The infrared (IR) region of the spectrum lies to the right of the red end of the visible spectrum. We are unable to see this light although certain animals such as the pit viper can, enabling them to hunt at night. IR radiation was first described by William Herschel in 1800.
He produced a solar spectrum by placing a glass prism in the path of the sun's rays and observed the changes, which took place when light of different wavelengths (different colors) fell onto the bulb of a sensitive thermometer. He noticed that the temperature increased as the thermometer was moved from blue to red, but he also found that the thermometer registered even beyond the red end of the visible spectrum.
Subsequent experiments showed that this portion beyond the red was composed of a similar type of radiation to visible light, in that it could be reflected, refracted and absorbed by materials, which would reflect, refract and absorb visible light.
How is IR radiation utilized by the analytical chemist? All covalent chemical bonds such as those in organic molecules (CH4 or methane) as opposed to ionic bonds, and those found in inorganic molecules (NaCl or common salt) absorb IR radiation, causing them to vibrate by stretching and contracting. The strength of the chemical bond between the atoms, which in turn is influenced by their atomic structure, determines which part of the IR spectrum the molecules absorb.
An analogous way of visualizing this phenomenon is demonstrated by Hooke's law of springs, where the amount of energy to start the spring oscillating is related to the strength of the spring and the mass on the end. In this case, it is the energy of the absorbed IR and the nature of the bond between the C and H atoms. The amount of energy in the IR beam is related to its wavelength; the smaller the wavelength the more energy. Although in this case only the exact energy required to cause vibration is absorbed. All other energies both smaller and greater have no effect.
Therefore, for a molecule with several different kinds of bonds (for example, a C-H and a C=O), one would expect to see at least two different absorption bands. Chemical bonds within a molecule are therefore said to exhibit characteristic IR absorptions. It is this property that is utilized by the analytical chemist. Chemists refer to these absorbencies as wavenumbers.
This is a more convenient way of discussing the frequency of the absorbed radiation, and is simply the number of waves in one centimeter. The final piece to the equation is how much of this radiation is absorbed. This is given by a simple law called the Beer-Lambert law that states the amount of IR absorbed is proportional to the concentration of the absorbing species and the distance the IR light has to travel through it.
IR spectroscopy uses an electrically heated glowbar as the IR radiation source, and this radiation is passed through the sample to the detector. The chemical constituents of the sample absorb some of the IR light at reproducible and specific wavenumbers.
The original method involved using a prism or diffraction grating to separate the individual wavenumbers and then detect them, portions at a time, as they were passed through the sample, and plot the absorbance against the wavenumber. This process was incredibly slow and, depending on the accuracy required, could take as long as 10 minutes per sample.
Modern Fourier transform infrared (FTIR) uses the Michelson interferometer. This nifty device utilizes a moving mirror, whose speed is monitored by a laser, which also acts as a wavelength reference. The detector then measures the summation of all the frequencies over time resulting in a time-dependent interference pattern called an interferogram.
A computer algorithm called a fast Fourier transform is then used to convert this signal to an absorbance spectrum. This is then ratioed to a background spectrum of the empty cell to remove the contribution of atmospheric contaminants such as CO2 and water vapor. This whole process takes as little as 1½ seconds per scan which allows for multiple scans on the same sample and for amplifying signal differences so that minute variations can be detected, giving greater accuracy.
Several years ago, Wearcheck purchased a new Biorad FTIR. This represented two major changes in the methodology for determining oil degradation and combustion by-products.
Spectral subtraction was replaced with computational interrogation of the IR spectrum and the resultant data trended.
The horizontal attenuated total reflectance (HATR) cell was replaced with a 100-micron transmission cell.
Spectral subtraction is essentially the subtraction of all the IR spectra data from a sample of the new oil from the IR spectra data of the used oil, essentially resulting in the difference between them. This difference in data is used by software routines to calculate numbers representative of the degree and type of degradation or contamination which has occurred (Figure 3).
There are various computer algorithms which can automate this process, thus removing any operator interpretation or bias. However these algorithms cannot address the strict requirement for using the correct new oil in the subtraction process.
Such a requirement complicates the overall laboratory procedure because the exact new fluid placed in the machine must be submitted, tracked, stored and correctly recalled and remeasured with all later samples from the same machine. Additionally, oil in machines is topped-up periodically to compensate for oil consumption which further complicates the subtraction process. Although the top-up oil may be the correct type, it may not be the same manufacturer, lot number or even the same blend. These complications will inevitably produce misleading or incorrect results.
Wearcheck has therefore adopted a simple trending methodology to eliminate the problems associated with spectral subtraction. In this method, areas under the IR curve are measured and reported. The key to successful implementation is the careful selection of appropriate areas to be measured. (This work was carried out by Biorad and is embedded in the FTIR control software.)
These measurements can then be compared to either the expected IR response from similar or identical machine components, or to the set of previous IR measurements from the same machine. If a particular parameter is maintaining a constant value and no adverse wear or performance degradation is apparent, then there is no reason why that parameter should be unacceptable. However, what is acceptable in one component performing a particular type of operation may not be acceptable in another component performing a different type of operation.
As long as the overall analysis with respect to trend remains constant, a state of normality is assumed. This trending methodology is already utilized in wear metal analysis by ICP. Using this method and the conversion of spectral data into numerical condition indicator data simplifies tracking and documentation and greatly reduces storage requirements because neither the new oil sample or its spectrum needs to be saved.
FTIR was initially performed using a horizontal attenuated total reflectance (HATR) cell. This was a horizontally mounted zinc selenide (ZnSe) crystal. The oil sample was placed on the cell and an IR beam bounced off the bottom of the crystal, penetrating slightly into the oil on each bounce. The sampling depth was effectively about two microns per reflection. This however had several major disadvantages:
The cell could be easily scratched by metal particles found in the used oil.
Low sensitivity was experienced due to the small beam penetration.
It was manually intensive, with potential exposure of the operator to cleaning solvents.
A more recent addition to sampling techniques is the transmission cell. This consists of two ZnSe crystals separated by a 100-micron spacer. The oil sample occupies the space between the two crystals and the IR beam passes directly through the cell and sample to the detector. This has the following benefits:
A twentyfold increase in sensitivity.
The filling and cleaning of the cell can be automated.
Cell damage is eliminated by an inline filter designed to remove particles big enough to scratch the crystals.
In all lubricating systems, organic compounds exposed to high temperatures and pressures in the presence of oxygen will partially oxidize (react chemically with the oxygen). There are a variety of by-products produced during the combustion process such as ketones, esters, aldehydes, carbonates and carboxylic acids, and the exact distribution and composition of these products is complex.
Some of these compounds are dissolved by the oil or remain suspended, owing to the dispersive additives in the oil. Carboxylic acids contribute to the acidity of the engine oil and deplete its basic reserve as neutralization takes place. The net effect of prolonged oxidation is that chemically the oil becomes acidic causing corrosion, while a physical increase in viscosity occurs.
FTIR determines the level of oxidation by a general response in the carbonyl (C=O) region of between 1,800 to 1,670 cm-1 (Figures 1 and 4). In this region, IR energy is absorbed due to the carbon oxygen bonds in the oxidized oil. Very few compounds found in new petroleum lubricants have significant absorbencies in this area. Monitoring this region is thus a direct measurement of the oxidation level, as compared to secondary technique such as the acid number (AN), which takes into account all the acidic species in the oil.
Figure 4. Absorption Wavenumbers for FTIR
In addition to oxidation products, nitration products are also formed when organic compounds are exposed to high temperatures and pressures in the presence of nitrogen and oxygen. These are generally in the form of nitrogen oxides such as NO, NO2 and N2O4. In addition to causing oil thickening and some of these products being acidic, nitration products are the major cause of the buildup of varnish or lacquer. An increase of the nitration index of an engine oil can indicate mistuning (incorrect fuel/air ratios) or improper spark timing. It can also reflect operating conditions, such as high loads and low operating temperature, as well as piston ring blow-by.
Nitration products can be monitored by FTIR because they have a characteristic absorbance between 1,650 to 1,600 cm-1, the region immediately below that of the oxidation products.
Sulfur compounds are typically found in crude oils and may also be used as additives in fuels and lubricating oils to achieve certain desired properties. Sulfate by-products such as SO2 and SO3 are formed by the oxidation of these sulfur-containing compounds. They subsequently escape into the lubrication system around the piston rings and seals and build up over of time.
These compounds increase the production of varnish and sludges and generally degrade the oil. They also react with the water formed during combustion to produce powerful inorganic acids such as sulfuric acid (H2SO4). These acids are neutralized by, and therefore deplete, the basic reserve in the additive package of the oil. Measurement of these compounds gives additional information on mistuned engines and ring failures. Sulfates are measured by FTIR in the same way as oxidation and nitration, by monitoring the increase in their characteristic IR absorbances, found between 1,180 and 1,120 cm-1.
Soot is created from the incomplete combustion of the diesel fuel. Burning a too-rich fuel/air mixture forms soot particulates. An increase in the soot content of the oil indicates combustion problems, or that the drain period may have been extended. Soot buildup is a problem in lubrication oils because it changes the viscosity and prematurely clogs the filters and oil galleries.
The FTIR analysis of soot is an exception to the general approach, that the area under the curve indicates the amount of other contaminant parameters, because soot lacks any specific IR absorption bands. Instead, the soot particles cause a general scattering of the IR radiation, which is more severe at higher wavenumbers. Therefore, soot loading is simply measured by taking the absorbance intensity at 2,000 cm-1.
Automotive fuels can consist of a wide variety of branched aliphatic compounds such as octane, aromatic compounds such as benzene, and many other compounds blended to produce a desired set of physical properties. This is especially so in countries such as South Africa where the fuel comes from either coal-derived or petroleum-derived sources. In addition, engine conditions will cause the overall composition of the fuel to change, partly due to incomplete combustion and partly due to the distillation of the lighter fractions.
In an ideal situation, the choice of the fuel remains constant and FTIR becomes a powerful tool in detecting fuel dilution. This is accomplished by measuring the absorption bands of the specific components of the fuel and the drop in the absorption bands of the oil as it is diluted (Figure 2).
In light of the real and nonideal situation, fuel dilution is usually determined by flashpoint measurements or gas chromatography (GC).
Additionally, FTIR is used as a screening tool for water and base number (BN) measurement of engine oils.
Even though water is a by-product of combustion, it tends to be a rather infrequent contaminant (approximately one percent of samples) because of the hot operating conditions in the average engine; however, it could indicate a coolant leak. When present, water readily dissolves or disperses in the oil. Water is a strong IR absorber and is easy to detect.
It also appears in a region of the IR spectrum where few other compounds that appear in petroleum oils will have significant absorbencies. Due to the increased difficulty of performing crackle tests on engine oils where boiling the fuel present might mask the crackle of any water, and because all these samples undergo IR testing as standard procedure, it was deemed prudent to use FTIR as a convenient screen for the presence of water. Of all engine samples submitted only five percent are trapped for an actual water test and of these, only 20 percent contain actual water.
Due to the potential health hazards and environmental considerations, it is not possible to measure actual BN on all engine samples that come into the laboratory. For this reason, IR is used as a screen and only those predicted to be below 6.0 will have an actual BN measurement.
The BN of an oil sample cannot be easily defined by IR analysis. BN depends on a wide range of factors with varying degrees of influence. Principle component regression and partial least squares (PCR/PLS) analyses are mathematical routines which have allowed the laboratory to predict the apparent BN of a sample. This method uses the whole spectrum instead of an individual peak or discrete area to derive a value for an unknown.
A series of 80 oils with duplicate BN values was used to create a training set, to establish the measurement criteria. These oils came from various engines, various lubricant manufacturers, differing grades and were not limited to any particular application. The software was used to break down the training set into a smaller set of principal components or factors. These factors were then integrated to predict the unknown. This process avoids unnecessary wastage and consequently only seven percent of samples have an actual BN measurement performed. Of these more than 70 percent fall in the 3-to-7 range.
Although FTIR provides a wealth of information about the condition of used lubricating oil, this information is complementary to that obtained by various other spectroscopic and physical property tests to give an overall picture of the condition of the oil and machinery in which it is used.
B.W. Cook and K. Jones. A Programmed Introduction to Infrared Spectroscopy, 1972.
B. George and P. McIntyre. Infrared Spectroscopy, 1987.
JOAP international condition monitoring conference, 1994.
Condition monitoring conference, 1994.
A. Geach. "Infrared Analysis as a Tool for Assessing Degradation in Used Engine Oils." Wearcheck Technical Bulletin No. 2.
J.R. Powell and A.M. Toms. "Molecular Analysis in Engine Condition Monitoring." Biorad presentation material.
J.R. Powell and A.M. Toms. "Molecular Analysis of Lubricants by FTIR Spectroscopy." FTS/IR Notes No. 114, 1997.
This article was originally published by Wearcheck Africa, a member of the Set Point group.