

O u

w

p

•o ■-I o

J= Iß (U Si

CO

o a U5 (U

Pi

c o

1-1

« •H •a c«

p.:

0

o «J CO

•H X

r--

0)

47

•»»»*» -■» «3k ^^

IHPUT OUTPUT

a) Inverter Schematic Diagram

200

100- _VT„ = 1.2V

3 0

-100

-200-

IN \

J I J L

,VIN = 0.4V

VIN = °'1V

J I I L 0 12 3 4

V(VOLTS) b) Output V-I Characteristics of Normal Device

< a.

'00 r

100 - -VIN = 1.2V

0 - /

-VIN = 0.4, 100 0.8V

200 " - i i I j 1 .. 1 , 1 i i i

0 12 3 4 V(VOLTS)

c) Output V-I Characteristics of Device with Open R^

Figure 74. Method Used to Detect Open Internal Resistors in the Inverter Circuits

48

+ litk

20

15

w o M > w p

c OS w

g 10 •z

%

s^ S

^ I

0.7

^ M I I tSl s

1 ffl !_

0.8 0.9

n

1.0 1.1 1.2

DOSE RATE FOR I = 200 ma lio' rad(Si)/s] sp

Figure 75. Truncation of the Word Switch Secondary Pho'tocurrent Thresholds with h„ and r_

49

i 11 yiw i

sn; 5fl<

»47n

12K f VW—i

VELINE DRIVER

TO OSCILLOSCOPE

125ft CABLE

i2on

1.5V

Figure 76.Experimental Method Used in LINAC Tests

of the Motorola Sense Amplifier

50

i f - afr i

Table 51 Summary of Low Power Transistors Tested

for Transient lonization Effects

Mfr. Device Type (a)

Construction Gold-Doped

Base Area

(cm")

Approx. No. Parts Tested

F3C 2N696 NPN PE No 3.43 x 10"3 150

FSC 2N2222 NPN PE No 8.47 x 10"4 150

FSC 2N2905A PNP No 9.35 x 10~4 150

FSC 2N709 NPN PE Yes 3.65 x 10"5 417 |

FSC 2N3960 NPN PE No 3.65 x 10"5 150 j

MOT MT 7111 Hex In v. 30 T

NPN PE IC Breakout (DI)

Yes 4.6 x 10~5 I

30

MOT MT 7113 Buffer BOT

NPN PE IC Breakout (DI)

Yes 3.03 x 10~4 30

TI TI711) Hex Inv BOT

NPN PE 3 IC Breakouts (DI)

Yes — _~ 87

TI 117113 Bufier BOT

NPN PE 3 IC Breakouts

Yes 87

(DI)

(a) PE - Planar Epi (b) DI = Dielectric

taxial Isolation

51

• ' ' fcflb Mai

IM lO

co H

tn u o

u H

u CD

§ EU

o J

o

4J co Q

4J C CD rJ M 3 CJ o u o

-C

o

& co

1—1

CO o ^) ■o*^ r» r*» 00 00 00 00 tH DN ON rH ON ON

CD *"> 1 1 1 1 1 1 1 1 1 I i i N !* o o o o o o o o o o o a

rmali

ra

d(S

X X

rH rH

X X

rH tH

X X

rH

X 1

rH r-t

X X 1

rH

X 1

rH rH

X X 1

r^ CM a\ co 00 O m 1 <* rH 1 1

1 r>- in i

o" -* NO vO 10 oo m co <r r^ vo o o zl tH rH rH rH rH tH vo CM rH CO rH rH

c ü CO 00 ON CM CM -o vO CM \D co o ■K O O ON CO O ON

00 00 r-. CM O a« O CM ~» m <r m in ON <t <r C» CO >tf ■~~.

X tH rH rH CM CM CM CM CM ON CM r^ O CO in rH CN rH CN CO co CM 2

4J c VO m -* •^- vo m ON r-> vO r-- -* CM NO O 01 • • CO • CO ON M co co o ON -* NO LO O O rH NO m ON in CO O O co o r* CD rH vO ■tf CM H CO rH \0 <t CO CO -H -* vO CN r-, tH <r CM 3 Ö0/-N o o Sl 1 1 1 1 1 1 1 1 1 1 1 1 1 i 1 ! 1 1 !

*J OtJ v^ ON CO 00 r^ ON * m ^r o o • o> CS| • o o • m co • rH o <• vO • ON . • CO

J2 <T rH CM in ON • rH CO 00 o r^ CM r-- CO CM vO ON ON O PL, CM rH r- vO CN rH CO rH CO 00 CN CO rH co CO CM rH ON r»-

> 0) p

■O /-N co oo vO CM ON <* NO m in r-- ^o m ^ <f rH O CM in

CO H CN O r» 00 <J- o CO vO 00 i^ NO 00 CO CO CM ON CM CM ^t TJ **-/ rH NO co CM tH 00 NO CM >£> CO rH rH

c tH (0 u

CO

ST rH f- ■* CM 00 in r~ r^ CNJ vO t"» ON O

c <-> ■H • o 00 cd < 00 vD CO O CM vO <r o es o CM vO CM Cfi vo r~ co r- ON

rH 00 in in CM vO CN ON CN m ON CO o o CN O rH VO <r rH CM rH

f—i o o o o o o O O O o o o CD u 00 00 00 ON rH -i a\ o> ON tH r-i tH t-H rH rH rH rH rH tH n »~~ o o o o o o o o o o o o o O o o o o o 3 0> TH rH tH tH rH tH rH rH tH rH tH rH rH rH rH rH rH rH tH rH ai u w O cd "—' X X X X X X XXX X X X X X X X XXX

a «i "ü X <0 CM CO co O vT -T CO O O O o m o o CO CO co <t <r

W M rH in in CO rH rH rH NO VO 00 00 rH MD vD r-» r~ H vO *

6 <D V. CO -H c a a a a a a a. D. Q. a CX ft ft 0, ft ft ft ft ft ft o «0 a a« a. a. co ft ft CO ft tn a ft oi ft en ft ft !/) ft-C M M H M M M M H M rH M M M M M M (H M tH

CO u

£*

01

•rl ft

H o H

CN * O o > S ß

O r-V

> >> vO CM o ON NO "^ H N-^ CD

01 H a> CM ON o ON Q VO CN CM r- CO H X fc! tJ a 25 a Z 2 O 0) O 3

CN CN CM CM CM n re PQ M

in X

•a o CO CO 0) u u 0)

T3

0) O C nj *-' w

•H CO D rJ

CO u CO C CO M *J

en co X o 3 in

en 4J ü a)

i<-i <4H a)

P^ (J CO

TJ c O o a) en

0 60 u a

•H (1) u 3 en

13 01 4J

ft ftTJ

M 01 u

^ cfl en eu 3 ft O 01

rH u cfl e u 0 cu

5 iw CO

52

^M. i i iL

C cd *J en C O U 4->

c 0) 0) E ^

•H rJ H 3

o a) o 60 u ct o t-> £ O PH AJ t/> >%

n T) «

S5 M

0) PU ro E m •H r<

H O S) Ü-I

rH 0) x> 6C 0) cd cd s-i H u a)

O 4-t 4-1 U

in c_>

c/)

< <. E E o o CO rH

ti i;

c/>

< <! : 6 E I O CN

II II

'S)

M M

3 0) U) t-> /—- O "3 ■H D. Cd c/3 X *~' w -o

cd M

a) o a)

•H p. > ►. 0) H a

00

m o

o-> o

\0 -J >0 N CO o oo oo m so m -J- cn

vt 0\ SO CN <t m

oo

CN cn r-

O r^

•<r m

m in

in m ON CT> <T CM <r 00 m cn ^-1 CM CN in m

en <t SO so <r CN

o in

cn 00 CN

fl <t o * m m m \o rH H o m

o CN

1 | o <r CN

II ii

IS> U pa

11 o <r rH

II II

C/5 CJ PQ

m CN

in so «a cn

cn m

m

o o 00 00 CJ\ rH Ol CJ\ rH O O O O O O O

V. •A X X X X X X

CN cn O <r cn o m o

rH m cn H rH SO rH sD

o o r-i r-\

> o o -t rH rH O rH X X

X m -T

co rH •O

sO

sO S3 CN

CN CN CN CN % CM

m O CM ?■. CN

O sO CTv

CN

H O ,--* !>« S-l

CTi ^^ 0) o '•H f - H UH

v; O 3 CN .u ru

53

Mte ^«tai

Table 54

Summary of Rank Correlation Coefficients for Various Screening Parameters versus I for Low-Power Transistors

PP

Device Type

Dose Rate [rad(Si)/s]

Rank Correlation Coefficient of I PP

vs:

C0B s BVCB0 hFEI rBI CTB

2N696 1.25 x 108 .013 -.099 .426 .176 .105 -.282

5.3 x 108 .079 -.123 .44 5 .248 .111 -.35'.

2N2222 3.0 x 109 -.334 .320 .317 -.066 .156 .413

1.35 x 1010 -.296 .184 .189 -.2 37 -.036 .501

2N2905A 1.3 x \0J .498 -.246 .249 . 501 .182 .013

6.0 x 109 .538 -.312 .168 .511 .317 -.082

2N3960 1.5 x 10iU -.305 .566 .478 .000 -.585 .038

6.0 x J0i0 -. 232 .299 .354 -.152 -.370 . 144

MT7113 1.3 x 1010 .180 .133 .295 .096 -.100 .284

6.4 x 10L° .133 -.075 .214 .180 .168 .076

Table 55

Relative Efficacies of Various Screening Parameters for Primary Photocurrents

PS

Device Type

2N969 2N2222 2N2905A 2N709 2N3960 MT7 111 MT7113

(1) TS CIB SE C0B C0B tSE

(2) SE V'cE(SAT) rs RV

CBO Ts CIB

rs

(3) BV CBO C0B VCE(SAT) rEI S L

BVCE0

(4) EBO C0B : Fi,i hFE

(5) VCE(SAT) hFEl

(6) C IB

v-

54

tfMto A. A

Table 56 of I fo

PP Low-Power Transistors Using Total Sample

Summary of MLR Predictions of I for 2N696, 2N2905A and 2N2222 PP

Device Type

Dose Rate [rad(Si)/s]

No. Screening Parameters

Prediction Errors (%)

Mean RMS Max Min

2"t96 1.25 x 108 3 1.2 11.5 33.3 -24.8

5.3 x 108 3 :.i 10.6 28.3 -25.9

1.25 x 10 12 1.2 10. A 36.8 -21.5

- - , ~8 5.J >: 10 12 0.7 9.C 31.7 -17. 2

2N2905A 1.3 x 10' 3 2.2 ] 4. 4 49.3 -34.1

6 , :o9 3 2.9 18.3 4 5.6 - 50. 6

1.3 x I09 12 l.J 11.2 45.6 -19.8

9 6 x 10 12 — 14.4 37.2 -28,7

2N2222 "* x 109 3 — 17.3 56.5 -26.8

1.35 x 1010 j — 17,7 52.8 -22.2

9 3 x 10 12 — 15.8 53.5 -22.8

1.35 xlO10 12 — 16.2 52.6 -23.6

Vv

1

in

_o m H

o

0) 4J c

a) o u c o

u u o u r* G 3 Pi

M-l o >. M

I t/1

/-N /-N •*\ y-N

M M M M /~\ V4 E W

tu h-i W

N— /

u Ä JS ^ J= X N^X N^ N.-^ N-^

<J o LT| rH OS ON CM

NO CM -* 1 cn 1 -* 00 -* NO 1 CM 1 m

* ' 1 • 1

o ^^ H < • tn ON ON o NO NO

(A »^ m n CM 1 <t 1 ON > 00 r^ NO ' CM

1 i

P. > 1 1

rH1«

14-1 in CM en CM r^ o

4-i

-* CM r^> 1 rH 1 cn 00 r-> r^ 1 CM i m

tn • • • 1 • ' • 4-1 C a>

•r-i CO NO CNI 00 m Ü on rH P> CO 1 00 1 CM

•H H CO r^ r^. 1 CM ! •* IW t * • 1 1 »4-1

0 u V? CM m •* 00

■>» en CM 1 r^ I m c PQ r-~ CM in 1 CM 1 m 0 4J • • • 1 1

•H 1 l i 4-i

rH 01 cn 00 en in •* U o r-. NO in I ON ON u m <r m o 1 rH I m o S5 » • • 1 i • u i 1 1

r*s c

o vO 00 rH oo ftf w NO CM m 1 00 1 ON

P4 CO -* Csl ! CN i <r JC ■ " • 1 i

o> <r m 00 rH pa rH m NJD 1 00 ! ° VJ NO m r^ 1 rr\

1 1 rH 1

<t 00 m 10 in

G. <t m O ° I m r-~ CO CO ) cn 1 rH

M1^ 1

o r—* r-, o o ai « ON 00 o r-l rH rJ ^~ o o rH o o 3 • i VJ ^o m r-l NO NO

< rH cn a» m CM O rH r-l O 01 o o CM ON Nf> rH rH

•H PL cr> en CM O ON r>» r-- > >s CM NO CM r^- en 4) H S3 V, 25 Z ?-' H H a r-l CM CM CM CM » '-4 ^i

56

,**-

H

n a

c at V-i

u 3 y o 0 s,

E •H rJ

PH

U 00 c in >w

a» en

JO a CO 0) H.H

•H <J-J

4-, D C

C_>

c o

ttJ r4 r4 o u

CO

CO

> o CM II

co CO

c

to co o to

to c

J"

u co iH in c r4 II II

0» G CM -<t O vß « N N 00 H X) CN si r^ <r m r^

rO i-l 00 rH O o\ CTi r^ CM CM r^ o CM r-~ O -» CM VJ3 co r-- <j 1/1 CC U1 CM -3- r^ <r <r r» N l/\ lC co on r-~

CM 0\ CM CM cr. CM m m a> o\ N OO VC r^ co co cr. CM \£> 01 rH <* lO CO CM <r ~- CM in r^ CM <r r-. CM s* ps

oo <r H nn h r-~ to CM co oo r- CM m r~- iH co t -

co <j <r o m H ro CM <t co oi co 01 r-, <j. cr» r~- o.

O oo vc cr. m co rH vO <t <T CM <f en f- a-, en oo 01

CO rH rH C'. 00 CO O co m m <t CM 0» r-^ O". co f-s cr.

i-i co -<t n iii '£ in in *<t co cr» r-% 0> r^ Oi X f-~ oo

C CO i-i CO CO CM -JHinO MC0 en oo os oo oc co

rH <■ O rH OI CO co m >n r-» vo f>> en r-~ oi r-~ r~~ oo

o be

a

<

CO iH <N o U II II u

iH iri vO iH sj- O LO CO <f CM cr. iH O, ~- en cr> !-■ 0i

O C cr. co co oo CO CH <J rH CM rH cr. r~~ cr. oi ro o\

m in <t CM c-4 m ol to UHO * c ui r - c. oo r- Oi

(q)änM3 H H

a ca c a <■■) o H H

H H cJ >') C C " O

H iH a «i o a °Q c

H Hi

(^^ < <; < p; pQ eq < < < K W « < < <; pq or. ec

s / H 41 / Jl 1 / r-,

•H S / -5 4-> P-l / !_j

O / cn 3/ ~ rl/ a) CJ -o fy yi 4-i co / ° ,3 ^ / a tf —

r-- r- r^ r» r-- r~- o c o o o o rH rH rH i-l rH rH

y. x x x y. x i-^ r-~ r^ r^ r-. r^

OO 00 CO CO CO CO o o o o o o rH rH rH rH rH rH

;<! X X M X X

CM CM c- CNI CM r J

rH rH rH --1 rH rH

00 00 CO 0C' CO CO o o o o o o rH r-H rH H rH rH

X X X X X X

\0 U1U3 vC * tl

ca

m c CO vi1

<f l ! rH ^U ^ CO

U) p. <r. U;

•H u 'H ,C OJ 0) u ^c rC

c: (- f; 3 b '? i. P: i-i

r-l r-l

a. o: cti nt •r( •H

rC U rl

a> CD H •j) W

3 "l-i c. cx Ul 0 o

.i": M 1-1 H 'J tj

Ifl ^

57

■o 01

nj 4-> « U/N «•HK a *o «^

(U -•o r-» r~- rH co r~- r- rH CO r^ r» i—t <4H VJ 0) O CH 01 Oi CM co iA o> CM CO m Oi CM CO m

a CM CM CM CM CM CM CM CM CM CM CM CM 0) CO -H 0£CM > G 0)

c* o t-i

ß ^ O 6>S •H ^ 00

m 4-1 CM 00 CO r^ vO -<f O rH vD <r o rH £ O M K -H o vD o o r^ \C rH 1T1 lO o\ o CO CO

T3 »-i CM CO -a- rH CM CO -d- rH rH rH rH 0> !-. u w fi-

-* o o CO u-l rH CM 00 o CO CO

CM

fe r-~ CO o vO r^. o O» «5 •* 00 CM rH CM CM CM rH c\l CJ rH rH >£. ■<t rH Vt

"w pr-

X£ \D H f». <r vC -» rH o in r-t CO \C in -tf rH o> m CM rH o rH a. m r^

o> 00 00 00 r^- 00 00 00 00 cr> 00 CTi 00 O <4H

o o

4J a ß CO & 01 o CJ <J -o T3 •o XI

 H

oi in 00 00 00 00 00 00 00 00 00 00 00 00 *-> -~- o o o O o o O o o o O o « ,-. rH rH rH rH rH rH rH rH rH rH rH ,—!

PH -H to X X X X X X X X X X X X

oi v—■ 01 X) m \D m vo m «JO in so U1 VD LPl VO o nj CM CM CM c-' CM CM a M ■ ■ • . . •

rH t-i rH rH rH rH

P. •H

u < < M m < < PQ pq < < m PH

4J ß 4) 01 rH

•V ,£> ß ctj PH PH (X p< PH PH PH &4 Pi PH PH PH 0) "H PH PH PH pn PH CM fX PH PH PH PH PH fl ^ M H H M M M rH M M H H H 0) HI n >

XI 01 4-1 u

>H X) 0) (H

o. 0) u 01 3

CO CM

u 01 £ 4-1 o 0) /-^ X CO 4-1 4-1

'4-1 XI 0

0) a u

H s 4-1

CJ ■H r*S

4-1 CJ cd p.

l X! it o

S CJ H

B II U3 o 0! lH PH

U 114 rH •H > X) »V

01 X)

01 4J 1 ^-N

C3 > r^ rH m o CM 3 CM

CJ II s^ rH C/J

CS (K W •H CJ M C'l to ^~* s^

ro 3 "^.

25 W rH XJ tl 0) « M *< 4J K M W cd H 4-1 H M 01 0) 0) 0) ß M r4 rH

0) 0) 01 01 M S > > ai tn to tn u 01 0) 0) 0) rH rH rH ^ JO 43 XI

cd cd cd to •H •H ■H u r4 rH M ß cd Cd cd 01 > > >

•H • CJ 01 4J 4-1 4-1

•H CD <H c c IM 0' y 0) 0) HH ,.C ,-j XI X3

01 .-' ; "* C ß :"^ (U 01 01 o - 0 P, a. a

01 0) 01 Pd -H "O X) XJ

Sä 3 H ß ß

58

ro O

o 4-1

NO c "J

tu •H

.-! !0 Si r.

CO rt H VJ

H

H H

1+4

C

> l-i

CO Fi n: 1

c/;

C

sa r! <^ r-- ^-1 r^ r^ m • • • - - - •

t> ---. cn rn <r CN CM r-i H O-l o m ON C^J C3 <0 X i I NO ? n

oä ra ? 52

c •Q 0

S-t -n ON ON a\ a\ CX) ON o O. ON c» Et; i.) O O O > ^> o O o o o O C

TJ <fl t-l i-H .—i ON Z* 30 t-i i-l rH ^-4 rH •—i r-4

G -H X X X <~f —4 rn X J< X X X X X CO > C*' r j a. o .—i O o CN] m cn o ON o

4-1 CÜ

c/i a CM CM o ,M "J- CN r-i m

' "

ON a-. ON ON ON O o ON ON ON ON r^

■w 0) O O o i-H —i O ,-J> O o O W 3 r~4 i—; r-l £> > > X X rH rH rH rH rH W ,H x x X CN 0 1 CN <f X X X X X 0 Ö r-. 00 zy "1 t—J -<t- in in m -1 CN in r^

• • • •

O

r-i <! ° —1 O o rH rH ^ ~~ 0Ü

O -H ON ON <j\ <J\ rH ON CN

4! o o O o o O o o ON CO rj 3 —1 r—i T~4 f-i rH rH rH r-\ O o eg ,.J X X

1, > J> > X X X X >'. rH rH

1) r. cn O rv. <r '-^ r^. CN") m X X '". >• ^ o CN CM r» '--) c-J > rH 0\ r-l s~* >—\ r^

*—^ E e X o $4 ND 00 NO

CO CN rH ro ^^ o a ■ X G C rt

C o O —4 cn 1 •H -H • H o I-H CN C, P4 P-,

•H 1—' v_^ 1 1 CU ^^ N_^ v-^

•W cn a X) "0 T) 0) a i~ 13 -a -a "O T3

•r-1 I—1 rH iH in ir. o T) *TD rH rH rH rH rH t-l 0 O 0 £ G ex rH ,-4 O o o O C IJ 43 X! rC 0 0 'I) O 0 -C rG rC J= rC cn CO cn « o. a "i rC JG cn cn cn in cn 3) U ru 0) Ü4 cn Bi U1 tn D a) 0) -o TD -H r—' H CO 0) )) (1) QJ 41 M-l CU 0) a) CU a) Q 4J .»-.) JJ ■H •H •H > > 4J 4.J 4-1 iJ u

CO m CO «4-4 U-4 13 e e CO (0 cfl CO CO 4-1 i^- 4-1 •H •H 0 4J 4-1 4-> 4-1 4.J 1/1

1 1 1 "0 O O

e o c

o o

'Si C/2 1 1

C/l 1

oo

1—i Ci O S s £J m NO rH o c o

0) 4-1

■1J V-. CO CO u CO U 0) U rH V rH iH

—I 1-1 S-i r-4 o 4-1 0 CU > u 01 H M IJ )J !H U-4

0) M CU M u-i 1 o 01 o U^

a H > C

i—i

H M-l

BO

o 4

4-i

0 c 4-1

0 3

03

cn 13

co a

CO •H U

a cn

G o u

rG Ü

■H

3

•H 3 V u

i) rC

o 4-1

cn

cu

cu r-l

T3 CU

■H

tJ o

59

Table 61

Some Rank Correlations for the y Response of the TI Inverter

Correlation Factors for y Response of Various Parameters

Initial Electrical O-State 1-State Parameter (Output Low) (Output High)

V0H -0.616 -0.427

XIN(1) -0.715 -0.493

TSK -0.647 -0.520

vos -0.679 0.738

Switching Time -0.438 -0.746 (Low to High)

Active Rise Time -0.693 -0.630

0.422 0.591 WT

hFE -0.703 -0.721

Rise Time -0.591 -0.771 (Low to High)

60

Table 62

Some Rank Correlations for the y Response of the TI Buffer

Correlation Factors for v Response of Various Parameters

1-State Response ]-State Response

Initial Electrical at 3 x 109 rad(Si)/s at 3 x 109 rad(Si)/s Parameter (Side A) (Side B)

XSK 0.436 0.462

ISINK 0.709 0.521

vos -0.705 -0.502

hFE 0.510 0.405

Active Rise Time 0.639 0.440

(Output voltage)

Switching Time 0.130 0.464

(Low to High)

Switching Time -0.414 -0.174

(High to Low)

IIN(1) 0.522 0.504

VBE(Q5) ;Note(a)] 0.550

(a)Special Lead Measurement

61

Table 63

Some Rank Correlations for the Ionization Response

Threshold of the TI A-O-I Gate

Electrical Rank Correlation with TI A-O-I Parameter Gate Ionization Response Threshold

xos -.284

Icc(o) .273

vos -.092

'pD -.323

Output Capacitance -.289

Stored Charge .170

hFE <V .126

(a) R, Resistance 0

-.248

(a)R(3 is the emitter-to-base resistor of the output transistor.

62

■^-

Table 64

MLR Results for TI A-O-I Gate Transient Response Threshold

Parameters Used For Regression

Prediction Error(a) (%) RMS Maximum

(b) R, Resistance (Calculated)

Propagation Delay Time

I Correlation Factor PP

x/Ios

16 45

(a)The first 70 units were used for regression coefficients which were then used to predict the values of all 140 units.

(b)R, is the emitter-to-base resistor of the output transistor.

63

Table 65

Some Rank Correlations for Ionizinp Rate Response of the Motorola Inverter

Correlation Factors for y Response

Initial Electrical

of various parameters

O-State 1-State Parameter (Output Low) (Output High)

V0H -0.161 -0.342

IIN<1) 0.042 -0.131

XSK 0.049 -0.160

vos 0.133 0.121

Switching Time 0.084 -0.304

(Low to High)

hFE 0.083 0.218

64

Table 66

Effects of Flectrical Screens on the Repression Results for the Motorola Inverter 1-State y Threshold

Description

Parameters Used For Repression

F Value

Prediction Error^(%)

RMS Ma::imum

MLR Based on Circuit Measurements (Devices #41 & fr'47 included)

MLR Based on Circuit Measurements

(Devices #41 & #47 excluded)(b)

hFF.

V0H R,(c) 4

hFE

R/C>

4

13.2

5.2

136

27

891

145

(a)The first 70 units were used to generate the regression coefficients, The 1-state y threshold was then predicted for (a) the 2nd 70 units when Devices #41 and #4 7 were included and (b) all 138 units when the two devices were excluded.

(b)The two devices were excluded because of faulty pretest electrical characteristics:

(1) Device #47 hac excessive output leakace current (112 pA at V = 3.5V)

(2) Device #41 had an open base-emitter resistor (R )

OH

(c)R^ is the external base-emitter resistor on the pull-up transistor.

65

Table 67

Some Rank Correlations for the y Threshold of the Motorola Buffer

Correlation Factor for Y Response

Initial Electrical

of Variov ts Parameters

0- State RtVonse 1- State Response

Parameter (Output Low) (Output High)

TSK -0.695 0.040

vos 0.217 0.234

V0H 0.331 -0.173

hFE -0.708 -0.052

Switching Time -0.532 0.131 (Low to High)

Switching Time 0.178 -0.027

(High to Low)

Stored Charge 0.546 0.096

(Peak Current)

R(a) 3

-0.267 -0.081

Depletion Width plus 0.381 0.193 Diffusion Length

(a) R„ is the emitter-to-base resistor on the output transistor

66

Table 68

Regression Results for the >k>torola Buffer y Threshold

Parameters Used For F

(a) Prediction Error (/')

Description Repression Value RMS Maximum

MLR 3ased on • hFF 27 15.7 36 Circuit Measure- . Stored ments for 0 -Stat« Charge Threshold Rise Time

'lev to hi gh) Switching ti"\e (Low tc High)

. R^(C)

MLR Base;; o.i .Circuit

• hFE

• VOL

2.1 99 964

Measurements for 1 state . V

OS Switching Time (High to Low)

Threshold (Device #40 included)

. Fail Time (High to Low)

• R4

"LR Rased on Circuit Measure-

• hFF

• VOL

2.1 56 163

ments for 1-St ate Threshold

(Device /'4Q excluded) {h)

■ vos Switching Time (High to Low)

- Fall Time (High to Low)

• P4(c)

(a) The first 70 units wer? used to generate the regression c efficients, and then predictions were made on all 140 units.

(b) Device #40 was excluded to observe the effect on the r._u: assion coefficients because the measured y Threshold for Device #40 was a factor of four (4) lower than the other devices.

(c) R and R, are the external base-emitter resistors for the output and pull-up transistors, respectively.

67

Table 69

Summary of Ionising Rate Data (Non-TTL Integrated Circuits)

Device Data Description Mean Worst Value Standard Deviation

Max /Min

TI Word

Switch

Inn @ 3.5X108*

PP

q& I @ 3.3x10 PP

I Threshold sp

3.59 mA

25.1 mA

8.9xl08*

4.1 irtA

30.4 mA 8*

7.7x10°

0.19 mA

2.2 mA 7*

7,2x10

1.33

1.5

1.4

Fairchilc

UA744

Op Amo

6* Response @ 4.3x10

Sat. Threshold 8*

Sat. Time @ 9x10

0.59V 7*

1.62x10

15.6 us

1.1V 6

7.0x10

41 us

1.5V

4.8xl06*

5.8 us

9.5

8

45

MOT

Sense Amp

Threshold

Ci-1 2 Threshold

2.26xl08*

ft A 1.36x10

7* 2.7x10

7* 1.5x10

7* 7.9x10

7* 5.1x10

12

18

* Dose rates in [rad(Si)/s]

68

Table 70

Some Rank. Correlation Factors for the Icnization Response of the Word Switch

Electrical

Parameter

Rank Correlation Coefficient

I 13 3.3 x 109 rad(Si)/s PP

I Threshold sp

Storage Time

(a) 90 ft Resistor

hFE

V BE

Capacitance

I Correlation pp Factor

C0FF

.308

-.166

-.240

-.167

-.251

-.039

.499

-.701 |

-.725

-.692

-.496

.097

-.695

.4S4

(a) Base-emitter resistor of the output transistor

69

Table 71

Multiple Linear Regression Results for the I Threshold of the TI Word Switch sp

Parameter Used For Regression

Prediction Error(a' (%) F Value

RMS Maximum

Storage Time

(b) 90 Q Resistor Value

hFE

Output Capacitance

3.91 11.5 71.1

(a) The first 70 units were used to generate the regression coefficients, and the coefficients were then used to predict all 140 units.

(b) Base-emitter resistor of the output transistor.

70

Table 72

Pome Rank Correlation Coefficients for the Ionizing Kate Response of the Motorola Sense Amp

Rank Correlation Coefficient

Init:ial Electrical Response Threshold Response Threshold

Parameter Channel 1 Channel 2

Power Supply Current .055 .127

V0L .057 .141

V0H .132 -.125

vos -.141 -.177

xos -.001 -.037

BIAS -.080 . 100

A0L .129 -.209

Recovery Time --.053 -.087

Channel Select Time -.108 -.238

Output Resistance -.207 -.149

71

Table 73

An Example of MLR Predictions for the Ionizing Rate Response of the Motorola Sense Amp

Electrical Parameter

Prediction Error^ (%) F Value

RMS Maximum

Power Supply Circuit

Output Current (IV)

V0H

V vos

A0L

Recovery Time

109 676 2.2

(a) First 70 units used to generate MLR coefficients. All 140 units used for prediction.

72

X! rt H

P.

x: 4J p.

o 1-1 O st

« < •u a C a) ai

•H x o •!-• •H IH IH <4-4 O

0) G 01 U CO

ß o

en 0)

erf

ß o •H 4-1 rt N

-H c 0

c o

•H 4-1 « ai >-i )J o u

ß rt e! ai S o in

(13 •u n) a ß o

•H

4->

rt N

•H ß O

W 4J c a

■H o

•l-l

U-l a) o O

ß o

—i ■u «

S-i

o U

ß CO a!

-a r-4 o

0) S-i

c o •H 4-1 R) M

4-1

cd

■H

Tj

4-J CCj

01 w ß o D. Ul Q)

C3 'r-1

O 01 H •H 4J

CO l-i 0) •H 4-1 e 4-1 0 rt •H 0) >-i ß H rt M W PM

r~- 00 ■-I 00 CN CO CN m CN CM rs O ro o> 00 r^ r-~ rH LO m CM

m vO m r^ in m m iH o CM CN O m o m m o ^l CN CN r-\

OJ e

•H H

^ M 0) 0) > u o ß u <n <\) 4-1 Pi •H

u c rt o 0) a

■H 4.J rt 4-1 CJ u CO Pi 1-1 4-1 to 3 S 3 < 4-1 01 u ►4 0. 1-1 CO rt t-t o o C M O

C/l L0 H <: M M M

73

in

tfl H

I o

< a. v

o 0) to c o a to 0) t6

C o •rl 4J tfl N

•H c o M

0!

o

3 to

F Value

For

Regression

n CM

RMS

Prediction

Error

For

Saturation

Threshold

o

F Value

For

Regression

i-H

00

RMS

Prediction

Error Fo

r Output Response

At A.3xl06 rad(Si)/s

O

Initial

Electrical

Parameters

Used Fo

r Regression

Saturation Time

Slew Rate

Power

Supply

Curr

ent

Input

Capacitance

Offs

et Current

74

-**-

Table 76

Failure Rates ror Parts Subject to Ionizing Rate Tests

Device

(a) Catastrophic Failure Rates (Percent /.I ,000 hours at 60°' Confidence)

""\ v Croup

Type ^\^^ Control Stressed

Hex Inverter

Dual Buffer

1.1

3.3

.63

1.4

(a) Drift failures were not included in the failure rate calculations, (1) for the reasons stated in Paragraph Id, Section V, Volume I and (2) because the particular circuits would still have functioned properly in normal practical applications.

75

SECTION IV

IONIZING RADIATION TOTAL DOSE HARDNESS ASSURANCE

1. INTRODUCTION

This section discusses the results of the work carried out on the

total dose aspects of. hardness assurance.

The primary objectives of the study were first, to identify certain

surface related electrical parameters which could be used as precursors

of radiation sensitivity and hence, which could enable one to predict

the expected total dose damage. The second objective was to assess the

feasibility of the low dose screening technique. Specifically, the ob-

jective was to test the assumption that the devices exhibiting the highest

radiation sensitivities during a low dose exposure are the ones most

likely to fail at higher doses.

Paragraphs 2 and 3 present the electrical and low dose screening

results, respectively, for a group of low-power transistor. Similarly

Paragraphs 4 and 5 perform the same functions for an operational am-

plifier.

2. ELECTRICAL SCREENING - LOW-POWER TRANSISTORS

a. Approach

The following material presents the results of the effort ex-

pended in evaluating electrical measurement techniques suitable for

screening low-power transistors for total dose effects. The devices

selected as vehicle? for the study were: 2N709 (npn), 2N930 (npn) and

2N2905A (pnp) .

The technical approach to the problem of evaluating electrical

screening parameters was to determine the rank correlation coefficients

between certain promising surface related, initial parameters and radia-

tion sensitivity. The use of the rank correlation technique immediately

implies that the scope of the program was primarily geared to the predic-

tion of the relative radiation sensitivities between devices of different

types and of a given type.

76

The rationale behind the selection of the correlation parameters

for the bipolar transistors was discussed in Paragraph 4-a, Section V,

Volume 1 and will not be repeated here. The summary of the results of

the rank correlation calculations are shown in Table 77 in a sort of

generalized form which gives an overview of the total dose task. (The

specific list which would also indicate the exact bias conditions is

prohibitively long.) As discussed in Paragraph 4, Section V, Volume 1

and as is evident in the tables, correlation was sought in the majority

of cases between certain surface dependent initial parameters (or com-

binations of these) and the "radiation sensitivity" of the device. The

radiation sensitivity was thought to be adequately represented by the

absolute or relative radiation induced change in the base current. For

increased measurement sensitivity the base current was measured at low

injection levels (with both fixed V and fixed I conditions) where the BE E

surface effects are dominant. Of course, there are other quantities of

practical importance in Table 77, besides the radiation sensitivities,

for which correlations with initial parameters were sought, e.g., the

absolute and relative changes in gain.

Total dose effects are extremely variable both between device

types and between de' i<"es o:: a given type. The three types of low-power

transistors, 2N930, 2N7Ö9 and 2N2905A, had high, moderate and low radia-

tion sensitivities, respectively. This is shown in Figure 77 where the.

mean value of the relative change in gain is plotted as a function of

dose. (The lowar end of the operating current range was used in these

plots for increased sensitivity.) The histograms of Figures 78 through

89 illustrate the variability in radiation sensitivity of presumably

identical devices. The radiation sensitivities are represented here by o LlE

low injection LI, Ig/Ig, '-(1/hp ) (= J—) > and hpE/h . Tables 78

through 80 show the mean and the covariance of these quantities at various

injection levels to give a better overall picture. (Note that the fol-

lowing bias conditions were applied during gamma exposures. 2N709:

Vnn = +10V and I = 20 uA, 2N930: V_. = +26V and I_ = 50 uA, 2N2905A:

V,„ = -30V and I„ = 40 uA.)

77

*4

b. Discussions and Conclusions

By scrutinizing the rank correlation tables the following con-

clusions can be drawn:

1. One of the promising initial parameters, the low frequency

1/f noise showed little correlation with radiation sensi-

tivity. A possible explanation may be that the 1/f noise

contains ?. current dependent term which prevents the deter-

mination of the density of the slow states in ungated de-

vices. Hence, any ranking of the devices on the basis of

noise will not necessarily mean a corresponding ranking in

the density of the slow states. This latter quantity was

the one with which correlation with radiation sensitivity

was anticipated. Apparently, the current dependence of the

1/f noise obscured any significant correlation.

2. Surprisingly the radiation induced AI (low injection) did B

not always correlate well with AT (high injection) espe-

cially at the high end of the operating current range for

the 2N709 and the 2N2905A. The correlation is quite good

for 2N930. Consequently the practice of using AI_ (low in-

jection) for predicting radiation sensitivity at high injec-

tion levels, just because AI (low injection) offers a great

increase in measurement sensitivity, has to be treated with

caution. The explanation of this effect is not clear at the

present time. i j

3. No significant correlation was found between the radiation

sensitivity and various initial parameters, e.g., burn-in

changes., I_, I„D„ (or BV ) measured at various tempera-

tures, etc. In fact, the correlation is practically non-

existent for the 2N930, slight for the 2N709, and moderate

for the 2N2905A. Even the moderate values («0.7 - 0.8) of

the rank correlation coefficients for the 2N29ÜS* are far

too low to expect any of the various initial parameters to

be meaningful screening parameters.

78

The lack of a strong correlation discusned above Ls perhaps

not too surprising. For example: it is well established

that the magnitude of the surface components of the pre-

and post-irradiation I is controlled by two primary factors,

namely, by the amount of charge within the- oxide and by the

density of interface states in the mid gap region. Either

of these factors can be lominant in certain cases. Also,

it has be, ,i shown in the past that in certain oxides there

was a correlation betweoii the pre- and post-irradiation den-

sity of the interface states (Ref. 18). In such cases, one

may expect a correlation between the initial I and the

radiation induced change in I as long as the surface com- A

ponent is always dominated by the interface states; this

'f-.U have been the case for the 2N2905A. It should be noted

that trie radiation induced excess base currents (good in-

dicators of radiation sensitivity) tended to be higher in

chose 2N2905A devices which had higher base currents initially.

In cons'crast, the amount of radiation induced charge accumu-

lation has not been found to correlate with the initial

amount of oxide charge (or the initial interface state den-

sity). Hence, whenever the oxide charge dominates the sur-

face component of either or both of the pre- and post-

irradiation Is, one does not expect to find any correlation

between the initial I„ and the radiation induced changes in D

I . Such might have been the case for the 2N709 and even B

more so in the 2N93C since the rank correlation coefficients

were very lew.

The consistently high correlation between h and Ah is r EO r E

not significant since contrary to expectations it does not

guarantee high correlation becween pre- and post-irradiation

gains, h and h , respectively.

The high rank correlatio-i between h and /'h , simply im-

plies that devices with higher initial gain will suffer in

general more gain loss. This is quite obvious e.g., whenever

79

the radiation induced base current increase, AI , is approxi- D

mately the same for all devices of a given type. Although

AI does vary among the devices (see histograms of AI in

Figures 78, 82, 86) apparently the spread is not large

enough to obscure the correlation between hv„n and Ah™. r EU et

5. The utility of the high correlation between h™. and hrc rtu rE seems to be limited and the results should be viewed with

caution. The correlation appears to offer a useful screening

technique against h„_, failure [defined by h__ < (h_,„) . ] rE rE rE min

since the initial h_„„ distribution can be simplv truncated. FEO '

Figures 90 through 92 containing 2N2905A, 2N709and 2N930

data show the futility of this approach since all of the low

gain devices do not fail first. (Incidentally, observe that

that the h versus Ah™ correlation is high in all cases; r EU r E

clearly a useless result especially for the 2N930 devices.)

It should be noted that the relatively high rank correlation

values for the 2N2905A and in one case for the 2N709 are

partially due to the fact that the gain degradations at the

higher currents were relatively small. The rank correlation

is expected to go down with higher relative gain loss.

Furthermore, and this may be the root of the problem, since

the degree of correlation between pre- and post-irradiation

gain is very much oxide dependent, one cannot be sure ahead

of time that a high h versus h correlation will exist FEO rE

for any given untested device type. This can be checked out

only by extensive experimental work which is just what we

are trying tc avoid by a simple electrical screen.

In other words, any given initial parameter which shows a

good correlation with radiation sensitivity only in certain

transistor types but not in others and whose behavior is

not predictable, will have only limited significance, if any.

Right now such seems to be the case with those initial para-

meters yielding reasonable good correlation in the 2N2905A

but not in the 2N709 or the 2N930.

80

6. Higher gain devices tend to suffer more relative gain loss

(Ah /h ) in certain-device types. The effect is clearly re. rt.

indicated for the 2N2905A, slightly indicated for the 2N709

and is completely absent for the 2N930. Again, the problem

of complete inconsistency in behavior among the device types

makes it impossible to generalize.

3. LOW-POWER TRANSISTORS - LOW DOSE SCREENING

The basic idea behind low dose screening is that devices which ex-

hibit a relatively large radiation sensitivity during a low dose exposure

are the ones which are most likely to fail after exposure to a larger

dose. The approach to test this hypothesis was rather direct; comparisons

were made of the tails of the histograms of radiation sensitivities at

low and at high doses. The radiation sensitivities were defined in terms

of I changes at both low and high injection levels. First, an arbitrary ß *

value of Al at high dose was selected and it was asserted that all de- B *

vices with values of AI > AI were to be rejected by the screen. The B — B

position of each AI of the devices in this group was then located on the o

low dose historgrams. Were they also located at the appropriate tail?

In general they were not and hence, the reliability of low dose screening

is very much in doubt. This conclusion was also supported by the rank

correlation coefficients between the quantities AI (low dose) and AI a a

(high dose). As shown in Table 81 the values of the correlation co-

efficients are much lower than might have been expected. i

In the following material, the results and conclusions for each de- |

vice type will be discussed separately. 1

a. 2N930 5 5

Histograms of AI„ @(3.0 x 10 rads) and AI„ @(1.0 x 10 rads) D 15

were compared at four test conditions. Devices in the part of the histo- 5 1

gram with the largest AI (?(3.0 x 10 rads) were generally scattered | i

through a relatively large part of the histogram showing the AI s at

1.0 x lO"' rads (see Figure 93). These data are summarized in Table 82

where it is shown, specifically, that to ,'liminate the N devices having 5

largest AI in tha R rows of the 3.0 x 10 rads histogram would require D

81

screening out M devices contained in the P rows of the 1.0 x 10 rads

histogram. Both Table 82 and Figure 93 emphasize the fact that low dose

screening does not seem to be a viable technique for our 2N930 transistors.

b. 2N2905A

The data for the 2N2905A can be treated in the same fashion as

that for the preceding device type, 2N930.

To eliminate the N devices having largest Al in the R rows of 6

the 5.6 x 10 rads histogram would require screening out M devices con- r

tained in the P rows of the 1.7 x 10 rads histogram (see Figure 94 and

Table. 83). On the basis of these data, the low dose screening does not

seem to be a viable technique for our 2N2905A transistors.

c. 2N709

Histograms of Al (1.7 x 10 rads) were compared with histograms 6

of AI_ (1.3 x 10 rads) at three bias conditions, Ir = 100 uA, 10 uA, and

3 uA. In the I_ = 100 uA and 10 mA data devices in the part of the 6

1.3 x 10 rads histogram with the largest AI were found in the higher 5

tail of the lower dose (1.7 x 10 rads) histogram. The best predictability

seems to occur for the 100 pA bias condition. The results at I = 3 pA E

showed significantly poorer predictability and were more like the data

analyzed for the 2N930 and the 2N2905A. One apparent difference in the

histograms of AI for the 2N709 is that the maximum values of AI were

1.5 to 4 times the mean, whereas for-the 2N930 and 2N2905A device types

the large values of AI were typically only 20% to 50% greater than the

mean. In other words, the coefficient of variation for the 2N709 distri-

bution was much larger for the 2N709 than for the other device types.

Table 84 summarizes the capability to eliminate the N devices in the R

rows of largest AI_ in the 1.3 x 10 rads histogram by screening out P 5 rows containing M devices in the 1.7 x 10 rads histogram. Figure 95

shows an example of the procedure used in the analysis. As seen in this

figure and in Table 84, the low dose screening does not seem to be a

viable technique for our 2N709 transistors.

82

"0

A. ELECTRICAL SCREENING - yA744 OPERATIONAL AMPLIFIER

The problems of predicting total dose damage for the pA744 opera-

tional amplifiers are in many ways similar to those encountered with the

bipolar transistors. The most sensitive parameters of an op amp in a

totai dose environment are the input bias currents which are simply the

base currents of the input transistors. It is important to note, hew-

ever, that the radiation induced increase in I is somewhat reduced, one

might say partially compensated, by the constant emitter current biasing

circuit.

The rationale behind the selection of the primary correlation para-

meters, the bias currents, was discussed in Paragraph 4-b, Section V,

Volume 1 in some detail and will not be repeated again . The radiation

sensitivity is defined by the absolute or relative changes in the bias

currents, similar to the procedure used for bipolar transistors.

The uA744 op amp exhibited a "medium" radiation sensitivitv in I- a

during total dose exposure. Although not very high it was by no means

negligible! (It should be noted that the op amps were irradiated at bias

voltages of +15V and -5V respectively in order to maximize the reverse

biases across the junctions during exposure.) The initial base current,

I distribution is illustrated in Figure 96, and the effects of the radia- B tion are shown in the AI and the I,, histograms in Figures 97 and 98.

B a The data emphasizes one very important message. The use of a failure

level definition of ID > 750 uA essentially a Honeywell specification will B - 5

reject 6%, 17% and 21% of the op amps after doses of 2.7 x 10 rads,

1.3 x 10 rads and 5.6 :< 10 rads, respectively.

Table 85 shows the rank correlation coefficients between the various

initial parameters (noise and others related to I„) and the radiation

sensitivities of the op amps. As seen, there is a definite hint of cor-

relation between the relative changes in the bias currents and the initial

parameters. However, the values of the coefficients (<. 0.J5) are not high

enough to have any statistical significance for screening purposes.

:■

83

1

5. LOW DOSE SCREENING - uA744 OPERATIONAL AMPLIFIER

The approach to test the applicability to the uA744 of the basic

premise of the low dose screening, "devices exhibiting high radiation

sensitivity at low dose are the ones most likely to fail at high dose"

was the same as presented" for bipolar transistors. Namely, the tails of

the histograms of the "radiation sensitivity" (measured by the change in

the bias current, I ) at low and high doses were compared to determine if

the devices maintained their relative positions in the two histograms.

As for the bipolars, the simple approach to low dose screening did not

work. Too many devices would have had to be eliminated after the low

dose test in order to make sure that those few exhibiting excessively

high radiation sensitivity at the high doses were indeed removed. This

conclusion is further supported by the following poor rank correlation

value between low and high dose radiation sensitivities.

Correlation Low Dose Sensitivity High Dose Sensitivity Coefficient

iID(2.7 x 105 rads) -I°] [£„(5.6 x 106 rads) -I°] 0.629

D D D a

After eliminating from consideration all devices with poor or suspect

data we found that those devices in the 5 rows of the histogram corres-

ponding to the largest values of AI at 5.6 x 10 rads were scattered al- B 5

most randomly through 15 rows of the highest AI s in the 2.7 x 10 rads

histogram. Precisely, to screen out the N devices in the highest R rows

of the 5.6 x 10 rads histogram requires elimination of the highest P

rows containing M devices in the 2.7 x 10 rads histogram. The results

are summarized in Table 86. Again, the conclusion is the same as for

the bipolar transistors i.e., the low dose screening did not seem to be

a viable technique, for our yA744 op amps.

84

T*

i

O

z «s

to

—i

CO C O

o •rl •P « ■H T>

Pi

< m o

z rt •o H 01 §

Pa o <u n X CT* •u z es

o

CO u V 01 3

3 60 •H ft.

oaiu/aiq Nivo aAiiviaa

85

"V

>o HI »

< o

1 m a ■J- y-i 1 f- -o r. IP i 1 «-• O «- .T ' ■J i

r-- -!■ *1 in IT in tf> * 1 u «** -* 1*1 *- r- •f -* f*l i

tfl */■■ i«' J-- a> I rsj U* in r rr <0 1 «1 .■»1 U <: -- . «■- in to n in \T. » JJ O wir <r *•>. fM i'» tit fcH j> * »r >.. 1

fi «i a- "j «» O CJ .»■

i/. IT n * LI iP <N T o <l .">

*4f- n •J ü i' p- *1 «n if» 4 ^ -j ^ i.i r*- P» f- '■- .j

4 n 1/ T> ^ «" V i/' L'^

in •r * a> IT -T) >r V a- 3' n in «n J\ 42 CJ

m tr *» < # 4 •T U1 J> in .n •-• o ■r U •+> o r\i » > O O <*■> ■0 ^ Ul ft /) t1' j •N .j

IT ■.'» *» •T -j r * J ■r. r- o ? u . 4V *■ ■J- M 7 - 4< 1.1 m cr .jn o ■*. o •T 4' n" ift ■J <■ f- r-. ■t ■'

m •* -» * .» •* ■t J f J- ^ ir v« t **j « »■- <", Ui :< «\j -,

o« T- r J n 1— -t ■J •J * «r« ."» s ■" m * * t J J J 4 ■j -* ■J 4 -» •• r ^

- -J 1/^ *-

O OOOOonÖQ.1 7» o f ■ o O O C «"» r> C.i r o c J « -r j r

i i i i i

: r-i -4 ■ J J VT'

O CJ I I

-.11 UJ

O O O O1 O t\ O O O -O Ci <3 ^ iT NN

O O " n o o n 9 O '.: 'j ^. a rf> C O "

' ^ -i■*) O O 0£<

- O '> fv V» O -O O O

o o r» O O O ' • w «a *i * ^ ♦ i

< 3.

CO

n

m u

M u •H • •> 4J to •H •a > a> •H u «J T+O

CO o C H V c« X

C in

O CM •H • iJ .H TO •H II •a

5 to o

a) Q x: v^/

*J

(fl c u

•H « in

c TO o 2: i-t u u TO c

21 TO IH

MJ « » «H a 1 O tsi

( g 9) TO JC

1 t-l 4-1 00

1 o tic 1 « C j CO o 1 1 1 t

31 1 1 1

• 00

1 ^ r>- 1 f-J

ai C 3 60

•H b

- I

J y-> if J\ iT.

ifi if» IP ■ f\ i/\ •

r*» O V* *' tf« * +< O * r> & W in 4 ^ »'[UN -« T> CD IT -t 4 fVJ fl 17" O IT -*i K :ir 4' *

•* < U-. "» •» h~ f> n» rg £ # i/. rfv »'■ IT

^»«)f\JC T rt I" >I ' <; <; — r- -■ -u a f '

4 •» »f «* * 'J^ u- "T» -' j -J O — O 3.» OC i.T <M J i o r> y

4 NT r <t :

\THT\ 4 + -? 4 4? -f 4 4 4 •? 4 **f

f- rj rr O r*> « o •» Pin IT ,* X m lA J r* in 4 IT J- u\ it

O O O O C O *■* -< f>l -i C lOflOf. C» O O O O O O ri r.' O O •■* —< -* • « -* -* -< ~» -' -• -* ~«

■P r\, o -o *"> O C* <7 *D •"• iACP -• 4 r- ^ n* n |i u n o UA A ir> u rj rt n iv K ^ ' if" '

<l «i n", N "J H T 4 o* O >■« *-t ' J M '•*■< "' 4 • I I t I I I I 1

O it O O O O O O O O X r*\ O U * f- *

U(^oOOOOOOOOMf-^-a;^*0-

in j' -u *c r-

) u in *_J UJ ui u; tu i en f^ 7 J 5 <t O

> 6 o o o a o o *y o -•

r»-\ if\ 0 »-«**- f" f» lJ 0- 91 O fl 01 Ü T O O >T 1" O O

v <fi n in « -f j T> m O O w O O O Ci O O O Q O

.^ IT —i o *M —• t* ^-* r> o ^4 -'i c o

3.

Oi II a)

*J •a •H « CO u c aivo w o

1-1 c o X

•H 4J IA OJ {SI

T3 rH

a 01 4) ^ 0) *J 0

a c v^

1-!

CO Ci ^ o 0) u 01 01

J3 >t-( U M-l

•H 60Q C

•r-l *U U 0 « h (0 4J M «) 0 3 4-1

r-l to iH •H M 10

O «!fl H M *N^ H m M o\

o «H l>- o z

r-l

s 1) (0 J2 M 4J M 0 60 4J G 0) O

SI

r-.

0) n 3 60

87

^

.-I >o l

-• I . re, t

O (V'

at -o |

UJ u.

of ^ I UJ W t a u i

I ? At o -» lt\ lA

■ .» * » IMAI ;»«• in r> i in to * « IM . </\ "> ■ ift J\ - •4 <D f tf\ PJ f l|\ i/\ t

■JT PJ ',

f.4 a I *Tt •* •

JJ »Ui i i/\ irt -a ■

P- r- BJ tT r « w* w u> ■

iM *• u\ u* . m j » -»

-» ^ # •* * * 4 •)

_• "i * ^ <M * t/> if ii>

? IT. .-^ * 3 in .4 -t

U' "I * 3 ^ tf- -4 •> Ml ^ J>

J «*■ J —'

f -J> t JP

* P- m tf>

m 4/ O f- ^ <M --• -r

H ?' t' IM< r> o o o ,c »n »A »r> iT «A P- *J * / »4 O UP»* ip w* «r* t u"\ •-»US 4J P- o p- a «j* r o * r <# * »n O Kt .f> J* <*» c-excp «r *■ «r ■* #

(.1 p- r r- «i ' * •* ■» t t ■ —* PJ P <• * * ■ r ■* ■* <? * ■

■ 'P. x

# <t 4 * -T >T *

?T* PJ PJ Pj »P INI PJ Pi PJ ■-* «-•"->-*-*.-* ~l »-I #•■ ~l -4 -*-*■-«•■*'■-• «-*•■'•.-• -I -.1 •-*••-« ^.-«-«.MW-t.M^*

IOOOOOOOOOOOOOOOO^OOOOOCOOOOOOOOOOOOOO i i i i i i i i i i i i I I r i t i I I i i I I t * I i f ■

UJ U' U U' l.j MJ U< J' LJ IU «j IJü UJ III UJ UJ UJ J! U UJ UJ Ul UI UJ Ul UJ lj> 1) UJ Ul tJ! Jj UJ UJ iL L: UI U: UJ •M a.- «J- I> ■# V *V J"i O p. P* O .f *»- C.l «I P -« -r *- .-i «*«•-•■* T •-« »p. «■*■* in V ^* i/^ 0 PJ IP T PJ P- 1 O ** C ^ <•• < M >r C 9 M <" /■ in 'j .'A IT. *- 'S PJ •* ■ ♦* "• -* t © OD ^ "H jp. «r o ri ui r- T PJ

« 4f PJ Pj PJ ^- P- 1 I ! (

oOoooooat *OT«.

OOOOOOOpOOP-Mf i^f-.s*jfr*»*p-jnjp-p- /j/MVPiwMO-t'no^coo'jio

pOooooo'i<HO«"nr6o-iir<i'tii'n^j!fK'fl^rtitnMfloft(^oo(

1-1 CO CO •o c cfl eu U

C/) vO

c o o M

■H JJ X «e

■H lO •a CM ca • Oi i-J

a» II J3 JJ 01

CO C o i-l p ^^ c f) CO

•H ^ JJ tl cd IH

■H CO M » CO > 4-1

C

J2 2 4J V

14-1

60 «w C •H

•H Q JJ CO <M n O «J

3 CO

U iH 0 i-l 4-1 M CO

•H Oca ss

J= M "»^ H b. too»

43 O 1^

0 i-l O 60 JJ C II CO O

•H IMW

• o oo <D I-l 3 50

88

■«

i .£ i c O

» p- ■*> ■ 0 *Ti If f IB .USi lA •

•H r- .* 3 f •I I «'V ^ I US ,|\ if- 'P - «D "V » T- ' i*t *■ sf n Ji U, Ji «i - O ^ «i -* f f. C ■* "' •P. 4% at ■/• iP r- -c -■■ rj o <M »-« ■ «■» tt\ «' w» ■ P- -v O -4J I

•A O f u . i -a C# C1 O- ' m m if N.t tn »■ * h- i -* D O JJ ' IP •? m j

a* ..-*: (*-■ si ,f. -C O -t J TJ , IT. .1 ■. *■ -f * r- o r^ s v

IP. 4 -# -f *» «\j -r> -t c* * o CP -f -y r if -t t * 4

.-* ,• fM ,f ^J

•I -r

i j

■i r ■; u tj

iv Kj p.] -M -v «•■! «inKi O t» O "-- O O a O f» ' I I I o .

o e» n o c» -) o ^ -

o o '? • i •

■ r,i ^ n, <t* n\ r" ,'J 4} V ' 0 O *> <> 15 ■*> S> •£> O <? O C> a

I O •-« t O ■•• O C *' —* -O ■•*« O Ö O

r o fu n '> N r< *5 •■< N O ^ n '

•H II

<3 0) 0)

a> s JC ^* u

0) c »J

•H 4)

c « o S

•H 4-1 u « c

■H V n u rt (!) £.» «4-1

w •H .G C) 4J

H-4 6C o C

4-> S2 «0 o n 4J *J to

3 •H 91

.-t S M

H U

M ON ^-^i *»» O <: «r^ 3. i-i fe < fvl o

o 11 iH

/—* 4J II w fnöO u

.c G 0

l-l

rH 1 U1 «a ■n

m rt M-l rA

n •H 00 4J X 0 •H 4-1 U) u-i (6 d CM

•rt <u • SC w r-l

00

60 •rl

89

^

60 C o 4~ < 41

w

o

• « > o w I O IM

o5 (

oo fM -«1 o o

O ru in * .-t <M o o o f^l **- *

(«IHN o o o o- * +i IN f< -*

o o o m I-NJ if >

o c»

a- *> <j -* r j> i^ c* 3« (SI ^ o O ■3"' rj V» O .1 IT cr J> rsj «".' CJ <.j o

p- ^v 'j fSI J r .t

1A 7 c ir f i f\l <-« O o ') o 0>> »Ti *> a a» »fi w ^ 0" <r r^- *•> "y

o o D cr ■.J o> •r; ■f « <»• -3 IP J) sC <-4 O '1 r- 7» 9 -1 K /l r-

*n T o i/* o ■■* <_» o u n «J T ^* k* -^ <-l ■ > r* V J f- J r- O r> ■ ) ■i o *-> t_> *-■ u r*-

nf\ T rr

-^ j <*1 ri n " O I o

~>J -v» r- f .r* V

o o o O O o D O

UJ UJ li' ÜJ U' ■*» r- ir> fv u^ * «A *». ■* -'

UJ UJ MI UJ UJ 'j ; JA (J «*. f- -n T •

, j r* » h- ,-, r .

I N (\i ffi * j> u f

O ^

H OC X ■» ■» CI O O O .-«.-« -<

> o o o *•« «-< -«

--)ooo©oor>:j ' "i p o o <*» o o o o a o a o o r» i i i i T i i i I i T t i I i i i I t i J i ( i

L> Ui I.IUIUI UJ tn'iJi t- <v -j. u.' OJ a ii'u. ii' JIU UJ u' ut u, ».')■> a1 f- o <f> >r *t A «j H (j j- n N a i/'. ^ "■ M H M o »1 c N *> I 'l* i *■ T L* O H '.V M « * .r *} f* T ft- O •-• M

«f*i-.^„HnjM(\j(vjNWivji j ^ ir\ r

o o o o o looo o

o o o o o c> p- t> • * j o o -J * r- ,- u; TI .o r- o N o o o o o o o«

0 C 'T tNOOOOOC'JOO»

-"> O O -1 D^Mf i m eg «|j h* ti »v ff o "^ r> *? Ci «^ O O '

0) • «• 0)

CO

c H om •H o JJ iH rt ■H X Tl

PS o • M



■J *-( 4-) o V) 0 V) r-i l-l fH o M 4J

W (EM

M (fl < S >4-l >J

0 H

§ O

M o\ OOZ O CM 4-1 w a) •rl X. SB u

CM 00

ft) M 3 00

90

TJ

> O | I

' I

o o « r-g

ifg3- :

» tl O I

i <■ «V f u o t «■ H" f

< rn —i ■ ■*> c c — -J C) ! r* ~t .-« c. u ■"'•"* **■• , J) » <^ "\ ^ * - 1) £

''-I > i/> <"^ *- T" O ", *- »N . M O ——* '? U -« f-rT5 -« 'O-lfTj J'iJNÖ-t

O1" t> «■ o ""- *- r- -M i> t\j

1 r- -» "W u"> /" *■ N *» * fo r- a» <r\ u- o ^ t- *-i■ & —i

i 'J U - U '•> .' i J -• J ^i r f\j j, M r

., i i*. tu -n 3- o f- -j i. ■*■ > o © O O -* o o t '-> v.» t>

J * *J o . < ni f j N m N / j *X r- «\ -o J't-rfHIJ^t^"' r .n T T) .♦ -n

- ^-•NMO-'^ M t D j r\ o -j o >- © tr © C3 <H

'0~iOOOf>00 o © o o o o o o o o '-■ a *> o o o o o o © o o o o c.i ^ o o o i i

Ui IM UJ 1» u; IL' UJ J UJ U 1. U1 1> <1 UJ u. Ui U Ut it' UJ it -1J LU U Ji i.t U.« 1 U li I C .jl U I Ij I LU U.I if' "J «■M^ ir^cift.noi; T»'%..»w^f-G'"'"*f»u"\r-(y>.-t*( r o ii ""> *\J .j -r> **> ■"><-» r* IT. f» c ■•***** IJ r- f> j-> "■ n M * A s j Ci ^ c * t ■& +- i* T'O'Mmsj /•■ o T c* o •-« fsj 4 in ;* ^ 4* o •-* *-J

M •* m a ** " i^ t^i •It |

i ■ !

0COO0OOt>OO0-*t>-d ^ Ob MS <f ^ ^ ^ I a

-< r\j *N "4 i\ 1^ «w *»J ^ r»\ ft "■"> (*> f° w> «1 f" r^ * <*

O © O O © © O © O Q O ' ■O f r* -o * -o «•« * vT «■ /) 0 J o—* 'J o—* -< «1 o © © o a $£ o

I

li

w (U

•H 4J •H >

•H 0! U T) 1-1 cd tn ^ c a)u"> w 0

r-

c 0 X

•H 4-1 0 « •

•H CO t3 rt n tf

01 i) en

43 0 4J O

P a 0 +-1

*m CD 4-! r* M

•H 4J u G (0 (JI > H

IU 4-l Z \ O CM r ***^ IJ^ :.

'! M 4-1 ■i,

60 i-t ,' O 60 ? W ß I .. CO 0 ■'

•H | W EC < M

■;

ro * 00

; 0) S 3 60

91

"V

\ i 0)

• IM !

o in *- • trt »ft

Z 41

uJ 4) '

<r JJ i — * i OO i UJ O ;

i n ^ - —

o o

r * Ift o o o

flo o »A -» ~« n NIC H «M ■**» 'ft o o -» ** ~* •*• K ft' <0 h- »■ O C- <0 K P- H f J 4 M O © U -• <~» -' -•

«I f«- f» _> o o O -4 CD O T» -i tfVF- A ~4 o o o -< <0 T K O Ut .| u\ -

w» 4> *»» t- «II *t J\ (1 <* o <-> -«

fl f- Jt

o O O 1\ >& f PI a T O <"> o

t .«o I tf ft

> T O *' 3J ' ,J *_• mt ■*., ».J C oj it* M K' -O .

$• *■ c 'v N i

o o O (i U ^ u •t <4 t* sr t* -•

* -* ^j r^

*- <j- ^ 3 ~

o o o o o

) -■> c o o -« A J t\j C' <?•

«i . IT. -* •* -T -T

O f** rJ <-> f) o S " \ T "I O i-< rsi —< -^ .-4 "•> O O o o o o

> <> O f-, o o o c» o o o o o o■ * joooooof'OOoorw

a UP.

0

tc at

^ lM u

-0 it

u ii' r^ o

vt 1 1.'

r t -0

.u

r\i "j

'IS ft 13

J b

(M

U-

rn ■J

UJ Jl .'i J

•LJ

r- a' o -,.. _. ** r *« -« «M A- 'M ■*# fv fy ,J

rj f\J m rt r« n r^ n' m l"i ■•> 4 't 1

>» > * •* * 1

"' oC o o O O O o O O f- ■J -J -u < - * - r- 4 X fli i/\ - r4 1 - r- - O f- *, O Cl o o o o o

! I ' 1 oo^>ôooor'j«-l^'M-<0'n*-*n'f- T-AO^ ^>*^*^)**'>^ôoo'ô'io

I I ■»

« CO c -a 4) a W u ÖT1 o o

1-1 rH 4J cd X i-( *d o S •

 4J

c 01 01

X. »4 u 0)

«w 60 «w n -H

•H o 4J CO <« K o

4-1 en ca 3 u

rH o

OO)

es * U ^ H

W WD

Ji cn o\

>*H ?5 O eg

^N

15 1 Wl 4J 00 i-i O 00 u G H <n o

SB -9 [xl

• 00

2 3 00

92

"V

5 I

S3

m rf\ a) F4 1 f^ o o o o F t o w ^ n O O <"•■ f\* rj m if, j .ft T *"1 *i r- O o <*J ' > -< O •M »: </> r- f- t- ' * trs ,$ ui i/i M T *• c; C< ** o n n .0 > ^f o n r- u r- T> •r. o fM r\ \r< j- i o 1- ■o O .0 O c* M -J o i/\ P* Ti ■M fT r f« tpi (.- "■» n rg *> ■T •J f« . n J r.-' "!' •r X C J> y VJ u O < > -' o * ... ') J ■■ r , -* ae Ti a U if f- -\J <% r.j r*l pi r c" V :.N ■j n- Otf'H o o • J (. > u ti *..- O .■* IT*. U O Ü" "^ -* r- <-J -J ff t r\* u »4 <\| *Nj f^ J- C> . u-. -r « t, ii1 K. o o —t —« (. -* . , _-. o ") '.i — O O O oj cr t> r- ^ T ■*" t *i T »J n "1

r ^ 3 ' r r n ■n * -o OO o O -* -« : o o '' '■ o O O '">

•> -f i+i m I* *i i< ,i: I-' rt *i », «i *i .TI (fl «\ «\ M in (^ '^ ^i i; r\; iV rvj "M i\) («j f, rvj f\j fVJ M oj \- v N o <~» o 7» o ^ o o o c o '' o o o o rj o o o o o 2 o *"> o t j n c *~> c > o *"> o r» o o o <^ I I I . I I 1 I t I t I I I I ( t 1 I I 1 1 I I I I I I I I I ( I I t I t I I

> ^J -4 T t>- t NJ - . -- ,■> ftl o V r -1 -4 > f^ .1 ■vj 'J r>i ■»- « r —. .') r o if* 7 J- ri -T iT ■n » 'i

Ifl ro -' l~i o --1 '■> ir T •* J1 > ■" i^ '"■ *' (^ f\l '- *' *^ '1 ' ' .-

-■' ■'■' iM •1 •^ * -J » ^ r> J- X> r- ^ 01

r< "* - »N. "J m »i «1 ^ -J »" .A ■0 •a r~ r- ai 1J V' tr - -• -' -* ** ** ^ - - - •r* 1

^ " -" .. ^ ** -*

c- -5 o - n o o o o o o - f- 0? E ■- r~ J - j r- r- H Ti - t*- ■f O O f- o o " o C o o o O

"i O o ' O O f- . ^ -t r- -T i> ^ ' ''»(^-*ooooo'^o•^,^

-T» o r-l

« c o o

•H • 4J !*1

<d •H 1

CO 0) « CO

0 a> a x, ^> 4J

CO Ö n

•H Cl> «H

0 CO C) S

T^

■J *J

5 C ,H 4) >-i t-i CO »4-1

U-l 0) -w x: Q 4J

««-1

60 O C

i-t CO u u « o H 4J 4J CO CO -H 3 B)

rH TO H U

H U

M o -^ ro

CQT> H Z O CM

II (11 x:

^N *J w (K 60

XI C

Öl < <3 a

CO m a) M O -H

4-1 n 8-5! w i-i -H M 6C U O Tt • M

4J « «) W C •o •H a) CO EC W u

m 00

01 u 3 60

■H fc

93

r

> o

Z1 UJ

UJ CC

m so

' I

0> a> < tf» if» <

o in t if* v\ f

> «N UA

i * t> » d i» J

fi rg i^ 1)

* (Si *

■* o o <•> ««1 »n

ff rsi «

& ■€ ■> •I* O 1 M "1 «

« IT V- '.1 T i, C> »Tt ■r n i

•H 4- (^ tf o tf>

:1 V .

<*■ Pi m T M { <o .-r o ~« » c

IM ■** rO ■*! CM " 4 ri u- f- tr> ~« -* .P r ifl T N or » T •* r~ ' in (M tt rvi <\i ,-vj "• IM f

) a Q . i^4 •

<-j rj ' O r- • * -J i i-J '""J

■ m

<J -*J N W 'i r *l -r. 4 »0« fM I^J M (M 1*1 *T '« -4J ,* r-, t|*. a-. oj (*■' f\j c* «,.„.> Of r«! «i ^ -j> r <\tfsi *M *■!

,f J* n r- i/\ -O l -j (fl ». N »I I .\ N N N -4

I CO I

o o O O i I I I I

U- Ul LI. 'Jl ' r- % r» w r- r- jj o ■

O r> C> O O O t"? O O O O ' t■ o■ o o o e TOOO fh. *k. pi.

mooooo f- *- a a- r> ■

'Jl l'. U. JJ 11J UJ t-i 'JJ i.i ■ -** *- <N h- -\j ^- *s.i i/i -j> i ><-| M f\| -4 - t 'J * J *l -^

o o o o o *- o K 4 ■ r» +■ ■* o •» a oo: »* v ji ■ ••* -« a m * o■■*•-. tn %r c- 4 o O ■# o © o f- 6 o

'3 O O O O n C C »* O O - if i ifl cj h- C n i\| f- in J N /\ J in rt o ^ o p w O O O O'.i ' '■

'»OOÔ-Ô«^*^* • IMT* *0 i»1M*'Ci 1 •"■> .1 -"> - * o o

(0

•H (0

•H CO > M

MA •H o !ft I-l c 0) X

CO ^D

£ • 0 m

T-t u H (0 H 01

13 w 5 o

o >-•

HI Ä 0) u M

m C IM

TH CO 3:

C 0 4-J

■H C u 0) CO h

•r< 0) H 'i-l t8 0-1 > •H

Ö (U Ä U-l W 0

or CO c h T-l o 4J u 10 CO u •H u to to 3 §

■H Wi iH H H <

pom H O <l CTl

CM IH a o <M ^*» I 0) <

Ä 3 M w oo C1 o be 4J C II tn o •H 1 MW rc

vC 00

a) t-i 3 HO

94

^Û

> o t

O I Z I < a. |

o * o o f\| -1,

r (* (\ "VJ

■# «J * o <r x» (T. f. "J

1 (M «V t- » <n (*-, r

j f, Ji •*■* J> * , r>i n Ki ft

s o r- ■<> -!

^ *n «^ 'i* >*%

- C -O C T> 4 *M •* "* t

j nj j M (t. tfl i

4 'V - > *-< *\ -T ;*J r>-

> .( r\t r-i T> r-

tj» C* o "iJ rj -

t- .-* C ip f ajcc i/ r- <v ■ fsl ro *g r-i ", '

!/> r o :f J"1

* j- r -< '■> .

<\J f\J <M • J ^J I

•9 -, -U *>

> rj o o o f, c> ^ ■ o c '* o O ("■ r:

U' U' '.'J '- rC O t» -* "* ■ ■ *! T r< *1 »T' •*■

-T lT\ ■!> "ft f- '"■ ' i -M .J /' X U -' -* O -\

o t> l a D o o o ."> o ■■•■■• * *' ■* o ?> 4> J> *

o -■•• o o o o M o o oooono

>jj UJ Ul r- r- n r (7 * ti» '11 ü-l U' w T C- O n - - ■*N u" t CM-»

f> p» t* a> o T' (?■ n» tK

1 o o ■?> o o c- tJ O O

O CJ ''-' OOOlHiO

o <t> o a o ci ■ «,i ^ ■:• T 3 c% h T - ■* "3 • ( o -< n ■ >'

. (0 (U •H 4J •« •H 0J > T3 •H « 4J M •H (DvO c o V i-l V)

X c o \o

•H • w to « •H II "O

PS (A 0

M

c <o O 3

•H *J tJ

si c •H 0) M M CD 0)

»■ > IM < «n t *J -H r? 3° — *■< .A «»-( X 00 o

e 'V •H M)

1 i- 4J J-l i . m o ' a. M 4J j *J Cffl 1 u;

if 3 (0 ■ <

T-i W ! U

UJ M h X H

! - o pa ' UJ M <; . -j »«« IT) 1 l/V POD »-- M W i = CM I o «w Z 1 Ul O es 1 * ^-v , <

g a» < 8 J3 3. —

L« . s- « ■7- 00 f» '—* o 0 M

> °u ß d "> a o o3S < H

"joo o

00

95 ? I

[

h- ■ 5/1 f*

O — U.' -•

> I o .> * *■ * ■*

<M rj rj «V m «-> *» r- * *>* M* ■» INt /Ni *M <M C p*. r^ si \r\ *> If» «' ^ U"N (MM «M N m ^rjOl- a cm» * « -o *S» f\J CM O* f\J

fin n\ t\i »> -4 M M N M N * f» if» f* r»

W N M N N

* IT« *. ft

If» T» ^ *\ ■* <

^■i T« -j (M *»l 1*1 MI 

«i -A; tn r- y\ (M

'H •* * O * X PJ *J *l Al * ft <\J -I -t

,r ,- -J m -H i-i «i -*\ .«■». ri .f T C IM K rt

i o o t (*1 c">

I OJ fS* fsj f-; *1 « IM

t I If» » 1

l_|L»4.»«.~l.-4«.>-4.-< oooooooooooco 3 o O O o ■ o r*oor)'-»oooooooooooor>t>'-> t I I 1 I •

_ IIJ ill ■ < o «> *, > * if n i

i i i i i i t . j UJ UJ JJ I*I it» i|i u: ■ r- tn ■♦ M -> '» T> >o

.n /» r- i) '* **> '.» -«

l l I t I I I l J UJ !li IN Ui UJ UJ U 'U UJ

K. *r* ^ oj -» > r* •4 *y *» ^ m m *>

UI "J a. u

1 | I t * o tn m * u-» o o « h- *■ o »L « OGOOO^OOOf- 0«-«-^-T'm-HPgrt)if\. - o c» o o o <t> o

OOOOOOO'pOOON** « <f N 01 N "(

I O 'J o o o o o qt O - O f>

-: ir* o M « m IT« iA^iOot">OOC»OOOCt)

■'■ O O J> O 1* -t »*> ♦• *- A* * ^ «> »I- *- ^ < o i p o o <r» o I " I ■»

lit'

Ö o

•H «

Id 4) ■H M-4 T3 «0

A 4J

A (0 ws u < M O 4J a» u n CM 3 55 W HNH

WS)'« .C 01

o« ^ CM M

J3 ß

O

«M 0) O "H vO

4J •

it" o -H a) w «o » •H o> a

ao ao

s 3,

•H

96

* ■- >Jt MMM

f

o ü>

> Z Mi I

»ft if* :l

o a » if» ■# ■

i*t c\ ' MT fi * mcir -I IM C ■*■<»»-

m m r

m m <■ -t O u

., mt» <■ m t-*t n# r

■4 ix ^ M A.I <SJ

O ff 4- n-, st ~ -' "*■ •* O ifi ',/• X O " - P» *>J (M (M 'M l\ i* ", O f - O f- .0 *

W iiM*i. i». t, "•*» ■*!■*■ i"0 -^ rf, •*> f*. "i *^ *\ ci ifl m ff1 'i i*j ( o o o o o n o o -o o ri o • > o *v* o r> o o o o ~» ft *> i l i i i t i i i i i . i i i i t i t • i i i • i l i i t i i

-. 0 ^ r» -j r-. j ■j -i -i m a- st -• «A ( n n ■JJ -i -n f\j r- ;; f~ »i -^ T> -** o -t o Lf\ 6 r _. " t

•o "" th f» is. ■"* ■i> 0 T> j- -J '■*■ 1 r- ». n •-' "•' ■*, if- 0 T T -' •NJ

A •*■ '■

G — f\ r r X -1 "J o

X * ■i ifi <*' ir> i/> r» iTi *r* ■D «o *> U -»> .n f» r- rs. r».

1 r- f^ f- 9 CO ■4 ■c -c U tT1 (> (T vr c> cr ■> - -"

o oo r- O O » + r- r» O -* m r- u> If - r«' «t i, \i\ cn h- O aj IA - -» - f>- X

1 6 -r o o o o oo

O O O O O O -* ■■•* O O O ■-* *M >f 'J X f ■ J- «J- ^ f. r- IT .' N rf^ <M -i M ü r, (J) M O 1' O O T

ft. O O r-i '1 O M Al fi H O M i J M ft( |n * «ö *• >A »4 X r* «i f* *•"» rsj fTi v4 sj O M O O O O O O

! -J.

c o •H *J «

•H •O

5 V x: tfl 4J u n »4-1

vt

c o 4J

•H c 4J a) « C •H D l-l M-l 1« «4-1

> Q

a) Ä 14-1

4J o 00 tfl C u

•H 0 4J 4-1 cfl »1 U •H u tfl (0 3 8

1-t u *-s 1-1 l-l

< 1 um ro n O ^^ a> II

ms M •z W < cs H

II Hi • M

Ä tfl *^N 4J •a K «A (14 DO u j: c **^ OsO

1 o <

Ul X IH 0) o •H

4-1 vO

1 •H >

U1

t-l •i-l II 60 4-1 o •H 0)

4-1 tfl tfl tn c 0

•H (U o BC w N-^

o\ 00

ID U 3 60

97

o o o *

o o.

— o* • •M

O I u I

I I I I

• -< 1 > O I tu I O u. i

I» I • • I

OK I ►- • I l/KM I

ui «• I

• I

o a) ui m

> a o — i

£4

o o*

o at

i I '

ft ft ft

*0 — «N

•*> I<^ *•* lAi«f Aj

m **> IM K O -« ■* *t> \f\ *\ r* m m *** ft

! 5

4MMÄ

» C *• X o * IM m in

9* m *■< < o- *

rg IM ^ < O « I

IM (M m I

*VfM i i**!l*l I f» O > IM OO I it IM i u> to t

INM

' g? w »

I« >i (M * I IM IM

» f* l-tf ' *— I*. t IV *M , ITS f* I o-o I IM,IM

4 IM

fg I

O tl> IM

IM It *M irt t/N CD g I- 01

MMN O fl* IM o vt «* IM IM IM

« * IM

f. -I * »- *,f» IM IM I o tnl r» rt'i IM N|IM «* IM'* ■o n-.-e IM IM;IM m to * «IM Mt IM«M IM IM* * «O IM,* IM IM IM -*ft I* ■n IM!* IM IM*M 0. ff. OT. « »i.in IM IM I'M tft 40,0 h- — -"" « M IM IM »M — «'* <0 MHM«

IM IM IM IM 4 Of- IM O —fM •« IM IM IM IM * 00'IM IM O O IM * IM IM IM IM

IMfMlMrgiMIMIMIMIMIMrMIMrglM^MIM'CglMfM'lMIMIMgM 00'OOOOOOIOOOOOCOOOOOOOOOOOOOOOoOO;0 00000|(

tf Ui UltJUi'ujttllJw UUU.UitLlb'U'UtU.lJUJUjUJUituU.UUJlJUJUJU |i> kü U U.' UJ U. U* UJ lu. "***"• ,UMtf*CC.»*»^>0 0 M^NOViHtffOKi'Wb'iNICf O *- C"i 01 ■* *** ff, vf — P*- rt ff- -O

«-« IN

I f I ili i ! oo oooooooooo*»»

1 • 1 i ' ' 1 1 i g> K tf\ m 43 >P IM IM "> 0" O »0 O U"t O O O O "O O O O O C ,0

oo o oo O OO'O O O O f tft

1 1 i» tf\ IM -> r~ o o o — — r- I- »- IM o o o o o o o o o «. a

! "*i""" i 1 1 i 1 ÜO o o o

1 I ooöioooorvff-

I i - 1 i 1 ! 1 ' 1

«MO' <*■ «Ni <M *- '* #■ ffl-jfft -* N -•.•# C.OOQOOCQOOO

1 \ r i i

Ho

I : • !

US

0 J2 CU 4J O (0 WD 0) •ri tnx: ^i SÖ Ä U H

o

01 u 60

•H

98

ä

o o o ■*

o t

o

11 11

jl l « • I to

i !i

o

4« II ijj m i x. t~ i

t i M I

I I I I

N I O I

Z I < iu I ;— <r- i O m I i) f- I r . i I

> i • i

O I

si

tu I I

|Ui UJ I l<LU i I

<y an o o.

o o

■4M. V r-'j* AA

r» c ■* «% 00 1 *• «Nj rs* -«N it <0 iA ,4H •■* r-< —» O O

«# §*£«• O O

_ _ O Du O O . *-* m ^ t|> *\ ■* f- QLF 'SJ **> t) n ** *♦ »£

a, o o *■ -J" •* «*"* ft* *## rw Mi ^ kT A A li^ 00 if* ■* f* «Ni

.-.^W «ÜB c> c o o o <">

o «of»

- K N V-Ö -Ö -C r- ■* «N« o o o-o o 0> H O N (^ r, i

iX j- « m -- o ■* o o c o o o o c O

* —i c> r* »A et A* >\ o r>t o O O O o o o o o o o

• •

OÜiO

( IM fsi rg »oolo IUJ Lu u.

— -. (Si

rsi f« o o'o

OJ N ^ '

o o' •oooooooocoo ili/U.iUiUUiU.U'UiVUtikUJ

o o o c o 10 lA * *** 4)

o o o o o rg ^ .-i -• CJ o o o ,o o

'"C IM (v rsi *v *M < o O O O O I

o r- » r-j rg o o •» o

m -« O o o o

l fst'P* I o o

oo o • •' • 0 0|0

o o o < • • • o oo i

T o o o i

i

I^H*«NM cgfs(fs*fNjf\i'ti<Nf\l**»

U. UJ lu U U Ultx.UL.UJiu

*•% t* r\ 4*\ r»> t*\ r\ r~> •* ,-t + -** ■■*

NNNI o o o <

u. u/ui A r- "1 ry »O •#

I I :

o « ■ • ■ » • •! • •

iFiUt

o o b o ^ «■

o * o o O O '

c^^tfvir»*- rt- ft* ** «o <** !

UJUJU. -^ <-• IN* m * r- ■ • i •

.1

i

Hl IM Ä O

o

Ö

w

T3

H w

Mm 0) o

o

01 o X • u m

1 @ t! *J o «d

3 B vi HU«

O SH

■H T3 > W 0) d! In

^! Q Xi U (i T3 •

* «J 2 « M o> qrt 2j fcC es

. o • (3 •H'-J •HtJfl rt CD cö O H U

0) 00 •HMO rt U 4J

•H O t0 V O -H •H Ä Ö W H tn^-v

J3 0) 4-1 10 o to o a Bv rt M u tu 0 J3 4J O 10 Ei] 0)

•H fnJ3

H

60

99

o

tn -* et

O o o *

l/>

M

K • 4 O > O u

> o III OIU

t|Q<n

I I i i l l i l I Zu I

o

4 111 — * 111 * X 5

•r

v> 1 in lk

1 Sü*

s ! »» i »- ^ • a Z « 1 • IK | ■o f* t z

•3

##

♦ 5«

^▼tn »A

■ MAmi£» IHo I or

, WT * » ■»

ft iff 5 0 ^ IT f ** c •sMc mCBw< #▼< •»•▼*

»■l^ ■r or o -• j ( «MB« m M *M Til •* * * <▼

^^k -0 f^ c 6- n. DB rr» fr. IM rn ^

IA m

mKn I «MiA ) -»,«1 i iA m

-W-a-\*

••!<• in in OM •» O tA IA — N O 0- lA * O l\< o in «1 ■#

9> IA >» * o -o r > 4- ■» «D IA • .*

IM * IA .«n 00 o * «4 ^ #n *MA IA -« rg • — * «* •n iA an «o ^ * O fl <M IA IA IA * e •»> H -0, a <A O -C -C IAA .«• IA * # IA »1 m in -• o *-,« * * •» ■» «A m 0> ^ * * 41 K * # lAlA lAlA

«> m n» m «n

i

oato OOlO

I tug. w 00 « «5 •> M *

oo'o ■—J '—L — L--L--L OOOOOC 3 0 0000000000000000000

(rCiMm — actAfNiCrf*- *»<(OiOriON^-ip <c »** o f- »A *% &■ . -*&,&Wtf*,**Qt<\it\tD-<i+ OCX wir f- O l« «^ ao —* -# «0 &■ *v •* 1

f

0» in

Si IA. «"I ■r

mi* an Ian

O O O O O.O O Op J U' li. If U> faJMJ ■ O IA fx w f- j- \ IA «? — m -e *

i«1«**»''-*r"«,«»iN« <r *< -* r* n* Aj '*M M i

O ^ ►• O ►■ * ■# IM I

1<«sf\r'W**,«*iAIA'lO«0^.CJ>^>-?-aDflO«OOOffff'(7'

I I I I I o r» o • ■ •

O -i o

i oe> o — —

«OK

-• O.O

OflJCOOO O IM o o o r

i "i "■ _r- i I i T o * — c o ,o

tie • -«

ll

a) <** J= o

<«-i -H T> o <d a)

C VAO «I mo 4J > rH X 0

a) 4i co

Ü «H

a) s 4J u o co cd U U fc CT» 4J i-l

iH cd l-l CJ*-v

■« o W 0 0) H

•H > w Q ^3 T3 « 01 Oi ^ O

c •

cd o «1 H fa

W O « wai

CM

01 u 3 60

100

X

... ^ 3. 1

a. 0

I?

L . .

; ,i -" - A i I i i i

II •j,. _

••• - >•

03-:t fT .* V f4 IT O

34 » -

■ I , , 1 '•' f '• " i- * ' .f ;i 4 i-

l' 7 'J T

i '7 4 r- ~ i i i i ' • »

:v.:

3 J

■- - i i -i ? i i I I I » 1 - 1 •> -1 r- --} 7 .-. ? • ... -j ,_- l I t I l l l I I I t l ft l

^-3-0- O * -'■ '.r - '.' ■ ' • - - 1) " t ..--'» "JO -• ' '. .* J f if u .

> \ • r; r- 4 5 J ,.• : ■1 ;o.ij

- O * O O ■} 1

<D J2 u

a) IM , £ o u 1» ! LM c o o

•H r-t *J •H « CO 4J H . fN 1-1 CO

^ ai OS .J p. n t) tom X o 4-> a) 1-1

fl) U X V CO rt n K o 4J V4 i CO f* 9 CD

iH 0) CO i-l a o M

o o

O X. 4J CO 00

 «t

u <u o

1 p

4) CO O Q

O .J

^ I

TJ 53 • CD <S PCW v~'

u 00 o

00 4J Ü ED ■3 -rl c « 0) 0) l-l u

crt

PR

< a «t 0) t-l CO 00 o O P

CO

<D CO O P

A 00 M «

PI

cu u 00

•H

»I

.1

101

1

a. *

■n

O o O 4

5 ! ••* t i • •>

>nl o Ü

HI I I

• m i

ZU[| « * ll ttl f» .1 ■ * w

• I

• .i

,?! - O « OO 1

**!

r IU

I IV»

> t « ■

25

IA O !>• M z

I • ■

K at H uiiu •

* f

! I

«i

in N I 4 I

•n W W

♦r» «A tf» A

I

m «*>

m m i»> O «• tf> •#•40

P. V

. MN K <o e

w* •«. r*

i Ä ■o

4> >

t

»*» H|«V

in ^ r- fa i I«» rv IV fi i i« ~ • «I I o

I w lf> I Pi iNi I 4- , <> i 4) IO

i •» m

0£i «M ift I»« 1*1

•** IM BOX 4 I»

kri

4 •0 *>r-

#kuN a> <o 4 * ■Bo «n u- e o

I WWio <M iv «v *♦> ♦ «• ■« IO 'Jb« "^JV

Mfcniä«ä fn*>o»*Ci**i*- «■■■'' (4n"in »i^fN NWWIT'I N IM »V IV IN IM f **

* «■ m • o p o o r

utuuiwluu * ■>•« r- «-• «n o tr &> *T> r? *•

............ to «r> <o r «r o» * c ■ « k «r «> * n• f. K poooooooooooooop

tf> in »V ! * «X M

r»A»

•*> <r M -I J3 ■»» IV M fM r- 4) 4 -• »<fN| fM <M M

»o 4i S MM

«>*» O I»

MN en« < In I» NNN r- »N O < ♦ f M M fM « 4 B r- * * IT *■. IJ fjf'NI

•t p » r- »\ N(M«NM

I <M o

r>o o SI IV IM

I» o t # Will O 4

ooooooooo " 1 I t r) I I

lü ui in IAI UJ üi u* III Ui 0'i«lf-t\i4)&«»ff'in

IV ft

?? tu IU r~ M c o

IM *4 *4 -I H I 1 I I I

e p .o• i» "■ 1 -^ *< *4 M I I I I I

JllMHrtMH

lOOOOOOOOO • T I i I i I .1 i

i ut IU |;j 111 u: U' iu UJ ui iw3ffMlfttiPimri-i

tr IT o o o o p •• -« • i i i i i i

Ul IU III IU (L U» UJ 4 f- O »«- « N O

5 tr 4 O

I ill UJ 1 N ' CM

0)

4J 0)

O 4J

10 <M

g° •H r-l ^> *i -rl TJ rt d oi uH h •H a pvo

•H 0) O l-J &rH

O. 4) 3 X

W tlvO

0) «J iO

OOI-o'-»iO«t*<llKl

I I i.»«0r^K4>4>iAi>^« mmfsllN«rt^4 p-io

» I ,1 I I

ODlvO>INN|i4KI«ll\r^N|tMP'4t'nf< 'ONOiil

i ! i *• m ~ i« •

I I I I ,1 I I I I I ,1 I I I I

i^o»ifiM*i^r-oooooh»i~

i p o o OOOOinOpo

i o o -» «()n(iiii«Niii4>i-4e'*>xopiii«H4Nli4t)|>aafi«>iio «IOMMOH «n fl) O

•a is » 4) CN PW *-'

r-l M <3 9) a

0) CO

o Q

I O

M 00 • o oo u a «

•H -H Ö 33 0) 0) M »4 M U <

CO

a cu U (0 oo o O Q U (0 > •H O

0) M 3 oo

01 (0 O Q

JC 00 ■H as

102

>

tt*.

r

!• >^

> u O a/

O <A ,*— •

1 i 11 i •» m

i in

i - o s

4 «J » O . 4 «u k -O

« .AI o

1» 'i U •

»4 i • *rt k V» ,i w t 4 * t- i> a

l/l ! 5 « i ■> • 4 U

«1

. 5 IM

ft , *i X i . . Ul a Si 1

;A *-•

*\ IA .-» « •i J «A rf* r- * IA vA

«A *?*

m »A •A r> ■A tn <e\ +~» u. <n iA IA u» «.' ■* M •A f/l <#VfsJ Pi r-* «A tf> 4 4^ *• O ** «x iA tA

1<1 -■* -O ifl .-»f » »A •# .. j o »A m

*-. fi»-»<r *-i iD u> uW lA t.-> « «i <r v

IA »**^

o o tA -A

«n O .t «A

to o <4 if>

3(?- i^/ ^ w r- «-r \T ./i «-« 4 -* tf»

, m* f>J r> > v oo *-• t t •# rf» ■ 4 v ^

O O o * »f »A »»> r- »A r* CD 4> N (^ ■#**■•# tf\

O -a <T -O tf\4l h 0 %1 J- * f *A ** ** <9

to ■?*

4- in o <o f-

I«

00 IP 0 o

1 I Ui LU *> o

m BC #n

III UJUiU.

* O *

I O

-I >» -r J •* "I ° J v * •» *^ #0> I«

Of» * 4)

> OA«

to 10 0 00 o o p o III»

artcpi^cgnir«f-^ 0C0O0Oo0Of>O0f)03O 4 111 I I I I I I I I i I " bj Ui Ui »J tu u: Ul iu Ji Hi I.U ui Ui ill ^HOQP>«ll-N/IP^ ♦ f" rt I idHhMliJOi'H /O*

•F

it» OK i L ' lOl-M

t* r- f- f- r* 3 O 3 O O C p

I I I i» I T ui Ui Ui u: Ui ui Ui W ?» 4 l> «I M » O O -« N 1 I" *t

r- t- K t- O o b o i i V i r ' ' Ui Ui Ui

.to* to «

»rîA-4^rtrgf^f4-H ^Hf-.^Hfum*« â^-o^r-^cog-v- ^ #-««-*-* *-i^4--«ri-i^(r-i

',' • ' ? ' • F' ' i ! • I i ■ I : i r 1 NOlMn^1! •*• WO«

v> « * K

I ! r-opt-oj~»t-i-o--pocbi

OpOOOOoPOU*Omt-IA>flJ0'4riOOpOOOO3~ — Op3opOJ0O j I I ♦*--- I I I I

ipifiHQOÔMNfifiOÔonbO ip>iOii00p-<H«o^ri^rUMtNii0^ri>iobôtiN*iriO'<oonbi

I

0) u 00

103

< J > ^ i c 1

O I Z I « "• — h- < ar- i .u "t i x

ift|P«MAr»-f'*<rf*«rtl*innrtPflCvNNfyNN^»,HM^-4« - f. « f <«\ »4

at Z UJ IU a o

i i i i • i i i

O O «9O4'0i}'Of l^rv«

... 0600000i]iyO*N«(l-

! ! ► w4 ** 6 »O -4 I

I S I p f I

«* >!* r» »A i - *[> CD p «v 9-

(•AnnOi)OoöoonO

> o oo oo > > o o 6oo < >

8 •rt 4.3 «t •rl U

> «I

! u cd M U » 3

a w fl o u

o a «5

«d •rl

a n M « 60 O 00 w a a

a» n 3,

•H

10A

O I t I

Zui I < O I UN 4 X *n I

• i

o > •» • o i ° 1 ° 12 11 ■*

1 4U>4 1 — "J 1 i o<M E 1 ul tn :t

K l

O O |

-• <o o o» o*l IN I

cr CO I «M rg ■ «•« co ■ -0 r» ■ >( r* ;

lA I CO

«4 rg ^ -J B3 0- tr ? J>Ö

! £

O 'M O r-

a *» M rg fg 'SJ 4 «C 00 «1 fi o rg rg rg

o c

r» rg » ^ o D i». ri —• O rf r- eo m r> rg eg -^ .f <M rg r- ^ r- rg M ~* rg ■f» 00 •O (T -O rg —' ~«

« n x. i i> b o i O- ^ m ; h- «C % < C'J fg 'M « *-* « o ■ r*- 02 kT- rg r\. 'V • ö JT O

rg rg rg ( -•NO- sf !if\ PH • (M (MN I O «C» o ■ *-4 »f\ rg t rg fg rg t tf\ r*\ rg i O Of -« . ig IM rg '

ONO: oj rgN i or IT -• f tr o a . -« rg ~< «

rg rg «n .-< r^ t> tr> r- to rg rg &

/ CD O (O 5 gr r- r- KM^N • r- r- tn > ■*■ * <o

: ■* o «o \ ■# O <n

<•*«> i ix »A rg * rg (v N t -^ -O yO

J rg rg fg ^ r- <7> r~ j •-* f> O JrgNH

HIMO

I rg rg #-i

' o -* to

»<rr-eo.or-r-0>

_i -NJ r-< rr» (T O •-» <0 .0 r* r- O1

)OOOOC-*000©OoOOOC,OOno'300©t>OOOOOOOOOOo60 i i i i i ,i i I i I i t I i i t i i t i < • i Jf t t i t t If r • ■ ' ' "

JujUJbJiuuliJujuujiuuJUj.uuuu-uti.'iiiaiUJUJui^jiuuJUJiiriiJUJUJUJUiujiiJiuiu ■^Cl/^lfl•T■flN^-'0(^(t^■■o•cl^J"'N^•-oo••t>0!D^^o■o«^■lJoaoOHfim iff rf\ Hlf* !»*.■(> fl^N <Vi««f O O N !r ■* p *NR fl « K f * tf O 4 JB;N ^H*» «N

rg oo (r

r« *4 r< ■Or- ^ -* -* +4

t I

o p ( r'

>f- r* r- ;i i i

4 O'Ain m 4 J (f, m h pj N H i i i i i i i I I ,i I i n

i a f»

ooooobo-if'Mairii^ - i r- r- .c O N r- o

OoOOOpO^<-trgi>-r>'DX' <c H * * O O h >

o o o o o <

J--Ü i r* (M to <*> (** fr -♦

• fM^oooooooopo

fim-tMoobooMO^bo

01 0) •H AJ •H >

•H 10 c cu

CO

c o

4J

Ö o

•H 4J /—s « 10

•H T3 U n> a) u >

vD (U O

■d -H 4J

X M (3 vO

■H • 4J m (fl V4 i: 4J CO 0) 3 CD H O J Q M •>—'

pace M <X < 1 M-l P. 0 o

% cu

u 4J t)0 0 00 u a (0 o

ON

cu n 3 60

•H

105

i ii ML

i

\ • ,1

X C 1

I

I ? t I I «J tw> 'I I »- *» I l 3 « ■ I w s I

of Z Ol in a. a

i (

r J * *■■ c x r» a r>- x> r< KC.

IS. *, &<* -g r t b ~

-.»«»; : -j r» p ■

-. "i «% ■ ■ r- 'j Vi i

• CO ; =0 »

X J1

f*> rsi

r* u rv (\.

: -M

j> r-

-n o Ai tf\ ■1 f 1

»»»«or-»!

O -*

4 " -■ w" .i •* ir. f?> r»

*M <M

-I

o ; ->i » •;

j o f I A4fJf

) .-I ■*.. -v, fj a-t (*M ^4 <% f -r - o ff N ■*» **

> o i *j — t o -•

fM (g • ox , •o ^ ■

-i c <A * I *M «M ; r» 41 t + r4\ OJ *SJ <

: r- (M ■-- « J) i ■ i^ a -. f*. 0.

'-" f* f .' /■ o ■* »

) «g —t

1 i f-

1» r-j if»

3 ft O C 0OOt>OOO0OpOCC00000t:030O0030C ooooopo

UJ Hi UJ UJ U. UJ'i: 0i LU rj UJ Ijl U' VJ UJ U' u. X' ui u. U' U' JJ At JJ a» tu nj <jj UJ j; x< ui UJ JJ ui ui Ik' LU oc^MTf^ff tfN^-r»'*\oöc*kr,'^r»- omin^nr'V'-tOOc-T*'^ ■CiP'£ff'*»4ir<NS'0

i • [ i I !

V

u 00 a o 4 Ö o

u (d >

6C

at u 4J

3

<d

vO O

X

vO

m ^^ PQ

(4-1

o

H 0) t>0 O.

31 co •H P. 50 O

00

<0 U 3 00

IN«MN3AQ^HO0ON60

106

J^

Table 77

Rank Correlation Coefficients for Total Dose Damage Prediction (Overview)

X vs. Y (a) 2N709 2N930 1

2N2905A

(1/f) Noise 0,3 0.2 0.3 |

I°vs. IB(dose)(b) 0.5-0.94 0.33-C.7 0.96-0.97

i° vs. i n° (h)

B B' B • (0.6-0.4) -(0.3-0.2) 0.8-0.7

B VS' ' V e) E ' 0.2-0.65 0.2-0.4 0.0-0.7

1/1° vs. U/I°-l/IB)(b)

AT (low ir.jection) vs. :.ID (high injection)

o

Ü.91-0.94 0.97-0.8 0.98-0.96

0.96-C.55-0.2,'e:' 0.98-0.90 0.9p-0.8Ö-0.55v=/

hFEO VS- hFE (d°Se) 0.56-0.80-0.95(e) 0.36-0.49-0. bS^' 0.97-0.97

hFE0 VS- WhFE (d0S6) 0.91-0.94 0.97-0.83 0.96-0.95

!'FEO VS- hFE/hFEO -(C.6-0.4) -(0.3-0.2) -(ü.8-0.7)

hFEO VS- ihFE/hFEO 0.5-0.6 0.3-0.2 0.8-0.7

i/hFEü vS. :(i/hFE) 0.2-0.7 -(0.4-0.2) -(0.75-0.6)

.0 (c) B

-0.5 -(0.6-0.55) 0.7-0.85

To ,-.(d,.c) lB (O -0.4 -0.4 0.7-0. 8

AlD (due to \T)(c) -0.4 -0.4-0 0.75-0.8

Alu (due Lo burn-in)^c^ b

u -0.3to0.3 -(0.3-0.2)

^BO <" BVEBO> -0.4 -0.3 0.6-0.75

"EBO (T) (°r K*J -0.4 -0.3 0.6-0.75

ÄIEB0 (due t, ■"., (or ,.BVEB0) -0.4 0.2 0.6-0.7

&lrnr (duf to burn-in) to1.1 0 —

2X930 (<!■"> *°** — 0.3 Pre- T° bu™~in( (T)

-0.4

— -0.4

(a)Unless otherwise noted, the Y parameter is Radiation Sensitivity change or the relative change in the low injection Ig.

(b)All Ig values w«re taken at relatively high injection .. -el. (c)All Ig values were taken at low injection level. (d)"T" indicate? ether than room temperature. (e)These multiple entries represent variation with emitter current.

correspond to the direction of increasing current.

the

Arrows

107

0\ o 55 CM

01 C «i

CO

a o 4J tfl

■H

iS*

CO n 4J to 3

O

o

U

in

*« rs.

> n r» SO *o CM r» CO rH • rH r^ en

8 r* CO co r» IT» CM <t CM 00 ■«» CO rH

•o id u

so r» vO «o H CM CM O 1 1 I 1 I 1 i-l o o o

rH O rH

o I-H

O

X X X X X X X

so VO o SO «n • m i-l «O H 00 vO CM CO «n rH r*. 00 m • • • « • • • • • • • •

CM H co CM

•3» CM o O p rH co rH

»•« o» a\ r-. 00 o> VO O oo o» • ST co rt

(0 > r^ so si- r*. m CM «* !-l Ul so >* CM •o o tt u

oo f* m N CM CM sQ I ! 1 1 • 1 o o c c o '© O iH rH rH rH H rH r^

X X X X rH «n o X X X

co lO IT» (»» r-» o oo «» m r^ r-» 00 3 • • • * • • • • • « • • • rH SO m 1-1 00 in rH o o o -» rH rH

x: •IT» so r^ NO o CO Ch m • SO • •

to > vO co •* CO CM r^ t-\ ao CM SO •* SO ■o o rH «n «* CO « CJ h

m Ov oo \o r> co CO o 1 I 1 I I 1 iH O o o o © o X

H rH rH rH rH rH

X X X X X X o m <t <t rH «n # <t rH o •» vO CM r^ ao CO CO CO CO CO • • • • * • ■ • • • • •

CO 00 «n rH ri i-l O o o o* m CO

w •xs a) M

m o H 1 1 I 1 1 1 I 1 I 1 1 1 X 1

1 1 1 1 i i 1 i

1 1 1

I I

1 1

i I 1

• H

> IT) <

p. 1 > m H H > M M H H H H H H

3B SI W1 4 3B H, *•"* ! o w CO V «1 o rF m

< 1 m M 1 M A rH

108

|s»

O co o\ z CM

o

>> 4J •T-\

P ■rl 4J •H tn C <a in

c o •rl u cd

•H

cd Bä

0)

u

tn 3

o 4J

(0 Q

e

»4

> o o

co iH rH rH

O CM CO

rH CM rH

o rH

VO i-H

CO rH

O rH

00 vO rH

vO rH

o CM

01 TJ (9 IH

vO O rH

X

vC

IT

r I ! o i-H

X

CO

St

1 o rH

X m CM

\o 1 o rH

X

r-

CO

IT> 1 o rH

X

CM

a»

rs

m CM

CM

CM CO

St CM

CO

m o ■-H

o

CT.

rH

o

rH CO

o St

O

CM I o i-H

X

a\ CO

CM

I O rH

X

vO

rH

CO I o i-H

X

<t

'/) J

OS H

vO O t-H

X

co

i-H

^5

> o o

i-H o vO

CO CO 00 00 rH

CM rH

00 rH

IT) rH

CM rH

rH rH

o CM

o CM

vO CO

1 o iH

X

ci

ro

1 O i-l

X

o CM

VD 1 o

X

o ro

1 o rH

X

00

vO

o o CM

Oi

ro CM rv

r--

CM

CO rH

o

CO CM

C

CO

o

CO

O

CM I O <-i

X

CO

CM

I o rH

X

CO

CO

O

X a*

vO

U5

X)

B-S

> O

O st

rH st co St st

r-i CO

rH CM

st CM rH

CM rH

VD ro

VO CO

o St

CO u

O i-H

X

o CO

1 o i-H

X

ro

CM

1 O iH

X

sf

vO 1 o rH

X

rH

CM

IT) 1 o rH

X

in o m rH

CO

CO iH

00

m

i-H

CM o

o CO

o vi-

es

UO VD

o

CM

O r-i

X

CM

CM

CO I o i-H

X rv

00

CO I o i-H

X

vD

St

K

rti

> Ü

vO ro *0 IT)

m vO

tr '•J

lT>

st CM CM

st CO CM

CM

CM

CM rH

CO vO

VD u")

CO in

U

o I-l

X

o t-H

DO

1 o i-H

X

<x>

00 1 O i-H

X

st

st

1 o I-H

X

r-.

CO

m i o rH

X

U">

CM

m VO

o CO

vO

rH

VD CO

o

o m o

vO

o

o CO

o

CO

O i—i

X

o%

' 00

CO I o rH

X

CO

St

ro I o rH

X

VO

CM

> m st

o

II

tu PO >

<

i-H

II

w

<

o o rH

II

w H

1 O rH

II

w

> st

o

u w «

>

<

rH

II

W M

<

o o rH

II

w H

t c

II

w

< 3.

o o

II

w H

H

II

W H

1 o rl

II

u

1 o

II

w

<

o o rH

II

w H

t r-l

II

w M

r-i

II

W M

3B2I HH 9I 3E ^q 3E ^\\

CO

PQ P. PQ 1 O PC M ! H

i

, o ti r-l

t-H >_•

109

o oo

< in O o\ CM a CM

IM 0

■H 01 C 0)

i/>

c o

5 0)

X.

■s 0)

«0

4J 01 3

O 4-1

« 4-1 Cfl O 14-1 o

3

(0

> O o

CO I-H

m H

o rH CM rH

• m

O» 1^

m

CO

CM i-H rH

rH rH

st

oo

•a at u

<o o rH

X

vO

m

1 o rH

X

CO

CM 1

1 O rH

X

oo

iH 1

st 1 o rH

X

o <M

CO

st

VO

m rH

rH m

o

CO m

o

vO

O

• CO

1 o i-H

X

st

CO

o i-H

X

r-s

vO

CM 1 O i-i

X

o

rH

m

cd

> 8

i-l CM i-H

CO rH

r-s CO rH

Ov

st

in

m

rH

m

Ok

m rH

CO rH

CO rH

u

so o iH

X

CO

rH

oo I o 1-1

X

CO

VO 1

00 1 o rH

X

CO

m

St 1 o rH

X

CM

H 1

a» rH

st

CM

vO

rH

r-i is.

o

»t VO

O

o 00

o

CO 1 o

X

CM

CO

CO 1 o rH

X

CM

st

CO 1 O i-H

X

m

m

to

cd

> o <M i-l

vO CO rH

rH l-H

St

m

St

st

st

CM

m

st

m

rH i-H CO i-H

vO CO

u IT» o iH

X

o

CO

1 o i-H

X

rH

m I

oo I o t-H

X

m

i-H 1

m I o rH

X

rH

o\ 1

rH

rH

St

rH

st

rH

i rs. 00

o

rH

o o

CO 1 o rH

X

CM

i-H

CO 1 o t-i

X

rH

CO

CO I c H

X

o

CM

0> •o cd

> 8

o m CM

o oo H

■ CO

CO

CM

1 1

i I I

1 1 1

H st

00 rH

CO st

u

m o iH

X

o

i o rH

X

00

H

o H

X

r^

m I

in I o

X VO

• m

I

oo

■

o

m rH

m CM

rH

1 .. 1

1 . •

I I

1 1 1

st 1 O i-H

X

oo

CO

CO 1 o rH

X

rH

st 1 O i-H

X

m

> St • O

H

W M >

< 3.

CO

II

W M

 m sr

©

N

W

< 3.

CO

II

w M

i o CO

II

w H

* CO

II

w H

© co

II

w H

1 o o CO

II

w w

i CO

II

W H

i o CO

II

M H

1 o o CO

:i

w

3B «i 38 flj 5B Mq 3B ^

PQ 01 M Q, - I

(4 M

O M H

K rH

110

_3^

Table 81

» »

Results of Rank Correlations Between Radiation Sensitivity (*,ow Dose) and Radiation Sensitivity (High Dose)

Device Type

Bias Condition for IB

Radiation Sensitivity: ID(irradiated)- I

Correlation Coefficient

Low Dose rad(Si)

High Dose rad(Si)

2N709 VBE = 0.45V

I = 3 vk

. VVE= 0.45V

IE - 3 ,A

1.7 x 103

1.7 x 105

3.0 x 105

3.0 x 105

1.3 x 106

1.3 x 106

1.3 x 106

1.3 x 106

0.678

0.663

0.893

0.821

2N930 VEH - 0.45V

I,, = 1 uA

VBE = °'45V

IE -lyA.

3.0 x 104

3.0 x 104

1.0 x 105

1.0 x 105

3.0 x 105

3.0 x 105

3.0 x 10°

3.0 x 10J

0.699

0.639

0.752

0.726

2N2905A VBE = 0.45V

lE * 3 Vk

VBE = 0.45V

IE - 3 yA

1.7 x 105

1.7 x 105

3.0 x ]05

3.0 x 105

5.6 x 106

5.6 x 106

5.6 x 106

5.6 x 106

0.329

0.675

0.296

0.681

k •

111

CM 00

■a

4)

o 9)

3 I»

a» .G 4J

u 0

•o <u u o u

cu 0)

4J

00 R

•H

u Cd p

o CO ON 5= CM

3

<*N « l-l a B

OH m o en

to CO CO o o CM rH rH rH CM NO •* CM CM CO CO CO C

o O CM ON r. 10 CO CO CO «» « ■» NO CM 00 00 00 c ts-a «» ■*"*-*•* CO NO vO vO CO CO IT» CM CO CO CO CO -4

x o H

/~\ a> rH a

CO O »» oo m r> «a- <t m o to • • • • • • • • • rH •J N N PI O CO i-l iH r-l IT» r- rH CO vO r* CM ON rH CM rH CM iH rH CM CM

85 o H

to CO » f» -^ u vO vO vo r» r>» *r\ ON a\ o» m m oo co •* m >n m vo m o

X to

as CO co ^ in irt O co co co ON o co m vo rsi r^ r>» o

rH rH

to 

Q rH rH CO r4 CO rH rH CM CM

o —i

Six X i-H CM CO ^ VI r-t CM CO «» rH CM CO rH CM CO «a- m vo co s

*~N /"^ > < ^N to a 1 O • 3 8 (0 -H

is u o i-4

o rH

TH -H II « T3 1 1 1 c w

O « W M w o > M M M \** \^ %• <*• n « « « w M M M

112

CO 00

■8 H

4-1 o

tn •u

d 0) U CM

•a CM

o

a» a)

4-> U o

60 in c

•H a' 4J m to o u o 4-1 tn 3 3 o

to u to ß

u <0

I 3

10

/~\ at

i-H a. S

«-i (0 m CM CM r~ r-. ON r-* r~ -* <T m o ** 0\ n GO

in o o CM -* r- p~ r^ o o <r vO o\ r>- B* f-1 CM r-~ r~- CM -* 00 rH rH <f «a- r~ vO 00 ON *s n)

* o H

^ QJ

rH

1 n-i to 00 MJ vt r^ CO CO m O co in r^- m co vO O Wl

CM O oo in rH vO CO in rH CO r^ CO rH in S-« i-4 -H i-l rH rH rH I-H rH r-A r-1 v—' tO

z « 0 H

(A •o tn to 0! n U

•H S vC Oi ON CM co <r in m r-. r-» m CO vO 00 m > co a» ON CO VO CM CM CM m m o ON CM CO o HI r-\ rH rH <-{ iH Q

X r^ tn • S CU v£> o o ON CM 00 a« ON CO CO 00 CM VO 00 iH o

P4 1-1 1-H rH ^ ^i rH rH i-H <-* i-H

tn a)

8) o •a ■H Z •* m \o 00 KO CO m rv. vO ON m m NO CM to > i-H CM rH CM i-H i-H CM -H CM v> 11

Q

o i-H

tf tn S 05 CM CO <J ~* KO t^- CO m \0 r- 00 CO <r in

m O

s-\ /-s

t •—\ *•> > in

o 1 < "I

<* (0 o o C CO rH CO

0 n 0) -H !! II II tO *■' W •H -i< W w W pq W t M M H >

c *—' v«* >*/ ^-* o ra PQ CQ pq o H H M

113

Table 34

Summary of Data Illustrating the Procedure and the Results of the Low Dose Screening (2N709)

Bias Condition

Rows *R

De- vices N

N(% of Total Sample)

Rows *P

De- vices M

M(% of Total Sample)

IB at 1 2 1.5 3 3 2.3

I_. - 10 mA E

2 4 3.0 6 9 6.8

5 7 5.3 7 11 8.3

6 11 8.3 10 33 24.8

7 15 11.3 10 33 24.8

8 23 17.3 12 47 35.3

XBat 1 1 0.75 1 7 5.3

T_ - 100 uA 3 4 3.0 1 7 5.3

4 7 5.3 3 9 6.8

7 11 8.3 9 22 16.5

9 14 10.5 9 22 16.5

11 20 15.0 12 55 41.3

XBat 1 1 0.75 1 7 5.3

IE «3yA 2 2 1.5 2 8 6.0

3 5 3.8 9 17 12.8

7 10 7.5 9 17 12.8

9 17 12.8 13 41 30.8

11 23 17.3 15 84 63.2

'»

*In these histograms, all devices with good data, but such large AI as to fall outside the histogram plot, were called row 1.

114

-*r-

Table 85

Rank Correlation Coefficients for Damage Prediction at 5.6 x 106 rads - uA744

Change in Bias Current».AI vs.: Rank

Correlation

'B -0.061

low frequency noise current 0.091

1° (22°C) - 1° (-50°C) 0.112

1° (75°C) - Ig (22°C) 0.157

1° (75°C) - 1° (-50°C) 0.190

1° (-50°C) -0.099

1° (75°C) 0.082

Relative Change in Bias Current, AI_/I° vs.:

S C. 546

low frequency noise current -0.316

1° (22°C) - 1° (-50°C) -0.403

1° (75°C) - 1° (22°C) -0.486

1° (75°C) - 1° (-50°C) -0.492

Ig (-50°C) 0.463

Ig (75°C) 0.237

115

Table 86

Summary of Data Illustrating the Procedure and the Results of the Low Dose Screening for the uA744 Op Amp

Rows R

Devices N

N« of Total Sample)

Rows P

Devices M

M(% of Total Sample)

1 2 1.4 13 56 39.7

2 5 3.5 14 74 52.5

3 9 6.4 14 74 52.5

4 14 9.9 15 106 75.2

5 20 14.2 15 106 75.2

116

-1_

REFERENCES

1. Sawyer, D.E., "Prevalent Error Sources in Tra; Delay- Time Measurements", IEEE Trans. On Nucl. Sei. NS-19, No. 6, 121 (1972)

2. Frank, M,, and Larin, F., "Effect of Operating Conditions and Transistoi Parameters on Gain Degradation", IEEE Trans, on Nucl. Sei., NS-12, 126 (1965)

3. Ramsey, C.E., Vail, P.J., "Current Dependence of the Neutron Damage Factor", IEEE Trans on Nucl. Sei. NS-17, No. 6, 310 (1970)

4. Gwyn, C.W., and Gregory, B.L., "Designing Ultrahard Bipolar Transistors", IEEE Trans, on Nucl. Sei. NS-18, No. 6, 340 (1971)

5. Rosenberg, C, Arimura, I., and Unwin, A.M., "Statistical Analysis of Neutron-Induced Gain Degradation of Power Transistors", IEEE Trans, on Nucl. Sei. NS-17, No. 6, 160 (1970).

6. Niehous, D., Frank, M., Stoll., E., and Swick, E., "Neutron Hardness Assurance for Power Transistors", IEEE Trans. On Nucl. Scie. NS-17, No. 6, 154 (1970). See also, "Neutron Screening Techniques for Hardened Bipolar Power Transistors', Bendix Corp. Technical Report, SAMS0 Contract NR F04701-69-C-0170, August, 1971.

7. Smith, D.M. and Hahn, L.A. "Neutron Hardness Assurance of Bipolar Transistors by Collector Resistance Technique", IEEE Trans. on Nucl. Sei. NS-19, No. 6, 125 (1972)

8. Larin, F., Radiation Effects in Semiconductor Devices, John Wiley & Sons (1968) p. 173

9. Shidd, W., Buchanan, B., and Dolan, R., "Radiation Effects on Junction Field Effect Transistors", IEEE Trans, on Nuc. Sei., NS-16, pp 87-95, Dec. 1969.

10. Buchanan, B., Dolan, R., and Roosild, S., "Comparison of Neutron Radiation Tolerance of Bipolar and Junction Field Effect Transistors", Proc. IEEE, Dec. 1967

11. Gregory, B.L. and Smits, F.M., "A Comparison of Radiation Tolerance of Field Effect and Bipolar Transistors", IEEE Trans, on Elec. Dev., May 1965 ~

117

Jk

REFERENCES (Continued)

12. Buchanan, B.L., Dolan, R.P. and Shedd, W.M. "Radiation Tolerance of Bipolar and Field Effect Transistors-as a Function of Lifetime and Doping", Trans, of Met. Soc. of AIME, 245, Mar. 1969.

13. George, W., "Radiation Hardened Junction Field Effect Transistors", AFCRL-70-0339, Scientific Report No. 2, Motorola, Inc., Prepared for AFCRL under contract No. F19628-68-C-0141, January 1970

14. Stein, H.J. and Gereth, R., "Introduction Rates of Electrically Active Defects in n- and p-Type Silicon By Electron and Neutron Irradiation', Jour. Appl. Phys, 39_ No. 6, 2890, May 1968.

15. Dacey, G.C., and Ross, I.M., "Unipolar 'Field-Effect' Transistor", Proc. IRE, Aug. 1953.

16. Graeme, Tobey and Huelsman (Editors), Operational Amplifiers, McGraw Hill, 1971.

17. Carr, E.A., "Simplified Techniques for Predicting TREE Responses", IEEE Trans, on Nucl. Sei., NS-12, Oct. 1965.

18. Sivo, L.L., Hughes, H.L., King, E.E., "Investigation of Radiation Induced Interface States Utilizing Gated-Bipolar and MOS Structures", IEEE Trans. Nucl. Sei.. NS-19, 313 (1972).

19. Thornton, C.G., and Simmons, CD., "A New High Current Mode of Transistor Operation", IRE Trans. On Electron Devices, Vol. ED-5, pp. 6-10, January 1968.

20. English, A.C., "Mesoplasmas and 'Second Breakdown' In-Silicon Junctions" Solid State Electronics, Vol. 6, pp. 511-521, September-October 1963

21. Portnoy, W.M. and Gamble, F.R., "Fine Structure and Electromagnetic Radiation in Second Breakdown", IEEE Transactions on Electron Devices ED-11, pp. 470-478, October 1964.

22. Scarlett, R.M., Shockley, W., and Haitz, R.H., "Thermal Instabilities and Hot Spots in Junction Transistors", Physics of Failure in Electronics (Edited by F.M. Goldberg and J. Vaccaro), pp. 194-203, Spartan Books Inc., Baltimore and Cleaver-Hume Press, London 1963.

23. Schafft, H.A., and French, J.C., "Breakdown Characteristics of Semiconductor Materials," Electro-Technology, Vol. 75, pp.77-82, June 1965

118

«Ji


24. Ford, G M., "Collector to Emitter Breakdown Related to Thermal Eunaway in Inhomogeneous Base Germanium Power Transistors", Solid State Design, Vol. 4, pp 29-36, June 1963

25. Budenstein, P.P., Pontius, D.H., and Smith, W.B., Second Breakdown and Damage in Semiconductor Devices, U.S. Army Missile Command"", Report No." RG-TR-72-15, April 1972.

26. Melchior, H., and Strutt, M. J. 0., "Secondary Breakdown in Transistors," Proc. IEEE (Correspondence), Vol. 52, pp. 439-440, April 1964

27 Egelkrout, D.W., "Second Breakdown and Radiation Effects in Semiconductor Junction Devices," The Boeing Company, D2-121704-1, December 1967.

28. English, A.C., and Power* H.M., "Mesoplasma Breakdown in Silicon Junctions", Proc. IEEE (Correspondence), Vol.51, p. 501, March 1963

29. Weitzsch, F., "Zur Theorie Des Zweiten Durchbruchs Bei Transistoren", Archiv Der Elektrischen Übertragung, Vol. 19, pp. 27-42, January 1965

30. Hakim, E.B., and Reich, B., "Effects of Neutron Radiation on Second Breakdown," IEEE Proceedings, (Correspondence) Vol. 52, p. 735, June 1964

32. Reich, B., Hakim, E.B., Hunter, E.T., "The Effects of Neutron Radiation on Second Breakdown and Thermal Behavior of Silicon Transistors", IEEE Transactions on Nuclear Science, NS-15, v. No. 6, Dec. 1968 '"

33. Caldwell, R.S, 28, 20 (1966)

Bowman, W.C., and Folsom, J.A., NEREM, Record

34. Carr, E.A., and Binder, D., IEEE Trans. Nucl. Sei., NS-16, No. 6, 120 (1969)

35. Tasca, D.M. and Peden, J,C, Feasibility Study of Developing A Non-Destructive Screening Procedure for Thermal Second Breakdown, General Electric Company Space Division Philadelphia, Pennsylvania, July 1971

119

■ ■ »A MM


36. Warner, R.M., Jr., Integrated Circuit Design Principles and Fabrication, Motorola, Inc, Semiconductor Products Division, McGraw Hill, 1965

37. Clark, L.E., and George, W.L., Current Crowding In Hardened Power Transistors, Motorola Semiconductor Products Division, Phoenix Arizona, (Paper presented at 1972 IEEE Meeting on Nuclear and Space Radiation Effects.

38. Reich, B., Hakim, E.B., "Bulk Reliability Effects in Semiconductor Devices, Current Crowding in Transistors", USAERDL Technical Report 2333, February 1963

39. Seminar on "Second Breakdown in Power Transistors", Marshall Space Flight Center, Huntsville Alabama, January IS, 1968, J. K. Morris, Chairman

40. Chang, Z.F., and Turner, C.R., "Characterization of Second Breakdown in Power Transistors", RCA Appl. Note SMA-21.

41. Buhanan, D., "Investigation of Current-Gain Temperature Dependence in Silicon Transistors", IEEE Transactions on Electron Devices, ED-16, No. 1 (January 1969)

42. Dixon, W.J., and Massy, F.J. Jr., Introduction to Statistical Analysis, McGraw Hill Third Ed. 1969

43. Private communication with Tom Auspech of Honeywell

44. Circuit Analysis of the X Switch, Honeywell Report No. WCCA-2-10, October, 1971

i t

120

distributed by: national technical information service u ... · national technical information...

Documents