ASTM D7418 Standard Practice for Set-Up and Operation of Fourier Transform Infrared (FT-IR) Spectrometers for In-Service Oil Condition Monitoring
5. Apparatus
5.1 Fourier Transform Infrared (FT-IR) Spectrometer - All FT-IR instruments suitable for use in this practice must be configured with a source, beamsplitter and detector suitable for spectral acquisition over the mid-infrared range of 4000 to 550 cm(-1). The standard configuration includes a room temperature deuterated triglycine sulfate (DTGS) detector, an air-cooled source, and a germanium-coated potassium bromide (Ge/KBr) beamsplitter, although a zinc selenide (ZnSe) beamsplitter may also be used. The FT-IR spectrometer's IR source and interferometer should be in a sealed compartment to prevent harmful, flammable or explosive vapors from reaching the IR source.

NOTE 1 - Photoconductive detectors such as mercury cadmium telluride (MCT) should not be used owing to inadequate linearity of the detector response.

5.2 Sample Cell - The sample cell employed for in-service oil condition monitoring is a transmission cell with a fixed pathlength that can be inserted in the optical path of the FT-IR spectrometer. Cell window material and cell pathlength considerations are stated below.
5.2.1 Cell Window Material - ZnSe is commonly used as the window material for condition monitoring and is recommended because of its resistance to water. Sample cells constructed of materials other than ZnSe may be used; however, to address all the various methods associated with condition monitoring, the window material should transmit IR radiation over the range of 4000 cm(-1) to 550 cm(-1). KCl and KBr are common cell window materials that meet this requirement but these are water-soluble salts and should not be used if oil samples containing moisture are frequently run through the cell, as contact with water will cause the windows to fog and erode rapidly. In addition, Coates and Setti have noted that oil nitration products can react with KCl and KBr windows, depositing compounds that are observed in the spectra of later samples. On the basis of this report, KCl and KBr windows should not be used with samples of gasoline or natural gas engine oils as well as other types of lubricants where nitration by-products may form due to the combustion process or other routes of nitration formation.
5.2.1.1 When ZnSe is used as the window material, the reflections of the infrared beam that occur at the inner faces of the windows cause fringes to be superimposed on the oil spectrum; these must be minimized using physical or computational techniques as presented in Appendix X1. Because KCl and KBr have lower refractive indices than ZnSe, the use of these window materials avoids observable fringes in the oil spectrum.

5.2.2 Cell Pathlength - The standard cell pathlength to be employed for in-service oil condition monitoring is 0.100 mm; however, in practical terms, pathlengths ranging from 0.080 up to 0.120 mm are suitable, with values outside this range leading to either poor sensitivity or non-linearity of detector response, respectively. The actual cell pathlength obtained can be determined from the interference fringes in the spectrum recorded with an empty cell or by recording the spectrum of a check fluid; details for calculating cell pathlength are presented in Appendix X2. The reporting units of the various in-service oil condition monitoring parameter test methods are based on a pathlength of 0.100 mm (see the respective test methods). Accordingly, all data must be normalized to a pathlength of 0.100 mm, either by multiplying all data points in the absorption spectra by a pathlength correction factor (spectral normalization) or by multiplying the results of the respective test methods by a pathlength correction factor (see 10.2). The normalization procedure is usually part of the software provided by instrument manufacturers.
5.2.2.1 Discussion - However, if sample dilution is employed (see Appendix X3), longer pathlengths may become suitable. For example, for dilution with odorless mineral spirits (OMS) in a 2:1 OMS:oil sample ratio, a pathlength of 0.200 mm has proven suitable.

NOTE 2 - For purposes of interlaboratory comparison of results, spectral normalization should be performed.

5.3 Filter (optional) - The use of a particulate filter with a mesh size of 0.100 mm or less to trap any large particles present in the sample is strongly recommended to prevent cell clogging.

5.4 Sample Pumping System (optional) - A pumping system capable of transporting oil to be analyzed into the transmission cell and of emptying and flushing the cell with solvent between samples may be used instead of manual cell loading. Commercial vendors offer various pumping systems that may differ in the type of pump, tubing, and transmission cell. Depending on the sample handling system employed and the viscosity of the oils analyzed, a wash/rinsing solvent may be run between samples to minimize sample-to-sample carryover as well as keep the cell and inlet tubing clean; commercial vendors may recommend specific solvent rinse protocols.
5.4.1 Hydrocarbon Leak Alarm - When a sample pumping system is used, an independent flammable vapor sensor and alarm system is strongly recommended The purpose of this alarm system is to alert the operator when a leak occurs in the tubing, connectors or transmission cell.

6. FT-IR Spectral Acquisition Parameters
6.1 The spectral acquisition parameters are specified below. Because the spectral resolution, data spacing, and apodization affect the FT-IR spectral band shapes, these specifications must be adhered to:
Spectral resolution: 4 cm(-1)
Data spacing: 2 cm(-1)
Apodization: Triangular
Scanning range: 4000 to 550 cm(-1)
Spectral format: Absorbance as a function of wavenumber

6.2 The number of scans co-added and hence the scan time will depend on the desired spectral noise level (see Section 12), whereby an increase in scan time by a factor of N will decrease the level of noise by a factor of N(1/2).

7. Sampling
7.1 Sample Acquisition - The objective of sampling is to obtain a test specimen that is representative of the entire quantity. Thus, laboratory samples should be taken in accordance with the instructions in Practice D4057.

7.2 Sample Preparation - Filtering the sample using a filter described in 5.3 prior to loading the cell with the sample is highly recommended. An exception to this recommendation may be made when oil samples are diluted (see Appendix X3).

8. Preparation and Maintenance of Apparatus
8.1 Rinsing, Washing and Check Solvents - A variety of hydrophobic solvents may be used to clean the cell and rinse the lines between samples as well as serving as a check fluid to monitor pathlength. Typical solvents include hexanes, cyclohexane, heptane or odorless mineral spirits (OMS). Health and safety issues on using, storing, and disposing of check or cleaning/wash solvents will not be covered here. Local regulations and Material Safety Data Sheets (MSDS) should be consulted.

8.2 Sample Cell and Inlet Filter - The cell should be flushed with the designated rinse/wash solvent at the start and end of analytical runs to clean the cell. Immediately following flushing of the cell, an absorption spectrum of the empty cell (see 9.1.2.2) should be recorded to check for build-up of material on the cell windows. If an inlet filter is used, the filter shall also be checked for particle build-up and its effect on sample flow rate.

8.3 Check Fluid and Pathlength Monitoring - The purpose of a check fluid is to verify proper operation of the FT-IR spectrometer/transmission cell combination, as well as any associated sample introduction and cleaning hardware. It is recommended that an absorption spectrum of the check fluid be recorded when a new or re-assembled cell is initially used and archived to disk as a reference spectrum against which subsequent spectra of the check fluid may be compared. The spectrum of the check fluid may also be used to calculate the pathlength of the sample cell to normalize all data to 0.100 mm and to monitor changes in the cell pathlength over time, where significant changes may imply wear or contamination on the cell windows and should prompt remedial action. To serve as a check fluid, a solvent must have consistent spectral characteristics (lot-to-lot) and a measurable (on-scale) IR absorption band for cell pathlength calculation; for more details, see X2.2. One IR manufacturer uses heptane, another uses OMS, and other commercial products are available.