Analysis of condition for ALD deposition of ferroelectric HZO

Abstract

Deposition of ferroelectric hafnium zirconium oxide (HZO) on semiconductor
samples with Atomic Layer Deposition (ALD) has proven to be a viable method
of production. But while the physical processes of ALD deposition is relatively
well know, there exists some gaps in knowledge about different parameters for
the ALD and the resulting depositions. This work is about analysing mainly how
the purge time and chamber temperature of a thermal ALD affects the deposited
HZO films thickness, defect density and general ferroelectric switching capabili-
ties. This involved depositing ferroelectric HZO using the different recipes, mea-
suring the thickness of the samples with an ellipsometer and testing the electric
characteristics with high frequency IV measurements. The results seemed to align
fairly well with known theory. Higher temperatures generally result in faster re-
actions but also increases the evaporation of the deposited. To counteract this,
shorter purge times should be used at higher temperatures, giving a ”sweet spot”
of temperature and purge time. Even though the results seemed to agree with the
theory to an extent, extrapolating clear linear trends were nearly impossible with
the measurements. But more importantly, all the samples showed ferroelectric
properties which was earlier hard to achieve with certain ALD depositions within
Lunds University. This speaks to the stability of the current method of deposition.
Another important result was that samples deposited with the same parameters
showed very similar thickness and electric properties, once again showing how
stable and reliable the ALD process seems to be.
Since there isn’t enough results to draw clear conclusions, further testing and
analysis is encouraged. This would mean testing other temperatures and purge
times, and perhaps also measuring other electric properties, like endurance.

Popular Abstract

It is fair to say that the microelectronics have become integral for the modern
society. Today you can’t walk down the street without seeing computer screens,
LED-lights and the vast amount of cellular almost everyone carry around in their
pockets. The devices are of course still improving as the market is always asking
for cheaper and faster products.
The improvement of the processing power for the devices the first couple of
decades since it’s invention has almost solely been possible by the down scaling
of the smallest component of a logistical circuit, the transistor. But in recent
years, this has proven harder due to the original transistor design not working well
when the devices are small enough to heavily experience quantum mechanical
disturbances. This in turn has introduced a heavier emphasis on innovation with
the physical design of the transistors and also what materials to use.
A metric that has gained a lot of traction because of the need for portable
devices is the ability to reduce power usage. This is due to the limited power
available in the batteries of the portable devices. The main way to reduce the
power consumption for computing is to minimize the leakage current for the tran-
sistors. This is hard to achieve with the current popular design of the transistor
as it carries an intrinsic lower limit on how much current will be generated from
a given voltage. To avoid this, new fundamental designs of the transistor is being
explored.
An electric property that might be of use for the transistors of tomorrow is the
ferroelectric property. In short a ferroelectric material is a material that can retain
a electric field after being exposed to an external electric field. This material can
also shift its electric field back and forward, making it possible to perhaps be used
as a method of storing information.
The ferroelectric phenomenon has been known for a long while but the ma-
terial that exhibited ferroelectric switching did not retain these properties when
scaled to the size of modern microelectronics. Thankfully the material hafnium,
which is currently being implemented as a material for modern transistors, has
been shown to be able to achieve ferroelectric properties during certain condi-
tions.
How to deposit hafnium in large stable processes is being debated, and one
very promising deposition method seems to be the atomic layer deposition (ALD).
The different metrics of how to deposit the hafnium using ALD is what is being
analysed in this work.
2

• author: Qvint, Teodor
• supervisor: Lars-Erik Wernersson
• organization: Department of Electrical and Information Technology
• course: EITM01 20221
• year: 2022
• type: H2 - Master's Degree (Two Years)
• subject: Technology and Engineering
• keywords: HZO, Ferro, ALD, Temperature, Purge time
• report number: LU/LTH-EIT 2022-871
• language: English
• id: 9092396
• date added to LUP: 2022-06-27 15:43:05
• date last changed: 2022-06-27 15:43:05