Photoacoustic Spectroscopy - The Future in Contamination Monitoring?

At Practicing Oil Analysis, we are always on the lookout for exciting new technology that has the potential to revolutionize the oil analysis industry. Recently we were made aware of some principal research conducted at the U.S. Department of Energy's Pacific Northwest National Laboratory (PNNL) in Richland, Wash., in which the ability to detect water in oil down to 50 ppm is reported. With notebook in hand, we set out to the Pacific Northwest to discover the secrets behind photoacoustic spectroscopy, an absorption spectroscopic technique similar to Fourier transform infrared (FTIR) spectroscopy.

The Status Quo
Water is one of the most abundant substances on earth. While vital for human life, the effects of water are quite the opposite on machine and lubricant life, resulting in oil degradation, corrosive wear and poor lubrication. Because of this, monitoring water contamination is one of the most important components of any oil analysis program. Current methods for determining absolute water concentrations in lubricating oils include FTIR, Karl Fischer titration, dielectric measurements and calcium hydride test kits. Table 1 summarizes the pros and cons of these test methods. Despite the importance of accurately determining water concentrations in oil, no one method can claim to be the ideal solution for all situations, as the table suggests.

The cure-all to these problems would appear to be a technology that (1) is readily adaptable to a compact, relatively low cost, robust field instrument - just like particle counters, (2) has precision and detection limits similar to a lab-based Karl Fischer moisture test and (3) doesn’t use any wet chemistry procedures. Photoacoustic spectroscopy (PAS) seems to meet all three of these criteria, and therefore, has great potential to revolutionize the oil analysis industry.

The Proof is in the Results
PAS is an absorption-based spectroscopic technique, similar to FTIR. However, FTIR measures a small signal on top of a large background (the difference between two transmitted light signals), while PAS has a very small background and thus a sensitivity that is perhaps 10 to 1,000 times greater than conventional absorption techniques such as FTIR. The background noise in PAS stems from absorption by the solvent and from an electromagnetic interaction between the laser light pulse (used to excite the water molecules) and the solvent matrix (in this case the oil) known as electrostriction. However, PAS is typically 10 to 20 times more sensitive than FTIR.

The research team at PNNL obtained samples of clean mineral-based Dextron-type transmission fluid and hydraulic fluid along with a polyol ester-based synthetic gas turbine engine oil to test their theories. Standard test samples were prepared by adding known quantities of water (by mass) to the oil samples, some of which were pre-dried using a molecular sieve to remove any trace of water originally present. The test samples were then analyzed using both FTIR and PAS to determine the lowest detection limits for each technique. In each case, the signal obtained was plotted as a function of the amount of water added to each sample.

Figure 1. Photoacoustic Signal Vs. Water Concentration (Reprinted with permission from Sensors and Actuators.)

Based on the reported results in Figure 1, it was concluded that within the range of water concentrations and oil types tested, the photoacoustic signal is directly proportional to the amount of water added, with lower detection limits of 45 ppm for the transmission fluid, 60 ppm for the hydraulic fluid and 515 ppm for the polyol ester engine oil. The lower sensitivity obtained for the polyol ester fluid is caused by two factors. First, the primary absorption band for water in a polyol ester matrix is shifted to higher frequency (~3625 cm-1) than in the mineral oils. Thus, the excitation frequency used for this experiment (3415 cm-1) was not at the absorption maximum for water in ester-based fluids. Second, two weak absorption bands related to the oil itself occur near the 3415 cm-1 excitation frequency, which increases the background noise in the measurement. Because PAS, like FTIR, is not molecule selective, any species present that absorbs light at the excitation wavelength will yield a PAS signal. By switching to a different excitation wavelength (such as the 3625 cm-1 band typically used to monitor water concentrations in ester-based fluids using FTIR), lower detection limits comparable to those obtained for the mineral-based transmission and hydraulic oil results would have likely been observed.

Figure 2. Layered Prism PAS Cell Used to Measure Water in Oil

By contrast, the best detection limit obtained by PNNL using FTIR - including 32 scan co-addition, a background offset correction and spectral smoothing - was 100 ppm for the transmission fluid samples. Compare this to 45 ppm using PAS, which indicates at least a two-fold increase in sensitivity for PAS compared to FTIR. The team at PNNL feels that with more development work, lower detection limits for PAS for water in oil may ultimately be as low as 10 to 20 ppm, making it 5 to 10 times more sensitive than FTIR.

The Future Looks Bright
The team at PNNL is currently seeking a commercial partner to help develop a PAS-based instrument for the detection of water in oil. The results of the experiments already conducted by PNNL clearly demonstrate a high potential for such an instrument. The combination of the relative simplicity and high sensitivity of the technique suggests a bright future for PAS in the oil analysis industry. With the ever-improving availability of diode lasers and high intensity flash lamps (required as pulsed excitation sources for a low-cost, small-scale PAS-based oil analysis instrument), along with PNNL’s continued development of specialty PAS cells and instrumentation, the opportunities are excellent for not only a lab-based instrument, but also small-scale, portable and even on-line PAS-based sensors and instruments.

Just like with FTIR, any molecular species that shows a characteristic absorption peak has the potential for detection using PAS. There may be opportunities to extend this PAS-based system to cover not just water, but other contaminants such as glycol and fuel, lube degradation by-products of oxidation and nitration, and even oil additives such as hindered phenols and ZDDP. Eventually, one may be able to buy a portable PAS-based diagnostic instrument, similar in size and cost to a portable particle counter that allows the detection of all these parameters and more - all at the press of a button!

Further Reading
1. Review of Scientific Instruments. (June, 1998). Vol. 69, pp. 2246-2258.
2. Sensors and Actuators. (2001). B 77, pp. 620-624.

For more information regarding PNNL’s photoacoustic work contact Eric Jurrus (licensing and commercialization, 509-372-4905, eric@pnl.gov), or Jim Amonette (technical issues, 509-376-5565, jim.amonette@pnl.gov).