NASA Technical Memo.randurn 105248L

AIAA-91-3525

p./s-

Neutron, Gamma Ray and Post-IrradiationThermal Annealing Effects on Power

_ Semiconductor Switches _(NASA-TN-I0524O) NEUTRON, GANNA PAY AND

POST-T_PA_IATI_N TH_ENAL ANNEALINg FFFECTS

ON POWER SEMICONDUCTOR SWITCHES (NASA)

i5 P CSCL 09A

N91-32#10

G_I33

Uncl,_s

0065767

G.E. SchwarzeLewis Research Center

Cleveland, Ohio

and

A.J. Frasca

Wittenberg University

Springfield, Ohio

1

Prepared for theConference on Advanced Space Exploration Initiative Technologies

cosponsored by AIAA, NASA, and OAI

Cleveland, Ohio, September 4-6, 1991

https://ntrs.nasa.gov/search.jsp?R=19910023096 2020-04-23T08:41:22+00:00Z

NEUTRON, GAMMA RAY AND POST-IRRADIATION THERMAL ANNEAUNG EFFECTS ON POWERSEMICONDUCTOR SWITCHES

G. E. Schwar-zeNASA Lewis Research Center

Cleveland, Ohio

A. J. Frasca

Wittenberg UniversitySpringfield, Ohio

Abst ct

The effects of neutrons and gamma rays onthe electrical and switching characteristics ofpower semiconductor switches must be knownand understood by the designer of the powerconditioning, control, and transmissionsubsystemof space nuclear power systems. The SP-100radiation requirements at 25 m from the nuclearsource are a neutron fluence of 1013 n/cm 2 and

a gamma dose of 0.5 Mrads. Experimental datashowing the effects of neutrons and gamma rayson the performance characteristics of power-typeNPN Bipolar Junction Transistors (BJTs), Metal-Oxide-Semiconductor Field Effect Transistors(MOSFETs), and Static Induction Transistors(SITs) are given in this paper. These three typesof devices were tested at radiation levels whichmet or exceeded the SP-100 requirements. Forthe SP-100 radiation requirements, the BJTs werefound to be most sensitive to neutrons, theMOSFETs were most sensitive to gamma rays,and the SITs were only slightly sensitive toneutrons. Post-irradiation thermal anneals at 300K and up to 425 K were done on these devicesand the effectiveness of these anneals are alsodiscussed.

Introduction

High efficiency and low specific-mass powerand control electronic components for future highcapacity power space systems, such as thoseproposed for the NASA Space ExplorationInitiative (SEI), will be required to function reliablyfor long time periods with operation most likely inharsh environments. For power systems using anuclear energy source, this harsh environment willinclude a wide energy spectrum of neutrons and

gamma rays emitted by the fission reaction. Theneutron fluxes and gamma dose rates to whichthe electrical components will be subjected, willdepend on both the type of shielding arrangementand the distance between the components andnuclear source. For example, for the SP-100delivering 100 kW at 25 m from the nuclearsource, the neutron flux is 4.5 x 104 n/cm2s andthe gamma dose rate is 8.15 rad/hr. These arerelatively low rates, but for a projected 7 yearmission, the neutron fluence is 1013 n/cm 2 and

the gamma dose is 0.5 Mrads.

Both semiconductor power switches andcontrol electronics will be required in the PowerConditioning and Control subsystem, and also inthe Instrument and Control subsystem for thenuclear reactor. The performance characteristicsof high power semiconductor switches andcontrol electronics subjected to high levels ofneutron fl'_=nces and gamma doses, and alsopossibly high temperatures, must be clearlyknown and understood by the designer of thesesubsystems if flawless mission operation isexpected.

The NASA Lewis Research Center has anexperimental test program underway to determineand assess the effects of nuclear radiation on theelectrical and switching characteristics ofcommercial and developmental-type powersemiconductor switches. The primary thrust ofthe experimental work is to clearly ident_y thoseswitch characteristics most sensitive to either orboth gamma rays and neutrons. The conductionof in-situ experimental tests enables thedetermination of the rates of degradation andfailure mechanisms of the switch'scharacteristics as a function of gamma dose rate

This paper fs dectared a work of theU.S. Government and is not subject to

copyright protection in the United States.

and total dose, and neutron flux and fluence at agiven temperature. The experimental resultsobtained under this program will permit the circuitdesigner to determine the useabilityof a particulartype of semiconductor switch for a specifiedmission application.

The types of semiconductor switches to beexperimentally tested under this program includepower diodes, transistors, and thydstors. Thetransistors Include the Bipolar Junction Transistor(BJT), the Metal-Oxide-Semiconductor Field EffectTransistor (MOSFET), the Insulated Gate BipolarTransistor (IGBT), and the Static InductionTransistor (SIT). The thydstors include both thephase-control and Inverter type Silicon ControlledRectifier (SCR) and the MOS Controlled Thyristor(MCT). All the power devices Investigated to dateused silicon semlconductors. A new andemerging semiconductor technology beingworked on today is silicon carbide (SIC). SiCswitches offer the real possibility of highertemperature operation than silicon switches. Asthese SiC devices become available, theirtolerance to radiation effects will beexperimentally Investigated.

This paper will give the experimental resultsshowlng the effects of neutrons, gamma rays, andpost-irradiation thermal anneals on power-typeNPN BJTs, N-channel enhancement modeMOSFETs, and N-channel SITs. A briefdiscussion of the effects of neutrons and gammarays on electronic materials and devices will firstbe given.

Gamma Ray and Neutron Effects

The physics of radiation damage in semi-conductor switches is quite complex and theeffects of radiation on a switch's electrical and

dynamic switching characteristics are dependenton the device's physics. The brevity of this paperdoes not allow a complete discusslon of all of theeffects associated with gamma rays, neutrons,and thermal annealing of radiation damage inelectronic materials and devices, so the Interestedreader is referred to reference (1) and thereferences cited therein. A brief description of theinteraction of gamma rays and neutrons withmatter follows.

Gamma rays are high energy photons orquanta and can be produced when a nucleusdecays from its excited state to another lower orground state. Gamma rays interact with matter inthree different ways: photoelectric effect,Compton scattering, and pair production.Electrons are emitted in each of these processes.In semiconductors and Insulators, ionizationproduces electron-hole pairs; an electron isemitted and a mobile hole is generated in thevalence band. About 3.6 eV is required to createan electron-hole pair In silicon. Charged particlessuch as electrons, protons, and alpha particlesInteract with atoms primarily by Rutherfordscattering (Coulomb scattering). An ionizedelectron Imparted with sufficient kinetic energycan cause additional ionization or displacement ofatoms. In sUicon the photoelectric effectdominates at photon energies less than 50 keVand pair production dominates at energies greaterthan 20 MeV with Compton scattering dominatingin the intervening energy range. Insemiconductor devices, the main effects ofionization are hole traps or positive charge build-up in the insulator (oxide) or passivation layer,electronic or interface states at the Interfacebetween the Insulator and semiconductor, andphotocurrents.

Neutron interactions with matter can bedivided into two categories: capture andscattering, in the capture process the targetnucleus absorbs the incident neutron to form acompound excited nucleus which subsequentlydecays to a stable nucleus through photonemission or charged particle emission(transmutation). In certain heavy nuclei thecompound nucleus splits (fissions) into twoseparate fragments accompanied by the emissionof neutrons and gamma rays. In scatteringinteractions the Incident neutron remains free at a

lower kinetic energy and part of the Incidentneutron's kinetic energy is transferred to thetarget nucleus known as the primary recoil orknock-on atom. In scattering Interactions thetarget atom in crystalline materials can bedisplaced from its lattice site and for mostmaterials about 25 eV is required to dislodge theatom. Displacement of atoms can cause eithersimple defects (vacancy-interstitial pairs) orcomplex local rearrangement of the atoms knownas cluster defects. Through thermal motion some

of thesedisplacedatomsbecomemobileandmigratethroughthe lattice structure until theyeither recombine as vacancy-interstitial pairs, orform Immobile stable defects, or escape to a freesurface. In crystalline substances these atomicdisplacements alter the periodicity of the lattice.This disruption of lattice periodicity in asemiconductor or Insulator leads to discreteenergy levels In the material's bandgap. Thepresence of these discrete energy levels In thebandgap can cause several processes to occur:generation, recombination, trapping, removal, andtunneling of electron and hole carriers. Besidesthese processes, radiation induced atomicdisplacements generate carrier scattering centerswhich affect carder mobility.

University of Cincinnati (UC) are used for theexperimental tests, data analysis, and assessmentof radiation damage. The OSU research reactorfacility is used for neutron tests and the peak inthe reactor's neutron energy spectrum is near 2MeV. The UC Cobalt-60 facility Is used forgamma ray tests and this faciltty consists of alarge water pool with the Cobalt-60 sourcelocated about fifteen feet below the water surface.The use of a submerged Cobalt-60 source allowsfor a significant amount of gamma backscatterwhich reduces the gamma energy to which thedevices are subjected. This broadened energyspectrum is typical of the energy for gamma raysemitted in a nuclear reactor due to fission

fragment decay.

Experimental Setup

In the experimental test program theapproach is to make all electrical and switchingmeasurements of the power semiconductorswitches under in-situ radiation and temperatureconditions whenever possible. The approachconsists of three main steps. The first step is todetermine the effects of gamma dose rate andtotal dose, and neutron flux and fluence underroom temperature conditions on the swltch'scharacteristics. Post-irradiation thermal annealsfor long periods of time at room temperature andshort-term anneals at elevated temperatures arealso part of this step. The next step is todetermine the effects of gamma rays andneutrons at elevated temperatures. The possibilitythat the effects of neutrons and gamma rays onpower semiconductor switches at temperatureswell above room temperature will cause lesspermanent radiation damage needs to be fullyInvestigated. Almost all of the devices Irradiatedto date were done at or near room temperature.The third and final step of the test program will beto integrate the semiconductor electronics andother electrical components In a circuit anddetermine degradation rates and failuremechanisms for the circuit when it is subjected togamma and neutron irradiation and operatingtemperature conditions.

The research facilities and equipment of theNASA Lewis Research Center, WittenbergUniversity, Ohio State University (OSU), and the

Almost all the devices Irradiated to date weredone at or near room temperature. Some initialneutron Irradiation tests have been conductedrecently at temperatures up to 365 K. Theprimary purpose of these tests was to test theadequacy of a newly fabricated high temperaturefixture. The room temperature neutron andgamma irradiation tests lasted from several hoursto days in order to obtain specific fluences ordoses at controlled fluxes or dose rates,respectively. Several of the gamma Irradiationtests were done at low dose rates to duplicate thegamma dose received by similar devices duringreactor tests. It was found that the gamma dosereceived during the reactor tests for fluences onthe order of 1013 n/cm 2 was only significant forpower MOSFETs whose gate-source voltage isquite sensitive to Ionization caused by gammarays.

The test fixture for the gamma and neutrontests at room temperature is made frompolyethylene and symmetrically holds up to fourdevices. The fixtures outer diameter is limited bythe reactor's port hole diameter of six inches.The fb<ture's size does not cause any limitationsIn the experimental tests since the time requiredfor the in-situ electrical measurements and

analyses sets the upper bound on the number ofdevices which can be irradiated. Both power andsense cables are provided for each test deviceand these cables extend outside either theneutron or gamma ray source to the testInstruments. The Inner surface and the fourindividual test sockets of the fixture are lined with

cadmiumtoabsorbthethermalneutrons,leavingonlytheeplcadmlumneutrons(0.5eV to 10 MeV).Displacement of atoms In silicon requires about25 eV so that the inclusion of thermal neutrons inthe fluence is not a fair representation of thedevice's capability in terms of neutron hardness.Thermal neutrons pdmadly cause neutronactivation, that is, transmutation, and minimizationof neutron activation is highly desirable so thatImmediate post-irradiation tests can be conductedoutside the reactor to monitor the effects of

thermal annealing at room temperature.

A curve tracer is used to make the electricalmeasurements before, during, and afterIrradiation. These measurements consist of theON-state current-voltage characteristic curvesandthe OFF-state leakage current-voltage breakdowncurves. The data from these curves are used toplot the switch's electrical parameters as afunction of either the neutron fluence or gammadose. These plots clearly identify those electricalparameters most sensitive to either gamma raysor neutrons, or both, by graphical illustration ofthe degradation rates. Special drive circuits aredesigned for the switching-time tests for each ofthe different types of switches. Presently, thetests for determining switching times areconducted only before and after Irradiation.

Upon completion of either a neutron orgamma Irradiation =test run, the electrical andswitching characteristics are monitored todetermlne any post-irradiation changes in devicecharacteristics at room temperature. Devicesfrom each test run are also subjected to thermalanneals at elevated temperatures to determineany changes in device characteristics caused byradiation damage. The thermal anneal times atelevated temperatures are sufficiently long toensure that any parameter changes havestabilized at the annealing temperature. Anychanges in device characteristics appear to becomplete in an hour at a specified temperature.

Experimental Results

Radiation and Annealina Effects in BJTs

The BJT is a current controlled device In

which the base current IB controls the outputcollector current IC. The ratio of IC to Ia is the DC

current gain, hFE. Physically, hFE represents thefraction of majority carders emitted by the emitterthat pass through the base as minority cardersand collected as majority carders by the collector.

Fast neutrons primarily cause duster latticedefects and they are the dominant radiationInduced damage mechanism In BJTs. Gammarays are also capable of generating lattice defectsby energetic electrons resulting primarily fromCompton scattering. Lattice defects cause theformation of Recombination-Generation (R-G)centers. As the R-G center density Increases witheither Increasing neutron fluence or gamma dose,the minority carrier recombination lifetimedecreases so that the rate of electron-holerecombination increases in the base. Decreasesin minority carder lifetime causes decreases inboth hFE and the switching storage time. Adecrease in switching storage time Is beneficialbecause it reduces the switching losses. Adecrease in hFE iS, of course, detrimental becauseIC decreases for a constant IB, or equivalently, IBmust Increase to maintain a given Ic. Aswitching-type power BJT has a relatively low hFE

at rated Ic and forward collector-emitter voltageVCE, SO any degradation in hFE impacts theswitch's operational use. A decrease in hFEdoesnot make a switching-type power BJT inoperativeas long as the device does not exceed itsmaximum continuous IB rating, but a decrease inhFE makes the BJT a more Inefficient switchbecause the switch's total losses increase as IBincreases to maintain a given IC.

The results given In Figures 1 through 12 arefor a D60T455010 power NPN transistor and withmanufacturer's ratings of hFE = 10 at IC = 50 Aand VCE = 2.5 V, maximum continuous IB = 20 A,VCBO = 500V, andVcE O = 450V. Figures 1through 9 show the effects of neutrons and post-irradiationthermal annealing onthe characteristicsof this transistor, while Figures 10 through 12show the effects of gamma rays and post-Irradiation thermal annealing.

Figure 1 is a plot of hFE versus theepicadmlum neutron fluence for three different ICvalues. These curves clearly show the rate ofdegradation of hFE with increasing fluence. TheIC = 10 A and 30 A curves are for a total fluenceof 1.65 x 1013 n/cm 2. The IC = 50 A curve

4

terminatesat 0.74x 1013n/cm2becauseabovethisfluence,IBexceeds20A. ExtrapolationoftheIc = 50 A curve shows that as the fiuenceapproaches 1013 n/cm 2, hFEapproaches unity.

Figure 2 is a different way of conveying thesame type of Information given in Figure 1.Specification sheets for BJT switches generallygive plots of hFE versus Ic to show thedependency of hFEon Ic. Figure 2 gives four hFEversus IC curves; the top curve is for pre-Irradiation conditions, while the next lower andtwo following curves are for Increasing fluencelevels. Like Figure 1, the results given in Figure 2clearly show the decrease in hFEwith increasingfluence. It should be noted that the totalbackground gamma dose of 37 krads shown inFigures 1 through 9 is due to fission fragmentdecay and this relatively low dose level had nosignificant effect on either the electrical orswitching characteristics of this transistor asverified by Independent gamma experiments.

Figure 3 shows the effects of post-neutronirradiation thermal annealing on hFE. In thisFigure the upper curve is the pre-irradiation hFE

versus IC curve and the lower curve is the post-irradiation curve for a fluence of 1.65 x 1013n/cm 2. The data points just above the lowercurve are for post-irradiation thermal annealingconditions at 300 K for various periods of time,and at 425 K after 6 months at 300 K. The datafor the 425 K thermal anneal was taken at 300 K.These results show that thermal anneals at 300 K

for long periods of time cause no significantincrease in hFE. Likewise, no significantimprovement in hFE iS caUSed by the 425 Kthermal anneal. These annealing results indicatethat the severe degradation in hFE caUSed byneutrons at 300 K is permanent and can not bereversed for thermal annealing temperatures upto425 K.

Figure 4 Is a plot of forward VCE versus I B forvarious values of IC. The lower set of threecurves is for pre-lrradiation conditions while theupper set of three curves is for a neutron fluenceof 1.65 x 1013 n/cm 2. The lower (pre-lrradiation)curves show that VCE for a given Ic does notchange for Ia z 5 A. The upper (post-Irradiation)curves show that VCE IS affected by neutrons asfollows: First, the post-Irradiation set of curves

are all displaced upward from the pre-irradiationset of curves so that this result shows that for a

given Ic and IB (i.e., a fixed hFE), VCE Increaseswith ttuence. Second, the post-irradiation curves

show that VCE increases as IB decreases. A lowVCE is required for low conduction losses in theswitch. Thus, the post-irradiation curves showthat low conduction losses can only be obtainedby using high IB values; but high IB means thebase circuit must supply higher currents, and thisresults in higher losses In that circuit. High IBvalues also cause increases in switching storagetimes, and this causes higher switching losses.Thus, even though the BJT Is operable at thisfluence level, its efficiency is degradedconsiderably.

Figure 5 is a plot of Ic versus the base-emittervoltage, VBE, for various values of IB. Comparisonof the pre-irradiation (dashed lines) curves wtththe post-irradiation (solid lines) curves shows thatneutrons have a negligible effect on the IC versusVBEcurves. The post-irradiation curves terminateat lower values of IC than the pre-irradiationcurves because of the decrease In hFE for thisfluence level.

In a reverse-biased p-n junction the driftcurrent dominates and this reverse current,

usually referred to as the leakage current,consists of two components. The onecomponent, the reverse saturation current,assumes a zero R-G rate in the depletion regionof the junction while the other componentassumes R-G does occur in the depletion region.Both current components increase as the minoritycarrier lifetime decreases (i.e., electron-holegeneration rate increases) so that an Increase inradiation induced defects causes an Increase inleakage current. Thus, for a specified reversebiased collector-base voltage with open emitter,

VCBo, and specified collector-emitter voltage wtthopen base, VCEo, the respective leakage currents,ICBO, and ICEO, increase as the radiation Induceddefects Increase, or equivalently, for specified

ICBO and ICEO, VCBO and VCEo, respectively,decrease as the density of defects increases.

Figure 6 is a plot of VCE O for ICEO = 10 _&.versus the eplcadmlum neutron fluence for fourdifferent D60T455010 devices. Two of the deviceswere tested at a flux of 7.55 x 108 n/cm2s for a

fluenceof 1.65x 1013 n/cm 2 and the other twowere tested for the same pedod of time but at halfthe flux to give half the fluence. The results showthat the device with the lowest pre-irradiatlon

VCE0 is the least affected by neutrons while thedevice with the highest pre-irradiation VCE0 is themost affected. The explanation of this wouldsuggest at least two possibilities. The first is thatthe collector of the device with the low initialVCE0is more heavily doped than the collector of thedevice with the high initial VCEO and concludethat the higher the collector dopant level, the lesssusceptible the collector is to neutron inducedlattice defects. The second possibility is that thedevice with low initial VCE0 haS current pathsexternal to the semiconductor, and conclude thatthese external leakage currents mask thesemiconductor leakage currents for low fluences,but as the fluence increases and more latticedefects are generated, then the semiconductorleakage current begins to exceed the externalleakage current. The acceptance of either orneither of these explanations would require theconduction of additional experiments. It shouldbe noted that the VCBO versus fluence curves forthese same devices show similar results.

In Figure 7, for three of the same devicesdiscussed in Figure 6, VCE0 iS held constant at400 VDC and ICEO iS plotted as a function ofepicadmlum neutron fluence. Figure 7 conveyssimilar Information as that given in Figure 6. Thatis, the device with the highest pre-irradiation ICEOshows only a small increase in ICE0 withIncreasing fluence, while the device with thelowest pre-lrradiation ICEO shows a verysignificant Increase in ICEO with Increasingfluence.

Figure 8 gives a set of ICEO versus VCE0curves for one of the devices listed in Figures 6and 7. The lower curve in Figure 8 is for pre-irradiation conditions, the upper isfor a t]uence of1.65 x 1013 n/cm 2, and the curves in betweenshow the effects of post-irradiation thermal

annealing at 300 and 425 K. ICEO cleady showsrecovery 48 hours after neutron irradiation andconsiderable recovery after 6 months. A 425 Kthermal anneal after the 6 months period at 300 K

causes ICEO to recover almost to its pre-irradiation value. It is possible that the ICEOrecovery shown for 6 months at 300 K occurred

before that time since no ICEO data was takenbetween 48 hours and 6 months. ICE0 showsexcellent recovery compared to hFE for similarannealing conditions. These ICEO annealingresults would Indicate that the radiation damagein the collector is considerably less permanentthan that in the base. At this time it Is not

obvious why the radiation damage in the collectorresponds to annealing treatment but this sametreatment is not a cure for the radiation induceddamage in the base.

The four states in the switching sequence ofa BJT are OFF, turn-on, ON, and turn-off. TheOFF-state, the non-conduction interval, is

characterized by VCBO and VCEO and theirrespective leakage currents, ICB0 and ICEO. TheON-state, the conduction interval, is characterized

by hFEat specified IC and VCE values. The turn-on time Is the sum of the delay dse and dse time.Durtng turn-on, excess minority cardersaccumulate In the base as the BJT goes from cut-off through the active to the saturation region.Excess minority carrier charge accumulatedduring turn-on must all be removed during turn-off. The turn-off time is the sum of the storageand fall times. For a BJT operating in thesaturation region, the storage time is the longestswitching time compared to the other threeIntervals. The excess carrier charge accumulateddudng turn-on and the rate of recombination ofthis charge during turn-off is proportional to theminority carrier lifetime in the base.

Figure 9 shows the prre- and post-irradiationstorage time for the two devices subjected to afluence of 1.65 x 1013 n/cm 2 as a function of !Bfor Ic = 40 A and collector supply voltage Vcc =45 VDC. The post-irradiation curve clearly showsa significant reduction of over 100% in storagetime due to the radiation Induced defects. Post-

irradiation thermal annealing procedures, similarto those shown in Figure 3 for hFE,did not causeany significant changes in the post-irradiationswitching times for these devices. The decreasein minority carrier lifetime caused by the neutron-induced atomic displacements cleady enhancesthe switching characteristics of BJTs but likewise,severely degrades the ON-state current gain andOFF-state forward blocking voltage capability.

The final three figures on BJTs show the

6

resultsof Co-60gammarayson the ON-andOFF-statesofthesametypeof devicesusedforthe neutron irradiations. Figure 10 is a plot of hFEversus gamma dose for three different IC values.These curves cleady Indicate that high doses ofgamma rays are capable of causing atomicdisplacements and generating lattice defects thatcause hFEto decrease with increasing dose.

In Figure 11, for the same device listed inFigure 10, the upper curve is the pre-irradiationhFE versus IC curve and the lower curve is thepost-irradiation curve for a gamma dose rate of92 krads/hr for a total gamma dose of 4.9 Mrads.The data points Justabove the lower curve are forpost-irradiation thermal annealing conditions at300 K for various periods of time, and at 425 Kafter 9.5 months at 300 K. The data for the 425 Kthermal anneal was taken at 300 K. Just like the

thermal annealing results for the neutronirradiated device, the annealing results for thegamma irradiated device show that both the long-term thermal anneals at 300 K and the 425 Kthermal anneal have no significant effect on therecovery of hFE. These experimental resultswould indicate that gamma irradiation at 300 Kcauses the formation of permanent damage siteswhich are unable to become mobile at a 425 K

annealing temperature.

Figure 12 gives two ICEo versus VCEo curvesfor the same device listed in Figures 10 and 11.The bottom curve In Figure 12 is the pre-Irradiation data and the top curve is the post-irradiation data. Gamma Irradiation to 4.9 Mrads

causes ICEO to Increase slightly and this changein ICEO iS almost completely reversed after a 9.5month thermal anneal at 300 K, Ana!ysls of theexperimental data shows that some of theIncrease in ICEO at 4.9 Mrads is caused byphotocurrent, which is not caused by radiationinduced damage. Thus, the response of ICEOandhFE to thermal annealing is the same for bothneutron and gamma ray induced damage causedunder 300 K conditions.

Radiation and AnnealinQ Effects in MOSFETs

The MOSFET, which conducts current bymajority carriers, is a voltage controlled device inwhich the gate-source voltage controls the outputdrain current. MOSFETs are of two types:

depletion- and enhancement- mode. Both typescan be either N- or P- channel devices.

Depletion-mode MOSFETs turn on with zero gate-source voltage and are normally-on devices, so agate voltage Is required for turn-off and tomaintain the drain-source blocking voltage in theOFF-state. Enhancement-mode MOSFETs are

normally-off devices because a gate-sourcethreshold voltage VGs(th(th)IS required for tum-on.Tum-on is accomplished by developing a chargeinversion layer in the semiconductor under thegate oxide to form the conducting channel. BothN- and P- channel enhancement-mode MOSFETshave been developed for power applications withN-channel devices having the largest selection ofcurrent and voltage ratings. Accordingly, N-channel enhancement-mode MOSFETs were thefocus of our irradiation studies, and thus, thesewill be the basis for the discussion of radiationInduced effects and the effects of post-irradiationthermal annealing. The devices Investigated werea high voltage, low current device (MTM15NS0);and a medium voltage, medium current device(IRF250). The manufacturer's specifications forthe MTM15N50 are: drain-source breakdown

voltage with gate.source shorted, BVDss = 500 V;continuous drain current, ID = 15 A, VGs(th) = 4.5V (max.), drain-source on-resistance, RDS(ON) -0.4 ohms iJ,_ax.) at Io = 7.5 A and VGS = 10 ;.The manufacturer's specifications for the IRF250are: BVDsS = 200 V, ID = 30 A, VGs(th) = 4 V(max.), and RDS(ON)= 0.085 ohms (max.)at ID =16AandVGs = 10V.

Neutron Irradiation of MOSFETs causes

RDS(_C_) to Increase for fluences greater than 1013n/cm. Radiation induced crystal defects causedby atomic displacements cause decreases inmajority carrier density (carrier removal), andcarrier mobility. The electrical conductivity Isproportional to the product of majority carderconcentration and mobility so a reduction inchannel and drift-region conductivity causes

RDS(ON)tO Increase.

Figures 13 and 14 show the effects of

neutrons on RDS(ON)for two MTM 15N50s and twoIRF250s, respectNely. The four test devices weresubjected to neutron fluxes from 3.78 x 10 7

n/cm2s to 2.27 x 109 n/cm2s for a total fluence of3.8 X 1013 n/cm 2 and a background gamma doseof 85 krads. No bias voltages were applied

dudngirradiationexceptdudngbrieftimepedodswhenthemeasurementsweremade.TheMTM15N50(500V/15A)devicesinFigure13beginto

• __ 1^13show=a definite Increase in RDS(ON_Deyono un/cm z while the IRF 250 (200 VT_ A) devices inFigure 14 show only a very small Increase inRDS(ON) beyond 1013 n/cm 2. These resultsIndicate that the high voltage devices are moresusceptible to neutrons than the medium voltagedevices but In either case, both are Immune toneutron tluences up to 1013 n/cm 2.

dose curves In Figure 15 show that the two MTM15N50s track each other quite well, and the sameholds true for the two IRF250s In Figure 16. A

comparison of the VGs(th) versus gamma dosecurves in Figure 15 with those In Figure 16 shows

that the response of VGS(th)to gamma rays is thesame. The curves in Figures 15 and 16 cleady

show that VGs(th) for both MOSFETs is verysensitive to gamma rays. These curves also show

that at 1.5 Mrads, VGS(th)has become negative forthe MTM15N50 and zero for the IRF250.

Gamma irradiation of MOSFETs primadly

causes VGs(th) shifts through the ionizationprocess. Ionizing radiation generates electron-hole pairs almost uniformly In the gate oxide.These electrons and holes are separated by anapplied or radiation Induced electric field in theoxide. The trapped holes In the oxide act like apositive gate-source voltage and cause chargeinversion in the semiconductor near the oxide-semiconductor Interface in the same way as anapplied positive gate-source voltage. The positivecharge build-up in an N-channel enhancement

mode MOSFET causes a negative shift in VGS(th)SOthat a less positive bias is required for turn-on.Sufficient positive charge build-up can cause

VGs(th) to be equal to or less than zero. ForVGS(th) s 0, the N-channel enhancement modeMOSFET behaves like an N-channel depletion-mode MOSFET and turns on when a drain-sourcevoltage is applied. Once the enhancement-modeMOSFET has reached this state, a negative biasis required for turn-off and maintenance of theOFF-state. Gamma rays can also cause theformation of interface or surface states which can

compensate the negative shift in VG th causeds( )by the positive charge build-up in the oxide. Asthe gamma dose increases, this compensationeffect can Increase to cause a decrease In the

rate of negative shift in VGs(th)to the point thatVGS(th)shifts positively.

The effects of Co-60 gamma rays on VGSthare shown in Figure 15 for two MTM15N50s andin Figure 16 for two IRF250s. The four deviceswere first subjected to a gamma dose rate of 33.8krads/hr to give a dose of 67.6 krads, and thenthe dose rate was increased to 62.1 krads/hr untila total dose of 1.5 Mrads was reached. No bias

voltages were applied during Irradiation exceptduring the test measurements. The VGS(th)versus

Rgure 17 shows the effects of post-irradiationroom and elevated temperature thermal annealingon VGs(th) normalized to its pre-irradlation valuefor the MTM15N50 and IRF250. The left portionof the plot gives a representation of the change In

the normalized VGs(th)for a gamma dose up to1.5 Mrads. The details of this portion of the graphwere previously presented in Figures 15 and 16.The mid-section of Figure 17 represents therecovery in the normalized V for each deviceafter 30 days at 300 K. After t his time periodGslth

VGs(th)is now greater than zero for each device.However, the VGs(th) required for turn-on isconsiderably less than that required for pre-Irradiation conditions. The right side of the plot inFigure 17 shows the change in the normalized

VGs(th)as a function of annealing temperature upto 425 K. Each device was annealed at a startingtemperature of 325 K and then in 20 K Incrementsup to 425 K for time periods of 3 to 7 hours ateach temperature. The devices were cooled toroom temperature after each annealing

temperature for the VGs(th) measurement. Thethermal anneals at the elevated temperatures

caused VGS(th) to increase slightly for theMTM15N50 but had no noticeable effect on theIRF250. Thus, it is concluded that the gammainduced positive charge build-up inthe gate oxide

is very stable because VGS()th shows only a smallImprovement after the 30 day, 300 K thermalanneal, and no significant Improvement forthermal anneals up to 425 K.

Radiation and Annealina Effects in SITs

The Static Induction Transistor (SIT) is a highvoltage, power Junction Field Effect Transistor(JFET). The SIT is a voltage controlled device Inwhich the gate-source voltage controls the output

draincurrent.TheSITis a normally-onswitchthatneedsnogate-sourcevoltagefor tum-on butrequires a gate voltage for both turn-off andmaintenance of the blocking state. The SIT is amajority carder device and can be thought of asa bar of n- or p- type semiconductor to form anN- or P- channel, respectively, in which thechannel resistance, and hence the drain current,is controlled by varying the gate-source voltage.Because the SIT Is a majority carder device, andalso because It has no gate oxide, the SIT is anextremely fast switch with switching times oftenths of microseconds for both turn-on and tum-off.

The current-voltage characteristics of a JFETare very dependent on the ratio of length to widthof the drain-source current channel. The

requirements for a switching-type power JFET arehigh drain current handling capability and low

RDS(QN) for the ON-state and high forwardblocking capability and low leakage current forthe OFF-state. Neither the short- nor the long-channel JFET is acceptable as a power device.The power JFET, such as the SIT, Is anIntermediate vertical channel device that exhibitspentode-like characteristics at high drain currentsand low gate reverse-bias voltages, and triode-likecharacteristics at low drain currents and high gatereverse-bias voltages. Even though high forwardblocking voltages can be achieved for SITs, theirON-state voltage drop is quite high and thus, SITshave relatively low current carrying capabilitycompared to power BJTs and MOSFETs withcomparable blocking voltages. The ON-statevoltage drop of the SIT can be reduced byoperating It in the bipolar mode of operation, thatis, by applying a positive gate bias to an N-channel device.

Because the SIT has no gate oxide it shouldbe relatively Insensitive to gamma rays comparedto the MOSFET. The principle radiation damagemechanism In SITs is atomic displacementsprimarily caused by neutrons. The effect ofatomic displacements in SITs is the same as in

MOSFETs; RDS(ON) increases with fluencebecause channel conductivity decreases due todecreases in carrier mobility and majority carderconcentration.

The 2SK180 N-channel SiT was selected for

neutron and gamma Irradiation tests. Themanufacturer's specifications for this device aregate-drain breakdown voltage = 600Vat ID = 0.1mA, gate-source breakdown voltage = 70 V at ID= 0.1 mA, ID (max.) = 20 A, ID (continuous) = 8A and RDS(ON)= 1.5 ohms at VGS = 0 V and ID =2A.

Figure 18 shows the effect of neutrons on

RDS(ON)for the 2SK180. The test devices weresubjected to neutron fluxes from 7.55 x 107 tO3.78 X 109 n/cm2s for a total fluence of 5.1 x 1013

n/cm 2 and a background gamma dose of 94

krads. The top curve for RDS(ON)versus fluencein Figure 18 Is for the SIT's normal mode ofoperation at VGS = 0, while the bottom curve isfor the bipolar mode of operation at IGS = 10 A.A comparison of RDS(ON) prior to Irradiation forVGS = 0 V (normal-mode) and IGS = 10 A(bipolar mode) shows that RDS(ON)decreasesalmost Fr_etimes when the SIT is operated in thebipolar mode. The curves in Figure 18 show that

for both operating modes of the SIT, RoS(ON)firstIncreases gradually as the fluence approaches1013n/cm2after which RDS(ON)begins a steeperclimb. A 300 K thermalanneal for 70 days

caused RDS(ON)to decrease about 10% from itspost-irradiation value. A 425 K thermal annealfollowing the 300 K anneal had no significant

effect on RDS(ON). Thus, the neutron induceddamage in SITs Irradiated at 300 K is quite stablesince the damage sites remain immobile forannealing temperatures up to 425 K.

The 2SK180s were gamma irradiated atgamma dose rates of 10.15 to 55.1 krads/hr for atotal dose of 1.8 Mrads. The only changedetected in electrical characteristics was about a

30% increase in RDS¿ON)at1.8 Mrads when thedevice was operated in the bipolar mode at IGS =10 A. No change was observed in ROS(ON)whenthe SIT was operated in its normal mode at VGS=0V.

Conclusions

This paper gives the experimental results forthe effects of neutrons, gamma rays, and post-Irradiation thermal anneals on the performancecharacteristics of power-type NPN BJTs, N-channel enhancement mode MOSFETs, and N-channel SITs. Each of these devices was exposed

to a minimumneutronfluenceof 1013 n/cm 2 andgamma dose of 0.5 Mrads, which are the SP-100requirements for the electronics at 25 m from thenuclear source. All of the devices were still

operational after being tested to radiation levelswhich exceeded the mlnimum levels. Althoughthe switches were stilloperational, one or more ofthe switch's parameters was definitely affected byeither the neutrons or gamma rays.

_The BJl"swere fo_jndto=_most sensitive toneutrons and least sensitive to gamma rays. Aneutron fluence of 1013 n/cm 2 caused a verysignificant decrease in both hFE and switchingstorage time, and a significant Increase in ICEOand ICBO. These same BJT parameters wereaffected by gamma rays, but gamma dosesgreater than 0.5 Mrads were required to causeany significant changes. The N-channelenhancement mode MOSFETs were found to bemost sensitive to gamma rays and least sensitiveto neutrons. The MOSFET parameter most

sensitive to gamma rays is VGS_th); VGS(th)approaches zero as the dose approacnes I Mrad.

RDS(ON)iS the MOSFET parameter most sensitiveto neu_ronsl but neutron fluences greater than1013 n/cm 2 were required to cause significantincreases In this parameter. The SITs were foundto be neither sensitive to neutrons nor gammarays for fluences to 1013 n/cm 2 and doses_Mrads, respectively. For fluences beyond 1013

n/cm 2, ROS(ON) did begin to show a significantIncrease, but no significant increase was observedin RDS(ON)for gamma doses up to 1.8 Mrads.

Post-irradiation long-term thermal anneals at300 K and short-term thermal anneals up to 425K were conducted on samples of each of theirradiated devices to determine the extent ofpermanent radiation damage. For the BJTs, thethermal anneals had hardly any effect on hFEandthe storage time, but the anneals significantlyaffected ICEO and ICBO. The thermal annealscaused ICEO and ICBO to recover to almost theirpre-irradiation values. For the MOSFETs the

thermal anneals caused VGS(th) to recoversomewhat from f[s post-irradiation value, butVGstrth_ never came close to regaining its pre-irra_ia'tion value. For both the MOSFETs and

SITs, the thermal anneals caused RDS(ON) todecrease only slightly below its post-irradiationvalue.

Acknowledament

This research was sponsored by the NASALewis Research Center under the High CapacityPower element of the Civilian Space TechnologyInitiative.

Reference

1. Schwarze, G. E. and Frasca, A. J., "Neutronand Gamma Irradiation Effects on PowerSemiconductor Switches" in Pro_. of the 25thJntersociety Enerav Conversion EnglneednaConference. held in Reno, NV, August 12-17,1990.

10

lO0I_ ic- 10 A DEVICE: 14

_""- n _ n

[ u_ A u.

r',,_",A 9_8 ",-,"A

0 lO12 1_ $

EPICAD_IUM NEUTRON _ (n_m

FIGURE 1, OC CURRENT GAIN • VCE - 2.$ VVERSUE FLUENCE FOR NPNTRANSISTOR DOOT455010 (450 V/S0 A). FLUX. 7YJ6xl0%_m_; GAMMADOSE. 37 IvlKI.

10TM

100

i °

i '°

D _

8 o o o _uA 1.7k_1oI" 4

n ,..17=io _" eo ¢_o_'n¢ 1.mcxlOI'_ 37

11 10 10o

OOU._TOR CURRENT (A)

FIGURE2. DCCURRENTGAINOVcE . :_.5V VERSUS COLLECTOR CURRENT FORVARIOUS _UM NEUTRON FLUENCES AND BACKGROUND GAMMA

0OSES FOR NPN TRANSISTOR 060T4S6010 (4S0 VF_OA); NEUTRON FLUX.7.r_xl0 a n_l.

100

i

!

e_O_o_o\ De_E:_,o

0_ e

Q PRE-RAD lO POST-PAD v_

POSToR/t_D;t. 2.3 dPOST.RAZ); t. 6 m_ _abPOST-RAD; T. 42S K v.,,_.A

lO

COLLECTOR CURRENT (A)

100

FIGURE 3 DC CURRENT GAIN O VCE - 2.5 V V_RSUS COLLECTOR CURRENT FORPRE-IRRADIATION, POST-IRRADIATION (FLUX. 7.S6x10 II rVcm_. FLUENCE- 1 ._ 1013 rye2, _ DOSE. 37 ktads), AND THERMAL ANNEAUNGCONDITIONS FOR NPN TRANSlSTC_ D60T456010. ALL kF.ASUREMENT5MADE AT 000 K.

I00i

S

,I0

.... PRE-IRRAD_llONIN36T-IRRADIATION

DL_CE 14

m_

_ A ...... Ic .3OA

v_. Ic =20 A

_._0A V ....... V ....... V ....... V

;o 2'oBASE CURRENT(A)

FIGURE 4. COLLECTOR - EMITTER VOLTAGE VERSUS BASE CURRENT FOR NPNTRANSISTOR O_0T4S$010 (450 V/50 A); NEUTRON FLUX - 7.6Sx10s n/cm_.FLUENCE. 1.$6x1013n/em_; TOTAL GAMMA DOSE - 37Iv "_

6O

50

O 3O

.... PRE-IRRADIATION IB - SA DE_ICE: 20POST-IRRAOiATIO N

O

/ /[] X o

-1A e

: i il

il ! ilO0 0-2 0.4 0.6 0.B 1.0 1.2 1.4 1.6 1,8 20

BASE - ENrrrER VOLTAGE (V_

FIGURE 5. BASE-EMITTER VOLTAI_MEVERSUS COLLECTOR CURRENT FOR NPNTRANSISTOR D00T455010 (450 VF_0A); NEUTRON FLUX . 7,55XI08 r,,_ml_,FLUENCE. 1.6Sx1013_. TOTAL GN4A_ DOSE_ 37 knKb

_00

!t

r----"---'--'V _ V_V

o,,\

_------_-_ O_ O_ O. O_

DEVICE FLUX_Wm%)

oO _ } 7.SSx,o '

• 18 1. 3.7_1xlO8• I, J

1012 10_3 10|4

EPICADI41UM NEUTRON FLUF..4iCE(n/_ _)

FIGURE 6. BREAI_:)OWN VOLTAGE BVcEO _ ICEo - 101._ VERSUS FLUI_..E FOR NPNTRANSISTOR D_OT455010 (4S0V,_0 A), GAMMA DOSE; (20 & 14). 37 kind;(18 & 11) - l&S Iv--r.

11

100

1

0

FIGURE 7,

FLUX

DEVICE (nlem_)

14 } 7'_10s

1011 10 _2 10 t_ 10 _

B=tCADI_UM NEUTRON FLUENCE (n_m _)

CURRENT ICE o • VCE o o 400 VDC VERSUS FLUENCE FOR NPN

TRANSISTOR D6OT4.r_10 (450 VF_0 A); TOTAL GAMIJ_ DOSE (20 • 14) -

37 b'ads; (18). 11L$ krads.

30

j,o

J

RGURE 8.

POST-PAD DEVICE: 20

O/Q POST-RAD;, t. 2.S m

r'l A POST-RAD; t. 48 hrs

°Z,/oy

//_IA _0 POST-RAD; t. $ moJ

/DO_ / _0_'t" POST*RAD;t.$m_&425 K

/i_" O's-+ _ _e PRE.RAD

/// o_ +_e _

..........100 _0 300 400 f_O 600 700 BOO

COLLECTOR-EkrrrER VOLTAGE. Vc£. O (V)

COLLECTOR-BdlTTER LEAKAGE CURRENT VERSUS COU..ECTOR-EMMITE R

VOLTAGE FOR NPN TRANSISTOR D60T455010 FOR PREdRRADtATION.

POST-IRRADIATION (FLUENCE. 1._1013 _ 2, GAkaVK DOSE. 37 kra_).

AND THERMAL ANNEALING CON_S. ALL MEASURF..ME_T$ kt_D_ AT

3001(

100

10

1

01

FIGURE g

DEVICE

D_

0 t4 ........ Q

111- •

.... PRE-IRRA_A'I1ON

FOSTJRRADIATION

$ 8 10 i2 t4 16

BASE CURRENT (A)

COLLECTOR CURRENT STORAGE TIME FOR I¢. 40 A AND VCC o 4B V

Vl_RSUS BASE CURRENT FOR NPN TRANSISTOR DS0T4S6010 (450 VFJ0 A);

NEUTRON FLUX - 7.§Sx10 s r_m:_,, FLUENCE. 1.1_5x1013 Nan 2, G_

COSE. 37 krad

100

AD

!

_v

FIGURE 10.

I¢=EOA

IC=30A

Ic.40 A

D_V1CE: 25

_O_o-o.oo O-D_

O- o []

_o

10 _ 10 _; 10 7

Co-SO OU_MM_ OOSE (ram)

IX: CURRENT GAIN • VC_ - 2.S V VERSUS Co-¢_ GAMMA DOSE FOR NPN

TRAN_STOR D60T455010 (450 V/50 A); GAMMA DOSE RATE. 92 krad_

100

_o

!

• _-. DEVICE: 25

e..

§'-'_O _o O'e,o. O

0 POST-RAD il_.FXr'] POST-RAD; t. 45 rain

PO_T-RAD; t. 2 d

0 PO_T-RA_; t o 40 d

X POST-RAD; t. g.S m_

Jr" POST-RAD; T. 425

II1_ g.S mm_ _ 300 K

...... ' ' 1" 100

_CTOR CURRENT (A)

FIGURE 11_ DC CURRENT GAIN • VC:E - 25 V V_R_JS COLLECTOR CURRENT FOR

NPN TRANSISTOR D_0T4.E_010 (450 V/r'J0 A) FOR PRE-IRRADIATION.

POST-IRRADIA'nON (GAMMA DOSE . 4.9 khKb) AND THERMAL ANNEALING

CONDiTiONS ALL MEASUREMENTS MADE AT 300 K.

12

DEVICE: 25

10 •

8 • PRE-RAD /

O POST-PAD /[] POST-RAD; t - 12 d

S & I=OST-RAD;t = 9.5me= //O POST-PAD; T-425K //

4 --,.s--. 3ooK _j

' itli r: - __..il_t_T_1 ,

0 2OO 40O 6OO 8OO 1000

COLUECTOR-E_ITTER VOLTAGE, VCE o (V)

FIGURE 12 COLIJECTOR-EIvITTER LJEAKAG_ CUP, tENT VERSUS COLLECTOR-EMTTER

VOLTAGE FOR NPN TRANSISTOR [_0T45_O10 (450 VF_0 A) FOR FIRE ,

IRRADIATION, POST-IRRADIATION (GAMMA DOSE. 4.9 M'lds) AND THERMAL

ANNEAUNG COnDiTIONS. ALL MEASUREMENTS MADE AT 300 K.

12

0.S

!0.4

Io.31 _---_(t_"_-_ -°_-_'-"

!_ o_

_ DEVICE0.1 O 1A2

!0.0_ ...... -lo,o--- lo1' lo': lO" lo"

EP_ADMIUM NEUTRON FLUENCE (Ncm_)

F1GURE 13. DP.AIN-SOURCE ON-STATE RESISTANCE FOR ID. 15 A AND VO8-" 10 VVERSUS EI_CADMUM NEUTRON FLUENCE FOR N-CHANNEL MTM15NS0(SO0V/15 A); NEUTRON FLUX. 3.7k107 TO 2,27x10 t n_m_, GAMMA OOSE

0.16

0.1(

0.0(

--[3-[3 - nn

DEV_En?

A 10

0.00 H- " 101o 10g! 101g 1013 10140EPICADMIUM NEUTRON FLUENCE (n_m _)

FIGURE 14. DRAIN-SOURCE ON-STATE RESISTANCE FOR ID. 30 A ANOVo$ "- 10vVERSUS EPICADMIUM NE_ FLUENCE FOR N-CHANNEL IRF23O(200 V/'J0 A); NEUTRON FLUX. 3.7ki07 TO 227x100 r_cm_, GAMMADOSE. I15krmb.

4

eo o0

,t ..o,¢

.t................. .......• 104 10 lO lO 7

!

c_eo GAMMA DOSE (.m_)

FIGURE 15. GATE-SOURGE THRESHOLD VOLTAGE VERSUS GAMMA DOSE FORN-CHANNEL _ 5Nrj0 (S00 V/15 A}; GAMMA DOSE RATE. 33.8 TO 62.1kra_Jhr.

4

3

0 104 I0 $ I0 '_Akm_x DOSE (rm_)

• DEVICE03• 6

107

FIGURE 16, GATE-SOURCE THRESHOLDVOLTAGEVERSUS GAMMA DOSE FORN-CHANNEL IRF'2_ (200 V/30 A); _ DOSE RATE . 33.11TO _..1 kradl#*'.

1.00

0.75

>

0.50

_ o._o

0._

|

_.,_

' FOST.F_I_D1DOSE

15 Mm_

ROOMTFJdP.

ANNEAL

ELEVATED TEMPERATURE ANNEAL; TIMEOF 3-7 HRS AT EACH TEMPERATURE.

O MTM15NS0- #9

3O DAYS

! ,300 325 350 376 400 425

ANNEALING TEk_ERATURE (K)

45O

FIGURE 17. NORMALI?I:O GATE- SOU'RCE THR EST4OLDVOLTAGE FOR N"CHANNELMOSFETs MTM15N_0 _00 V/15 A) AND IRF250 (200 V/3O A) FORPOST-IRRACIATION _GAkIMA DOSE. I S kk "_) AND THERMAL ANNEALINGCONDITIONS. ALL kr=J_UREMENTS MADE AT 300 K.

VGB-0V

DEVICE: 2

,:P[]

[3n7.0.0. -" r'nO"

_0

IO_. 10 A

oOo.O"

...Ooo_O -O

....... 1011 10'2 1013 10"

E_UM NEUTRON FLUENCE (n,'cm ;_

FIGURE 16. DRAIN-9OURCE ON-STATE RESISTANCE FOR ID- 16 A VERSUSEPICADIAUM NEUTRON FLUENCE FOR N-CHANNEL 2SK180 (600 V_ A);NEUTRON R.UX. 7._I0 ? TO 3,76xI0 _ n_zs. GAMMA DOSE - 94 kn_dl

13

Form ApprovedREPORT DOCUMENTATION PAGE OMBNo 0704-0188

public reporting burden for this collection o_ info_a"_'_ Is estimated to average 1 hou¢ _ response, Including the tiT_ for :viewing instructions, searching existing data Ioorces,

gati'mdng and rna]ntalnlng the data needed, artd completing and reviewing ti_ collec6on o_ information. Send oomment= regarding if=is borden estimate or any other ispecl of this

of information, including suggestions for redudng this burden, to Washington I.leadquarleca Services, Diceclorate for informaEon Operations and Reports, 1215 Jefferson

Davis Highway, Suite 1204, Acllngton, VA 22202-4302. and to the Off'me of Management and Budget, PN>erwork Reductk_ Project (0704-0188), Washington, DC 20503.

1. AGENCY USE ONLY (Leave b/an/c) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

Technical Memorandum

" ' _ $. FUNDING'I_UMBERS4. TITLE AND SUBTITLE

Neutron, Gamma Ray and Post-Irradiation Thermal Annealing Effects on

Power Semiconductor Switches

s. AUTHOR(S)

G.E. Schwarze and A.J. Frasca

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)

National Aeronautics and Space Administration

Lewis Research Center

Cleveland, Ohio 44135 - 3191

,, ,,,,,,,,

B. SPONSORING/MONITORING AGENCY NAMES(S) AND ADDRESS(ES)

National Aeronautics and Space Administration

Washington, D.C. 20546- 0001

WU- 590- 13- 3I

8. PERFORMING ORGANIZATIONREPORT NUMBER

E-6577

10. SPONSORING/MONITORING

AGENCY REPORT NUMBER

NASA TM-105248

AIAA-91-3525

11. SUPPLEMENTARY NOTES

Prepared for the Conference on Advanced Space Exploration Initiative Technologies cosponsored by AIAA, NASA, and

OAI, Cleveland, Ohio, September 4-6, 1991. G.E. Schwarze, NASA Lewis Research Center; A.J. Frasca, Wittenberg

University, Springfield, Ohio 45501. Responsible person, G.E. Schwarze, (216) 433 - 6117.

12". DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

Unclassified - Unlimited

Subject Category 33

"13. ABSTRACT (Maximum 200 words)

The effects of neutrons and gamma rays on the electrical and switching characteristics of power semiconductor switches

must be known and understood by the designer of the power conditioning, control, and transmission subsystem of space

nuclear power systems. The SP-100 radiation requirements at 25 m from the nuclear source are a neutron fluence of13 2

10 n/cm and a gamma dose of 0.5 Mrads. Experimental data showing the effects of neutrons and gamma rays on the

performance characteristics of power-type NPN Bipolar Junction transistors (BITs), Metal-Oxide-Semiconductor Field

Effect Transistors (MOSFETs), and Static Induction Transistors (SITs) are given in this paper. These three types of

devices were tested at radiation levels which met or exceeded the SP-100 requirements. For the SP-100 radiation

requirements, the BITs were found to be most sensitive to neutrons, the MOSFETs were most sensitive to gamma rays,

and the SITs were only slightly sensitive to neutrons. Post-irradiation thermal anneals at 300 K and up to 425 K were

done on these devices and the effectiveness of these anneals are also discussed.

14. SUBJECT TERMS

Radiation damage; Radiation effects; Power transistor; Solid state switch;

Semiconductor switch; MOSFET; SIT; Power conditioning

17. SECURITY CLASSIFICATIONOF REPORT

Unclassified

18. SECURITY CLASSIFICATION

OF THIS PAGE

Unclassified

NSN 7540-01-280-5500

19, SECURITY CLASSIFICATION

OF ABSTRACT

Unclassified

15. NUMBER OF PAGES

1416. PRICE CODE

A03

20. LIMITATION OF ABSTRACT

Standard Form 298 (Rev. 2-8g)

Prescribed by ANSI Std. Z3g-18

298.102