Characterizing the destructive potential of corrosive contaminants
By Chris Muller, technical director of Purafil
The control of airborne molecular contamination (AMC) continues to grow as a requirement for all advanced semiconductor manufacturing. With more fabs using copper processing and 300 mm wafers, and as we move into the 65 nm technology node, the requirements for AMC control are becoming more stringent. Most AMC control specifications call for levels to be at or below 1 part per billion (ppb) for the target contaminants, and verifying attainment and maintenance of these levels can be an expensive proposition if real-time monitoring is an option being considered.
 |
There are a number of real-time monitoring technologies available that can measure at low- to sub-ppb levels; however, they can cost upward of $20,000 per point in a facility monitoring system. Due to the large investments required in terms of equipment, supplies, and personnel, manufacturers are looking for lower-cost AMC monitoring options that can still provide information relevant to the protection of processes and materials.
There are a number of “semiquantitative” analysis techniques being used for AMC monitoring. This refers to those analytical techniques that provide quantitative information on environmental air quality, but do not measure specific contaminants. One of these techniques involves the use of passive or real-time reactivity monitoring.
Reactivity monitoring characterizes the destructive potential of an environment with respect to corrosive contaminants. The growth of corrosion films on specially prepared copper, silver, and other metal-plated sensors provides an indication of the type and level of essentially all corrosive chemical species present in the local environment. Both passive and real-time corrosion monitors are currently available, and each can be used to gather information on gaseous contaminants and their levels in the environment.
A current method for real-time reactivity monitoring utilizes a quartz crystal microbalance (QCM) coated with a reactive metal that is then attached to an oscillator.1 As the metal corrodes, the oscillation frequency of the crystal decreases, and this frequency change can be converted to a corrosion film thickness corresponding to severity levels applicable to semiconductor manufacturing2 and other comparable industry standards.3, 4, 5
Environmental reactivity monitors (ERMs) using QCM technology are used to measure the damaging potential of the environment toward processes and products from, for example, sulfur compounds and inorganic acids. They provide real-time data concerning the level of AMC present in the local environment by measuring the mass accumulation on the sensors resulting from the reaction of corrosive contaminants with the metals. The mass increase is described in terms of the corrosion film thickness measured in angstroms (Å). This is a highly sensitive method that can provide the reactivity data for contaminant levels of less than or equal to 1 ppb.
A primary use of ERMs has been for the development of AMC baseline data and to verify attainment of specified cleanliness levels in areas where wafers are handled, stored, or processed. They allow for monitoring the performance of chemical filtration systems installed in air-handling units (see Figure 1). A classification system has been developed to correlate film thickness to environmental air quality, and to define acceptable levels of AMC inside a facility. 6, 7 These air-quality classifications are shown in Figure 2.
 Figure 2: Reactivity monitoring guidelines |
It has been established that air quality in the wafer environment should be Class C2/S2 or better, with no evidence of active sulfur or inorganic chlorine contamination. Where monitoring results indicate Class 1, there is little else that can be done, economically, to improve the environment. These standard classifications are currently being used as part of several fabs’ overall contamination-control programs to assess and verify AMC levels.
The ERM described here is generally suitable for measuring and detecting low levels of AMC in the outdoor air and in the ambient cleanroom environment; however, more precise measurements will be required for advanced semiconductor manufacturing and as the wafer environment continues to shrink.8 Although these ERMs have shown sensitivities to various chemical species at levels at or below 1 ppb, there are many applications where airborne AMC limits are moving into the part-per-trillion (ppt) range. Thus, reactivity monitoring could be more widely applied if detection levels could be reduced by an order of magnitude or more.
Even as real-time reactivity monitoring is being integrated into the contamination-control programs of many manufacturers, the next generation of reactivity monitors is now evolving. Advancements in materials and manufacturing techniques have provided for the development of a new, more sensitive, and lower-cost reactivity monitoring technique based on the use of microcantilever sensor elements. This monitor has the potential to significantly lower detection limits for the same corrosive species.
 Figure 3. Pictured here is 150-micron-long cantilever containing a vibration sensor. Photo courtesy of LG Electronics Institute of Technology. |
Cantilever structures are the simplest micro-electromechanical systems (MEMS) that can be easily micromachined and mass-produced. The microcantilever beam (Figures 3 and 4) is a member of the class of electromechanical sensors that includes the quartz-crystal microbalance (QCM), the surface acoustic wave (SAW) device, and other resonating sensor structures.
Microcantilever sensors (MCS) have many advantages over conventional sensors, chief among them is their size, extreme sensitivity, reaction time, and low-power consumption. For sensors that require the measurement of resonant frequency, microcantilevers have a much higher sensitivity because of the low mass and thickness compared to QCM and SAW sensors.
An MCS oscillated at its resonance frequency can be used as a mass detector. AMC (for example, sulfur dioxide and chlorine) can react with a metal coating on the cantilever, and the mass change can be determined from the shift in resonance frequency. This mass change can, in turn, be converted into an output signal indicative of the reactivity level of the local environment-typically measured in angstroms for consistency with the classifications in Figure 2.
A number of metallic coatings are already being used in other electromechanical sensors, and can be readily employed on their much smaller microcantilever cousins. Detection limits for gaseous contaminants have yet to be fully explored, but low parts per billion to parts per trillion are realistic. The fact that the cantilever can be readily produced with sub-micron thickness favors a high sensitivity where, for example, 1 Hz in resonance frequency can correspond to a mass change of 1 picogram. 9
Experimental corrosion sensor
A project was initiated to develop a working microcantilever-based corrosion sensor that would employ, as a minimum, copper- and silver-coated microcantilevers in a distributed load configuration operating in dynamic mode.10 (Nanocantilevers were considered for this but were ultimately rejected in favor of microcantilevers, which were readily available, as were the prototyping services.)
Several microcantilevers were chosen for evaluation, and an example of one design is shown in Figure 5. This is a commercially available non-contact (vibration) mode atomic force microscopy (AFM) probe that operates at a resonance frequency range of 32 to 35 kHz and has a minimum detectable mass density of less than 1 ng. A proprietary process was used to apply solid-phase materials (such as copper) to the microcantilever.
 |
A simple detection circuit was built that included the microcantilever shown in Figure 5, an oscillator, a processor, and an EEPROM, as well as a power supply to supply current to the components arranged in the circuit. The oscillator was used to detect and receive a signal associated with the reaction of the microcantilever to a corrosive gas. The shift in resonance frequency due to mass loading can be calculated and was shown to be linear with the frequency change over the selected range.11 Examples of calculated frequency shifts versus mass loading are shown in Figures 6A and 6B.
 |
The sensor was exposed to a corrosive gas in the range of 1 to 2 ppb generated by permeation devices. Over time, this resulted in a mass gain due to the formation of copper sulfide reaction products on the surface of the sensor. This was converted to corrosion film thickness, in angstroms, and an example of the data obtained is shown in Figure 7.
 |
A number of commercially available microcantilevers of various dimensions and geometries are currently being evaluated for their mass sensitivity over a given frequency range in order to optimize their dose response characteristics for low levels (less than or equal to 0 ppb) of a number of corrosive gases. Detailed information on this evaluation and experimental data for the candidate sensor will be presented in a future report, along with details of the prototype corrosion monitor.
Technologies currently available could be used to produce cantilevers, analog processing, and even data transmission on a single chip. Small size provides the ability to be employed in microenvironments, which will be a primary application area for this type of sensor. Information about the corrosive nature of an environment can provide the basis of warranties for electronic process and control equipment. Board-level or chip-level corrosion monitors based on microcantilevers can find application in almost any electronic device or any processor-based device.
As microelectronic and semiconductor devices continue to decrease in size to the deep sub-micron range, their susceptibility to damage or failure due to exposure to low- and sub-parts per billion levels of corrosive gases will continue to increase. This requires the development of a more sensitive reactivity monitor to protect these devices and to help assure productivity and profitability. The microcantilever sensor has shown promise as being able to meet these requirements.
References
1. England, W.G., et al., “Applications of a Real-Time Electronic Contact Corrosion Monitor,” Proceedings of Advances in Instrumentation and Control, Vol 46: pp 929-955, Instrument Society of America, Anaheim, 1991.
2. Muller, C., “Airborne Molecular Contamination,” Semiconductor Manufacturing Handbook, ed. M.H. Geng, McGraw-Hill 2004.
3. Standard: ANSI/ISA S71.04-1985, “Environmental Conditions for Process Measurement and Control Systems: Airborne Contaminants,” Instrument Society of America, Research Triangle Park, NC, 1985.
4. Standard: IEC 60654-4 (1987-07) “Operating Conditions for Industrial-Process Measurement and Control Equipment. Part 4: Corrosive and Erosive Influences,” International Electrotechnical Commission, Geneva, Switzerland, 1987.
5. Standard: JEIDA-29-1990, “Standard for Operating Conditions of Industrial Computer Control System,” Japan Electronic Industry Development Association, Tokyo, 1990.
6. Muller, C., “Evaluating the Effectiveness of Airborne Molecular Contamination Control Strategies with Reactivity Monitoring,” Journal of the IEST, Volume 45, 2002 Annual Edition.
7. Stanley W.B.M. and Muller C.O., “Corrosive Gas Kills Product: Reactivity Monitoring Proves Environment and Tracks Episodes,” InTech, Vol 50, No. 6, pp. 30-33, June 2003.
8. http://www.itrs.net/Common/2005ITRS/Yield2005.pdf
9. http://www.zurich.ibm.com/st/nanoscience/mass.html
10. Lang, Hans Peter, Hegner, Martin, and Gerber, Christoph, “Cantilever Array Sensors,” Materials Today, Vol. 8, No. 4, April 2005, pp. 30-36, London.
11. Thundat, Thomas G., et al., “Microsensors to Monitor Missile Storage and Maintenance Needs,” U.S. Department Of Energy Contract DE-AC05-84OR21400, Oak Ridge National Laboratory, Oak Ridge, TN, October 1997
Christopher O. Muller is the technical director for Purafil, Inc. in Doraville, Georgia, a manufacturer of gas-phase air-filtration media, filters, equipment, and air-monitoring instrumentation. He is responsible for technical support services and various research and development functions. Prior to joining Purafil, he worked in the chemical process and pharmaceutical manufacturing industries in quality assurance/quality control. He has written and spoken extensively on the subject of environmental air quality and the application/use of gas-phase air filtration, and counts over 90 papers and articles, more than 20 seminars, and seven handbooks to his credit. He can be reached at [email protected].