Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

DOPING CONCENTRATION MEASUREMENT TECHNIQUE IN TFSOI DEVICES

IP.com Disclosure Number: IPCOM000007858D
Original Publication Date: 1996-Nov-01
Included in the Prior Art Database: 2002-Apr-30
Document File: 3 page(s) / 161K

Publishing Venue

Motorola

Related People

Hank H. Shin: AUTHOR [+3]

Abstract

During development of thin film silicon-on- insulator (TFSOI) devices, there were die-by-die and wafer-by-wafer threshold voltage (Vt) variations: they may be due to a doping concentration issue and/or other process integration/material problems. To clar- ity the mechanism, the doping concentrations should first be known. However, SOI devices have unique characteristics of 1) buried oxide (BOX) layer (1 k-10 kA: thick oxide), thus floating devices, and 2) very thin Si thickness (5 1000 A), and thus nearly fully depleted. Because of these material restrictions, no bulk method is straightforwardly applicable to meas- ure doping concentrations (Na for NMOS and Nd for PMOS). In addition, SOI devices are either enhancement-mode (n+pn+ for NMOS and p+np+ for PMOS) or accumulation-mode (n+nn+ for NMOS and p+pp+ for PMOS), which requires a different kind ofmeasurement methodology.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 45% of the total text.

Page 1 of 3

Technical Developments

DOPING CONCENTRATION MEASUREMENT TECHNIQUE IN TFSOI DEVICES

by Hank H. Shin, Dieter K. Schroder and Syd R. Wilson

  During development of thin film silicon-on- insulator (TFSOI) devices, there were die-by-die and wafer-by-wafer threshold voltage (Vt) variations: they may be due to a doping concentration issue and/or other process integration/material problems. To clar- ity the mechanism, the doping concentrations should first be known. However, SOI devices have unique characteristics of 1) buried oxide (BOX) layer (1 k-10 kA: thick oxide), thus floating devices, and 2) very thin Si thickness (5 1000 A), and thus nearly fully depleted. Because of these material restrictions, no bulk method is straightforwardly applicable to meas- ure doping concentrations (Na for NMOS and Nd for PMOS). In addition, SOI devices are either enhancement-mode (n+pn+ for NMOS and p+np+ for PMOS) or accumulation-mode (n+nn+ for NMOS and p+pp+ for PMOS), which requires a different kind ofmeasurement methodology.

  A literature survey showed that two techniques were proposed solely for SO1 devices [1,2]. Henderson et al.'s method [l] is for an accumulation-mode (n+nn+), fully-depleted, floating NMOS device and uses l/Gm vs. Vg in inversion (slope) and l/GmZ vs. Vg in depletion (Vg axis intercept) at Vbs=O V to get Na. This method is not applicable to the enhancement-mode. Park et ails method [2] needs both enhancement-mode (n+pn+) and accumula- tion-mode (p+pp+), fully-depleted, floating CMOS devices. This is theoretically very complicated and measures many parameters (IV and CV) based on the values of Tsi and ambiguous V, (=f(Qs Qk, Q ox, etc.)). They assumed p concentrations of the two different devices are the same. This method is not practically feasible, either. Therefore, a new meth- odology to measure the doping concentration ofthe enhancement-mode, SO1 devices easily and accu- rately is necessary.

  To overcome these two material restrictions, a simple technique, a modified version ofthe so-called "Threshold Voltage Method" [3,4], was developed

using a large-geometry (Lee1 pm), body-tied, H-gate NMOS device and additional positive body bias. However, this technique requires very speci- fied sample preparations. As diagrammed in Fig. 1-A for the NMOS case, if the same PO4 mask is used during both NMOS Vt implantation for p dop- ing and PMOS source/drain pf implantation, intrin- sick regions which have never been exposed to implanted boron atoms exist at the mask bounda- ries. These intrinsic regions between p and pf doped regions have such high resistance that decent elec- trical contacts between them are impossible to obtain. Because of this, Vt is not sensitive to body bias change, uncharacteristically high body bias is required to see measurable Vt changes, and thus data are not correct. On the other hand, if a larger P25N maskset is used during NMQS Vt implanta- tion (Fig. l-B), these intrinsic regions are removed with p doping a...