Irradiation of silicon power rectifiers with electrons of 12 MeV energy has been carried out. Minority carrier lifetime ~, forward voltage V F, reverse recovered ...
Radiat. Phys. Chem. Vol. 25. Nos. 4,--6, pp. 827--84l. 1985
0146-572-t/85 $3.00 + .00 Pergamon Press Ltd
Printed in Great Britain.
ELECTRON IRRADIATION EFFECT ON MINORITY CARRIER LIFETIME AND OTHER ELECTRICAL CHARACTERISTICS IN SILICON POWER DEVICES Ip.G. Fuochi , P.G. Di Marco , A. Monti 2G.M Bisio E. Di Z i t t i ,3B. Passerini and S. Tenconi I I s t i t u t o FRAE-CNR, Via de' Castagnoli I, 40126 Bologna Italy 2Dipartimento di Ingegneria Bioflsica ed Elettronica, Universita' di Genova, Via Opera Pia 11A, 16145 Genova 3ANSALDO, Divisione Elettronica Industriale, Via N. Lorenzi 8, 16152 Genova, I t a l y
ABSTRACT I r r a d i a t i o n of s i l i c o n power r e c t i f i e r s with electrons of 12 MeV energy has been carried out. Minority c a r r i e r l i f e t i m e ~ , forward voltage VF, reverse recovered charge QnR, reverse recovery time t~n for the diodes, c i r c u i t commutated t u r n - o f f time tq, and on-state voltage V~ for the t h y r i s t o r s are measured as a function of dose. Power diodes and t h y r i s t o r s obtained fromIxI04 Gy. A dose rate e f f e c t on the e l e c t r i c a l characteristics of the devices using pulses of d i f f e r e n t duration is analyzed. Annealing studies are carried out at 150 eC, 200"C and 360"C to assess the s t a b i l i t y of the defects produced by the electron bombardment by monitoring the variation of the e l e c t r i c a l characteri s t i c s of the irradiated devices in the temperature range of i n t e r e s t . DLTS measurements performed on electron irradiated power r e c t i f i e r s have revealed a complex defect pattern. The EI defect level (Ec-0.17 ev) is the principal recombination center that controls l i f e t i m e following room temperature i r r a d i a t i o n . The energy levels and capture cross sections of these i r r a d i a t i o n induced-defects are reported. This study confirms that l i f e t i m e control in s i l i c o n power devices is feasible by high energy electrons. The major advantages of t h i s technique over metallic diffusion or 60Co ~ - i r r a d i a t i o n methods are: better q u a l i t y , lower processing cost and higher device y i e l d s . Annealing a f t e r i r r a d i a t i o n is important to ensure long-term device s t a b i l i t y .
KEYWORDS Electron i r r a d i a t i o n e f f e c t ; l i f e t i m e control; defect level; defect annealing. INTRODUCTION The switching performances of s i l i c o n power r e c t i f i e r s and t h y r i s t o r s are strongly dependent upon the minority c a r r i e r l i f e t i m e . A controlled reduction of t h i s parameter is necessary for high-frequency operation of these de827
828
P . G . Fc:octll et al.
vices. Research in the f i e l d of i n d u s t r i a l production of power semiconductor r e c t i f i e r s is oriented toward technologies f o r l i f e t i m e control which allow to reduce the manufacturing costs. Lifetime control has been f i r s t achieved by high-temperature d i f f u s i o n of deep level i m p u r i t i e s , such as gold or platinum (1,2), which form recombination centers inside the forbidden gap of the semiconductor. In recent years a growing i n t e r e s t has developed in the use of high energy electron or 7 - i r r a d i a t i o n as new technology f o r l i f e t i m e control (3-7). In f a c t , the i n t e r a c t i o n of high energy (MeV) r a d i a t i o n with the s i l i c o n device causes the displacement of s i l i c o n atoms from t h e i r normal l a t t i c e pos i t i o n s , forming vacancies and i n t e r s t i t i a l s . These defects, through migrat i o n inside the c r y s t a l , can form more complex l a t t i c e defects with i m p u r i t i es or other vacancies and act as recombination centers. In t h i s note the e f f e c t s of electron bombardment on power r e c t i f i e r s using a 12 MeV l i n e a r accelerator are analyzed. In p a r t i c u l a r the dependence of the main e l e c t r i c a l c h a r a c t e r i s t i c s , c a r r i e r l i f e t i m e s , r e s i s t i v i t y of the s t a r ting m a t e r i a l , energy l e v e l s of deep centers and t h e i r densities on the i r r a d i a t i o n dose and dose rate are described. Annealing studies are carried out to assess the s t a b i l i t y of the recombination centers introduced by e l e ctron bombardment by monitoring the v a r i a t i o n of the c h a r a c t e r i s t i c s of the i r r a d i a t e d devices in the temperature range of i n t e r e s t . The electron accelerator energy of 12 MeV used f o r l i f e t i m e control has been already tested (5,8) with good r e s u l t s . I t has been noted that the increase of the electron energy up to 12 MeV improves the t r a d e o f f curves (5). This has been correlated with the f a c t that the higher energy p a r t i c l e s cause more atom displacements (9). Anyway there is a l i m i t to the maximum electron energy to be used f o r i r r a d i a t i o n of s i licon devices; t h i s l i m i t is set up by the high energy gamma r a d i a t i o n of bremsstrahlung which, approaching the energy required f o r a (7, n) reaction in s i l i c o n , w i l l induce r a d i o a c t i v i t y in the doped s i l i c o n , ~ e f f e c t that is undesirable.
EXPERIMENTAL SECTION
Device Fabrication The devices were a l l obtained from neutron transmutation or phosphorus-doped float-zone s i l i c o n s l i c e s . The r e s i s t i v i t y of the s t a r t i n g material was 655~cm f o r t h y r i s t o r s and 120~cm f o r diodes. Details of r e c t i f i e r and f a s t - s w i t c h i n g t h y r i s t o r f a b r i c a t i o n process have been given elsewhere ( 6 , 8 ) ; Fig. I summarizes them. The l a s t process step was to a l l o y the s l i c e s onto molybdenum discs 2 mm t h i c k and to provide Ni-Au contacts. A f t e r standard beveling and contour passivation with an organic termosetting polymer reverse recovery character i s t i c s and o p e n - c i r c u i t voltage decay (O.C.V.D.) were measured on a l l samples by a standard laboratory set-up (6,8,11). The samples not Au- or Pt-diffused were then i r r a d i a t e d at room temperature with 12 MeV electrons using an L-band l i n e a r accelerator and then measured again.
I r r a d i a t i o n and Dosimetr~ The i r r a d i a t i o n technique and the technical c h a r a c t e r i s t i c s of the accelerat o r have been described elsewhere (8,10). Studies on the e f f e c t of the i r r a d i a t i o n dosage were carried out using pulses of electrons of I #s duration, while f o r the dose rate experiments pulses of d i f f e r e n t duration were used (50 ns, 200 ns, I Us, 2 # s ) . The u n i f o r m i t y of the i r r a d i a t i o n on the device area was checked by thermoluminescent dosimetry, using LiF c r y s t a l s . Detailed description of dosimetry has been given in a recent paper (8). The modified
Electron Irradiation Effect
on Minority Carrier
829
Fricke dosimeter (10 -2 M Fe(NH4)2S04 in 0.8 N H2SO4 solution saturated with 02 ) was used to c a l i b r a t e a graphite charge c o l l e c t o r . Absorbed doses were in the range 0.25-4.5 x104 Gy (5wIO12to 9xI013 e-/cm 2) f o r f i n i s h e d and alloyed devices.
Device Characteristics High speed t h y r i s t o r s (HST), diode r e c t i f i e r s (DR) and f a s t recovery diodes (FRD), manufactured by Ansaldo S.p.A., Genoa, I t a l y , not gold or platinum d i f f u s e d , with the f o l l o w i n g s t a r t i n g s p e c i f i c a t i o n s were used: a) HST t u r n - o f f time tq= 80 to 160 #s (125~C) on-state voltage VT = 1.4 to 1.45 V (1600 A, 25"C) peak reverse voltage VR~ peak forward o f f - s t a t e voltage
b) DR
I
= 1450 V (75 mA, 125"C)
Venu )
reverse recovery time tRR = 15 to 20 #S (150eC) forward voltage VF = 1.3 V (1800 A, 2 5 ~ ) peak reverse voltage ~mu = 2600 V (50 mA, 175:C) reverse recovered charge Qmn= 2000 to 3200 #C (500 A, 65 A/#s,
150"C)
c) FRD reverse recovery time tnn= 10 to 15 #s (150"C) forward voltage VF= 1.2 to 1.3 V (1200 A, 25"C) peak reverse voltage Vnnu = 2200 to 2600 V (50 mA, 150*C) reverse recovered charge QnR= 1200 to 2000 #C (500 A, 40 A/#s, 150°C) Normally t h y r i s t o r HST and diode FDR are made f a s t switching by gold and platinum d i f f u s i o n r e s p e c t i v e l y . The f i n a l goal of data combinations a f t e r met a l l i c d i f f u s i o n is: HST: tq down to 25 #s, VT ~ 2 . 0 V while maintaining Venm ,VoRu~1400 V. FRD: tnn down to 4 #s, V~ ~ 2 . 3 V, ~nM = 2600 V and Qnn= 190 #C. When using electron i r r a d i a t i o n rather than gold or platinum d i f f u s i o n we a i med at the same f i n a l goal. The diode DR is not manufactured f o r high frequency operation and is used without any m e t a l l i c d i f f u s i o n but with electron bombardment we t r i e d to improve i t s c h a r a c t e r i s t i c s to the f o l l o w i n g s p e c i f i c a t i o n s : DR: tnedown to 4 #s, VF ~ 2 . 0 V, Vnau = 2600 V and Qen ~420 uC, so that i t could be used as a f a s t recovery diode.
Testing The r e c t i f i e r c h a r a c t e r i s t i c s that were tested are: I - minority c a r r i e r l i f e t i m e s (reported in the t e x t as ~tt f o r low-level and ~ t f o r h i g h - l e v e l of i n j e c t i o n ) determined from O.C.V.D. measurements; 2- t h y r i s t o r s on-state (VT) and diodes forward voltage (VF); 3- reverse recovered charge (QmM); 4- t u r n - o f f time tq f o r t h y r i s t o r s and reverse recovery time tnn f o r diodes. For optimum device operation i t is important that the c r i t i c a l parameters VT, VF, Qnn, tq and tRn be small, in order to keep as low as possible the power d i s s i p a t i o n in the on-state and during the commutation of the device and to allow high frequency operations. Moreover i t is important to have a reverse recovery current waveform without abrupt changes so as to prevent high voltage t r a n s i e n t s across the power devices, I t happens that a reduction of r e s u l t s in a decrease of tq, tnn and QNa but is also accompanied by an undesirable increase in Vr and V~. Thus the choice of the appropriate l i f e t i m e becomes an accurate " t r a d e o f f " between opposing requirements.
Deep Level Measurement The deep levels introduced by the electron i r r a d i a t i o n were studied by the
R.PC 25;416-g
830
P.G. FUOCHIef alo
technique of deep level t r a n s i e n t spectroscopy (D.L.T.S.) (12), coupled with a computerized system f o r f a s t capacitance t r a n s i e n t analysis (13). This t e chnique allows one to determine the spectrum of recombination centers, t h e i r concentration and t h e i r cross-section.
RESULTS AND DISCUSSION
Dose Effect on the Device C h a r a c t e r i s t i c s Figure 2 shows the dependence of reciprocal low- and h i g h - l e v e l l i f e t i m e measured by the O.C.V.D. method (current density = 0.4 A/cm2) in diodes as a function of i r r a d i a t i o n dose. The decrease of ~ is in good agreement with the relationship
1/~ = 1 / ~ o + k~D
(1)
which d e f i n e s the r a d i a t i o n damage c o e f f i c i e n t k, ~ o b e i n g any m i n o r i t y c a r r i e r l i f e t i m e of t h e u n i r r a d i a t e d d e v i c e and D the e l e c t r o n dose. Equation ( I ) may be applied to t u r n - o f f time tq f o r t h y r i s t o r s and rearranged to 1/tq= 1 / t q o + ktD (2) and in t h i s form, t h e t h y r i s t o r data h a v e o e e n p l o t t e d in F i g . 3. These p l o t s yield k~: L = 1.1x10-~cm2/s at 25°C k~wL = 7.2x10-~cm2/s at 25°C kt = 3.0x10-Bcm2/s at 125°C Th~se values of k obtained a f t e r 12 MeV i r r a d i a t i o n , performed at room temperature using pulses of I #s length, compare favorably to values reported by other authors ( 5 , 7 , 1 4 ) . I t was found that the electron i r r a d i a t i o n was also very e f f e c t i v e in reducing the stored charge Qmn in the devices; t h i s resulted in a greatly improved t u r n - o f f time tq and reverse recovery time t H f o r the t h y r i s t o r s and the diodes r e s p e c t i v e l y , as shown in Figs. 3 and 4 and in Table I. I t can a l so be seen from these plots and Table that Qnn, tq and tnRcan be c o n t r o l l a bly reduced within the rated values by adjusting the r a d i a t i o n dose. This also makes the DR diode compare very favorably with the FRD diode. Moreover i t can be concluded t h a t in order to obtain a Qnn, tq and tnR approximately equal to t h a t of a good platinum or gold diffused device, a r a d i a t i o n dose of between 0.5 - 1.0xi04 Gy should be used. TABLE I Device
Device Characteristics a f t e r Electron I r r a d i a t i o n
Dose
tq at 125°C
tiR
QH
(104 Gy)
(~s)
(~s)
(~C)
VT at 25°C
(V)
VF at 25°C
HST
0.48 0.72
23 20
260 (a) 176 (a)
1.83 2.23
FRD
0.70 0.90
4.2 (b) 4.0 (b)
242 (b) 228 (b)
1.88 2.01
DR
0.53 1.04
4.4 (a) 3.35(a)
443 (a) 250 (a)
1.55 2.21
(V)
(a) at 125°C (b) at 150°C Test conditions f o r the HST t h y r i s t o r s and the DR diodes are the same as reported in Figs. 3 and 4 r e s p e c t i v e l y , f o r the FRD diodes were: VF at 1200 A, QM~t 500 A, d i / d t =-40 A/#s, reverse voltage -100 V. Increasing the r a d i a t i o n dose f u r t h e r resulted in even better QnR ,tq and tnn values, but i t was found t h a t a reduction of these values is accompanied by
Electron Irradiation
Effecton Minority Carrier
831
an undesirable increase in the on-state voltage Vxfor the t h y r i s t o r s and in the forward voltage drop VF f o r the diodes, Figs. 3 and 4. In high-current high-voltage devices, V~ and VF are c r i t i c a l because they determine the power d i s s i p a t i o n in the device while carrying the forward current or being in the on-state. There is a t r a d e o f f between VF and Qu f o r d i odes and Vr and tq f o r t h y r i s t o r s which depends on the i r r a d i a t i o n c o n d i t i ons. The VF-QNntradeoffs f o r platinum-diffused, ~ - i r r a d i a t e d and f o r room tem perature 12 MeV electron i r r a d i a t e d diodes are plotted in Fig. 5.
Radiation Effect on the S i l i c o n R e s i s t i v i t y n-Type s i l i c o n s l i c e s normally used in the f a b r i c a t i o n of commercial high-power r e c t i f i e r s and t h y r i s t o r s were examined to measure any e f f e c t s of the rad i a t i o n . These are reported in Table 2. As i t can be seen, the deep recombination centers induced by electron i r r a d i a t i o n influence the r e s i s t i v i t y of the s t a r t i n g material increasing i t with the dose.
TABLE 2
I r r a d i a t i o n Effects on n-type S i l i c o n R e s i s t i v i t y
Sample
1 2 3 4 5
Dose (104Gy)
* R e s i s t i v i t y (~-cm) pre-radiation post-radiation
0.70 0.72 0.95 1.35 1.36
61.2 63.7 126.0 126.0 127.0
62.3 64.2 135.0 142.0 145.0
* R e s i s t i v i t y was measured by the four point
probe technique.
Dose Rate Effect on the Device Characteristics and Defect Spectrum A batch of FRD diodes and HST t h y r i s t o r s were i r r a d i a t e d with doses between 0.6 and 2.4 x 104 Gy at room temperature, using pulses of electrons of d i f f e rent duration. The aim of t h i s experiment was to analyze the e l e c t r i c a l behaviour of the devices as well as the defect spectrum and concentrations in the samples as the energy of the electron was varied. In f a c t a d i f f e r e n t
TABLE 3 Power FRD Diode Characteristics at 25°C (Test Cond i t i o n s : Vr at 1200 A, Recovery at 500 A, d i / d t = -40 A/~s, Reverse Voltage -100 V) Sample P.L. Dose ( ~ s ) (104Gy) D1 D2 D3 D4 D5 D6 D7 D8 D9
1 1 1 0.2 0.2 0.2 0.05 0.05 0.05
0.62 0.57 0.69 0.65 0.67 0.82 0.83 0.85 0.84
~wt
(#s) 64 171 133 186 113 134 150 138 171
8.8 9.0 7.4 9.2 9.0 6.5 7.9 7.6 7.9
~tt
(#s) 23 78 66 65 39 50 120 92 144
2.9 3.2 2.5 3.0 3.0 2.8 3.1 3.5 2.9
Qn
(~c)
989 126 1583 134 1343 98 1600 142 1319 132 1288 99 1654 116 1579 119 1595 117
Into
(A) 224 244 244 247 239 245 245 246 245
79 82 69 84 81 69 76 77 76
(V) 1.24 1.31 1.26 1.27 1.28 1.23 1.32 1.28 1.27
1.43 1.74 1.69 1.59 1.69 1.46 1.79 1.65 1.71
Columns ~ and ~ r e f e r to measurements taken before and a f t e r i r r a d i a t i o n respectively.
832
P.G. FL,'OCH!er al, TABLE 4 Power FRD Diode Characteristics at ISO°C. Other Conditions being the same as Table 3
Sample P.L. Dose (us) (IO Gy)
~HL (~s) a
DI
I
0.62
D2
1
0.57
D3 D4 D5
I 0.2 0.2
0.69 0.65 0.67
D6
0.2
D7 D8 D9
0.05 0.05 0.05
~LL (;Js) b
a
S'-S 14T3
QnR (/JC)
b
a
Inn (A)
b
a
VF (V)
b
a
b
1.~6
249 111 243 128 247 124
1,~6 1.14 1.07 1.10 1.14
1610 249
2SO 113
1.03
1914 276 1893 278 1929 276
248 118 247 118 248 117
1.18 1.82 1.13 1.66 1.11 1.66
19--8 24~5 12-32 2~8
24"-5 ITO
177 15.2
682 30.0
1930 307
250 125
149 11.9 200 14.1 108 13.7
544 22.0 695 26,9 404 26.6
1680 240 1918 322 1635 299
0.82
157 1 1 . 4
463 2 o . 8
0.83 0.85 0.84
195 12.6 166 13.9 195 13.1
960 26.4 679 23.8 979 23.1
1.67
1.53 1.56 1.63 1.42
i
Columns ~ and ~ refer to measurements taken before and after i r r a d i a t i o n respectively.
TABLE 5 Power HST Thyristors Characteristics (Test Conditions: VT at 1600 A, Recovery at 600 A, d i / d t = -75 A/~s, Reverse Voltage -lOO V; tq at 900 V, dV/dt= 200 V/~s, 400 A, di/dt= -20 A / p s Reverse Voltage -50 V) Sample
P.L. (~s)
T1 T2 T3 T4 T5
1 1 1 0.2 0.2
Dose (IO4Gy) 0.66 0.67 0.67 0.61 0.58
tqat 125°C (~s)
Qn~t 125°C (~C)
a
b
a
b
92 160 120 116 84
27 23 29 27 27
-
464 452 419 401 431
VT at 25°C (V) a
b
1.41 1.63 1.39 1.59 1.41 1.68 1.42 1.72 1.39 1.66
Columns ~ and ~ refer to measurements taken before and aft e r i r r a d i a t i o n respectively. ~ABLE 6 Energies, Concentrations and Cross-Sections ( G ) in the n-Base Region of FRD Diodes ( I r r . Dose 2.4 x 104 Gy) Lifetime Control
#-irr.
ET(eV)
T (K)
E1=0.17 E2=0.19" E3=0.23 E4=0.27 E5=0.33 E6=0.35 E7=0.41 E8=0.44
84 96 118 141 172 178 217 233
NT (cm"3) at pulse width of 50 ns 200 ns I /Js 2)is
G (cm 2)
0-14 7.6x1012 4.1x10 12 5 5x1012 6 7x1012 I1. " 1011.3~3x 1011 1 12 12 12 6 1.7x1012 . l x l O il ~ . l x l O ~ J 1 . 2 x i 0 2x10 -1 11 .9x101] 4.2x1011 3.1x10 2.4x10 3.7x1011 3.6x10 l1 "3.6xlO11*3.5xlO I1 4.6x1011 5.7x1011 4x10-16 1.6x1012 1.1x10 12 1 l x l O 12 1.1x10124xlO "15 --
*2.9xlO11*11.8x10
~-irr.
E1 : 0 . 2 0 E2:0.22 E3=0.36
5.5x1012 1.6x1011 5. lx1011
Pt doping
E1 =0.20 E2=0.33
8.2x1012 3.2x1011
2.3x
"These defects are estimated since t h e i r presence is not evident after i r r a
Electron [~adiafion E ~
on Minod~ Career
833
d i a t i o n . They w i l l become c l e a r l y v i s i b l e during subsequent annealing.
pulse length implies a d i f f e r e n t d i s t r i b u t i o n in the peak energy of the e l ectrons from the L-band Linear Accelerator (10). Measurements of the elect r i c a l c h a r a c t e r i s t i c s f o r the diodes and f o r the t h y r i s t o r s were consequent l y made before and a f t e r i r r a d i a t i o n f o r d i f f e r e n t pulse lengths. The r e s u l ts are given in Table 3, 4 and 5 r e s p e c t i v e l y . I t can be seen that in a l l cases there are no relevant d i f f e r e n c e s caused by i r r a d i a t i o n dose rate in the measured c h a r a c t e r i s t i c s which depend mainly upon the accumulated dosage. I t is also i n t e r e s t i n g to note that electron i r r a d i a t i o n reduces the spread in the values of the recovery c h a r a c t e r i s t i c s . The DLTS spectra obtained from the diodes i r r a d i a t e d with 2.4 x 104 Gy (3.5 x I013~/cm) are shown in Figs.6, 8, 9. The properties of the defect i n troduced by the electron bombardment are summarized in Table 6. From the data reported here i t appears evident that the defect pattern produced by electrons is f a r more complex than that of Pt-diffused or ~ - i r r a d i ated devices. At least four peaks are c l e a r l y d i s t i n g u i s h a b l e in the DLTS spectra (EI , E3, E6 and ET) and some of them have been already i d e n t i f i e d and reported in the l i t e r a t u r e : a) the EI defect l e v e l is the well-characterized A-center (O-V p a i r ) located at E= - 0.17 eV (14,15) f o r which w~4 have estimated a room temperature electron cross section of the order 10 cm2. I t is present with d i f f e r e n t con centrations, depending on the pulse width; b) the E3 defect l e v e l located at 0.23 eV is coupled with the E7 located at Ec - 0.41 eV. They have been i d e n t i f i e d as the double negative ( [ V - V ] : ) and single negative ( [ V - V ] - ) charge state of the divacancy (16) r e s p e c t i v e l y . The electron capture cross sections at E3 and E7 that we have calculated are (2.01 O.3)x I0-15cm2 and ( 4 . 0 ± O.5)x I0-15cm2 r e s p e c t i v e l y . Their concentrat i o n s seem to be quite stable f o r the 200 ns, I #s and 2 Us pulse lengths, only the 50 ns pulse shows a s i g n i f i c a n t increase. The other l e v e l s reported in Table 6 are not well characterized. The E2 defect level located at 0.19 eV is not evident a f t e r i r r a d i a t i o n , probably i t is hidden by the t a i l of the EI center. I t becomes c l e a r l y v i s i b l e only a f t e r the device has been annealed f o r 60 min at 360°C, when EI has annealed to 35 % of i t s o r i g i n a l value, as i t can be seen from Figs. 7, 8 and 10. The presence of E4 and E5 defect l e v e l s , located at Ec - 0.27 eV and at Ec - 0.33 eV r e s p e c t i v e l y , can only be deduced, but while E4 appears more v i s i b l e during the annealing process, E5 remains hidden by the more populated E6 defect l e v e l . The E6 defect level located at Ec - 0.35 eV becomes more evident as the pulse duration increases and an electron cross section of ( 4 . 0 ± 0 . 6 ) x I0 -IBcm 2 has been calculated f o r i t . The EI defect level f o r i t s concentration and i t s capture cross section f o r electrons appears to be the dominant recombination center that controls minor i t y c a r r i e r l i f e t i m e a f t e r room temperature i r r a d i a t i o n .
Annealing Studies The thermal s t a b i l i t y of the defects created by electron i r r a d i a t i o n , on which the e l e c t r i c a l parameters depend, is of great importance in order to ensure long-term service l i f e of the i r r a d i a t e d devices. Thus the e f f e c t s of the annealing process were examined via periodic forward voltage drop VF , reverse recovery time tse and DLTS measurements during isothermal annealing at 360°C. From previous studies (8) i t is known that annealing of the devices at 150°C does not a f f e c t t h e i r e l e c t r i c a l c h a r a c t e r i s t i c s , even a f t e r 1500 h. At 200°C the diodes did not show any s i g n i f i c a n t v a r i a t i o n s in V~ and t n and t h i s was also true f o r the parameter VT f o r t h y r i s t o r s . The tq howewer showed a marked increase a f t e r 100-150 h of annealing (8). So the annealing of the samples was carried out under argon atmosphere at 360°C, temperature that has been shown to be very e f f e c t i v e f o r k i n e t i c studies. Figures 6-10 show the DLTS spectra of FRD diodes taken r i g h t a f t e r electron i r r a d i a t i o n with 50 ns, I #s
834,
P . O . FUOCHI er a,/.
and 2 #s pulses and a f t e r subsequent annealing f o r 10, 30, 60, 120 and 250 min at 360°C. The EI , E3, and E7 l e v e l s show a s i m i ! a r rate of decay f o r both the 50 ns and I #s pulses while f o r the 2 #s pulse they e x h i b i t a f a s t e r decay. After 250 min a l l the three centers have annealed out almost completel y . After 60 min, as the EI level has annealed to 35 % of i t s o r i g i n a l value, the emergence of the E2 defect l e v e l becomes c l e a r l y v i s i b l e . A f t e r 250 min of annea!ing E2 showed an almost twofo!d increase in defect concentration. The E4 defect level has a s l i g h t increase in defect concentration a f t e r 60 min of annealing f o r both the 50 ns and I #s DLTS spectra, then i t anneals to 40 % of i t s o r i g i n a l value. As previously said the E5 defect l e v e l never becomes c l e a r l y v i s i b l e even during the annealing process, being always hidden by the growing EB l e v e l . The strong decrease in defect concentration of the E3 and E7 l e v e l s a f t e r 60 min of annealing appears to be s t r i c t l y related with the corresponding increase in the E6 defect concentration, that reaches i t s maximum a f t e r that time. This was also observed by Evwaraye et a t . (15). In the same time the peak at 0.35 eV s h i f t e d toward higher energy. The s h i f t represents a temperature change of 8 K. With subsequent annealing the defect concentration decreases and the peak goes back to i t s i n i t i a l energy value. A f u r t h e r defect ! e v e l , E5 , located at Ec - 0.44 eV, 233 K in temperature sca!e, appears in the DLTS spectrum only a f t e r 60 min of annealing and reaches i t s maximum concentration af t e r 120 min. I f we c o r r e l a t e the e l e c t r i c a l c h a r a c t e r i s t i c s with the behaviour of the defects, we must f i r s t point out the d i f f e r e n c i e s between VF and recovery measurements. The recovery performances are more related with l i f e time in the n-base region (where the defect introduced can be analyzed), than VF , which depends also on i n t e r f a c e e f f e c t s of m e t a ! l i z a t i o n . This is probably the reason f o r the rapid i n i t i a l decay of VF in a l l diodes (Fig. 11), while DLTS spectra show only l i t t l e v a r i a t i o n , ! i k e the tRn charac t e r i s t i c s . Figure 11 shows marked changes in the slopes of VF curves, while tRnvaries more uniform!y. The behaviour of tnR and of the EI center during the f i r s t 10 min of annealing is a f u r t h e r confirmation that EI plays the maj o r role in influencing the recovery c h a r a c t e r i s t i c s of the FRD diodes. At longer annealing time i t is d i f f i c u l t to e s t a b l i s h the contributions of the various centers in c o n t r o l l i n g the m i n o r i t y c a r r i e r l i f e t i m e .
CONCLUSIONS The present study confirms that l i f e t i m e control in s i l i c o n power devices is f e a s i b l e by high energy electron i r r a d i a t i o n . I t has been demonstrated that t h i s technique is very e f f e c t i v e f o r t a i l o r i n g the forward voltage drop and the on-state voltage in f a s t recovery diodes and high speed t h y r i s t o r s respec t i v e l y and t h e i r recovery c h a r a c t e r i s t i c s by monitoring the r a d i a t i o n dose. Moreover high energy electron i r r a d i a t i o n has s i g n i f i c a n t processing advantages over Au- or Pt-doping and y - i r r a d i a t i o n : (a) a b e t t e r control of the i n duced defects inside the c r y s t a ! l a t t i c e and hence a greater u n i f o r m i t y in the recovery c h a r a c t e r i s t i c s of the s i ! i c o n r e c t i f i e r s ; (b) the defects produced are stable up to 150°C, which is generally higher than the working temperature of such devices; (c) the f a c t t h a t i r r a d i a t i o n could be performed a f t e r f a b r i c a t i o n of the device is completed makes t h i s process clean and simp!e, and i t gives the p o s s i b i l i t y to r e p a i r and rework them thus bringing to higher device y i e l d s . Annealing treatment a f t e r i r r a d i a t i o n has been proved to be very useful or even necessary in i r r a d i a t e d devices to remove surface damage and r a d i a t i o n induced charge thus ensuring long-term device s t a b i l i t y . By using pulses of d i f f e r e n t duration i t has been shown t h a t there are no relevant differences in the DLTS spectra and in the main e l e c t r i c a ! parameters of samples studied here due to d i f f e r e n t dose r a t e . Any observable v a r i a t i o n depends main!y upon the t o t a l dose. DLTS and VF ,tnR measurements before and during annealing of the FRD diodes lead to the conclusion t h a t the dominant recombination center which i n f l u e n ces the recovery c h a r a c t e r i s t i c s immediate!y a f t e r i r r a d i a t i o n is the defect
Electron Irradiation Effect on Minority, Carrier
835
leve! EI (E c - 0.17 eV). Other types of defects instead seem to control VF . Uncertainty s t i l l e x i s t s concerning the i d e n t i f i c a t i o n of the dominant recombination center as the annealing proceeds. More work needs to be done in t h i s f i e l d to establish the effectiveness or c o n t r i b u t i o n s of the various centers in c o n t r o l l i n g the device c h a r a c t e r i s t i c s .
REFERENCES I. J. M. F a i r f i e l d and B. V. Gokhale, Solid State Electron., 1966, ~, 905 2. M. D. M i l l e r , H. Schade and C. J. Nuese, I n t . Electron Devices Mtg. 1975, 180, paper 8.7 3. P. Rai-Choudhury, J. Bartko and J. E. Johnson, IEEE Trans. Electron Devices, 1976, 2__33, 814. 4. B. J. Baliga and E. Sun, IEEE Trans. Electron Devices, 1977, 24, 685. 5. R. O. Carlson, Y. S. Sun and H. B. A s s a l i t , IEEE Trans. Electron Devices,
1977, 24, 1103. 6. G. M. Bisio, M. I c a r d i , E. Di Z i t t i , M. Portesine and S. Tenconi, IEEEPESC Conference Records, 1981, 284. 7. W. R. Fahrner, D. Braunig and E. Borchert, Phys. Stat. S o l . ( a ) , 1982, 72, 79. 8. L. Barberis, M. I c a r d i , M. Portesine, S. Tenconi, P. G. Di Marco, A. Mart e l l i and P. G. Fuochi, Radiat. Phys. Chem., in press. 9. J. W. Corbett, G. D. Watkins, Phys. Rev., 1965, !38, A 555. 10. A. Hutton, Quad. d e l l ' Area Ric. d e l l ' Emilia-Romagna, 1974, ~, 57; A. Hutton, G. Roffi and A. M a r t e l l i , ibidem, 1974, ~, 67. 11. G. M. Bisio, M. I c a r d i , M. Portesine and S. Tenconi, ESSDERC '80 Europhysics Conference Abstracts, vo1. 4N, pp. 49-50,1988. 12. D. V. Lang, J. Appl. Phys., 1974, 4 j , 3023. 13. G. M. Bisio, E. Di Z i t t i , G. D o n z e l l i n i , G. Parodi and E. Zerbini, 85th Annual Meeting of A . E . I . , Riva del Garda, October 14-18, 1984, paper 16.1 and references t h e r e i n . 14. S. D. Brotheron and P. Bradley, Semiconductor S i l i c o n , 1981, 779. 15. A. O. Evwaraye and B. J. Baliga, J. E1ectrochem. Soc., 1977, 124, 913. 16. A. O. Evwaraye and E. Sun, J. Appl. Phys., 1976, 47, 3776.
P.G. FUOCHIeral.
836
float-zone neutron transmutation or phosphorus doped silicon 120 ~Lcm end 6SJ~cm .... Gallium diffusion
]
Phosphorus diffusion
t
, ,,,,,
I Auor Pt diffusion
I
l
Alloying
I
[ Fig.
I.
!i
] I,
I
,".,,n,,.. ]
Testing
!
I I'"' irr/diitiin 1 i
Testi n o
]
Power r e c t i f i e r s f a b r i c a t i o n steps.
1
Dose (104 Gy) 2
3
4
./
60
50
."40 i
%
~30 20 /
10
/ Dose (1013elect rons/cm2)
Fig. 2. Variation of c a r r i e r l i f e t i m e ~ L L ( i ) and TNL(e) at 25°C with electron dose.
Electron Irradiation Effect on Minority Carrier
O.5
4f
R
Dose (104Gy) %O
,
1.5
,
.
83"/
I°L
3
g/ 5,0 4200
%0%~%%0
~ ~" 0
I
J Dose (lOt3electron=/¢m 2)
Fig. 3. Electron i r r a d i a t e d t h y r i s t o r s : dose dependence of ~(o)at 1600 A, I/tq(m) and QnM(o). Test conditions: 125°C, tq atdV/dt = 200 V/#s, 900 V reapplied voltage, 400 A f o r ward current, current t u r n - o f f rate d i / d t of 20 A/#s and reverse vottage -50 V; Qnn at 400 A forward current, current t u r n - o f f rate d i / d t of 50 A/#s and reverse voltage -100 V.
I
Dose (104 Gy) 2 3
4 6OO
500
4OO
-2
300 Qg 0 2OO
I00
4
6
8
Dose (1013electrons/cm 2)
Fig. 4. Electron i r r a d i a t e d DR diodes: dose dependence, V~ (o) at 1800 A, tnn(m) and Qan(o). Test conditions: 125°C, recovery at 800 A peak forward current, current t u r n - o f f rate d i / d t of 65 A/Us and reverse voltage - 100 V.
838
P.G. FL'OCHt eta/.
i
i --opt ~ )o
o
~
mP,o
o ~ (~c)
mio
4d,o
Fig. 5. VF-Qntradeoff curves in DR diodes Pt-diffused ( o ) , ~ - i r r a d i a t e d (m) and electron irradiated (Q); VF at 240 A/cmLand Qnn at 100 A/cm2 peak forward current, current t u r n - o f f rate diF/dt of 8 A/cm2#s and reverse voltage -100 V, junction temperature 125°C.
o-tlI 0.61
°-,°tt I 0.30 -~
o.,oli i ./,.::t .... Ark olll %,,.,,#, ,.-%,._ ~.,.,.~'-;.'.,~,~,4"J" 75
11"7
1~
201
243
285
Tefltp4wature (K)
Fig. 6. DLTS spectra of an FRD diode, irradiated with pulses of 50 ns width, (2.4xI04 Gy t o t a l dose), under various annealing conditions at 360 C. Annealing time: ( ) O, ( - - - ) 10 and ( - - - ) 30 min.
Electron Irradiation Effect on Minority Carrier
839
0"231
~ 0.20] v --
0.17-
"3
0.13"
,~
0.10-
I
/
I O.07- I I
J 0.03- y 0 75
11'7
159
285
2~,3
201
Temperature (K)
Fig. 7. Conditions are as in Fig. 6. Annealing time: ( ) 60, F - - ) 120 and (-.--) 250 min. 0.74 0.66 0,57. 0,49
~
0.41
"3 0.33"
'I'!li' I
"~ 0.24. 0.16' ~\
t-~
.-.-.
0.08. 0 75
117
159
201
243
264
Temperature (K)
Fig. 8. DLTS spectra of an FRD diode irradiated with pulses of I #s width, 2.4xi04 Gy t o t a l dose, under various annealing conditions at 360°C. Annealing time: ( ) O, ~ - - ) 10, (-.--) 30 and (--,.--) 60 min.
P . O . FUOCHI etal,.
840
0.77 0.68 0.60 u,.
'•
0.51-
0.43-
0"17t ~
,tl o
75
11~7
159
201
2i3
" --285
Tempereture (K)
Fig. 9. DLTS spectra of an FRD diode i r r a d i a t e d wlth pulses of 2 ~s width, 2.4x104Gy t o t a l dose, at various annealing times at 360°C: ( ) O, (----) 10 and (-.--) 30 mln. 0.15
0.13' 0.12 - -
U. ¢& ¢
0.10 0.08
"3 0.07
U~ k.J
0.05 0.03 t
',i -,~. . . . .
0.02~
ot
75
-~-
117
./J
\ '\, .-~.~._ .P'~\ \ \
.,...,.-/
--,--~
159
251
.\,,
243
285
Temperature (K) Fig.
10.
60, ~ - - )
Conditions
are
as i n
Fig.
120 and (--.--) 250 min.
9.
Annealing
time:
(
Electron Irradiation Effect on Minority Carrier
841
i
i
o . s~~ - '~ -
, vF at 2s "c
T
....
tRRat 1S0 "C
L
,,
T
Time
(hours)
Fig. 11. Evolution during annealing at 360°C of V~ and tnn f o r FRD diodes i r r a d i a t e d with electron pulses of d i f f e rent duration: (®) 50 ns, (B) 200 ns, (¢~) 1~s, (e) 2 ~s. The parameters are normalized to the expected value V~ = 2.3 V and taa= 4 ~s. Test conditions: V~ at 1200 A, tam at ITM = 350 A, d i ~ / d t = -80 A/~S, VR = -50 V. I r r a d i a t i o n dose:2.4x104 Gy.
