From: Jimmy Chang, Senior Development Engineer

cc: Sia Parsa, Process Engineering Manager

Subject: 2008 Year-End Report

Date: 15 January 2009

The highlights of my work in 2008 were: startup of three CENTURA Chambers - Metal Etch Chamber, Post Etch Strip Chamber, and Cooling Chamber; startup of Picosun Atomic Layer Deposition Tool; and special process developments and support for a BLMA company. The yearly summer intern program was successful. The internist has been exposed to many aspects of the semiconductor processing, with emphasis on Chemical Mechanical Polishing. The number of Engineer Testing Request has been increasing over the year, and so is the degree of difficulty, e.g. SiGe deposition on square glass plates, phosphorus doping of extra thin square substrates for solar panel.

Startup of CENTURA Metal Etch, Post Etch Strip, & Cooling Chamber

Applied Materials’ CENTURA platform is a cluster tool for multi-process capability. In the industry, it is has the excellent reputation of robot reliability. Our CENTURA system had a DPS chamber, dedicated for deep silicon etch, an MxP+ chamber for deep submicron oxide etch, and a flat finder chamber when it was first installed. Last year, with the ingenuity of our equipment staff and the assistance of Applied Materials’ field support, a DPS chamber for metal etch (aluminum film and associated cap layers, e.g. Ti, TiN), a strip chamber for post metal etch passivation and photo-resist strip, and a cooling chamber were installed. It completed the full function of the CENTURA system with all its six process slots utilized.

Extensive tests were performed to characterize the newly installed chamber processes. The results are showed in Table I for Aluminum processes; and in Table II for passivation and photo-resist strip processes.

CENTURA-MET Aluminum Etch Process Characterization
	Main Etch	Over Etch
Etch Rate (Å/min)	6744	4361
Non-Uniformity
Within Wafer	3.38%	3.60%
Wafer to Wafer	0.77%	2.97%
Selectivity
LTO (Tystar12)	9.0	6.7
PSG (Tystar12)	8.4	6.6
DUV Photo Resist	2.1	2.0
I-Line Photo Resist	2.1	1.7

Table I. CENTURA-MET Aluminum Etch Process Characterization Results

Test wafers for aluminum etch process were prepared using the standard process of CPA sputtering system that deposits 6 to 7 kÅ of Aluminum film. Under the Aluminum film, there was a layer of thermal oxide (1 kÅ) that insulated the film from the Silicon substrate for electric measurement. It also made post-etch visual inspection easier because of the oxide layer’s distinguish purplish color. For oxide selectivity tests, LTO and PSG films were prepared using Tystar12. For photo-resist selectivity tests, DUV and I-Line resists were used.

The Aluminum film thickness was converted from the film sheet resistance measured using four-point-probe (4PTPRB). This technique was developed by the summer internist (see Microlab Annual Report 2007). Two patterned wafers were used to verify the conversion accuracy. They were etched first. The film thickness was then measured using step profiler ASIQ. The difference of the average film thickness between two methods was less than 2%. The sheet-resistance conversion method saved both time and cost for photo-lithography patterning.

All the test results were well within the specification provided by Applied Materials. One interesting result was that the oxide selectivity for the over etch step is lower than the main etch step. The reason was that the over etch step needed to clear some cap layers, e.g. Titanium, which was more difficult to etch. However, our test showed that both main etch and over etch process can clear Titanium layer of 300Å thickness.

CENTURA-STRIP Process Characterization
	Passivation	PR Strip
Etch Rate (Å/min)	6006	44954
Non-Uniformity
Within Wafer	3.89%	7.86%
Wafer to Wafer	3.02%	2.24%

Table II. CENTURA-STRIP Passivation & STRIP Process Characterization Results

After metal etch process, the exposed Aluminum surface needs be passivated and the photo-resist removed before unloading to room atmosphere. Otherwise, the process residual gases trapped in the photo-resist would react with water in the room air and form acid that would corrode the etched-out fine structures. This is extremely important for the deep submicron devices.

The STRIP process uses both steam and oxygen which can remove the etch residual much effectively than just oxygen along. The process heated the wafer to 250°C and applied over 1000W of RF power. All these features ensured the removal of all etch byproducts completely. The process also improved the electrical conductivity. The test results showed that the sheet resistance of the Aluminum film decreased by 5-10% after STRIP process.

The COOL chamber is used to lower the temperature of the hot wafer coming from the STRIP chamber. If a hot wafer is send back to the plastic cassette, it will deform the slots. Using a deformed cassette will cause problems such as: error in wafer detection, scratched wafer surface, even wafer breaking. Because the whole CENTURA system is under vacuum, the heat transfer efficiency is low. It needs the COOL chamber that uses a water-cooled chuck to cool hot wafers.

In the past, the CENTURA system was operated under manual mode for deep silicon etch and oxide etch. Because the whole Aluminum etch process requires the use of four chambers: Flat finder, MET etch, STRIP, and COOL, manually moving the wafer is cumbersome. Automatic robotic sequences have been set up and tested successfully.

Our baseline engineer has used the tool to re-etch the ASML assignment marks. All main etch parameters has been checked. Some other detailed parameters, e.g. sidewall profiles, require cooperation with photo-lithograph staff. The manual has been drafted and under management review. All the three newly installed chambers should be available for lab member to use in January 2009.

During the process characterization tests, a few problems listed below were found. I worked with our equipment staff and Applied Materials’ field service and all problems were resolved.

· Metal etch chamber: Gate valve slipped during etch process.

Chamber dome overheated during long etch.

· Post etch strip chamber: Oxgen pressure too low.

Faulty DI water fittings for steam generator.

· Cooling chamber: First wafer failed initialization after long period of idling.

Startup of Picosun Atomic Layer Deposition Tool

Atomic Layer Deposition (ALD) process is well known for it superior uniformity, step coverage, and low temperature requirement. Microlab acquired an ALD tool from Picosun last year. It is designed for metal oxide deposition. The precursor vapor, which contains organic molecules with desired metal atoms, is injected into the process chamber using a precisely controlled pulse valve. The precursor vapor then absorbed (chemically or physically) on the wafer/sample surface. Due to the design of the precursor chemistry, the absorbed film is only one molecule in thickness all over the wafer/sample surface. After un-reacted precursor vapor is purged away, the oxidizing vapor is introduced into the chamber. It reacts with the organic part of the absorbed molecules and oxidizes the metal atoms to form a single atomic layer of oxide. The process then repeats itself until the desired metal oxide film thickness is reached.

	Al2O3	TiO2
Process Temperature	300°C	280°C
Precursor	TMA (Trimethylaluminum)	TTIP (Titanium Tetrakis Isopropoxide)
Precursor Temperature	Room Temperature	80°C
Oxidizer	DI Water Vapor	DI Water Vapor
Pulse/Purge Time	0.1/4 seconds	0.2/4 seconds
Deposition Rate	1 Å/cycle	0.25 Å/min
Non-uniformity	<2%	<5%

Table III. Picosun ALD Process Parameters & Test Results

The new Picosun ALD is set up for Aluminum oxide and Titanium oxide processes. The deposition process is fully automatic using a touch screen PLC controller. To decrease the wafer/sample loading/unloading time, and to keep the process chamber clean, a load lock has been installed. It needs some manual controls.

Both Aluminum oxide and Titanium oxide processes were developed. The process parameters and the test results are shown in Table III. Several lab members have been trained to use the tool. The films were tested by lab members using Surface Charge Analyzer, HF vapor etch, and IV characterization with good results.

Summer Internship – Chemical Mechanical Polishing Characterization

Each summer, Microlab offers internship to high school female students. The purpose of the internship is to expose them to engineering environment so they are interested in choosing engineering major in college.

The intern project for past year was to characterize the effects of process parameters of Chemical Mechanical Polishing tool. Our CMP tool is an industrial production machine. But unlike the production environment, which is continuous operation, it may stays in idle mode for days. Furthermore, because various films are allowed to be polished, the polishing parameters need to be adjusted accordingly based on the research requirements. The project results could provide a guideline for lab members to achieve best polish condition for their applications.

In the test wafer preparation stage, the internist learned the operation of sinks for wafer cleaning, LPCVD furnaces and P5000 TEOS system for oxide and poly-silicon deposition. A numerical model was developed on the EXCEL spreadsheet to show the wafer movement related to polish table and wafer carrier’s rotation speed. A matrix based on Statistical Design of Experiment was used for all the CMP tests. The internist presented the results successfully using Power Point software in a monthly Microlab staff meeting.

The internist also tested a new type of Chemical Mechanical Polishing (CMP) pad provided to Microlab by a vendor. The pad is non-conventional. Instead of using felt-like composite material, it uses high-density plastic bumps on the pad. The design is similar to the polishing rods used in post CMP cleaning tools. However, the new CMP pad broke wafers at an extremely high rate and the test was aborted.

Engineer Test Requests

An important BLMA company was developing a proprietary product in Microlab. All the process staffs were involved in the project. In my area, I developed and supported the deposition of stacks of various standard and non-standard films. I also in charged of the etching processes that demanded extremely high selectivity.
Deposition of Si-Ge alloy on square glass substrates. Processed three batches of 12 each with 3 different alloying levels, plus test runs.
Using ASMT to coat a protective lay of transparent film on photo-sensing chips (multiple requests for over 40 chips)
POCl3 doping of ultra-thin solar panel substrates using Tystar13.
Multiple batches of several boxes ultra-clean oxide wafers and Surface Charge Analyzer measurements.
Reworked an MEMS Exchange run: Strip poly-silicon (processed wrongly by other lab) with minimum lose of oxide under layer; re-deposit correct poly film.
Many regular requests for Low Stress Nitride, Low Temperature Oxide, Poly and Amorphous Silicon, and PECVD oxide.

Process Support & Miscellaneous

Provided general process support to lab members.
Working with equipment staff in trouble-shooting furnaces and plasma etchers.
Graded equipment quizzes, train and/or qualified lab members on various tools.
Lead monthly laboratory safety tours. Substitute in the processing session of the orientation as needed.
Wrote, contributed, or revised the following equipment manuals:

Chapter 7.x - Centura MET Chamber (Draft)

Chapter 6.24 - Picosun ALD Atomic Layer Deposition (Draft)

Chapter 2.2 - Dummy Wafer Preparation and Rework for Tystar LPCVD Furnaces and Lam Etchers

Chapter 5.2 - Tystar2 MOS Clean Dry/Wet Oxidation & Anneal Atmospheric Furnace

Chapter 6.29 - Oxford Plasmalab 80plus PECVD System

Chapter 4.28 - Matrix 106 Resist Removal System