Enhanced gate-induced floating-body effect in PD SOI MOSFET under external mechanical strain

(1)

Enhanced gate-induced

ﬂoating-body effect in PD SOI MOSFET under external

mechanical strain

Chih-Hao Dai

a

, Ting-Chang Chang

a,b,c,

⁎

, Ann-Kuo Chu

a

, Yuan-Jui Kuo

a

, Shih-Ching Chen

b

,

Chih-Tsung Tsai

b

, Wen-Hung Lo

b

, Szu-Han Ho

d

, Guangrui Xia

e

, Osbert Cheng

f

, Cheng Tung Huang

f

a

Department of Photonics, National Sun Yat-Sen University, Kaohsiung, Taiwan, ROC

b

Department of Physics, National Sun Yat-Sen University, Kaohsiung, Taiwan, ROC

c

Center for Nanoscience and Nanotechnology, National Sun Yat-Sen University, Kaohsiung, Taiwan, ROC

d

Department of Electronics Engineering, National Chiao Tung University, Hsinchu, Taiwan, ROC

e_{Department of Materials Engineering, the University of British Columbia, Vancouver, Canada} f

Device Department, United Microelectronics Corporation, Tainan Science Park, Taiwan, ROC

a b s t r a c t

a r t i c l e i n f o

Available online 16 July 2010 Keywords:

SOI GIFBE

Electron-valance band tunneling Strained silicon

The influence of tensile mechanical strain on gate-induced floating-body effect (GIFBE) in advanced partially depleted SOI n-MOSFETs was investigated. Both drain current and mobility enhance after applying strain due to the reduction of average transfer effective mass. However, it was found that the GIFBE becomes serious under the mechanical strain. To explain this phenomenon, wefirst clarify the mechanism of GIFBE using different operation conditions. The experiment results indicate that the GIFBE can be attributed to the anode hole injection (AHI) rather than the widely accepted mechanism of electron band (EVB) tunneling. Based on the AHI model, the enhanced GIFBE under the mechanical strain is mainly due to the narrowing of band gap induced by the strain in the poly-gate.

1. Introduction

Silicon-on-insulator (SOI) CMOS devices are attractive since they provide high current drivability and reduced junction capacitance as compared to bulk-Si devices[1]. However, as the gate length scales below the 100 nm, it becomes critical to realize the high drive current due to the degradation of carrier mobility caused by the required increase in channel doping[2]. Therefore, the use of strained silicon technique offers an alternative method to enhance the SOI perfor-mance through an increased channel mobility. In addition, with the aggressive shrinking of gate oxide thickness, the strain technique can also reduce the gradually increasing gate tunneling current in advanced SOI devices [3]. According to a previous study, the increasing tunneling current in partially depleted (PD) SOI MOSFETs has been reported to cause a newfloating-body effect in the linear operation region, phenomenon referred to as “linear kink effect” (LKE) or“gate-induced floating-body effect” (GIFBE)[4]–[6]. Howev-er, the influence of mechanical strain on GIFBE has not been studied yet. Besides, there are few studies that confirm the origin of the GIFBE in PD SOI MOSFETs by using systematical operation conditions. Therefore, the purpose of this work is to clarify the mechanism of

GIFBE for PD SOI n-MOSFETsﬁrst. Then, based on our new model, we further study the inﬂuence of mechanical strain on GIFBE. The experimental results demonstrate that anode hole injection (AHI) from the poly-gate is the main dominant mechanism responsible for GIFBE. Based on the AHI model, it was found that the mechanical strain has a dramatic impact on the GIFBE of PD SOI n-MOSFET. 2. Experiment

Using 65 nm SOI CMOS technology, PD SOI n-type MOSFETs are employed the T-gate structure to investigate the GIFBE mechanism. The SOI wafers were fabricated using SMART-CUT technology. The silicon_{film and buried oxide thicknesses for the devices are 75 nm} and 145 nm, respectively. The gate oxide with a thickness of 12 Å was grown by in-situ steam generation (ISSG), with a channel doping concentration of about 3 × 1018cm−3. The channel currents follow in theb110Ndirection on (100) substrates. In this work, devices with a channel width (W) of 1 um, and a length (L) ranging from 1μm to 0.2μm were selected. The devices with a floating or a grounded body were measured to study the phenomenon of LKE. To further investigate the GIFBE, temperature-dependent electrical character-istics were performed at temperatures ranging from 303 K to 453 K. In order to study the in_{fluence of the mechanical strain on GIFBE, the} tensile strain along the channel length was introduced by mechanical bending as follows; the thickness of the silicon substrate was reduced from 800μm to 50 μm using a Struers RotoPol-21 polisher.

⁎ Corresponding author. Department of Physics and Institute of Electro-Optical Engineering, Nation Sun Yat-Sen University, 70 Lien-hai Rd., Kaohsiung 80424, Taiwan, ROC. Tel.: + 886 7 5252000x3708; fax: + 886 7 5253709.

E-mail address:[email protected](T.-C. Chang).

Contents lists available atScienceDirect

Surface & Coatings Technology

(2)

Subsequently, the silicon substrate was adhered to a metal foil and thenﬁxed on a bending form, as shown in theFig. 1. All experimental curves were measured using a Keithley 4200 semiconductor param-eter analyzer.

3. Results and discussion

Fig. 2(a) shows the transfer characteristics in the linear region (VD= 50 mV) for a PD SOI n-type MOSFET underﬂoating-body (FB)

and grounded-body (GB) operations. It can be clearly observed that an excess of drain current and a second gm peak appear in the FB device. To realize the second peak in this unusual gm_–VGcurve, the IB–VG

transfer characteristic was also measured for comparison. Obviously, the IB under the GB operation increases rapidly beyond the gate

voltage of 0.9 V, corresponding to the second gm peak. This phenomenon indicates that additional carriers (hole) accumulate in the body beyond this voltage under the FB operation. Therefore, the body potential increases exponentially, which causes the threshold voltage (VT) to reduce, resulting in a higher drain current (Idlin) and

the second gm peak in the linear region.

To examine the origin of additional carriers (hole) caused by the gate leakage current (IG),Fig. 3(a) illustrates the IGand IBversus gate

voltage (VG) for different N+poly-gate area (1μm2to 0.2μm2) while

the source/drain (S/D) is grounded. Clearly, the IBhas a signiﬁcant

enhancement as the IGincreases, meaning that the source of IB is

strongly dependent on the IG. Furthermore, the linear relationship

between IBand IGis also obtained as shown in the inset ofFig. 3(a).

The IB–VGcurve inFig. 3(a) shows that the VGcorresponding to the

rapid increase of IBfor different N+poly-gate areas always locates

around 0.9 V. This indicates that additional carriers only can be generated under certain vertical electricalﬁeld regardless of the poly-gate area.

It was proposed that the IGconsists of electron tunneling from the

conduction band (ECB), electron tunneling from the valence band (EVB), and hole tunneling from the valence band (HVB) as shown in

Fig. 3(b)[7]. In fact, the hole accumulation in the p−substrate has been explained due to EVB and HVB tunneling currents under a FB condition. However the EVB tunneling model was widely accepted to explain the LKE phenomenon because the HVB is typically very small for NMOS[5].

Fig. 1. The illustration of sample bending on the holder with curvature radius R = 30 mm.

Fig. 2. (a) Transfer characteristics for PD SOI n-MOSFETs under GB and FB operations. (b) The gm as a function of the gate voltage under the FB operation in the linear region (VD= 50 mV). IBis measured under the GB operation.

Fig. 3. (a) IG–VGand IB–VGcurves for different N+poly-gate areas. Inset displays the linear

relationship between IBand IGfor different N+poly-gate areas. (b) The schematic diagram

(3)

InFig. 4(a), the schematic band diagram of EVB shows that the electrons tunnel from the valence band of the Si substrate to the poly-gate conduction band under a suf_{ficiently large vertical electric field.} As a result, holes accumulate in the body due to thefloating body, which leads to a rise in the body potential. However, the other mechanism, that of AHI[8], also shown in theFig. 4(a), is also a possible model to explain the LKE. In AHI, when the gate oxide is thin enough, electrons can tunnel from the inversion layer to the poly-gate (anode) and generate hot holes by impact ionization in the poly depletion region. Then these hot holes can inject over or through the anode/oxide interface energy barrier and traverse the oxide layer to the body (cathode). Consequently, the HVB tunneling current becomes significant due to the increase of AHI tunneling current from the poly-gate to the p−substrate. Both mechanisms generate holes that accumulate in the body, but there are two key differences between the EVB tunneling and AHI model. One is the source of electrons: the electrons in the EVB tunneling model come from the band, but the source of electrons in the AHI model is from the conduction band. The other difference is that the intensity of the vertical electricfield in the EVB tunneling model is larger than that in the AHI model. This is because the VGin the EVB tunneling model

needs to be larger than the silicon band gap (1.1 V) to make the band energy level in the silicon substrate align with the gate conduction band[7]. Our experimental results that identify the rapid increase of IB

for different N+poly-gate areas located at around 0.9 V imply that the AHI model is more likely to be responsible for the GIFBE.

To further verify that the AHI model is the dominant mechanism of GIFBE, two operation conditions illustrated inFig. 4(b) are performed to distinguish the different sources of IG that result in hole

accumulation. We measured the IG and IB versus VGof the device

under S/Dfloating and grounded operations. The result under the S/D floating operation is shown inFig. 5, where it can be clearly observed that the gate leakage current becomes insignificant because the S/D cannot supply sufficient minority carriers (electrons) to the inversion layer. The IBis not evident (~ 0.1 pA) in this operation. On the contrary,

both currents show very pronounced increases under the S/D grounded operation. The results prove that the origin of GIFBE can be attributed to the electrons in the inversion layer, rather than the electron-hole pair separated in the valence band of the P−substrate. Therefore, the AHI model is the dominant mechanism responsible for GIFBE.

The second gm peak introduced inFig. 2was also analyzed as a function of temperature in the range from 303 K to 453 K.Fig. 6(a) shows that the second peak amplitude becomes smaller and shifts left as the temperature increases. This reduction is ascribed to an increase in recombination rate in the source and drain junctions as temper-ature increases[9]. To investigate the second gm peak shift, the IB–VG

curve as a function of temperature was also measured and shown in

Fig. 6(b) for comparison. It shows that the left shift trend of IBagainst

temperature is similar to that of the second gm peak. However, IBhas

no signiﬁcant change after a VTcorrection for different temperatures,

as shown in the insert of Fig. 6(b). The results provide further evidence that hole injection from the anode is direct tunneling, because it is independent of temperature.

To further investigate the inﬂuence of tensile strain on GIFBE,

Fig. 7(a) presents the transfer characteristics of the SOI n-MOSFET under the GB operation with mechanical strain, as compared to the un-strained device (without strain). The result indicates that the Idlinand gm improve 3.0% and 3.4%, respectively. The enhancement

can be attributed to the reduction in the effective mass and intervalley scattering caused by the tensile strain effect [10]. In addition, the transfer characteristics of the devices under the FB operation both with and without tensile mechanical strain are

Fig. 4. The schematic diagrams of (a) EVB tunneling and AHI models for an ultra thin gate oxide of PD SOI n-MOSFET and (b) the device under S/Dﬂoating and grounded operations. Fig. 5. IG–VGand IB–VGcharacteristics for PD SOI n-MOSFET under S/Dﬂoating and

(4)

shown inFig. 7(b). Under the mechanical strain, the position of GIFBE shifts toward a lower VGand the magnitude of second gm

peak increases obviously. Furthermore, after the second peak, the gm for the strained device recovers its initial trend (under the GB operation trend), as compared to the un-strained device.

To realize the inﬂuence of the strain on the second gm peak clearly, the IGand IB versus VGwith and without mechanical strain were

measured in Fig. 8. It shows that the left shift trend of IB under

mechanical strain is similar to that of the second gm peak. Besides, the IG decreases slightly under the mechanical strain. This opposing

reduction is due to the strain-induced change in the conduction band offset between Si and the SiO2 gate dielectric[3].

Based on the AHI model, the IB generally decreases as the IG

decreases. However,Fig. 8indicates that even if the IGdecreases under

the mechanical strain, the IBstill increases signiﬁcantly. This unusual

result is mainly related to the narrowing band gap induced by the mechanical strain effect[11]–[13]. Because the impact ionization rate is exponential to the energy band gap, the narrowing band gap causes more additional holes to be generated in the poly depletion region, leading to the increase of IB and a signiﬁcant second gm peak in

strained SOI devices. Besides, after the second peak, the gm for the strained device recovers its initial trend. This implies that the enhanced IBcauses the body potential to reach the junction diode

threshold voltage. Therefore, for the strained device, holes cannot further accumulate in the body, resulting in the gm recovery after the second peak[4].

4. Conclusion

In this work, wefirst demonstrate that the EVB tunneling model is not the dominant mechanism for the GIFBE. When oxide thickness is reduced to only 12 Å, both IGand IBare insignificant under the S/D floating

operation. Therefore, the GIFBE can be attributed to the electrons of the inversion layer supplied from the S/D. This result indicates that the increase of HVB tunneling current induced by AHI is the dominant mechanism responsible for the GIFBE. By analyzing IBas a function of Fig. 6. (a) Transfer characteristics as a function of temperature varying from 303 to

453 K under the FB operation and (b) IB–VGcharacteristics of the PD SOI n-MOSFET

with body-contact for different temperatures. The inset shows the IB–VGafter a VT

correction, where the VTis obtained from linear extrapolation from the gmmax.

Fig. 7. Transfer characteristics for PD SOI n-MOSFETs with and without mechanical strain under (a) the GB operation and (b) the FB operation.

Fig. 8. IG–VG and IB–VGcharacteristics for PD SOI n-MOSFET with and without

mechanical strain. The inset shows the schematic diagrams of the AHI model under mechanical strain.

(5)

temperature, we also prove that the hole current (IB) is direct tunneling

from the poly-gate. This is because IBis independent of temperature after

a VTcorrection. Based on the AHI model, it was found that the tensile

strain has a strong impact on the GIFBE. This is owing to the band gap narrowing induced by the strain effect in the poly-gate.

Acknowledgments

Part of the work was performed at the United Microelectronics Corporation. The work was supported by the National Science Council under Contract NSC-98-3114-M-110-001, and NSC-97-2112-M-110-009-MY3.

References

[1] J.C. Sturm, K. Tokunaga, J.P. Colinge, IEEE Electron Device Lett. 9 (1988) 460.

[2] T. Ghani, K. Mistry, P. Packan, S. Thompson, Stealer M, Tyagi S, VLSI Tech Dig, 2000 174.

[3] W. Zhao, A. Seabaugh, V. Adams, D. Jovanovic, B. Winstead, IEEE Electron Device Lett. 26 (2005) 410.

[4] A. Mercha, J.M. Raﬁ, E. Simoen, E. Augendre, C. Claeys, IEEE Trans. Electron Dev. 50 (2003) 1675.

[5] P.G.D. Agopian, J.A. Martino, E. Simoen, C. Claeys, Microelectron 38 (2007) 114. [6] M. Cassé, J. Pretet, S. Cristoloveanu, T. Poiroux, C. Fenouillet-Beranger, F. Fmleux, C.

Raynaud, G. Reimbold, Solid-State Electronics 48 (2004) 1243. [7] W.-C. Lee, C. Hu, IEEE Trans. Electron Devices 48 (2001) 1366. [8] K.F. Schuegraf, C. Hu, IEEE Trans. Electron Devices 41 (1994) 761.

[9] L. Vancaillie, V. Kilchytska, P. Delatte, L. Demeus, H. Matsuhashi, F. Ichikawa, D. Flandre, IEEE Int. SOI Conf. (2003) 78.

[10] S.E. Thompson, M. Armstrong, C. Auth, S. Cea, R. Chau, G. Glass, IEEE Electron Device Lett. 25 (2004) 191.

[11] F.A.I.R. RB, J. Appl. Phys. 50 (1979) 860.

[12] T. Irisawa, T. Numata, N. Sugiyama, S. Takagi, IEEE Trans Electron Dev. 52 (2005) 993.

[13] Y.J. Kuo, T.C. Chang, P.H. Yeh, S.C. Chen, C.H. Dai, C.H. Chao, Thin Solid Films 517 (2009) 1715.