Infrared spectroscopy offers many unique advantages for measurements within an industrial environment. Historically, the technique has been used for a broad range of applications from the composition of gas and liquid mixtures to the analysis of trace components for gas purity analysis. In cases where a well-defined method of analysis is involved, such as the measurement of combustion gases for carbon oxides and hydrocarbons, simple filter-based instruments, or non-dispersive infrared (NDIR) photometers are widely used. These offer simplicity of measurement concept at a potentially lower cost. However, they are rigid in design (dedicated to single measurement) and are moderately expensive to maintain in terms of the need for frequent calibration. For more complex measurements, involving multiple chemical species for either environmental or process stream applications, the use Fourier transform infrared (FTIR) instruments has become relatively popular. FTIR instruments offer flexibility in terms of the ability to customize and optimize the measurement for a wide variety of applications. However, this comes with a significant penalty in terms of cost, size, “robustness’, and an associated operational complexity.
An ideal alternative is proposed in this article, where an applications specific analyzer is offered with many of the attributes of an NDIR instrument in terms of size and cost, yet with many of the desirable attributes of an FTIR spectrometer. Like an NDIR analyzer the spectrometer works with a predefined set of wavelengths or wavelength regions with band passes optimized for the specific application. Up to 256 of these wavelength regions may be selected to define an analytical system, however, for most average applications as few as ten may suffice. Like an FTIR instrument, the individual wavelengths used are uniquely modulated, and the final spectral intensities are computed via a Fourier transform. The application and the required wavelengths for that application may be defined by the end-user, and may be customized for optimum performance for that application. Multiple analytes (components) are handled by an integrated chemometrics function (implemented in firmware), and the method, once finalized, is recorded on a medium that is similar in appearance to a compact disk (CD). The technology utilized is known as applications configured encoding (ACE). This provides a means to standardize the method, where multiple copies of the method may be readily generated for use on other ACE analyzers. Furthermore, unlike an NDIR analyzer, the spectrometer is easily reconfigured in the field for a different analysis, simply by changing the ACE disk.
This article describes the fundamentals of the ACE technology. Although the system may be configured to accommodate measurements in other spectral regions, merely by using different source and detector combinations, this article will primarily focus on the mid infrared. The analyzer is capable of addressing the needs of a wide range of applications based on the type of sample handling that is interfaced to the system. Different types of analyses are addressed in the text, however, from a practical point of view, special reference is made to use the use of the analyzer for gas phase monitoring applications.
Infrared spectroscopy is one of the oldest spectroscopic measurements that are used to identify and quantify materials in industrial and environmental applications. Traditionally, for analyses in the mid infrared, the technologies used for the measurement have been limited to fixed wavelength (NDIR –Non-Dispersive Infrared) filter-based methods and scanning methods based on either grating spectrometers or interferometer-based (FTIR) instruments. The last two methods have tended to be used more for instruments that are resident in the laboratory, and filter instruments have been used mainly for process, field-based and specialist applications, such as anesthesia gases in the hospital operating room.
Very few scanning grating instruments were ever used for mid infrared applications in the nonlaboratory applications, and no further commentary in regard to their application outside of the laboratory will be made. Originally, the same applied to FTIR instruments, although in the past 10 years there has been a strong desire to move instruments away from the laboratory and towards point of use applications, either at-line, on-line or field-based. These are primarily associated with the great flexibility associated with measurements made on an FTIR instrument, and the performance characteristics of spectral resolution, speed and optical throughput, when compared to traditional grating instruments.
Moving an FTIR instrument out of the benign environment of a laboratory to the more alien environment of either a process line or that of a portable device is far from trivial. Major effort has to be placed on the instrument design in terms of both ruggedness and fundamental reliability of components. Furthermore, issues such as environmental contamination, humidity, vibration and temperature are factors that influence, negatively the performance of FTIR instruments. As a consequence, FTIR instrumentation that is to be successfully applied in process and field-based application MUST be designed to accommodate these factors from the start. This comes at a premium, in terms of design requirements and cost. Issues are packaging and making the system immune to effects of vibrations and temperature variations, the need to protect delicate optical components, such as beam splitters, and the overhead associated with making a high-precision interferometric measurement.
At the other end of the scale there are the filter-based instruments. With the exception of the scanning filter instruments based on CVF (circular variable filter) technology, originally designed by Wilks, and later marketed by Foxboro, most filter instruments feature one or more fixed wavelength (based on center wavelength) measurements. Various designs have been made ranging from filter wheel based instruments, where a common detector is used, to systems with combinations of filters and detectors. Neither of these situations is ideal in terms of performance and stability. However, in spite of that filter, or NDIR, instruments account for the vast majority of instruments used for process and field-based applications. In most cases, with the exception of certain multi-component commercial analyzers, there is little customization and most commercial filter instruments a designed for fixed applications with a fixed set of common analytes, such as total hydrocarbons, carbon oxides for combustion applications, halothanes for anesthesia gas applications, etc. Few of these applications are truly challenging, and the filter photometer meets the basic needs of these applications in terms of the simplicity of the fundamental measurement concept and the general low cost requirement. However, in spite of the perceived simplicity of such analyzers, practical implementation problems exist.
For many years in the past, the norm in process analyzer implementation was that NDIR analyzers were used for simple well-defined applications, and process gas chromatographs (GCs) were used for handling complex mixtures. However, process GC has a huge overhead in terms of support (the need for carrier gases) and maintenance (replacement of columns, etc.), and in general process engineers live with this, but prefer the ease of implementation of an optical method. This has resulted in a desire to shift towards the use of FTIR instruments in certain process applications, where an advanced spectrumbased analyzer can handle more complex analyses. However, this complicates the maintenance issue because although implementation may be easier, the methods development and support now requires the support of a spectroscopist. As a result, very few FTIR implementations can be made in process and field-based applications without the back up of a support laboratory. This in many ways negates the benefits of going to a spectral/optical method of measurement.
Given that there is a continuing need to measure complex mixtures in a non-laboratory situation, and given the desire to have a good practical spectral solution it is time to look at alternative technologies. There are indeed a number of new technologies for infrared-based measurements that have been proposed in recent years. One particular technology, known as applications configured encoding (ACE), which has been developed and implemented in an Advanced Photometric Analyzer (APA), offers a good practical compromise between the NDIR and FTIR. Conceptually, the technique that can be considered as a hybrid approach, providing the performance and flexibility benefits of FTIR, while providing the simplicity, ease of implementation and low cost associated with an NDIR-based system.
In brief, as currently applied, the APA is a photometrically simple system, where the incoming infrared beam from the sample is imaged on to a diffraction grating based spectrograph. The dispersed radiation from the grating is imaged across an aperture above the surface of a rotating encoder disk. The encoder disk has a series of reflective tracks, which are spatially located within the dispersed grating image to correspond to the wavelengths and wavelength regions used for the analysis. Each track has a pattern that produces a reflected beam with a unique sinusoidal modulation for each individual wavelength. The reflected beams are brought to an image on a single detector, which generates a signal that forms a discrete interferogram. The intensity contribution for each wavelength component is obtained by applying a Fourier transform to the interferogram. The results, in terms of analyte concentrations are determined from the measured intensities by the use of an embedded classical least squares function. The heart of the system is the encoder disk, which has the appearance of a CD with a series of concentric barcode-like bands, is used to optimize and customize the analyzer for a specific analysis. In its current configuration, covering the spectral range from 3.0 to 5.0 microns, the analyzer has a baseline noise level of better than 10-5 absorbance, and as an example is providing sub-ppm detection limits for hydrocarbon-based species with a 1 meter sample pathlength.
As an example, the analysis of mixed hydrocarbon gases, containing methane, ethane, propane and butane can be considered to be a challenging application for NDIR, and for a process stream containing such a mixture, one would have to resort to either process GC or an FTIR, if a spectral-based solution is required. The problem is handled by the APA system, confirming that the system can provide FTIR performance, while possessing the positive attributes of an NDIR system. This application will be discussed further with example performance data.