Study Reveals Factors That Affect Particle Counting Accuracy

Thomas S. Wanke, Milwaukee School of Engineering Paul W. Michael, Milwaukee School of Engineering Michael A. Mccambridge, Milwaukee School of Engineering
Tags: particle counting, oil analysis

Particulate contamination interferes with the ability of hydraulic fluids and lubricants to minimize friction and wear. This concentration of contaminants has a direct impact on a system's performance and reliability, and must be controlled to an appropriate level. The quantitative determination of particle contamination requires precise measurement. Automatic particle counters that utilize the light extinction principle are widely used for this purpose. Accuracy of particle count data can be affected by sample preparation techniques and sensor design.

Occasionally, particle counters detect phantom particles that cannot be removed by filtration. This article examines the possible role of base oil and additive selection in the appearance of phantom counts. Filtered Group I and Group III base oils were doped with the additive components of an engine oil formulation and particle levels were monitored before and after filtration using an MP Filtri LPA-2 on-line automatic particle counter. The results show that base oil selection has minimal bearing on the appearance of phantom counts, while additives may be a factor.

Additionally, the results from three different particle counters were compared. All three automatic particle counters produced phantom counts when a polydimethylsiloxane antifoam additive was blended into the base oil. A direct imaging laser particle counter classified antifoam particles as water droplets. After filtration, the particle counts at four microns remained elevated and there was a five-ISO code spread in the results produced by the three particle counters. The six- and 14-micron counts were within one or two ISO codes of each other, which is within the allowable difference between particle count runs specified in ISO 11500-1997.

Table 1. Ranges for ISO 4406 Scale Numbers (Abridged)

Background
Measuring Fluid Cleanliness
Automated white light and laser particle counters are used to quantify the contamination level of hydraulic fluids. The basic concept of these particle counters is simple: a beam of light is projected through a narrow stream of the sample fluid, and when a particle blocks the light, a voltage pulse that is proportional to the equivalent diameter of the particle is produced. Unfortunately, these sensors can be susceptible to producing phantom counts.1

Recently, LaserNet Fines (LNF), a direct imaging laser particle counter, was developed by the United States Naval Research Laboratory and commercialized by Lockheed Martin.2 This instrument is an optically based particle analyzer that uses an artificial neural network to analyze pixilated images of wear particles that are projected onto a charge-coupled device (CCD) camera image sensor. LNF measures the distribution of particles from four microns to 100 microns and classifies contaminants larger than 20 microns based on their shape. Through the use of mathematical algorithms, LNF is able to identify images of:

  • Cutting wear particles

  • Fatigue wear particles

  • Sliding wear particles

  • Water and air bubbles

  • Fibers and nonmetallic contaminants

The ISO Contamination Code
ISO 4406 describes a method of coding the level of solid particles in an oil sample.3 The code number corresponding to a contamination level involves three scale numbers, which permits the differentiation of the dimension and distribution of the particles. The first scale number represents the number of particles per milliliter (ml) of fluid = four microns (µm). The second scale number represents the number of particles per ml of fluid = six µm. The third scale number represents the number of particles per ml of fluid = 14 µm. The range between the upper and lower limits for each scale number is a factor of two (Table 1). This also means that for each increase of one ISO code number, the number of particles approximately doubles. For example, an ISO code of 17/15/12 indicates there are 641 to 1,300 particles = four µm/ml, 161 to 320 particles = six µm/ml, and 21 to 40 particles = 14 µm/ml of fluid.

Table 2. API Base Oil Classifications

Changes in Base Oil Dewaxing Processes
For years, solvent-refined paraffinic oils were the primary base stocks used in hydraulic fluids and engine oils. Over the past decade, demand for reduced emissions and enhanced oxidation stability in diesel and passenger car motor oils has led many refineries to convert base oil production from solvent extraction to hydrogen treatment and catalytic dewaxing processes. In hydrotreating and catalytic base oil dewaxing, the feedstock reacts with hydrogen in the presence of a catalyst at high temperatures (752°F / 400°C) and pressures (3,000 psi). This results in:

  • Removal of compounds containing sulfur, nitrogen and oxygen.

  • Conversion of aromatic hydrocarbons to saturated cyclic hydrocarbons.

  • Breaking up of higher molecular weight polycycloparaffins into lower molecular weight saturated hydrocarbons.

The lubricating oils produced by this method have higher levels of saturated hydrocarbons and lower levels of sulfur. The American Petroleum Institute (API) categorizes these lubricating oils based on the sulfur and saturated hydrocarbon content (Table 2).4 These catalytically processed or hydrocracked base oils provide significantly higher oxidation life of traditional solvent refined paraffinic oils.5 This improvement in oxidation stability is the direct outcome of reduced sulfur and unsaturated hydrocarbon levels. While eliminating these components improves oxidation stability, it also reduces the solvency of the base oil and, to some extent, additive solubility.6 The effect of base oil solubility on particle count results is further examined by comparing Group I and Group III base oils.

Table 3. Additive Descriptions and Concentrations

Additive Descriptions and Concentrations
Additives are an integral part of modern hydraulic fluid formulations. Some additives such as dispersants, detergents, antiwear agents and corrosion inhibitors react with metals and oxidation products to reduce wear and maintain system cleanliness. Other additives such as foam inhibitors and viscosity index improvers enhance the physical properties of lubricants critical to performance. Table 3 describes the additives evaluated in this study and the percent by weight of additives blended into the base oils.

Table 4. Results of Particle Counter Validation Tests with ISO 2806 Medium Test Dust

Experimental
Procedure
In this study, particle counts for blends of individual additive components were compared to a fully formulated diesel engine oil dispersant-inhibitor (DI) additive system. These additives were mixed into commercial Group I and Group III paraffinic base oils at concentrations similar to what might be found in a typical engine oil. The reservoir was charged with 10 gallons of base oil and the pump flow was adjusted to 10 gallons per minute to achieve a nominal circulation rate of one "turn" per minute. An MP Filtri LPA-2 on-line particle counter with dual laser sensors was used to monitor the fluid's contamination level at five-minute intervals throughout the test. Bottle samples were also collected for off-line analysis.

The base oil was circulated through a three-micron Beta >200 filter until it reached ISO -/16/13 or cleaner. After the fluid reached ISO -/16/13 and 100°F, the filter-bypass valve was opened and the additive was measured in the cone-bottomed reservoir. After = 30 minutes of mixing via circulation, the filter bypass valve was closed and filtration was initiated. (Note: The filtration step was omitted if introducing the additive did not raise the ISO code above -/15/12.) Filtration was discontinued when the ISO code reached -/15/12 or 30 minutes, whichever came first. If the fluid did not reach ISO -/15/12 or cleaner within 30 minutes, the system was drained and flushed with mineral spirits prior to charging with fresh base oil.

Particle Counter Calibration
Calibration was verified with an ISO 2806 medium test dust secondary reference fluid prior to analysis.7 The results are shown in Table 4.

Table 5. Particle Count Results for Additives Blended into Group I and Group III Base Oils

Results
The unfiltered Group I and Group III base oils were relatively clean prior to introducing the additive. In both instances, the base oils were found to be cleaner than a typical commercial lubricant (Column 1 in Table 5) for each base oil type. As can be seen in Table 5, blending additives into the base oil increased particle counts to varying degrees (Column 2 in Table 5) for each base oil type. The greatest increase can be seen with the diesel engine oil additive and the polydimethylsiloxane antifoam. In terms of absolute counts, the effect of the remaining additives was less significant. More importantly, these other additives responded to the three-micron filtration in a predictable manner.

To compare particle counter results, a second experiment was performed on the Group I base oil / DI additive combination. In this series of tests, six bottle samples were collected from the reservoir at approximately 10-minute intervals. Particle count tests were performed using the MP Filtri LPA-2, Hiac 8000A and LaserNet Fines particle counters. As seen in Table 6, addition of the DI package to the Group I base stock at 120°F increased the initial particle count (sequence No. 1) in all size ranges (sequence No. 2 and No. 3).

After a half hour of mixing with the filter bypass loop open, the authors commenced three-micron filtration. Filtration reduced the six- and 14-micron ISO codes. The allowable difference between runs with one sample per ISO 11500-1997 is shown in Table 7. The three particle counters meet this criteria at six and 14 microns. In the four-micron range, the results varied by nearly two orders of magnitude. Interestingly, the LaserNet Fines (LNF) particle counter reported the presence of hollow spherical particles. These particles were classified as water droplets by the LNF artificial neural network.

Table 6. Group I Base Oil and DI Package, Comparison of ISO Codes for Three Instruments

Discussion
The results indicate the silicone antifoam agent is the most likely source of phantom counts in light-blockage laser particle counters. Foaming occurs when gas rises to the fluid surface and forms stable bubbles that do not immediately break. Silicone antifoams function by forming an insoluble micelle within air bubble walls that reduces the surface tension of foam and causes the thinning and collapse of the bubble wall. To accomplish this, an antifoam must have a surface tension lower than the fluid, be insoluble in the fluid, and disperse into small droplets within the fluid.8 The insoluble nature of these additives can lead to the appearance of phantom counts in light-blocking laser particle counters.

According to data provided by the additive supplier, the DI package contains six ppm silicon when diluted in base oil, the finished, blended concentration. In polydimethylsiloxane antifoam additives, R = CH3 and there are a large number of repeating groups (n). Based on the structure of the repeating groups, dimethyl silicone compounds contain approximately 38 percent silicon, plus carbon, oxygen and hydrogen. Therefore a finished lubricant that contains six ppm silicon actually contains 16 ppm silicone compounds. Because the density of silicone (0.971 g/cm3) is greater than that of mineral oil (approximately 0.876 g/cm3), 16 ppm weight is equivalent to ~14 ppm (vol).

While 14 ppm may not seem like enough to affect particle count results, each ml of fluid formulated with silicone compounds at a treat rate of 14 ppm by volume contains 14 x 106 cu µm of silicone. Assuming the silicone forms six-micron spherical micelles, each micelle contains 113 cu µm per spherical "particle." Thus, if all of the silicone antifoam in the DI package were to form six-micron micelles, the particle count at the six-micron level would exceed 120,000 particles/ml. The DI additive produced counts in excess of 120/ml at = 14 µm, 4,000/ml at = six µm and 21,000/ml at = four µm, 15 minutes after additive addition.

The total volume of these particles is less 1.3 x 106 cu µm, assuming the particles are spherical and there is no significant overlap between the size ranges. This accounts for less than 10 percent of the total 14 x 106 cu µm of silicone present in the diesel engine oil. Because there was no appreciable increase in the foam stability of the fluid after filtration, and ICP analysis indicated the change in silicon content was one ppm or less, it appears that 90 percent of the silicone antifoam additive remains dispersed in the fluid at a particle size less than four µm.

Table 7. Allowable Difference between Particle Count Runs per ISO 11500©. This material is reproduced from ISO 11500-1997 with permission of the American National Standards Institute on behalf of the International Organization for Standardization (ISO).

Conclusion
Filtered Group I and Group III base oils were doped with the components of an engine oil formulation. The results show that base oil selection has minimal bearing on the appearance of phantom counts, while additive selection can be a factor. The results from three different particle counters are compared, and all three automatic particle counters produced phantom counts when a polydimethylsiloxane antifoam additive was blended into the base oil. A direct-imaging laser particle counter classified these antifoam particles as water droplets.

After filtration through three-micron media, the particle counts in the = four-micron media remained elevated and there was a five-ISO code spread in the results produced by the three particle counters. The = six and =14 µm counts were within one or two ISO codes of each other, which is within the allowable difference between particle count runs specified in ISO 11500-1997.

Acknowledgments
The authors gratefully acknowledge the contribution of materials by Benz Oil and Afton Chemical Corporation. The fine work of MSOE undergraduate research assistant Matt Loeffler is also appreciated.

References

1. Michael, P.W. and Wanke, T.S. "Surgically Clean Hydraulic Fluid - A Case Study." Proceedings of the 47th National Conference on Fluid Power. National Fluid Power Association, Milwaukee, WI 1996. p. 129-136.

2. Reintjes, J., Tucker, J., et al. "LaserNet Fines Wear Debris Analysis Technology: Application and Mechanical Fault Detection." AIP Conference Proceedings, No. 657B, 2003. p. 1590-7.

3. ISO 4406:1999 Hydraulic fluid power - Fluids Method for coding the level of contamination by solid particles.

4. American Petroleum Institute Publication 1509. "Engine Oil Licensing and Certification System." 13th Ed., 1995.

5. Michael, P.W. "Standards for Hydraulic Fluid Testing." Handbook of Hydraulic Fluid Technology, G.E. Totten, Ed., 2000. Marcel Decker, NY. p. 1189.

6. Givens, W.A. and Michael, P.W. Fuels and Lubricants Handbook, G.E. Totten, Ed., 2003. ASTM International, West Conshohocken, PA. p. 335.

7. ISO 11171:1999 Hydraulic fluid power - Calibration of automatic particle counters for liquids.

8. Friesen, T.V. Transmission-Hydraulic Fluid Foaming, SAE Technical Paper 871624, 1987.


About the Author
About the Author
About the Author
Create your own user feedback survey