ROBOTICS AND RADIATION HARDENING IN THE NUCLEAR INDUSTRY by VegasStreetProphet

VIEWS: 0 PAGES: 210

									ROBOTICS AND RADIATION HARDENING IN THE NUCLEAR INDUSTRY




                              By

                     LAURENT P. HOUSSAY




          A THESIS PRESENTED TO THE GRADUATE SCHOOL
      OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
            OF THE REQUIREMENTS FOR THE DEGREE OF
                       MASTER OF SCIENCE

                    UNIVERSITY OF FLORIDA


                             2000
 Copyright 2000

       by

Laurent P. Houssay
To my Parents
                               ACKNOWLEDGMENTS



       My sincere thanks go to the members of my supervisory committee, Chairman

Professor James S. Tulenko, Dr. G. Ronald Dalton and James L. Kurtz, for their guidance

and advice.

       Special thanks go to Mu Yang for his constant support. Profitable discussion and

dosimetry data from Dr. Georgi Georgiev are acknowledged and appreciated. Thanks

also go to the members of the Electrical Communication Lab, Jason M. Cowdery, Charles

Overman and Bruce Horton for their valuable help and advice. Corrections provided by

Denielle Caparco were very helpful and appreciated.

       The work on this thesis project was as a part of the U.S. Department of Energy

sponsored program in Applications of Robotics for Hazardous Environments. This

program supplied much of the funding to complete this work.




                                           iv
                                              TABLE OF CONTENTS


                                                                                                                             page

ACKNOWLEDGMENTS.................................................................................................. iv

LIST OF TABLES ........................................................................................................... viii

LIST OF FIGURES..............................................................................................................x

ABSTRACT...................................................................................................................... xii



1 - OPERATIONAL ENVIRONMENTS IN THE NUCLEAR INDUSTRY.....................1

   Fuel Fabrication .............................................................................................................. 1
   Reactor System Operation............................................................................................... 2
   Spent Fuel Handling and Storage In the Power Plant..................................................... 4
   Spent Fuel Disassembly and Waste Processing.............................................................. 4
   Waste Handling and Storage........................................................................................... 5
   Decontamination and Decommissioning. ....................................................................... 6


2 - USE OF ROBOTIC SYSTEMS IN THE NUCLEAR INDUSTRY ..............................9

   Need for Robotics Systems ............................................................................................. 9
   Mobile Robots for Routine Monitoring and Surveillance ............................................ 10
   Inspection and Maintenance of Reactor Components................................................... 13
     Inspection and Cleaning of a Typical Steam Generator............................................ 13
     Inspection of Reactor Vessel..................................................................................... 20
     Pipe Inspection.......................................................................................................... 24
     Underwater Inspection.............................................................................................. 25
   Handling and Processing of Waste ............................................................................... 28
   Decontamination and Decommissioning ...................................................................... 30
     Surface Cleanup ........................................................................................................ 30
     Tank Cleanup ............................................................................................................ 33
     Decommissioning...................................................................................................... 43
   Post Accident Operation ............................................................................................... 48




                                                                v
3 - RADIATION EFFECTS...............................................................................................51

   Definition and Units in Nuclear Engineering................................................................ 51
     Radioactivity ............................................................................................................. 51
     Activity...................................................................................................................... 51
     Decay constant, Mean-life, Half-life......................................................................... 52
     Energy ....................................................................................................................... 52
     Dosimetry.................................................................................................................. 52
   Types of Radiation and their Interaction....................................................................... 53
     Photons: Gamma and X-rays .................................................................................... 53
     Beta: Electron and Positron....................................................................................... 55
     Heavy Charged Particle ............................................................................................ 56
     Neutron...................................................................................................................... 57
     Conclusion................................................................................................................. 58
   Radiation Effects on Passive Elements ......................................................................... 59
     Inorganic Materials ................................................................................................... 59
     Organic Materials ..................................................................................................... 61
     Optical Material ........................................................................................................ 65
     Electronic and Electrical Components ...................................................................... 68
     Mechanical and Electromechanical Components ..................................................... 74
   Radiation Effects on Semiconductors........................................................................... 75
     Physical Effects on Semiconductors......................................................................... 76
     Technology Families................................................................................................. 80
     Discrete Components ................................................................................................ 96
     Optoelectronics ......................................................................................................... 99
     Digital Integrated Circuits....................................................................................... 102
     Analog Integrated Circuits ...................................................................................... 104


4 - RADIATION HARDENING TECHNIQUES ...........................................................107

   Definition of Failure.................................................................................................... 107
   Minimal Approach...................................................................................................... 108
     “Split” Technique.................................................................................................... 108
     Maintenance and Repairs........................................................................................ 109
     Shielding ................................................................................................................. 110
   Radiation Hardening Strategy..................................................................................... 112
     Modification of an Existing Design........................................................................ 112
     Innovative Design ................................................................................................... 113
   Radiation-hardened Components ................................................................................ 114
     Definition ................................................................................................................ 114
     Presentation of Radiation-hardened Components ................................................... 115
     Advantages and Limitations of Radiation-hardened Components.......................... 116
   Use of Commercial Off-The-Shelf Component.......................................................... 118
   Radiation Hardening Design Technique ..................................................................... 121
     Selection of Components ........................................................................................ 121
     Radiation Tolerant Design ...................................................................................... 122


                                                                 vi
      Annealing................................................................................................................ 123
      Biasing..................................................................................................................... 123


5 - RADIATION TESTING.............................................................................................125

   Radiation Types and Testing Facilities....................................................................... 125
     Cobalt-60................................................................................................................. 125
     Spent Fuel ............................................................................................................... 127
     X-ray Machines and Particle Accelerators ............................................................. 128
   Testing Conditions ...................................................................................................... 129
     Dose Rate ................................................................................................................ 129
     Temperature ............................................................................................................ 131
     Biasing..................................................................................................................... 132
     Other Parameters..................................................................................................... 133
   Dosimetry.................................................................................................................... 133
     Dose Rate Detectors................................................................................................ 133
     Total Dose Sensors.................................................................................................. 134
   Testing Procedure........................................................................................................ 136
     Number of Samples................................................................................................. 136
     Testing Conditions .................................................................................................. 136
     Testing Equipment .................................................................................................. 137
     Choice of Test Parameters. ..................................................................................... 138
     Norms ...................................................................................................................... 138


6 - CONCLUSION...........................................................................................................140

APPENDIX A: COMPARISON OF LEAD AND TUNGSTEN SHIELD.....................145

APPENDIX B: TOTAL DOSE TESTING OF A GaAs AMPLIFIER............................147

APPENDIX C: TOTAL DOSE TESTING OF A BANDPASS FILTER........................158

APPENDIX D: TOTAL DOSE TESTING OF AN OPERATIONAL AMPLIFIER......168

APPENDIX E: TOTAL DOSE TESTING OF A GALLIUM ARSENIDE MIXER ......177

LIST OF REFERENCES .................................................................................................186

BIOGRAPHICAL SKETCH ...........................................................................................198




                                                                vii
                                                LIST OF TABLES



Table                                                                                                                   Page

1: Radiation environment in a CAGR nuclear reactor........................................................... 3

2: Radiation environment around a PWR on load ................................................................. 3

3: Radiation environment around a PWR during refueling.................................................... 3

4: Dose rate during fueling operation..................................................................................... 4

5: Nuclear environments associated with waste processing................................................... 5

6: Typical dose rate of various medium and high level wastes ............................................. 6

7: Dose rate for typical CAGR decommissioning tasks......................................................... 7

8: Typical decommissioning environments of a PWR........................................................... 7

9: Photon tenth value thickness in cm for Al, Fe, Pb and concrete........................................54

10: Range of electron in aluminum........................................................................................56

11: Range (µm) of alpha particles in Al, Pb, water and air ...................................................57

12: Range of protons in aluminum.........................................................................................57

13: Radiation damage thresholds on metals...........................................................................60

14: Radiation damage thresholds on ceramics.......................................................................60

15: Radiation tolerance of plastics.........................................................................................62

16: Radiation damages on coatings........................................................................................63

17: Radiation damages on adhesives .....................................................................................64

18: Radiation effects on lubricants.........................................................................................65

19: Radiation damages on window glasses............................................................................66


                                                          viii
20: Radiation damages thresholds on resistors ......................................................................69

21: Radiation damages on capacitors.....................................................................................70

22: Radiation damages on connectors, switches and relays...................................................73

A-1: Comparison of lead and tungsten shields ......................................................................146

B-1: Output data versus total dose .........................................................................................156

C-1: Results of the bandpass filter irradiation .......................................................................166

D-1 :Evolution of the DC offset with the total dose. .............................................................176




                                                           ix
                                                 LIST OF FIGURES



Figure                                                                                                                       Page

1: Internal structure of a steam generator............................................................................... 14

2: Trapping zones at Si-SiO2 interface................................................................................... 79

3: Radiation effects on n-channel MOS devices.................................................................... 86

4: Effect of biasing on threshold voltage of MOS devices. ................................................... 90

5: Attenuation provided by lead and tungsten shields ...........................................................111

6: Radiation tolerance by families of components.................................................................119

B-1: Dose Map of the University of Florida Cobalt Irradiation ............................................149

B-2: GaAs amplifier test configuration..................................................................................151

B-3: GaAs amplifier testing board.........................................................................................152

B-4: University of Florida Cobalt Irradiation facility............................................................153

B-5: GaAs amplifier in the irradiation chamber ....................................................................154

B-6: Window of the testing at the end of the experiment ......................................................155

B-7: Variations of power with total dose...............................................................................156

B-8: Variations of magnitude with total dose........................................................................157

C-1: Dose Map of the University of Florida Cobalt Irradiation ............................................160

C-2: Bandpass filter prepared for a radiation test ..................................................................162

C-3: Bandpass filter circuit ....................................................................................................163

C-4: Bandpass filter test configuration...................................................................................163

C-5: University of Florida Cobalt Irradiation facility............................................................164


                                                           x
C-6: The bandpass filter in the irradiation chamber...............................................................165

C-7: Results of the bandpass filter irradiation .......................................................................167

D-1: Dose Map of the University of Florida Cobalt Irradiation............................................170

D-2: Operational amplifier testing board...............................................................................171

D-3: Operational amplifier test configuration........................................................................172

D-4: University of Florida Cobalt Irradiation facility............................................................173

D-5: Operational amplifier in the irradiation chamber ..........................................................174

D-6: Evolution of the frequency response with the total dose...............................................175

D-7: Evolution of the DC offset with the total dose ..............................................................176

E-1: Dose Map of the University of Florida Cobalt Irradiation.............................................179

E-2: GaAs mixer testing board...............................................................................................180

E-3: GaAs mixer test configuration .......................................................................................181

E-4: Testing setup in the irradiation room.............................................................................181

E-5: University of Florida Cobalt Irradiation facility ............................................................182

E-6: The GaAs mixer testing board .......................................................................................183

E-7: Evolution of the amplitude versus total dose .................................................................184

E-8: Evolution of the phase difference versus total dose.......................................................185




                                                         xi
                  Abstract of Thesis Presented to the Graduate School
                 of the University of Florida in Partial Fulfillment of the
                    Requirements for the Degree of Master of Science

                ROBOTICS AND RADIATION HARDENING IN THE
                           NUCLEAR INDUSTRY

                                            By

                                    Laurent P. Houssay

                                       August 2000


Chairman: Professor James S. Tulenko
Major Department: Nuclear and Radiological Engineering

       A review of the robotic tools in the nuclear industry is presented. The complexity

and efficiency of these systems has improved greatly during the last decade. However,

the degradations induced by radiation often limit the capability of robotic systems and

shorten their lifetime. The effects of radiation on robotic equipment are presented. The

damages produced by radiation on electronic devices are particularly dramatic and are

studied thoroughly. Several methods to upgrade the survivability of electronic circuits

are developed. These techniques increase the lifetime of a robotic system but their

development is often costly and time consuming. Finally, the importance of a good

testing protocol is stressed. This guide is destined for engineers who are not familiar with

robotic systems for nuclear applications and who need an overview of the impact of

radiation on these systems.




                                            xii
                            CHAPTER 1
         OPERATIONAL ENVIRONMENTS IN THE NUCLEAR INDUSTRY



       Not only is radiation hazardous to humans but it is hazardous to electronics as

well. When designing a system for radiation environments, the total dose, the type of

radiation and sometimes the dose rates are major factors that will limit the lifetime and

the reliability of the equipment. In nuclear science, there is radiation in every step of the

fuel cycle. The real concern is for the spent fuel and the decommissioning of nuclear

facilities where the highest radiation fields are found. The data presented in this section

comes from an evaluation made in English facilities [1] and from the RADECS 99

(RADiation Effects on Components and Systems) technical book [2-3]. This data gives

the dose rate in several facilities for typical work on the equipment. The following dose

rates are only meant to indicate an order of magnitude. This information will allow

engineers to calculate the total dose requirement for the system over its expected lifetime.

All indicated dose rates are referenced to energy deposited in silicon.




                                      Fuel Fabrication


       Due to the wide variety of nuclear reactor fuels it is very difficult to give an

accurate estimate of the dose rate in the fuel fabrication process. It is known, however,

that mixed oxide fuel (Uranium + Plutonium) have higher radiation levels when

compared to uranium based fuels. This is due to the plutonium that has its own activity


                                              1
                                             2


and to small concentration of radioactive impurities. The main source of radiation of new

fuel is low energy gamma rays. The penetrating power of these photons is weak and the

dose rate is at its highest when the fuel is in powder form. The photon dose rate ranges

from 10-4 Gy/h to 10-1 Gy/h with a common value is 10-3 Gy/h. The contribution of

neutrons that come from the decay of transuranics associated with plutonium is much

weaker and is between 10-6 Gy/h and 10 -4 Gy/h. In most cases the effects of radiation on

robotics from new fuel elements are close to zero.




                                Reactor System Operation


       The dose rates existing in the core of a fission reactor are the highest dose rates

associated with nuclear power production. It is also a place where the neutron dose rate

causes significant damage to electronics. The neutron flux is so high that it creates single

event upsets, in much the same manner as cosmic rays in space. The equipment that can

be used in such an environment must be highly hardened against radiation damages.

Table 1 indicates the gamma and neutron dose rates when the reactor is on load and

shutdown. The values of dose rate are typical for a commercial advanced gas cooled

reactor (CAGR), and are close to the ones from a pressurized water reactor (PWR).
                                            3


Table 1: Radiation environment in a CAGR nuclear reactor
                  State of the        Gamma dose    Neutron dose         Neutron flux
Location
                  reactor             rate (Gy/h)   rate (Gy/h)          (n/cm2/h)
                                         7
                  On load             10            107                  1017
Within the core                          4
                  Shut-down           10            5. 10-1              1010
Outside the       On load             102           10                   5. 1010
 radial shield    Shut-down           10-3          Negligible           104
Above the         On load             10            1                    1010
                                            -3
 pressure dome Shut-down              5. 10         Negligible           10
Coolant loop
                  On load             5. 10-1       Negligible           Negligible
 area
Above
                  On load             5. 10-4       3. 10-4              Negligible
 operating deck
Source of data: see references [1, 2, 3]

       The following Tables 2 and 3 indicate additional typical dose rate values for a

PWR on load and during refueling/maintenance operation.




Table 2: Radiation environment around a PWR on load
Location                       Gamma dose rate (Gy/h)        Neutron dose rate (Gy/h)
Pressure vessel annulus        102                           3. 102
Coolant loop area              5. 10-1
Outside the loop area          2. 10-3                       2. 10-4
Source of data: see references [1, 2, 3]



Table 3: Radiation environment around a PWR during refueling
Location                               Gamma dose rate (Gy/h)
Lower intervals of the pressure vessel 2. 10-1
Upper intervals of the pressure vessel 7. 10-2
Neutron shield pad                     >103
Around the steam generators            1.5 10-1
Coolant loop area                      3. 10-3
Source of data: see references [1, 3]

       When the reactor is on load, the gamma spectrum consists mainly of 1 to 5 MeV

photons resulting from Nitrogen-16 decay. The energy is weaker (between 0.3 and 1.3
                                              4


MeV) when the reactor is shut down, with the major contributor being Cobalt-60 from

the activated steel structures.




                    Spent Fuel Handling and Storage In the Power Plant


        In many reactors, fuel elements are removed from the reactor and placed in a fuel

storage pool. The fuel is stored there until its elements with the shortest half-life have

decayed. After their activity has significantly decreased, the fuel elements are shipped to

a waste processing plant.




Table 4: Dose rate during fueling operation
                        Gamma dose rate           Neutron dose rate     Neutron flux
Location
                        (Gy/h)                    (Gy/h)                (n/cm2/h)
Within the fuelling                                                     From 10 11 to 10 12
                        105                       102
 machine                                                                 near the fuel
                        From 10 3 to 104
In the storage pond                               Negligible            Negligible
                         near the fuel
Source of data: see references [1, 2, 3]



                       Spent Fuel Disassembly and Waste Processing


        Waste processing activity consists of many steps and offers a wide range of

radiation levels. It is impossible to have a generic estimate of dose rates because each

facility, technology and type of fuel will have different results. Nevertheless, it is known

that the gamma rays give the highest contribution to the dose. Mechanical stripping and

cutting of the fuel elements generates about 103 Gy/h. Chemical processing is less

aggressive with 10 2 Gy/h. Vitrification of high-level waste is the final step and generates
                                              5


up to 104 Gy/h, which is extremely high. Fuel reprocessing is often executed in a very

harsh environment and the radiation is only one of the many aggressive factors. Due to

high dose rates and the abundance of remote operations, waste processing employs the

greatest array of robotic systems in the entire nuclear industry. The following Table

describes several waste processing environments.




Table 5: Nuclear environments associated with waste processing
                               Maximum gamma dose rate Maximum neutron dose rate
Operation
                               (Gy/h)                      (Gy/h)
                                  3
CAGR dismantling facility 10                               10-3
Storage pond                   103                         10-3
Shearing                       103                         10-3
Dissolution                    103                         10-3
Solvent extraction             102                         10-3
Plutonium finishing            2. 10-2                     4. 10-4
Uranium finishing              5. 10-3                     4. 10-4
Vitrification                  104                         4. 10-4
Source of data: see references [1, 3]



                               Waste Handling and Storage.


       There are two categories of waste: high and low-level wastes. High-level waste is

spent fuel that has been processed and vitrified. The risks involved with this kind of

waste are significant due to the hazardous compounds that are concentrated. For

example, the gamma dose rate at a distance of 1 meter from an unshielded vitrified

element is up to 200 Gy/h. The storage of high-level waste is a major undertaking that is

governed by many regulations. A complex setup of shielding, monitoring and remote

handling is required since the activity of this type of radioactive material is extremely
                                               6


high. Each equipment requirement should be linked to the expected work environment,

because a wide range of dose rates can be found.

          Low-level waste is less of a threat to robotics and to the health of workers. This

waste is composed of lightly contaminated waste such as gloves or concrete from

decommissioning. The contaminants are radioactive elements with low activity and short

half-life. Low-level waste is usually contained in metallic drums and stored in protected

sites on surfaces or superficially buried. Dose rates associated with low-level waste do

not create damage to electronics most of the time. Unlike high-level waste, lighter

shielding and packaging systems are required and used to reduce the exposure to workers.

The next Table indicates several dose rates at the surface and at 1 m from various nuclear

wastes.




Table 6: Typical dose rate of various medium and high level wastes
                                                            Dose rate at 1 m distance in
Waste description              Dose rate at surface (Gy/h)
                                                            air (Gy/h)
Medium level sludge            6                            4.5 10-1
Cemented waste                 2. 10-3                      10-4
High level vitrified           1.8 103                      2. 102
Cemented resins (ions
                               Up to 3. 10-1                Up to 1.5 10-2
 exchangers)
Filter elements in concrete    Up to 1. 10-1                Up to 5. 10-3
Miscellaneous                  Up to 5. 10-3                Up to 2. 10-4
Source of data: see references [3]



                           Decontamination and Decommissioning.


          Many nuclear power plants as well as “cold war” nuclear weapons production

facilities built in the 1950s, have reached the end of their lifetime. These buildings
                                             7


require decontamination, which can be accomplished by cleaning up the radioactive

contamination and then decommissioning the facility. The dose rate at 1 cm from 2

grams of spent fuel is on the order of 10 Gy/h. Certain parts of the equipment have been

activated by the neutron flux over the years. These parts of the reactor are less and less

activated as the distance from the core increases. The next Tables indicate typical doses

rate associated with the decommissioning of a CAGR (Commercial Advanced Gas

Cooled Reactor) and PWR (Pressurized Water Reactor) (Mol Belgium).




Table 7: Dose rate for typical CAGR decommissioning tasks
Location                       Gamma dose rate (Gy/h)
Debris vault                   50
In-reactor steel components 1 to 10
Graphite stack                 3. 10-2
Diagrid region                 2. 10-4 to 2. 10-3
Boiler region                  10-4 to 10-3
Vessel concrete                10-4
Standpipe region               10-4 to 10-3
Source of data: see references [1, 3]



Table 8: Typical decommissioning environments of a PWR
Location                           Dose rate (Gy/h)
De-activation pool                 10-5
Reactor pool                       2. 10-5
Reactor building                   2. 10-5
Shipping area                      2. 10-5
Primary circuit                    4. 10-1 to 1.2 10-2
Building hot spots                 2 to 1.2 10-2
Shipping area near waste drums 2 to 3. 10-1
De-activation pool near fuel       101 to 103
Vessel mid plane                   102 to 103
Source of data: see references [3]
                                             8


       There are currently many storage tanks in the US that require clean up due to

leaks and structural problems. The liquids and sludge contained in these tanks are highly

radioactive. For example, the Department of Energy (DOE) owns the Hanford storage

tanks, which contain 232,000 cubic meters of mixed hazardous waste, with a total activity

of 9.25 1018 Becquerel (2.5 108 Curies) [4]. This means that one-gallon generates an

activity of 4 Curies. For safety and access reasons, robotic tools are the only safe means

to cleanup such facilities.
                               CHAPTER 2
            USE OF ROBOTIC SYSTEMS IN THE NUCLEAR INDUSTRY




                                Need for Robotics Systems


       Today robots are widely used in the nuclear industry. Their main application is to

perform automated and repetitive work or to execute hazardous tasks that are dangerous

to human beings. First, profitability is the motivation to switch from a regular worker to

an automated system. Second, safety of the worker and regulation are issues that should

not be ignored. In nuclear science, protection of workers became a catalyst for the

development of robotics. Today, the regulation 10 CFR 20 from the Nuclear Regulatory

Commission indicates that an occupational worker cannot receive more than 50 mSv per

year for the full body dose [5]. Due to this dose regulation more workers have to be

employed to accomplish a mission. Once the maximum dose has been reached the

employee must stop working immediately. A study published in 1987 [6] indicates that

worker exposure costs more than $500,000 per man-Sv. On the contrary, a utility

executive said in 1990 that every dollar spent on robotics is doubled in return [7]. In

nuclear power plants, the reactor must be shut down or at least be brought to a fraction of

its maximum power to allow human intervention near the core. Robots shorten the

maintenance time and the number of workers needed; this reduction generates many

additional savings: less protective clothes, waste and paperwork. Due to the aging of

nuclear reactors, increased inspections and repairs are needed at a deeper level than ever


                                             9
                                            10


before. Finally, an estimation of the hazardous environment in an emergency situation

avoids putting personnel at risk. Although human safety is the main reason for the use of

robots, profitability is also a strong motivator. Furthermore, a remote system is often the

only way to enter a very high radiation field. The use of robotics in the nuclear industry

is obvious due to the factors mentioned.

       This chapter is dedicated to the discussion of several robots that were designed or

used in the nineties. In most cases the robots are designed for “exceptional use,” which

means that they are not running continuously but rather only when they are needed.

Inspection and repair robots are a good example of such tasks. Since these systems are

not permanently kept in a radiation field, their hardening is a minor concern compared to

their task efficiency. In contrast, a robot whose aim is to manipulate radioactive

equipment on a daily basis has a much greater need for radiation-tolerant components.




                 Mobile Robots for Routine Monitoring and Surveillance


       The SIMON (Semi-Intelligent Mobile Observing Navigator) robot is a mobile

monitoring and surveillance robot developed in 1990 and introduced to the DOE’s

Savannah River site the same year [8]. There were three requirements present in the

design of SIMON to avoid the need for human inspection of this nuclear facility. These

requirements were to measure radiation, to measure temperature and to retransmit

televised views of the area. The results were a great success in that SIMON located 20

spots of beta contamination that were undetected by “human” surveys [9]. SIMON is

equipped with the three-wheeled base of the Cybermotion K2A robot. Equipment used

with SIMON includes radiation detectors, temperature sensors and a camera mounted on
                                             11


a telescoping mast. The positioning equipment consists of optical encoders on the motor

and the wheels, infrared beams and ultrasonic pulses to determine the position relative to

its docking station, and sonar and bumpers are used for collision avoidance. SIMON

navigates into a room by either following a preprogrammed path or it is controlled

manually to perform its monitoring task. It returns to the dock to recharge its batteries

when necessary. A computer program called “Dispatcher” choreographs the

preprogrammed path. Ultra-high frequency radio communication (UHF) links the host

computer to the robot. Once the “Dispatcher” is downloaded, SIMON can still run if the

radio link is out of range. Three on-board computers manage the resources of the robot.

The electronic components that are used in SIMON were chosen because of their natural

radiation resistance. A total dose radiation hardness of 200 Gy (20 krad) is achieved at a

reasonable cost using this design approach. The reliability is also enhanced by a self-

diagnostic capability. SIMON received a US patent in 1994.

       The robot MACS (Mobile Automated Characterization System) was designed in

1996 as the second generation of SIMON robots [9]. This time the goal was to develop a

contamination map of nuclear facilities for decontamination and decommissioning

purposes. This task is performed automatically and in a reliable manner as compared to

workers equipped with portable detectors. MACS is equipped with highly sensitive

detectors and transmits the dose information in real time to the host station. The data

measured is incorporated into an interface called RadMap, which generates a color-coded

map of the contamination on the floor.

       The robot ARIES (Autonomous Robotic Inspection Experimental System) was a

robot developed by the University of South Carolina for the DOE (Department of
                                             12


Energy) [10]. The DOE stores several thousands of steel drums that contain low-level

radioactive waste. The drums are stacked and stored in long aisles. A weekly inspection

of the packaging integrity is required. This work is tedious and long-term exposure must

be avoided, although the radiation level is not hazardous to workers. The goal of ARIES

is to locate each drum and to perform a visual inspection to find paint blisters, rusted

areas or any other sign of container degradation. Since a bar code is located on every

drum, a database containing all the drum characteristics can be updated consistently and

regularly. When a damaged drum is found, its contents are repacked. ARIES was

developed after the SWAMI robot (Stored Waste Autonomous Mobile Inspector) [11].

SWAMI was designed at the DOE Savannah River site for the same application as the

ARIES robot. The two robots have fairly similar capabilities, however ARIES is more

modern and more sophisticated. ARIES uses the K3A mobile platform manufactured by

Cybermotion. This platform, associated with an elaborate sonar and light navigation

system, allows the robot to navigate in a narrow aisle of stacks of drums. The camera

positioning system (CPS) is a module that characterizes each drum. The module includes

a camera, a bar code scanner and a strobe light. Whenever ARIES arrives in front of a

drum stack, the CPS extends up and takes two pictures of the drum. The on-board

computer then processes the images and updates the database while the module moves up

to the next drum. Once the stack is completely inspected, the robot moves to the next

stack or to its dock to recharge its batteries. The system has been successfully tested in

several DOE facilities including the Fernald site and INEEL (Idaho National Engineering

and Environmental Laboratory). A radiation hardening study concluded that a radiation-

hardened version of ARIES was not needed since the radiation level in the storage facility
                                            13


was too low to jeopardize the reliability of the equipment. The major shortcoming of the

present system is its low drum survey speed, which allows inspection of only one-third of

the expected 10,000 drums per week, due to battery problems [12].




                   Inspection and Maintenance of Reactor Components



Inspection and Cleaning of a Typical Steam Generator


       The degradation of the steam generator performance is a technical challenge both

on the secondary and the primary side. In a steam generator, the primary hot water from

the reactor enters the generator at its base and moves to the other side through a bundle of

tubes. The tubes heat the secondary water that fills the vessel and then generates steam.

Figure 1 shows the internal structure of a steam generator. The sludge that builds up on

the secondary side of steam generators reduces its thermal efficiency. This accumulation

decreases the power output of the power plant and can cost millions of dollars in lost

revenues. Lower thermal efficiency is not the only consequence of the sludge. Another

major problem is the corrosion generated by the chemical elements that make up the

sludge, which may lead to cracks in the tubes [13]. Fretting leaks due to the presence of

foreign objects may also require the complete replacement of a steam generator. The

major challenges of the sludge removal are the extremely difficult access inside the steam

generator and the radiation field existing in this area. One cleaning method for removing

the sludge is chemical cleaning. Unfortunately, this process is expensive and generates

mixed chemical plus radioactive wastes that are difficult to store. A widely used cleaning
                                                14


technique consists of highly effective water jets that are powerful enough to remove soft

as well as hard sludge.

                                             Steam




      Support plates                                                         Tube




                                                                      Secondary water



                                                                            Tubesheet
        Divider plate
                                                                             Manway
 Primary coolant inlet                                                       Primary coolant outlet


                         Primary hot water             Primary cold water


                         Figure 1: Internal structure of a steam generator



       The robot CECIL (Consolidated Edison Combined Inspection and Lancing) is

designed for steam generator maintenance [14]. It is limited to cleaning the bottom part

of the steam generator, between the tubesheet and the first support plate. Michael

Reinhardt [15] provides an clear description of the CECIL setup, “CECIL 4 rides on a rail
                                            15


installed through the steam generator inspection port. The robot can move longitudinally

along the rail assembly and extend and retract a flexible lance. This lance can be driven

between the columns of the U-tubes, and the robot can rotate about its axis to position the

lance at any point within the tube bundle above the tubesheet and below the first support

plate. Water jets are introduced into the tube bundle by way of these flex lances and the

robot’s barrel sprays. The flex lances are the transport mechanism for inserting spray

nozzles and a video probe deep into the tube bundle.” The system is remotely controlled

by a workstation installed in a low radiation zone. The operator can monitor and control

the position of the robot and the lance in real time. The water used in the process is

continuously pumped and processed in order to reduce waste volume. Several types of

lances are available, each one for a different task. One of the most interesting group of

lances is not used for cleaning the sludge but to locate and remove small objects

accidentally lost deep in the tube bundle. These objects can create significant damage

leading to fretting leakage and must be removed. The FOSAR (Foreign Object Search

and Retrieval) lance is equipped with one of four grappling tools to retrieve these objects

[16]. This system has been demonstrated successfully several times [17]. The first

cleaning test of CECIL 4 was carried out at Indian Point 2 in 1989. Later in 1991, the

FOSAR system removed a weld rod deep inside a steam generator at the Salem-1 reactor

[18].

        The robots called UBIB (Upper Bundle in Bundle) and UBHC (Upper Bundle

Hydraulic Cleaning) were introduced in 1996 [19]. These robots offer a considerably

increased cleaning and inspection efficiency compared to the CECIL system. The

CECIL robot is limited to the region between the tube sheet and the first support plate of
                                             16


the steam generator, whereas the UBIB and UBHC robots can access the upper bundle

region without drilling holes in the shell. UBIB is the upper bundle in the bundle

inspection system that is dedicated to remote visual inspection of steam generators. The

UBIB system is installed, like CECIL, through the hand hole at the bottom of the steam

generator, and its mast is extended to the upper region. A video camera placed at the end

of a wand can then inspect the bundle. A complete inspection of a steam generator can

be completed in one day and provides information for the cleaning process. The UBHC

robot is the cleaning version of the UBIB robot. UBHC is deployed the same way as

UBIC but is equipped with 20 nozzles in addition to a camera and lighting system. These

powerful water jets deliver 3000 psi water at a rate of 70 gallons-per-minute while the

position of the equipment is secured by inflating support bladders. The complete

cleaning operation lasts two days or even less if a previous inspection by UBIB did not

require a full bundle cleaning [20]. An operator located in a control station located in a

low dose area remotely controls the inspection and cleaning operations. The inside of the

steam generator and the location of the robot are monitored in real time. The system also

provides many safety features that stop the process in case of problems. Both the UHBC

and UBIB systems are currently designed to treat Westinghouse Model 44, 51 and F

steam generators. UBHC has been commercialized since February 1999 and has

successfully completed numerous cleaning operations in more than four countries.

       A study assessing the occupational radiation exposure in European nuclear

reactors [21] indicates that the highest doses experienced in a PWR (Pressurized Water

Reactor) are for steam generator primary side work, where the median exposure value

exceeds 200 man-mSv. Since the maintenance is performed remotely, the majority of the
                                             17


worker exposure occurs during the installation of the inspection and cleaning equipment.

Therefore the total dose received by the workers is primarily proportional to the number

of steam generators opened rather than on the duration of the maintenance. The major

contribution to the exposure during maintenance comes from the installation and the

removal of the nozzle dams in the steam generator. The dose rate in the steam generator

bowls is high: from 50 mSv/h to 250 mSv/h. Nozzle dams are installed in the inlet and

outlet nozzle of the steam generator to isolate the primary loop that is flooded during

refueling. The replacement of the workers by a robot for this specific task considerably

decreases the personnel exposure. At least two systems have been designed to meet this

challenge. Both systems use a remotely controlled robot positioned in the channel head.

With the oldest design, several cameras and lights are mounted on the robot, which

allows the operator to control a manipulator arm very accurately. The robot is located

outside the steam generator and operates through a man-way. The nozzle dam is then

positioned carefully by the arm and secured with bolts. This method has been

successfully tested with several handling robots: with a Schilling hydraulic manipulator

in 1991[22], with the ROMA (Remotely Operated Manipulator Arm) robot in 1990 and

the ROSA III robot in 1992 [23]. The need of a handling robot has become obsolete,

thanks to the system developed, tested and designed by Foster-Miller [24]. A robotic arm

is avoided, because movement of a bulky robot at the bottom of the steam generator is

difficult and because the complexity of the installation requires highly trained personnel.

The idea behind this new robot is to design a system exclusively for the installation of the

nozzle dam. In this system the dam is fully integrated in the robot to provide a faster

installation. The bolting system is also unique and fits both the robot and the dam for a
                                             18


quicker and safer operation. The robot can be used on any existing nozzle and is installed

without personnel entering the steam generator. The manipulation of the system is

greatly simplified compared to the previous design and does not require extensive

training for the operators. Foster-Miller successfully tested this robot on a mockup in

1994 and has shown significant exposure and critical path time reduction.

       Once the nozzle dam is installed and secured the inspection or the maintenance of

the primary side of the steam generator can begin. An eddy current probe performs the

inspection of the inside of the tubes. This probe is sent into the tube by a manipulator

robot from the bottom of the cold or hot leg of the tube bundle. The deterioration of the

tubes or their leaks can then be documented. The most challenging operation is to clean

the inside of the tubes. When the reactor is working, deposits accumulate inside the

tubes, increasing the flow resistance through the steam generator and reducing the total

efficiency of the process. The chemical cleaning method consists of injecting a chemical

solution to attack and dissolve the deposits. This technique generates a large volume of

mixed waste that is undesirable. The SivaBlast mechanical cleaning system from

Siemens is a unique solution that is both clean and efficient [25]. The principle is simple:

steel beads are blasted from the cold leg of the primary side and are collected at the hot

leg with the waste. A generator provides the pressure to a blasting nozzle, a vacuum tool

receives the beads and a reclaimer separates the beads from the debris. This vacuum tool

avoids any contamination such as dust and small wastes. The beads are then washed and

reused. This method has already been used on non-nuclear applications. The first

challenge is operating remotely on each of the tubes, the second challenge is to clean all

3550 tubes contained in a single steam generator on schedule. A robot with a strong
                                             19


dexterity can solve the first challenge. The goal is to access all tubes of the cold leg and

their corresponding end on the hot leg. Initially, the Telbot robot was used, then the

Flexivera manipulator system followed. A description of those two systems is given in

the next paragraph. The second challenge is to improve the speed of the cleaning

operation. The software ROBCAD is a program that optimizes the robot path planning.

This method allows a speedier cleaning operation since the robot cleans all the tubes

corresponding to its arm configuration. Finally, the design of the blasting tool was

developed to include two nozzles. The robot interface consist of a user interface to

control the manipulator, a ROBCAD based simulation system, and several monitors to

follow the operations. The role of the operator is greatly simplified by the high degree of

autonomy offered by the hardware and the software. Three camera sets provide many

views of the cleaning operation; views of the steam generator man-ways, views of the

two cold and hot sides of the steam generator and views of the vacuum and blasting tools.

This SivaBlast system has been tested in the Point Lepreau power plant in 1995: it

removed 787 kg of deposits from 8209 tubes.

       A description of several mechanical arms that are mainly used for steam generator

inspection or maintenance is next described. The ROSA III robot is the evolution of the

ROSA I and II models [26]. This robot is a remotely controlled maintenance and

inspection robot. ROSA III is a mechanical arm that can be equipped with tools for

general-purpose work. It is mainly used, however, for steam generator operations

through the manhole port to reduce personnel exposure. The control station consists of a

highly graphic interface allowing maximum flexibility for the operator. ROSA III is

available from Westinghouse Electric. The Cobra system was specifically designed to
                                             20


increase maintenance productivity and to reduce personnel exposure when workers

access the inside of a steam generator [27]. It is a lightweight electric manipulator that

inspects and repairs the lower tube sheet. Cobra was tested in 1992 and is deployed by

B&W Nuclear Service. The Telbot robot is a 6 degrees of freedom mechanical arm

developed for application in areas inaccessible for humans. Its main advantage is its

great flexibility; each arm can rotate 360 degrees continuously. It lacks electrical

components, having neither a cable nor a motor in its arm. All sensitive elements are

protected in the base that contains drive motors, drive gears, encoders and cabling.

During steam generator maintenance, the arm is installed through the man-way, while the

base remains outside. The control cabinet can be associated with this base or located in a

safer area. The Flexivera manipulator system was exclusively designed for steam

generator operation. It is small, lightweight and easy to install in the steam generator

bowl. The four axes of the manipulator arm are controlled by a computer system located

outside of the radiation area and linked to Flexivera by a cable. It is now one of the most

efficient robots for inspection and maintenance of the primary side of steam generators.



Inspection of Reactor Vessel


       Specific regulations require reactor vessel inspections on a periodic basis. The

goal is usually to look for cracks and to probe the welds visually with an ultrasonic probe.

These operations imply several challenges. First, all the operations are accomplished

underwater since the reactor is completely flooded. Secondly, all the equipment in the

reactor has to be taken out prior to inspection. That means that the reactor is out of

service during the maintenance period. Immobilization of the reactor can cost from
                                              21


$500,000 to $1,000,000 per day; therefore, the duration of the process must be minimized

in order to increase the profitability of the plant. The third challenge is the accessibility

of the welds. Early design of reactor vessels did not include inspection guidelines, nor

provide easy access to sensitive parts of the vessel. For example the inspection of the

Yankee Rowe PWR reactor vessel was challenging because of a permanent thermal

shield, which stands 2 inches inward from the vessel wall [28]. In such a case, the NRC

(Nuclear Regulatory Commission) would grant relief and waive a requirement for an

inspection of these inaccessible parts. Approval for relief and waiver is now more

difficult to obtain, especially for a license renewal.

       The inspection of BWR (Boiling Water Reactor) vessels is often much more

challenging than the PWR inspections because of their highly complex internal structure.

BWR’s contain hardware that stays in the vessel and occupies a great deal of space,

which makes it difficult to deliver the ultrasonic probes or the detectors at the right

location. In Tsuruga unit-1 for example, an inspection tool must pass through a gap of

only 14 cm to access some of the welds [29]. This restriction leads to some technical

difficulties that are a result of a shroud that covers the entire bottom half of the reactor

vessel, up to 12 meters below the water surface. This challenging inspection was

resolved by the use of a complex telescopic mast and arm setup. The upper part of the

vessel is probed with the upper mast in its highest position and the lower mast deployed.

The operator then moves an ultrasonic detector with a mechanical arm placed at the end

of the mast. The inspection of the middle part of the vessel is the most delicate operation

since it requires the insertion of an arm through the narrow gap between the shroud and

the vessel wall. After the insertion is accomplished, the operator moves down the upper
                                             22


part of the mast. The entire setup has 9 degrees of freedom and is attached to a platform

that moves on a ring guarder around the vessel. The operator controls the system from

the platform without the help of a visualization system. About 75% of the formerly

inaccessible weld can then be accessed, probed and documented. The Tsuruga unit-1 was

successfully inspected with this tool in June 1991.

       The internal parts of a PWR are completely removed during the maintenance

operations. Consequently, access to the welds is much less a problem than for a BWR.

The design of inspection robots for PWR is more focused on maintenance costs and time

reduction. These savings can be accomplished by optimizing the installation and

operation duration (critical path time), by reducing the personnel and the hardware (polar

crane) needed for the inspection. The techniques that have been used in the past are

based on the same principle: a rigid telescopic mast attached to a massive base at the top

of the vessel. The movements of the arm and the mast make it possible to inspect the

entire vessel. The major disadvantages of such a system are its weight and bulk. A

central mast manipulator (CMM) from Siemens weighs 16,000 kg and requires three

trucks for transportation. An automated reactor vessel inspection system (ARIS) from

B&W Nuclear Technology is also bulky and weighs 13,000 kg. It is obvious that the

installation of such equipment on the top of the reactor vessel creates handling problems

that cost time and money. The installation and calibration of the system often requires

more than three days [30]. The personnel needed for this inspection are numerous and

are kept from other tasks. A lighter system has been designed and used by the PAR

company: the inspection module weighs a total of 1800 kg. The major difference with

the previous technology is that a bridge is not needed at the top of the vessel. Instead of
                                             23


being controlled from a platform at the top of the reactor, the mast is fixed on a tripod

structure attached to the vessel wall completely under water. This system is much

cheaper and easier to handle than the previous tools but its weight still requires the use of

the polar crane during the installation. The Mitsubishi advanced ultrasonic testing

machine (A-UT) is an innovative method that has revolutionized the reactor vessel

inspections. This robot does not have a central mast but has a small base held to the

vessel wall by four suction cups. This A-UT robot scans the vessel and then swims or

moves to its next location using its 8 wheels and 6 thrusters. Since this robot does not

have an external location reference, it uses several lasers to determine its orientation and

its position inside the vessel. The A-UT weighs only 300 kg and requires little

installation and little handling (no polar crane). The disadvantages are the complexity

and the reliability of the displacement system (28 degrees of freedom) but also the fact

that the unit has to swim to the surface to change the ultrasonic tools. The URSULA

(Ultrasonic Reactor Scanner Un-like Aris) robot that appeared in 1995 represents the

most highly developed system available today [31]. The general design is fairly close to

the Mitsubishi A-UT; however, URSULA uses three suction cups to attach itself to the

vessel wall while a six-degree of freedom arm scans the welds with an ultrasonic module.

The robot uses thrusters to swim to its location and uses lasers to keep track of its

position. All the modules of the robot are neutrally buoyant to minimize the load on the

thrusters and the vacuum pump. URSULA is so small that it can enter the reactor

through personnel hatch and two units can work together in the vessel, offering

considerable timesavings. The robot can choose from several ultrasonic tools and

calibrate them while being underwater. The ultrasonic head allows the inspection of all
                                             24


types of welds: circumferential, longitudinal, bottom head, nozzle inner radius, nozzle to

vessel, nozzle to pipe and safe end to elbow. All the sensors and control equipment is

installed in the robot, therefore, only the signal and communication lines link the robot to

the operator with fiber optic cables. This operator monitors the operations through video

cameras and a graphical interface while an efficient automated system gives significant

autonomy to the robot. The time reduction for the inspection of a typical four loop

Westinghouse plant is reduced to four days by using two URSULA robots, whereas eight

to ten days was usually required.



Pipe Inspection


       Reactor vessel welds are not the only welds that are under stressful conditions.

BWR and PWR reactors have pipes that also need inspection of their welds. Two types

of pipe inspection are available depending on the access possibility; welds from the

inside or the outside of the pipe. For outside inspection of the pipes the challenge, once

again, is the access. In BWR for example, jet-pump risers are located between the reactor

vessel wall and a shroud in a gap often smaller than 30 cm. The concerns with such

pump risers are the welds at and near both ends of the riser elbow. A robot has been

specially designed to access and inspect these welds. This robot is controlled from the

top of the reactor and is lowered into position between the core shroud and the reactor

vessel with a long-handle pole. The robot then uses a pneumatic clamp that attaches the

unit to the pipe. The inspection is accomplished with an effector that moves the

ultrasonic sensors to scan the pipe around the pipe and the elbow. Pipe crawlers are

widely used for inside inspection of pipes. These small robots use tracks, wheels or
                                             25


thrusters to navigate deep in the pipe. Their task is often limited to a visual inspection of

the pipe with a CCD camera [32] but a few robots are equipped with ultrasonic

transducers or radiation detectors [33]. The challenge for these robots is to navigate deep

into the pipe, through turns and elbows. The shape of the pipe crawler robot adapts to the

pipe. These robots are often flexible and can be as small as 2 inches of diameter. The

connection with the operator is realized with a highly flexible wire that carries the power,

controls and the signals. For difficult access like with a reactor vessel nozzle, the pipe

crawler can be delivered in the pipe by another robot. The nuclear industry is one of the

many users of pipe crawlers. A wide variety of sizes and models are now available as

well as required equipment. The operation of pipe crawlers presents little risks. Many

years of experience have improved their reliability.



Underwater Inspection


       Water provides sufficient shielding against radiation at a very low cost. Many

nuclear power plants operations are conducted under water for these reasons. Many

robots are therefore designed to operate or navigate under water. Previous technology

used long poles or telescopic arms to perform visual inspection and to retrieve small

object [34]. This method is not only costly, but also requires substantial time and some

personnel exposure [35]. Submersibles provide a great deal of help to reduce these costs.

First, submersibles can provide underwater views of an operation that would be very

helpful; to a polar crane operator for example. Second, submersibles require little

maintenance and mission preparation. Third, they are very flexible, easy to use and fast;

which is very useful in emergency conditions. Finally, submersibles are affordable, easy
                                             26


to upgrade and offer little risk of damaging reactor equipment. However, there are few

shortcomings; submersibles are often limited to visual inspection and their size cannot

allow them to go in very tight spaces. Their payload is small so they cannot handle a

large load of onboard equipment. The radiation hardening strategy of submersibles is

simple, all of the electronics except for the CCD camera is removed from the robot. The

control system is located in the operator’s desk and is linked to the robot with cables.

This method limits the risk of failure; moreover, all the submersible parts are secured and

linked to the surface so that no foreign object can be lost by accident. For most

applications, only two workers are needed to operate a submersible. A cable tender is

located at the top of the pool and his task is to submerge the robot into the water and

make sure that the umbilical cable does not get tangled. The operator is located far away

from the robot, to insure a low-radiation field. Minimal training is required for the

operator and the controls of the robots are simple. The more one displaces the joystick,

the more thrust is applied. Several monitors provide the operator with all necessary

views and parameters. The in-vessel visual inspection of reactor elements is the primary

application of submersibles. Their maneuverability and small size is particularly useful

in BWR where the access is limited by the internal hardware. Since submersibles are

used in many areas like the petroleum industry, ocean exploration, or the military, many

models of submersibles are available from several companies around the world. One of

the first submersibles used in the nuclear industry was the MiniRover [36]. It was used in

the fall 1987, in Salem-1 for underwater reactor core inspection [7]. Deep Ocean

Engineering Inc. manufactures three types of robots for the nuclear industry [37]. The

Firefly robot is a small sized (190×153×356 mm) submersible that can inspect confined
                                             27


spaces, such as beneath the BWR core plate. It is equipped with a CCD camera. The

Dragonfly robot is a version that is larger than the Firefly. It can carry additional

equipment such as cameras or lasers. Both the Firefly and the Dragonfly have a body

geometry that prevents being affected by thermal water currents. The Phantom 500 robot

is the biggest submersible made by Deep Ocean Engineering Inc. It is capable of visual

inspection but can also retrieve small objects in a reactor or perform simple ultrasonic

testing. The size of the robot is a critical factor for BWR inspection. Toshiba has

developed a very small (15 cm diameter, 20 cm long) ROV (remotely operated vehicle)

that navigates underwater to inspect BWR core internals [38]. Four thrusters assure the

propulsion of this ROV that is linked to the operator via a small flexible and buoyant

cable. This robot is limited to visual inspection because it only has a CCD TV camera.

The recovery of such robot after a problem in an inaccessible part of the reactor is a risky

operation, potentially dangerous for the reactor. Therefore a support system has been

developed to deliver the ROV in the bottom region of the reactor vessel. The robot is

held at the end of a long pipe, similar to a control rod guide tube. In addition of holding

and releasing the ROV, this support system is equipped with two cameras, a light, and the

responsibility of managing the cable unwinding mechanism. This technique has many

advantages: it allows a much easier recovery of the ROV in case of problems, it avoids

interference between the cable and the ROV cable and finally the additional cameras

provide an extensive view of the work. From the radiation hardening point of view, the

only radiation sensitive element of this system is the CCD camera and this has been

certified up to 200 Gy. This equipment has been successfully tested in actual reactors

and has demonstrated excellent timesavings.
                                             28




                            Handling and Processing of Waste


       Manipulators are essential in a nuclear environment. They are used in many other

disciplines to process repetitive and delicate work. Their mission in the nuclear industry

is different: they replace the human arm where the radiation level compromises the safety

of the personnel. For a long time the manipulation of hazardous material has been

executed by a master/slave system. The operator manipulates a master arm that is

mechanically connected to a slave robot in a hot cell. The robot reproduces every

movement of the operator. The advantages of this system are its simplicity and

affordability. The shortcomings are a low payload and a limited distance between the

operator and the robot. Teleoperated systems have been developed to avoid such

problems. In such systems, the symmetry of movement is not reproduced mechanically

but electrically. Sensors are measuring the displacement of a master arm, then the

information is sent to actuators on the slave robot. Since the cost of custom-built robot is

often excessive, the strategy is to use commercially available robots. Unfortunately, the

use of robots in other industries is different; the manipulators are used to perform a

preprogrammed task over and over, and are not controlled in real time by operators.

Some modifications are thus needed. For radioactive environments, the requirements for

a teleoperated robots are a good sealing of the parts to avoid contamination, the

installation of a force feedback system, an acceptable level of radiation hardness, a good

payload/mass ratio, a high reliability and modularity for easier maintenance and an easy

integration into embedded equipment. The installation of a radiation-hardened force

feedback system is very important when upgrading a commercial manipulator to its
                                             29


nuclear equivalent. Since the manipulator is used in place of a human arm, it is not

sufficient to simply visualize an operation to be able to operate correctly. A good

example is the handling of an egg. An operator can see how well a gripper holds an egg

on a monitor but this is not enough. This operator needs a force feedback system to avoid

squeezing the egg too hard and avoid breaking or dropping it. This feedback is made

with force and torque sensors and requires a significant amount of electronic and controls

algorithms. A commercial manipulator has many advantages including low cost, high

payload, speed and good reliability. The NEATER 760 (Nuclear Engineered Advanced

TeleRobot) developed by AEA Technology is a modification of the commercial Puma

762 robot [39-40]. The on-board electronics of the Puma were redesigned to a tolerance

of 106 Gy of total dose. A modular design also allows a quick replacement of a failing

unit. The maximum load of the NEATER robot is 20 kg at a maximum reach of 1.4 m.

The same strategy has been used by the French CEA for the modification of the

STRÄUBLI RX90 robot [41]. The CEA has also developed a computer based

teleoperation control system named TAO2000. This program integrates the geometry of

the robot and the non-linearity of its sensors to provide a virtual model of the robot and

its trajectory. An automated path planning and an efficient man-machine interface can

then be created and improve the quality of the teleoperation. The TAO2000 system is

also used with the BD250 dexterous arm also developed by the CEA. Unlike the

modified RX90, the BD250 is not a modification of a commercial robot [42]. It was

designed to meet unique requirements in the nuclear industry. The BD250 is a 7 degrees

of freedom mechanical arm. A succession of roll and pitch axes provides a very high

dexterity to this robot. The payload is 25 kg, while the mass of the robot is 75 kg; the
                                             30


payload to mass ratio is very good. The total length of the robot is 1.4 m and the arm can

be introduced into 25 cm diameter fitting. A master arm located with the operator in a

safe area controls the BD250. The on-board electronics and the force feedback sensors

are radiation resistant up to 10 kGy. Two BD250’s can be associated to form a high-

performance mobile working platform. This dual arm system is suspended by a mobile

crane and is much more mobile, lightweight and smaller than any other dual manipulator

robot.




                          Decontamination and Decommissioning



Surface Cleanup


         Despite the constant care and maintenance of the equipment in a nuclear facility,

it may happen that radioactive contaminants are spilled and must be removed. More

often some facilities and hardware need to be decontaminated after being in contact with

radioactive contaminants during operation. In most cases the spill is small enough that

the workers are protected by mobile shielding equipment and can clean up the spill.

However, the need to reduce personnel exposure and the ALARA (As Low As

Reasonably Achievable) concept are pushing the use of robotics for cleaning and

decontamination work. Conventional, general-purpose robot or custom-made actuators

as well as cleaning robots are used depending on the cleaning needs and the contaminated

facility geometry. Several examples are given below.

         A master slave manipulator has been specially designed to cleanup a flooded

radioactive waste area at Nine Mile Point 1. The cleaning operation was required to
                                            31


remove all contaminated materials including the barrels and to decontaminate the area to

acceptable levels. A robot called TROD (Tethered Remote Operating Device) consisting

of a remote manipulator Gamma 7F with 6 degrees of freedom and 2 meters reach was

specially manufactured by RedZone robotics [43]. This mechanical arm is mounted on a

moving base that navigates in the contaminated area by an existing conveyor system

fixed to the ceiling of the room. This method of navigation provided greater access to the

waste, and better maneuverability than a ground based vehicle. The Gamma 7F master

slave manipulator is hydraulically powered and controlled by an operator, its radiation

resistance is greater than 10 5 Gy. The slave arm replicates the movements of the six

degrees of freedom master arm. Even thought no force feedback is provided, the master

arm was designed to increase the operator comfort and efficiency. The TROD system

was deployed for 7 months in 1990. It accomplished many tasks including removing

debris and barrels as well as decontaminating walls and floor. The use of TROD was

efficient and saved between 1180 and 1960 man mSv of personnel exposure. Cleaning

operations rarely require custom-made robots, like those in the previous example. In

most cases commercial robots are adequate, sometimes with a few modifications. That is

the case with the cleanup of radioactive tank nozzles by the ANDROS Mark VI robot at

the Susquehanna nuclear power plant [44]. In that case, the robot was introduced into a

waste-processing tank containing radioactive resin, where its goal was to unplug a flow-

mixing nozzle with high-pressure water. Remotec manufactures the ANDROS robot as

well as other general-purpose robots. This mobile platform includes 6 tracks, a

manipulator arm with a 16 kg maximum load and two video cameras. The unique design

of the tracks allows the ANDROS to move on very uneven ground. The control
                                             32


electronics are installed on board, sealed in the main body. No radiation hardening is

required since the dose rate in the tank was as low as 0.03-0.05 Gy/h. The operator

controls the ANDROS via a long and flexible cable. This cable includes the control lines

and video signals but no power lines since the robot carries its own batteries. ANDROS

was modified for this application by the addition of a high-pressure nozzle on its

mechanical arm. The robot was lowered into the tank through a man-way where an

additional camera was added to provide a general view of the tank internals. Once the

nozzles was unplugged, a test was performed to check them. The robot were then

removed and decontaminated. This application of ANDROS was successful and reduces

the personnel exposure from 30 mSv to 4 mSv. This application resulted in money

savings, plus the fact that the ANDROS can now be used for other applications in the

plant. In the next example, two robots; Scarab IIA and Scavenger, were used together for

a cleaning operation. In the Waterford 3 plant, about 1 cubic meter of spent resin tank

was spilled in a pump room [45]. A remote intervention was needed because of the high

level of radiation on the floor; up to 1 Sv/h. It was impossible to simply collect the waste

because of the many obstacles located in the room. A two-robot strategy was applied.

First the Scarab 2A gathered the waste at the lowest point of the room. Secondly the

Scarab robot used a water nozzle spray to slurry the waste while the Scavenger robot

vacuumed the floor. The waste was then evacuated to a shielded container in a room next

door. The operation was a success.

       Cleaning operations are far more complex after a nuclear accident. In Chernobyl

for example, the damages are so large and the radiation level so high that

decontamination of the site is not planned in the near future. After the Three Mile Island
                                             33


loss of coolant accident in 1979, fuel debris remained in the reactor. Most of the fuel

contained in the reactor was removed in the eighties, but about 20 tons of fuel remained

in a unreachable zone. The access of this debris area was very challenging since the fuel

rubble was located at the bottom of the reactor vessel, beneath the lower core support

structure. An innovative system debris called the airlift system (ALS) [46] was designed

to remove the debris. A long tube ending in a articulating nozzle was installed in the

vessel. The lower end of the tube including the nozzle was introduced in 171.5 mm

access hole to access the debris area. When air is injected into the nozzle, the fuel rubble

is displaced by the flow of air and water. A lift tube sucks out this mixture. If the

upward water velocity is higher than the falling velocity of the fuel debris in the tube then

the rubble is transported upward. The separation of the fuel and water is accomplished in

a separation chamber located above the nozzle, in the reactor vessel. The fuel is then

collected in a bucket that is regularly lifted and emptied. The ALS system was tested on

a full size mock-up at the Idaho National Engineering Laboratory. The results were very

encouraging and better than anticipated.



Tank Cleanup


       The DOE produced several dozen millions of gallons of liquid mixed waste

during the cold war. This waste is both highly toxic and highly radioactive. It came from

the production of nuclear weapons, with the chemical processing used in separation

facilities. At that time, the storage of the waste was not addressed because of the weak

environmental regulations, a shortage of money and the immediate priority of the arms

race. Single shell concrete tanks containing up to 4160 m3 of hazardous waste were built
                                            34


to address the storage of the waste for a short period of time. Such tanks are located at

the production sites owned by the DOE (e.g., Hanford Site, Oak Ridge Reservation,

Savannah River Site, Idaho National Engineering and Environmental Laboratory

(INEEL), Fernald Site, etc). These tanks have been in use much longer than originally

planned and many of these tanks are now leaking and are creating a major threat to the

environment. Such tanks do not meet today’s requirement for toxic and radioactive

storage. The waste contained in the single shell tanks has to be removed and stored in

more appropriate storage. The removal of the waste is challenging for many reasons.

First, the waste is extremely toxic, corrosive and radioactive, and is sometimes

inflammable as well as explosive. Second the tanks access is limited to only tank risers

as small as 18 inches in diameter. Third, the inside of the tank is sometimes filled with

pipes and equipment that limits the maneuverability inside the tank. This equipment

consists of cooling pipes, thermocouples and other monitoring devices. Fourth, a

requirement is that little or no new elements should be added to the waste volume. Fifth,

a decontamination procedure has to be created to avoid the spread of contaminants to the

environments. Finally, the consistency of the waste is very complex. Tanks contain

liquid, soft sludge and hard sludge. The composition of each tank is different; sometimes

a crust is formed on the top layer of waste. Trapped gases can sometimes move a

significant amount of waste. A sudden eruption of gas can damage the internal

equipment of the tank [47]. The removal of waste from in-tank storage is a long-term

process that starts with an evaluation of the tank and the definition of the waste. The

fluid part of the waste is then pumped outside the tank. Other methods are used to

remove the various forms of sludge.
                                             35


       The characterization of the tank containment is important to find leaks and to

evaluate the aging condition of the tank walls. The chemical composition of the waste,

its radiation activity and temperature must also be determined, as well as the sludge

consistency and depth. The DOE has developed remote tools that perform visual

inspection, ultrasonic testing and sampling. The easiest way to monitor the inside of a

tank is to lower a video camera through the 127 mm diameter of one of the tank risers.

This method has been used in cooled waste tanks at the Savannah River site [48]. In that

case, the camera was used to monitor the new in-tank precipitation process that added

high temperature and risks of explosion to the list of potential hazards. For reliability

reasons the system was kept very simple. The camera housing consists of a camera in a

vertical position and is directed to a mirror above it. Two powerful halogen lights are

installed under the mirror. The bottom part containing the mirror and the lights can rotate

providing a circular view of the tank. The mirror can also rotate on a horizontal axis and

is constantly heated to prevent fogging. The camera can be radiation-hardened up to 10 5

Gy or not, depending on the environment. The controls of the camera system consist of a

video monitor and switches for camera positioning. The camera housing can be lowered

in position by two systems. A telescopic metallic mast is often used but is too heavy in

many applications. The other system utilizes a motorized cable reel, which is

lightweight, compact and simple. The electrical cables feeding the video and the other

equipment are wrapped around the cable reel. Both systems have been tested on mock-

ups in 1992 and were then used in the field. The DOE Savannah River site has ever more

demanding requirements for tank inspection since the tank risers have a diameter of only

51 or 76 mm [49]. There are no appropriate commercial inspection tools for such small
                                              36


diameters. A few video cameras are available but the lighting is too weak for tank

inspection. The available sampling tools are not useful since they present risks of

contamination. Specialized equipment has therefore been designed for this unique

application. When introduced in the tank riser by a pole, the camera is maintained in a

horizontal position by an electromagnet. Once in the tank, the electromagnet is released

and the camera is secured horizontally by a spring while the lighting system rotates out of

the camera housing. Two visual inspection systems have been built; one for the 51 mm

inspection and the other one for the 76 mm inspection. Several sampling cups have also

been built with the appropriate bag-out and containment modules to prevent

contamination. The remote tank inspection system (RTI) is a five-degree of freedom

manipulator attached to a telescopic mast, which is lowered vertically from a 12 inch

tank-riser [50]. The vertical mast can go as deep as 12.5 m below the ground level and

the arm has a reach of 1.8 m. Several end effectors can be attached to the manipulator to

accomplish a wide variety of missions [51]. The visual inspection end effector consists

of two cameras, two lights and a light positioning system. In operation, the first camera

provides a general colored view of the environment and the second gives a high-

resolution close-up view of details in black and white. The ultrasonic end-effector is

used to detect corrosion or other defects in the tank walls or internal equipment. The

sampler end effector collects about 120 millimeter of waste at a determined depth.

Several samples can then assess the different layers of sludge. The characterization of

each sample will give the chemical composition, the activity, the density, the consistency

and the viscosity. Finally, the last end effector is a laser that gives a 3D view of the

inside of the tank to determine the location of foreign objects. The setup using the RTI
                                              37


(Remote Tank Inspection) system was tested in 1991 in a mock-up facility and was the

basis of a light-duty utility arm (LDUA) that was developed in the first half of the

nineties at the DOE Hanford site [52-53]. One of the improvements brought by the

LDUA was a bag out system that prevented the spread of contamination in the

environment during and after the removal of the robot from the tank. This work led to

the design of the modified light-duty utility arm (MLDUA) in the second half of the

nineties. Only the MLDUA has been used in an actual tank. The MLDUA description is

continued on page 40.

       Once the characterization of a tank and its sludge is realized, the removal of the

waste must be accomplished. For the majority of the tanks, the removal strategy is as

follows. First, the most fluid part of the waste is pumped out of the tank. The remaining

wastes left are the soft and hard sludges that cover about one foot of the tank’s bottom.

Second, a sluicing tool dislodges and mixes the sludge, then evacuates it. This sluicing

tool can be autonomous or be manipulated by the MLDUA or a mobile robot from inside

the tank. This mobile robot can also move the waste inside the tank to provide more

efficient pumping. Third, when the tank is empty the walls are scarified to remove the

thin contaminated layer of concrete that was in contact with the waste. There are no

major challenges with the transfer of liquids from one tank to another. Available

technology can be used without modifications. Plugging of the transfer mechanism must

be avoided however. A PulsAir system has been used in a few of the Oak Ridge tanks to

separate the waste components by particle weight [54]. The principle is to inject air

bubbles into the waste through large plates at the bottom of the tank. This makes the

lightweight particles available at the surface for safe transfer. After the liquids are
                                             38


removed, the sludge remains. A method, widely used now, uses pressurized water to

dislodge and mix the sludge that is then collected and pumped out of the tank. The

borehole-miner extendible nozzle sluicing system operates at a pressure up to 20.7 MPa

[55]. It consists of a telescopic mast lowered through a tank riser that supports a

maneuverable high-pressure nozzle. The nozzle can be extended up to 3 m from the mast

centerline. This hydraulic motor driven chain and rod system allows the nozzle to get

closer to the sludge or the walls. The entire mast can rotate on its axis to cover the

totality of the tank. The nozzle sprays clean water or recycled slurry in order to minimize

the volume of added waste. A jet pump integrated or separated from the mast then

collects the dislodged sludge. A remote control monitoring system has been installed to

assist the operator at his console. The information given by the encoders are also

converted into an intuitive graphical view of the inside of the tank. The borehole miner

extendible nozzle sluicing system was successfully used in the summer of 1998. Five

tanks were cleaned up in less than 3 weeks with less than 20% of waste dilution. A

Scarab III robot then inspected the tank [56]. The Scarab III robot is an evolution of the

Scarab II robot previously discussed in this document. The aluminum parts were

exchanged for stainless steel elements for a greater resistance to the corrosive sludge of

the tank. The cross section was also reduced to pass through the 18 inch tank riser.

Finally, metallic wheels replaced the tracks. These modifications resulted in an increase

in weight, which required more powerful motors. The role of the Scarab robot is to take

samples and to inspect the inside of the tank. To accomplish its mission the robot has a 2

degrees of freedom gripper that can handle small tools and objects. Three cameras are

installed on the unit. One monitors the back of the robot, particularly the tether. Another
                                             39


one is placed in the front to control the operation of the arm. These two cameras are

black and white, the last one is more sophisticated and provides a general view of the

robot’s environment. This inspection camera is mounted on a telescopic mast at the

central part of the robot. When the inspection of unreachable tank walls is required, the

telescopic mast is deployed up to 83 inches and provides the appropriate views. The

supporting equipment outside the tank consists of an operator console and the

deployment and containment module (DCM). The control desk consists of monitors,

switches, joysticks and a VCR in a relatively compact and easily portable station. The

DCM is used to lower the robot into the tank and assures the maintenance, repair and

decontamination. To realize this mission, the DCM unit is built around a glove box.

This containment cell includes the lifting mechanism, washers, sprays and repair tools

and is designed to minimize the deployment time. The SCARAB III unit provides an

easy, versatile and cost effective tool for tank inspection. Although it is very successful

for this mission, it is also limited by its low performance for inspection and sampling.

During a performance test, several sluicing tools were installed on the Scarab III and

revealed a poor efficiency in sludge removal. In that particular case, the Houdini robot,

(see page 40), is a more efficient choice.

       The confined sluicing end effector (CSEE) element is an alternative to the

borehole miner. The CSEE is used to dislodge and vacuum the sludge. In the CSEE,

nozzles are located at the end of three blades that are rotating. The nozzles eject

pressurized liquid or air in the direction of the inlet of a vacuuming tube. This tube is a

part of the hose management arm (HMA) that consists of a vertical mast connected to a

8.53 m long two section boom. Once lowered into the tank through one tank riser, the
                                             40


HMA provides pressurized fluid to the CSEE and pumps out the dislodged sludge to the

outside of the tank. Since the HMA is flexible but does not have an active mobile part, a

manipulator must position the CSEE in the tank. Two actuators are used; the MLDUA

and the Houdini robot. The MLDUA is the modified light duty utility arm. It is an 8

degrees of freedom hydraulic arm [57]. Its reach in the tank has a radius of 5.03 m and

has a maximum payload of 90.72 kg. The challenge of introducing such a massive

mechanical arm into the tank through the small tank riser has been accomplished by using

a massive vertical positioning mast (VPM). Not only does the VPM manage the

installation and removal of the MLDUA but it also prevents any contaminant from

polluting the environment. A hydraulic portable unit (HPU) provides the MLDUA with

hydraulic fluids. The HPU houses the hydraulic oil pumps, reservoir, filters, cooler and

also all the electronics that control the MLDUA. The operator of the MLDUA is located

in a trailer that provides video camera displays of the cleaning operation, joystick

controls and computer assistance. Like many manipulators, the MLDUA can be

programmed to scan an area of the tank. Its repeatability is better than 0.7 cm. The

MDLUA is efficient and reliable but has the inconvenience of being bulky, heavy and

costly. The Houdini system offers greater access to the internal part of the tank and an

improved efficiency with lightweight and versatile equipment. The Houdini system is a

mobile robot that is inserted into the tank in a folded position through tank risers as small

as 61 cm [58]. Once on the tank floor, Houdini is automatically deployed. It consists of

a mobile platform moved by two parallel tracks. A collapsible plow blade in front of the

robot can move sludge in the tank to the CSEE effector. A Schilling TITAN manipulator

is fixed on the platform. This mechanical arm can retrieve objects in the tank or
                                             41


manipulate the CSEE end effector. Two cameras and their lighting provide views of the

operations. A tether linked to the outside of the tank feeds the robot with hydraulic oil,

power lines and controls. This tether is covered partially in Kevlar and offers all the

strength needed for the installation and recovery of the robot. Outside of the tank, the

tether management and deployment system (TMADS) is located right above the tank

riser [59]. The TMDAS is used to lower and remove the robot in the tank. It is equipped

with a decontamination system and provides gloved access to the robot for repairs. The

TMDAS also manages the tether and is the interface between the tether, the hydraulic

lines, power lines and signal lines. The on-board electronics of Houdini is kept to a

minimum to avoid radiation damage and the control elements are located in the power

distribution and control unit (PDCU) instead. The PDCU manages all the electric lines of

Houdini. It provides the robot with the appropriate voltage and is the interface between

the robot and the operator console. The operator sits in a control trailer where critical

parameters like hydraulic pressure and filter condition are displayed. The internal views

of the tank are monitored while the operator controls the actuators of the robot with a

joystick and switches. The Houdini system is currently in its second version. The first

version called Houdini 1 was delivered to Oak Ridge National Laboratory in September

1996, where it underwent testing in a cold mock-up facility. The first “hot” application

occurred in June 1997. Several improvements aimed to increase the reliability of the

system led to the second version of the system called Houdini 2 [60]. The Houdini has

shown an impressive efficiency especially when associated with the MDLUA. The

CSEE is not the only end effector that is used by both the MLDUA and the Houdini. The

characterization end effector (CEE) takes samples of the tank wall to later determine its
                                             42


radioactivity and physical degradation. The purpose of the gunite scarifying end effector

(GSEE) is to remove the thin layer of material on the wall that was contaminated by the

liquid waste. When a lattice of pipes covers the inside tank, the limited access does not

allow any use of robotics inside the tank. Five C-tanks at the Oak Ridge site have been

cleaned up with a unique technique. In this technique, the retrieval equipment is

completely external to the tank and consists of pumps and two charge vessels [61]. The

technique carries out a two-step process to remove the waste and the sludge. In the first

part of the process, the liquid waste is pumped into the charge vessels. The second part,

the same volume of liquid is re-injected into the tank through two high-pressure nozzles.

The movement of liquid dislodges the sludge and mixes the waste. If this process is

cycled many times, no sludge remains attached to the bottom of the tank and after the last

cycle, all the waste is simply pumped out of the tank. The cleaning of the C-tank W-21,

W-22 and W-23 by this method was successfully carried out. Up to 95% of the waste

was removed by this pulse-jet mixing technique. The remaining sludge is then cleaned

by a more aggressive technique. This method requires little maintenance since almost no

moving parts are involved. The hardware however, is bulky and heavy and needs

extensive decontamination. Another system, the Tarzan locomotor, avoids these

drawbacks. Tarzan is a mobile platform designed to deliver a Schilling manipulator into

a full tank containing a dense array of vertical pipes [62]. The principle of motion is

simple; two grippers at each ends of the robot help Tarzan to grasp the pipes. Once

attached to a column, Tarzan deploys itself and grasps the next one. At this point, it

releases the previous gripper and looks for the next supporting pipe. When the working

location in the tank is reached, the two grippers are secured in position and the
                                             43


manipulator arm can operate. Tarzan is hydraulically powered; it includes three high

torque actuators for in-plan motion and one vertical translator. Three cameras provide

views of the operations. Each gripper requires two different views: an overview of the

room to locate the next pipe and a close-up view for an accurate grasp of the pipe. A

single camera and a mirror offers both views simultaneously and a monitor displays these

views. Each of the two grippers is equipped with this system. An additional camera

monitors the manipulator arm operation. The robot is linked to the operator and to the

hydraulic unit by a flexible tether. The cameras use a black and white image tube and

can operate without failure at a dose rate of 10 3 Gy/h and up to 10 6 Gy of total dose. The

entire robot is designed to tolerate a total dose of 105 Gy in the very harsh environments

of the tank. The first application of Tarzan is planned for the tanks of the West Valley

Demonstration Project in the state of New York.



Decommissioning


       With many nuclear facilities reaching the end of their lifetimes, the

decommissioning effort is a new priority. The reduction of employee exposure as well as

cost reduction requires the uses of remote tools for contaminated parts [63]. Once the

core elements are removed, the decommissioning of the rest of the building can be

accomplished in a classical way. The classical decommissioning strategy is to dismantle

large structures from top to bottom, then transport the components to a disassembly line

where they are then cut into smaller pieces and packaged. For example the dismantling

of the Niederaichbach power plant in Germany followed the following steps: dismantling

of upper neutron shield, removal of pressure tube units, dismantling of lower neutron
                                             44


shield, dismantling of moderator vessel and dismantling of thermal shield [64]. The

Idaho National Laboratory has developed a computer based planning system called

DDROPS (Decontamination, Decommissioning, and Remediation Optimal Planning

System) [65] that plans and optimizes the decommissioning operation. As a result, an

improved robotic pathway is generated and personnel exposure is reduced which leads to

savings in time and money. The remote dismantling is a very difficult operation for

remote operators. These operators need to maneuver delicate tools accurately by using a

video screen or sometimes though a thick leaded glass window. They do not have the

direct “feeling” of the tool like any other worker. In those conditions, even unscrewing a

bolt requires all the dexterity and concentration of the operator. The decommissioning of

a facility is a long working operation. The remote equipment is constantly running on the

job for a long period of time. To avoid extra costs and excessive maintenance time, the

robot must be made of robust and easily changeable off-the-shelf parts. The cutting,

sawing, drilling of small metal parts, in the disassembly line in particular, can be

accomplished by modified off-the-shelf tools. This means that holding arms must be

added as well as video cameras and that parameters such as unit current, voltage or

hydraulic pressure are monitored. The removal of elements directly from the reactor

structure is too specific to be realized by available technology. A heavy-duty dismantling

robot must be designed to ensure efficient and reliable remote operations. During the

decommissioning of the Niederaichbach unit, the following remote tools were developed.

The main development was a rotary manipulator that could use one of 63 configurations

or tools to dismantle the inside of the reactor. The crane manipulator was designed to

remove the parts disassembled by the rotary manipulator and to transport them to the
                                             45


disassembly line. The ring saw, located under the reactor vessel, cut and operated with

the rotary manipulator to dismantle the cylindrical steel moderator vessel. The ring saw

was then replaced by the band saw to remove the thermal shield that protected the

moderator vessel. The decommissioning of the CP-5 research reactor at the Argonne

National Laboratory in 1997 had a different strategy. The idea was to use the same

manipulators on different platforms. This system is called dual arm work module

(DAWM) [66-67]; it consists of two Schilling Titan II mounted on a 5 degree of freedom

articulation platform. The DAWM module can be installed on a ground mobile vehicle

called Rosie, or on a platform suspended to an overhead boom or to a crane. This last

configuration was used in the CP-5 reactor and renamed the dual arm work platform

(DAWP). In the CP-5 case, the platform was suspended by a crane and used Titan III

manipulators that were easily decontaminated. The DAWP robot undertook the complete

dismantling of the reactor. Its operation also involves the cutting of the metallic parts and

removal of the graphite blocks and lead bricks. The DAWM and DAWP platforms

include 5 on-board computers based on a UNIX development system. Several cameras

provide a good view of the manipulator operation. The electronics did not have a

radiation-hardening requirement. A bundle tether contains the power, signal and control

lines. The operator is located in a low-radiation room where monitors, graphical

interfaces and joysticks and switches assist the operation. The decommissioning of the

CP-5 reactor has been successfully completed with the DAWP system. The main area for

improvement in the next application is the user interface, which is critical for a lengthy

operation like decommissioning. A similar system is planned for the remote dismantling

of a spent fuel reprocessing facility in Karlsruhe, Germany [68]. In this building, several
                                            46


hot cells are accessible from the top. A dual master/slave arm system supported by a

crane will operate in the hot cell until completion of decommissioning. This system is

called the manipulator carrier system (MCS). The mechanical arms are two

electromechanical master slave manipulators (EMSM) with 8 degrees of freedom, a load

capacity of 100 kg and a reach of 2.8 meters. Several tools are available for the EMSM:

a shear, grinder, saw, drill-machine; spray for decontamination and radiation detectors.

The unit also supports video cameras, microphones and other measuring devices. All the

major unit of the MCS and the crane are redundant to avoid a need to intervene in the

contaminated environment. This equipment has been improved during and after realistic

tests in a mock-up facility carried out from 1995 to 1997. The actual demolition of the

hot facility will be achieved in 2003. The robots that were described above are designed

for the complex job of dismantling the reactor vessel. There is a requirement for a track-

based excavator to manipulate the rubble in a more efficient manner as well. The Haz-

Trak robot can meet this requirement [69]. This vehicle is similar to a commercial

excavator; it has the required power and efficiency. A single operator remotely controls

this robot. Force feedback allows the user to “feel” the reaction of the machine when it

finds buried objects for example. The control electronics are located inside the machine

and has no radiation hardening requirements. Modular design however, allows an easy

replacement of the equipment. The robot is power independent because it uses its own

diesel motor. The communication link to the operator is either an optic fiber or a RF link.

The operator has access to several remote views of the operations and also has displays of

the robot parameters; oil pressure, temperature, etc. The control desk features the usual

tools of an operating desk: joysticks, switches, monitors, graphical user interface and
                                            47


VCR. A task recall feature is installed in addition to the force feedback system. This

capability allows the operator to “teach” the robot up to 255 routines some running as

long as 10 minutes. These routines are then executed automatically, under the

supervision of the operator. The weakness of the HAZ-TRACK is its low radiation

hardness that limits its uses to the final phase of decommissioning. A German firm: Mak

System Gmbh has designed a heavy manipulator vehicle that is radiation resistant up to

10 kGy from Cs 137 by use of tungsten shielding [70]. This vehicle, called SMF, uses a

mobile platform similar to a small tank. A powerful diesel engine drives the tracks over

big obstacles. A massive 6 axis hydraulic arm is mounted on the platform. This

manipulator can lift up to 250 kg and has a reach of 3 meters. The end-effector of this

arm can be a gripper as well as any required decommissioning tool. A radio-frequency

link from the robot to the control unit can be located as far as 1 km from the vehicle and

up to 10 km with additional relays. The SMF unit includes four radiation resistant

cameras. Two computers are used to manage the modules of the robot. A special

tungsten container protects the electronic system from Cs 137 and allows for a reliable

operation in a radiation environment as hazardous as 100 Gy/h. An additional back-up

computer provides duplicate protection and allows the recovery of the vehicle in case of

failure of the master computer. A high-level data exchange system offers reliable

communication with the operator. A fail-safe concept has been followed during the

design of the SMF vehicle and guarantees the reliability of the unit without the use of

radiation-hardened components. The SMF robot has been available since March 1994

from Kerntechnische Hilfsdienst Gmbh in Germany.
                                             48


                                 Post Accident Operation


       When an unexpected problem occurs in a radioactive environment, it becomes a

potential hazard. A fast and versatile tool for situation assessment is needed without

putting personnel at risk. General-purpose robots like the Remotec ANDROS or Rovtech

Scarab are fitted for such jobs. Their versatility comes from their ability to navigate in

unstructured environments, their efficient vision system and their mechanical arm, which

allows for a wide range of actions [71]. The Chernobyl accident on April 26th 1986 has

stressed the need for robots for intervention purposes in cases of nuclear accidents. The

first robots that were sent to Chernobyl failed because of a lack of radiation hardness or

because their tethers became stuck in the rubble. The Russians were the first to design a

robot especially for the site of the Chernobyl accident. The Mobot-ChHV was on site as

soon as August 1986 and successfully cleaned the roof [72]. It was equipped with

electromechanical actuators, but did not have any on-board electronics. In the following

years many robots have been developed and have accomplished a wide variety of tasks at

Chernobyl [73]. Among them are a video inspection robot, a boring and drilling robot, a

dust and air cleaning robot and a few dismantling robots. The latest robot designed to

explore the inside of the Chernobyl sarcophagus is the result of US-Russian

collaboration. This robot is called Pioneer; it uses many features of the successful

Redzone Houdini vehicle [74]. Pioneer uses a fully electric, track-based platform. The

modularity of the vehicle allows easy transport of each separate module into the

Chernobyl building and final assembly close to the work location. Pioneer carries a

remote viewing system, a concrete sampling drill, a manipulator arm, a sensor package

and a plow bucket in the front of the vehicle. The original goal of this unit was to take
                                             49


concrete samples from the wall and floor. The concrete sarcophagus that was built after

the accident seems to have a structural weakness due to the high radiation level.

However, radiation is not the only challenge for a vehicle in the sarcophagus; light is

almost absent and the rubble of the accident also makes the building unfriendly for

remote operation. The robot contains a viewing system with a color camera radiation-

hardened up to 10 MGy of total dose and a 1 kGy/h dose rate. Three additional cameras,

shielded, with lead provide a 3D map of the environment. These cameras and the four

150 W lights are adjustable and are controlled by the operator. The concrete sampling

system consists of a rotational motor, a linear thrust actuator, a 6 axis force-torque sensor

and a diamond cutting bit [75]. The drilling machine was designed with a priority for

reliability. The system was kept as simple as possible; the on-board electronics were

minimized and designed to operate without failure to an accumulated dose of 10 kGy and

a dose-rate of 35 Gy/h. The entire sampling drill can be mounted vertically or

horizontally to sample the floor or the wall. The drilling system is controlled remotely by

the operator. The environmental sensor package measures the temperature, humidity, and

gamma radiation level as well as thermal and epithermal neutron fields. The plow bucket

is used to clear the way or to move up to 91 kg of debris at a time. The manipulator has

the same 6 degrees of freedom arm as the ANDROS Mark V-A robot and has a

maximum payload of 45 kg and a reach of 1.68 m. Like the Houdini vehicle the control

electronics of Pioneer is not onboard but in the control station to avoid radiation damage.

However, the viewing system, the drilling tool and the mechanical arm control

electronics are installed onboard. A tungsten shielding box protects them against

radiation from Cs 137 and allows a total hardness of 1 kGy to 10 kGy total dose depending
                                            50


on the incidence of the gamma field. The viewing system, the moving platform, the

mechanical arm and the drilling system are powered and controlled via a tether to a

power distribution and control unit (PDCU). The PDCU is separated into different

modules for easier transport. The control console provides all the tools for easy

operation: monitors, switches, joysticks and graphical user interfaces. The maximum

distance between the operator and the robot is 500 m.
                                     CHAPTER 3
                                 RADIATION EFFECTS




                       Definition and Units in Nuclear Engineering



Radioactivity


       Radioactivity occurs with the natural decay of an unstable atomic species into a

more stable specie. The decay of an atom is usually associated with the emission of

particles or photons. This emission of particles or photons is called radioactivity and the

initial atom is qualified as radioactive. The decaying of a radioactive atom can yield

other radioactive atoms. This succession of decay can last several centuries and finally

ends with a stable element.



Activity


       The number of disintegration of a radioactive source is measured by its activity.

A source activity of one becquerel (Bq) indicates that only one atom of the source

disintegrates per second. Several particles may be emitted per decay; however it does not

mean that only one particle is emitted per second. The becquerel is the legal unit but is

very small. Another unit of measurement is called the curie (Ci): 1 Ci = 3.7 1010 Bq.

The curie is the amount of radiation from 1 gram of radium.




                                             51
                                              52




Decay constant, Mean-life, Half-life


         The decay rate is measured by the decay constant λ; the probability of particle

decay per unit time. The decay constant is an absolute constant for each particular

radionuclide. There is no variation with other parameters such as temperature or

pressure. The mean life τ of a radionuclide is then τ = λ -1. The half-life T of an element

is the amount of time needed to decay to half of the initial number of a radionuclide

without adding any new element. The relation between the half-life and the decay

constant is: T = ( ln 2 ) / λ.



Energy


         When a particle or photon is emitted after decay or any other reaction it has a

specific energy E. For photons, this energy is proportional to their frequency ν: E = h ν.

For material particles like electrons, this energy is kinetic energy. In nuclear engineering,

the energy of particles is measured in electron Volt (eV). The legal unit is the Joule (J),

and the relationship between the two units is 1 eV = 1.6 10 -19 J.



Dosimetry


         The interaction of particles and the matter they travel into causes a loss of energy.

The energy absorbed by a material per unit of mass of this material is called the dose.

One Gray (Gy) is the absorption of one joule per kg of material. A non-legal but

common unit of dose is the rad (rad), the correspondence is 1 Gy = 100 rad. A dose is
                                              53


always referenced to a mass of material. Consequently, 1 Gy in water does not represent

the same effect as 1 Gy in silicon. The radiation effects on the human body depend on

the type of particle. The quality factor (Q.F.) is multiplied by the absorbed dose in Grays

to obtain the equivalent dose in sievert (Sv).

•   For gamma, X-rays and beta         Q.F.=1          1 Gy ⇔ 1 Sv
•   For neutron and protons            Q.F.=10         1 Gy ⇔ 10 Sv
•   For alpha particle                 Q.F.=20         1 Gy ⇔ 20 Sv

Another unit commonly used for equivalent dose is the rem: 1 Sv = 100 rem.




                          Types of Radiation and their Interaction



Photons: Gamma and X-rays



       Definition. Gamma rays and X-rays are identical: they both are photons with very

short wavelengths. The only difference is their origin; gamma rays come from a nuclear

interaction, X-rays come from electronic or charged-particle collision. They interact

identically: they lightly ionize and penetrate deeply into the matter; however, they do not

create any activity after interaction as long as their energy is less than 10 MeV. The

number of photons exponentially decreases with the target thickness. When a collimated

beam of mono-energetic photons enter perpendicularly into a target with an intensity I0,

the intensity of uncollided photons at depth x of the target is I (x) = I0 e-µx. The

coefficient µ is the attenuation coefficient and depends on the photon energy and on the

target, especially the electron density in the target. The only protection against photons is

a thick sheet of material with a large charge atomic number Z, like lead. A way of
                                            54


comparing the absorption of photons in different material is to use the tenth value

thickness (T.V.T.). The T.V.T. is the thickness of material needed to attenuate by a

factor of ten a collimated beam of mono-energetic and uncollided photons. Table 9 gives

several values of TVT for different materials and photon energies.




Table 9: Photon tenth value thickness in cm for Al, Fe, Pb and concrete
Energy (MeV)       Aluminum          Iron               Lead               Concrete
0.05               21                1.6                0.25               18
0.1                50                8.1                0.37               49
0.2                70                20                 2.03               77
0.5                101               35                 13                 111
1                  139               49                 29                 153
1.5                170               60                 39                 188
2                  198               69                 44                 218
3                  241               81                 48                 265
4                  274               88                 48                 306
5                  300               93                 47                 338
Source of data: see references [1]

       Three principal types of photon interaction occur: photoelectric effect, photon

scattering and pair production.



       Photoelectric effect. The photoelectric effect is the absorption of the incoming

photon’s energy by an outer shell electron. This electron is then ejected from the atom

with kinetic energy equal to the difference of the photon energy and the electron binding

energy. This interaction also results in emission of luminescence X-rays and Auger

electrons. The photoelectric effect is the dominant interaction for low energy photons (<

0.5 MeV).
                                             55



       Photon scattering. Photon scattering is by definition the scattering of an incoming

photon by an electron. This scattering can be coherent (the photon energy is conserved)

or incoherent (the photon energy is partially transferred to the electron). In both cases the

photon has its trajectory modified and the electron is ejected from the atom. The most

common scattering is Compton scattering.



       Pair production. This interaction is dominant at high energy and occurs only if

the photon energy is greater than 1.022 MeV. In the electric field of a nucleus or an

electron, a photon is spontaneously annihilated and converted into a electron-positron

pair. The positron and the electron have a total kinetic energy equal to the difference of

the initial photon energy and 1.022 MeV.



Beta: Electron and Positron


       A β- particle is a free electron, a β+ particle is a positron. Apositron is what is

known as antimatter; a positron has the same weight as an electron but its charge is the

exact opposite. A positron does not travel very far because it is quickly annihilated by an

electron from the material and results in two photons. When travelling into a material,

the Coulomb force of the bond electrons interacts with the β particles. The exchange of

energy that occurs excites the atoms or ionizes them. In that case the energy is

transferred to the ejected/excited electron. The succession of accelerations and

decelerations also generates the emission of photons called Bremsstrahlung photons. The
                                            56


range of beta particles is limited; usually a simple aluminum sheet can stop them (Table

10). Beta particles do not create any new radioactivity in a material.




Table 10: Range of electron in aluminum
Energy (MeV) 0.01          0.03    0.05 0.1         0.25    0.5      1       3         5
Range (µm)        0.06     0.6     1    5           20      60       160     550       940
Source of data: see references [1]


Heavy Charged Particle


       Heavy charged particles are ions like protons (H1+), alpha particles (He 4++) or any

other ionized atom. These particles are absorbed principally by scattering from an atomic

electron and from atomic nuclei. The stopping power is the rate of energy loss of the ion

per unit length, the mean range is defined as the distance the ion travels before coming to

rest. This range is usually small and does not exceed a tenth of a millimeter for alpha

particles (Table 11) and a millimeter for protons (Table 12). Heavy charged particles do

not create any new radioactivity in a material. Heavy charged particles that meet in

nuclear interactions are stopped by the plastic package around components.

Consequently they are not a threat to electronic systems. The effects on plastic or surface

can be significant however.
                                             57


Table 11: Range (µm) of alpha particles in Al, Pb, water and air
Energy (MeV)       Aluminum         Lead               Water                Air
0.01               0.1              0.05               0.2                  240
0.1                0.6              0.4                1                    1330
0.5                1.8              1.5                3                    3310
1                  3.3              2.5                5                    5520
2                  6.6              4.6                11                   10800
5                  22               14                 37                   36700
Source of data: see references [1]



Table 12: Range of protons in aluminum
Energy (MeV) 0.1             0.3     0.5           1.         3          5          10
Range (µm)        0.7        2.7     5.3           14.5       78         182        604
Source of data: see references [1]


Neutron


       Neutrons are uncharged particles with approximately the same mass as protons.

Neutrons are classified into three categories depending on their energy; thermal neutrons

(0.025<E < 0.5 eV), intermediate neutrons (0.5<E < 10 keV) and fast neutrons (E>10

keV). Since Coulomb forces cannot interact, neutrons are very difficult to stop. Fast

neutrons lose their energy by elastic scattering with the atoms. The transfer of energy is

greatest when the neutron collides with a hydrogen atom. This explains why materials

containing a lot of hydrogen, like water, are the best shields against neutron. Fast

neutrons are slowed down after multiple scattering; they become thermal neutrons. The

probability of nuclear reactions in this range of energy is much higher. The number of

protons or neutrons of the target nucleus can be modified and lead to radioactivity if this

new element is unstable. A nuclear reaction also generates a number of other particles

like gamma, beta or alpha. The damage created by gamma rays on robotics and
                                              58


electronic systems, are far more damaging than the damage by neutrons. In addition

neutrons are much less common than gamma rays. This explains why the effects of

neutrons are not a concern when it comes to designing a robotic or electronic device for

nuclear application. The only exceptions are the in-core and near-core applications.



Conclusion


        The previous paragraph presented radiation types and their interaction with

matter. This presentation shows that gamma rays are the biggest threat to robotics and

electronic systems for terrestrial nuclear environments. Alpha and beta particles can be

stopped with a light shield and are therefore non-threatening to the reliability of a system.

This is not true outer space.

        When gamma rays interact with material, they create two effects. The first effect

is ionization. Photoelectric effect, Compton scattering and pair production eject electrons

from the atoms of the material. These ejected electrons can create secondary reactions.

The result is a track of ionized atoms in the bulk of the material. The second effect is

atomic displacement. Sometimes the atom receives so much kinetic energy at the site of

interaction that it leaves its initial location in the material. This displacement creates

additional atomic movement on its track that may result in a cluster of defects into the

atomic lattice. The immediate and long-term results of ionization and atomic

displacement strongly depend on the material. The next few paragraphs will discuss

these effects.
                                             59


                          Radiation Effects on Passive Elements


       This section describes the effects of gamma rays on a wide range of materials

other than semiconductors. It is extremely difficult to define a level of failure. The

performance of a material is defined with many physical properties: mechanical,

electrical, thermal, optical, etc. A slight change in any of these properties may cause no

effect or may have tremendous effects on a system. Therefore the following values of

radiation damage threshold are only an estimation of radiation effects on materials. The

data presented in this chapter come from various origins. K.U. Vandergriff from Oak

Ridge National Lab has published an extensive set of data on radiation damages on

materials [76]. Several other sources give some interesting information on radiation

effects [1,3,77,78].



Inorganic Materials



       Metals. Metals are immunized against damages from gamma rays. The metallic

structure is very resistant to radiation. Exposition of metals to very high dose rate

generates some heat that may be indirectly damaging to the system. After long-term

exposure (several decades) to a very high dose rate, some defects may be detected like an

increase in tensile and yield strength and a decrease in ductility. These defects can be

annealed, and the metal would recover its mechanical properties. Table 13 gives some

damage thresholds for commonly used metals.
                                              60


                     Table 13: Radiation damage thresholds on metals
                     Metal                       Threshold level (Gy)
                     Aluminum and its alloys 5. 1011
                     300 series stainless steel 1. 1011
                     400 series stainless steel 5. 1010
                     Iron                        3. 1010
                     Copper                      2. 1010
                     Brass and bronze            1. 1010
                     Nickel and its alloys       1. 1010
                     Beryllium copper            6. 109
                       Source of data: see references [76]


        Ceramics. Ceramics are used as a dielectric in capacitors, and also as coatings to

replace plastic coatings. Ceramics are more resistant to radiation than organic materials

but are not as tolerant as metals. The radiation effects on ceramics are a dimensional

swelling and therefore a decrease of the density. Table 14 shows some damage

thresholds for common ceramics.




                    Table 14: Radiation damage thresholds on ceramics
                       Ceramics               Threshold level (Gy)
                       Alumina                5. 1010
                       Silicon carbide        6. 108
                       Mica                   5. 107
                       Quartz                 2. 107
                       Glass, flint           2.5 105
                       Glass, borosilicate 1. 105
                       “Vycor” glass          5. 104
                       Source of data: see references [76]

        See the next paragraphs for a description of the radiation effects on specific

materials like glass, crystals and optical fibers.
                                             61


Organic Materials



       Polymers and plastics. Polymers are long, chain-like molecules, with a relatively

small number of chains per unit volume. Radiation produces two effects on polymers:

cross-linking and chain-scission. Cross-linking occurs when two molecules are bound

together by the radiation effects on electrons. The opposite effect also occurs: chain-

scission, which shortens molecules to smaller chains. Both of these effects occur

simultaneously in polymers. The predominance of one effect varies with polymers and

experimental conditions. Few plastics are actually irradiated to improve their mechanical

properties but in most cases radiation degrades plastics. The damages result in cracking,

blistering, embrittlement and an increased sensitivity to mechanical stress. Some

polymers are particularly sensitive to radiation. Teflon (PTFE: polytetrafluorethylene) is

degraded after only 100 Gy. The use of Teflon is prohibited in radiation environments.

Halogenated polymers and fluorocarbons release corrosive chemical like HCL and HF

when irradiated. Not only do these gases affect the polymer but also the other equipment.

Such polymer categories includes PVC, PVDF, Teflon and Viton. PVC does not suffer

important degradation under irradiation but the hydrogen chloride slowly shows its

effects and completely destroys the integrity of the plastic. Temperature influences the

rate of degradation as well as the composition of air. An increased concentration of

oxygen in the air produces more oxidation and increases the number of chain-scission.

Table 15 gives the radiation damage classifications of a few plastics. See reference [76]

for a complete set of radiation test data.
                                            62


                         Table 15: Radiation tolerance of plastics
            Radiation resistance              Polymer
                                              Glass fiber phenolics
                                              Asbestos filled phenolics
                                              Epoxy systems
                                              Polyurethane
            Highest radiation resistance      Polystyrene
                                              Mineral filled polyester
                                              Mineral filled silicones
                                              Furane-type resins
                                              Polyvinyl carbazole
                                              Polyethylene
                                              Melamine-formaldehyde resins
                                              Urea formaldehyde resins
            Moderate radiation resistance
                                              Aniline formaldehyde resins
                                              Urfilled phenolic resins
                                              Silicone resins
                                              Methyl metacrylate
                                              Unfilled polyesters
            Poor radiation resistance         Cellulosic
                                              Polyamides
                                              Teflon
              Source of data: see references [76]


       Coatings. Organic coatings consist of a thin polymeric film, used for esthetic

purposes and protection against corrosion. They are also used in radiation environments

to provide an easier decontamination. Like any polymer, organic coatings suffer

degradation from radiation. The damages may result in cracking, blistering,

embrittlement and/or surface flaking. The degradation on coatings varies a lot with their

composition and depends on factors such as; temperature, type and preparation of the

surface. Table 16 indicates the radiation damages on several coatings for different

surfaces.
                                            63


Table 16: Radiation damages on coatings
Polymer base            Surface                Dose (Gy)              Damages
Epoxy                   Steel                  6.7 106                No failure
                        Concrete               9.4 106                No failure
Furan
                        Steel rod              8.4 106                No failure
                        Concrete               9.4 106                No failure
Modified Phenolic
                        Steel rod              8.7 106                Severely embrittled
                        Concrete               6.7 106                No failure
Silicon alkyd
                        Steel                  6.7 106                No failure
                        Concrete               8.7 106                Failed blistered
Styrene                 Steel                  8.7 106                Failed cracked
                        Steel (wet)            8. 105                 Failed cracked
                        Aluminum               2.1 106                Failed blistered
Vinyl
                        Concrete               1.1 107                Borderline failure
                        Aluminum               2.1 106                Failed blistered
Vinyl chloride          Concrete               1.1 107                Failed blistered
                        Steel                  8.7 106                Failed blistered
Source of data: see references [76]

       Unlike any other material, coatings are threatened by alpha particles. Alpha

particles have a limited range that is comparable to the thickness of the coating film. The

entire bulk of polymers may then be damaged by the alpha particles. Alpha particles are

more ionizing than gamma rays and therefore create more damage. If coatings are

certified with gamma particles, the damages will show up earlier than expected when

used in alpha environments.



       Adhesives. Adhesives are organic compounds that are used to hold two structures

together. Adhesives are under mechanical stress when loaded. Radiation damages the

chemicals in the adhesives and decreases the number of adhesive bonds. Some adhesives

are more radiation sensitive than others are. The degradation may be accelerated with

vibrations, temperature and/or a higher concentration of oxygen in the air. Table 17

illustrates the radiation damage of a few useful adhesives.
                                              64


Table 17: Radiation damages on adhesives
Adhesives                                            Radiation damage threshold (Gy)
Neoprene-nylon-phenolic                              5. 105
Neoprene-phenolic                                    106
Epoxy, epoxy-thiokol, nitrile-phenolic               5. 106
Epoxy-phenolic, vinyl-phenolic, nylon-phenolic       107
Source of data: see references [76]


        Elastomers. Elastomers are polymers used for their compression and elongation

properties. Elastomers are used for seals, o-rings, gaskets, diaphragms and insulation.

Mechanical properties like tensile strength, compression set and elongation at break are

affected by radiation. The radiation degradation depends on the polymer base and the

type and concentration of additives. Some additives like amines and phenols may also

protect elastomers from the effects of radiation. Radiation damages are unlikely to occur

on elastomers after a total dose less than 10 kGy. See reference [76] for a complete set of

radiation test data.



        Lubricants. Oil and grease are used in moving mechanisms to reduce friction, but

also to cool and to prevent corrosion. Lubricants are organic materials made of natural or

synthetic oil and a minority of additives. The type and concentration of additives control

the characteristic of the lubricant: viscosity, thermal conductivity, heat capacity,

corrosiveness, chemical and temperature stability. These parameters are affected by

radiation because lubricants are radiation sensitive like any other organic compound. The

results of the chemical degradation of the organic molecules are an increase in viscosity

that may ultimately lead to a polymerization to a solid state and a destruction of the

additives that result in modified physical properties. It is found that synthetic lubricants
                                             65


are more radiation resistant than natural lubricants but few exceptions exist. The more

radiation tolerant lubricant are polyphenyls, poly(phenyl ethers) and alkylaromatics, but it

is often better to operate unlubricated in very high dose rate environments. Table 18

gives a few examples of radiation damage threshold of lubricants. See reference [76] for

a complete set of radiation test data.




Table 18: Radiation effects on lubricants
Radiation damage threshold (Gy)       Lubricant
  4
10 and below                          No significant radiation damages
  4      5
10 to 10                              Aromatic phosphates, silicones, aliphatic esters
  5      6
10 to 10                              Diesters, aromatic esters
  6      7
10 to 10                              Mineral oils, aliphatic ethers
  7      8
10 to 10                              Alkylaromatics, poly(poly ethers), polyphenyls
  8
10 and above                          No lubricant available
Source of data: see references [76]


Optical Material


       Optical materials are transparent because the energy gap between the valence

band and the conducting band is greater than 3.1 eV. Light photons whose energy is less

than 3.1 eV cannot excite carriers. Therefore the material does not absorb the

corresponding light wavelengths. Radiation creates defects that have new energy levels

in the gap, allowing greater carrier excitation and light absorption. This effect is called

darkening or browning of the transparent material under radiation.



       Glass window. Remote operators visualize the inside of a hot cell through a thick

leaded window that absorbs radiation. This window also attenuates visible light,
                                             66


requiring a very powerful lighting source in the hot-cell. Any darkening of a window due

to radiation affects the ability of the operator and is undesirable. The addition of less

than 2.5% of cerium oxide in the glass composition prevents the effects of darkening.

Unfortunately this additive and the lead produce a yellow tint that increases with the lead

concentration and affects visibility. A hardening technique is to put several glass sheets

between a cell and the outside with increasing lead concentration and decreasing cerium

concentration. Another drawback of the cerium and lead mix is the risk of electrostatic

discharge (ESD). When exposed to very high dose rate, a leaded window may suddenly

release its accumulated static charge. This ESD produces cracks that may affect the

integrity of the window. An ESD is more likely to happen on windows with low

conductivity. Therefore the addition of conducting elements to a window composition

reduce the risk of ESD. Table 19 presents the effects of radiation on several natural and

cerium-treated glasses.




Table 19: Radiation damages on window glasses
                          Total     Average light        Light transmission after dose at
Material                  dose      transmission              various wavelengths
                          MGy       before dose       400 nm 500 nm 600 nm 700 nm
Crown-2                   10        98%               0%        3%        25%       46%
Crown-2 protected         10        98%               60%       86%       88%       89%
Dense flint-2             50        94%               0%        1%        11%       21%
Dense flint-2 protected 10          91%               45%       83%       85%       86%
Purified fused silica     50        100%              89%       89%       89%       89%
Quartz                    10        99%               35%       30%       31%       56%
Styron 690                10        75%               0%        2%        28%       56%
Vycor                     0.2       99%               0%        0%        0%        1%
Vycor protected           5         99%               24%       24%       36%       61%
Source of data: see references [76]
                                             67



       Camera lens. The radiation tolerance of hardened CCD camera range from 100

Gy to 10 kGy. Camera lenses are altered by the same darkening effect as the window

glass, but are more tolerant than the camera’s electronics. The quality of the anti

darkening treatment is adapted to the expected lifetime of a camera to minimize costs and

maintenance.



       Optical fiber. Today remote operations require several dozens of transmission

and signal lines between an operator and a robotic system. Due to thick cables containing

a great number of conductors, the mobility and the maneuverability of a robot is limited.

It is very useful to replace such cables with optical fibers. Optical fibers have many

advantages compared to metallic wires. A single fiber can carry as much information

that thousand of wires, it is less sensitive to perturbations and does not have any ohmic

attenuation. Unfortunately radiation induces an optical attenuation even at low dose.

Commercial optical fibers have been tested and showed degradations up to 100

dB/km/Gy. The majority of optical fibers are therefore not suitable for applications in

nuclear environments. It is found that pure-silicate optical fibers are more tolerant to

radiation. When the core of the fiber does not contain any doping, the attenuation is

greatly decreased. This improvement is particularly significant for the infrared

wavelengths where most of the transmissions occur [79]. Such radiation-hardened

optical fibers can show an attenuation of 0.1 dB/m after 1 MGy of exposure. This type of

fiber can be very useful where short length of fibers are needed like in hot cells. There

are other performance factors that also encourage the use of optical fibers. First, the

attenuation saturates with the exposure, typically after 1 MGy. Second, an almost
                                              68


complete recovery occurs at room temperature after less than one week without

exposition to radiation. This is very interesting for maintenance robots that do not stay

permanently in a radiation field. Third, a pre-irradiation of the fiber greatly decreases the

fiber sensitivity to radiation. Therefore pure-silicate optical fibers are suitable for

operations in a nuclear environments like maintenance or where short cable length can be

used. The challenge is to build the corresponding radiation-hardened conversion

optoelectronics.



Electronic and Electrical Components



       Vacuum tubes. Vacuum systems were used in early electrical systems. Vacuum

tubes are now rarely used in electronics because of their bulk and power requirements.

Despite these drawbacks, vacuum tubes are still the only equipment available that can

operate under an extremely high dose rate of radiation. The hardness of the vacuum

devices originates in their simplicity; they consist of metallic parts enclosed in a glass

bottle. A mechanism with light detection or amplification is much less affected by

radiation than with solid state devices. Vacuum tubes can tolerate total dose up to 1

MGy. Many radiation-hardened cameras use vacuum tubes that avoid total dose and dose

rate effects. Some darkening may alter the optical lense but the radiation tolerance of

these cameras is determined by the tolerance of the associated electronics.



       Crystal. A crystal is a passive element used to control the frequency of an

oscillator. Radiation creates defects on the crystal lattice that results in frequency shifts

and degradation of the Q value. The frequency shifts are very small and are usually
                                              69


negligible. The frequency shift is worst with natural quartz because it contains

impurities. In that case, the frequency change can be larger than 1 ppm after only 100 Gy

of total dose. For some applications, it may be essential to the reliability of a system to

have a very stable crystal frequency. It is best in this case to use synthetic quartz crystals.

The rate of change per unit of dose can then be as low as 1 part in 1010 per Gray at 10

kGy [80].



       Resistors. The resistor manufacturer uses different technologies that produce

different resistance to radiation. Most of the resistors are very tolerant to radiation but

oxide film resistors can fail as soon as 10 Gy. It has been found that high resistance

value resistors are more sensitive than resistors with a low resistance value. The

radiation induces chemical degradation of the materials in the resistor that leads to a

decrease in the resistance. Table 20 gives several examples of resistor technologies and

their radiation hardness.




                   Table 20: Radiation damages thresholds on resistors
              Resistors                                Threshold level (Gy)
              Precision wire-wound ceramic bobbin 106 - 1010
              Metal film                               105 - 109
              Precision wire-wound epoxy bobbin        104 - 107
              Carbon film                              104 - 107
              Other film                               103 - 105
              Composition                              102 - 106
              Oxide film                               10 - 104
               Source of data: see references [76, 78]
                                             70


       Capacitors. A capacitor consists of two conducting surfaces separated by a

dielectric insulator. Many capacitor technologies are used and their radiation tolerance

differs greatly. Since the conducting surfaces are usually metallic, they are not affected

by radiation. The dielectric damage, however influences the capacitance value, the

leakage current and the breakdown voltage. All these parameters are affected after

exposure of electrolyte capacitors to radiation. Electrolyte capacitors are the most

radiation sensitive capacitors and can fail as soon as 100 Gy. Organic dielectrics are also

not very tolerant to radiation and their exposure results in leakage current and dielectric

loss. Tantalum capacitors show both radiation-induced conductivity and stored charge

variation because charges are generated when radiation is absorbed. Glass and ceramic

capacitors are the most resistant to radiation but are limited to small capacitance values.

The degradation of capacitors depends on the bias during irradiation. The bias influences

the trapping of charges created by radiation that results in long term conductivity. Table

21 shows damage threshold levels for several capacitor technologies.




                        Table 21: Radiation damages on capacitors
                         Capacitors         Threshold level (Gy)
                         Glass              105 - 108
                         Paper              105
                         Mica               104 - 107
                         Ceramic            104 - 108
                         Tantalum           103 - 105
                         Polyester          103 - 107
                         Polycarbonate      102
                         Electrolyte        102
                       Source of data: see references [76, 78]
                                             71


       Inductor. An inductor is a solenoid made of a metallic wire wound around a

former. The wire is very radiation resistant but the insulator around the wire is not and

may create short circuits. The deformation of the former under radiation may also

change the overall inductance of the inductor. The radiation resistance of an inductor,

therefore, depends on the hardness of the former and the insulator materials and lies

between 10 and 10 6 Gy.



       Cables. Cables consist of one or several metallic conductors insulated by organic

compounds. The conductors are used to transmit electrical data and power and are

radiation tolerant. The insulators prevent short-circuits and protect against the external

environments but are radiation sensitive. The number of connectors and the cable

thickness depends on the application. For mobile robots the umbilical cord is vital

because it carries the power signal and control lines. In such applications, the cable

flexibility is important and the cable integrity must be intact even after contact with

corrosive chemicals and mechanical stress. Most of the commercially available cables

(excluding Teflon) would not present significant degradation up to a total dose of 1 MGy.

At higher dose the insulator become brittle, chips, or peels and becomes much more

sensitive to mechanical stress. The most radiation tolerant cables use PEEK and

polyimide that will not fail up to 70 MGy. Polyurethane rubbers are resistant up to 50

MGy and are more flexible. Inorganic insulators like ceramics, glass or mica can be used

in high dose rate environments. Flexibility is then accomplished by coiling the wire like

a telephone cord. When a cable carries a very high frequency signal, the dielectric
                                             72


changes in the insulator are a large concern, but radiation-hardened RF cables are

commercially available.



       Thermocouples. Thermocouples are devices that measure temperature. They

consist of two metallic wires soldered at both ends. A natural difference of potential

occurs when the two solders are at different temperature. The simplicity of this device

makes it one of the most radiation resistant instruments. Thermocouples are even used in

reactor cores. The only limitation is the resistance of the wire insulation.



       Transformers. Transformers modify the AC voltage between two electrical

networks or provide isolation. Transformers consist of two coils of wires wounded on an

iron former. This element is critical since a change in the transforming ratio may damage

or disable an entire system. Radiation damages the enameling on the wires. This leads to

shorts and changes in the transforming ration. Radiation-hardened transformers using

resistant enameling is a requirement above 1 MGy. At very high dose the magnetic

properties of the transformers are also affected.



       Connectors, switches and relays. Connectors, switches and relays are as sensitive

to radiation as their polymeric components are. Plastics and polymers suffer shrinking,

cracking and other degradation of their mechanical and insulation properties that result in

shorts or loss of integrity. Internal movement in relays, external mechanical actions,

stresses on switches, and connectors also accelerate the degradations. The release of

chemical compounds may also affect the quality of electrical contacts. An appropriate
                                             73


selection of the organic compounds or the use of ceramics prevents degradation due to

total dose. Table 22 shows damage threshold levels for several connectors and switches.




Table 22: Radiation damages on connectors, switches and relays
Component                                         Dose to produce 25% damage (Gy)
Connector, polystyrene                            6. 107
Connector, polyethylene                           9. 105
Connector, duroc ceramic                          3. 106
Connector, melamine plastic                       3. 106
Relay, switch base, asbestos filled phenolformald 1. 107
Relay, switch base, unfilled phenolformald        1. 105
Source of data: see references [76, 78]

        The decontamination of robotic equipment is realized after operation in radiation

environments. Connectors must be designed to avoid the trapping of contaminants in

inaccessible parts. The effects of alpha particles are also a cause of concern for small

plastic parts.



        Circuit boards. Insulating and conducting mechanical supports to electronic

circuits is provided by circuit boards. The conducting parts are metallic and do not suffer

from damage due to radiation. The main board is sensitive to the total dose if it is made

of polymers. The resulting mechanical degradation may cause distortion, cracking or

modification of the insulating properties. The radiation tolerance of the board is a

function of polymer type but damages are unlikely to be significant at doses lower than

100 kGy. An important exception is Teflon (PTFE) used in UHF boards that is

particularly sensitive to radiation above 100 Gy. The integrity of circuit board is assured

by using a glass fiber circuit board or any other radiation resistant material.
                                             74




Mechanical and Electromechanical Components



       Ball bearings. Friction is eliminated in moving parts by the use of ball bearings.

A ball bearing consists of a cage containing metallic balls between two metallic annuli.

The cage is made of plastic or metal. Metallic cages should be the only cages used in a

radiation environment. The choice of a radiation tolerant lubricant is crucial since many

lubricants lose their viscosity after less than 10 kGy of total dose. Synthetic lubricants

are more resistant to radiation than natural lubricants, however any containing Fluor

compounds are very sensitive.



       Motors. Motors work under mechanical, thermal and electrical stress. They are

designed to operate under these stressful conditions, but radiation effects are rarely taken

into account. Motors are not exclusively made of metals. They also contain a wide

variety of organic compounds like lubricants, seals and elastomers. These parts are

radiation sensitive and the harsh environment may accelerate damages to a motor. The

radiation breaks the chemical links in molecules. This results in a change in the

mechanical properties. Temperature, vibrations and the mechanical stress in the motor

aggravate these effects further. A design with radiation tolerant compounds prevents

failure and extends the radiation tolerance of motors.



       Magnets. Magnetic components are used in electromechanical equipment as well

as non-volatile data storage devices. Hard magnetic materials are highly resistant to

radiation. Soft magnetic materials are more sensitive and radiation can actually improve
                                              75


their magnetic property [76]. Magnetic data storage devices like magnetic tapes or disks

are very tolerant to radiation. Damage to magnetic materials can be initially observed

after a total dose of 106 Gy.



       Thermal insulation. Thermal insulation is realized by low-density foam made of

polymers. Radiation damages the polymers by releasing some gas that alters the thermal

conductivity slightly. The effects on insulation are negligible, but the degradation of

mechanical properties of the foam may raise concerns.



       Mechanical sensors. A wide variety of mechanical sensors measure displacement,

pressure, acceleration, vibration, etc. Metallic sensors are very resistant, as long as the

insulators and the connectors are not damaged. This is the case for strain gauge and coil

solenoid sensors. Piezoelectric sensors are also radiation tolerant and operate without

failure up to 100 kGy. Many mechanical sensors also use some semiconductor elements,

particularly accelerometers, pressure gauges, etc. Since semiconductors are very

sensitive to radiation, an assessment of the hardness of these sensors is required.




                            Radiation Effects on Semiconductors


       Radiation effects of gamma radiation on semiconductor devices are described

here. The effects of neutrons, protons, beta particles and heavy charged particles are not

studied since these effects raise little concern in robotic systems for nuclear

environments. The bibliography for this section includes the excellent books from

Holmes-Siedle and Adams “Handbook of radiation effects” [77] and “The effects of
                                             76


radiation on electronic systems” from Messenger and Ash [81]. Two guides from

Benemann [82] and from Sharp and Garlick [1] also contributed to the writing of this

section.



Physical Effects on Semiconductors.



       Displacement damages. Displacement damages are created by the ejection of an

atom from its original location in the lattice. A momentum transfer from a particle to the

atom creates this displacement. Gamma particles do not have any mass and therefore

cannot directly displace an atom. Gamma ray interactions, however, produce secondary

electrons. If the energy of the secondary electron is greater than the displacement energy,

then the target atom is expulsed from its initial position. The kinetic energy of the

knocked atom depends on the secondary electron energy and on the initial photon energy.

With low energy photons, it is unlikely that an atom receives enough energy to be

knocked-out. The displacement effects are therefore minor for low energy particles, but

increase with energy. High-energy particles like neutrons, protons or electrons create

much more displacement damages than gamma radiation, but their effects are not studied

here. When an atom is ejected from its position, it creates a vacancy in the lattice. The

ejected atom may recombine with a vacancy or stay in an interstitial position in the

lattice. The vacancies are mobile and combine with other vacancies or with impurities of

the semiconductor. High-energy photons give rise to clusters of defects and low-energy

photons only produce single point defects. The interstitial atoms are not as electrically

active as a complex of defects.
                                             77


       Defects introduce intermediate energy levels in the gap between the conducting

band and the valence band. These band-gap defects disturb the transport of electrical

charges by several reactions [83]. First, generation and recombination of electron-hole

pairs degrade the minority carrier lifetime. Second, the trapping and compensation

effects change the majority carrier density and decrease the carrier mobility [84]. The

reduction of minority carriers lifetime affects particularly minority carrier devices like

bipolar transistors and diodes. The reduction of carrier mobility affects all

semiconductors but is often a secondary problem compared to ionization damages or

minority carrier lifetime reduction.



       Ionization damages. Ionization is the creation of electron-hole pairs along the

track of the gamma particle and the secondary electrons. Only a few eV are needed to

ionize an atom and no transfer of momentum is involved. The incident particle energy is

therefore less critical for ionization than for displacement effects. Electrons leave the site

of interaction more rapidly than holes due to their higher mobility. Any solid has a

temporary increase in its conductivity due to this carrier generation. After electron-hole

generations, electrons and holes travel in the bulk under the influence of the local electric

field. The mobility of electrons is much higher than the mobility of holes, but both

charge carriers may get into defects of the lattice called traps. Charge carriers

accumulate around traps and create a local charge build-up. These traps can be single

point defects or a mismatch of interface surfaces. This accumulation of charges is

dramatic in insulators that are used to induce an electric field in the device. In

semiconductors, the trapped charges can be excited back in the conduction band because
                                              78


of the narrow band gap. In an insulator like SiO2, the band gap is much larger and the

trap energy levels are above the conduction band. It is unlikely that an electron will

recombine with a trapped hole in an insulator in a short period. An important exception

is the electron tunneling effect that annihilates trapped holes near Si-SiO2 interface. If

charges are accumulated in the insulator, it changes the electric field and the

characteristics of the device. The most common insulator used in semiconductor devices

is SiO2. A controlled growth of SiO2 is processed during the manufacturing of a device.

The bulk of SiO2 has a low concentration of defects, the Si-SiO2 interfaces, however,

contains many defects due to mismatches in the bonds between the two planes. These

two types of defects produces two categories of charge trapping; Qot is the total charge

that results from oxide trapped charges, Q it is the total charge that results from interface

trapped charges. The sign of Q it T can be negative or positive depending on the Fermi

level at the interface. The trapped charges due to interface defects are located less than

20 nm from Si-SiO2 interfaces (see Figure 2 from source [77]). The bulk of silicon

provides electrons that annihilate the closest holes by the tunneling effect. The buildup

of charges and interface effects is particularly dramatic for MOS (Metal Oxide

Semiconductor) and CCDs (Charge Coupled Devices) but is secondary for bipolar

devices.
                                                   79


                                               Distance from Si-SiO2 interface (Log scale)


  Bulk Si-SiO2 with few deep traps                                   Bulk SiO2
                                      20 nm

                                                           Long-lived hole trapping
 Bulk Si-SiO2 with many deep traps     5 nm



                                       2 nm                                                     Trapping
                                                                           Hole annihilation and
              Near interface SiO 2                                          electron tunneling
                                      0.2 nm                Slow exchange

              Near interface SiOx              Fast exchange

                                           0
                                                                Si-SiO 2 interface
                  Near interface Si

                                      0.2 nm

                           Bulk Si                                     Bulk Si

                         Figure 2: Trapping zones at Si-SiO2 interface




        Annealing of defects. The damages caused by atomic displacement and

ionization can be recovered partially or even completely by a phenomenon called

annealing of the device. The physics of annealing is not very well understood. Many

parameters can influence the efficiency of the annealing, the temperature level is

particularly crucial. Usually a greater temperature allows a faster recovery. The risk,

however, may be to damage the device with excessive heat. Many devices show

annealing at room temperature. The biasing also plays a key role during annealing. If the

rate of damage recovery from annealing is greater than the rate of damage creation (when
                                              80


the device is turned off for example) it is possible to use annealing as a part of a

hardening method.



       Dose rate effects. When a photon penetrates into silicon, it deposits energy along

its track. The energy required to create an electron –hole pair in silicon has an average of

3.6 eV. These charge carriers contribute in conduction and react like any other electron

or hole. The contribution of these charge carriers to the total current in a device is called

photocurrent. A dose rate of 1 Gy/s in silicon is equivalent to the creation of 4. 10 9

electron-hole pairs per µm3 of silicon. This means that a photon dose rate of 1 Gy/s on 1

µm3 of silicon produces of 644 pA of current. It is then obvious that the dose rate effects

for silicon devices are negligible in the large majority of cases. There may be a concern

for devices that work with very low current in a very high dose rate (i.e., video cameras).

In that case the dose rate is so high that total dose failures are rapidly reached. These

results are also true for all semiconductor devices like germanium and gallium arsenide

devices, but may vary in magnitude.



Technology Families



       P-N Junction devices. P-N junctions are used in many applications such as

switching, rectification, voltage reference (zener diode) or in optoelectronics. P-n

junctions are naturally radiation tolerant. The effect of radiation comes from both

minority carrier lifetime reduction resulting from displacement damages in the bulk and

also from charge trapping in the oxide layer insulating the junction. The lifetime
                                             81


reduction of minority charge carriers increases the reverse leakage current, increases the

forward voltage drop and modifies the breakdown voltage. These changes are very

limited and often negligible under a total dose of 1 kGy. These limited changes in the

junction characteristic often do not cause any trouble in their application.



       Bipolar technology. A bipolar transistor consists of two p-n junctions put

together in a single device. Two combinations are possible: n-p-n or p-n-p with an oxide

layer insulating the device. The bipolar technology is known for its natural radiation

resistance often greater than 10 kGy. Two types of damages occur during irradiation.

First, particles create defects in the bulk of silicon by displacement damage. These

defects operate as recombination centers for minority carriers and shorten their lifetime.

The second origin of damages lies in the trapped charge in the oxide passivation layer.

These trapped charges introduce new interface states at the frontier SiO2-Si interface that

also decreases minority carrier lifetime and increases the junction leakage current. The

impact of oxide trapped charges on bipolar devices is much smaller than for MOSFETs

since the oxide is not an active part and because the surface doping is much greater for a

bipolar transistors than for MOSFETs. Both bulk and interface defects decrease the

overall gain and increases the leakage current. The influence of one effect over the other

depends on the photon energy, the type of silicon, temperature, bias and the transistor

geometry. Gain reduction affects linear integrated circuits while leakage current affects

digital circuits. Both displacement and ionization damages have been described earlier in

this text. It is found that the gain loss due to displacement damage strongly depends on

the base width. The thinner base region that is increasingly used in modern integrated
                                             82


circuits increases the radiation hardness of bipolar circuits. The photon energy also plays

a role since different gain values are observed after Co-60 and x-ray irradiation for the

same total dose [85]. The defects due to charge trapping in the oxide strongly depend on

component batches and on manufacturers. Since the oxide layer is used as insulator, its

geometry and thickness are not critical parameters for manufacture and a wide range of

effects can be expected. The effects of interface traps on gain loss are more obvious at

low current because the surface recombination of carriers takes out an important fraction

of minority carriers. This occurs because low currents are often surface currents. It is

possible to avoid such problems by an appropriate doping. The leakage current is also an

effect of the charge build-up of the Si-SiO2 interface. The trapped charges form a surface

channel that conducts a small current between collector and base. The leakage current

strongly depends on the impurity concentration in the oxide and on the bias during

irradiation. The geometry (vertical or lateral) and the type of transistor (p-n-p or n-p-n)

seem to play a role in the radiation hardness of a bipolar device although the dispersion of

results makes it difficult to draw any conclusive ranking. The bias is also a key

parameter, although it sometimes can improve or worsen the radiation sensitivity.

       The most important radiation behavior with bipolar technology is that not only the

total dose effects depend on the dose rate during irradiation but also that a high dose rate

sometimes gives smaller defects than a low dose rate. This simply means that a failure

dose is often underestimated by several orders of magnitude when the test is

accomplished with a higher dose rate than for the actual application. This phenomenon is

due to the time required for oxide trapped charges to slowly moves to the Si-SiO2

interface [86]. On the other hand, this effect is not found for all bipolar devices,
                                             83


sometimes a high dose rate gives more damage than a low dose rate [87, 88]. The fact

that an important fraction of the traps remain in the oxide bulk and does not shift to the

interface complicates this effect. This explanation is consistent with the strong variation

of the dose rate impact on total dose damages with the oxide thickness. It is found that

the difference of effects between high dose rate and low dose rate irradiation is greatly

reduced by using a thinner oxide layer. MOS devices do not suffer from this effect

because of the small thickness of their oxide. A testing at very low dose and at the exact

dose rate conditions as the operational environments is often impractical and too costly.

There are two alternatives. The first one is to over-test the bipolar device. This means

that if a device is certified up to 100 kGy with a high dose rate, it may operate without

failure up to 1 kGy with a low dose rate. The difficulty is to find the most accurate

overtest factor that converts the high dose rate results to low dose rate conditions. This

method may be very inaccurate when testing a complex electronic board because of the

complicated way components interact, and because the non-linear relation between the

high and low dose rates. A high dose rate can give an overestimated radiation resistance

for a MOS devices but underestimated results for a bipolar devices. The second

alternative is to irradiate the device under test at elevated temperature with a medium

dose rate [89]. It is found that an elevated temperature accelerates the shift of trapped

charges to the interface and gives closer results to a low dose rate exposure in a shorter

time. The limitation of this method is that an excessive temperature is needed when

working at high dose rate. A medium dose rate and a longer exposure time are not

avoidable. Secondly, it may be difficult to maintain a uniform temperature for a circuit

board under test. If a board under test includes MOS devices, the heat heals their
                                            84


damages under radiation and an unexpected failure may occur in real application with

low dose rate.

       A fraction of the damages can be recovered after exposure by the appropriate

annealing. A room temperature annealing however shows little recovery rate. This is

because the energy levels of the defects are important and are stable at room temperature.

An elevated temperature is required to anneal the device: between 100°C and 200°C to

suppress the interface states and between 150°C and 300°C for the trapped charges. The

highest temperature is needed to remove the bulk damages that result from displacement

of atoms: between 200°C and 450°C.



       JFET. Junction field effect transistors (JFET) have strong radiation tolerance that

is even greater than bipolar devices. Unfortunately, their use is not very common

because of their limited power and voltage. Their low noise and high input impedance

make them very useful for pre-amplification of signals. Unlike the metal oxide

semiconductor FET, the current flow is not controlled through an oxide insulator but

directly on the surface of the semiconductor. The charge trapping that is characteristic of

damages in MOSFETs is then avoided. JFETs are also majority carrier devices that are

not affected by minority carrier lifetime reduction as with bipolar transistors. Radiation

slightly affects the parameters of JFETs such as transconductance, pinch-off, and on-

resistance. The most sensitive parameter is the gate-to-source leakage that can increase

after a total dose of 10 kGy. This may become a problem for low leakage applications

after a few kGy. The noise characteristic is often the reason for using a JFET, this noise

voltage rises slowly with integrated dose integrated dose [1, 90]. JFETs using Gallium
                                             85


Arsenide materials are extremely resistant to radiation, even more than their silicon

counterpart (see page 108).



       MOS technology. Unhardened MOS (Metal oxide semiconductor) devices are

very radiation sensitive. Their tolerance is usually less than 100 Gy and their use is often

to be avoided in radiation environments. Radiation effects on MOS devices are mainly

due to ionization damages. Most of the damages occur in the SiO2 insulator between the

gate and the channel. Radiation creates electron-hole pairs in the oxide. The electric

field across the oxide separates electrons and holes. The electron moves rapidly out of

the oxide due to its higher mobility, the hole drifts slowly in the oxide and both may be

trapped in defects. The generation of electron hole pairs from irradiation therefore results

in a build up of charges in the oxide. The consequence is that a higher gate voltage is

needed to produce the same electric field across the oxide. This results in a shift ∆V in

the drain current – gate voltage characteristic of the device. Two types of traps generate

two different effects. The contribution from long-lived deep trapping of holes near the

Si-SO2 interface give rise to a total positive charge Qot that causes a simple translation in

the current-voltage characteristic. The defects at the interfaces and their resulting trapped

charge Qit create a voltage shift but also a distortion of the current-voltage curve (see

Figure 3 and [91]).
                                              86


 Log (I ds)

              N-channel




                                                                        Pre-irradiation
                                                                        20 Gy
                                                                        50 Gy
                                                                        80 Gy



                                                                                  Vg (V)
              -1                    0                    1                    2
Figure 3: Radiation effects on n-channel MOS devices. Drain current versus gate voltage
for several total doses for a n-channel MOSFET.



       For a p-channel MOS device, Qot and Q it are both positive when the gate voltage

is at the threshold voltage. These charges in the insulator weaken the effects of the

electric field in the oxide. A lower voltage is required to get the pre-irradiation result.

This modification of threshold voltage increases continuously with the accumulated dose

and makes it impossible to turn-on the device. For an n-channel device, Qit is negative

when the gate voltage is at the threshold voltage. This means that the trapped hole charge

Qot and the interface charge Qit have opposite contributions that contribute to a complex

behavior. The effect of Qot far an n-channel device is the same as in a p-channel MOS; it

results in a lowered threshold voltage. The contribution of Q it may shows up at low

radiation dose rate and/or without biasing. The decrease of oxide charges Qot due to
                                             87


annealing at room temperature allows the negatives charge Qit to actually increase the

threshold voltage. This reaction is called “turnaround” and is due to the competition of

several processes. The build-up of Qot is a faster process than the creation of Q it. The

positive charges are created faster than negative charges, resulting in a positive total

charge. The low dose rate provides more time for annealing of trapped holes, the value

of Qot becomes comparable to Qit, because charge creation and annealing takes more

time. The total charge in the oxide then goes from positive to increasingly negative. The

result of the low dose rate behavior is an increased voltage threshold. The turnaround is a

decrease followed by an increase of the threshold voltage. If this succession of threshold

voltage changes remains within acceptable limits, the radiation tolerance of components

may be extended by the turnaround effect. If a device is tested with a high dose rate and

its design is modified to adapt to the lower threshold voltage, then an exposure to a low

dose rate that results in an increase of a voltage threshold may lead to unexpected failure.

       The shift in voltage threshold for p-channel and n-channel MOS devices affects

the logic compatibility of a system: logic 1 can be taken for logic 0 and vice-versa. The

voltage threshold change is not the only effect of radiation on a MOS device. Another

effect of gamma radiation is the growth of quiescent current (supply current when the

gate is not changing state). This increased leakage current is the result of the voltage

shift in n-channel devices. With n-channel MOS, no drain current exists initially for

Vg=0 V. After irradiation, the negative voltage shift allows an increasing amount of

drain-current to pass in the channel even at Vg=0 V. The result is a greater supply

current consumption of a MOS circuit. If the current exceeds the maximum output of the

power supply, it may create a failure and/or a destruction of power components. The
                                              88


measurement of the supply current allows monitoring and prediction of the degradation

on MOS devices. The leakage current that exists in a transistor that is supposedly “off”

reduces the potential difference across its channel and reduces the difference between the

“on” and “off” states. The problem of logic compatibility is one of the main origins of

failure in MOS devices. The change in the slope of the Id-Vg characteristic results in a

reduced trans-conductance of the device. The smaller current drive increases switching

time and delays since it takes more time to accumulate charges to change the logical state

of a device [92]. This effect results in logic timing errors that limit the operational

frequency of a device. A microprocessor may fail at its nominal clock frequency, but can

operate properly at a lower frequency. Finally, the oxide located between the gate and

the channel is not the only source of problems. All modern MOS devices use thick oxide

layers as insulators to protect the active zones from interacting with the connectors or the

package. This zone of the component is the location of a parasitic parallel transistor near

the so called bird’s beak region of the lateral oxide insulator and is due to trapped holes

in the oxide layer. Radiation generates leaks in this oxide that may shunt two conducting

elements and produce a failure earlier than other effects at the gate. Displacement

damages exists in MOS devices but are always swamped by the effects of ionization.

Displacement damages decrease the minority carrier conduction but do not affect MOS

components that are majority carrier devices.

       The biasing of a device plays a major role in the radiation damage in MOS

devices. When a voltage is applied to the gate during irradiation, the electron-hole pairs

generated by the gamma rays are swept away by the intense electric field in the oxide. It

is improbable that an electron and a hole recombine. When the device is not powered, no
                                             89


force is applied on the carriers. Electron and holes travel randomly in the oxide and the

Coulomb force between the two charges favors recombination. The biasing increases the

rate of damage in the oxide; it is therefore strongly recommended not to power-up a

device under radiation when its use is not required to operate a system. In many aspects,

the radiation effects on an unbiased MOS device are comparable to the effects from

irradiation at low dose rate. In theses two conditions, the damage rate is comparable to

the rate of reparation due to annealing at room temperature. Ultimately, the rates of

damage and reparation are the same. This effect produces a saturation of the oxide-

trapped charge Qot. The contribution of the interface trapped Q it charge may then be

significant and also saturate. When the device is biased, the contribution of Qit only

shows at low dose rate. Figure 4 gives the threshold voltage shift ∆V as a function of the

total dose for biased and unbiased MOS n or p channel. The graphs show the

contribution of the voltage shift due to oxide traps ∆Vot and to interface traps ∆Vit.
                                                    90


     ∆V     N-channel MOS biased                             ∆V   N-channel MOS unbiased
                                     ∆Vit
                                                                                               ∆Vit
                                                                                              ∆VT
 0                                                       0
                                             Dose                                             ∆Vo t Dose

                              ∆VT


                                    ∆Vo t




     ∆V     P-channel MOS biased                             ∆V   P-channel MOS unbiased

 0                                                       0
                                            Dose                                           ∆Vo t    Dose

                                                                                             ∆Vit
                                                                                            ∆VT
                                    ∆Vo t

                        ∆VT
                                     ∆Vit



               Figure 4: Effect of biasing on threshold voltage of MOS devices.



          One of the strategies used to extend the lifetime of equipment under radiation is to

have a dynamic bias. It is rarely possible to leave a device unbiased at all times to reduce

radiation damage. The alternative is to power up a device only when needed. If a device

is to be used permanently, it may be possible to have a periodic bias. If the damage rate,

when the device is unbiased, is smaller than the rate of recovery due to annealing at room

temperature, then the damages that occur during the biased period can be annealed during

the unbiased period. This periodic bias consists of a square wave signal at low

frequency. If a dynamic bias cannot meet the system requirements, a reduction of the

bias voltage may significantly reduce the radiation damages [93].
                                             91


       Trapped positive charges in the oxide have an energy level that makes their

annihilation by electrons improbable in the short term. This does not mean that recovery

of defects is impossible. The accumulation of positive charges in the oxide creates a

built-in field that attracts electrons and drives positive charges away. The reaction that

heals the radiation damage is called annealing and is particularly important for MOS

devices. The annealing of trapped charges is a probabilistic reaction that varies

exponentially with time. That means that most of the damages are annealed immediately

after irradiation. The rate of damage recovery depends on many factors like temperature,

bias, trap energy level, and the type of radiation environments. Annealing is a process

that exists in permanence as soon as damages are created. At high dose rate, the trapped

charges accumulate much faster than the rate of recovery, and the contribution of

annealing is therefore negligible. The hole-trapping rate is comparable to the annealing

rate at low dose rate or when the device is unbiased, a significant fraction of the damages

is therefore annealed during exposure, even at room temperature. These general rules

about annealing are unfortunately simplified. In reality many parameters affect the

response of a MOS device. The trapped charges in the oxide Qot recover much faster

than the interface trapped charge Q it that can only be annealed at high temperature. This

leads to some interesting behavior. In an n-channel MOS device irradiated with a high

dose rate, the threshold voltage shift is due to positive trapped charges. The annealing

removes the positive charges more easily than the negative charges due to interface traps.

The total trapped charge may decrease, become zero and sometimes become negative.

This effect is called “rebound” or “super-recovery” and indicates that the positive trapped

charges are annealed but the negative interface charges remain. The opposite effect
                                             92


exists with p-channel MOS devices; in that case, Q it and Qot are both positive and the

recovery is slower than with n-channel MOS [94]. After irradiation, new interface states

are created, that trap more charges to counteracts the annealing of positive charges. In

few cases, the annealing worsens the post irradiation damage, this effect is called

“reverse-annealing”. Annealing at room temperature exists but is always slower than

annealing at high temperature. Temperature increases the recovery rate but should

remain low enough to avoid destruction of the device. Two types of thermal annealing

are possible; isothermal annealing keeps a device at a constant temperature for the entire

annealing time. Isochronal annealing uses a constant gradient of temperature; the

temperature increases at a constant rate up to a temperature limit. Isothermal annealing

gives more time for the damages to be annealed, but cannot heal the defects with the

highest energy levels. Isochronal annealing covers the entire range of trapped energy

level, but the limited duration of this operation does not allow a complete recovery of the

defects. The process of annealing is not very well understood but it seems that hydrogen

atoms play an important role. The biasing during annealing is also a key parameter. It is

found that positive bias allows a faster recovery whereas a negative bias may create the

opposite effect, that is reverse annealing. The amplitude of this biasing should be

carefully chosen to optimize the recovery rate. It is found that the recovery rate is further

increased by superimposing a high frequency signal on the continuous bias. Again the

frequency and amplitude of the superimposed signal should be chosen carefully to

accelerate the recovery [95]. A combination of thermal annealing and biasing

significantly extends the radiation tolerance of a device. The challenge is to optimize the
                                              93


numerous annealing parameters that vary with radiation environments and MOS devices

[96].

Hardened MOS devices are commercially available [ 97, 98, 99]. The difference between

hardened and unhardened components is in the processing of the device. The purity of

the material is stringently controlled to reduce the number of potential traps and special

care is given to the growth of the oxide where most of the damages are created [100].

Since the largest contribution to failure is given by the trapped charges Q it and Qot, the

manufacture of hardened MOS devices uses special steps to prevent the formation of

defects that lead to bulk and interface traps. The goal is not only to prevent the build-up

of charge but also to allow the fastest recovery after irradiation. It is found that a

reduction of the oxide thickness offers a large gain in radiation hardness. First, a thinner

oxide decreases the number of holes that can be trapped. Second, the small thickness of

oxide allows a larger fraction of trapped holes to be annealed by the tunneling effect.

The following parameters, given by Holmes-Siedle and Adams [77] are critical in the

manufacturing steps of radiation-hardened MOS devices:

•   Material preparation and cleaning

•   Gate oxide growth (thickness, temperature, ambient)

•   Gate oxide annealing (temperature, ambient)

•   Gate electrode (material, deposition conditions)

•   High temperature processing (interlayer dielectrics, passivation, packaging)

The manufacturing of commercial MOS components is directed toward profitability. It is

unfortunate that many of the parameters that allow a more affordable device are directly

opposite to the ones that improve radiation hardness. The increasing miniaturization of
                                             94


MOS devices however, indirectly benefits the radiation dose tolerance of a device. It is

often found in the literature that new semiconductor devices are more radiation sensitive

than the older generation. This is true for space and military applications where the

threats are single-event upsets and dose-rate effects. However, this is not the case for

nuclear applications where the concerns are total dose effects. In the recent years, the

size of MOS elements has considerably decreased to allow a greater integration for better

performance. This reduction in size not only reduces the size of the oxide but also

requires a more elaborate, controlled process. The manufacture of such devices requires

more processing steps with extremely pure materials, in a controlled environment. The

elaboration of highly integrated MOS devices is therefore comparable with the

fabrication of a total dose radiation-hardened components. This should result in an

improvement of radiation hardness for robotic uses in terrestrial operation. The opposite

is true for space based applications, where single event upsets (SEU) can occur.



       CMOS/SOI and CMOS/SOS technologies. Silicon on insulator (SOI) and silicon

on saphire (SOS) are two technologies that have been used to harden electronics against

dose-rate effects and single-event upsets. SOI and SOS devices reduce the bulk of silicon

available for electron-pair generation and thus reduce the amount of photocurrent at high

dose rate. SOS and SOI technologies also provide a good protection against latchup from

single event upsets that occur in space. These two technologies have been applied to

military and space applications to address specific needs of radiation hardness. Recently,

SOI chips have been developed in VLSI (Very Large Scale Integration) circuits for two

reasons. First, SOI devices need less volume than bulk MOS. Second, SOI technology
                                             95


offers a reduced parasitic capacitance that allows a greater operating frequency. The fact

that a device uses a SOI or SOS technology does not offer any guarantee of radiation

hardness regarding the total dose effects. Many SOS and SOI chips however, have been

designed for military and space application with a total dose tolerance requirement

[101,102, 103]. This hardness does not come from the use of SOI or SOS technology

but from special design and processes that make the device less radiation sensitive. In

other word, unhardened SOI/MOS and SOS/MOS devices are as radiation sensitive as

bulk MOS devices unless their manufacturing includes hardening design and processes.

The latest VLSI circuits that use SOI technology require a complex manufacturing

process that is close to the complexity of a hardened device. These VLSI circuits

consequently offer some gain in radiation hardness.



       BiCMOS. BiCMOS technology combines the high speed of a bipolar transistor to

the low current consumption and small size of the CMOS technology on a same chip.

Bipolar transistors are known to have greater radiation hardness than MOS devices.

Unfortunately the combination of the two technologies does not give a hardness greater

than the weakest point. Threshold voltage shift and edge leakage in MOS devices as well

as leakage in bipolar transistors are the source of failures at a dose as low as few hundred

Grays [104].



       GaAs technology. Gallium Arsenide (GaAs) devices are increasingly used in

high-speed circuits. The GaAs material is a relatively pure semiconductor but has a high

density of surface defects that forbids its use as a MOSFET. GaAs devices therefore use
                                             96


Schottky junction (metal semiconductor junction) instead of an oxide insulator at the

gate. GaAs components are more resistant than their silicon equivalent. They often can

operate above 100 kGy without significant problem. A GaAs field effect transistor (FET)

does not have any oxide that can trap charges. Little threshold voltage shift resulting

from charge trapping is experienced. In GaAs heterojunction bipolar transistors, the

defects are comparable to the ones resulting from irradiation of a silicon bipolar device.

Radiation creates defect centers in the emitter and base materials. A build-up of charges

in the surface passivation layer results in an induced space charge region. These two

effects increase the base current and lead to a small decrease of current gain [105]. GaAs

devices should be the first choice to harden electronics as their radiation hardness is

rarely worse than 10 kGy.



Discrete Components



       Diode. Diodes are naturally radiation resistant. The total dose effects result in an

increased leakage current, a larger forward voltage and a modified breakdown voltage.

The worst cases of total dose damage threshold voltages are 1 kGy for rectifying diodes

and 100 kGy for switching diodes but in most cases no significant characteristic change is

recorded under 1 MGy. The forward voltage variation usually stays within 5% of its

initial value. The most radiation resistant diodes are the small volume, low power ones

because the increase of forward voltage can create some power dissipation problems for

high-power diodes.
                                            97


          Voltage reference diode. Zener and avalanche diodes are diodes whose

breakdown voltage is used as voltage reference. Zener diodes have heavily doped

junction and a low breakdown voltage. Avalanche diodes are lightly doped and have

greater breakdown voltage than Zener diodes. The radiation effects are the same as for

conventional diodes. The variation of breakdown voltage is positive or negative but

generally stays within 0.1% of the initial value. The damage threshold ranges from 10

kGy to 1 MGy of accumulated dose. It is found that high power and high voltage

reference diodes are more radiation sensitive than low power and low voltage reference

diodes.



          Schottky diode. Schottky diodes are simply a metal-semiconductor junction

whose critical parameters are the junction capacitance and breakdown voltage. Both

parameters do not suffer from radiation effect under 1 MGy of total dose. Thus ,Schottky

diodes are naturally radiation resistant.



          Microwave diode. Microwave diodes comprises Impatt, Baritt, Trapatt, Gunn,

tunnel and PIN switching diodes and are naturally radiation resistant like any other types

of diodes. Most devices have a resistance to a total dose greater than 1 MGy, especially

the ones using Gallium Arsenide. The most sensitive microwave diodes are the PIN

diodes whose junction capacity and breakdown voltage can be slightly affected with less

than a few kGy of total dose.
                                               98


          Varactor diode. A varactor diode has a p-n junction capacitance that varies with

the applied voltage. The capacitance of the junction is not affected by radiation at a total

dose less than 100 kGy and the change is independent from the bias. Radiation induced

current may contribute to the destruction of the diodes at a doses as low as 100 Gy

depending on the doping profile of the varactor.



          Bipolar transistor. The case of bipolar devices has already been treated page 94.

The major effects of radiation on bipolar transistors are a gain loss and an increase of

leakage current. The degradations are greater for power bipolar transistors, especially at

low currents. The dose threshold damage varies from 100 Gy (for high power transistors)

to 10 MGy. No significant annealing is observed at room temperature. The bias affects

the damage creation rate but in an inconsistent way. The damages that result from a high

dose rate exposure of a bipolar transistor are smaller to the ones that can be found with a

low dose rate. This behavior should always be kept in mind when testing a bipolar

device.



          Junction field effect transistor. Junction field effect transistors (JFET) are very

radiation tolerant. Parameters like trans-conductance and leakage current are only

slightly affected by radiation. The overall radiation tolerance often exceeds 1 MGy. A

rapid annealing is observed in the first few hours at room temperature



          Metal semiconductor field effect transistor. Gallium Arsenide metal

semiconductor FETs (MESFET) are increasingly used for microwave applications. They
                                             99


suffer from the same type of damages as silicon FET but at a smaller scale. Pinch-off

voltage and leakage current change slightly ofter a dose accumulation greater than 1

MGy.



       Metal oxide semiconductor field effect transistor. Metal oxide semiconductor

FETs (MOSFET) are particularly radiation sensitive and their response to a radiation

stress is complex. The two major effects are a shift in the threshold voltage and a

modification of the supply current. The damage on MOS transistors strongly depends on

the temperature, bias, dose rate and type of transistor (n or p). A failure can occur as

soon as 100 Gy. An important annealing effect is observed at room temperature. This

annealing can heal or worsen the damages from irradiation. Radiation-hardened

MOSFET transistors can resist up to 10 kGy of total dose.



Optoelectronics



       Two publications from Lischka and coworkers [106, 107] described the total dose

effects of radiation on light emitting diodes, photodiodes and laser diodes.



       Photodiode. Photodiodes are diodes generating a current proportional to the

incident intensity. Radiation affects the sensitivity of the device and increases the dark

current. Detection of low intensity light is more and more difficult and ends with a

complete failure of the device. The damage increases gradually but a total failure rarely

occurs before 1 MGy. The spectral response is usually unchanged. Photodiodes are
                                            100


therefore relatively radiation tolerant. Radiation-hardened photodiodes are available but

show little hardness improvement for nuclear environments.



       Phototransistor. Phototransistors are bipolar transistors whose base current is

produced by the incident light. Radiation affects the gain and rapidly reduces the output

after a total dose ranging from 10 Gy to 10 kGy. Phototransistors are therefore less

attractive than photodiodes because of the higher radiation sensitivity.



       Light emitting diode. Light emitting diodes (LEDs) produces a visible light when

polarized. The output light intensity decreases much faster than the other parameters

characteristic of a diode. High –quality epitaxial Gallium Arsenide LEDs are more

sensitive than other diodes. Their light output decreases significantly after a few

hundreds Grays. Other diodes may be affected after 1 kGy of total dose or support a total

dose of 1 MGy without damages.



       Opto-coupler. Opto-couplers are used to isolate two circuits. They consist of a

light emitting diode optically linked to a photo-sensor like a photodiode or a

phototransistor. Opto-couplers radiation sensitivity comes from the damage occurring in

the LED and the photo-sensor. The radiation response of an opto-coupler is therefore a

combination of the effects on the different elements. It is no surprise, that opto-couplers

using phototransistors are more sensitive than the ones using photodiodes. The output

may fall between 1 kGy and 1 MGy depending on the device.
                                             101


       Laser diode. A laser diode is a laser that uses the solid state properties of silicon

to emit a light. The output power usually falls rapidly with the total dose, the radiation

effects are comparable to thoses of LEDs. The conversion of electric power into light is

directly proportionnal to the minority carrier lifetime of the semiconductor. The radiation

resistance ranges from 100 Gy to 10 kGy.



       Charge coupled device. Charge-coupled devices (CCDs) are used in cameras as

the main light sensor. CCD are based on a MOS technology and have the same limitation

on radiation tolerance. Typical failures of CCD cameras occur after 100 Gy of total dose.

CCDs are one of the few devices that are sensitive to the dose rate of radiation. A

“snow” effect degrades the quality of the image. This effect is due to the noise added by

the radiation to the electron hole pairs generated by visible light and can show up with

dose rates as low as 1 mGy/h. A dose rate of 1 kGy/h is usually the upper limit of

usability for a commercial CCD camera [108]. This effect disappears once the camera is

removed from a radioactive area. Total dose effects degrade the charge transfer

efficiency [109] and increases the dark current. In most cases the image is suddenly lost,

but it may happen that saturation of pixels, loss of lines and “smearing” effects show up

before total failure [110]. Annealing effects are observed at room temperature in some

devices and not in others, depending on the device. When testing a CCD camera, it is

often found that the CCD element is not the most sensitive element. In fact, the failure

level of the additional electronics is also very radiation sensitive [111, 112].
                                            102


Digital Integrated Circuits



       Bipolar logic. The bipolar logic family includes many types of circuits like

transistor- transistor-logic (TTL) and emitter-coupled logic. This family is generally

more resistant to total dose than their MOS equivalent. The most sensitive parameter is

the gain of the transistors. The major problems, therefore, occur at low current where

most of the gain degradation occurs. The output voltage, input current, switching time

and fan-out can be degraded by radiation. Total dose failure can be expected within the

range 100 Gy to 1 MGy. The integrated-injection logic circuits are used for their low

power requirement. Unfortunately, their very low current makes them particularly

sensitive to gain degradation and their radiation should be asseseds before integration.



       JFET, MESFET logic. JFET and MESFET circuits are naturally radiation

resistant since they do not use a gate oxide. Gallium Arsenide MESFETs are used in

very fast circuits and MESFETs as well as JFETs show a radiation resistant greater than

100 kGy.



       MOSFET logic. Today, logic components are principally using MOS technology

for low power consumption and integration capabilities. The effects of radiation on MOS

devices have been reviewed (see page 98). Different types of problems show up with an

increasing total dose, the following example gives typical failure and total dose values.

Some noise and switching reduction may show up at about 10 Gy. A threshold voltage

shift creates an increase in quiescent current after 50 Gy. A significant switching speed

reduction is noticeable after 100 Gy. Finally a logic failure due to a change of logic state
                                            103


resulting from the threshold voltage shift occurs at about 300 Gy. Depending on the

device, each one of these steps may generate a failure of the device. The resulting range

of damage threshold dose is 50 Gy to 500 Gy [113]. Many hardened MOS circuits are

now available, but only a few exceed 10 kGy of total dose tolerance.



       Microprocessor. Microprocessors are often the most critical part of a system.

The majority of these devices uses MOS technology and are therefore very radiation

sensitive. The most common causes of failure are excessive quiescent current and

extended signal propagation time. The typical range of damage threshold dose for

unhardened processors is 10 Gy to 500 Gy. Few hardened microprocessors are available,

but their availability is limited and their cost extremely high. The current dynamism of

the personal computer market has pushed the development of always smaller and faster

microprocessors. The manufacture of such small scale processor requires a very

elaborate process that is comparable in many aspects to the processing of hardened

devices.



       Memory. Random access memories (RAM) are using MOS technology and are

therefore sensitive to radiation failure. The leakage current of a single cell, multiplied by

the total number of memory cells can give rise to an important quiescent current. The

refresh time increases and the noise behavior can also lead to a failure. Total dose failure

typically occurs between 50 Gy and 5 kGy. Radiation-hardened memory chips are

available.
                                             104


       Erasable Programmable Read Only Memory. Erasable Programmable Read Only

Memory (EPROM) chips are based on a MOS technology. These EPROMs are very

sensitive to radiation because total dose effects modify the cells logic conditions and

modify the stored information. Such a failure can occur between 10 Gy and 200 Gy of

total dose. An EPROM is often the most sensitive element of digital circuits that leads to

the logic failure of a microprocessor, for example.



Analog Integrated Circuits


       Bipolar analog circuits are typically more radiation resistant than the some

devices based on MOS technology. On the other hand analog devices often work at low

current and require high gain. It is found that radiation affects bipolar transistor gain

especially at low current. The study of radiation effects on bipolar circuits is thus

unavoidable. Typical total dose damage threshold is 100 Gy to 10 kGy for bipolar linear

circuit [114] and 10 Gy to 100 Gy for MOS linear devices.



       Operational amplifiers and comparators. Operational amplifiers (Op-amp) are

commonly used in modern circuits. Op-amps can include bipolar, MOSFET or JFET

devices in their design. It is therefore hard to predict their failure, the threshold damage

value ranges from 50 Gy to several MGy. Typical impacts of radiation are reduction of

open-loop gain and input offset voltage change [115]. The input current, offset voltage

and supply current may change depending on the device. Op-amps including MOSFET

transistors, are known to be more sensitive to radiation.
                                            105


       Voltage regulator. A voltage regulator is a component used to stabilize DC

voltage. Radiation does not strongly affect the output voltage and quiescent voltage; they

tend to fall slightly. The output current capability, however, may quickly fall to zero.

Typical damage thresholds range from 1 kGy to several MGy.



       Analog-to-Digital Converter. Commercial analog-to-digital converters (ADC) are

often sensitive to radiation, their tolerance varies from 100 Gy to 10 kGy. Two

techniques of conversion are used: successive approximation or a ramping technique.

AD converters are fairly complex circuits that include both analog and digital

components. The most sensitive parameters of ADCs under radiation are the linearity

and the zero scales values [87, 88]. Depending on the design and the technology used, a

radiation resistance ranging from 100 Gy to more than 10 kGy can be found. Radiation-

hardened ADCs have been developed using SOS and SOI technology [116]. With this

technology, it is possible to reach a total dose 1 MGy without failure [117].



       Sample and hold. Sample and hold devices are used in analog to digital

conversion to sample a voltage and keep its value while the ADC estimates its digital

value. The sensitive parameters are the output voltage drop, the slew rate, the input bias

current and the charge transfer. All these parameters are affected by radiation. The most

sensitive sample and hold devices fail at about 1 kGy. More radiation tolerant devices

can be found depending on their design and technology.
                                            106


       Timer. Timers are widely used in many applications where a time reference is

needed. The oscillating frequency is not very affected by radiation. The reset current

and the low level output voltage, however, are more sensitive. Damage thresholds can be

as low as 100 Gy depending on the technology.



       Multiplexer. Multiplexers allow the connection of a line to multiple inputs by

switching from one input to another. This is generally done before an analog to digital

conversion. A multiplexer includes a digital and an analog circuit. The digital circuit is

often the most sensitive element of the device. Damage thresholds for multiplexer rank

from 50 Gy to 1 kGy depending on the technology. Radiation-hardened multiplexers are

available on the market [118].
                                   CHAPTER 4
                        RADIATION HARDENING TECHNIQUES



        This chapter describes the different methods used to harden electronic systems

from gamma radiation. No discussion on hardening techniques for materials and

mechanical devices is given since their radiation sensitivity is negligible compared to

semiconductor devices, or because the hardening methods are trivial (replacement of

Teflon by another plastic for example).




                                     Definition of Failure


        The word failure is a generic term that has many different meanings. For the

same device, a parameter change of 10% due to radiation may be insignificant in one

application but can cause a complete breakdown in another application. A failure is easy

to define when a device ceases to operate abruptly. This type of failure often occurs with

digital circuits. Even if the overall system fails suddenly, there may be some clues in the

circuit that a failure is imminent. In most cases, the degradations due to radiation

increase gradually. In that case, a failure is defined as a limitation of the device

capability that is not acceptable for a reasonable fulfillment of the device mission.

Arbitrary limits or quality norms are then used to fix threshold values that help to define a

failure. It is often better to talk about failure as the incapability to fulfill the mission of a

system, than to talk about single device degradation. In other words, the degradations of


                                               107
                                             108


a single device are relative to the impact of these degradations on the entire system

capabilities. For example the current leakage of a diode may increase by several orders

of magnitude over the manufacturer specification and still allow an acceptable operation

as a rectifier diode. In nuclear applications, a failure level is always defined in term of

total dose. This is inaccurate; it is found that electronic systems have a broad response to

radiation. MOS devices are more resistant at low dose rate and the opposite is sometimes

true for bipolar transistor. Therefore a failure level should be specified in terms of total

dose for a given dose rate. The type of radiation is also an important factor. Exposure

with a x-ray machine, Co-60 or Cs-137 does not always give the same results.

Temperature plays a role too, but room temperature is usually the reference.




                                     Minimal Approach


       This section describes some affordable but limited techniques that can be used to

extend the lifetime of robotic and electronic system under radiation without using any

complex radiation hardening work.



“Split” Technique


       The easiest way to make a system tolerant to radiation is to avoid the use of

electronics in a high dose rate area. The active mechanical parts are exposed to radiation

but the associated electronics are protected in a less aggressive area. The two parts of the

system are connected with wires. This technique is useful when the electronics do not

need to be in the radiation zone. The length and number of wires may be a problem,
                                            109


especially when the two parts are far from each other. The level of the signal and the

noise may be affected. The majority of intervention robots are now using an umbilical

that links the robot to its control station. The on-board electronics are kept to a minimum

because all the control is achieved at the control desk, far from radiation. This technique

is not compatible with the increasing complexity of robotic systems. Robots use more

and more sensors that need on-board processing. Furthermore, on-board electronics are

required to increase the robot “smartness“ and autonomy in order to reduce the workload

of the operator. Tethers often limit the range of robots and may cause a failure when the

wires get stuck. The removal of the umbilical cable imposes the creation of a radio-

communication system and a complex on-board sensing and decision center. Modern

robotic systems have not reached this point yet.



Maintenance and Repairs


         The hardening process is costly and time consuming. The cost and benefits of a

hardened system should always be examined. In some cases the failure and replacement

of a system is more profitable than upgrading this system to a radiation-hardened version.

Industrial CCD cameras for example are available at a relatively low cost, whereas

hardened cameras are expensive. The repeated replacement of commercial cameras may

therefore be more cost effective than the cost of a hardened camera and its maintenance.

On the other hand, this technique generates more radioactive waste and decontamination

needs.

         Another way of reducing maintenance costs is to minimize the number of

components that are exposed to radiation. A smaller number of chips reduces the chance
                                            110


of failures and minimizes maintenance and repairs. The radiation sensitivity study is also

less complex. Few systems are designed so that the sensitive electronics can be removed

and replaced easily, by hand or by another robot. Sensitive elements should be regrouped

in modules of the same radiation tolerance. Such a modular design is necessary to

minimize maintenance and repair time. When the maintenance of equipment is

accomplished on a periodic basis, it is helpful to match the expected lifetime of

electronics in this equipment with the maintenance period. A self-diagnostic capability

of the system “health” is very useful as well as an on-board dosimetry system using

radfet dosimeters. A preventive replacement of electronics avoids a retrieval of the robot

in difficult conditions and maintains the operability of the robot at its maximum. The

failure history and the maintenance executed on a system should therefore be carefully

documented in order to estimate the remaining lifetime of the equipment components.



Shielding


       One hardening method consists of protecting the sensitive electronics by a shield

that reduces the flux of incoming photons. The best materials to stop photons are those

with atoms having a high charge number Z; uranium is an excellent shield, for example.

But attenuation is not the only criteria. Since shield materials have high Z numbers, their

mass number and density is also extremely high. The weight of a shield rises very

rapidly and can be a major concern for mobile robots. For robotic applications, the best

shield is therefore the material that gives the best attenuation for the smallest weight. The

most commonly used shielding materials are lead (Pb) and tungsten (W). It is very

important to perform a study that demonstrates which shield is the best for a given shield
                                                      111


geometry and photon energy. Figure 5 gives the variation of the attenuation provided by

a spherical shield of inner diameter 10 cm with the weight that is added when increasing

the shield thickness (see appendix). It is very interesting to note that lead is the best

shield for 0.5 MeV photons, lead and tungsten provide the same attenuation for 1 MeV

photons, but tungsten is the best shield for 2 MeV photons. It is also obvious from this

graph that an attenuation by a factor of ten generates a shield weight of 8 kg for 0.5 MeV

photons, 25 kg for 1 MeV photons and about 45 kg for 2 MeV photons. The use of

shielding as a hardening method therefore greatly decreases the payload of the robot and

would not be acceptable for space mission.




                                                  Shield weight (kg)
               1.E+00
                        1                                   10                              100
 Attenuation




               1.E-01



                            Shield material and
                            photon energy
                                 Pb 0.5 MeV
                                 W 0.5 MeV
                                 Pb 1 MeV
                                 W 1 MeV
                                 Pb 2 MeV
                                 W 2 MeV


               1.E-02


Figure 5: Attenuation provided by lead and tungsten shields. Spherical shield of 10 cm
inner diameter versus shield weight for 0.5, 1 and 2 MeV photons.
                                             112




                               Radiation Hardening Strategy



Modification of an Existing Design


       Robots designed to work in a nuclear environment are often a modification of a

more generic model. The electronic circuit associated with this unit is therefore already

designed but requires some modification in order to become radiation resistant. The first

step is to find the most radiation sensitive components in the original design. Two

approaches are possible at this point. The most costly and time-consuming method is to

test every single component of the circuits. The relevant characteristics under radiation

are measured and the component is judged acceptable or not for a hardened circuit. This

method requires an extended testing effort over a long period of time. Another drawback

of this strategy is that information is not available about how the components interact

with each other. Again, a major characteristic change in one component does not

necessarily result in a failure, it depends on the tolerance of the system. Although

acceptable, this restrictive technique is not necessary. A more practical approach is to

irradiate the original board (a separation in modules is possible). When a failure occurs

during irradiation, the board is tested to find out the origin of the failure and which

element is responsible. This component is then replaced and the irradiation continued

until the next failure. This process can last as long as the total dose requirement. This

method gives an accurate idea of where the sensitive elements are without over-testing

the rest of the board. A ranking of components by sensitivity is possible to achieve the

hardening priorities. This technique is inexpensive because it does not requires a long
                                             113


testing campaign and therefore saves money in personnel, equipment, and in irradiation

facility usage. A functional failure of a device does require a complex measurement

setup for detection, usually only a few maintenance instruments are needed. This

strategy has a minor drawback; the short time between a failure in the irradiation facility

and the testing of a board may be enough for some annealing at room temperature to

occur. This may result in some misdiagnoses due to the change of characteristic due to

annealing. Some cases the board under test may even re-operate properly. The only way

to avoid this problem is to create an on-line setup that tests the board in real time. This is

often impractical to achieve due to the high number of parameters to monitor

simultaneously. Once the list of sensitive components is noted, the hardening process of

these components can take place. This strategy reaches its limit when an important

fraction of the circuit must be completely redesigned to become radiation-tolerant, this is

particularly true when the total dose requirement reaches a high dose rate (greater than 10

kGy). In that case, it is often better to forget the initial design and to build a brand new

system with radiation tolerant components that achieves the same function.



Innovative Design


       When the strategy described above shows that a major fraction of an existing

board must be redesigned to become radiation tolerant, the designer does not have any

choice but to start from scratch. The electronics destined for a new robot specially

designed for a nuclear application are unique and must be designed from scratch. The

designer has then the freedom to incorporate any radiation tolerant component and

hardening technique he wants. A first design is tested to assess its radiation resistance. If
                                            114


the total dose requirement is not met, the initial design is then modified. This process

may take several design and testing cycles depending on the designer experience. The

existence of radiation testing databases helps the designer to choose the right

components, but their content is often limited and out of date. Another technique to

develop a radiation-hardened circuit is to use a simulation program like SPICE. This

software simulates every element of a circuit to determine the conditions of operation.

Manufacturers now make SPICE models of their devices available. A designer can then

modify these models to incorporate the degradation of the device characteristics. A good

understanding of a device response to radiation is needed to define an accurate model.

But SPICE is a powerful tool to simulate the response of a device to radiation in different

conditions of biasing, dose rate or temperature. An innovative design specifically

oriented to radiation resistance generally results in high radiation hardness but requires

more research effort than the simple modification of an existing product. The designer

must have a thorough understanding of both radiation hardening and electronics for the

specific application of the device (power IC, digital IC, analog IC, high frequency IC,

etc).




                             Radiation-hardened Components



Definition


        The terms “radiation-hardened” (rad-hard) components denotes the

semiconductor devices that are designed with a specific process that makes them

voluntarily resistant to radiation. The manufacturer, through detailed specifications
                                           115


certifies this radiation hardness. Radiation-hardened components are not commercial off-

the-shelf (COTS) components that may or may not be radiation resistant but whose

design did not include any radiation resistance requirement. A radiation-hardened

component does not differ in other respects from a commercial component. Many rad-

hard devices have a pin to pin equivalent components available. Their package, biasing,

current, power and other characteristics are similar to commercial devices. The only

differences between COTS and rad-hard components beside their resistance to radiation

are their testing, warranty and price. Rad-hard devices can be several orders of

magnitude more expensive than their COTS equivalent.



Presentation of Radiation-hardened Components


       Radiation-hardened components are available from a limited number of

manufacturers. Some of the major retailers are Intersil, Honeywell and UTMC.

Radiation-hardened components are generally developed for space and military

application that have more stringent requirements than nuclear facility application. Rad-

hard components for space applications are tolerant to single event upsets (logic errors

resulting from high-energy ions) and total dose from electrons and protons and total dose

from low dose rates of gamma radiation. These components are often highly specialized

and integrated, and their application in robotic operation is often limited. The gamma

total dose requirement of these space components rarely exceeds a few kGy. Military

components are resistant to the electro-magnetic-pulse (EMP) and to the consequences of

an atomic explosion. Military rad-hard devices resist short-lasting high dose rate of

neutrons and gamma rays. Their gamma total dose tolerance is rarely greater than few
                                            116


kGy. Particle accelerators used in high energy physics facilities also need some

radiation-hardened electronics that is usually destined to the particle detectors but these

devices are rarely commercially available. To be useful in a nuclear environment, a

radiation-hardened component should be resist a total dose of gamma radiation greater

than 10 kGy. After this overview it is obvious radiation-hardened devices are not

designed to match the needs of nuclear operations. In a few cases the specifications of a

radiation-hardened components may overlap with the requirements for a device operating

in a nuclear environment. However, the benefits of using radiation-hardened components

for nuclear applications are generally marginal and limited.



Advantages and Limitations of Radiation-hardened Components


       The main advantage of radiation-hardened components is their reliability and

consistency. The manufacturer certifies the radiation resistance of a rad-hard device.

There is no need for testing under radiation because the characteristics are already

available and guaranteed. The specifications are also consistent for every batch of

devices; that is rarely the case for COTS components. If a radiation-hardened component

fails prematurely, the responsibility may be transferred from the designer to the

manufacturer when the specifications were not respected. In other words, rad-hard

components bring relief to a designer. The behavior of rad-hard devices under radiation

is known by prior testing and the device characteristic parameters are guaranteed to stay

within limits.

       Unfortunately, rad-hard components suffer from numerous limitations. As

discussed earlier in this text, radiation-hardened components are not designed for nuclear
                                            117


applications. Their total dose tolerance to gamma radiation is therefore rarely greater

than 10 kGy. The manufacturing of a radiation-hardened device is a small business with

a limited number of customers and a decreasing number of manufacturers. The

development of a new rad-hard device is extremely costly and the profitability uncertain.

Consequently the choice of rad-hard components is extremely limited. Only a few

models are available in each component family. If the specifications of a rad-hard device

do not meet the requirements of a circuit, there is often no other choice but to modify this

circuit because no alternative model is available. Another fact further limits the use of

rad-hard devices. The components that are the closest match to the nuclear environment

requirements are the components developed for military applications. Most of these

devices where developed during the cold war area. This technology is now old and does

not allow a high level of performance. The end of the cold war coincided with the end of

radiation hardening research funding. It is now rare to see new radiation-hardened

devices coming on the market. Many rad-hard components are not manufactured any

more. Retailers often keep a stock of devices that they sell at a very high price.

Availability is then a major issue and can significantly increase the time and production

costs. Delivery times greater than one year are current. The fact that several components

were developed for military applications often forbids their use outside the country. A

robot hardened with rad-hard components is then limited to the American market because

its exportation is illegal. Designers of radiation-hardened systems located overseas

simply do not have access to radiation-hardened components for security reasons.

Finally the major shortcoming of rad-hard components is their price. A single rad-hard

component is often several orders of magnitude more costly than their commercial
                                            118


equivalent. This is due to several factors. First the research and development of a new

radiation hardening device or process is a time consuming operation that requires a high

level of competence with the appropriate technology resulting in a greater cost than for a

traditional device. Secondly, the processing of rad-hard devices requires a high level of

cleanness and purity as well as additional manufacturing steps that increase the overall

cost. Third the market of rad-hard components is a small fraction of the electronic

business. The number of applications and customers interested in radiation-hardened

device is limited and the production is similarly small. Therefore high research testing

and production costs as well as the small volume of production explain the high price of

rad-hard components.




                      Use of Commercial Off-The-Shelf Component


       Commercial off-the-shelf (COTS) components are components that do not

include any special design, treatment or modification that makes them more radiation

tolerant. COTS devices are available everywhere with generic specifications.

       COTS components have many advantages. First their price is affordable

compared to radiation-hardened devices. Numerous COTS devices can be tested or

destroyed without any impact on the design cost of a radiation-hardened system. Second,

the availability is rarely an issue. COTS components are produced in large volumes and

sometimes by several manufacturers for the same model. Every retailer has a large stock

of devices and delivery time can be as low as few hours. Third, the choice of technology

and models is unlimited. Unlike rad-hard devices, numerous models of COTS exist to

execute a given function. A variety of versions can fulfill a mission with different
                                                         119


package, speed, power and accuracy properties. The use of COTS also allows the

integration of the latest technologies. The exportation of radiation-hardened electronics

based on COTS does not create any security issue.

           COTS are commercial devices that are not designed to be resistant to radiation.

This only means that radiation tolerance was not an issue during the design of such

device. In fact, the radiation tolerance of COTS components can be significantly greater

than rad-hard components, although the majority of COTS components are more

sensitive to radiation. Chapter 3 of this report described the radiation effects on

semiconductor devices. It gave a description of the radiation tolerance for the major

families and technologies of semiconductors. A summary is given in Figure 6 [81].


                                                No degradation     Minor damage    Major damage



       Linear ICs (Si-MOS)

        Digital ICs (Si-MOS)

        MOSFET transistors

      Digital ICs (Si-CMOS)

      Digital ICs (SOS/SOI)

  Power transistors (bipolar)

      Linear ICs (Si-bipolar)

      Digital ICs (Si-bipolar)

         Bipolar transistors

               Signal diodes

          Reference diodes

            JFET transistors

          Linear ICs (GaAs)

          Digital ICs (GaAs)

                                 1     10          100                1000          10000         100000   1000000
                                                                 Total dose (Gy)


                            Figure 6: Radiation tolerance by families of components.
                                             120


       It is clear that the use of MOS devices is to be avoided in radiation tolerant

circuits, but the use of bipolar devices can bring an important gain in radiation hardness.

A few years ago, the quality and the reliability of components were not as good as today.

The development of a military specification was required to select the best devices.

Today the quality and consistency of COTS components has greatly improved. The

semiconductors have fewer defects because of more elaborate processes and the higher

purity of materials. Today’s COTS components are more reliable than the military

components from the past. The difference of quality between military and commercial

components is now reduced. The high concentration of defects that made COTS devices

very sensitive to radiation few years ago does not exist today. Today, the improved

quality of COTS components allows them to be used in radiation fields that would have

been destructive yesterday. An appropriate selection of COTS components can now

offers a better hardness than rad-hard components.

       COTS components are very useful for hardening a system but they should be used

carefully. First, the tolerance of COTS components is generally unknown. The

technology of a device (MOS or bipolar) allows the designer to estimate the order of

magnitude of its radiation resistance, but only tests gives an accurate idea of its exact

tolerance. Therefore it is important to build a database of radiation tolerant components,

to have a list of key components whose radiation tolerance is high and proven. Secondly,

the radiation hardness of a component is unique. Large variations in radiation sensitivity

may be found between identical models from different manufacturers and even between

different batches. This is because each manufacturer has its own process to build a

device and a small change in the fabrication has a tremendous effect on radiation
                                             121


tolerance. To avoid any bad surprises, a significant number of chips from the same batch

should be tested to assess the radiation tolerance of this batch. Each new batch should be

tested in order to guarantee the right level of tolerance. The level of confidence in the

radiation resistance of a batch is directly linked to the variance found when testing the

batch.




                           Radiation Hardening Design Technique


         When creating a new design or modifying an existing circuit, several techniques

are used to make a system tolerant to radiation. The radiation hardening process is then a

succession of design, testing and improvements to reach the radiation tolerance goal.



Selection of Components


         A radiation tolerant circuit consists of radiation tolerant elements. The success of

a radiation hardening process is linked to the judicious choice of components. With rad-

hard components or COTS elements the radiation sensitivity of a circuit is generally

weaker than its weakest part. It is important to know what are the sensitive elements to

concentrate the improvement effort on these sensitive points. It makes no sense to use

components tolerating more than 100 kGy of total dose and parts failing at 1 kGy in the

same circuit. Their use increases the cost of research and complicates the maintenance.

In fact, all the elements of a given module should last roughly the same time before

failing. This minimizes the research effort and simplifies the maintenance because there

is no need to find out which are the failed components and which are the healthy ones.
                                             122


When the system fails due to the damages of one component it means that the rest of the

equipment is going to fail too in a short period of time. Therefore all the parts need to be

replaced and no complex testing is required.

       The designer has to trade-off between many requirements: the overall

performance of the system, its cost and its radiation hardness. MOS devices are very

efficient but their radiation hardness is too weak in many cases. Rad-hard components

are limited in models and availability and their prices are very dissuasive to their use.

Bipolar and JFET devices are usually the best choice because they offer acceptable

performances, radiation tolerance and affordability.



Radiation Tolerant Design


       A circuit is tolerant to radiation when the modification of its internal parameters

do not affect the efficiency of the system. In other words, the system should tolerate big

changes and stay in stable operation regardless of the accumulated dose. One way of

achieving this stabilization, is to use feedback loops. Linear circuits such as operational

amplifiers are more stable when they include a feedback system that decreases the gain

but also reduces the radiation sensitivity. The operating point of a MOS transistor should

be kept as low as possible since high input voltage increases the degradation of MOS

devices. On the other hand, the operating point of bipolar transistor should be kept as

high as possible because the gain at low current is more affected by radiation than the

gain at high current. The leakage current of the system increases with the accumulated

dose. This is particularly dramatic for MOS devices. The supply current capability

should therefore be able to supply a larger current than in normal operation. If not, the
                                            123


power supply may reach its limit and a protection circuit may shut down the power. Each

stage of a circuit should tolerate the degradation of the previous element. The leakage

current is a particularly sensitive point since it causes most of the failures in MOS

devices. The impact of the threshold voltage shift should also be limited.



Annealing


       When a device includes MOS device, the annealing effect plays a major role. The

complexity of annealing on MOS devices has been described in chapter 3. The annealing

effect varies a lot depending on the device type, the biasing and the exposure dose rate.

In most cases however, a period of time without radiation or in a low dose rate area

contributes to the survivability of a system. A mobile robot should alternate work in high

dose rate area and in low dose area to maximize the benefits of annealing effects. When

annealing at room temperature does not yield enough benefits, a heating device can be

installed on the component to be annealed. The heating setup usually consists of a

resistor glued on the components with a thermistor to regulate the heat and sometimes a

temperature sensor [111]. The device is the heated up to 100°C or even 150°C. A study

of the long-term impact of the temperature on the system should be undertaken to

determine the maximum temperature allowed. Eventually a RF signal can be connected

to the device since RF biasing is felt to accelerate the repair of radiation damages.



Biasing


       Biasing is an important factor in radiation hardening. A biased device has a much

shorter life than an unbiased one. In many cases a device shows some recovery when it
                                            124


goes from biased to unbiased operation, even when continuing to be irradiated. This

happens in unbiased operation when the rate of recovery is greater the damage creation

rate. Unfortunately if a device is not biased it simply means that it does not work and

therefore does not fulfill its mission. Therefore, it is only possible to have an unbiased

component when the use of this component is not required. An operator should then have

the possibility to turn-off a module when its use is not necessary. Another possibility is

to install a “wake-up” system that powers a module only when its use is required by the

system and turn it off once the work is done. Additionally, if a device needs to be biased

at any time then it can be beneficial to reduce the amplitude of the bias. Many systems

can now work with low voltage without affecting efficiency and this allow some gain in

radiation tolerance for little development effort. Another technique is to have two similar

systems in parallel. When one system is running, the other one remains unplugged,

suffers little damages and often shows some recovery due to the annealing effect. After a

given period of time, the two systems switch operations and this cycle is repeated

indefinitely.
                                      CHAPTER 5
                                  RADIATION TESTING




                           Radiation Types and Testing Facilities


       There are many ways to irradiate a sample. Each method has its advantages and

shortcomings. Appendixes B through E give few examples of radiation testing

accomplished by the Author at the University of Florida. Two families of irradiation

techniques are available. The first one uses radioactive sources, whose decay produces

emission of photon with discrete energies. Theses sources requires shielding at all time

and their intensity decreases exponentially with time. The second family of irradiation

technique uses artificial sources of radiation like x-rays machine and particle accelerators.

These tools give a broader spectrum of photon energies and their intensity is adjustable.

Although artificial sources are versatile, they are rarely used for total dose testing.



Cobalt-60


       Cobalt-60 sources are certainly the most common way to irradiate samples.

Cobalt-60 is a radioactive element produced by neutron irradiation of Cobalt-59 in a

reactor. Cobalt-60 emits two photons per decay; their energies are 1.17 MeV and 1.33

MeV. Many testing norms require Cobalt-60 for irradiation. Cobalt-60 from steel

structures is the main origin of radioactivity in areas where a robot operates. Testing with




                                             125
                                            126


a Cobalt-60 source simulates accurately the real conditions of use. Sealed sources of

Cobalt-60 are available at a reasonably low cost. An irradiator facility with important

shield and access protection is required. Lead as well as concrete or steel can constitute a

good shield. The dose rate decreases with the inverse of the distance to the source

squared. The variations of the dose rate are then very important near the source. The

uniformity of the dose rate is best when the sample is located as far from the source as

possible. This is particularly important for large samples because the dose can vary

significantly from one point to another. The internal volume of an irradiator facility

allows a better uniformity when irradiating a large sample. Unfortunately, the weight and

volume of the shield often limit the internal volume of the irradiator facility. The

problem of the volume available for irradiation can be solved by putting the source at the

bottom of a pool filled with water. Water provides an inexpensive shield that allows an

important volume to be irradiated. The samples are then lowered to the pool bottom.

The photons from Cobalt-60 have a great penetrating power. Several samples can be

stacked and still receive a uniform dose. The sources of radiation are usually designed

with a cylindrical symmetry. The radiation field is then uniform around this source.

However, the walls and ceilings of the irradiation chamber can present some back

scattering that alters the uniformity of the radiation distribution. The contribution of low

energy can also be significantly increased. When the source intensity is too high for an

application, lead bricks are placed between the source and the sample. It may be useful

to place a sheet of plastic between the lead bricks and the sample to avoid the

contribution of secondary electrons to the total dose. Cesium-137 is an alternative source

to Cobalt-60. The photon energy of Cesium-137 is 0.622 MeV and requires less
                                            127


shielding than Cobalt-60. The half-life is longer, being 30 years for Cesium-137 and only

5.27 years for Cobalt-60. Less shielding and a longer half-life contribute to make

Cesium-137 an attractive source of radiation. The reason for its limited use is that the

photon energy from Cesium-137 does not reproduce the exact operational conditions of

radiation. In real environments the major contribution to the dose comes from Cobalt-60.

The use of Cesium-137 for testing will give slightly different results than with Cobalt-60.

The difference comes mostly from the influence of shielding and metallic parts that

attenuate Co-60 and Cs-137 photons differently, but also from displacement damages that

are energy dependent.



Spent Fuel


       After being used in reactors, the spent fuel elements are stored in pools where the

water protects the surroundings from the effects of radiation. The radiation characteristic

from these spent fuel elements varies a lot depending on their constitution and age. The

energy spectrum is broad and the dose rates can be very high. Irradiation in spent fuel

storage pool offers an important volume and a variety of dose rates. Some modification

of the existing equipment is needed but the overall cost is low. Although the types and

energy spectrum of the radiation are varied, fuel elements recreate roughly the

environment existing in some nuclear facilities. A comparison of radiation effects on

different technology of semiconductors with a Cobalt-60 source and with spent fuel

elements has been carried out by AEA Technology in Great Britain [2]. This study

shows a divergence lower than the measurement error after a Cobalt-60 irradiation and a
                                            128


spent fuel irradiation. There is therefore no detectable difference in the degradation

induced by Cobalt-60 source and by a spent fuel element.



X-ray Machines and Particle Accelerators


       X-ray machines produce photons with an average energy lower than 100 keV. At

this energy, the interaction rate is very dependent on the material charge number. This

can lead to an energy deposition very different from a source like Cobalt-60. Therefore

total dose tests using an X-ray machine gives inaccurate results of total dose effects.

Unpredicted failure would result in nuclear environment after a testing using an X-ray

machine. X-ray machines are sometimes used to produce photons in the 10 keV range

for space applications. The space environment and the nuclear environment are very

different and the results from space testing are not exploitable for devices destined to

nuclear applications.

       Linear accelerators produce X-rays with greater energy than X-ray machines.

They also tend to have larger testing area. The University of Florida has used its 4 MeV

Van de Graaff accelerator to test the motors of the Remotec ANDROS robot. Linear

accelerators are adjustable; they can deliver a very high dose rate as well as a much lower

dose rate.
                                              129


                                      Testing Conditions



Dose Rate


        The dose rate during testing is a critical issue. In real environments a device is

destined to run for several months or even years without maintenance intervention. If the

dose rate is low, the exposure can last a long time causing any failure. A test cannot

reproduce these environmental conditions. In most cases, it is not realistic to put a device

under test for more than one month. The design of a radiation tolerant system requires

repetitive testing of numerous devices. It is not acceptable to wait for months to find out

if a device meets the radiation tolerance requirements. Secondly, the room available in

irradiation facilities is limited. If the testing lasts a long period of time, several items are

tested simultaneously. This requires a lot of space and is often not possible. A test with a

similar dose rate as the actual conditions of use is therefore impractical. In most cases,

the tests are an accelerated test; it simulates the conditions of use in a much shorter period

of time.

        This accelerated testing implies a higher dose rate that can cause important

problems. High dose rates produce an photocurrent due to electron-hole pair generation.

This photocurrent modifies the characteristics of a semiconductor device and can lead to

misunderstandings of the total dose effects. These dose rate effects are easily detectable;

a simple comparison of the measurements before and few seconds after exposure

indicates the presence of photocurrent. These effects also disappear immediately after

the end of the exposition and are usually undetectable in a real dose rate environment. A

good example of such dose rate effects is the snow that appears on pictures acquired by a
                                             130


CCD camera exposed to radiation. High dose rates also create permanent effects that

influence dramatically the response of systems to radiation. Many bipolar devices are

less affected by radiation when exposed to a high dose rate than to a low dose rate. This

effect is explained is chapter 3 and the conclusion is that testing in high dose rate

conditions can lead to serious misleading estimations of the total dose effects. There are

two ways of avoiding this incorrect failure prediction. The first one is to use a device to

only a fraction of what the failure dose is, when this failure dose is given by a test at high

dose rate. This technique uses a preventive approach but does not yield any quantitative

results of the exact radiation effects. A two-step method proposed by R.D. Schrimpf uses

a high dose rate test at elevated temperature that yields close results to a low dose rate

test. The first step is an irradiation using a very high dose rate: greater than 1.8 kGy/h.

The samples are irradiated up to the specified total dose; this phase is destined to check

for isolation-related failures. The second step is an irradiation of a second set of samples

with the same high dose rate but at an elevated temperature. This temperature should be

in the range 100-125°C. This second step is destined to check for gain-related failures.

This procedure gives a good idea of the possibility of a dose rate effects when testing a

bipolar device. Dose rate effects also affect MOS technology. The annealing effect that

occurs at low dose rate is the main origin for the differences of total dose damages when

using a high and a low dose rate. In many cases testing using high dose rate gives a

pessimistic estimation of the device radiation tolerance. For a n-channel MOS device

however, the gate threshold voltage shift is negative at high dose rate and positive with

low dose rate. The “turnaround” effect (chapter 3) may even complicate the response to

radiation. The dose rate also affects the radiation tolerance of organic materials. When
                                              131


irradiating a sample of polymers, a high dose rate dose not give enough time for the

oxygen atoms to react with material. At low dose rate however, reaction is possible and

results in more degradations.

        The impact of dose rate on total dose testing is very important but is often

underestimated because of the difficulty of avoiding the use of accelerated testing. The

results of a total dose testing are correct for a limited range of dose rate and it is difficult

to estimate the real total dose failure threshold in a different environment. The common

strategy to avoid this problem is to over-test a device. A device that is to be replaced

after 10 kGy in a low dose environment is tested up to 100 kGy with a high dose rate. If

this device does not fail after the high dose rate testing, it may be reasonable to assume

that the radiation tolerance of the device exceeds 10 kGy in a low dose rate environment.

It is also very useful to irradiate a device with a reasonably low dose rate for an

acceptable period of time in order to check if the trend of the radiation damage follows

the results of the high doe rate testing. If different results are obtained, it may be a clue

that dose rate effect exist.



Temperature


        Temperature has an important impact on the characteristics of a semiconductor

device. The manufacturer specifications are always given for a specific temperature,

usually the room temperature around 295°K. The irradiation of a sample generates heat

that can increase the temperature of the sample. A dose rate of 3.6 kGy/h is equivalent to

a power dissipation of 1 Watt. The heat generated by radiation is negligible in the
                                             132


majority of cases. At very high dose rate, the increase of temperature can be significant

and can affect the semiconductor.

       The annealing effect is very sensitive to temperature. This is particularly

important when testing MOS devices. If a system is going to be used in an environment

with elevated or low temperature, the testing procedure should try to recreate the same

conditions. If this simulation is not achieved, the estimation of the radiation degradation

on MOS devices may be inaccurate. When a heating device is mounted on a component

to generate an important annealing effect, good ventilation should be provided. The size

of irradiation chambers is often limited and the thick walls provide an excellent thermal

shield. Therefore, when a heating device is installed, the temperature increases rapidly

and can affect devices that remain cooler in a real environment.



Biasing


       Biasing can have a tremendous impact on electronic equipment. A biased device

is degraded more rapidly by radiation than a non-biased device. Therefore the biasing

conditions of a test must be the same as during the real use of the system. The amplitude

and duty cycle of the biasing must be identical. It is incorrect to irradiate a chip without

connection to a power supply and to periodically test this device to check for failure.

This method should be avoided because it gives results that are not even close to the one

expected in a real application.
                                             133


Other Parameters


        When testing a robotic device, the electronics are not the only system to test.

Organic materials are affected by radiation too. They are under constant attack from

numerous factors: chemical, thermal and mechanical. Organic compounds can also

generates corrosive and even toxic and explosive compounds when irradiated. These

organic compounds must not be neglected and their testing should be carried out with the

appropriate care. Vibrations, humidity, chemical aggression and temperature affect the

mechanical parts of a robot and motivate extensive testing to recreate the harshest

conditions




                                          Dosimetry


        The evaluation of the dose received by a system is important to determine

accurately its radiation resistance. Two categories of detectors are available: the first

category measures the incident dose rate, and the second one evaluates the accumulated

dose.



Dose Rate Detectors


        Dose rate detectors are based on the principle of ionization chamber. When a

photon enters a chamber filled of gas, the molecules of gas are ionized along the particle

track. If an electric field is applied, the ions can be collected at the electrodes. The

measurement of the voltage across the electrode gives a pulse whenever a photon reaches

the detector. With a Geiger-Müller detector, the amplitude of the pulse is independent of
                                             134


the photon energy since the entire volume of gas is ionized. After averaging, the

measurement of the current at the output of the chamber indicates the number of photons

entering the detector. This can be useful with mono-energetic sources, but Geiger-Müller

detectors cannot measure accurately the dose rate when the energy spectrum is broad. A

proportional chamber however is able to indicate the energy of the incoming photons.

The number of ionized gas molecules is proportional to the energy of the incoming

photon. Therefore the collected charge and the pulse amplitude have a linear relation

with the photon energy. The value of the averaged current is therefore proportional to the

energy deposition rate, i.e., the dose rate. The detector is put at the location of the

irradiated sample to measure the dose rate. There are two ways to calculate the total dose

with an ionization chamber. The first way is to measure the dose rate and to multiply this

dose rate by the irradiation time to get the total dose. The second way is to integrate the

output of the detector over the irradiation time, the total dose can then be directly read.

The use of ionization chambers is easy and needs little additional equipment to refill the

gas chamber. One inconvenience of these detectors is their important volume that dose

not give a good accuracy for a precise location.



Total Dose Sensors


       Total dose sensors are based on properties of solid. Three types of detectors are

commonly used: thermo-luminescent dosimeters (TLD), film dosimeters and MOS

dosimeters. This dosimeters are placed on the irradiated sample and removed at the end

of the experiment or after a period of time. In general several sensors are used in the

same time and the total dose is the statistical average of the detected values.
                                             135




       TLD. Thermo-luminescent dosimeters (TLD) are crystalline materials like LiF.

This crystal contains defect and electron and holes traps that are filled during irradiation.

After exposure, the crystal is heated and frees electron and traps. Light is emitted when

electrons and traps recombine. This light is detected by reading equipment. The curve of

the light output versus temperature is known as the glow curve. Its integral is

proportional to the total dose. TLD are available in different forms but their size is as

small as few millimeters. The major inconvenience of TLD is the cost of the reading

equipment but TLD are reusable.



       Film dosimeter. Film dosimeters are thin plastic films containing dyes. These

dyes become colored when irradiated; the optical density varies with the accumulated

dose. A measurement of the optical density at a given light wavelength and a calibration

curve gives the total dose received by the film. Film dosimeters can be very small in size

and their use and reading is very simple. Their accuracy is better than TLD but is

affected by the temperature and the relative humidity.



       MOS dosimeter. Metal oxide semiconductor (MOS) dosimeters, also called

radfets, uses the threshold voltage shift of MOS devices that was described in chapter 3 to

measure the accumulated dose. These detectors can measure total doses in the range 10

mGy to 10 kGy. These detectors are very useful because they allow a remote monitoring

of the total dose. This can be very useful to measure the total dose accumulated on a

remote system in real time, whereas other systems need to get the detector back.
                                             136




                                     Testing Procedure



Number of Samples


       It is known that radiation effects on semiconductor components vary a lot

depending on the manufacturer and between batches. It is often difficult to get the batch

number from the manufacturers, particularly for COTS components. The results present

an important statistical distribution and a significant number of samples should be tested.

Practically, it is very costly to test a great number of samples. The test data is then given

with a lower confidence level. Fortunately the quality of COTS components has greatly

improved over the years the results usually stay within acceptable limits. However,

bipolar transistor batches contain a few components called “mavericks” whose electrical

characteristics are very far from the rest of the batch. These mavericks can cause

important problem if they are not detected in a screening.



Testing Conditions


       The principal parameters that can affect the results of a test are described above.

The idea behind a testing is to recreate the realistic environmental conditions of use. This

is particularly difficult for a mobile robot that must move in a radiation area. In that case,

the incident dose rate can vary a lot and the photon energy too. The best and worst case

conditions should be determined to indicate the most appropriate testing conditions.
                                            137


       The variations of the system parameters should be recorded on line. This is

particularly important because this technique gives an accurate description of the

degradation process. It indicates if a changing parameter is saturating or not. It also

allows the designer to make immediate decisions concerning eventual modification

without waiting for the end of the irradiation to check for the results. The technique that

needs to remove periodically a component from the irradiation area to undergo some test

to check for failure is a bad method. The short period of time between the removal from

the source and the measurement of the device parameters is enough for some annealing to

occur. On the other hand this technique is usually simpler and less costly than building a

complex setup of instruments that accomplish the on-line testing.

       Once the irradiation is completed, the device should remain under test to check

for annealing at room temperature for a short period. An important recovery at room

temperature can be very valuable information that can be exploited to extend the

survivability of a device.



Testing Equipment


       Today a variety of testing equipment is available that facilitates the testing

operations. The generalization of computer based instrument control allows a computer

to control several instruments or a data acquisition boards simultaneously. The

instruments are connected to the PC via a General Purpose Interface Bus (GPIB) that

controls up to 15 instruments in the same time [119]. Computer programs like LabVIEW

from National Instruments make the programming of the instruments and the

development of the testing procedure a relatively easy operation. Radiation testing
                                             138


normally does not require any other instrumentation than that available in most electronic

laboratories. In most cases the instrumentation remains outside the radiation area and is

connected to the device under test with cables. These cables should be radiation tolerant.



Choice of Test Parameters.


        When irradiating a device, several electrical characteristics of a devices can be

affected. There is a big difference between testing a wafer and testing an encapsulated

device. Wafer testing gives some internal parameters that are not accessible once the

chip is packaged. For example, it is not possible to determine the gain loss on the first

stage of a complex bipolar VLSI. The test designer can only measure external

parameters like the supply current.

        When testing a single chip it is theoretically possible to measure every accessible

parameter. In most cases however, many parameters are irrelevant because they do not

affect the operation or are not affected by radiation. It is the test designer’s decision to

monitor the acquire the appropriate measurements. When testing a module including

many devices, the setup that would monitor every interesting measurement would be too

complex. Only a limited number of parameters can be acquired. These parameters

should be strategically located in the circuit to locate a failure. Again it is the designer’s

responsibility to find out which data acquisition is the best.



Norms


        Many organizations have regulated the radiation testing into norms. Most of them

are oriented toward military and/or space application. Few of them however give a
                                            139


standard method that can be used for nuclear application. The US Department of

Defense has published the military standards Mil-Std-833D that is the most accepted

norm today [120]. Only two of the methods listed in this document concerns total dose

effects from gamma radiation:

   •   Method 1019.4 (Mil-Std-833D): Ionizing radiation (Total dose) Test procedure.

   •   Method 1022 (Mil-Std-833D): MOSFET Threshold voltage.

The American Society for Testing and Materials (ASMT) also published a group of

standards that are adapted to the nuclear industry:

   •   ASTM E1249-88: Practice for minimizing dosimetry errors in radiation hardness

       testing of silicon electronic devices using Cobalt-60 [121].

   •   ASTM E1250-88: Methods for application of ionization chambers to assess the

       low energy gamma component of Cobalt-60 irradiators used in radiation hardness

       testing of silicon electronic devices [122].
                                      CHAPTER 6
                                     CONCLUSION



       The nuclear business is now a frequent user of robotic systems. The principal

reason for their use is the reduction of human exposure. In some nuclear facilities, there

is no alternative but to use remote systems. Over the past decade, the implementation of

new robotic systems has demonstrated that robots are not only successful in reducing

human exposure, but also their efficiency demonstrated important cost savings.

       The use of monitoring and surveillance robots has considerably increased the

efficiency and the quality of radiation surveys. Inspection robots are now able to scan an

entire reactor vessel in a few days during the critical period of power plant maintenance.

The inspection and cleaning of steam generators formerly was a challenging operation

because of high radiation fields and accessibility problems. Today these operations are

achieved with a limited exposure of personnel and the quality of the cleaning has

considerably extended the lifetime of steam generator while maintaining the thermal

efficiency. The waste processing of fuel elements is certainly the most hazardous

operation in the nuclear fuel cycle because of the very high dose rates that are found. The

use of remote systems is imperative for this operation. These systems used to be

essentially mechanical. Today, an increasing number of electronic circuits improve the

control and the security of the waste processing operations. An important number of

decontamination and decommissioning robots were developed over the past decade.

Their versatility and efficiency allow operations with almost no human intervention on


                                           140
                                             141


site. This improvement has considerably reduced radiation exposure of workers while

limiting the decommissioning costs. The reliability of emergency intervention robots has

greatly improved. Their survivability level is now enough to execute some important

work in harsh environments likely after a nuclear accident. Almost every application of

robotic systems demonstrates an important reduction of human exposure while

decreasing the cost of the operation. These two factors explain the popularity of remote

systems.

       Several research projects are now developing more sophisticated and intelligent

robots. Major efforts are underway to improve the man-machine interface to simplify the

task of the remote operators and to provide faster and more accurate work. The nuclear

industry has also benefited from the development of robotics in non-nuclear activities.

The majority of surveillance and maintenance robots are working in areas where the

radiation dose rate is low. These robots do not require any radiation tolerance because

they are not exposed to significant accumulated doses of radiation. The improvement of

remote systems destined for high radiation environments is more limited. These robots

usually do not carry their own electronic system but are controlled by an operating center

located in a low radiation area via an umbilical cable. The on-board electronics are often

very primitive and limited. The reason for these limitations is the difficulty of

developing radiation tolerant systems. Radiation damages nearly all electronic materials.

The damages are most noticeable for organic materials and semiconductors. An

appropriate selection of organic materials can improve their radiation tolerance. This

hardening process does not create any problems because of the wide variety of organic

materials available today. It is more difficult to build a radiation tolerant electronic
                                             142


system. The majority of semiconductor components are very sensitive to radiation. The

principal damages result from the trapping of charge carriers into insulators that modify

the characteristics of a device. The metal-oxide technology (MOS) is widely used today

for its integration capability and its low power consumption. Unfortunately, MOS

devices are strongly affected by radiation. The buildup of charges in the gate oxide

creates a voltage shift in the current voltage characteristic of the device. This effect

results in an important increase in the leakage current and in other degradations. Bipolar

devices are more radiation resistant than MOS devices. They also are affected by a build-

up of charge in their oxide layer similar to MOS devices and they suffer from

displacement damages in the bulk of the material. These two effects result in a lower

current gain, especially at low current. Devices based on junction field effect transistors

and semiconductors built in Gallium Arsenide show the best radiation tolerance. The

design of a radiation tolerant system requires a good knowledge of the radiation effects of

each individual component. Many factors influence the build-up of radiation damage.

One of the most important factors is the annealing effect. The heating of a device allows

a partial or complete recovery of the degradations. The impact of the annealing is

particularly important for MOS devices but exists for any semiconductor device even at

room temperature. The biasing also plays an important role. A biased device is degraded

by radiation much more quickly than an unbiased one. A reduced biasing therefore

extends the lifetime of a device. A complicating factor is the non-linearity of the total

dose damages with the dose rate. The combination of all these complex effects makes the

prediction of a failure very difficult. The first step when designing a radiation tolerant

circuit is to select components that are known to be radiation resistant. A good
                                             143


experience in radiation tolerant circuits is then very helpful as well as the consultation of

database of radiation tolerant components. Additional methods use the annealing effect

and the biasing to improve the radiation resistance. Radiation testing is required to

estimate the tolerance of a device. It is accomplished in irradiation facilities that recreate

the environmental conditions of use. The process of hardening a device to radiation is

then a repetition of testing and improvement to reach the tolerance goal. The

development of radiation-hardened system is therefore a costly operation. On the other

hand the design of a radiation tolerant system has never been easier and this trend will

probably continue. The understanding of the degradation effect is now understood to a

greater extent than a decade ago. Many manufacturing techniques that add to radiation

hardening were developed and are now implemented that extend the survivability of

commercial components. In spite of a compromised future for the use of rad-hard

components, an increasing number of commercial components show natural resistant to

radiation. The process and the design of new highly integrated components are similar to

the one for rad-hard components. The reduction of the active size of semiconductor

devices makes them more sensitive to single-event upsets but more tolerant to total dose

effects. In fact, the impact of these single-event upsets is becoming an increasing issue in

the aerospace industry.

       The past decade has seen the introduction of new robotic systems that improved

the costs and efficiency of nuclear operations. Now these robots are routinely used in

nuclear facilities. The market for robotics for nuclear application is expanding

continuously and creates a new category of workers specializing in operating these

remote systems. The next decade is going to see an increasing use of electronic systems
                                            144


in high radiation areas. Today, systems working in high radiation dose rates use

mechanical equipment with a limited efficiency. The introduction of more controls and

sensors will improve the efficiency and reduce costs of operations in high radiation fields.

A major breakthrough will be made when the communication between the operator and

the robot will be achieved without umbilical cable. The development of a radiation-

hardened radio-communication system by the University of Florida contributes to

research on wireless remote systems and holds promise for the future.
              APPENDIX A
COMPARISON OF LEAD AND TUNGSTEN SHIELD
Table A-1: Comparison of lead and tungsten shields
 0.5 MeV
   mu t     B.F. Pb    B.F. W    Atten Pb   Atten W thick Pb       thick W    weight Pb 5cm weight Pb 10cm   weight W 5cm   weight W 10cm
     1        1.2       1.2      4.41E-01   4.41E-01   0.57          0.40           2.28          8.63           2.64            10.14
     2        1.4       1.4      1.89E-01   1.89E-01   1.15          0.80           5.08         18.25           5.69            21.09
     4        1.7       1.8      3.11E-02   3.30E-02   2.29          1.61          12.46         40.66           13.20           45.55
     7         2        2.3      1.82E-03   2.10E-03   4.01          2.81          28.77         83.02           28.43           89.14
    10        2.3       2.6      1.04E-04   1.18E-04   5.73          4.02          52.65        137.16           49.15          141.76
    15        2.7       3.1      8.26E-07   9.48E-07   8.59          6.02         113.21        257.49           98.21          251.79
    20        3.1       3.6      6.39E-09   7.42E-09  11.45          8.03         205.51        421.23          168.83          393.17
  1 MeV
   mu t     B.F. Pb    B.F. W    Atten Pb   Atten W thick Pb       thick W    weight Pb 5cm weight Pb 10cm   weight W 5cm   weight W 10cm
    1         1.4       1.4      5.15E-01   5.15E-01   1.28          0.81           5.82          20.63          5.76            21.31
    2         1.7       1.8      2.30E-01   2.44E-01   2.56          1.62          14.55          46.52          13.37           46.07
    4         2.3       2.6      4.21E-02   4.76E-02   5.12          3.24          43.18         116.43          35.19          106.96
    7          3        3.6      2.74E-03   3.28E-03   8.95          5.68         122.94         275.57          88.30          230.65
    10        3.7       4.6      1.68E-04   2.09E-04  12.79          8.11         261.17         514.13         172.07          399.36
    15        4.8       6.3      1.47E-06   1.93E-06  19.18         12.17         665.15        1131.90         398.79          799.55
    20        5.9       8.2      1.22E-08   1.69E-08  25.58         16.22        1350.62        2089.37         762.43         1376.54
  2 MeV
   mu t     B.F. Pb    B.F. W    Atten Pb   Atten W thick Pb       thick W    weight Pb 5cm weight Pb 10cm   weight W 5cm   weight W 10cm
    1         1.4        1.4     5.15E-01   5.15E-01   1.96          1.19          10.07          33.75           9.10           32.53
    2         1.8        1.9     2.44E-01   2.57E-01   3.92          2.39          27.77          80.59           22.48          72.80
    4         2.5        2.7     4.58E-02   4.95E-02   7.84          4.77          94.61         222.15           65.36         179.82
    7         3.7         4      3.37E-03   3.65E-03  13.73          8.35         305.68         586.36          182.38         418.94
    10        4.8        5.4     2.18E-04   2.45E-04  19.61         11.93         701.26        1184.34          382.41         772.16
    15        6.8         8      2.08E-06   2.45E-06  29.41         17.90        1927.99        2857.88          960.71        1674.84
    20         9        11.3     1.86E-08   2.33E-08  39.22         23.87        4096.56        5610.10         1934.46        3059.31
Comparison of shield weigt for a spherical shield of inner radius 5cm and 10 cm
in function of photon energy and for different thickness of lead and tungsten
This calculation includes the build-up factor (B.F.)




                                                                         146
                                APPENDIX B
                   TOTAL DOSE TESTING OF A GaAs AMPLIFIER

                                 University of Florida
                  Department of Nuclear and Radiological Engineering
                                 (Revised 06/25/00)



                                         Summary


       The GaAs amplifier showed acceptable performance after being irradiated to

approximately 7 Mrad. The circuit should be able to be used in a radiation environment

without radiation effects problems.




                                        Introduction


       Many electronic components suffer degraded performance in radioactive

environments due to the disruptive effects of radiation. Of primary concern in

radioactive environment is the effect of gamma radiation on electronics. The

Department of Nuclear and Radiological Engineering (NRE) at the University of Florida

is currently working on remotely controlled and robotic components to perform

operations within nuclear radioactive facilities. In support of such research, the

Electronic Communications Laboratory (ECL) at the University of Florida is providing

assistance in the design and evaluation of radiation tolerant circuits and components for




                                            147
                                             148


use in such systems. The radiation testing is provided at the 600 Ci University of Florida

Cobalt Irradiation Facility (UFCIF).

       This paper reports on the testing of a GaAs high power amplifier, which will be

used in the wireless transceiver if radiation tolerance can be verified. The power out

versus power in ratio was studied versus total radiation dose absorbed.




                                       Radiation Source



Type


       University of Florida Cobalt-60 Irradiation Facility was used to irradiate the

devices to be tested. The facility produces gamma radiation approximately 1.25 MeV in

energy. A Monte Carlo simulation of the Cobalt’s gamma quanta interaction with the air

and lead walls of the hot cell indicates, that the energy spectrum of the gamma quanta

inside the hot cell is almost pure Cobalt-60 spectrum. The low energy component of the

gamma spectrum, due to the X-ray fluorescence of the lead walls, is negligible.



Dosimetry


       The dose calibration of the Irradiation Facility, i.e., determination of absorbed

dose rate (in krad/h) vs source-to-detector distance, was performed using multiple devices

and methods. Different kind of Radiachromic films were used to determine the dose rate

as a function of the distance from the 60Co source. The film MD-55 was used in the

range 0 - 50 krad total dose, the less sensitive HD810 film was used in the range up to
                                                 149


500 krad, and the film batch 8F8 was used for very high doses – up to 10 Mrad. The

optical density of the film irradiated was read by Radiachromic Reader Model FWT 92D.

Besides the Radiachromic films, two kinds of Ion Chamber was used in dose calibration

measurements: Model 9010 Radiation Monitor with Model 9060 Converter and Model

10X5-0.6 Ion Chamber. In addition, dose rate calculations were performed by using

MShield4 code. All these measured and calculated data were used to determine the dose

map of the Irradiation Facility. The dependence of the dose rate on the distance from the
60
     Co source in a horizontal direction is shown in Figure B-1.

                    15.6+/-0.6
                                                                   28.2+/-1.0

                    21.9+/-1.9
                                                                    44+/-5


                                                                    65+/-6


                                                                    107+/-9



                                                                    230+/-34


                         Distance [in]
           8   7”    6       5   4   3   2   1


                                                                    Cobalt source




Figure B-1: Dose Map of the University of Florida Cobalt Irradiation Facility as of June,
2000 (dose rate in krad/h).
                                             150




                                 Test Setup and Procedure



Test Circuit and Electrical Testing


       This test serves to verify the characteristics of the SNA-586 power amplifier.

This circuit uses a GaAs technology that is believed to have good radiation tolerance. To

verify the tolerance of this component, a 2.4 GHz oscillator generates a sinusoidal input

signal. This signal is split to provide both a reference signal and an input for the

amplifier. A logarithmic amplifier and a subsequent voltmeter is used to measure the

power of the reference. An identical setup monitors the power out of the GaAs amplifier

under test. This ratio is monitored versus total radiation dose absorbed by the GaAs

amplifier. Additionally, a phase shifter is applied to the reference to generate quadrature

2.4 GHz references. These are used to mix in quadrature the output of the GaAs

amplifier baseband. DC voltmeters are used to measure the amplitude of the output

signal mixed to DC. This measurements provides an indication of changing phase

response in the GaAs amplifier under irradiation. The change in the power out at DC

versus the change in power out for the entire bandwidth , as determined by the

logarithmic amplifier, gives some indication as to non-linear changes in the amplifier that

may produce noticeable output power in harmonics of the 2.4 GHz reference. The test

elements are described by Figure B-2.

       The four DC voltages to be measured are connected to the analog ports of a

Motorolla 68HC11 EVBU. The 68HC11 exchanges information with a PC through a

serial port. The acquisition is controlled by a LabVIEW program and follows several
                                            151


steps. First, the user specifies the sampling period on the PC interface and then the

acquisition starts. The 68HC11 executes a program that performs an analog conversion

of each of its eight analog input and sends the coded value of the voltage to the PC

through a RS232 link. The PC receives the data from the EVBU board, then a virtual

instrument stores and displays the new voltage. This process is repeated indefinitely at

the defined period until the user stops the process. The testing board is shown in Figure

B-3.




                      Figure B-2: GaAs amplifier test configuration
                                             152


        -




                         Figure B-3: GaAs amplifier testing board




Test Environment


       The irradiation testing of circuit was carried out at the 600 Ci University of

Florida Cobalt Irradiation Facility (UFCIF) (Figure B-4). Ambient temperature

throughout the test is at room temperature. To verify its tolerance, a GaAs amplifier was

irradiated, and its functional characteristics were monitored versus total absorbed dose.

In this test, the GaAs amplifier was irradiated to device failure or to a radiation level

acceptable for device operation. The testing of the circuit involved the monitoring and

recording of data by a personal computer (PC) equipped with the instrumentation

software LabVIEW.
                                           153




               Figure B-4: University of Florida cobalt irradiation facility




Biasing


       The GaAs amplifier was powered during the test.



Goals of Radiation Testing


       The GaAs amplifier performance was studied as a function of total radiation dose

absorbed. The results will determine whether or not the amplifier is capable of

withstanding radiation environments, or whether some other demodulation scheme must

be designed.
                                           154




                                   Data Representation




                  Figure B-5: GaAs amplifier in the irradiation chamber



       The testing apparatus consists of a single circuit board (Figure B-5). The data

was acquired in real time from inside the Co-60 irradiator. No shielding was used. The

irradiation lasted about thirteen days at a dose rate of 232 Gy/h (23.2 krad/h). The

testing was stopped three times for few hours to allow for other experiments. In these

cases, the amplifier and the wires were removed from the irradiation chamber and then

put back few hours later. Each time the experiment was restarted a slight degradation of

the output signals was observed. Such small variation can be attributed to a change of
                                           155


impedance in the cables and not to an annealing effect which would usually improve the

quality of the signal.

        The four DC outputs signals provided by the testing board were sampled every

four minutes. No dramatic change was observed. Figure B-6 shows the window of the

LabVIEW interface at the end of the experiment.




              Figure B-6: Window of the testing at the end of the experiment



        Even thought four voltages were monitored during the testing, only three

parameters were tested during this irradiation: the input power (Pow in), output power

(Pow out ) measured from a logarithmic amplifier, and also the magnitude voltage obtained

from quadrature signals. A summary of the results is shown in Table B-1:
                                                               156


Table B-1: Output data versus total dose
Total Dose (Mrads) power in (dBm) power out (dBm)                                magnitude(V)
0.000                -11.0            0.61                                       0.2359
1.001                -11.0            -0.47                                      0.2368
2.000                -11.0            0.07                                       0.2412
3.000                -11.0            0.07                                       0.2368
4.000                -11.0            -0.47                                      0.2368
5.000                -11.0            -0.47                                      0.2343
6.001                -11.0            -1.56                                      0.2197
6.883                -11.0            -1.56                                      0.2197

                         A loss of about 2 dBm can be observed at the end of the experiment; this

reduction represents 37% less gain than the initial power value. The gain loss at one

Mrad is acceptable. Figure B-7 shows the variations of the input power (Pow in) and

output power (Pow out ) with the total dose:

                                                          Total dose (Mrad)
                     0             1           2           3               4           5        6   7
                2




                0




                -2
 Power (dBm)




                -4


                                                       Power in      Power out
                -6




                -8




               -10




               -12



                                       Figure B-7: Variations of power with total dose
                                                            157


                       The magnitude also shows little change with the total dose: Figure B-8. The

GaAs amplifier tested showed no indications of unacceptable performance after being

irradiated to approximately 7 Mrad. This radiation testing determined that the component

is acceptable for use in future wireless communications circuit designs which are to

operate in hot cell environments. It is expected that the GaAs amplifier circuit may be

used without foreseeable problems in radiation environments.

                0.30




                0.25




                0.20
 Magnitude(V)




                0.15



                                                             magnitude(V)
                0.10




                0.05




                0.00
                       0          1          2          3              4        5         6          7

                                                        Total dose (Mrad)


                                  Figure B-8: Variations of magnitude with total dose



                                                   Reviewed and Approved

                                                      James S. Tulenko

                                                   Chairman and Professor

                                                  Co – Principal Investigator
                               APPENDIX C
                 TOTAL DOSE TESTING OF A BANDPASS FILTER

                                 University of Florida
                  Department of Nuclear and Radiological Engineering
                                 (Revised 06/25/00)



                                         Summary


       The Bandpass filter showed no indications of failure effects after being irradiated

to approximately 45 Mrad. The circuit should be able to be used in a radiation

environment without radiation effects problems.




                                        Introduction


       Many electronic components suffer degraded performance in radioactive

environments due to the disruptive effects of radiation. Of primary concern in

radioactive environment is the effect of gamma radiation on electronics. The

Department of Nuclear and Radiological Engineering (NRE) at the University of Florida

is currently working on remotely controlled and robotic components to perform

operations within nuclear radioactive facilities. In support of such research, the

Electronic Communications Laboratory (ECL) at the University of Florida is providing

assistance in the design and evaluation of radiation tolerant circuits and components for




                                            158
                                             159


use in such systems. The radiation testing is provided at the 600 Ci University of Florida

Cobalt Irradiation Facility (UFCIF).

       This paper reports on the testing of a bandpass filter, which will be used in a

demodulation circuit design if radiation hardness can be verified. The frequency

response was studied versus total radiation dose absorbed.




                                       Radiation Source



Type


       University of Florida Cobalt-60 Irradiation Facility was used to irradiate the

devices to be tested. The facility produces gamma radiation approximately 1.25 MeV in

energy. A Monte Carlo simulation of the Cobalt’s gamma quanta interaction with the air

and lead walls of the hot cell indicates, that the energy spectrum of the gamma quanta

inside the hot cell is almost pure Cobalt-60 spectrum. The low energy component of the

gamma spectrum, due to the X-ray fluorescence of the lead walls, is negligible.



Dosimetry


       The dose calibration of the Irradiation Facility, i.e., determination of absorbed

dose rate (in krad/h) vs source-to-detector distance, was performed using multiple devices

and methods. Different kind of Radiachromic films were used to determine the dose rate

as a function of the distance from the 60Co source. The film MD-55 was used in the

range 0 - 50 krad total dose, the less sensitive HD810 film was used in the range up to
                                                 160


500 krad, and the film batch 8F8 was used for very high doses – up to 10 Mrad. The

optical density of the film irradiated was read by Radiachromic Reader Model FWT 92D.

Besides the Radiachromic films, two kinds of Ion Chamber was used in dose calibration

measurements: Model 9010 Radiation Monitor with Model 9060 Converter and Model

10X5-0.6 Ion Chamber. In addition, dose rate calculations were performed by using

MShield4 code. All these measured and calculated data were used to determine the dose

map of the Irradiation Facility. The dependence of the dose rate on the distance from the
60
     Co source in a horizontal direction is shown in Figure C-1.

                    15.6+/-0.6
                                                                   28.2+/-1.0

                    21.9+/-1.9
                                                                    44+/-5


                                                                    65+/-6


                                                                    107+/-9



                                                                    230+/-34


                         Distance [in]
           8   7”    6       5   4   3   2   1


                                                                    Cobalt source




Figure C-1: Dose Map of the University of Florida Cobalt Irradiation Facility as of June,
2000 (dose rate in krad/h).
                                           161




                                Test Setup and Procedure



Test Circuit and Electrical Testing


       The bandpass filter of interest is manufactured by Mini-Circuits and has part

number BP-21.4 (Figure C-2). Determination of this component tolerance is needed for

the development of wireless communications components and architectures. The

preliminary circuit design utilizes a bandpass filter, which has not been proven under

radiation testing condition to function properly subject to gamma radiation. For this test

sinusoidal signal ranging from to 5 MHz to 15 MHz generated by an external waveform

generator was input into a frequency doubler (outside the radiation chamber) (Figure C-

3). The output is connected to the bandpass filter located inside the radiation chamber.

The frequency and amplitude of the output were monitored using an digital oscilloscope.

       The frequency response of the filter has been plotted every 15 minutes during the

irradiation time. A function generator and an oscilloscope, both controlled by LabVIEW

have been used to generate the frequency response. Every 15 minutes, the function

generator produced a sinus waveform. This signal entered the filter via the frequency

doubler and the RMS voltage of the output was measured by the oscilloscope. This

operation was performed 50 times at each step of frequency between 5 and 15 MHz.

Since a frequency doubler was connected between the function generator and the filter,

this corresponds to a range of 10-30 MHz. The test elements are described by Figure C-

4.
                                           162


       At each frequency step the auto-set function of the oscilloscope was used to get

the best possible signal. There are drawbacks to this technique. First, when the

amplitude of the original signal is weak, the oscilloscope cannot trigger the sine-wave.

Secondly, sometimes the harmonics generated by the frequency doubler are bigger than

the expected signal. At this point the oscilloscope triggers on the harmonics and not on

the right sinewave.




                 Figure C-2: Bandpass filter prepared for a radiation test
                     163




     Figure C-3: Bandpass filter circuit




Figure C-4: Bandpass filter test configuration
                                              164



Test Environment


        The irradiation testing of circuit was carried out at the 600 Ci University of

Florida Cobalt Irradiation Facility (UFCIF) (Figure C-5). Ambient temperature

throughout the test is at room temperature. To verify its tolerance, a bandpass filter was

irradiated, and its functional characteristics were monitored versus total absorbed dose.

In this test, the bandpass filter was irradiated to device failure or to a radiation level

acceptable for device operation. The testing of the circuit involved the monitoring and

recording of data by a personal computer (PC) equipped with the instrumentation

software LabVIEW.




                Figure C-5: University of Florida cobalt irradiation facility
                                            165



Biasing


       The bandpass filter is a passive circuit that does not need any biasing. A sine

wave was connected all the time to the input however.



Goals of Radiation Testing


       The bandpass filter was studied versus total radiation dose absorbed. The results

will determine whether or not the filter is capable of withstanding radiation

environments, or whether some other demodulation scheme must be designed.




                                   Data Representation




                Figure C-6: The bandpass filter in the irradiation chamber
                                           166


       The testing apparatus consists of only one circuit board (Figure C-6). Since the

hardness of the filter was expected to be good, the dose-rate of the testing was been

chosen to be high in order to save time for other experiments. The filter was placed very

close to the source, with a corresponding large inaccuracy in the dose-rate estimation.

The dosimetry indicates a dose rate between 2.1 and 0.8 Mrad/h. In a conservation step,

the dose rate of 0.8 Mrad/h was chosen. A summary of the results is shown in Table C-1.




Table C-1: Results of the bandpass filter irradiation
Gain      time dose
(dB)      (hr) (Mrad)
frequency
                          10.0 12.5 14.0 15.0           15.7     18.7    24.5    28.7     30.0
(MHz)
          5.20 4.16       -13.77 -14.73 -31.12 -46.27   -32.08   -2.82   -3.17   -24.57   -38.37
          10.20 8.16      -13.75 -14.64 -31.15 -53.47   -32.57   -2.78   -3.18   -24.24   -37.71
          15.20 12.16 -14.03 -14.67 -32.00 -48.20       -32.93   -2.83   -3.18   -24.62   -39.52
          20.21 16.16 -13.77 -14.49 -31.33 -47.64       -32.25   -2.80   -3.19   -24.48   -39.21
          25.21 20.17 -14.02 -14.68 -32.76 -48.62       -32.74   -2.83   -3.19   -24.94   -38.11
          30.21 24.17 -14.05 -14.67 -31.02 -49.71       -32.39   -2.79   -3.17   -24.43   -38.77
          35.21 28.17 -13.79 -14.73 -31.61 -48.49       -32.68   -2.83   -3.18   -24.56   -39.30
          40.22 32.17 -13.79 -14.77 -31.70 -47.84       -33.07   -2.85   -3.16   -24.45   -40.11
          45.22 36.17 -14.05 -14.67 -31.68 -49.47       -32.81   -2.83   -3.17   -24.53   -40.03
          50.22 40.18 -13.80 -14.66 -31.18 -51.27       -32.67   -2.83   -3.17   -24.55   -41.10
          56.72 45.38 -13.78 -14.68 -31.83 -50.55       -32.64   -2.83   -3.15   -24.50   -40.16
                                           167




                  Figure C-7: Results of the bandpass filter irradiation



       The bandpass filter tested showed no indications of failure effects after being

irradiated to approximately 45 Mrad. These results were expected because of the good

hardness performances of passive components. This radiation testing determined that the

component is acceptable for use in future wireless communications circuit designs for hot

cell environments. The bandpass filter may be used without foreseeable problems in

radiation environments.



                                    Reviewed and Approved

                                        James S. Tulenko

                                    Chairman and Professor

                                   Co – Principal Investigator
                              APPENDIX D
            TOTAL DOSE TESTING OF AN OPERATIONAL AMPLIFIER

                                 University of Florida
                  Department of Nuclear and Radiological Engineering
                                 (Revised 06/25/00)



                                         Summary


       The operational amplifier showed no indications of failure effects after being

irradiated to approximately 6 Mrad. The circuit should be able to be used in a radiation

environment without radiation effects problems.




                                        Introduction


       Many electronic components suffer degraded performance in radioactive

environments due to the disruptive effects of radiation. Of primary concern in

radioactive environment is the effect of gamma radiation on electronics. The

Department of Nuclear and Radiological Engineering (NRE) at the University of Florida

is currently working on remotely controlled and robotic components to perform

operations within nuclear radioactive facilities. In support of such research, the

Electronic Communications Laboratory (ECL) at the University of Florida is providing

assistance in the design and evaluation of radiation tolerant circuits and components for




                                            168
                                             169


use in such systems. The radiation testing is provided at the 600 Ci University of Florida

Cobalt Irradiation Facility (UFCIF).

       This paper reports on the testing of an operational amplifier, manufactured by

National semiconductor and with part number LM6172. The frequency response and DC

offset were studied versus total radiation dose absorbed.




                                       Radiation Source



Type


       University of Florida Cobalt-60 Irradiation Facility was used to irradiate the

devices to be tested. The facility produces gamma radiation approximately 1.25 MeV in

energy. A Monte Carlo simulation of the Cobalt’s gamma quanta interaction with the air

and lead walls of the hot cell indicates, that the energy spectrum of the gamma quanta

inside the hot cell is almost pure Cobalt-60 spectrum. The low energy component of the

gamma spectrum, due to the X-ray fluorescence of the lead walls, is negligible.



Dosimetry


       The dose calibration of the Irradiation Facility, i.e., determination of absorbed

dose rate (in krad/h) vs source-to-detector distance, was performed using multiple devices

and methods. Different kind of Radiachromic films were used to determine the dose rate

as a function of the distance from the 60Co source. The film MD-55 was used in the

range 0 - 50 krad total dose, the less sensitive HD810 film was used in the range up to
                                                      170


500 krad, and the film batch 8F8 was used for very high doses – up to 10 Mrad. The

optical density of the film irradiated was read by Radiachromic Reader Model FWT 92D.

Besides the Radiachromic films, two kinds of Ion Chamber was used in dose calibration

measurements: Model 9010 Radiation Monitor with Model 9060 Converter and Model

10X5-0.6 Ion Chamber. In addition, dose rate calculations were performed by using

MShield4 code. All these measured and calculated data were used to determine the dose

map of the Irradiation Facility. The dependence of the dose rate on the distance from the
60
     Co source in a horizontal direction is shown in Figure D-1.

                         15.6+/-0.6
                                                              28.2+/-1.0

                         21.9+/-1.9
                                                               44+/-5


                                                               65+/-6


                                                               107+/-9



                                                               230+/-34


                              Distance [in]
                8   7”    6       5   4   3   2   1


                                                               Cobalt source




Figure D-1: Dose Map of the University of Florida Cobalt Irradiation Facility as of June,
2000 (dose rate in krad/h).
                                           171




                                 Test Setup and Procedure



Test Circuit and Electrical Testing


       The testing apparatus consists of one circuit board (Figure D-2) containing the

amplifier which was placed inside the radiation chamber. A sine wave is connected to

the input of the operational amplifier by a waveform generator. The frequency of the

waveform steps from 1 KHz to 15 MHz. This input sinusoid is set 2 V peak to peak. An

oscilloscope measures Vrms and DC offset of the amplifier output. The determined

frequency characteristics supply the indicator of proper device operation and is monitored

an a personal computer with a LabVIEW program on a regular basis. The configuration

of this test is shown in Figure D-3.




                      Figure D-2: Operational amplifier testing board
                                            172




                   Figure D-3: Operational amplifier test configuration




Test Environment


       The irradiation testing of circuit was carried out at the 600 Ci University of

Florida Cobalt Irradiation Facility (UFCIF) (Figure D-4). Ambient temperature

throughout the test is at room temperature. To verify its tolerance, an operational

amplifier was irradiated, and its functional characteristics were monitored versus total

absorbed dose. In this test, the operational amplifier was irradiated to device failure or to

a radiation level acceptable for device operation. The testing of the circuit involved the

monitoring and recording of data by a personal computer (PC) equipped with the

instrumentation software LabVIEW.
                                           173




               Figure D-4: University of Florida cobalt irradiation facility




Biasing


       The operational amplifier was powered during all test.



Goals of Radiation Testing


       The operational amplifier performance was studied versus total radiation dose

absorbed. The results will determine whether or not the operational amplifier is capable

of withstanding radiation environments, or whether some other demodulation scheme

must be designed.
                                           174




                                   Data Representation




              Figure D-5: Operational amplifier in the irradiation chamber



       The testing apparatus consists of only one circuit board (Figure D-5). The data

was acquired in real time from the inside of the Co-60 irradiator. The irradiation was

performed at a dose rate of 145 Gy/h (14.5 krad/h) and to a total dose of 6 Mrad. The

two outputs signals provided by the testing board were sampled every 15 minutes.

       No change has been observed on the RMS amplitude spectrum after an integrated

dose of 6 Mrad. Figure D-6 shows the evolution of the spectrum with the total dose.
                                          175

            -




           Figure D-6: Evolution of the frequency response with the total dose



       The DC offset, however slightly decreases with the irradiation. There is about 1

mV of difference between the spectrum at the beginning of the irradiation and 6 Mrad see

Table D-1 and Figure D-7.




Table D-1: Evolution of the DC offset with the total dose.
Frequency (Hz) 1.00E+3 4.81E+3 2.31E+4 1.11E+5               5.34E+5   2.56E+6   1.23E+7
Vdc @ 0 Mrad 6.10E-3 5.83E-3 6.62E-3 6.46E-3                 6.35E-3   5.84E-3   8.36E-3
Vdc @ 3 Mrad 4.89E-3 4.71E-3 4.64E-3 4.93E-3                 5.32E-3   5.69E-3   6.51E-3
Vdc @ 6 Mrad 5.11E-3 5.13E-3 4.19E-3 5.43E-3                 5.05E-3   5.44E-3   7.30E-3
                                           176




                Figure D-7: Evolution of the DC offset with the total dose



       The operational amplifier tested showed no indications of failure effects after

being irradiated to approximately 6 Mrad. This radiation testing determines that the

component is acceptable for use in future wireless communications circuit designs for hot

cell environments. The operational amplifier circuit may be used without foreseeable

problems in radiation environments.

                                      Reviewed and Approved

                                         James S. Tulenko

                                      Chairman and Professor

                                   Co – Principal Investigator
                              APPENDIX E
            TOTAL DOSE TESTING OF A GALLIUM ARSENIDE MIXER

                                 University of Florida
                  Department of Nuclear and Radiological Engineering
                                 (Revised 06/25/00)



                                         Summary


       The mixer showed no indications of significant failure effects after being

irradiated to approximately 2 Mrad. The circuit should be able to be used in a radiation

environment without radiation effects problems.




                                        Introduction


       Many electronic components suffer degraded performance in radioactive

environments due to the disruptive effects of radiation. Of primary concern in

radioactive environment is the effect of gamma radiation on electronics. The

Department of Nuclear and Radiological Engineering (NRE) at the University of Florida

is currently working on remotely controlled and robotic components to perform

operations within nuclear radioactive facilities. In support of such research, the

Electronic Communications Laboratory (ECL) at the University of Florida is providing

assistance in the design and evaluation of radiation tolerant circuits and components for




                                            177
                                             178


use in such systems. The radiation testing is provided at the 600 Ci University of Florida

Cobalt Irradiation Facility (UFCIF).

       This paper reports on the testing of a GaAs mixer MMIC 2.5 GHz Direct

Quadrature Modulator, manufactured by RF Microdevices and with part number RF2422.

A 100 MHz reference signal was mixed single band to 2.4 GHz. This signal was mixed

down in quadrature to study magnitude and phase changes of the recovered waveform

versus total radiation dose absorbed by the modulator.




                                       Radiation Source



Type


       University of Florida Cobalt-60 Irradiation Facility was used to irradiate the

devices to be tested. The facility produces gamma radiation approximately 1.25 MeV in

energy. A Monte Carlo simulation of the Cobalt’s gamma quanta interaction with the air

and lead walls of the hot cell indicates, that the energy spectrum of the gamma quanta

inside the hot cell is almost pure Cobalt-60 spectrum. The low energy component of the

gamma spectrum, due to the X-ray fluorescence of the lead walls, is negligible.



Dosimetry


       The dose calibration of the Irradiation Facility, i.e., determination of absorbed

dose rate (in krad/h) vs source-to-detector distance, was performed using multiple devices

and methods. Different kind of Radiachromic films were used to determine the dose rate
                                                       179


as a function of the distance from the 60Co source. The film MD-55 was used in the

range 0 - 50 krad total dose, the less sensitive HD810 film was used in the range up to

500 krad, and the film batch 8F8 was used for very high doses – up to 10 Mrad. The

optical density of the film irradiated was read by Radiachromic Reader Model FWT 92D.

Besides the Radiachromic films, two kinds of Ion Chamber was used in dose calibration

measurements: Model 9010 Radiation Monitor with Model 9060 Converter and Model

10X5-0.6 Ion Chamber. In addition, dose rate calculations were performed by using

MShield4 code. All these measured and calculated data were used to determine the dose

map of the Irradiation Facility. The dependence of the dose rate on the distance from the
60
     Co source in a horizontal direction is shown in Figure E-1.

                          15.6+/-0.6
                                                              28.2+/-1.0

                          21.9+/-1.9
                                                               44+/-5


                                                               65+/-6


                                                               107+/-9



                                                               230+/-34


                               Distance [in]
                 8   7”    6       5   4   3   2   1


                                                               Cobalt source




Figure E-1: Dose Map of the University of Florida Cobalt Irradiation Facility as of June,
2000 (dose rate in krad/h).
                                            180




                                Test Setup and Procedure



Test Circuit and Electrical Testing


       The testing apparatus consists of one circuit board (Figure E-2) containing the

mixer which was placed inside the radiation chamber. A 100 MHz oscillator generates a

sinusoidal input signal, which will be in quadrature. This quadrature signal pair is fed

into the in-phase and quadrature port of the modulator. An LO of 2.4 Ghz is used to

single-band mix the signals to radio-frequency. The output of the modulator is split and

mixed down by unradiated mixers. Two down conversion mixers are used, with their

coherent 2.4 GHz Los in quadrature, to recover the quadrature components of the

modulating 100 MHz signal. Figure E-3 shows the configuration of the test. Changes in

magnitude and phase of the recovered signals are monitored versus total dose by an

oscilloscope. This oscilloscope is controlled by a personal computer (PC) via a GPIB

bus. A LabVIEW program control the acquisition and store the data (Figure E-4).

              -




                          Figure E-2: GaAs mixer testing board
                      181




   Figure E-3: GaAs mixer test configuration




Figure E-4: Testing setup in the irradiation room
                                             182



Test Environment


       The irradiation testing of circuit was carried out at the 600 Ci University of

Florida Cobalt Irradiation Facility (UFCIF) (Figure E-5). Ambient temperature

throughout the test is at room temperature. To verify its tolerance, a GaAs mixer was

irradiated, and its functional characteristics were monitored versus total absorbed dose.

In this test, the GaAs mixer was irradiated to device failure or to a radiation level

acceptable for device operation. The testing of the circuit involved the monitoring and

recording of data by a personal computer (PC) equipped with the instrumentation

software LabVIEW.




                Figure E-5: University of Florida cobalt irradiation facility
                                           183



Biasing


       The GaAs mixer was powered during the test.



Goals of Radiation Testing


       The GaAs mixer was monitored as a function of total radiation dose absorbed.

The results will determine whether or not the mixer is capable of withstanding radiation

environments, or whether some other demodulation scheme must be designed.




                                  Data Representation




                       Figure E-6: The GaAs mixer testing board
                                           184


       The testing apparatus consists of one testing board (Figure E-6). The data was

acquired in real time from the inside of the Co-60 irradiator. The irradiation was

performed at a dose rate of 160 Gy/h (16 krad/h) and to a total dose of 2 Mrad. The two

outputs signals provided by the testing board were sampled every 20 minutes.

       No change has been observed on the amplitude of the two channels after 2 Mrad.

Figure E-7 shows the evolution of the amplitude with the total dose.




                 Figure E-7: Evolution of the amplitude versus total dose



       The phase between the two signals did not change either. The variation of phase

were always smaller than 1 degree and did not vary continuously. Figure E-8 shows the

evolution of the phase difference between channel 1 and 2 with the total dose.
                                            185




             Figure E-8: Evolution of the phase difference versus total dose



       The GaAs mixer tested showed no indications of failure effects after being

irradiated to approximately 2 Mrad. This radiation testing determines that the component

is applicable for use in future wireless communications circuit designs for hot cell

environments. It is anticipated that the GaAs mixer circuit may be used without

foreseeable problems in radiation environments.



                                     Reviewed and Approved

                                        James S. Tulenko

                                     Chairman and Professor

                                    Co – Principal Investigator
                                LIST OF REFERENCES


[1] R.E. Sharp, and D.R. Garlick, “Radiation effects on electronic equipment, a
    designers’/users’ guide for the nuclear power industry,” AEA-D&W-0691, 1994.

[2] J.C. Boudenot, C. Barillot, C. Brisset, M.C. Calvet, T. Fillon, and M. Tavlet, “Les
    environments radiatifs et leurs effets a l’aube du XXIème siècle,” Journée technique
    RADECS 99. Abbaye de Fontevraud, France, September 13th 1999. pp. I-74 – I-78,
    1999.

[3] R. Sharp and M. Decreton, “Méthodes de durcissement appliquées au nucléaire
    civil,” Journée technique RADECS 99. Abbaye de Fontevraud, France, September
    13th 1999.

[4] W.T. Alumkal, “177 tanks, 177 problems - dealing with Hanford's primary cleanup
    challenge,” Nucl. Eng. Int. vol. 40, no. 488, pp. 34-36, 1995.

[5] Code of Federal Regulations, Title 10, Energy-Part 20, Standards for Protection
    Against Radiation, Section 20.1201, Occupational dose limit for adults, 1993.

[6] F.A. Marian and H.T. Rowan, “The Cost-Benefit of Robotic Devices in Nuclear
    Power Plants,” Public Service Electric and Gas Company, 1987

[7] “Robots deliver real benefits for PSE&G,” Nucl. Eng. Int. vol. 35, no. 435, pp.120-
    121, 1990.

[8] P.J. Weber, and C.W. Vanecek, “SIMON combines radiation hardness with
    computer power,” Nucl. Eng. Int. vol. 35, no. 430, pp. 34-38, 1990.

[9] "MACS makes maps of contamination," Nucl. Eng. Int. vol. 41, no. 503, pp. 18-20,
    1996.

[10] J.S. Byrd, R.O. Pettus, “ A robotic inspector for low-level radioactive waste,”
     Proceedings of the 1996 2nd Specialty Conference on Robotics for Challenging
     Environments. Albuquerque, New Mexico, June 1-6 1996. pp. 276-282, 1996

[11] F. Brack Hazen, “Robotics for waste storage inspection: a user’s perspective,”
     Proceedings, 42nd Conference on Robotics and Remote Systems. New Orleans,
     Louisiana, June 1994. vol. 1, pp. 147-154, 1994.



                                          186
                                          187


[12] R.O. Pettus and P. McCarty, “ARIES: The implementation of an autonomous
     mobile robotic system for the inspection and characterization of stored low-level
     nuclear waste,” Proceedings ANS 8th International Topical Meeting on Robotics and
     Remote Systems. Pittsburgh, PA, April 25-29 1999.

[13] A.T. Ashton and J. Gay, “SID seeks sludge and foreign bodies in steam generator
     while...,” Nucl. Eng. Int. vol. 38, no. 464, pp. 20-24, 1993.

[14] J. Malaugh and D. Monaghan, “CECIL lances Bruce’s boilers,” Nucl. Eng. Int. vol.
     39, no. 476, pp. 23-25, 1994.

[15] M. Reinhardt, “CECIL grows up,” Nucl. Eng. Int. vol. 36, no. 449, pp. 44-45, 1991.

[16] B. Murchie and M. Reinhardt, “Plucking a loose weld rod from the depths of a
     boiler,” Nucl. Eng. Int. vol. 40, no. 487, pp. 26-27, 1995.

[17] F. Puzzuoli, S. Allen, B. Murchie and L. Spencer, “In-bundle mechanical and
     chemical cleaning of BRUCE A steam generators,” Proceedings, 42nd Conference
     on Robotics and Remote Systems. Washington, D.C., November 1994. vol. 2, pp.
     22-28, 1994.

[18] H.T. Roman, “Robots in PSE&G’s nuclear plants – Experience and future
     projections,” Proceedings, 40th Conference on Robotics and Remote Systems.
     Chicago, Illinois, November 1992. vol. 2, pp. 124-130, 1992.

[19] M. Reinhardt, K. Barbito and K. Baughman,” Going a bundle on steam generator
     inspection and cleaning UBIB & UBHC,” Nucl. Eng. Int. vol. 42, no. 518, pp. 38-
     41, 1997.

[20] R.K. Lewis, “ Proactively monitor and maintain steam generators with remote visual
     inspection and cleaning systems,” Proceedings ANS 8th International Topical
     Meeting on Robotics and Remote Systems. Pittsburgh, PA, April 25-29 1999.

[21] I.R. Brookes and K.E. Schnuer, “Occupational radiation exposure in European light
     water power reactors 1981-1991,” Radiation protection no. 56, Commission of the
     European Communities. 1994.

[22] “No jumpers needed.” Nucl. Eng. Int. vol.36, no.439, pp. 30, 1991.

[23] “Installing nozzle dams 100% remotely,” Nucl. Eng. Int. vol.38, no. 464, pp. 17-19,
     1993.

[24] A.T. Ashton. “Robotic installation of steam generator nozzle dams,” Proceedings,
     42nd Conference on Robotics and Remote Systems. New Orleans, Louisiana, June
     1994. vol. 1, pp. 79-86, 1994.
                                           188


[25] N. Bains and B. Majarais, “Robotic cleaning of steam generator tubes at the primary
     side,” Proceedings ANS 8th International Topical Meeting on Robotics and Remote
     Systems. Pittsburgh, PA, April 25-29 1999.

[26] “ROSA III: the Westinghouse workhorse,” Nucl. Eng. Int. vol.36, no.449, pp. 42-
     43, 1991.

[27] “Franco-US co-operate on Cobra/Aramis,” Nucl. Eng. Int. vol.37, no.452, pp. 39,
     1992.

[28] “Inspecting the inaccessible at Yankee,” Nucl. Eng. Int. vol.37, no.452, pp. 27-31,
     1992.

[29] K. Tokunaga, T. Aoki and T. Sasaki, “Development of unique ISI equipment for
     RPV,” Proceedings, 40th Conference on Robotics and Remote Systems. Boston,
     Massachusetts, June 1992. vol. 1, pp. 28-35, 1992.

[30] J.B. Fallon, S.B. Shooter, C.F. Reinholtz and S.W. Glass, “Design of an underwater
     robot for nuclear reactor vessel inspection,” ASCE Specialty Conference,
     Proceedings 1994. Albuquerque, New Mexico, February 26 – March 3 1994. pp.
     311-319, 1994

[31] S.W Glass, M.L. Levesque, G.J. Engels, F.C Klahn and D.B. Fairbrother, “Under-
     water robotic tools for nuclear vessel and pipe examination,” Proceedings ANS 8th
     International Topical Meeting on Robotics and Remote Systems. Pittsburgh, PA,
     April 25-29 1999.

[32] R.M. Wiesman, “GRI’s internal inspection system for piping network,”
     Proceedings, 40th Conference on Robotics and Remote Systems. Chicago, Illinois,
     November 1992. vol. 2, pp. 109-115, 1992.

[33] S.I. Schreiner, “Internal pipe inspection devices for use in radiation survey
     application,” Proceedings, 42nd Conference on Robotics and Remote Systems.
     Washington, D.C., November 1994. vol. 2, pp. 16-21, 1994.

[34] A. Gebel and P. Fontcuberta, “Equipment for the recovery of U.D.O.s from storage
     and reactor pools,” Proceedings, 42nd Conference on Robotics and Remote Systems.
     New Orleans, Louisiana, June 1994. vol. 1, pp. 87-94, 1994.

[35] W.D. Willis, M.O. Anderson and M.D. McKay, “The remote underwater
     characterization system,” Proceedings ANS 8th International Topical Meeting on
     Robotics and Remote Systems. Pittsburgh, PA, April 25-29 1999.

[36] J. Rizzo, “Small underwater/land robotic vehicles and related technology,”
     Proceedings, 39th Conference on Robotics and Remote Systems. San Francisco,
     California, November 1991. pp. 10-15, 1991.
                                            189



[37] D. Rosen, ”Reducing critical path IVVI time through the use of submersibles,”
     Proceedings ANS 8th International Topical Meeting on Robotics and Remote
     Systems. Pittsburgh, PA, April 25-29 1999.

[38] M. Kimura, H. Okano and H. Shimada, “The 15 cm inspection robot,” Nucl. Eng.
     Int. vol.40, no.495, pp. 52-53, 1995.

[39] A. Fryatt, “Radiation-tolerant robots improve nuclear maintenance,” Nuclear News.
     pp. 54, 1994.

[40] E. Abel and C.J.H. Watson, “A NEATER advance in active handling technology,”
     Nucl. Eng. Int. vol.35, no.430, pp. 26, 1990.

[41] P. Desbats, C. Andriot, P. Gicquel, J.P. Viallesoubranne and C. Souche, “Using
     industrial robots in nuclear fuel reprocessing facilities force feedback teleoperation
     control,” Proceedings ANS 8th International Topical Meeting on Robotics and
     Remote Systems. Pittsburgh, PA, April 25-29 1999.

[42] P. Desbats, R. Cammoun, F. Littmann, T. Jouan de Kervanoel, J.M. Idasiak and C.
     Andriot, “Development of a new dexterous arm for telerobotic maintenance of
     nuclear facilities,” Proceedings ANS 8th International Topical Meeting on Robotics
     and Remote Systems. Pittsburgh, PA, April 25-29 1999.

[43] W. Gerriets, “TROD cleans up at Nine Mile Point 1,” Nucl. Eng. Int. vol.37, no.452,
     pp. 32-34, 1992.

[44] G. Boughman and S.L. Jones, “Robotic cleaning of radwaste tank nozzles,”
     Proceedings, 40th Conference on Robotics and Remote Systems. Chicago, Illinois,
     November 1992. vol. 2, pp. 68-72, 1992.

[45] “Cleaning up a spent resin spill,” Nucl. Eng. Int. vol.43, no.529, pp. 20-22, 1998.

[46] B. Burdge, “TMI defueling project fuel debris removal system,” Proceedings, 40th
     Conference on Robotics and Remote Systems. Boston, Massachusetts, June 1992.
     vol. 1, pp. 10-14, 1992.

[47] P.A. Titzler, D.E. Legare and H.G. Barrus, “Handford site tank 241-SY-101,
     damaged equipment removal,” Proceedings, 41st Conference on Robotics and
     Remote Systems. San Francisco, California, November 1993. pp. 7-12, 1993.

[48] E.K Anderson, C.W. Robinson and F.M. Heckendorn, “Specialized video systems
     for use in waste tanks,” Proceedings, 40th Conference on Robotics and Remote
     Systems. Chicago, Illinois, November 1992. vol. 2, pp. 164-169, 1992.
                                          190


[49] T.A Nance, C.W. Robinson and A.S. Preston, “Tools for inspecting and sampling
     waste in underground radioactive storage tanks with small access riser openings,”
     Proceedings ANS 8th International Topical Meeting on Robotics and Remote
     Systems. Pittsburgh, PA, April 25-29 1999.

[50] B.L. Griebenow, C.C. Conner and C. Christensen, “Remote inspection of
     underground storage tanks,” Proceedings, 40th Conference on Robotics and Remote
     Systems. Boston, Massachusetts, June 1992. vol. 1, pp. 84-88, 1992.

[51] B.L. Griebenow, C.B. Selleck, L.A. Strope and J.B. Burke, “Imaging, inspection,
     and characterization systems for underground storage tank remediation,”
     Proceedings, 39th Conference on Robotics and Remote Systems. San Francisco,
     California, November 1991. pp. 145-150, 1991.

[52] W.R. Jaquish, E.J. Shen and J.A. Yount, “Robotics technology demonstration
     program for underground storage tank remediation,” Proceedings, 39th Conference
     on Robotics and Remote Systems. San Francisco, California, November 1991. pp.
     133-138, 1991.

[53] B.L. Burks, G.A. Armstrong, P.L. Butler and P.T. Boissiere, “Combined long-reach
     and dexterous manipulation for storage tank applications,” Proceedings, 39th
     Conference on Robotics and Remote Systems. San Francisco, California, November
     1991. pp. 139-144, 1991.

[54] P.W. Gibbons, “Overview of the tanks focus area HLW tank retrieval activities
     (remote operations),” Proceedings ANS 8th International Topical Meeting on
     Robotics and Remote Systems. Pittsburgh, PA, April 25-29 1999.

[55] J.A. Bamberger, G.F. Boris and D.G. Alberts, “Using the borehole miner
     extendible-nozzle sluicing system to dislodge and nix settled sludge and supernate
     in the old hydrofracture tanks during remediation at Oak Ridge National Laboratory,
     Oak Ridge, Tennessee,” Proceedings ANS 8th International Topical Meeting on
     Robotics and Remote Systems. Pittsburgh, PA, April 25-29 1999.

[56] M.W. Noakes, B.L. Burks, D.D. Falter and D. Vesco, “SCARAB III remote vehicle
     deployment for waste retrieval and tank inspection,” Proceedings ANS 8th
     International Topical Meeting on Robotics and Remote Systems. Pittsburgh, PA,
     April 25-29 1999.

[57] R.L. Glassell, S.M. Killough, P.D. Lloyd, L.J. Love, J.D Randolph, S.D. Van
     Hoesen, J.A. Blank, B.L. Burks, R.E. DePew, W.H. Glover, D.D. Falter and D.P.
     Vesco, “Use of the modified light duty utility arm to perform nuclear waste cleanup
     of underground waste storage tank at Oak Ridge National Laboratory,” Proceedings
     ANS 8th International Topical Meeting on Robotics and Remote Systems. Pittsburgh,
     PA, April 25-29 1999.
                                         191


[58] H. Schempf, “Houdini: in-tank mobile cleanup robot,” ASCE Specialty Conference
     on Robotics for Challenging Environments. Albuquerque, New Mexico. pp. 286-
     295, 1994

[59] D.D. Falter, B.L. Burks, W.H. Glover, F.B. Hazen, D.P. Vesco, C.L. Fitzgerald,
     R.L. Glassell, D.J. Kington, S.D. Van Hoesen and A. Slifko, “Deployment and
     performance of the Houdini remotely operated vehicle system in the gunite and
     associated tanks at Oak Ridge National Laboratory, Oak Ridge, Tennessee,”
     Proceedings ANS 8th International Topical Meeting on Robotics and Remote
     Systems. Pittsburgh, PA, April 25-29 1999.

[60] A. Slifko, M. Catalan, F.B. Hazen and D.D. Diedre, “Houdini-II lessons learned and
     application improvements for enhanced operations in UST’s,” Proceedings ANS 8th
     International Topical Meeting on Robotics and Remote Systems. Pittsburgh, PA,
     April 25-29 1999.

[61] P. Murray, “Emptying of C-tanks at Oak Ridge reservation utilizing pulse jet
     mixing,” Proceedings ANS 8th International Topical Meeting on Robotics and
     Remote Systems. Pittsburgh, PA, April 25-29 1999.

[62] M. Catalan, B. Thompson and T. Denmeade, “Remote tool delivery system for the
     WVDP HLW tanks,” Proceedings ANS 8th International Topical Meeting on
     Robotics and Remote Systems. Pittsburgh, PA, April 25-29 1999.

[63] L.C Bares, “Decommissioning? Why not use a robot?” Nucl. Eng. Int. vol.40,
     no.488, pp. 37-38, 1995.

[64] D. Orwantschke, “Dismantling Niederaichbach. A lesson in remote-controlled
     decommissioning,” Nucl. Eng. Int. vol.40, no.493, pp. 16-18, 1995.

[65] K. Kostelnik, R. Meservey and M. McKellar, “Optimal planning of
     decontamination, decommissioning, and remediation,” Proceedings ANS 8th
     International Topical Meeting on Robotics and Remote Systems. Pittsburgh, PA,
     April 25-29 1999.

[66] M.W. Noakes, “Dual arm work module development and applications,”
     Proceedings ANS 8th International Topical Meeting on Robotics and Remote
     Systems. Pittsburgh, PA, April 25-29 1999.

[67] D.J. Cox, “Requirements for modular dual-arm robot architecture for use in
     deactivation and decommissioning tasks,” Proceedings ANS 8th International
     Topical Meeting on Robotics and Remote Systems. Pittsburgh, PA, April 25-29
     1999.

[68] K. Hendrich, H. Wiese and G. Katzenmeier, “One year ahead of remote dismantling
     of the German prototype spent fuel reprocessing plant Karlsruhe,” Proceedings ANS
                                           192


     8th International Topical Meeting on Robotics and Remote Systems. Pittsburgh, PA,
     April 25-29 1999.

[69] “Remotely operated excavator,” Nucl. Eng. Int. vol.37, no.450, pp. 41-43, 1992.

[70] “Making tracks to clean-up after accidents,” Nucl. Eng. Int. vol.39, no.476, pp. 26,
     1994.

[71] S.A. Bailey, “Hanford U plant railroad tunnel deployment,” Proceedings ANS 8th
     International Topical Meeting on Robotics and Remote Systems. Pittsburgh, PA,
     April 25-29 1999.

[72] A.F. Batanov, “The experience of application of remote control systems produced
     by Moscow state technical University for removal consequence of radioactive
     damages,” Proceedings ANS 8th International Topical Meeting on Robotics and
     Remote Systems. Pittsburgh, PA, April 25-29 1999.

[73] A. Ivanov, A. Ivanov, A. Borovoi and A. Kluchnikov, “Main results of using of
     remote controlled systems inside the Chernobyl “sarcophagus” rooms and some
     unsolved problems,” Proceedings ANS 8th International Topical Meeting on
     Robotics and Remote Systems. Pittsburgh, PA, April 25-29 1999.

[74] A. Slifko, M. Catalan and T. Denmeade, “PIONEER, a robot for structural
     assessment, of the Chernobyl shelter,” Proceedings ANS 8th International Topical
     Meeting on Robotics and Remote Systems. Pittsburgh, PA, April 25-29 1999.

[75] T. Camill, J. Teza, A. Ghavimi and F. Serrichio, “A concrete-sampling system for
     the Chernobyl unit 4 reactor,” Proceedings ANS 8th International Topical Meeting
     on Robotics and Remote Systems. Pittsburgh, PA, April 25-29 1999.

[76] K.U. Vandergriff, “Designing equipment for use in gamma environments,”
     ONRL/TM-11175, 1990.

[77] A. Holmes-Siedle and L. Adams, “Handbook of radiation effects,” Oxford Science
     publication, 1993.

[78] J.L. Bostock and F.R. Sias, “Radiation hardening for terrestrial robots,” Topical
     report, 1994.

[79] S. Coenen, M. Decreton, “Feasibility of optical sensing for robotic in highly
     radioactive environments,” IEEE Transactions on Nuclear Science. vol. 40, no. 4,
     pp. 851-856, 1993.

[80] S. Dowling, M. McGovern and R.H. West, “A comparison of gamma and X-
     radiation effects on high quality quartz oscillator crystal,” First European
                                           193


     Conference on Radiation and its Effects on Devices and Systems, RADECS 91. La
     Grande-Motte, France, September 9-12 1991. pp. 277-280, 1991.

[81] G.C. Messenger and M.S. Ash, “The effects of radiation on electronic systems,”
     Van Nostrand Reinhold, New York, 2nd edition, 1991.

[82] A.E. Benemann, “A guide to total dose effects on electronic components,”
     Teleman/Entorel, TM41/Siemens-03-31-93-02, 1993.

[83] G.C. Messenger, “A summary review of displacement damage from high energy
     radiation in semiconductors and semiconductor devices,” First European
     Conference on Radiation and its Effects on Devices and Systems, RADECS 91. La
     Grande-Motte, France, September 9-12 1991. pp. 35-40, 1992.

[84] J.R. Hauser and S.E. Kerns, “Circuit related issues due to radiation in hostile
     environments,“ Journal of Electronic Materials, vol. 19, no. 7, pp. 671-688, 1990.

[85] R.N. Nowlin, D.M. Fleetwood, R.D. Schrimpf, R.L. Pease and W.E. Combs,
     “Hardness-assurance and testing issues for bipolar/BiCMOS devices,” IEEE
     Transactions on Nuclear Science, vol. 40, no. 6, pp. 1686-1693, 1993.

[86] A.H. Johnston, B.G. Rax and C.I. Lee, “Enhanced damage in linear bipolar
     integrated circuits at low dose rate,” IEEE Transactions on Nuclear Science, vol. 42,
     no. 6, pp. 1650-1659, 1995.

[87] G.U. Youk, “Dose rate effects of a bipolar A/D converter,” IEEE Radiation Effects
     Data Workshop 1996. pp. 38-43, 1996.

[88] C.J. Lee, B.G. Rax and A.H. Johnston, “Hardness assurance and testing techniques
     for high resolution (12 to 16 bits) analog-to-digital converters,” IEEE Transactions
     on Nuclear Science, vol. 42, no. 6, pp. 1681-1688, 1995.

[89] R.D. Schrimpf, “Recent advances in understanding total-dose effects in bipolar
     transistors,” IEEE Transactions on Nuclear Science, vol. 43, no. 3, pp. 787-796,
     1996.

[90] V. Radeka, S. Rescia, P.F. Manfredi, V. Speziali and F. Svelto, “JFET monolithic
     preamplifier with outstanding noise behavior and radiation hardness
     characteristics,” IEEE Transactions on Nuclear Science, vol. 40, no. 4, pp. 744-749,
     1993.

[91] B. Sigfridsson, “Application of a new MOS-transistor model to total dose radiation
     effects in CMOS-devices,” First European Conference on Radiation and its Effects
     on Devices and Systems, RADECS 91. La Grande-Motte, France, September 9-12
     1991. pp. 35-40, 1992.
                                           194


[92] O. Flament, E. Dupont-Nivet, J.L. Leray, J.F. Pere, E. Delagnes, A. J. Auberton-
     Herve, B. Giffard, G. Borel and T. Ouisse, “High total dose effects on CMOS/SOI
     technology,” First European Conference on Radiation and its Effects on Devices
     and Systems, RADECS 91. La Grande-Motte, France, September 9-12 1991. pp.
     178-182, 1992.

[93] O. Quittard, F. Joffre, C. Oudéa, L. Dusseau, J. Fesquet and J. Gasiot, “Effects of
     input bias on different commercial technological lines of CMOS inverters with
     respect to the cumulated dose.” IEEE Radiation Effects Data Workshop 1998. pp.
     137-141, 1998

[94] I. Yoshii, K. Hama, K. Maeguchi, S. Takatsuka andH. Hatano, “Total-dose
     characterization of high performance radiation-hardened 1.0 µm CMOS sea-of-gates
     technology,” IEEE Transactions on Nuclear Science, vol. 37, no. 6, pp. 2089-2096,
     1990.

[95] S. Mandal, “Concurrent high frequency ac bias annealing of MOS devices working
     in nuclear environments,” Master’s thesis, University of Florida, 1994.

[96] J.S. Tulenko, D. Ekdahl, L. Utley and H. Hamilton, “On-line annealing processes
     for hardening electronic components in mobile robots for radiation environments,”
     Proceedings, 39th Conference on Robotics and Remote Systems. San Francisco,
     California, November 1991. pp. 183-186, 1991.

[97] C. Chabrerie, O. Musseau, O. Flament, J.L. Leray, J.C. Boudenot, B. Shipman and
     H. Callewaert, “ Post-irradiation effects in a rad-hard technology,” IEEE
     Transactions on Nuclear Science, vol. 43, no. 3, pp. 826-830, 1996.

[98] D.B. Kerwin and J.M. Benedetto, “Total dose and single event effects testing of
     UTMC Commercial rad-hard gate arrays.” IEEE Radiation Effects Data Workshop
     1998. pp. 80-85, 1998

[99] J.M. Benedetto, A. Bishop, M. Martin and N. Haddad, "High total dose response of
     the UTMC gate array fabricated at Lockheed Martin Federal Systems," IEEE
     Radiation Effects Data Workshop 1998. pp. 86-90, 1998

[100] O. Flament, J.L. Leray, F. Martin, E. Orsier, J.L. Pelloie, R. Truche and R.A.B.
     Devine, “Effect of rapid thermal annealing on radiation hardening of MOS devices,”
     IEEE Transactions on Nuclear Science, vol. 42, no. 6, pp. 1667-1673, 1995.

[101] O. Coumar and R. Gaillard, "Total-dose characterization of CMOS/SOI - ZMR
     technology," First European Conference on Radiation and its Effects on Devices
     and Systems, RADECS 91. La Grande-Motte, France, September 9-12 1991. pp.
     199-204, 1992.
                                          195


[102] V. Berland, A. Touboul, E. Dupont-Nivet and J.L. Leray, "Effect of x-ray
     radiation on MOSFET's (SIMOX) LF excess noise," First European Conference on
     Radiation and its Effects on Devices and Systems, RADECS 91. La Grande-Motte,
     France, September 9-12 1991. pp. 159-163, 1992.

[103] P. Chen, H. Wu, G. Li, Y. He, P. Li and Y. Mao, "Total dose radiation effects on
     the hardened CMOS/bulk and CMOS/SOS," First European Conference on
     Radiation and its Effects on Devices and Systems, RADECS 91. La Grande-Motte,
     France, September 9-12 1991. pp. 249-253, 1992.

[104] R.L. Pease, W. Combs and S. Clark, “Long term ionization response of several
     BiCMOS VLSIC technologies,” First European Conference on Radiation and its
     Effects on Devices and Systems, RADECS 91. La Grande-Motte, France, September
     9-12 1991. pp. 114-118, 1992.

[105] L. Silverman, G. Kuehner, “Rad-hard successive detection microwave logarithmic
     amplifiers employing GaAs monolithic chips using heterojunction bipolar
     transistors,” IEEE Radiation Effects Data Workshop 1995. pp. 71-77, 1995.

[106] H. Lischka, H. Henschel, W. Lennartz and H.U. Schmidt, “Radiation sensitivity
     of light emitting diodes (LED), laser diodes (LD) and photodiodes (PD),” First
     European Conference on Radiation and its Effects on Devices and Systems,
     RADECS 91. La Grande-Motte, France, September 9-12 1991. pp. 404-408, 1992.

[107] H. Lischka, H. Henschel, O. Kohn, W. Lennartz and H.U. Schmidt, “Radiation
     sensitivity of light emitting diodes (LED), laser diodes (LD) and photodiodes (PD),”
     Proceedings of the European Conference on Radiation and its Effects on
     Components and Systems, RADECS 1995. pp. 560-563, 1995.

[108] R.E. Sharp and A.A. Dumbreck, “Radiation testing solid-state sensor-based
     CCTV cameras the easy way,” First European Conference on Radiation and its
     Effects on Devices and Systems, RADECS 91. La Grande-Motte, France, September
     9-12 1991. pp. 363-367, 1992.

[109] J. Cluzel, A. Guiga, R. Tronel, M. Vilain, M. Vie, B. Azais and J.P. Boulet,
     “Radiations effects on CCD, “First European Conference on Radiation and its
     Effects on Devices and Systems, RADECS 91. La Grande-Motte, France, September
     9-12 1991. pp. 357-361, 1992.

[110] R. Payat and G. Cerdan, “Image processing in hostile environment. Experimental
     CCD cameras tests, results for robotic operations, “First European Conference on
     Radiation and its Effects on Devices and Systems, RADECS 91. La Grande-Motte,
     France, September 9-12 1991. pp. 348-352, 1992.
                                           196


[111] M. Kais and M. Robiolle, “Radiation hardening of an industrial CCD camera,”
     Proceedings ANS 8th International Topical Meeting on Robotics and Remote
     Systems. Pittsburgh, PA, April 25-29 1999.

[112] B.A. Pellegrino, “Radiation testing of Ca-Zoom PTZ-IV,” Proceedings ANS 8th
     International Topical Meeting on Robotics and Remote Systems. Pittsburgh, PA,
     April 25-29 1999.

[113] F. Joffre, J. Buisson, D. Mijuin, J.P. Brunet, M. Marceau and M. Perrault,
     “Equipment to study the gamma total dose effects on components of
     microcomputing systems,” First European Conference on Radiation and its Effects
     on Devices and Systems, RADECS 91. La Grande-Motte, France, September 9-12
     1991. pp. 569-573, 1992.

[114] C.I. Lee, A.H. Johnston and B.G. Rax, “Total ionizing dose effects on voltage to
     frequency converters,” IEEE Radiation Effects Data Workshop 1998. pp. 117-120,
     1998.

[115] J.M. Bosc, G. Sarrabayrouse and F.X. Guerre, “Total dose effects on elementary
     transistors of a comparator in bipolar technology,” Proceedings of the European
     Conference on Radiation and its Effects on Components and Systems, RADECS
     1995. pp. 223-229, 1995.

[116] F. Baille, G. Borel, B. Commere, F. Roy, C. Delmas and C. Terrier, "Multi MRad
     hardened 8 bit/20 MHz flash ADC," First European Conference on Radiation and
     its Effects on Devices and Systems, RADECS 91. La Grande-Motte, France,
     September 9-12 1991. pp. 271-276, 1992.

[117] N. Van Vonno, H.D. Davidson and R. DeMair, "Radiation-tolerant 8-bit flash
     A/D converter implemented in silicon-on-sapphire," First European Conference on
     Radiation and its Effects on Devices and Systems, RADECS 91. La Grande-Motte,
     France, September 9-12 1991. pp. 241-248, 1992.

[118] P. Leszkow and M. Decreton, “Radiation tolerant techniques of multiplexing
     analog signals in a nuclear environments,” First European Conference on Radiation
     and its Effects on Devices and Systems, RADECS 91. La Grande-Motte, France,
     September 9-12 1991. pp. 254-257, 1992.

[119] "IEEE standard digital interface for programmable instrumentation : IEEE
     recommended practice for code and format conventions for use with ANSI/IEEE
     Std 488-1978," Institute of Electrical and Electronics Engineers, Wiley-Interscience,
     New York, 1983.

[120] “Military standard. Test methods and procedures for microelectronics,” Mil-Std-
     833D, US Department of Defense, 1991.
                                           197


[121] “Methods for application of ionization chambers to assess the low energy gamma
     component of Cobalt-60 irradiators used in radiation hardness testing of silicon
     electronic devices,” ASTM E1250-88, American Society for Testing and Materials,
     1988.

[122] “Practice for minimizing dosimetry errors in radiation hardness testing of silicon
     electronic devices using Cobalt-60,” ASTM E1249-88, American Society for
     Testing and Materials, 1988.
                              BIOGRAPHICAL SKETCH



       Laurent Houssay was born in France in 1976. He graduated in 1996 from the

University Institute of Technology (IUT) of Toulouse, Physical Measurements

Department. He spent the following two years at the National Engineering School of

Physics in Grenoble (ENSPG). The ENSPG is one of the nine engineering schools of the

National Polytechnic Institute of Grenoble (INPG). He then came to the Nuclear &

Radiological Engineering Department at the University of Florida in 1998 as a part of an

agreement while finishing his degree requirement; he graduated from ENSPG in 1999.

He then worked for two years in the Robotics Laboratory of the Nuclear Engineering

Department under the supervision of Pr. James Tulenko while pursuing a Master of

Science degree in engineering hysics.




                                          198

								
To top