This article was printed in CONTROL's April 2009 edition.
Greg McMillan and Stan Weiner bring their wits and more than 66 years of process control experience to bear on your questions, comments, and problems. Write to them at firstname.lastname@example.org.
By Greg McMillan and Stan Weiner, PE
In this last part of our “Secret Life” series, we bring back experts Scott Broadley and Bob Garrahy at Broadley-James Corp. (BJC) and Jim Gray and John Wright at Rosemount Analytical Inc. (RAI) to wind up our discussion of one of the grand old tools of process control—the pH electrode.
Stan: In our first book, How to Become an Instrument Engineer, we had a chapter called “How Not to Do Packaged Equipment.” It’s amazing that 30 years later some of it is still applicable today.
Bob: Many of the early OEM electrodes were of poor quality, rather like car radios until the 1990s. Others were simply “one size fits all” and often ill-suited for particular applications. OEM suppliers eventually got better at supplying what customers wanted, so many of the sales and service providers have gone away.
Greg: What happens to glass electrodes as they age?
Jim: Alkaline ions (or lithium ions) continually move out of the glass. The leached gel layer on the exterior surface of the glass electrode increases in thickness, which increases the response time and resistance of the glass measurement. The time to 98% of the final response is dramatically longer for an aged electrode. While there is some reduction in span (efficiency), often the user mistakenly measures a much shorter than actual span during calibration because the electrode is not left in the buffer solution long enough to see the full response.
Stan: We’ve seen improvements in the longevity, durability and ease of installation of pH electrodes. What’s next?
Bob: The future is factory calibration and the “Holy Grail” of a constant reference potential.
John: The details (date, offset and span adjustments) of factory calibration and subsequent calibrations will be stored on a chip in the electrode. We can then get a pretty good idea of what is or is not going on from that data.
Scott: Wireless transmission will eliminate wiring termination mistakes and ground loops. Only about 10% of installations suffer from ground potentials and noise, but these problems are difficult to find. When we connect a lab meter to the electrode, the noise or shift in pH often goes away, indicating we could get rid of the problem by eliminating the wiring.
Bob: Ground problems often go away when an electrode connected to a wired transmitter is inserted in a glass beaker containing the process solution. The problem reappears when a wire in the solution is connected to a local metal surface. This test indicates a ground loop via the metal process piping or vessel wall. Agitator and pump motors are often the source of noise.
Greg: In a recent comparative test of conventionally wired versus new WirelessHART pH transmitters on a single-use bioreactor, we had some interesting test results.
Scott: The bioreactor had a mammalian cell culture that was very sensitive to pH. We asked for a resolution setting of 0.01 pH. We mistakenly got 0.01%, which translated to about 0.001 pH.
Greg: To our surprise, we were able to control to within 0.002 pH of set point while reducing the number of communications by 60% by using exception reporting with the incredibly fine 0.001 pH resolution setting in the WirelessHART transmitter, which is beyond the stringent pH control requirement of bioprocesses. I expect that some chemical processes would need a 0.01-pH to 0.05-pH resolution setting, while a 0.1-pH resolution setting should be good enough for chemical waste treatment. In any case, the resolution setting should be greater than the noise amplitude, but less than 1/5 of the allowable deviation from set point. Coarser resolution settings correspond to a greater reduction in communications and commensurate increases in battery life.
Scott: Confirming our suspicions that noise from grounding problems could be eliminated with wireless transmitters, we found that spikes in the conventional wired transmitter did not appear in the WirelessHART transmitter pH signal.
Greg: The initial results from bioreactor wireless pH and temperature control tests can be found at www.controlglobal.com/0904_CTtests.html. An enhancement was developed for the PID algorithm called PIDPlus to provide good control despite a variable sample time. The wireless PIDPlus does about as well as wired traditional PID for bioreactor pH control for a 0.01 pH (0.1%) resolution. More impressive is the improvement shown by the PIDPlus compared to the traditional PID for glucose control and for generic self-regulating (continuous) and integrating (batch) processes where significant analyzer measurement delay has been introduced. This advance in the PID for all loops is a result of maximizing the utility of the new wireless technology. (See the Feb 9, 2009, entry at www.modelingandcontrol.com.)
Scott: Wireless measurements will also report and alert users when problems develop instead of waiting for an automated maintenance system query.
John: There will be a commensurate increase in the type of diagnostics we can provide as we get smarter about detecting and interpreting the impedance changes in the glass and reference electrode. As users become more aware of the actual problem, we’re seeing an interest in being able to fix rather than replace them.
Greg: I hope we’ll also see better options for automatic compensation of the effect of temperature on the solution pH. The standard temperature compensation offered for electrodes is for the change in millivolts generated per pH by the glass electrode per the Nernst equation. Most people are unaware that the solution’s pH also changes due to changes in the water and acid and base dissociation constants with temperature. For strong acids and bases, their respective dissociation constants are well beyond the normal pH range. For these processes, what we see is mostly the effect of temperature on the water dissociation constant that can be approximated to be -0.03 pH/oC for basic solutions.
Stan: Whether a customer sees more glass or reference problems depends on the type of application, control limits, and how long the electrode has been in service. Often we hear about reference problems in the pharmaceutical industry perhaps because the control limits are so tight. For example, we have the following input from Rick Cooley after taking the pH survey.
Rick: By far the biggest pH measurement problem I encountered was electrode junction potential. From a training perspective, the biggest issues were a lack of understanding that temperature compensation only corrected for the change in the Nernst response and not the actual solution’s change in pH with temperature, and a lack of understanding of measurement uncertainty and its relationship to control limits. Just because a vendor says the electrode and/or meter is accurate to ±0.01 pH units, this doesn’t mean users can set their control limits based on that value. The actual application can have a significant affect on the system’s measurement accuracy, particularly due to the impact various process conditions have on the electrode’s performance.
Greg: Except for some non-aqueous and high-pH (> 12 pH), and high-temperature (> 65 °C) applications that wreacked havoc with the glass, most of the problems I’ve seen are drifting reference potentials due to coatings, contamination and slow equilibration. With better diagnostics and users sharing results, we’ll all become smarter. For example, a trend of glass resistance is an indicator of a deterioration of the glass surface due to aging or chemical attack. For more info on pH, check my modeling and control website, www.modelingandcontrol.com.