 |
|
| |
Gebruik deze url om te verwijzen naar dit item / Use this url to link to this item:
http://hdl.handle.net/1979/971
|
Titel
Title | Low-k dielectric reliability in copper interconnects |
Auteur
Author | Li, Yunlong |
Promotor
Supervisor | Maex, Karen |
Co-promotor
Co-supervisor | Groeseneken, Guido |
Juryleden
Members of the jury | Froyen, Ludo
Willems, Yves
Afanasiev, Valeri
Puers, Robert
Beirlant, Jan
Mertens, Robert Pierre
Zsolt, Tõkei
Barth, Hans-Joachim |
Datum verdediging
Date doctoral defence | 21-Sep-2007 |
| Summary | The interconnect module is one of the most important processing modules in a ULSI CMOS technology. It takes care of the transmission of the signals between the active devices in semiconductor chips. In the past, aluminum was used as metal interconnection material, whereas SiO2-based dielectrics were used as insulators between the metal lines. By the continuous scaling of the critical dimensions (CD) the RC-time constants of these interconnects were becoming higher than those of the active devices. In order to decrease these delay times, copper has been introduced as interconnect metal (to reduce R), whereas lowdielectric constant (low-k)materialswere introduced as insulator (to reduce C). One of the important aspects of this introduction which needs investigation is the potentially reduced inter-metal dielectric reliability, which is caused by the soft low-k materials and the shrinking line-spacings, and thus higher electrical field, in the copper interconnect system.
In this PhD work, the reliability of inter-metal low-k dielectrics in copper damascene structures was studied. First of all, the leakage mechanisms and the properties of the metal/dielectric interfaces were investigated.
It is found that the leakage current is caused by Frenkel-Poole emission for low-k dielectricswith sealing diffusion barriers, and the possibility of Schottky emission can be excluded based on the metal/dielectric interface characterization. During a bias stress, the defect distribution within a low-k changes, by which the leakagemechanism changes to a so-calledmodified Frenkel-Poole mechanism.
Next, the time-dependent dielectric breakdown (TDDB) of low-k dielectrics was investigated. It is demonstrated that the TDDB lifetime margins of low-k dielectrics are shrinking with CD scaling and the introduction of new low-k materials. For inter-metal dielectrics, the Poisson area scaling which is commonly observed in gate oxides, is not valid in low-k dielectrics due to a strong layout dependence of the breakdown spots. Moreover, it was found that there exists an interaction between electromigration tests under high stress current densities, and the low-k degradation.
In a third part the influence of a number of critical processing steps on the low-k reliability was investigated. Three steps were identified as being critical: 1) CMP and post CMP clean; 2) metal diffusion barrier integrity; 3) low-k plasma etch/ash. The major degradation factors include copper contamination and moisture uptake. By annealing the samples with moisture, the low-k reliability can be partly restored, but not to a level which is possible to be comparable to an optimized integration scheme. |
| Samenvatting | De interconnectiemodule is een van de belangrijkste procesmodules van een ULSI CMOS technologie. Zij zorgt voor de transmissie van de signalen tussen de actieve devices in halfgeleiderchips. In het verleden werd hiervoor Aluminium gebruikt als interconnectiemateriaal, en een SiO2-gebaseerd dielectricum als isolator tussen de interconnectiebaantjes. Door de verdergaande schaalverkleining lopen de R more/meer ... |
Summary for the
general public | With the semiconductor technology advancement, various new materials are introduced into integrated chips (IC) to improve the chip performance. Low-k dielectrics are a kind of such new materials to reduce the time delay in signal transmission. However, the influence of new low-k materials on IC reliability needs detailed investigation. This PhD work investigates the fundamental reliability issues more/meer ... |
Inhoudstafel
Table of Contents | Contents
1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 State of The Art 7
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Copper Damascene Integration . . . . . . . . . . . . . . . . . . . 9
2.2.1 Single Damascene Integration . . . . . . . . . . . . . . . . 11
2.2.2 Dual Damascene Integration . . . . . . . . more/meer ... |
| URI | http://hdl.handle.net/1979/971 |
| Type | Doctoral thesis
Doctoraatsthesis |
| Accessibility | Publicly accessible |
| Toegankelijkheid | Publiek toegankelijk |
Bestanden in dit item / Files in This Item:
| Bestand / File |
Grootte / Size | Formaat / Format |
|---|
| PhD_Thesis_YunlongLi.pdf | 10513Kb | Adobe PDF | View/Open |
|
|
|
