Updated Respons of Devices from the FSG and FSP Radiation-Hardened Power MOSFET
Families to 1-MeV Equivalent Neutrons
James E. Gillberg (Member IEEE) Fairchild Semiconductor
Somerville, NJ 08876 USA Donald I. Burton (Member IEEE) Fairchild Semiconductor
125 Crestwood Road
Mountaintop, PA 18707 USA Jeffrey L. Titus (Senior Member IEEE) & Noel Hubbard
NAVSEA Crane
300 Highway 351
Crane, IN 47522-5001 USA
C. Frank Wheatley (Life Fellow IEEE)
tudePrivate consultant
181 Middle Road
Drums, PA 18222 USA
Abstract— This paper updates our 2001 workshop paper and provides neutron test results of six device types reprenting the Star Power Gold, FSG, and the Star Power, FSP, radiation-hardened power MOSFET families manufactured by Fairchild Semiconductor. Both, the FSG and FSP families, employ a stripe-cell topology with devices having rated drain breakdown voltages from 30 volts to 250 volts.
Keywords—neutron; power; MOSFET; displacement damage; star power; star power gold; stripe-cell; rad-hard
I.I NTRODUCTION
Many military and space systems are designed to operate in various types of radiation environments. Potential radiation sources are neutrons and protons. Both of the radiation sources produce ionization and displacement damage. If the system or its individual components are not hardened to withstand the environments, failure could occur.
Displacement damage is usually not a significant concern especially on components that are fabricated with highly doped, shallow junctions, which are ud in most of today's complex integrated circuits. Power MOSFETs incorporate a thick, lightly doped, epitaxial layer, making them more susceptible to displacement damage. Displacement damage can cau significant changes in the power MOSFETs electrical performance, for example, an increa in the on-resistance; breakdown voltage, and/or drain-to-source leakage current.
Many of the power MOSFET technologies have not been characterized to displacement damage effects. Therefore, a ries of tests were designed to characterize displacement damage effects of veral power MOSFET devices from four different radiation-hardened families manufactured by Fairchild Semiconductor.
II.B ACKGROUND
In our 2001 workshop paper, we reported upon twelve device types (from the FR, FS, FSP, and FSG families) that were expod to neutrons at the Pennsylvania State University Breazeale TRIGA Mark III nuclear reactor and electrically characterized [1]. Table I provides a condend summary of tho twelve device types and the levels of neutron exposure.
This paper extends our initial databa by providing neutron test results for six additional device types from the FSG and FSP radiation hardened power MOSFET families. The FSG family of devices is rated for higher LET performance. Data are provided in graphical format and reprent an average of four test samples of each device type. Data for the FSGL033, FSGL035, and FSP033 are reproduced in this paper to provide a compendium of results on the FSG and FSP families.
domTABLE I. D ESCRIPTION OF P REVIOUSLY T ESTED D EVICES Device
Type
Rated
BV DSS
Channel
Type
Neutron
Level Neutron
Level
Neutron
wishyouwerehereLevel
FSGL033a 30
V N 1.1x1014 9.9x1014 2.8x1015
FSGL035a 60
V N 1.1x1014 9.9x1014 2.8x1015
FSPL033a 30
V N 1.1x1014 9.9x1014 2.8x1015
FSL13A0 100
V N 1.0x1013 1.0x1014 1.0x1015母语英语
FSL9130 100
V P 1.0x1013 1.0x1014 1.0x1015
FSL923A0200 V P 1.0x1013 1.0x1014 1.0x1015
2N7397 250
V N 1.0x1013 1.0x1014 1.0x1015
FSL33A0 400
cappuccino什么意思V N 1.1x1012 1.1x1013 3.3x1013
FSL430 500
V N 1.1x1012 1.1x1013 3.3x1013
FRL130 100
V N 1.0x1013 1.0x1014 1.0x1015
FRL9130 100
V P 1.0x1013 1.0x1014 1.0x1015
FRL234 250
V N 1.0x1013 1.0x1014 1.0x1015
a Data from the device types are reproduced in this paper to
provide a compendium of results on the FSG and FSP families. 0-7803-7544-0/02/$17.00 © 2002 IEEE
Neutrons do not exhibit electrical charge (they are electrically neutral) and penetrate deeply into most materials becau they do not electrically interact with the charged particles (protons and electrons). Displacement damage occurs when a neutron strikes an atom and imparts enough energy to displace that atom from its lattice position forming a vacancy. Neutrons can cau significant displacement damage to the atomic lattice and even cau fission reactions, if neutron ab
bruslssorption occurs. Indirect ionization can occur when a neutron collides with the atomic nuclei producing recoils, which then produce ionization along the recoil paths. Ionization during neutron exposure often occurs due to gamma ray contaminated neutron beams.
III. T EST S ETUP
多哈回合Neutron testing was performed upon ninety-six devices reprenting six different device types (16 samples for each device type) in TO-39 type packages. All devices were electrically characterized prior to any neutron exposure. Four of the sixteen samples of each device type were treated as controls. The other twelve samples of each device type were gmented into three groups of four samples, which were then placed into anti-static bags. Tho bags were nt to Pennsylvania State University for neutron exposure. Each bag was expod to a specified level of neutrons. The first bag was expod to a neutron fluence of 1013 2cm n −⋅, the cond to 1014 2cm n −⋅, and the third to 5x1014 2cm n −⋅.
Throughout this paper, neutron fluence is expresd in units of neutrons per square centimeter and is provided as a 1-MeV equivalent for silicon as determined by the dosimetry. Table II provides a description of the additional device types tested and the rated drain-to-source breakdown voltage (BV DSS ), channel type, and neutron exposure levels.
Irradiated devices were electrically characterized at NAVSEA Crane using a Tektronix-370 curve tracer. Measurements consisted of the following DC electrical tests:
• rever gate-to-source leakage current (I GSS ),
• drain-to-source leakage current (I DSS ), • drain-to-source breakdown voltage (BV DSS ),
•
gate threshold voltage (V TH ), and drain-to-source on-state voltage (V DSON ), which was ud to calculate the drain-to-source resistance (R DSON )
TABLE II. D ESCRIPTION OF A DDITIONAL D EVICE T YPES
Device Type Rated BV DSS
Channel
Type Neutron Level Neutron Level Neutron
Level
FSPL130 100 V N 1.1x1013 1.1x1014
5.4x10
14FSPL134 150 V N 1.1x1013 1.1x1014 5.4x1014FSPL230 200 V N 1.1x1013 1.1x1014 5.4x1014FSPL234 250 V N 1.1x1013 1.1x1014 5.4x1014FSGL134 150 V N 1.1x1013 1.1x1014 5.4x1014FSGL234 250 V N
1.1x1013 1.1x1014 5.4x1014
IV. T EST R ESULTS
Electrical measurements of the different test parameters were performed under similar test conditions as specified in the datasheets of each tested device type. Electrical specifications (datasheets) for the device types were available at /products .
A. I GSS
Figs. 1 and 2 graphically display the averaged gate leakage current (I GSS @ V GS = ±30 V). Fig.1 shows devices from the FSG family while Fig. 2 shows devices from the FSP family. Averaged values of I GSS ranged from 2 nA to 5 nA. The manufacturer's specification limit for I GSS is 100 n
A. Devices are specified to meet this specification limit after exposure to a neutron fluence of 1013 2cm n −⋅. The data showed that none of the devices exhibited any notable change in I GSS and continued to operate within the specification limit.
Figure 1. Averaged value of I GSS of tested devices from the FSG family.
Figure 2. Averaged value of I GSS of tested devices from the FSP family.
B. I DSS
Figs. 3 and 4 display the averaged I DSS (V DS at 40% of the device's rated BV DSS ). Fig 3 shows the device respon of the FSG family while Fig. 4 shows the device respon of the FSP family. Figs. 5 and 6 display the averaged I DSS at 80% of the devices rated BV DSS for the same devices. The manufacturer's specification limit for I DSS is 25 µA. Devices are specified to remain below 25 µA even after exposure to a neutron fluence of 1013 2cm n −⋅. The data show the devices exhibited notable changes in I DSS but remained below the 25 µ
A specification limit.
Figure 3. I DSS of devices from the FSG family measured at 40% rated BV DSS
Figure 4. I DSS of devices from the FSP family measured at 40% rated BV DSSanyother
lf
Figure 5. I DSS of devices from the FSG family measured at 80% rated BV DSS
Figs. 3-6 show that I DSS increas with increasing levels of neutron exposure and that the change in I DSS is more vere for tho devices having a higher blocking voltage. The rate of degradation appears to be linear.
C. BV DSS
Figs. 7 and 8 display the measured drain breakdown voltage (BV DSS ) at an I DS of 1 mA. The data show that the breakdown voltage increas slightly with neutron do caud by an increa in bulk resistivity due to displacement damage. Devices remained above their specified BV DSS
rating.
Figure 6. I DSS of devices from the FSP family measured at 80% rated BV DSS
Figure 7. Averaged measured BV DSS
of tested devices from the FSG family
Figure 8. Average measured BV DSS
of tested devices from the FSP family
D. V TH
Figs. 9 and 10 graphically display the average measured threshold voltage (V TH ) of the FSG and FSP devices, respectively, when measured at an I DS of 1 mA. Figs. 11 and 12 display the average measured threshold voltage (V TH ) of the same devices, except the data were measured at an I DS of 5 mA. The obrved threshold voltage shifts are consistent with possible charge build-up in the gate oxide. To examine the threshold respon more cloly, limited ts of
subthreshold curves were recorded.
Fig. 9 Averaged V TH
respon of tested devices from the FSG family
Fig. 10. Averaged V TH
respon of tested devices from the FSP family
Fig. 11. Averaged V TH respon of tested devices from the FSG family
Figs. 13 and 14 graphically show subthreshold respons of the FSGL134 and FSPL134, respectivel
y. Ionization damage incident to neutron exposure can result in oxide traps and interface states being formed. Formation of oxide traps should cau the subthreshold respon to shift to the left along the V GS -axis and formation of interface states should cau the subthreshold slope to decrea. The data suggest that oxide traps dominated causing the threshold voltage to shift to the left. Interface states buildup appears small but
becomes more noticeable at a higher neutron do.
Fig. 12. Averaged V TH
respon of tested devices from the FSP family
Fig. 13. Subthreshold respon of the FSGL134 at a V DS
of 0.1 volts.
Fig. 14. Subthreshold respon of the FSPL134 at a V DS
of 0.1 volts.
E. R DSON
nomatter
R DSON was determined by dividing the measured V DSON by I DS . V DSON was measured using a V GS of 10 and 12 V. Three values of I DS were lected for each gate bias. Figs. 15-18 graphically prent the average R DSON of the FSGL033, FSGL035, FSGL134, and FSG234, respectively. In Fig. 18, R DSON at a neutron fluence of 5.4x1014 was approximately 7 ohms, but obtaining an accurate measurement of V DSON was impeded by thermal dissipation, which impacted the measurement. Resistance increas at higher temperature.
Figure 15. Averaged R DSON
respons of the FSGL033.
Figure 16. Averaged R DSON
respons of the FSGL035
Figure 17. Averaged R DSON respons of the FSGL134
Figs. 19-23 graphically prent the average R DSON of the FSPL033, FSPL130, FSPL134, FSPL230, and FSPL234, respectively. Measurements were taken at gate bias of 10 and 12 volts and I DS was lected bad upon the each device's current rating. In Fig. 23, the measurement of V DSON at a fluence of 5.4x1014 resulted in a R DSON of approximately 6 ohms. Like the FSGL234, the large change in the on-resistance after neutron exposure produced sufficient thermal dissipation to cau the junction temperature to ri during the measurement resulting in an unstable condition.
Figure 18. Averaged R DSON
respon of the FSGL234
Figure 19. Averaged R DSON
respon of the FSPL033
Figure 20. Averaged R DSON
respon of the FSPL130
