Commercially available hardware and software have transitioned digital image analysis from space exploration to everyday life. For example, stores use scanners to read bar codes printed on product packaging.

Applications of digital imaging in lubricant analysis stepped into the spotlight at international technical meetings in 2017. Two speakers demonstrated ways that digital image analysis adds value to standard lubricant tests for wear and corrosion.

Wear Scars Get Deep

Nicole St. Pierre, technical development manager with Nye Lubricants in Fairhaven, Massachusetts, informed her audience of test results for antiwear additives at the Society of Tribologists and Lubrication Engineers annual meeting in Atlanta.

St. Pierre undertook an ambitious study to compare five antiwear chemistries (Table 1, Page 28) at two treat rates (1 and 3 percent by weight) in three base oils (polyalphaolefin, di-ester and polyol ester).

She used four-ball wear testing (ASTM D4172) to compare antiwear performance. In the test, three lubricated steel bearings are clamped snugly by a ring inside a cup while a fourth bearing rotates against them for one hour. Afterward, an optical microscope is used to measure the length and width of each wear scar. The average scar diameter for three bearings is used to evaluate lubrication.

St. Pierre told the audience that she was inspired to use a profilometer instead of an optical microscope to measure the wear scars. This laboratory technique uses an instrument called a profilometer to measure vertical heights and depths of bumps and grooves, as well as their shapes or profiles, on a solid surface.

The earliest profilometers had a stylus or a small tool with a sharp diamond tip that glided across the rough surface of a test piece on a stage that the instrument moved back and forth below the stylus. State-of-the-art optical profilometers use an LED or other light source to illuminate the surface of a test piece, optics to analyze reflected light and computers to generate digital images of rough surfaces and analyze their features.

First, it was convenient to use a high-resolution bench-top optical profilometer to automatically measure wear scars with appropriate accuracy and repeatability, she explained.

Second, I was looking for a way to see if we could differentiate between similar-performing wear additives. This profilometer [Taylor Hobson, 1/10 Angstrom vertical resolution] automatically measures wear scar diameters and volumes. Historically, the lubricants industry has only looked at wear scars two-dimensionally. There is an entire dimension that has never been considered which may drastically affect how antiwear additives are chosen for each application.

Third, wear scars are rarely symmetric and oftentimes have jagged edges, which make it difficult to accurately measure scar diameters. By looking at the total wear volume of a scar, we can account for the jagged edges of the scar.

And fourth, she continued, this profilometer uses digital image analysis. It collects more representative data in 3-D and generates pictures of surfaces. This allows you to see how the test specimens are wearing in 3-D. By looking at the wear scars, we can see what type or types of wear (pitting, spalling, galling, etc.) are present under specific test conditions. That information will help formulators improve formulation capabilities.

St. Pierre used a commercial statistical software package to organize and analyze test data. She calculated numerical models to fit four-ball data. These models fit her data with regression coefficients of 87 percent for wear scar diameters and 72 percent for wear scar volumes.

Results for wear scar diameters and volumes for 1 percent and 3 percent additive treat rates in PAO, relative to a control with no antiwear additives, are shown in Typical wear scar images are shown on pages 30 and 31. For example, average wear scar diameter and volume for the PAO control were 0.79 millimeters and 3.2 million cubic microns, respectively. With 1 percent treat rate of amine phosphate, scar diameter and volume were reduced by 46 percent (to 0.46 mm) and by 92 percent (to 130,000 cubic microns), respectively.

St. Pierre explained that it is important to analyze wear in three dimensions (scar volume) as well as two dimensions (scar diameter). She observed many cases in which scars with similar diameters had significantly different volumes. In one instance, 3 percent treat rates of ZDDP and phosphate ester in PAO reduced scar diameters by the same amount (16 percent), but scar volumes were smaller for phosphate ester than ZDDP (53 percent versus 45 percent reduction, respectively).

In the ester base oils, there were even larger differences in wear scar volumes for additives that had similar 2-D wear scar diameters, she continued. Some 3-D wear scars showed a 33 percent difference for ZDDP vs. molybdenum phosphorodithioate (27 and 60 percent reductions, respectively); and a 63 percent difference for the molybdenum phosphorodithioate vs. amine phosphate (97 and 34 percent reductions, respectively).

I think thats a real game changer, St. Pierre exclaimed. Bigger differences could be seen with other additives or additive combinations.

She also warned her audience about cases, such as 3 percent treat rate of methylene bis(dibutyldithiocarbamate) in PAO, in which scars were much deeper (100 percent increase in 3-D) than expected on the basis of diameter (16 percent increase in 2-D).

I think that imaging analysis of test specimens will allow formulators to have a better understanding of how additives or a combination of additives will perform in each lubrication regime, which will allow them to better select the appropriate additives for the specific application that they are formulating for. This, in turn, will lead to longer component life, less frictional heat generation and improved fuel economy for vehicles, she concluded.

Corrosion in 3-D

While St. Pierre used off-the-shelf hardware and software to make and evaluate digital images of wear scars, other scientists and engineers are inventing new laboratory equipment and software for digital image analysis of corrosion on metal surfaces.

Jason Galary, applied science and tribology manager at Nye, introduced his patent-pending inventions to the audience at a meeting of the European Lubricating Grease Institute in Helsinki in May. Galary has developed laboratory equipment to create digital images of corrosion on metal surfaces, plus software to accurately analyze the images and measure the two-dimensional area of corrosion.

Almost every metal can undergo corrosion, he reminded meeting attendees. Rust, the most familiar type, appears when metals undergo electrochemical reactions. Rust forms on steel when iron atoms react by giving some of their electrons to oxygen and/or water molecules at the steel surface. Yellow metals such as copper and bronze can corrode by reacting with acids, bases or other aggressive chemicals to form undesirable byproducts. Various types of chemical additives have been developed to protect metals by inhibiting corrosion under specific conditions.

Next, Galary reviewed laboratory tests for evaluating corrosion and the corrosion resistance performance of lubricating greases. In three widely used ASTM tests, D1638 (also referred to as the EMCOR test), D1743 and D5969, bearings are coated with grease and then exposed to conditions that promote corrosion, such as heat, humidity and chemicals.

Laboratory personnel then visually evaluate bearing surfaces for evidence of corrosion, including discoloration, deposits and pitting, and categorize grease performance according to a generic rating system. Human error can limit the accuracy and reproducibility of this visual method of evaluation.

Galary described his equipment as a machine vision system that uses computational algorithms to detect and classify corrosion and accurately determine the percentage of the surface area that is affected.

He used special lighting and camera equipment to make digital photographs of bearing surfaces (above and left). His greatest challenge was edge detection-defining the edges or boundaries of corrosion in digital images. To solve this problem, Galary developed an algorithm that detects the edges of discoloration, pits and deposits of corrosion byproducts, and calculates the percent of corroded surface area.

To validate his digital image analysis equipment, Galary tested NLGI grade 2 greases, meeting MIL-PRF-32014 specifications, from two manufacturers. Greases were formulated from polyalphaolefin base oil (140 cSt at 40 degrees Celsius), lithium soap thickener and corrosion inhibitors (an amine borate and barium dinonyl naphthalene sulfonate).

Galary performed the EMCOR test, in which bearings were filled with 10 grams of grease and partially immersed in 3 percent synthetic sea water. The bearings were run at 83 rpm with no load for eight hours, followed by 16 hours motionless (ASTM D6138), for one week. Then they were disassembled, cleaned and evaluated visually and digitally.

The results showed that Galarys equipment produced good repeatability in this study, and his computational method accurately detected relatively small differences in grease formulations.

Galary concluded that his inventions show potential for improving the use of standard corrosion test methods to formulate greases, compare lubricants, differentiate between raw materials and even help provide greater insight into corrosion prevention properties of lubricating greases.

Mary Moon, Ph.D., is a chemist with hands-on R&D and management experience formulating, testing and manufacturing lubricating oils and greases and specialty chemicals. She is skilled in industrial applications of tribology, electrochemistry and spectroscopy. Contact her at mmmoon@ix.netcom.com or (267) 567-7234.