Abstract

Surface Acoustic Wave (SAW) Technology technology is an effective means to quantify the performance of gas-phase chemical filters used to reduce surface molecular contamination (SMC). Implementing chemical filtration technology to tackle SMC issues can stimulate process yield, whether that implementation takes place at the wafer, equipment, or facility level. Without monitoring, chemical filters will either be replaced while still effective, resulting in filter costs that are unnecessarily high, or replaced after the filter is spent, resulting in potential yield loss. Quantitative analysis of molecular contamination is essential in verifying the effectiveness, efficiency, and lifetime of chemical filters used to reduce SMC. Most importantly, this analysis will also help determine what is the acceptable (or non-acceptable) level of molecular contamination for a particular process, area, or facility. Particle Measuring Systems' AiM® uses SAW technology to measure, in real-time and with high sensitivity, the effect of molecular contamination as it interacts with the product surface.

Introduction

The time for gas-phase chemical filtration is present or looms ominously on the horizon for most major aspects of semiconductor manufacturing. SMC is representing a larger percentage on the yield loss pareto chart. Because SMC has the ability to create very thin chemical coatings on product surfaces, as well as potentially react with that surface, it is important to minimize the impact to the product surface. One of the most effective means to minimize this damage is by filtering SMC through the use of gas-phase chemical filtration. The removal of AMC from the airborne environment is accomplished through physisorption (physical adsorption) or chemisorption (adsorption via chemical reaction) with the filter media. This filtration can occur at the wafer, equipment, and even the facility level. Chemical filters are commonly used on SMIF/FOUP pods and wafer stockers to minimize SMC at the wafer level. Original Equipment Manufacturers (OEM) are implementing chemical filtration into their equipment to protect wafers during processing. A common example of equipment level filtration is on photolithography steppers and scanners. To protect wafers while in the queue to be processed, some fabs are incorporating chemical filtration into both makeup and recirculation air handlers.

In addition to incorporating gas-phase chemical filtration as a preventative measure to eliminate or reduce SMC, it is beneficial to be able to quantitatively assess the performance and lifetime of these measures. SAW technology can evaluate the performance of chemical filtration at the wafer, equipment, and facility level. Contamination concentrations that can cause product yield loss are extremely low. Therefore, it is necessary to have a sufficiently sensitive monitor to respond to small changes in gas phase concentrations. Based on the updated 2002 International Technology Roadmap for Semiconductors (ITRS), current recommended limits for residual carbon from organic contamination in a gate stack are on the order of 0.36 ng/cm2.1 For example, IPA contamination levels less than 0.3 ng/cm2 have been shown to detrimentally affect time zero dielectric breakdown (TZDB) reliability of gate oxides.2 A high sensitivity monitor must be used in order to identify and measure such low contamination levels. In addition, real-time monitoring techniques need to be utilized to minimize product damage in between samples, or in between waiting to receive data analysis from previously taken samples. Real-time monitoring will also allow for possible identification of the event source based on equipment or process logs.

Method of Operation

SAW technology operates on the principle of piezoelectricity. As an alternating voltage is applied to a quartz crystal, a standing surface acoustic wave is produced. The frequency of the acoustic wave depends upon what is interacting with the surface of the crystal, so as mass is adsorbed or chemically reacts with the SAW surface, the frequency of the wave propagation is reduced. By measuring the difference between an exposed SAW crystal and a sealed reference crystal, mass accumulation on the exposed crystal can be determined. This mass accumulation, or surface molecular contamination (SMC) provides a quantitative measurement that can be correlated to contamination effects on product surfaces. Data referenced below has been gathered using Particle Measuring Systems' AiM molecular contamination monitor. The AiM utilizes a 200 MHz SAW sensor with a sensitivity of 0.02 ng/cm2/Hz and a time resolution of 1 minute. With a 1 Hz reduction in the frequency, 0.02 ng/cm2 of mass has accumulated on the sensor surface. In addition, SAW sensors can be coated with various materials in order to better mimic the product surface at key steps in the process. Typical coatings include SiO2, copper, and silver. SiO2 is a good representation of optical component surfaces in photolithography equipment as well as witness wafers that are currently used in semiconductor manufacturing to monitor AMC. Copper and silver coatings are used to monitor acidic gas contamination, with silver more selective to SOx compounds, and copper more selective to F, Cl, and other inorganic halogens. Copper coatings can also be used to mimic production wafers for those companies who are involved in copper interconnect processing.

Gas-Phase Chemical Filtration

Selection of gas-phase chemical filtration for reducing AMC is a difficult challenge. Maximizing the number of airborne contaminants a given filter targets, choosing a filter with a long lifetime and with a high efficiency over the majority of that lifetime, and reducing the cost of ownership of chemical filter usage are important parameters to be evaluated. Selection of the appropriate filter is not the end of the story. Once a filter has been installed in a system, engineers must have confidence in the filter performance. Blind confidence based on recommendations by chemical filter manufacturers is inadequate at best, and so the engineer must have in his/her monitoring arsenal a quantitative method to evaluate these parameters. Particle Measuring Systems' AiM monitor allows engineers to monitor chemical filter performance to better understand and control their processes.

Chemical Filter Efficiency for Acidic Gas Reduction on an ASML 248 nm Stepper

The installation of gas-phase chemical filtration on process tools allows for a reduced or contamination free environment within the tool enclosures. Once a chemical filter is installed, the AiM is used to quantitatively evaluate the efficiency over the chemical filter lifetime. In order to evaluate the efficiency, the contamination rate without the use of a chemical filter must be known and can be measured at the filter intake or with no chemical filter installed. This is essentially a worst-case scenario, i.e., how much contamination could potentially be interacting with product surfaces when a chemical filter has failed or reached breakthrough prematurely. Once that value is known, efficiencies can be determined over the lifetime of an installed chemical filter. In most cases, however, chemical filters are replaced prior to this worst-case scenario, so waiting until an efficiency of near 0% is achieved is not necessarily the appropriate time to replace a chemical filter.

Figure 1 (Download this paper for all tables and figures.) (290.6 KB) below demonstrates the efficiency evaluation of a chemical filter installed on an ASML 248 nm stepper using a silver coated SAW sensor. The silver coated SAW sensor was used to evaluate sulfur oxide contamination downstream of the chemical filter. After determining the contamination rate with no chemical filter installed, the efficiency of an old chemical filter and a new chemical filter were determined.

With no chemical filter installed, the mass accumulation rate is 9.94 ng/cm2/day. The monitoring environment downstream of the old chemical filter has a contamination rate of 6.34 ng/cm2/day and a new chemical filter has an average rate of 0.23 ng/cm2/day. SOx removal efficiencies are then calculated to be 36.2% for the old chemical filter, and 97.7% for the new chemical filter (see Table 1).

In addition to determining the efficiency for the chemical filters at various stages over the lifetime, the high sensitivity of the SAW sensor allowed detection of filter material outgassing during the initial installation of the new chemical filter. This outgassing pattern was related to temperature shifts in the facility.

Filter Lifetime Evaluation

Chemical filter lifetime has traditionally been based on experimental techniques. Total time the filter has been in service, the number of wafers processed, or set monthly or yearly replacements are included in some of the filter change-out procedures used by various companies and they are based on trial and error. To replace a chemical filter prior to the end of its useful lifetime is wise because the risk to production material is low, but this can be costly based on the number of chemical filters used in a facility. Not replacing a chemical filter after its useful lifetime has expired can create significant jeopardy to product, and costs can greatly exceed that which would have been required to replace the chemical filter at the appropriate time. These are not optimal techniques because they do not provide a quantitative reason for the filter change and do not take into account variables that affect chemical filter lifetime. The most critical variable affecting chemical filter lifetime is what contaminant concentration is acceptable for a given process or facility. Higher tolerable contaminant concentrations downstream of a filter will extend the lifetime of a chemical filter. Other common variables that affect chemical filter lifetime include the contamination concentration in the air, chemical vapor pressure, affinity for chemical reaction with impregnants within the filter, residence time, temperature, pressure, moisture, etc. As these variables change over time (minutes, hours, days, months), filter performance and lifetime change as well; the ability to collect real-time data is a useful tool for monitoring these changes and their effect on product material.

By placing an AiM downstream of a chemical filter, or in between filters if used in a serial arrangement, quantitative data can be provided that will allow engineers to establish firm guidelines to justify replacing or not replacing chemical filters. Engineers will be able to access the total contamination accumulation, as well as the contamination rate and removal efficiency, in order to judge the damage that may be occurring on product surfaces.

Activated Carbon Filter Monitoring in a Mini-Environment

Activated carbon is a typical chemical filter used in semiconductor manufacturing facilities, due to the large surface area per unit mass within a carbon particle. Activated carbon filters target a wide range of contaminants such as organics, hydrocarbons, siloxanes, acidic species, etc, and chemical impregnants can be used to target other specific contaminants.

Implementing activated carbon filtration can have significant benefits in the reduction of SMC. As seen in Figure 2 (Download this paper for all tables and figures.) (290.6 KB), an activated carbon chemical filter was installed in a mini-environment that previously had not implemented chemical filtration. A background level of contamination was established prior to installing the chemical filtration, and the contamination rate was ~1.1 ng/cm2/day. Upon installing the chemical filter, the mini-environment becomes significantly cleaner. In fact, mass is actually desorbed from the SiO2 surface as the mass on the surface equilibrates with the reduced contamination levels in the mini-environment. As equilibrium conditions are reached, mass accumulation rates on the sensor surface are near zero. The sensitivity and real-time feedback from the AiM give immediate confidence that the contamination levels have been minimized due to the installation of chemical filters and that the product is being exposed to a significantly cleaner environment.

Monitoring Effectiveness of Gas-Phase Chemical Filters

As mentioned previously, a significant number of factors can be used in the chemical filter selection process. There are a number of gas-phase chemical filter manufacturers with what seems to be a limitless choice of filter media and materials. Understanding the contamination reduction needs for a given process or facility greatly reduces the choices available, but evaluations prior to purchase should consider how effective each filter is at actually reducing or eliminating these contaminants.

p>SAW sensors allow contamination mass and rate information to be gathered from the sensor surface. In addition, the sensor can then be analyzed using TOF-SIMS or another off-line analytical technique to verify if the contaminant species of interest is present on the surface of the SAW sensor. A strategy can also be implemented that uses witness wafers in conjunction with the AiM monitor's real-time data. The witness wafers are pulled and sent for thermal desorption GC-MS analysis on regular intervals or when contamination rates exceed process limits.

Chemical Removal of Organic Contamination in a TEL Diffusion Furnace

There have been many documents detailing the detrimental results of organic contamination prior to gate oxide formation.3,4 Increasingly, there is interest in chemical filter installation in diffusion areas to reduce organic contamination that may impact film growth. As mentioned previously, even very low levels of contamination can be cause for concern and parts per trillion (ppt) levels of contamination can accumulate on silicon dioxide surfaces very quickly. For example, 34 ppt of dibutylphthalate (DBP) when exposed to a SiO2 surface will exceed the ITRS recommended limits of 0.36 ng/cm2 in approximately 10 minutes. Figure 3 (Download this paper for all tables and figures.) (290.6 KB) diagrams the organic contamination monitoring locations in a TEL diffusion furnace evaluated to understand the source and potential risk of organic contamination to oxide growth. Monitoring location A is in the service area and near the furnace air intake. Location B is inside the furnace near the load/unload port and evaluates the chemical filtration efficiency of the air supplied to the furnace, and location C represents the contamination environment outside the furnace in the process area near the load/unload port. Table 2 (Download this paper for all tables and figures.) (290.6 KB) summarizes the contamination accumulation rates for each monitoring location.

Location A represents the largest risk for the introduction of molecular contamination to the wafers processed through this TEL furnace, indicated by a very large contamination rate as measured on a SiO2 coated SAW sensor. Location B is an appropriate location for monitoring the efficiency and lifetime of the chemical filters installed on this equipment. The efficiency of the installed chemical filters during the evaluation was 85% (assuming all air in the furnace comes from the service area). If the chemical filters fail or are not quantitatively monitored, significant production material may be in jeopardy as service air may interact with the product surface. Location C is an extremely clean environment and indicates that wafers can be transported and exposed in this location without serious threat to the product surface from organic contamination.

Facility Monitoring of Chemical Filters

Chemical filters are being integrated into large sections of facilities for a number of reasons. Processing a semiconductor device (integrated circuits, MEMs, etc.) from wafer start to packaging can take significant amounts of time, from weeks to many months. Fabs are installing chemical filters into make-up and recirculated air handling systems due to the prolonged exposure of wafers while progressing through the lengthy number of steps needed to become a sellable product. In addition, certain process areas can be very sensitive to molecular contamination. Copper CMP processes expose the wafers to damage from corrosion; instead of installing gas-phase chemical filtration on individual tools, a CMP bay may invest in filtration for the entire area.

A diffusion and photolithography area were examined with an AiM during chemical filter replacements for reduction in molecular contamination within these bays. Data gathered during the first month was used to evaluate the baseline conditions, and as shown in Figure 4, both areas were similar in that the contamination accumulation rate was moderate and stable. At the beginning of the second month, a carbon chemical filter was replaced in the makeup air in the photolithography bay only. At the beginning of the third month, secondary carbon chemical filters were replaced in the recirculation air in both the photolithography and diffusion bays. As a result of the chemical filter replacement, the contamination rate decrease in the photolithography bay over the entire evaluation was 35%, and the decrease in the diffusion bay was 20%.

As shown in Figure 4 (Download this paper for all tables and figures.) (290.6 KB), monitoring a cleanroom environment with an AiM provides real-time molecular contamination data for product wafer exposure. The AiM can identify and quantify changes in contamination accumulation and rate conditions due to modifications in chemical filtration, airflow rates and patterns, relative humidity, temperature, etc., at the facility level.

Summary

Surface acoustic wave technology offers a real-time, high sensitivity monitor that allows quantitative measurements to be made for chemical filter effectiveness, efficiency, and lifetime. Locating SAW sensors downstream of chemical filters, or in-between sequential chemical filters, provides information that allow engineers to determine and monitor filter performance, and optimize filter lifetime. Through the use of SAW technology, Particle Measuring Systems' AiM contributes to cost savings by quantifying chemical filtration performance parameters and identifying contamination events which can be investigated and corrected leading to yield improvement.

(Download this paper for all tables and figures.) (290.6 KB)

AiM® is a registered trademark of Particle Measuring Systems, Inc.

Contact us if you need more information or have questions.

Learn more about our microcontamination monitoring services. **References((

1. 2002 Updated International Technology Roadmap for Semiconductors, Front End Processes, 52.
2. Motai, K., et. al., The effect of Isopropyl Alcohol Adsorption on the Electrical Characteristics of Thin Oxide, Jpn. J. Appl. Phys., 37 (1998)1137-1139.
3. Sugimoto et. al., Influence of Organic Contamination on Silicon Dioxide Integrity, Jpn. J. Appl. Phys., 39 (2000) 2497-2502.
4. De Gendt et. al., Impact of Organic Contamination on Thin Gate Oxide Quality, Jpn. J. Appl. Phys., 37 (1998) 4649-4655.
5. Akiyama, T., Takahashi, H., and Gomi, H., Monitor System for Clean Dry Environment Using Surface Acoustic Wave Sensor. 21st Annual Tech. Meeting on Air Cleaning and Contamination Control, Japan Air Cleaning Association, Apr. 2003, 46-48.