A Study of Charge Control and Gate Tunneling in a Ferroelectric-Oxide-Silicon Field Effect Transistor
- PDF / 82,781 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 42 Downloads / 191 Views
A Study of Charge Control and Gate Tunneling in a Ferroelectric-Oxide-Silicon Field Effect Transistor Yih-Yin Lin, Yifei Zhang1, and Jasprit Singh Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, MI 48109-2122 1 Integrated Device Technology, Duluth, GA 30097 ABSTRACT It is known that conventional metal-oxide-silicon field-effect transistor (MOSFET) devices will have gate tunneling related problems at very thin oxide thicknesses (dox ≤ 20 Å). In this paper we discuss the potential of polar and non-polar high-dielectric films for gate tunneling suppression and charge control. We also examine ferroelectric thickness effects in ferroelectricoxide-silicon field effect transistors (MFISFETs). Our formalism is based on a blocking-layer model for the ferroelectric film and a self-consistent solution of the Poisson and Schrödinger equations. We show that the polarization effects of ferroelectrics can allow greater controllability of the silicon interface (mobile) charge density and also the high dielectric constant effectively suppresses gate tunneling probability. In addition, the effects of ferroelectric film thickness are quite important in a MFISFET device and allow a small control of the threshold voltage. Results will be presented for the capacitance-voltage curve, tunneling probability, and leakage currents. As an example, we find for a sheet charge of 1013 cm-2, the gate tunneling probability in a MOS structure is 10-6 (dox = 13 Å), while that in a MFISFET is 1014 with the same equivalent oxide thickness. Our studies also show that details of the blocking layer model (dead layer thickness, coercive field, and polar charge) can be estimated from C-V measurements. INTRODUCTION The electronic industry is in part driven by the shrinking gate length of MOSFETs. However, it is now quite clear that once gate lengths approach ~ 0.05 µm the Si/SiO2 structure will no longer be viable for transistors. The reason is the high gate leakage current that occurs once the oxide thickness approaches ~ 20 Å. In addition, doping fluctuation [1] is expected to create serious problems as channel lengths shrink to approximately 500 Å. For example, only ten dopant atoms are contained in a 0.1-µm channel for the implant dosage of 1012 cm-2 at a depth of 0.01 µm, resulting in random distribution. To overcome this problem, in recent years there has been a growing interest in ferroelectric thin films. Ferroelectrics are materials which have a large dielectric constant (in the range of ~100ε0). The high dielectric constant makes it a possible gate dielectric material to avoid tunneling since it is much thicker than a SiO2 film with the same gate capacitance. It is known that ferroelectrics like barium strontium titanate (BST) can have very high polarization charge ~ 1014 cm-2 [2,3]. This polar charge may be developed during growth or can be induced by an external electric field. It should be noted that this polar charge can be used to enhance the modulation charge formation with less doping. In devices
Data Loading...
