Using Corrosion Coupons in your Data Centre
The concept of using corrosion coupons to characterise potentially damaging data centre environments, was first introduced in the Instrument Society of America Specification ISA-71.04-1985, Environmental Conditions for Process Measurement and Control Systems: Airborne Contaminants. This standard was updated in 2013 although the basic procedure, still followed, dates back well into the last century.
This standard divided environmental corrosion risk into four bands based on the extent of corrosion of a pure copper surface over a one month exposure period. The range and identification is from mild (Class G1) to severe (Class G4) and is determined by the thickness of the corrosion film created. This concept has been adapted and extended by ASHRAE (ASHRAE 2011. Gaseous and Particulate Contamination Guidelines For Data Centers) and adopted by most major IT manufacturers to gauge the effect of this type of damage on equipment reliability. When lead was removed from solder connection technologies, ASHRAE introduced silver coupons to better reflect the changes in circuit card metallurgies.
Each of the four Classes is associated with a limit of film thickness growth for each metal over a one month interval. ASHRAE strongly advise that Class G1 is the preferred requirement for IT locations, with appropriate filtration of incoming air if this level cannot be naturally achieved.
The stated method for determining film thickness is by quantitive electrochemical/coulometric reduction of the copper corrosion film back to pure metal from which film thickness can then be determined.
Other potential methods include using a microbalance, comparative colour testing of corrosion films (Fujitsu) and dedicated crystal oscillator microbalance detectors (IBM Corrosion Management for Data Centers 2011).
All methods have their strengths and disadvantages. Film thickness monitoring needs to be a compromise between speed, accuracy, simplicity and ease of measurement. For example, accurate weight measurement in the nanogramme region requires a laboratory-based instrument in a controlled humidity environment. Samples then need to be carefully handled and transported between sampling location and laboratory, and stabilised before measurement. This is demanding and time consuming. Coulombic reduction requires careful surface preparation and handling. Neither method offers insight into the chemical nature (and hence the cause) of corrosion. Yet, you would need this information to eliminate the cause. The accuracy limitations of such methods has been highlighted/summarised here.
RITEL offers an alternative way of quantifying corrosion film thickness: that is we measure the penetration depth of high energy electrons into the corrosion coupon using an electron microscope, hence giving an indication into which 'G' level the environment falls. This is achieved by measuring the relative intensities of X-ray emission peaks for copper and silver together with those for the corrosion products (if any) that are formed. This offers a huge advantage - we can identify all the chemical species present, including those that are causing the corrosion. Further advantages are that this method requires far smaller samples which can be conveniently transported for analysis and are less susceptible to damage in handling and transit, and a far simpler cleaning process to create the initial pure metal surface.
Simultaneously with these two corrosion measurements, adjacent coupons also collect airborne particulates and fine magnetic particles. The latter which are frequently found in urban environments, have the potential to disrupt high speed digital signals (when in significant concentrations). These four coupons combined, analysed after a month in place, give an all round indication of what is circulating in data hall cooling airflows, and provide a simple and convenient environmental health check.
More information on RITEL's Corrosion Coupon Testing