Higher School of Economics, Moscow, Russia. Zelenograd, Moscow, Russia

Transcription

1 Advanced Materials Research Online: ISSN: , Vols , pp doi: / Trans Tech Publications, Switzerland Hardware-Software Subsystem for MOSFETs Characteristic Measurement and Parameter Extraction with Account for Radiation Effects Konstantin O. Petrosyants 1 3 *, Igor A. Kharitonov 1,3 and Lev M. Sambursky Moscow Institute of Electronics and Mathematics, National Research University Higher School of Economics, Moscow, Russia 2 Institute for Design Problems in Microelectronics, Russian Academy of Sciences, Zelenograd, Moscow, Russia 3 Science and Research Institute for Microelectronics and Instrumentation, Moscow, Russia * Corresponding author eande@miem.edu.ru Keywords: Measuring system, MOSFETs, radiation effects, test structures, SPICE models, parameter extraction Abstract. Hardware-software subsystem designed for MOSFETs characteristic measurement and SPICE model parameter extraction taking into account radiation effects is presented. Parts of the system are described. The macromodel approach is used to account for radiation effects in MOSFET modeling. Particularities of the account for radiation effects in MOSFETs within the measurement and model parameter extraction procedures are emphasized. Application of the subsystem is illustrated on the example of radiation hardened 0.25 μm SOI MOSFET test structures. Introduction Complementary MOS (CMOS) technology is the most widely used for fabrication of radiation-hard electronic equipment for aerospace, wireless communication, nuclear physics and energy control, medicine, military and other special applications. Various design and technology concepts, like Radiation Hardness by Design (RHBD) techniques [1] or Silicon-on-Insulator/Sapphire (SOI/SOS) technologies [2] may be implemented to improve radiation hardness of MOSFET ICs. The key point of radiation hardened IC design is using adequate MOSFET models and their characteristic measurement and SPICE models parameter determination. For today this problem is solved only for nonirradiated devices and ICs. For the case of irradiated devices and ICs the problem of account for radiation effects is under consideration and measurements for transistor parameter determination are very complicated, time-consuming and expensive. In this paper we describe a hardware-software subsystem designed for measurement and processing of MOSFET electrical characteristics and extraction of SPICE model parameters with account for radiation effects. The automation procedures are included into this subsystem to reduce the amount of measurements and cost. Hardware-Software Subsystem Schema The schema of the hardware-software subsystem for measurement and processing of MOSFETs characteristics and extraction of SPICE models parameters taking into account radiation effects is shown in Fig. 1. The hardware part includes: instruments for MOSFETs electrical measurements; radiation test equipment; sets of MOSFET test structures. All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans Tech Publications, (# , Pennsylvania State University, University Park, USA-18/09/16,07:31:31)

2 Advanced Materials Research Vols The software part includes: IC-CAP [3] tool for control of measurement equipment and extraction of MOSFETs SPICE macromodel parameters; SPICE macromodel library for conventional (bulk) and SOI/SOS MOSFETs with account for radiation effects; TCAD tool for MOSFET process and device simulation, virtual test structures generation. Fig. 1. Hardware-software subsystem schema MOSFET Parameter Degradation under Irradiation Conditions Physical structures of bulk (a) and insulated substrate (SOI/SOS) (b) MOSFETs and their macromodels taking into account radiation effects are presented in Fig. 2. As a result of ionizing radiation exposure, the main MOSFET parameters are degraded: threshold voltage V T (see Fig. 3), carrier mobility μ (see Fig. 4,a), subthreshold slope S (see Fig. 4,b), and leakage currents (see Fig. 4,c) [4]. For bulk MOSFETs the most critical radiation-induced effect is latch-up. This effect is modeled by a subcircuit consisting of Q1, Q2, R1, R2 (see Fig. 2,a). For SOI/SOS MOSFETs the most critical are radiation-induced currents modeled by parasitic transistors M side and M botm (see Fig. 2,b). Macromodel Approach for MOSFETs with an Account for Total Dose Effects The macromodel approach to create a compact model for MOSFETs with an account for total dose effects is common for different types of MOSFETs. In includes: 1) Replacement of a simple MOSFET model with a macromodel. 2) Addition of circuit elements to account for radiation effects such as leakage currents. 3) Including functional dependency of component parameters on dose into the macromodel using the well known physical equations. The model for a MOSFET with account for radiation effects is created in the form of a macromodel [5] [8] on the basis of a standard MOSFET SPICE model, available in a simulation tool (BSIM, BSIMSOI, EKV, MM) (see Fig. 2). The main transistor component M F reflects the behavior of the front Si SiO 2 interface. Other components are added to account for various radiation-induced static and dynamic leakage currents. Model parameters for M F and other components are made dependent on total dose using the known physical equations [4] and approximations. Fitting factors in these expressions are just the set of static radiation parameters of the macromodel.

3 752 Advanced Measurement and Test III (a) (b) Fig. 2. Physical structures of bulk (a) and insulated substrate (b) MOSFETs and their macromodels taking into account radiation effects A Modified MOSFET Characteristic Measurement Procedure The modified electrical measurement procedure is formed on top of the standard one that is launched several times during the whole process. The modification is necessary to characterize the components of the macromodel that replaces a standard SPICE model to account for radiationinduced leakage currents. The Standard Procedure. In the course of the standard measurement procedure a number of standard electrical characteristic curves of a number of length- and width-varied MOSFET devices under test are measured under the control of IC-CAP and processed therein: Id-Vg curves: I D,I B vs. V various drain voltages V D, various body voltages V B ; Id-Vd curves: I D,I B vs. V various gate voltages V G, various body voltages V B ; diode curves for p-n-junctions; I-V curves for parasitic bipolar transistor structure. The Modified Procedure [7] with an Account for Total Ionizing Dose. The known radiationhard test transistor structures (H- or O- or R-type) are essentially needed in addition to common linear (called F- or I-type) (see Fig. 5) devices to do the model parameter extraction adequately. The said special structures are designed in such a way that the lateral radiation-induced leakage currents are effectively suppressed or their active silicon areas have no lateral sides at all.

4 Advanced Materials Research Vols Fig. 3. Threshold voltage shift with dose: nmos (left), pmos (right) [4] a) Mobility b) Subthreshold slope c) Leakage current Fig. 4. Typical MOSFET curves degradation with dose [4] a) O-type b) R-type c) H-type d) I-type Fig. 5 Layout of different types of radiation hardened MOSFETs (O-, R- and H-types) used to suppress leakage currents compared to the linear type (I-type) 1) The unirradiated front MOSFET model is fully measured with the standard procedure, parasitic MOSFETs being switched off. 2) A shortened (minimum, see below) set of electrical curves is measured on the devices under test after each of the scheduled radiation exposure levels the device has undergone. All the three interfaces (front, bottom, and sidewall) are activated sequentially with the help of special test structures, so that the radiation-induced leakage currents are separated. For every single dose level a shortened set of electrical curves is measured for the front, sidewall and bottom MOSFETs: Id-Vg curves: I D vs. V a few drain voltages V D, fixed body voltage V B = 0; Id-Vd curves: I D,I B vs. V a few gate voltages V G, fixed body voltage V B = 0. Automation of MOSFET Curve Measurement. Automation of curve measurement and data processing is highly desirable to reduce time of operation and human error. A measurement data acquisition and formatting suite was developed by tying up of IC-CAP and LabView tools with instrumentation. All the curve measurements of a single transistor are carried out without reconnection of the device under test, which further reduces risks of human error and device damage.

5 754 Advanced Measurement and Test III Spice Model Radiation Parameter Extraction Strategy In our work we use three methods for SPICE model parameter extraction from MOSFET characteristics: 1) with dedicated procedures described in MOSFET SPICE model user manuals; 2) with ISExtract software tool built in TCAD; 3) with universal IC-CAP extraction tool (Fig. 1). In practice, the most frequently used method is based on the universal IC-CAP extraction tool. The standard extraction flow inside IC-CAP sequentially invokes measurement data to identify the whole set of device model parameters group by group. Modification of the extraction procedure is necessary to identify macromodel components radiation parameters. At the measurement stage a set of characteristic curves is measured for every component of the macromodel. It is therefore possible to identify the model for each component separately, and then combine the resulting models into a single macromodel card. For MOSFETs under radiation, we developed the following strategy of SPICE model parameter extraction: 1) The full set of macromodel parameters is at first extracted for the unirradiated device. 2) Among all the model parameters for MOSFET sub-components a limited number of radiationdependent parameters is selected: threshold voltage, mobility, subthreshold slope and their factors (depending on the selected models). 3) For each radiation dose D i IC-CAP is used for extraction of the set of V T (D i ), μ(d i ), S(D i ) etc. This procedure is repeated for all the doses D i : i=1,n. 4) The sets of V T (D i ), μ(d i ), S(D i ) etc. are approximated by analytical functions of the type: a + exp(b D). 5) The determined functions are embedded into the MOSFET SPICE macromodel description. 6) The MOSFET macromodel with radiation-dependent parameters is included into the SPICE model library. Example of the Subsystem Application The extraction strategy is illustrated on the example of SOI MOSFET with gate dimensions L/W = 0.25/8 μm. Separate experimental characteristic curves for the main transistor M F (Fig. 6,d) and parasitic bottom (Fig. 6,b), sidewall (Fig. 6,c) transistors were obtained. The total I-V-curves for SOI MOSFET (Fig. 6,a) are the sum of partial curves (Fig. 6,b,c,d). The set of macromodel parameters was determined by means of the described extraction strategy. The dependencies of parameters on dose D were approximated with analytical functions of the type: a + exp(b D) (see Fig. 7), where a and b are fitting factors. The error between measured and simulated I-V-characteristics is 10 15%. Conclusion A universal subsystem for measurement of MOSFETs characteristics and extraction of their SPICE model parameters under radiation influence was developed. The subsystem was designed for bulk and insulated-substrate MOSFETs. The subsystem provided three methods of SPICE-parameter extraction for MOSFETs of different type and layouts. In order to account for radiation effects in MOSFET modeling, the macromodel approach was proposed in conjunction with introducing the model parameter dependencies on dose. An automated procedure for measurement and processing of MOSFET characteristics and extraction of SPICE model parameters was established and validated on radiation hardened SOI/SOS MOSFET test structures. This work was carried out within the framework of the Basic Research Program (Project No. T3-108) at the National Research University Higher School of Economics (HSE) and Russian Foundation for Basic Research grant

6 Advanced Materials Research Vols a) total b) bottom c) side d) front Fig. 6 Measured (circles) vs. simulated (lines) characteristic curves for SOI MOSFET with gate dimensions L/W = 0.25/8 μm Fig. 7 Several radiation-dependent parameters of the front MOSFET model vs. dose References [1]. D. G. Mavis, D. R. Alexander, Employing radiation hardness by design techniques with commercial integrated circuit processes, Proc. of AIAA/IEEE Digital Avionics Systems Conference 1 (1997) [2]. R. Schwank, Space and Military Radiation Effects in Silicon-on-Insulator Devices, 1996 Intl. IEEE SOI Conference, Short Course, Sanibel Island, Florida (1996) [3]. Agilent 85190A, IC-CAP 2006 User s Guide. [4]. Ionizing Radiation Effects in MOS Devices & Circuits. Ed. By Ma T. P. and Dressendorfer P. V. John Wiley and Sons, [5]. K. O. Petrosjanc, A. S. Adonin, I. A. Kharitonov, M. V. Sicheva, SOI device parameter investigation and extraction for radiation hardness modeling with SPICE, IEEE Proc. on Microel. Test Structures 7 (1994) [6]. Petrosjanc K. O., Kharitonov I. A., Orekhov E. V., Sambursky L. M., Yatmanov A. P., A Compact SOI/ SOS MOSFET Macromodel Accounting for Radiation Effects, Proc. of IEEE East-West Design & Test Intl. Symposium (EWDTS'07), Yerevan, Armenia (2007) [7]. Petrosjanc K. O., Sambursky L. M., Kharitonov I. A., Yatmanov A. P., SOI/SOS MOSFET Compact Macromodel Taking into Account Radiation Effects, Russian Microelectronics 40(7) (2011) [8]. K. O. Petrosyants, I. A. Kharitonov, L. M. Sambursky, V. N. Bogatyrev, Z. M. Povarnitcyna, E. S. Drozdenko, Simulation of Total Dose Influence on Analog-Digital SOI/SOS CMOS Circuits with EKV-RAD macromodel, Proc. of 10th IEEE East-West Design & Test Intl. Symposium (EWDTS'12), Kharkov, Ukraine (2012)