Grease Analysis

Monitoring Grease Serviceability

and Bearing Condition
BY BILL H ERGUTH, HERGUTH LABORATORIES The second philosophical issue, that of whether the grease
sample is truly representative, is slightly more complex to

A pproximately 90 percent of all bearings are lubricated resolve. In this case, the technician needs to be aware of the
with grease. But how much do you know about the grease differences in information that can be obtained from grease
or greased bearings in your plant? Grease analysis is certainly located at the bearing’s raceway interfaces, compared to grease
not for every bearing . . . maybe it’s not even for most bear- that has been pushed out and is around the outside area of the
ings. But when you have a need to know, a thorough analysis housing, in much the same way as sample point location is vital
of the grease in question can prevent headaches and save when taking used oil samples. Generally, the grease sample of
money. interest is the grease doing the work at the contact interfaces, in
Historically, the analysis of grease has been confined to new the load zone of the bearing. This grease sample will have the
grease testing for product acceptance and quality control. most evidence of wear, contamination and degradation and in
Technically, this was due to the sample size required to general will be the most representative, although it will likely
perform conventional ASTM (American Society of Testing also be the most difficult to obtain.
and Materials) methods on grease samples. However, over the
last couple of decades, new analytical methods have made it
Analyzing Used Grease Samples
possible to profile the serviceability of grease using as few
Changes in Grease Consistency
milligrams of sample as possible (0.00003527 of an ounce).
Grease is made up of base oil, a gelling agent or soap thick-
ener (sometimes called filler) and additives, which perform in
Sampling In-service Greases much the same way as oil additives. The consistency of grease is
The philosophical issues with sampling used greases are controlled by the type and ratio of the gelling agent to the oil
twofold. First, if the whole bearing must be sent to the lab just and its viscosity. Grease can harden or soften in service due to
to get at the grease under the shield, what good is the analysis? the effects of contamination, loss of oil or mechanical shearing.
Second, if you can get to the bearing grease without disman- The classic way in which the consistency of a grease is meas-
tling the bearing, is the sample reliable and representative of the ured is using the cone penetration test (ASTM D217). In this
condition of both the grease and the bearing? test, the grease is heated to 25°C (77°F) and placed below the
These two problems are quite different. If a plant has tip of the test cone, as shown in Figure 1. The cone is allowed
hundreds of greased bearings that are so hard to get to that they to drop into the grease; the amount of penetration is measured
must be removed and sent to the laboratory for grease analysis, by the penetrometer in tenths of a millimeter. The greater the
the data obtained on a few bearings may need to be extrapo- penetration, the thinner the grease consistency. The test is often
lated across the entire bearing population to draw appropriate repeated after “working” the grease to replicate the effects of
conclusions. An example of this might be to identify the root mechanical shearing. In fact, worked penetration of new
cause of a number of repeat bearing failures plant-wide. greases is the property upon which the NLGI (National
Experience shows that within a given application and environ- Lubricating Grease Institute) grease consistency classification
ment, most of the failures, if they are lubrication-related, will system is based, as shown in Table 1.
stem from the same root cause. Therefore, it is worth the effort Cone penetration using ASTM D217 requires a large volume
to take a group of bearings with varying service life and submit of sample and is not normally performed on used grease
them for analysis. Once complete, the test results can be used samples. An alternative method, ASTM D1403 uses ¼ or ½
to make an informed decision about grease selection, regreasing the sample volume of ASTM D217 making it more amenable
intervals, common wear mechanisms and typical cleanliness to used grease analysis.
levels. Continued on Page 20

18 Practicing Oil Analysis • March - April 2002

NLGI Worked Description of Continued from Page 18
Number Penetration mm Consistency A more modern alternative to cone penetration, for esti-
000 445-475 Semi-liquid mating changes in the consistency of used greases is Thermal
00 400-430 - Gravimetric Analysis (TGA). TGA analysis measures the mass
0 355-385 Viscous liquid of a substance in relationship to temperature and is used to
1 310-340 - determine the loss of material with increasing temperature. The
2 265-295 Will flow analysis can be carried out in an inert atmosphere such as
3 220-250 - nitrogen or a reactive atmosphere such as oxygen. Typically, a
4 175-205 Resist flow few milligrams of the sample is weighed and heated under
5 130-160 - controlled conditions. The weight loss at specific temperatures
6 85-115 Hard and wax like allows the technician to evaluate the oil/gelling agent ratio as
compared to new (unused grease), as well as the presence of
Table 1. NLGI Numbers for New Grease
volatile compounds such as water, allowing any significant
change in gelling agent chemistry to be determined (Figure 2).
Courtesy Koehler Instrument Co.
Antioxidant Levels in Grease
Greases, like oils, contain a variety of additives. Antioxidant
levels are of particular interest in identifying the useful life of
the grease. Differential Scanning Calorimetry (DSC) is a
modern analytical method for measuring the onset of oxidation
in used grease (ASTM D5483). When compared to the new
reference grease, the test can be used to determine the
remaining useful life (RUL) of a grease. This test is analogous
in the information it seeks, if not in methodology, to the
RPVOT test commonly used to determine the RUL of turbine
oils and other lubricating oils.
DSC works by placing a sample of used grease into a test cell.
The cell is heated and pressurized with oxygen. When the
grease starts to oxidize, an exothermic reaction occurs, which
liberates heat. By measuring the onset of the reaction in the
used grease (commonly called the induction point) in reference
to the new grease, an estimate of the oxidation stability of the
grease can be made (Figure 3).

Figure 1. Standard Penetrometer Cone Grease Viscosity

The viscosity of grease is often misunderstood. The viscosity
typically listed on a new oil data sheet is usually the kinematic
viscosity of the oil used in blending the grease measured, using
the standard ASTM D445 method. The kinematic viscosity of
the base oil is important in ensuring the correct grease,
containing the correct grade of oil is used for lubrication
purposes. However, we can also measure the viscosity of the
grease itself. Since a grease is non-Newtonian, we can only
measure the apparent viscosity because the viscosity of a non-
Newtonian fluid changes with shear stress (see “Understanding
Absolute and Kinematic Viscosity” by Drew Troyer on page 6).
The apparent viscosity of a grease is determined using ASTM
D1092. This test measures the force required to force the grease
through an orifice under pressure. As such, this test is an ideal
way of determining the flow characteristics of grease through
pipes, lines, dispensing equipment as well as its pumpability.
Figure 2. Example of a TGA Grease Profile Continued on Page 22

20 Practicing Oil Analysis • March - April 2002

The sample pan and reference Continued from Page 20
pan are located on two identical Rheology measurements of grease may
heating elements. These are
enclosed in an oxygen rich soon replace both the cone penetration
environment and heated. While and the apparent viscosity measure-
heating, the enthalpic difference ments. Rheology is the study of the
between the two pans is
measured. When the grease
deformation and/or flow of matter when
sample starts to oxidize, an it is subjected to strain, temperature and
exothermic reaction occurs, time. A rheometer only requires a few
which generates heat. This grams of sample to perform the analysis,
increased heat in the used sample
relative to the new sample can yielding much more information than
be measured, allowing the onset the cone penetration or the apparent
of oxidation to be determined. viscosity measurements. This makes the
Onset of oxidation of grease
can be measured by the DSC
rheology measurement an ideal test for
method. When compared to new small amounts of used grease.
grease, the time it takes to
oxidize the used grease is an Dropping Point
indication of the remaining useful The dropping point of a grease is the
life of the antioxidant additive
contained within the grease. By temperature at which the grease changes
knowing the service life of the from semi-solid to a liquid. The
used grease sample and the dropping point establishes the
depletion of the antioxidant
additive as measured by DSC, the
maximum useable temperature of the
appropriate regreasing schedules grease, which is typically set at 50°C to
or the relative effectiveness of 100 °C below the experimentally
different greases at resisting
determined dropping point. Dropping
oxidation can be determined.
point can help establish if the correct
grease was supplied or is in use, and
Figure 3. DSC Scan of Used Grease Sample determine if a used grease is good for
Differential Scanning Calorimetry (DSC) continued service.

Contamination of Used Grease

Many bearings fail prematurely due to
contamination. Grease contamination
can come not only from common envi-
ronmental contaminants such as dirt
and water, but also cross-contamination
from other grease types. This is a major
issue with greases because many gelling
agents are incompatible, resulting in
either a significant change in consistency
(either thicker or thinner), or a separa-
tion of the oil from the gelling agent.
There are a number of ways to deter-
mine the presence of contaminants in a
used grease sample.
Contamination from water or other
grease types can be identified by Fourier
Transform Infrared Spectroscopy (FTIR)
Analysis (Figure 4). FTIR can also
measure gelling agent type and
concentration, along with oxidation
by-products.
If cross-contamination with different
Figure 4. Fourier Transform Infrared Spectrometry (FTIR) Analysis types of grease is suspected, it may also

22 Practicing Oil Analysis • March - April 2002

be feasible to perform elemental analysis (after acid digestion) to look for
common metals present in the gelling agent. For example, a grease that is
supposed to be an aluminum complex grease, but has become contaminated
with a calcium sulfonate complex grease will show both aluminum and calcium
in spectroscopic analysis, indicating a contamination problem.

Alert and Alarm Management Methods

The methods for establishing criteria, and methods for alert and alarm values
vary according to the experience of the user. Alert values are those considered to
be above or below the norm, while alarm values are those beyond a safe oper-
ating level. One of the primary concerns in establishing a new database is setting
the criteria at a level that will alert the operator to a change in the machine or
lubricant, give that operator enough advance warning to make a timely deci-
sion, yet realistic enough so that premature alerts do not occur.

Absolute Values
Absolute values, also referred to as fixed or hard numbers, may be assigned to
any characteristic. These values are based on the equipment type and grease type
and grade. In some cases, fixed values may be obtained from the original equip-
ment manufacturer. In cases where there are no recommended values, the fixed
limit may be set using the experience of the laboratory with the specific lubri-
cant and machine combination. It is important to remember that hard number
alert values are a place to start a program, which contain many unknown
factors.
Continued on Page 24

Elemental Analysis ICP-AE Parts-Per-Million

Sample Fe Al Cu Si
8 0 0 0 39
1 8 7 3 17
3 11 0 0 18
6 12 4 28 29
2 18 6 80 17
4 31 5 34 18
11 32 0 0 48
5 40 0 120 28
18 68 0 0 10
10 84 0 10 43
7 89 0 32 34
17 144 29 458 230
9 163 0 24 40
12 234 98 540 323
14 300 43 234 167
13 345 87 340 830
15 378 16 253 230
16 628 28 8 1110
Avg. of Normal 36 2 28 27
STDEV of Normal 31 3 39 12
STDEV of Normal + Avg. 67 5 67 39 OK
2x STDEV + Avg. (Normal) 98 8 106 51 ALERT
3X STDEV + Avg. (Normal) 129 11 145 63 ALARM

Table 2. Setting Alarms for Metals Analysis

Practicing Oil Analysis • March - April 2002

Continued from Page 23 on obtaining correct and consistent operating time values,
During the initial phase of the program, it is not uncommon normally hours. It is also more likely to be invalidated by other
for the alarm values to remain unchanged and invalidated for a variables such as inconsistent sampling techniques.
substantial period of time. If the alarm value set is not appro-
priate for the machine in its unique configuration, the risk of Setting Limits
machine failure remains higher than acceptable. In the sample seen in Table 1, it is easy to see which of the
samples have high iron, aluminum, copper and silicon. The
Percent Change alarms set are based on 18 samples from various parts inside of
For some tests, such as oil versus gelling agent, it is more a bearing cavity from different wheel bearings. Many times (as
appropriate to set values on a percentage change rather than in this case) a data evaluator must make an initial judgment
standard deviation. An advantage of this type of alert is that it about what is considered normal. After sorting the data set by
does not require valid statistical populations if the baseline is iron, it is clear there is a break at 144 ppm. Considering all of
considered. Many percentage alarms can be converted to the samples lower than 144 ppm of iron as normal, the basis for
absolutes when the baseline value is a known quantity or the the analysis is established.
test has published typical values.
■ OK Samples: The average (Avg.) of all normal data is added
Statistical Analysis to the standard deviation (STDEV) of all normal data. These
Statistical analysis of wear metals is effective on mature data- samples are considered OK.
bases. This requires a statistically valid population (typically 30
data points or more); therefore, it is normally based on like ■
Abnormal Samples: Two times the STDEV of all normal
equipment in a group rather than a single piece of equipment. samples plus the Avg. These samples are considered ALERT.
Once sufficient historical data for the single machine is avail-
able, statistical analysis may be applied to the machine alone. ■ Critical Samples: Three times the STDEV of all normal
samples plus the Avg. These samples are considered ALARM.
Trend Performance or Rate of Rise
Setting alert values based on the data’s rate of rise, or on Measuring Low Concentration of Water
the slope of the curve, for a specific wear metal above a FTIR can identify the presence of water. However, it is not
predetermined minimum threshold value can be accomplished sensitive to low levels. Water in the parts-per-million (ppm)
after the initial three sets of data are entered into the database. range can be measured using a variation of ASTM D6304 -
The logic behind the three histories is simply in that it takes a Standard Method for Determination of Water in Petroleum
minimum of three points to calculate a curve. While this can Products, Lubricating Oils, and Additives by Coulometric Karl
provide additional information to the analyst, it relies heavily Fischer Titration. This method allows for the distillation of

Figure 5. Ferrographic Analysis (left) Shows a Large Concentration of Abrasive Silica in the Grease. The Ferrogram on
the Right Shows Severe Cutting Wear. This Information Helps Identify the Root Cause of Premature Bearing Failures.

24 Practicing Oil Analysis • March - April 2002

water using a distillation tower at 120°C (248°F) into a titration vessel where it is
solublized in toluene and sparged with nitrogen. The toluene/water mixture is then
titrated using Karl Fischer Reagent as per ASTM D 6304. The levels of detection
using this method are in the low ppm range.

Wear Debris and Particle Contamination

Conventional methods for measuring wear debris are ferrographic analysis and
elemental analysis. While the quantitative estimation of wear debris is difficult in a
used grease sample using elemental analysis, because of the difficulties of obtaining
a representative sample, ferrographic analysis, which by its very nature is a
qualitative technique, is ideal in determining the active wear mechanism and
severity of the problem in grease-lubricated bearings.
Ferrographic analysis on used greases is carried out by extracting the wear debris
from the sample and analyzing it visually using an optical microscope, similar to the
way ferrography is used for used oil samples (Figure 5). By looking at particle
morphology, it is often possible with ferrographic analysis to identify the root cause
of premature bearing failures, allowing appropriate corrective action to be taken.

Modern methods of analysis for used grease samples are rapid, sophisticated and
require only a fraction of the sample volume necessary in the past. Sound,
cost-saving maintenance decisions can be made using grease analysis as the basis for
preventive and predictive programs.

Method Determines
D217 - Cone Penetration Consistency of Grease
D942 - Oxidation Stability Useful Oxidation Life
D972 - Evaporation Loss Determine Operational Due to Evaporation
D1092 - Apparent Viscosity Flowability, Pumpability
D1264 - Water Washout Characteristics Water Washout Resistance
D1403 - Cone Penetration Quarter and Half Scale Supplement D217 (if not enough sample)
D1404 - Estimation of Deleterious Particles Abrasive Contamination
D1742 - Oil Separation Storage Stability
D1743 - Corrosion Preventive Properties Corrosion Resistance
D2266 - Wear Preventive Characteristics (Four-Ball Method) Performance Test
D2265 - Dropping Point Useable Temperature
D2509 - Load-Carrying Capacity (Timken Method) Performance Test
D2596 - Extreme-Pressure Properties (Four-Ball Method) Performance Test
D3704 - Wear Preventive Properties (Block on Ring) Performance Test
D5483 - Oxidation Induction Pressure DSC Useful Oxidation Life
D6185 - Compatibility of Binary Mixtures Compatibility
D5185 - Inductively Coupled Plasma Spectroscopy Evaluate Additives, Wear Metals and
Contaminants
Additional Tests
Method Purpose
Wear Particle Analysis Evaluate Wear Debris and Contamination
Fourier Transform Infrared Evaluate Grease Type and Contamination
Thermal Gravimetric Analysis D5967, Appendix 4 Evaluated Filler vs. Oil and Ash Levels

Table 2. Common ASTM Tests Used for Grease Analysis

Practicing Oil Analysis • March - April 2002