Do dopant diffusion and drift decide semiconductor device degradation and dimension limits?

Igor Lubomirsky, David Cahen

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

We explore chemical and physical limits to semiconductor device miniaturization. Minimal sizes for space charge-based devices can be estimated from Debye screening lengths of the materials used. Because a doped semiconductor can be viewed as a mixed electronic-ionic conductor, with the dopants as mobile ions, dopant intermixing across a p/n junction presents a chemical limit. Given a desired lifetime, simple relations can be derived between size and dopant intermixing for reverse- or forward-biased devices. Mostly, conditions for significant dopant mobility are far from those where the material is used. Thus, it is generally held that elemental and III-V-based p-n junctions are immune to this problem and persist because of kinetic stability. Indeed, we find this to be so for Si in the foreseeable future, but not for III-V- and II-VI-based ones. The limitation is more severe in structures with very thin undoped layers sandwiched between doped ones or vice versa, where even 1% intermixing can be critical. This decreases lifetime nearly 100 times. For example, for structures containing a 10 nm critical dimension, none of the components can have an average diffusion coefficient higher than 10-24 cm2/s for a 3 year lifetime. Ways to overcome or mitigate this limitation are indicated.

Original languageEnglish
Pages (from-to)559-565
Number of pages7
JournalSolid State Ionics
Volume136-137
DOIs
Publication statusPublished - Nov 2 2000

Fingerprint

Semiconductor devices
semiconductor devices
Doping (additives)
degradation
p-n junctions
Degradation
life (durability)
miniaturization
space charge
screening
diffusion coefficient
conductors
Electric space charge
Screening
kinetics
Ions
Semiconductor materials
electronics
Kinetics
ions

ASJC Scopus subject areas

  • Electrochemistry
  • Physical and Theoretical Chemistry
  • Energy Engineering and Power Technology
  • Materials Chemistry
  • Condensed Matter Physics

Cite this

Do dopant diffusion and drift decide semiconductor device degradation and dimension limits? / Lubomirsky, Igor; Cahen, David.

In: Solid State Ionics, Vol. 136-137, 02.11.2000, p. 559-565.

Research output: Contribution to journalArticle

@article{3a0f165b88a643bf81468fdab62a25be,
title = "Do dopant diffusion and drift decide semiconductor device degradation and dimension limits?",
abstract = "We explore chemical and physical limits to semiconductor device miniaturization. Minimal sizes for space charge-based devices can be estimated from Debye screening lengths of the materials used. Because a doped semiconductor can be viewed as a mixed electronic-ionic conductor, with the dopants as mobile ions, dopant intermixing across a p/n junction presents a chemical limit. Given a desired lifetime, simple relations can be derived between size and dopant intermixing for reverse- or forward-biased devices. Mostly, conditions for significant dopant mobility are far from those where the material is used. Thus, it is generally held that elemental and III-V-based p-n junctions are immune to this problem and persist because of kinetic stability. Indeed, we find this to be so for Si in the foreseeable future, but not for III-V- and II-VI-based ones. The limitation is more severe in structures with very thin undoped layers sandwiched between doped ones or vice versa, where even 1{\%} intermixing can be critical. This decreases lifetime nearly 100 times. For example, for structures containing a 10 nm critical dimension, none of the components can have an average diffusion coefficient higher than 10-24 cm2/s for a 3 year lifetime. Ways to overcome or mitigate this limitation are indicated.",
author = "Igor Lubomirsky and David Cahen",
year = "2000",
month = "11",
day = "2",
doi = "10.1016/S0167-2738(00)00414-8",
language = "English",
volume = "136-137",
pages = "559--565",
journal = "Solid State Ionics",
issn = "0167-2738",
publisher = "Elsevier",

}

TY - JOUR

T1 - Do dopant diffusion and drift decide semiconductor device degradation and dimension limits?

AU - Lubomirsky, Igor

AU - Cahen, David

PY - 2000/11/2

Y1 - 2000/11/2

N2 - We explore chemical and physical limits to semiconductor device miniaturization. Minimal sizes for space charge-based devices can be estimated from Debye screening lengths of the materials used. Because a doped semiconductor can be viewed as a mixed electronic-ionic conductor, with the dopants as mobile ions, dopant intermixing across a p/n junction presents a chemical limit. Given a desired lifetime, simple relations can be derived between size and dopant intermixing for reverse- or forward-biased devices. Mostly, conditions for significant dopant mobility are far from those where the material is used. Thus, it is generally held that elemental and III-V-based p-n junctions are immune to this problem and persist because of kinetic stability. Indeed, we find this to be so for Si in the foreseeable future, but not for III-V- and II-VI-based ones. The limitation is more severe in structures with very thin undoped layers sandwiched between doped ones or vice versa, where even 1% intermixing can be critical. This decreases lifetime nearly 100 times. For example, for structures containing a 10 nm critical dimension, none of the components can have an average diffusion coefficient higher than 10-24 cm2/s for a 3 year lifetime. Ways to overcome or mitigate this limitation are indicated.

AB - We explore chemical and physical limits to semiconductor device miniaturization. Minimal sizes for space charge-based devices can be estimated from Debye screening lengths of the materials used. Because a doped semiconductor can be viewed as a mixed electronic-ionic conductor, with the dopants as mobile ions, dopant intermixing across a p/n junction presents a chemical limit. Given a desired lifetime, simple relations can be derived between size and dopant intermixing for reverse- or forward-biased devices. Mostly, conditions for significant dopant mobility are far from those where the material is used. Thus, it is generally held that elemental and III-V-based p-n junctions are immune to this problem and persist because of kinetic stability. Indeed, we find this to be so for Si in the foreseeable future, but not for III-V- and II-VI-based ones. The limitation is more severe in structures with very thin undoped layers sandwiched between doped ones or vice versa, where even 1% intermixing can be critical. This decreases lifetime nearly 100 times. For example, for structures containing a 10 nm critical dimension, none of the components can have an average diffusion coefficient higher than 10-24 cm2/s for a 3 year lifetime. Ways to overcome or mitigate this limitation are indicated.

UR - http://www.scopus.com/inward/record.url?scp=16744367376&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=16744367376&partnerID=8YFLogxK

U2 - 10.1016/S0167-2738(00)00414-8

DO - 10.1016/S0167-2738(00)00414-8

M3 - Article

VL - 136-137

SP - 559

EP - 565

JO - Solid State Ionics

JF - Solid State Ionics

SN - 0167-2738

ER -