ASTM F996-1998(2003) Standard Test Method for Separating an Ionizing Radiation-Induced MOSFET Threshold Voltage Shift Into Components Due to Oxide Trapped Holes and Interface Statee.pdf

1、Designation: F 996 98 (Reapproved 2003)Standard Test Method forSeparating an Ionizing Radiation-Induced MOSFETThreshold Voltage Shift Into Components Due to OxideTrapped Holes and Interface States Using the SubthresholdCurrentVoltage Characteristics1This standard is issued under the fixed designatio

2、n F 996; the number immediately following the designation indicates the year oforiginal adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. Asuperscript epsilon (e) indicates an editorial change since the last revision or re

3、approval.1. Scope1.1 This test method covers the use of the subthresholdcharge separation technique for analysis of ionizing radiationdegradation of a gate dielectric in a metal-oxide-semiconducter-field-effect transistor (MOSFET) and an isola-tion dielectic in a parasitic MOSFET.2, 3, 4The subthres

4、holdtechnique is used to separate the ionizing radiation-inducedinversion voltage shift, DVINVinto voltage shifts due to oxidetrapped charge,D Votand interface traps,D Vit. This techniqueuses the pre- and post-irradiation drain to source current versusgate voltage characteristics in the MOSFET subth

5、resholdregion.1.2 Procedures are given for measuring the MOSFET sub-threshold current-voltage characteristics and for the calculationof results.1.3 The application of this test method requires the MOS-FET to have a substrate (body) contact.1.4 Both pre- and post-irradiation MOSFET subthresholdsource

6、 or drain curves must follow an exponential dependenceon gate voltage for a minimum of two decades of current.1.5 The values given in SI units are to be regarded asstandard. No other units of measurement are included in thistest method.1.6 This standard does not purport to address all of thesafety c

7、oncerns, if any, associated with its use. It is theresponsibility of the user of this standard to establish appro-priate safety and health practices and determine the applica-bility of regulatory limitations prior to use.2. Terminology2.1 Definitions of Terms Specific to This Standard:2.1.1 anneal c

8、onditionsthe bias and temperature of theMOSFET in the time period between irradiation and measure-ment.2.1.2 doping concentration, NN-orP-type doping, is theconcentration of the MOSFET channel region adjacent to theoxide/silicon interface.2.1.3 inversion current, IINVthe MOSFET channel currentat a g

9、ate-source voltage equal to the inversion voltage.2.1.4 inversion voltage, VINVthe gate-source voltage cor-responding to a surface potential of 2fB.2.1.5 irradiation biasesthe biases on the gate, drain,source, and substrate of the MOSFET during irradiation.2.1.6 midgap current, IMGthe MOSFET channel

10、 currentat a gate-source voltage equal to the midgap voltage.2.1.7 midgap voltage, VMGthe gate-source voltage corre-sponding to a surface potential of fB.2.1.8 oxide thickness, toxthe thickness of the oxide of theMOSFET under test.2.1.9 potential, fBthe potential difference between theFermi level an

11、d the intrinsic Fermi level.2.1.10 subthreshold swingthe change in the gate-sourcevoltage per change in the log source or drain current of theMOSFET channel current below the inversion current. Thevalue of the subthreshold swing is expressed in V/decade (ofcurrent).2.1.11 surface potential, fsthe po

12、tential at the MOSFETsemiconductor surface measured with respect to the intrinsicFermi level.3. Summary of Test Method3.1 The subthreshold charge separation technique is basedon standard MOSFET subthreshold current-voltage character-istics. The subthreshold drain or source current at a fixed drainto

13、 source voltage, VDS, is measured as a function of gatevoltage from the leakage current (or limiting resolution of the1This test method is under the jurisdiction of ASTM Committee F01 onElectronics and is the direct responsibility of Subcommittee F01.11 on Quality andHardness Assurance.Current editi

14、on approved May 10, 1998. Published September 1998. Originallypublished as F 996 91. Last previous edition F 996 92.2For formulation of subthreshold charge separation technique see McWhorter, P.J. and P. S. Winokur, “Simple Technique for Separating the Effects of Interface Trapsand Trapped Oxide Cha

15、rge in MOS Transistors,” Applied Physics Letters, Vol 48,1986, pp. 133135.3DNA-TR-89-157, Subthreshold Technique for Fixed and Interface TrappedCharge Separation in Irradiated MOSFETs, available from National TechnicalInformation Service, 5285 Port Royal Rd., Springfield, VA 22161.4Saks, N. S., and

16、Anacona, M. G., “Generation of Interface States by IonizingRadiation at 80K Measured by Charge Pumping and Subthreshold Slope Tech-niques,” IEEE Transactions on Nuclear Science, Vol NS34, No. 6, 1987, pp.13481354.1Copyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, P

17、A 19428-2959, United States.measurement apparatus) through inversion. The drain currentand gate voltage are related by IDa 10VG. When plotted as logIDversus VG, the linear I-V characteristic can be extrapolatedto a calculated midgap current, IMG. By comparing the pre- andpost-irradiation characteris

18、tics, the midgap voltage shift,D VMGcan be determined. The value ofD VMGis equal to DVot, whichis the voltage shift due to oxide trapped charge. The differencebetween the inversion voltage shift, DVINV, andD VMGis equalto DVit, which is the voltage shift due to interface traps. Thisprocedure is show

19、n in Fig. 1 for a p-channel MOSFET.4. Significance and Use4.1 The electrical properties of gate and field oxides arealtered by ionizing radiation. The time dependent and dose rateeffects of the ionizing radiation can be determined by compar-ing pre- and post-irradiation voltage shifts, DVotandD Vit.

20、 Thistest method provides a means for evaluation of the ionizingradiation response of MOSFETs and isolation parasitic MOS-FETs.4.2 The measured voltage shifts, DVotandD Vit, can providea measure of the effectiveness of processing variations on theionizing radiation response.4.3 This technique can be

21、 used to monitor the total-doseresponse of a process technology.5. Interferences5.1 Temperature EffectsThe subthreshold drain currentvaries as the exponential of qfB/kT, and other terms whichvary as a function of temperature. Therefore, the temperatureof the measurement should be controlled to withi

22、n 62C, sincethe technique requires a comparison of pre- and post-irradiation data. At cryogenic temperatures, this test methodmay give misleading results.45.2 Floating Body (Kink) EffectsFloating body effectsoccur in MOSFETs without body (substrate) ties. This testmethod should not be applied to a M

23、OSFET without asubstrate or substrate/source contact.5.3 Short Channel EffectsTo minimize drain voltage de-pendence on the subthreshold curve, a small drain measure-ment voltage is recommended but not necessary.5.4 Leakage CurrentBecause the MOSFET midgap cur-rent is below the capabilities of practi

24、cal current-voltagemeasurement instrumentation, extrapolation of the subthresh-old swing is required for the determination of a MOSFETmidgap voltage. Extrapolation of ideal linear MOSFET sub-threshold current-voltage characteristics is unambiguous, be-cause of the constant subthreshold swing. An exa

25、mple of nearideal subthreshold characteristics is given in Fig. 2, where thesubthreshold swing is relatively constant between 1011and106A. Nonideal subthreshold characteristics, that are aberra-tions from the theoretical linear subthreshold swing, cancomplicate the subthreshold swing extrapolation t

26、o the midgapvoltage. For subthreshold characteristics that have multiplesubthreshold swings, the value of the midgap voltage would bedependent on the values of the subthreshold current fromwhich the extrapolation is made. Nonideal subthreshold char-acteristics are caused by MOSFET leakage currents t

27、hat can beeither independent of, or a function of, gate-source voltage.5.4.1 Junction Leakage CurrentThis leakage current isfrom the drain to the substrate and is independent of gate-source voltage. Junction leakage current masks the actualMOSFET channel subthreshold current below the leakagecurrent

28、 level. Junction leakage current is easily distinguishedfrom the channel subthreshold current as is shown in Fig. 2 bythe flat section of the drain current, ID, below 1011A. Thisfigure also shows the advantage of using the source current, IS,for extrapolation. The source current is not affected by j

29、unctionleakage so that a measure of the MOSFET channel current isobtained to the instrumentation noise level. However, if there isnot a separate source and substrate contact (for example, powerMOSFETs), the drain current must be used. Only the part of thesubthreshold curve above the junction leakage

30、 or instrumenta-tion noise level should be used for extrapolation. A minimumof two decades of source or drain current above the leakage ornoise is required for application of this test method.5.4.2 Gate LeakageGate leakage can be any combinationof leakage from the gate to source, drain, or substrate

31、.Typically this leakage will be a function of the gate-sourcevoltage. If gate leakage is greater than 1.0 A for anygate-source voltage, the test method should not be applied.Gate leakages less than 1.0 A can still cause nonidealFIG. 1 Determination of Radiation Induced Voltage Shift forp-Channel MOS

32、FETFIG. 2 Near Ideal Subthreshold Characteristics from ann-Channel TransistorF 996 98 (2003)2subthreshold characteristics. The minimum value of the sub-threshold source or drain current used for extrapolation to themidgap voltage must be above any changes in the subthresholdswing that can be attribu

33、ted to gate leakage. Plotting the log ofthe gate leakage along with log source and drain current on thesame graph, will aid in the determination of gate leakageeffects on the drain and source subthreshold swing.5.4.3 Edge Leakage CurrentMost microcircuit MOSFETsuse an open geometry layout so that io

34、nizing radiation induceddrain to source leakage can occur in n-channel devices outsideof the intentional MOSFET channel. The effect of this edgeleakage on the subthreshold swing is dependent on the aspectratios and threshold voltages of the intentional and parasiticMOSFETs. The aspect ratio of the p

35、arasitic MOSFET wouldusually be much smaller than a standard width MOSFETlayout. Thus, when the MOSFET channel is in strong inver-sion, the channel current will typically dominate. However, asthe channel current is reduced, edge leakage can go from aminimal fraction to dominating the measured drain

36、or sourcecurrent if the parasitic MOSFET inversion voltage is less thanthe intentional MOSFET. This effect can be observed in themeasured subthreshold characteristics as a deviation from theideal linear subthreshold curve that is a function of thegate-source voltage. Examples of parasitic MOSFET ind

37、uceddeviations from the ideal linear subthreshold swing are given inFig. 3 and Fig. 4. In Fig. 3, the subthreshold swing changesfrom the initial swing near inversion to a much larger mV/decade swing. In Fig. 4, a more pronounced deviation isshown. The section of the subthreshold curve that should be

38、used for extrapolation to the midgap voltage is shown in bothfigures. The upper section of the subthreshold curve above thelower current level deviations was used. Any lower currentchange in the subthreshold swing from the initial subthresholdswing below strong inversion should be considered a paras

39、iticMOSFET induced deviation. Only the part of the subthresholdcurve above this deviation should be used for extrapolation asis shown in Fig. 3 and Fig. 4. Some n-channel MOSFETs mayhave post-irradiation edge leakage sufficiently large to preventany observation of a subthreshold swing. The subthresh

40、oldcharge separation technique cannot be applied to thesesamples. A minimum of two decades of source or drain currentabove any subthreshold swing deviation is required for appli-cation of this test method. Open and closed (annular) geometrylayouts can be used to separate edge leakage current from th

41、eMOSFET channel current.5.4.4 Backchannel and Sidewall Leakage in a SOIMOSFETBackchannel leakage arises from a parasitic MOS-FET located at the interface between the epitaxial silicon andthe insulator. Sidewall leakages arise from the parasitic MOS-FET formed at the edges of the intentional MOSFET.

42、Theseparasitics distort the subthreshold curve in the same manner asdescribed in 5.4.3.6. Apparatus6.1 To measure the subthreshold current-voltage character-istics of a MOSFET, the instrumentation required consists of,as a minimum, two voltage sources and four ammeters.6.2 The power supplies are use

43、d to apply voltage to the gateand drain of the MOSFET. The ammeters are used to measurethe gate, drain, source, and substrate currents.6.3 For MOSFETs that have a common source/substratecontact, only three ammeters are required.6.4 For a typical digital microcircuit MOSFET the voltagesources and amm

44、eters should meet the specification given inTable 1.6.5 For a power, parasitic field oxide, or high voltage linearMOSFET, the maximum voltage requirement for the gate-source power supply can be substantially larger. Field oxidefield effect transistor (FET)s may have pre-irradiation thresholdvoltages

45、 of several hundred volts. The gate-source powersupply is required to be such that the MOSFET drain-to-sourcesubthreshold current can be measured from leakage into stronginversion. The resolution of a gate-source power supply mustbe at least 0.5 % of the maximum gate-source voltage, for theMOSFET su

46、bthreshold current-voltage measurement.6.6 The test fixture, containing the MOSFET under test(DUT), and the cabling connecting the test instrumentation,FIG. 3 Example of a Parasitic MOSFET Induced Deviation Fromthe Ideal Linear Subthreshold SwingFIG. 4 Example of a Parasitic MOSFET Induced Deviation

47、 fromthe Ideal Linear Subthreshold SwingTABLE 1 Minimal Instrumentation SpecificationsDrain-source power supply 610 V, 0.01 V resolutionGate-source power supply 610 V, 0.005 V resolutionAmmeters 610 mA, 10 pA resolutionF 996 98 (2003)3must be designed for low current measurements specified inTable 1

48、. Probe stations can be used for MOSFETs in waferform as long as a current resolution of 100 pA can bemaintained.6.7 The test fixture may be the same or different from thefixture(s) used during irradiation and anneal (irradiation andanneal fixture(s). Any handling of the DUT must be performedso as t

49、o minimize static discharge or other voltage transientsthat may damage (alter current-voltage characteristics) theDUT.6.8 For control of measurement and data recording, thesubthreshold current-voltage characteristics can be measuredwith a programmable tester having the proper current andvoltage capabilities, or with computer control of independentpower supplies and ammeters. To minimize apparatus setup,programmable testers are usually selected.7. Test Specimens and Sampling7.1 Test Specimens This test method can be applied toMOSFETs in wafer form usin