Spectroscopy is the study of how light and matter interacts. Light can be broken up into several regions as illustrated in the electromagnetic spectrum in Figure 1.
Spectroscopy is perfectly suited for accurate real-time analysis and monitoring of continuous and batch processes. The near infrared (NIR), as well as the ultraviolet and visible (UV-VIS) regions of the spectrum provide a wealth of information about many chemical and physical properties to help operators control their processes better. No other technology matches the acquisition speed and the range of measurements returned by a fiber optic-based spectrometer. A comprehensive set of data that takes hours of laboratory analyses to acquire can be available in less than a minute.
Figure 1: The electromagnetic spectrum between 190 nm to 25,000 nm
Early spectrometers used prisms to separate the wavelengths of visible light. However, as scientists moved from studying visible and near-infrared light to the mid-infrared region, they discovered that prisms and the gratings of the early 1900s were slow, noisy, and lacked resolution. The solution for mid-infrared spectroscopy was the interferometer designs of Michelson.
However, in the late 20th century the electronics, optics, and other components inside a spectrometer system improved. Modern dispersive spectrometers use highly engineered gratings to separate the wavelengths of light and reduce stray light. A diffraction grating is essentially an aluminum-coated mirror with thousands of parallel and equally spaced grooves etched into its surface. As such, a grating is one of the most precise objects ever made. How gratings work is described in all good freshman level college physics textbooks.
In a peer-reviewed study, Coates found that both Dispersive Analyzers (DG-NIR) and Fourier Transform Spectrometers (FT-NIR) have equivalent performance (Coates, 1994 NIR news DOI:10.1255/nirn.250). This study shows that with modern technology many of the marketing claims by manufacturers of the advantageous of FT-NIR Spectrometers are no longer valid. How do you decide which spectrometer technology is the best choice for your analysis?
Which Technology is a Better Choice?
FT-NIR is a powerful technique, especially in a laboratory setting where samples can be introduced to the spectrometer without using optical fiber. FT’s also work well for measuring samples in reflection such as powders or solids. However, when considering NIR for a process application that involves a “clear” liquid or gas using optical fibers, FT-NIR instruments lose many of their advantages.
The bandwidth of spectral features varies with the state of the sample. Gases and vapors have spectral features (lines) that are very narrow and often require high resolution to see. Liquids and solids have very broad spectral features (bands) due to the hindered rotation of the molecule in its matrix. Thus for clear liquid hydrocarbon samples (the bulk of NIR applications), bandwidths range from 2 nm to 12 nm. Aqueous samples are even broader. Thus for condensed samples, the higher resolution of FT-NIR is not required nor desirable since high resolution always comes at the cost of signal-to-noise ratio (SNR). The latter property for chemometric analysis of complex hydrocarbon mixtures is more important for accurate analysis than resolution.
A scanning grating dual-beam (DG-NIR) spectrometer actually produces a superior signal-to-noise ratio (i.e., sensitivity for lower detection limits) and greater stability to withstand ambient air fluctuations. In addition, the dual-beam operation provides the added benefit of superior long-term stability required of process analyzers that must operate unattended for weeks. Dual-beam operation is easy to implement in a grating instrument but difficult in an FT-NIR analyzer. Most FT-NIR analyzers do not operate in dual-beam mode hence require more frequent referencing to compensate for photometric drift.
Near-Infrared (NIR) is the region of the electromagnetic spectrum from 750 nm to 2600 nm. Molecules containing C-H, O-H, and N-H bonds absorb NIR radiation in specific
regions or at specifi c wavelengths. Each molecule has a unique spectrum. These absorbances are used to measure (predict) chemical and physical properties of the sample. Applications for NIR (n-line, at-line and laboratory) spectroscopy are numerous and cover a wide range of industries included but not limited to: chemical, refining, pharmaceuticals, polymer, semiconductor, and agricultural.
Fourier transform (FT-NIR) spectroscopy does not record the spectrum of the sample directly but records an interferogram based upon time domain measurements. A spectrum is produced by performing a Fourier transform of the recorded interferogram. Albert Michelson knew how to do this but lacked the recording electronics and computers we now have.
This uses Dual-Beam, Dispersive Grating technology for NIR measurement. This state-of-the-art design eliminates, or minimizes many of the problems associated with previous grating spectrometer designs, while retaining Guided Wave’s time proven dual-beam design and built in multiplexing. For this technical note, we will describe a scanning grating system that incorporates the following features: post-dispersion, dual-beam, high efficiency blazed plane grating with on-axis aberration free, high optical throughput transmissive optics.
A single spectrum consists of two scans, one for the sample channel immediately followed by a reference channel scan, i.e. dual-beam spectroscopy. The ratio of these two scans provides a stable spectrum. Dual-beam operation in a spectrometer removes nearly all common mode drift problems in detector, lamps and electronics. This reduces the required frequency of collecting reference (zero) spectra from hours to weeks which also reduces operator involved maintenance. It also improves baseline stability and spectral quality. Beam switching between sample and reference fibers or between sample channels does not involve moving optical components, hence there is no noise introduced in the dual-beam or multichannel operation.
Advancements in Spectroscopy
Since Coates’ report in 1994, dispersive analyzers have taken on another level of advancement with the development of Guided Wave’s full spectrum post dispersive planar grating, dual-beam (DG-NIR) technology. These advancements applied to NIR process
analyzers will ensure accuracy due to the reduction of stray light and excellent signal-to-noise ratio.
Post-Dispersed Design Lowers Black-Body Impact
Post-dispersed design means that ambient or black-body radiation will be dispersed like all other radiation, such that at any given wavelength, its impact will be much smaller.
Plane Grating Improves Efficiency
Most commercial grating spectrometers use concave holographic gratings because the optical systems are very simple. Concave gratings always introduce off-axis aberrations such astigmatism and coma which rob light from the image. Holographic gratings are hard to blaze, thus not very bright at the angles of use, again robbing light from the image. Instead of using a concave grating GWI uses a highly blazed plane grating which is a more efficient design. Collimation and focusing are provided by a pair of triplet achromatic lenses. These lenses produce a virtually aberration-free image of the source fiber in the focal plane of the spectrometer. The result is a spectrometer of exceptional brightness, i.e., throughput, hence, a very high signal-to-noise ratio.
FT-NIR vs. DG-NIR
FT-IR (not NIR) spectrometers are definitely superior to grating spectrometers in the energy limited infrared region. However, the near infrared is not energy limited so many of the advantages of FT technology do not apply. This has led to many misconceptions or myths (listed below) associated with NIR spectrometer technologies.
In the NIR region, FT-NIR spectrometers offer no significant advantages over DG-NIR spectrometers, and many times are not as accurate,
efficient or economical as DG-NIR multi-channel, dual-beam analyzers.
FT-NIR Misconceptions and Facts
FT-NIR is a newer technology
|The fundamental technology of FT systems and dispersive analyzers were both developed in the 1800s.(Michelson interferometer – 1887, Henry Joseph Grayson grating ruling engine – 1899). Both technologies became feasible for process applications with the development modern telecom fibers and detectors, high quality optics, and the advent of the PC. Both use the same high quality optics, detectors, fibers, and light sources.|
FT-NIR has easier calibration transfer
|Both FT-NIR Systems and DG-NIR Analyzers can directly transfer calibrations between channels. The method of light dispersion is not relevant to thr success of calibration transfer. Instrument-to-instrument repeatability in terms of the fundamental characteristics (bandwidth, stray light, wavelength axis accuracy) are key in successful calibration transfer. FT-NIR will use their laser source to maintain wavelength accuracy, while DG-NIR instruments use temperature compensated filters with NIST traceability.|
FT-NIR has lower error in calibrations due to better wavelength resolution
|In the near infrared region the small increase in resolution by FT-NIR does not translate into lower error calibrations. (Armstrong, 2006 Applied Engineering in Agriculture. 22. DOI:10.13031/2013.20448)|
|TRUE Somewhat: Fellgett Advantage – scan time||FT-NIR measure all wavelengths simultaneously while scanning grating systems measure one wavelength at a time. This theoretically gives the FT-NIR a “multiplexed” advantage which improves the SNR. However, since all of the light falls on the FT detector, it is often driven non-linear and the light must be attenuated. The reality is that grating spectrometers can have a SNR that is equal to or superior to a comparable FT-NIR system.|
Jacquinot Advantage – higher light throughput
|If the FT-NIR system is configured to measure through a fiber optic cable, then the aperture or throughput of light is limited by the diameter of the fiber optic cable which is essentially the same for both types of instruments. This eliminates any potential advantage for FT-NIR online process monitoring.|
Connes Advantage – wavelength accuracy
|FT-NIR systems use a single NeHe laser to verify the wavelength accuracy. FT-NIR wavelength accuracy depends on the precision of the alignment between the laser beam and the white light source and the number of zero crossings measured of the laser fringe, i.e., the resolution at which the spectrum is recorded.
Dispersive analyzers use NIST traceable standards to check the accuracy of multiple points along the wavelength range. The wavelength accuracy is limited by the precision of the NIST standards and the reproducibility of the grating drive mechanism. (Armstrong, 2006
Applied Engineering in Agriculture. 22 DOI: 10.13031/2013.20448)
Additional Considerations and Comparisons: Analyzer Validation
An important consideration for successful process monitoring is the ability to continually monitor the accuracy and precision of the system, thus ensuring the analyzer is producing validated spectra for your applications.
With FT-NIR analyzer validation is often done using external fluids (Pentane and Toluene) which is rarely available and is an expensive consumable. Pentane is a wash fluid. Spectroscopic Grade Toluene is required as the validation sample. Industrial grade Toluene cannot be used for this purpose. Validation can be automated or run manually but requires additional plumbing to inject the sample on the probe. Thus validation reduces the analyzer up-time and adds complexity (potential failure points) to the sample handling system. With a Guided Wave’s DG-NIR analyzer the validation system is simple and requires no maintenance or consumables. Using the optional Stability Monitoring System (SMS) in the analyzer, there is no need to interrupt the other channel operation. It provides automatic and continuous analyzer validation according to ASTM methodology.
The ongoing costs and ease of use associated with any instrument is an important consideration. Both FT-NIR and DG-NIR use tungsten-halogen lamps as the light source and an InGaAs detector.
For DG-NIR the lamp replacement is typically every six months and it is the only consumable needed. The replacement of this light source can be completed by any person in a matter of seconds, as lamps are all pre-aligned.
FT-NIRs also require that the lamp be periodically replaced. Furthermore, the laser has a finite lifetime and occasionally needs replacement. Replacing the laser not necessarily simple as its alignment to the white light beam from the lamp is critical. Thus laser replacement is often done by a factory trained service engineer.
Multiplexing Facility - The Input Module
FT-NIR spectrometers can be multiplexed (multi-channel operation) but to do so often requires fiber multiplexers with moving optical elements. It is not possible to move optical elements and not introduce some noise in the system. On-line spectroscopy often requires SNRs > 105 which exceeds the capability of moving optical element multiplexers. The cost of the additional hardware limits the number of channels to between 2 and 8.
An alternate method of multiplexing is stream switching. This involves an extractive sample system with motor operated valves and possible cross contamination in the sample cell. This is a slow, high maintenance approach.
Guided Wave’s DG-NIR analyzers have built in multiplexing with no moving optical elements. Thus there is no degradation in the SNR. A twelve channel DG-NIR system can switch between samples in seconds.
Process analyzers are expected to operate 24/7 with minimal maintenance. Moving parts in an analyzer are therefore always looked on with suspicion. Fortunately, most modern spectrometers have long mean time between failure (MTBFs) on the order of years. Both FT-NIRs and scanning grating spectrometers have critical moving parts. FT-NIR spectrometers have one or two oscillating mirrors that provide the phase encoding of the spectrum. If these mirrors do not move smoothly or fall out of alignment, faulty spectral results can and do occur. Similarly, scanning grating spectrometer must rotate the grating precisely and measure that rotation with a precision optical encoder. Again failure of the mechanism can result in bad spectra. However, the reliability of FT mirror mechanisms and grating drives are exceptional with years of trouble free service expected from both.
Conclusion: DG-NIR Advantage
When considering NIR for an application that involves a “clear” liquid or gas, a dispersive NIR Spectrometer, DG-NIR is the superior choice. By developing NIR analyzers with dual-beam, post-dispersive planar gratings, the engineers and scientists at Guided Wave have advanced the state-of-the-art in dispersive NIR technology. By incorporating these advancements without compromising the dual-beam operation, Guided Wave can offer to the market a DG-NIR analyzer with superior accuracy and resolution. The DG-NIR advantage is due to the way these analyzers are designed to control light and minimize all forms of aberrations DG-NIR system has been carefully optimized to provide exceptional signal-to-noise ratio, excellent long-term photometric and wavelength stability, built-in multiplexing, and ease of maintenance.
Guided Wave has been a leader in online, process monitoring for over 35 years. Established in 1983, Guided Wave was recognized as an industry leader when it delivered the first fiber optic-based NIR analyzers. Today Guided Wave has NIR analyzer installations on six continents and in more than 50 countries, with thousands of analytical instruments sold worldwide for the Chemical, Refinery, Pharmaceutical, Polymer, Semiconductor, and Sterilization industries.
Guided Wave is the only process NIR manufacturer that provides a complete optically matched system, yielding the best throughput efficiency and long-term performance, exceeding industry standards. These workhorse analyzers have been industry-proven for over three decades with individual analyzers in the field typically running 24/7 lasting more than 10 years with >99% uptime.