Detecting and Controlling Water in Oil

Marianne Duncanson, ExxonMobil
Tags: water in oil, oil analysis, contamination control

Moisture is considered a chemical contaminant when suspended or mixed with lubricating oils. It presents a combination of chemical and physical problems for the lubricant and machinery, respectively. The potential problems, states of existence and methods for measuring moisture are discussed here.

Effects of Water on Equipment and Lubricants
The effects of water are insidious. Failure due to water contamination may be catastrophic, but it may not be immediate. Many failures blamed on lubricants are truly caused by excess water. The following are some of the effects of water on equipment:

  • Shorter component life due to rust and corrosion

  • Water etching/erosion and vaporous cavitation

  • Hydrogen embrittlement

  • Oxidation of bearing babbitt

  • Wear caused by loss of oil film or hard water deposits

Rust and Corrosion
Water attacks iron and steel surfaces to produce iron oxides. Water teams up with acid in the oil and corrodes ferrous and nonferrous metals. Rust particles are abrasive. Abrasion exposes fresh metal which corrodes more easily in the presence of water and acid.

Water Etching
Water etching can be found on bearing surfaces and raceways. It is primarily caused by generation of hydrogen sulfide and sulfuric acid from water-induced lubricant degradation.

Erosion occurs when free water flashes onto hot metal surfaces and causes pitting.

Vaporous Cavitation
If the vapor pressure of water is reached in the low-pressure regions of a machine, such as the suction line of a pump or the pre-load region of a journal bearing, the vapor bubbles expand. Should the vapor bubble be subsequently exposed to sudden high pressure, such as in a pump or the load zone of a journal bearing, the water vapor bubbles quickly contract (implode) and simultaneously condense back to the liquid phase. The water droplet impacts a small area of the machine’s surface with great force in the form of a needle-like micro-jet, which causes localized surface fatigue and erosion. Water contamination also increases the oil’s ability to entrain air, thus increasing gaseous cavitation.

Hydrogen Embrittlement
Hydrogen embrittlement occurs when water invades microscopic cracks in metal surfaces. Under extreme pressure, water decomposes into its components and releases hydrogen. This explosive force forces the cracks to become wider and deeper, leading to spalling.

Film Strength Loss
Rolling element bearings and the pitch line of a gear tooth are protected because oil viscosity increases as pressure increases. Water does not possess this property. Its viscosity remains constant (or drops slightly) as pressure increases. As a result, water contamination increases the likelihood of contact fatigue (spalling failure).

The effects on lubricating oil can be equally harmful:

  • Water accelerates oxidation of the oil

  • Depletes oxidation inhibitors and demulsifiers

  • May cause some additives to precipitate

  • Causes ZDDP antiwear additive to destabilize over 180°F

  • Competes with polar additives for metal surfaces

Maximum Recommended Water Concentrations
Oil, unless it is dried, contains some dissolved water. Figure 1 shows the amount of dissolved water that can be found in ISO 220 paper machine oil and ISO 32 turbine lubricant before it turns cloudy.

Figure 1. Dissolved Water as a Function of Temperature
in Paper Machine Oil and Turbine Oil

Table 1 helps determine the relative life of mechanical equipment versus the amount of water in the lubricant. To use the chart, estimate the current moisture level in the system along the y-axis, move toward the right to the target moisture level. The top of the chart gives the estimate of how much the life of the oil is extended. For example, by reducing moisture from 2,500 ppm to 156 ppm, machine life is extended by a factor of 5.

Life Extension Factor
Current Moisture Level
Table 1. Moisture Life Extension Method

Tests for Water in Oil
The guidelines in Table 1 help only if it is known how much water is in the oil. There are several qualitative and quantitative tests to determine water content. The easiest one to perform is a simple visual test. An ISO 68 turbine lubricant was observed at room temperature with controlled amounts of water. Table 2 shows the results of the test.

Amount of water, ppm
Appearance of oil
Bright and clear
Trace of translucent haze
Slight translucent haze
Translucent haze
Opaque haze
Opaque haze with slight water drop out
Table 2. Visual Check of Water in Turbine Oil

Bear in mind that several factors can affect the cloudy or hazy appearance of the oil. First, as the oil sits, it will clear up and the oil may become supersaturated. Second, dye and dark-color oil can mask cloudiness.

Visual Crackle Test
A test that can be performed on-site is the crackle test. It is a quick control test that is performed by heating the oil in a small metal pan using a Bunsen burner or hot plate. It is heated rapidly to 100°C and the technician listens carefully for the number of pops or crackles. It is not run on hazy oil unless there is a doubt as to whether the haziness is caused by water or some other substance.

Noria has copyrighted the following technique for running a visual crackle test. Here are the instructions from the Web site at

  • Maintain surface temperature on a hot plate of 300°F (135°C).

  • Violently agitate oil sample (such as in a paint shaker) to achieve homogenous suspension of water in oil.

  • Using a clean dropper, place a drop of oil on the hot plate.

If no crackling or vapor bubbles are produced after a few seconds, no free or emulsified water is present.

If very small bubbles (0.5 mm) are produced but disappear quickly, approximately 0.05 percent to 0.1 percent water is present.

If bubbles approximately 2 mm are produced, gather to center of oil spot, enlarge to about 4 mm, then disappear, approximately 0.1 percent to 0.2 percent water is present.

For moisture levels above 0.2 percent, bubbles may start out about 2 to 3 mm then grow to 4 mm, with the process repeating once or twice. For even higher moisture levels, violent bubbling and audible crackling may result.

The method is not quantitative. Hot plate temperatures above 300°F induce rapid scintillation that may be undetectable. The method does not measure the presence of chemically dissolved water. Different base stocks, viscosities and additives will exhibit varying results. Certain synthetics, such as esters, may not produce scintillation. Refrigerants and other low boiling-point suspensions may affect results. False positives are possible with entrained volatile solvents and gases.

Wearing protective eyewear and long sleeves is suggested, and the test should be performed in a well-ventilated area.

Calcium Hydride Test
A convenient way to determine water concentration in the field is by using a calcium hydride test kit (Figure 2). Water reacts with solid calcium hydride to produce hydrogen gas, which is directly proportional to the amount of water present in the sample. The water content of the sample is measured by the increase in pressure in a sealed container. These test kits are reported to be accurate down to 50 ppm free or emulsified water.

Figure 2. Calcium Hydride Test Kit

All of the water must come into contact with the calcium hydride. Viscous oils physically prevent water from mixing with calcium hydride whereas polar additives chemically attract water molecules to hold the water in solution.

On-line Sensors
There are several on-line sensors that measure water while equipment is operating (Figure 3).

Figure 3. On-line Impedance-type
Moisture Sensor

Some sensors measure the temperature and relative water saturation of petroleum and synthetic fluids and fuels. A probe senses water at a representative point of the system. The devices change capacitance as water concentration increases and decreases. Results are read as percent water saturation. Another technology monitors the humidity in the sump or reservoir headspace (Figure 4). Relative humidity of the headspace air has been found to correlate to lubricant moisture levels.

Figure 4. On-line Headspace
Moisture Sensor

By monitoring water content below the saturation level, these units allow action to be taken prior to the formation of free water, thus preventing problems such as additive depletion, oil oxidation, corrosion and reduced lubricating film thickness.

Temperature changes affect saturation. An oil with 200 ppm water may be suitable for use at an operating temperature of 180°F, but if the equipment cools down to 60°F, saturated water can be released as potentially damaging free water. Testing the oil in-service and correcting for temperature allows operators to discover and correct water problems before they reach the stage where water drops out.

One drawback of saturation meters is that temperature, additives, contaminants and wear particles affect saturation point. In addition, saturation meters are unable to quantify water content accurately when water is above the saturation point, typically 200 to 600 ppm for industrial oils. Despite these limitations, saturation meters can be a useful trending tool to determine moisture, provided they are used frequently and routinely.

Another sensor technology is based on the absorption of infrared light (Figure 5).

Figure 5. Single-channel Infrared
Moisture Sensor

One channel measures the amount of water in the oil while the other is a reference channel. The infrared absorption is determined from the difference between these two channels at the target spectral band for water. This absorption, using a calibration curve, is used to estimate the amount of water in the oil sample as traditionally presented in ppm or percent. According the manufacturer, it will read concentrations to two percent water.

Quantitative tests for water include Karl Fischer, water by distillation and FTIR. Karl Fischer (Figure 6) is accurate from 1 ppm to 100 percent and is relatively quick and inexpensive. The oil sample is titrated with a standard Karl Fischer reagent until an end-point is reached. The difference in test methods is based on the amount of sample used for the test and the method used to determine the titration end-point.

Figure 6. Coulometric Karl Fischer
Titration Analyzer

ASTM D1744, a volumetric method, is reliable and precise, but there can be reproducibility problems at low water concentrations (200 ppm or less). Soaps, salts from wear debris and sulfur-based additives react with the Karl Fischer and can give a false positive. In fact, a new, clean, dry antiwear (AW) or extreme pressure (EP) oil may give a reading of as much as 200 to 300 ppm.

ASTM D6304, a coulometric titration method (Figure 6), is more reliable than D1744 at low water concentrations and is less prone to interference effects, although again, AW and EP additized oils can show as much as 100 ppm of water.

The most reliable method is ASTM D6304. The oil sample is heated under a vacuum so that any water present in the sample evaporates. Water vapors are condensed and dissolved into toluene, which is then titrated. Because the additives and other interfering contaminants remain in the oil, the condensed water in the toluene is a true indication of water present in the sample.

Water by Distillation
Water by distillation measures the amount of water boiled off in a special still (Figure 7).

Figure 7. Distillation Method for Determining Moisture Levels
Figure 7. Distillation Method for Determining Moisture Levels

The classic method for determining water-in-oil is the Dean and Stark distillation method (ASTM D95). This test method is fairly cumbersome and requires a comparatively large sample to ensure accuracy, which is why it is rarely used in production-style oil analysis labs today.

As the oil is heated, any water present vaporizes. The water vapors are then condensed and collected in a graduated collection tube, such that the volume of water produced by distillation can be measured as a function of the total volume of oil used. It can detect between 500 ppm and 25 percent water.

Fourier Transform Infrared
FTIR can be an effective method for screening samples containing in excess of 1,000 ppm of water, provided a correct new oil baseline is available for spectral subtraction. However, due to its limited precision and comparatively high detection limits, FTIR is not adequate in many situations where precise water concentrations below 1,000 ppm or 0.1 percent are required.

Control of Water Sources
Now that the amount of water in the oil has been determined, how does one control it?

First, control the source of water contamination. Following are common sources of water into lubricating oil and suggestions on how to control them:

  • Manage new oil properly.

  • Use desiccant breathers or other tank headspace protection.

  • Avoid shafts, fill ports and breathers when washing down machines.

  • Avoid high-pressure sprays around seals if possible.

  • Maintain steam and heating/cooling water system seals.

  • Periodically inspect rotary steam joints for leaks; replace seals and/or correct alignment as appropriate; install flinger flanges to direct steam away from labyrinth seals.

  • Repair heat exchanger leaks.

  • Prevent washdown water from entering vents and reservoir covers.

  • Properly install and seal covers and hatches.

  • Watch for condensation caused by cold water lines located close to a hot reservoir.

  • Gutter water to divert flow away from reservoir hatches.

  • Install secondary seals or V-rings on critical systems.

  • Use and maintain high-quality shaft and wiper seals.

  • Prevent contamination from conden-sation by using a bladder-type breather on vents.

  • Install desiccant air breathers on vents

  • Prevent water from entering new oil by storing drums indoors. If they must be stored outdoors, keep them in a shed or under a tarp, or store them on their sides.

  • Install a vapor extractor or fan to remove humidity from large reservoirs.

  • Periodically drain water from low points in system.

Moisture can be an insidious problem for the equipment operator. With general precautions to prevent contamination, and an appropriate understanding of the methods for and a plan to detect the presence of moisture in mechanical systems, the deleterious effects of moisture can be avoided. Coupled with an effective moisture removal approach, lubricant and machine life may be extended appreciably, providing the equipment operator with one more lever to use in the pursuit of reliability.

Editor’s Note:
This article was originally published in the Lubrication Excellence 2005 Conference Proceedings, Noria Corporation.

Troyer, D. “The Visual Crackle - A New Twist to an Old Technique.” Practicing Oil Analysis magazine, September-October 1998.

About the Author