IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 41, NO.

12, DECEMBER 1994

2357

On the Universality of Inversion Layer Mobility in Si MOSFET's: Part I-Effects of Substrate Impurity Concentration
Shin-ichi Takagi, Member, IEEE, Akira Toriumi, Masao Iwase, and Hiroyuki Tango
Abstract-This paper reports the studies of the inversion layer mobility in n- and p-channel Si MOSFET's with a wide range of substrate impurity concentrations to 10l8 cm-"). The validity and limitations of the universal relationship between the inversion layer mobility and the effective normal field (E,E) are examined. It is found that the universality of both the electron and hole dependences of mobilities does hold up to l O I R cm-3. The the universal curves are observed to differ between electrons and holes, particularly at lower temperatures. This result means a different influence of surface roughness scattering on the electron and hole transports. On substrates with higher impurity concentrations, the electron and hole mobilities significantly deviate from the universal curves at lower surface carrier concentrations because of Coulomb scattering by the substrate impurity. Also, the deviation caused by the charged centers at the Si/SiO2 interface is observed in the mobility of MOSFET's degraded by Fowler-Nordheim electron injection.

I. INTRODUCTION
HE inversion layer mobility in Si MOSFET's has been a very important physical quantity as a parameter to describe the drain current and a probe to study the electric properties of a two-dimensional carrier system. Therefore, much study [ 11 since the 1960's has revealed dominant scattering mechanisms determining the mobility. However, a comprehensive understanding of the inversion layer mobility, which includes the quantitative description near room temperature, the effect of substrate impurity, the difference between the electron mobility and the hole mobility and the effect of surface orientation, is still insufficient. On the other hand, it has already been reported that the electron and hole mobilities in the inversion layer on a (100) surface follow the universal curves at room temperature independent of the substrate impurity concentration or the substrate bias when plotted as a function of effective normal fields, E,R [2]-[9]. E,ff is defined by the following equations. where y is the elementary charge, E S ~is the permitivity of si, Ndpl is the surface concentration of the depletion charge,
Manuscript received June 3, 1994. The review of this paper was arranged by Associate Editor K. Tada. S. Takagi is with the Solid State Electronics Laboratory, Stanford University, Stanford, CA 94305 USA. A. Toriumi, M. Iwase, and H. Tango are with the ULSI Research Laboratories, Research & Development Center, Toshiba Corporation, 1 Kamukai Toshiba-cho, Saiwai-Ku, Kawasaki 210, Japan. IEEE Log Number 9405895.

N , is the surface inversion carrier concentration. Here, 7 is a key parameter in defining E,R and it has been reported that, in order to provide the universal relationship, the value of q should taken to be 1/2 for the electron mobility [2] and 1/3 for the hole mobility [3]. This relationship has been often utilized as a precise mobility model in device simulators [lo]-[ 121. In spite of its usefulness, however, the origins of the universality, the value of 71 and the effective field dependence of the universal curves have not been fully clarified yet. The aim of this paper (Part I) and the companion paper (Part 11) is to study the applicability and the physical meaning of the universal relationship. While the Part I demonstrates the validity and the limitation of the universality of the electron and hole mobilities on (100) surface experimentally, the Part I1 examines the physical meanings of E,ff and 7 , based on the new experimental findings regarding the surface orientation dependence. In this paper we concentrate on the effect of substrate impurity on the electron and hole mobilities. The motivations to study the effect of substrate impurity are twofold. One is to examine the validity of the universal relationship over a wide range of substrate impurity concentration. The substrate impurity concentration changes Ndpl in (1) and the resultant E,R, independent of N,. Therefore, the value of 71 that offers the universal relationship can be determined experimentally by comparing the mobilities on the different impurity concentrations. However, the systematic study of the universality over a wide range of substrate impurity concentrations has not been done sufficiently. Such an extensive verification of the universality allows us to characterize the E,ff dependence of the universal curve quantitatively. The second motivation is to examine the influence of substrate impurity scattering. Although Coulomb scattering by substrate impurity is considered to degrade the mobility on higher impurity concentration substrates, the quantitative characterization has been still lacking in spite of the practical importance in the scaled MOSFET's. From the above motivations, we investigate the electron and hole mobilities with the substrate impurity concentrations of 1015 to 10" cm-3 systematically in terms of the universal relationship.
11. SAMPLE PREPARATION AND MEASUREMENTS

N-channel and p-channel MOSFET's used in this paper were fabricated on (100) Si wafers. The substrate impurity concentration was varied from to 10'' cmP3, using

0018-9383/94$04.00 0 1994 IEEE

___~

2358

IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 41, NO. 12, DECEMBER 1994

boron or phosphorus ion implantation, followed by a long and high temperature annealing (1 190C 60 min). This annealing allowed the impurity profile to be considerably flat to a 3 pm depth from the Si surface. All the devices were of the surface channel type. The gate oxide was grown to a thickness of 25 nm in dry oxygen at 900C. The gate materials were n+ and p+-poly Si for n- and p-channel MOSFET's, respectively. The channel length, L , and the channel width, W , of the measured devices were 200 pm and 100 pm, respectively. The effective mobility in the inversion layer, pLeff, was determined from the drain conductance gd in the linear region.
m 0
-

10'

NAc c i 3 I

I
.9d

1
10'

ELECTRON
= +
I

was measured at the drain voltage, V d of 50 or 10 mV. The ( V , ) , was determined directly surface carrier concentration, N,? through gate-channel capacitance Cgc( V,), measurement [ 141, [151.

1
I

0.1
EFFECTIVE FIELD [ MV/cm

1.0

YNS(V,) =

.I_,C,c(Vg)~~Vg.
1;

(3)

Fig. 1. Electron mobility in inversion layer at 300 K and 77 K versus effective field E,=, as a parameter of substrate acceptor concentration, S I . Here, Ec,fris defined by E,n. = q . r/ . .V*)/ES; with of 1/2.

The measurement frequency was selected to be as low as 0.4-1.0 kHz in order to avoid the influence of the resistive component of the channel [15]. The value of Eefi was determined from (l), (4), and (5).
(4)

Here, 4~ is the bulk Fermi energy, Nsub is the substrate impurity concentration, k~ is the Boltzman constant, and n, is the intrinsic carrier concentration. Nsut, was determined from the minimum capacitance in the high frequency (100 kHz) C-V curves of the MOS diodes. We have confirmed, using the process simulator (TOPAZ) [16], that the variation in the impurity concentration within the depletion layer is less than cmP3. -around 30% for the doping of boron more than The main issue regarding the universal relationship is to determine the value of 77 in (1). This can be done experimentally from the experimental mobilities with the different substrate impurity concentrations. When the universal relationship holds and the value of 11 is chosen appropriately, the mobility should be described as the single universal curve against E,n, independent of the substrate impurity concentration. If 77 is incorrect, the mobilities with the different substrate impurity concentrations are to have different values for a same E,tf value, because 71 changes the weights of N , and NdPl.

are taken to be 1/2 and 1/3, respectively. As seen in Fig. 1, the electron mobilities at 300 K are represented by the universal curve as a function of E,R in the range of 0.05 to 1.5 MV/cm independent of the substrate impurity concentrations by choosing the value of -77 to be 1/2. Similarly, the hole mobilities are also represented by the universal curve as a function of E,ff in the range of 0.05 to 1 MV/cm by choosing the value of 77 to be 1/3, as observed in Fig. 2 . These results confirm us that universal relationships of the electron and hole mobilities on (100) surface do hold up to the substrate impurity concentration of 1 0ls cm-3. It is confirmed simultaneously that the values of 71 for the electron and hole mobilities are 1/2 and the substrate and 1/3, respectively, for a wide range of Eeff impurity concentration. These values are in agreement with those reported previously [ 2 ] , [SI. Moreover, the electron and hole mobilities at 77 K also have the universality in high Eee, which was defined by the same value of 17 as at 300 K. It is, however, observed that the electron and hole mobilities with higher substrate concentrations exhibit a significant deviation from the universal curves near the threshold voltage. This issue will be discussed in Section 111-C in detail.

B. Effectitve Field Dependences of Universal Curves for Electrons and Holes

We have obtained the universal curves that hold for the impurity concentrations of 10l5 to 10" cm-3 for both the electron and hole mobilities in Section 111-A. However, the origin of the Ecff dependences have not been fully understood yet. Especially, the E,ff dependence of the hole mobility has been studied little so far. In this section we study the E,ff dependence of the universal curve with emphasis on the difference between the electron mobility and the hole mobility. From Fig. 1, the characteristics of the E,ff dependence of the electron mobility are summarized as follows. 1) At 300 K, the mobility is proportional to E : , at E,tf lower than 0.5 MV/cm over a one order of magnitude Eeffrange. 2 ) At

RESULTS 111. EXPERIMENTAL

A . Unitiersal Relationship o f Electron and Hole Mobilities

Figs. I and 2 show the Eeffdependences of the electron and hole mobilities in the inversion layer, respectively, at 300 K and 77 K. The parameter is the substrate acceptor concentration, N.4, or the substrate donor concentration, N p . V d in the measurement of gel was taken to be IO mV. Here, the values of rI for the electron and hole mobilities

TAKAGI et al.: ON THE UNIVERSALITY OF MOBILITY: PART I

2359

I
0

5000 I

ELECTRON

: 2000
5
U

I000
. * -

..- .....-..
. a ..

2 2 500
> k I

200
0

EXPERIMENT - MODEL
1

397K 447K
I

100
I
I

0.1

1.0

EFFECTIVE FIELD [ MV/cm 1

20

0.1
EFFECTIVE FIELD [ MV/cm

1.0

Fig. 4. E,fl dependences of electron mobility in the inversion layer on (100) surface in the range of 77 K to 447 K. Substrate acceptor concentration was 3.9 x 1OI5 ~ 1 1 1 Here, ~ ~ . E,fi is defined with 1) of 1/2. Open circles show the experimental data. The solid curves were calculated using (6H8).

Fig. 2. Hole mobility in inversion layer at 300 K and 77 K versus effective Here, E e ~ field, E,n as a parameter of substrate donor concentration, ,Yo. is defined by E p = ~ q . (NCv,,,,1 7 .N , ) / E s , with IT of 1/3.

. . . . . . . .

Total Mobility
0.1 1.0 EFFECTIVE FIELD [ MV/cm 1 Fig. 5. dependences of hole mobility in inversion layer on (100) surface in the range of 30 K to 447 K. Substrate acceptor concentration was 5.2 x 1OI5 cm-. Here, E,fi is defined with 11 of 1/3. Open circles show the experimental data. The solid curves were calculated using (6)-(8).

EFFECTIVE FIELD

Eeff

Fig. 3. Schematic diagram of E,E or AV, dependence of mobility in inversion layer by three dominant scattering mechanisms.

300 K, the mobility decreases steeply at E,R higher than 0.5 MV/cm. 3) At 77 K, the mobility is roughly proportional to E ; : at high E,R. In contrast, the E e dependence ~ of the hole mobility in Fig. 2 as described below. 1 ) At 300 K, the E,R dependence does not exhibit a single power law over any E e range ~ and is a little stronger than E ; , . 2 ) At 300 K, the change in a slope at high Eeffis not so marked as for the electron mobility. 3 ) At 77 K, the hole mobility is nearly proportional to E , : , which
is rather weaker than that for the electron mobility. In order to understand these E,R dependences quantitatively, it is necessary to characterize the universal curve in terms of scattering mechanism. Fig. 3 shows a schematic dia(or N , ) dependences on the basis of a general gram of the l ? , ~ understanding of the inversion layer mobility. According to this diagram, the universal curve can be divided into phonon scattering term and surface roughness scattering term. If this is true, the difference in the E,R dependence between the electron mobility and the hole mobility can be ascribed to surface roughness scattering, because the difference becomes larger at high E,tf and/or at low temperature. In order to examine this interpretation, the temperature dependences of the electron and hole mobilities were measured.

Figs. 4 and 5 show the E,R dependences of the electron mobility with N A of 3.9 x 1015 cm- and the hole mobility with N o of 7.8 x 1015 cm-. Here, V d in the measurement of g d was taken to be 50 mV, because we focused on higher N , region. The almost no temperature dependence of the hole mobility in the range of 30-77 K means that the mobility in this temperature region is limited only by surface roughness scattering. It has been, therefore, confirmed that the mobility limited by surface roughness scattering has remarkably different Eeff dependences between electrons and holes. With increasing temperature, on the other hand, the E e dependences ~ of both the electron and hole mobilities approach roughly This fact shows that the mobility limited by photon scattering follows the same power law for both electrons and holes. Based on the above findings, a calculation using a simple model was performed. The mobility limited by phonon scattering, &,h, was determined by

2360

IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 41, NO. 12, DECEMBER I994

Here, the constant coefficient A was taken to be 2.0 x IO5 and 6.1 x lo4 for the electron and hole mobilities on (loo), respectively, where p , E e ~ and T should be in cm2/V.s, -0.3, was taken from V/cm, and Kelvins. The power of the results in Figs. 4 and 5. The power of T , -1.75, was determined from the temperature dependence of the electron mobility at E,R of 0.2 MV/cm, where the contributions of surface roughness scattering and Coulomb scattering are so small that the temperature dependence of only phonon scattering can be determined. The mobility limited by surface roughness scattering, ,uL,,, was modeled separately for the electron and hole mobilities. For the electron mobility
p5r= B . E, ;'.

1 0 '

IOi

1
IOii

IOi2 Ns
[ cm-2

(7)

6. Electron mobility component determined from deviation from univerHere, the parameter, 7 , was determined to be 2.6 so that the Fig. sal curve, p,,,,,lorr,l,, versus surface carrier concentration, S,, , as a parameter total mobility at 77 K fitted the experimental data. The constant of substrate acceptor concentration, x {!,.r,,,l<,,,Ll,was calculated from (9). coefficient, B, was taken to be 4.5 x 10". The reason why y is slightly different from the power at 77 K, which was 104 roughly -2, is that phonon scattering still has an influence on the electron mobility at 77 K. As for the hole mobility, on the other hand, the experimental mobility at 30 K was used as psr. It should be noted that, as seen in Fig. 5, psr for holes cannot be represented by the single power law of E,R. Using Matthiessen's rule, the total mobility, ptot, is described by
A \ ~

d = Fpl;

+ P,l'

(8)

300 K
NS= 2x10"

\ \ ;

The solid line in Figs. 4 and 5 represent the calculated results. A good agreement between the experimental and calculated mobilities is obtained at moderate or high E,R, where Coulomb scattering is negligible. This fact confirms us that the difference in the E e dependence ~ between the electron mobility and the hole mobility is attributed to the difference in the E,tf dependence of pYr. Also, the deviation of the hole mobility at 300 K from is understood by considering that surface roughness scattering affects the hole mobility over a wider E,R range even at 300 K.

'oi016

L
1017 10'8

NA,ND

bm31

Fig. 7. Dependences of pr,,ul~l,,,l,for electrons and holes in inversion layer at 300 K on substrate acceptor concentration, -V.~i and substrate donor concentration, S n , respectively. Here .\*, is 2 x 10" cm-".

Ee2 '

C. Deviation jinm the Universal Curves

As already seen in Figs. 1 and 2, the electron and hole mobilities deviate from the universal cures near the threshold voltage even at 300 K. Moreover, this deviation becomes larger as the substrate impurity concentration increases. These facts suggest that the mobility is considerably affected by Coulomb scattering associated with substrate impurity. It has been pointed out [17]-[191, on the other hand, that the application of a finite drain voltage can lead to a decrease in carrier concentration near the drain and a resulting lower p e at ~ low N,. In order to minimize this error, V d was selected to be as low as 10 mV for an accurate characterization of the mobility at low N,. The behavior of the mobility component corresponding to the deviation, i~,o,~olnt,,was studied quantitatively using (9), in order to examine whether the deviation is caused by substrate impurity scattering.

Here, jLcoulomb was determined as a function of N,, because the Coulomb scattering rate should be characterized as function of N,, which is directly related to the screening effect and the electron energy [20]. Fig. 6 shows the N, dependence of pcoulomb for the electron mobility at 300 K as a parameter of N,q. It is seen that pco,lorn~ increases in proportion to N:', independent of Moreover, p,,,,lomt, is found to decrease with an increase in N.4. Fig. 7 shows the substrate impurity concentration dependences of p,,,lomb of the electron and hole mobilities at N , of 2 x 10l1 cm-2. At 300 K, the deviation from the universal curves is distinctly observed in the substrate impurity concentration higher than 7.2 x 1016 cm-3 for the electron mobility and higher than 1.6 x 10l6 cm-3 for the hole mobility. The inversely linear relationship between /L,o,lomh and the substrate impurity concentrations confirms that the substrate impurity is the main Coulomb scattering center. With substrate impurity concentrations lower than 5 x 10" cmP3, however, other Coulomb scattering centers are responsible. Fig. 8 shows p,-o,lomb of the electron mobility at 77 K at N, = 2 , 5 and 10 x 10" cm-'. Here, pcoulomb was determined by subtracting the contribution of p p h estimated

TAKAGI et al.: ON THE UNIVERSALITY OF MOBILITY: PART I

2361

1 0 '

I
:

ELECTRON

NS
9

[cm-*]

lxlo'2
5x10" 2X1011

. I

\ 1 x 10"

m
0.1 1 E F F E C T I V E F I E L D Eeff C M V / c m 3

NA=~.~x~~'~c~TI-'

7.2 x 1016 cm-3 has the weaker N A dependence. It is known [l], [20] that there are three types of Coulomb scattering centers that can affect the inversion layer mobility; substrate impurities, interface state charges and charges trapped in SiO2. Thus interface state charges or trapped charges are responsible for bcoulomb with N A lower than 7.2 x 1016 cmP3. In addition, it is found that the N , dependence of pcoulomb in this N A region becomes weaker. This weaker N , dependence is also considered to come from the change in the type of the dominant scattering centers. The Coulomb interaction between

0
\

IO'

.A.
I\

N i : t

300 K

ELECTRON

.
10'-

j ; :

- 2 .
0

F N INJECTION
~~.~x~o"cm-'

*.

k.

2
10'

a deviation from the universal curve can be examined by studying the mobility of MOSFET's degraded by carrier injection, because carrier injection into Si02 is known to generate interface charges [2 11. Therefore, the inversion layer mobility after Fowler-Nordheim (FN) tunneling injection into Si02 was studied. Electrons were injected into Sios from the inversion layer in n-channel MOSFET's. The injection current was maintained at a constant value (1inj = 2.65 x loP4 A/cm2. Fig. 9 shows the E,R dependence of the electron mobility before and after FN tunneling injection as a parameter of the number of injected electrons per unit area, Ninj. After the injection the deviation from the initial curve is clearly observed and becomes larger at lower E,R (thus lower N,), similar to the mobility on higher impurity concentration substrates. Also, the deviation becomes larger with an increase in Ninj. Fig. 10 shows the relationship between pCoulomb at N , of 2 x 10l1 cmP2and the threshold voltage shift, A&, estimated from CV curves. The number of interface charged centers per unit area, Nint, is also shown in the horizontal axis. Here, Nlnt was calculated from AV,, under the assumption that all the

generated charges are located at the interface. bcoulomb is found to be inversely proportional to Nint. It is confirmed from this result that the mobility degradation is caused by the generated interface charges. IV. CONCLUSION This paper has reported the studies of the inversion layer mobility in n- and p-channel MOSFET's over a wide range of substrate impurity concentrations from the viewpoint of a universal relationship against the effective field, E,R. Universality has been found to be maintained up to a substrate impurity concentration of 10'' cmp3. It has also been found that the E e dependence ~ of the universal curves can be explained over a wide temperature range by the combination of phonon scattering and surface roughness scattering. While phonon scattering provides the same E,R dependence between the electron mobility and the hole mobility, a difference in the E,E dependences of the mobility limited by surface roughness scattering has been ob-

2362

IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 41, NO. 12, DECEMBER 1994

served. It has been concluded that surface roughness scattering affects the electron and hole transports differently. On the other hand, a significant lowering from the universal curve, which is attributable to substrate impurity scattering, has been observed even at 300 K for the substrate impurity concentrations higher than 7.2 x and 1.6 x cmP3 for the electron mobility and the hole mobility, respectively. The inversion layer mobility limited by substrate impurity scattering has been found to be in proportion to the N: for both electrons and holes. It has also been experimentally confirmed in stressed MOSFETs that a marked deviation from the universal curve can be ascribed to Coulomb scattering associated with the generated interface charges.

Shin-ichi Takagi (M93) was bom in Tokyo, Japan, on August 25, 1959. He received the B.S., M S , and Ph.D. degrees in electronic5 engineenng from the University of Tokyo, Tokyo, Japan, in 1982, 1984, and 1987, respectively He joined the Toshiha Research and Development Center, Kawasaki, Japan, in 1987, where he has been engaged in research on the device physics of Si MOSFETs. including the carrier transport in inversion layer, the impact ionization phenomena, the hot camer degradation and the electnc properties of Si/SiOz interface. He is currently d Visiting Scholar at Stanford University, Stanford, CA, where he i r working on the Si/SiGe heterostructure devices D r .Takagi is a member of the IEEE Electron Device Society and the Japan Society of Applied Phyws

ACKNOWLEDGMENT
The authors wish to thank K. Nishinohara and N. Shigyo for their valuable comments concerning the measurement method. They would like to acknowledge N. Konishis contribution of performing the process simulation. They are also grateful to M. Yoshimi, K. Natori, and T. Wada for their encouragement throughout this work. They are indebted to F. Umibe for reviewing the original manuscript and suggesting revisions in its English.
Akira Toriumi received the B S. degree in physics and the M.S. and Ph D. degrees in applied physics from the University of Tokyo, Tokyo, Japan, in 1978, 1980, and 1983, respectively. He Joined the Toshiba Research and Development Center, Kawasaki, Japan, in 1983, where he has been engaged in research on the MOSFET device physics. Dunng 1988-1990 he was a Visiting Scientist at the Massachusetts Institute of Technology, where he studied the physics of quantum effect devices. He is currently working on the physics and . technology of 0.1 irm CMOS, thin Si02 reliability, andthe quantum effects of very small silicon devices. Dr. Toriumi is a member of the Physical Society of Japan, the Japan Society of Applied Physics. and the American Physical Society.

REFERENCES
T. Ando, A. B. Fowler, and F. Stem, Re\,. Mod. Phys., vol. 54, pp. 437-672, 1982. A. G. Sabnis and J. T. Clemens, in IEDM Tech. Dig., 1979, pp. 18-21. S. C. Sun and J. D. Plummer, IEEE Trans. Electron Devices, vol. ED-27, pp. 1497-1508. 1980. N. D. Arora and G. S. Gildenblat. IEEE Trans. Electron Devices, vol. ED-34, pp. 89-93. 1987. J. T. Watt and J. D. Plummer, in Symp. 0 1 7 VLSl Technol. Tech. Dig., 1987, pp. 81-82. S. Takagi, M. Iwase, and A. Toriumi, in IEDM Tech. Dig., 1988, pp. 398401. S. Takagi. M. Iwase. and A. Toriumi, in Exr. Ahstr. 22nd Conf. Solid State Devices and Materials, 1990, pp. 275-278. C. L. Huang and G . Sh. Gildenhlat. IEEE Trans. Electron De).ices, vol. 37. pp. 1289-1300, 1990. K. Lee. J. S. Choi, S. P. Sim, and C. K. Kim, IEEE Trans. Electron Devices, vol. 38, pp. 1905-1912. 1991. S. W. Lee. IEEE Trans. Computer-Aided Design, vol. 8, pp. 724-730, 1989. H. Shin, G. M. Yeric, A. F. Tasch, Jr.. and C. M. Maziar. Solid-state Electron., vol. 34, pp. 545-552, 1991. V . M. Agostinelli, Jr., H. Shin. and A. F. Tdsch, Jr., IEEE Trans. Electron Devices, vol. 38, pp. 151-159, 1991. S. Takagi, A. Toriumi, M. Iwase, and H. Tango, to be published in IEEE Tram. Electroti Delices. C. G. Sodini. T. W. Ekstedt, and J. L. Moll. Solid-State Electron., vol. 25, 1982, pp. 833-841. P. D. Chow and K-L, Wang, IEEE Trans. Electron Devices, vol. ED-33, pp. 1299-1304, 1986. M. Nakamura, N. Aoki, N. Konishi, and H. Amakawa, Extended Abstracts, the 39th Spring Meeting, 1992, The Japan Society ofApplied Pl7ysics and Related Societies no. 2, 29a-ZL-10, p. 668. S. Selberherr, W. Hansch. M. Seavey, and J. Slotboom, Solid-State Hectroii., vol. 33, pp. 1425-1436, 1990. C.-L. Huang and G. Sh. Gildenblat, Solid-State Electron., vol. 36, pp. 61 1-615, 1993. K. Nishinohara, H. Tanimoto, N . Konishi, S. Takagi, and N. Shigyo, in Int. Workshon on VLSI Process a ~ De1ic.e d MOCiplinn Tech. Din.. 1993. pp. 17617;. F. Stem and W. E. Howard, P h w Rev., vol. 163, pp. 816835. 1967. S. Manzini, .I. Appl. Phys., vol. 57, pp. 41 1 4 1 4 , 1985.

Masao Iwase was born in Aomon, Japan, in 1963. He joined the Toshiba Research and Development Center, Kawasaki, Japan, in 1982, where he has been engaged in the development of beam Iithography and submicrometer CMOS technology. He is currently in the Process Rewearch Department at Toshiba ULSI Research Laboratones. Mr. Iwase is a member of the Japan Society of Applied Physics.

Hiroyuki Tango was born in Saitama Prefecture, Japan, on February IO, 1942 He received the B S degree in electncal communication engineering and the M S and Ph.D. degrees in electronics engineering, all from Tohoku University, Sendai. Japan. i n *iL2 1965, 1967, and 1970 In 1970, He joined the Toshiba Research and Development Center. Kawasaki, Japan, where he studied MOS device physics, MOS on bulk, and MOS on SOS integrated-circuit technology. From 1979 to 1983 he researched SOS LSI technoloev as the Head of an SOS technology group at Toshiba. From 1984 to I9ik he supervised the research and development of MOS device physics, device simulation, MOS-LSI process. and MOS devices of SO1 technology. Since 1988 he has supervised the R&D of DRAM, EEPROM, and CCD devices. Dr. Tango is a member of the Japan Society of Applied Physics and the Institute of Electronics, Information, and Communication Engineers.