Document jgpvxozK9LZvnNomprrDv7Q75

I) l transforming IR spectroscopy ftiV Dr. R. O. Kag*l, senior anal, specialist, and Or. 5. T. King, analytical specialist. Analytical Laboratories, Dow Chemical Co. OUIUER TRANSFORM SPECTROSCOPY is a vi tions of the technique. Present day Fourier fable analytical technique with a bright Transform systems use a dedicated digital future and an illustrious past As we knowcomputer to process the data. In real problem it today, the Fourier Transform spectro solving situations, there must be an immedi meter is the end product of a most suc ate data display to adjust parameters, and in cessful mating, as A. J. Mittledorf (1) puts current systems spectra are either plotted or it, "between an old optical instrument, some almost instantaneously displayed on an oscil what older mathematical principles, and a loscope. new computer approach.' P. D. Fellgett is credited with introducing The optical instrument is the interfero Fourier Transform spectroscopy with his 1951 meter, which, at the time of its invention by dissertation at the University of Cambridge A. Michelson in 1887, generated perhaps one on the multiplex principle, the first numeri of the most important experimental results cally Fourier-transformed interferogram, and ever obtained. The result, of course, was the his work on infrared detectors. The high ener Michelson-Morley experiment that later gy throughput of the interferometer was proved instrumental in shaping modem theo pointed out by P. Jacquinot and the results ries of light and matter. The mathematical of these works were quickly exploited by H. principle is that of French mathematician, A. Gebbie, G. A. Vanasse, and J. Strong who Baron Jean Baptiste Fourier, by which a com demonstrated the utility of the digital com plex summation of cosine waves, an inter- puter to Fourier Transform far infrared inter- ferogram, is transformed into an optical ferograms. These early developments are spectrum. described in detail in an excellent account The mathematical manipulations are un by Lowenstein (2). wieldy, time consuming, and until the advent The theory of Fourier Transform spectros of small, fast digital computers presented a copy has been treated in detail by many formidable barrier to most practical applica authors (3) and will be outlined here only I f L F MICHCLSON Interferometer radiation is partially transmitted and partially reflected to the two mirrors Mi and M>. The reflected beams will either constructively or destructively recombine at the beam splitter depending on the position of the moving mirror M,. For monochromatic radiation, the detector will see a cosine signal whose amplitude is a function of mirror position. I L uce 034U19 '(A)^ti^rogpj^|nA|a 1 nt I Sfroma ii^n|pf%(mMreo.|ution feaaa^MWM %t#rf*rogrni)! "fwfflSlliS o.o ' -\l -tZ'-Y-k' 7V&,'' WF'f&itsZfjl . , ` .?. 0.71 1200 WHITT^ PY/K n |^ o' < 1 ; : ' ' i| uv' j J . $ - ;t-,,- Mi ` ` I, ~:i: 'a T ` V* 2600 rfej in a qualitative sense. The Michelson inter ferometer consists of two mirrors and a beam splitter. The beam splitter transmits_hal_of all incident radiation from a source to a mov ing mirror and reflects half to a stationary mirror. Each component reflected by the two mirrors returns to the beam splitter, where the amplitudes of the waves are combined. If the two mirrors are equidistant from the beam splitter, the amplitudes constructively combine. But if one mirror is moved a dis tance equal to A/4 (assuming monochromatic incident radiation of wavelength A), the emerging beam will be the result of two 180" out-of-phase beams that recombine destruc tively. Thus, for monochromatic incident radiation, the detector will see a cosine sig nal whose amplitude is a function of mirror position. We can extend this simple treatment to the case where polychromatic radiation enters the interferometer. The signal received at the detector is a summation of all the inter ferences as each wavelength component inter acts constructively or destructively with every other component. The resulting signal is an interferogram, which is a complex pattern of light amplitude (energy) as a function of distance traveled by the mirror. The relation ship between the intensity of the interferogram as a function of mirror travel, I(x), and the intensity of the source as a function of optical frequency, I(k), is given by a cosine Fourier Transform, I!: I(x) = / I(o) COS(2*X)dt> -- Oft and the inverse transform I(o) = / I(x) cos(2irt>x)dx -- 00 relates the interferogram to the optical spec trum. Taking advantage of Fellgett and Jaequlnot There are two major advantages of Fourier Transform systems over conventional disper sive spectrometers and these advantages are based on the speed and the throughput of the two instruments. The multiplex, Fellgett's advantage, has to do with the fact that all wavelengths of light simultaneously reach the detector. The interferometric signal (S) is directly proportional to the observation time (T) while the noise (N) is proportional to T'/`. The signal-to-noise ratio is, therefore, proportional to T7*. In a dispersive instrument each resolution element (M) is observed sequentially for an average time, T/M. The noise is proportional to -(T/M)1'* and the signal-to-noise ratio, to THE IASIC - experimental arrangement for emission experiments is quite simple. > Solid samples are mounted on a metal stand f between ? two vertical heaters and contained in an insulated bird-house enclosure. The sample temperature is monitored by a calibrated thermistor readable to --0.25 C. TOP VIIW SAMPLE HEATERS r ucc 034020 INOUSTRlAt, BISEARCH--OCT 17J 55 0.0 SFECTRUM of a black speck (0.25-mm dia) found in molded plastic sample, 300 scans, top. Spectrum of portion from dead insect found in storage bin of granulated polymer feed stock, 300 scans, lower. Photo shows dead insects found in bin. (T/M)vl. Thus, an advantage of Mv* is realized for the interferometer over the dis persive instrument operated at an equivalent signal-to-noise ratio. Jucquinot's advantage is concerned with the light throughput of the interferometer compared with a dispersive instrument at a given resolution. The interferometer has a circular aperture of approximately 30 mm in diameter and has no slits. The slit area for a dispersive instrument operating at 1 cm-1 resolution is approximately 1 mm2. Comparing these figures, one arrives at a throughput advantage of about 2000 for the interferometer over the dispersive instrument. However, in practice, this figure is never realized because the solid angle of view is about 50 times greater for the dispersive instrument than for the interferometer. A realistic estimate of the throughput advantage is closer to 40 times greater for the interfero meter. Many applications of Fourier Transform spectroscopy take advantage of the capacity of the instrument to produce spectra in energy-limited situations and conditions under which conventional dispersive instruments fail to function satisfactorily. A big emphaii* on the small view The analysis of micro samples is perhaps one of the most important applications of Fourier Transform techniques. Here, both Fellgett's and Jacquinot's advantages are ex tremely important, because when working with the minimum amount of sample that will still give a usable spectrum, one sacrifices more than 99% of the incident radiation in some cases. This is truly an energy limited situation. Using a 6X beam condenser and a mask with a 0.05 mm diameter hole in it with Fourier Transform techniques, however, we | I 1 I 1 fc. -,5V t ."if. L can obtain acceptable spectra of polyethylene film with an estimated 40 nanograms of sam ple in the beam. A combination of microscopy-micro infrared techniques has been successfully applied to the identification of microscopic particles iso lated from a variety of substrates, including molded plastics, paper, and integrated cir cuits. The cause of delamination in plastic laminates is easily traced to specific sub stances by examining microscopic chips from the delaminated regions. Similar techniques have been applied to micro potassium bromide pellets containing nanogram quantities of sample, to dilute solu tions containing sample in the sub-microgram range, and to minor components isolated by the various chromatographic methods. Another application that illustrates the ability of Fourier Transform spectroscopy to cope with energy limitations deals with the detection of minor additives, blowing agent residues, plasticizers, and so forth, in thick plastic films and foams. No sample preparation is necessary with this type of sample other than to cut the material physically and mount it in a sample holder. Polyethylene foam has a relatively simple infrared spectrum and lends itself rather nicely to this kind of study. Bands due to Freon fluorocarbon gas, a blowing agent, en trapped in the foam are easily observed in the SMALL FLAKES (<0J3 mm) in a delaminated plastic can be examined by micro Fourier Transform techniques. The cause of the delamination can be quickly identified. Soda: 1 mm/dh>. polyethylene matrix as long as several hours after foaming. Trace amounts of blowing agent residues, such as cyanuric acid, a de composition product of azobisformamide, axe also easily observed by this technique. The effect of antioxidant on foam was studied by computer ratioing spectra of the foam before and after exposure to UV light. The spectrum of the foam before exposure was stored in the computers memory (disc) and retrieved as necessary during the study which spanned several days. Since essentially no sample preparation is necessary in these cases, the data are collected in a few minutes as opposed to work-up times as long as an hour associated with conventional techniques. SCALE EXPANDED spectrum (4 cm'1 ret.) of Z-cm-thick Si slug containing 122 ppba oxygen. Based on S/N, estimated lower limit of detection is 8 ppba. <&!* NIKON EXTENDING MAN'S VISION IN QUALITY CONTROL It takes a sophisticated quality in optics. Optics that make the approach to produce the difference in quality control components and assemblies from production to final that make up today's high- assembly. technology systems. And Producing the finest more sophisticated inspection instruments available for and measuring equipment to science and industry is a make sure the demanding constant challenge. tolerances are met. Wherever quality control is Nikon inspection and needed, on land, beneath the measuring systems range seas, or in space--we are from simple to highly complex. meeting that challenge. Many of them are compatible Extending your vision is our with the superb Nikon main concern. photographic systems. All are products of Nikon pioneering The finest optics in the world. Nikon Instrument Division, Ehrenreich Photo-Optical Industries, 623 Stewart Ave., Garden City, N.Y. Tel. (516) 248-6200. In Canada, Anglophoto Ltd., Ontario. DIAL FOR DATA (free); MO/ttl-OW Nikon 6C An extremely versatile, moderately priced, bench model with IT" diameter ? screen and standard magnification range from -j* 5x to 5(J0x. Inspects and jf measures contours. * surface textures, and . | cavities. (Measuresdepth optically without physical contact with the specimen).! Measuring suites, overlay J charts, and fixtures { provide capabilities as 1 varied a* the needs of industrial inspection and metrology. Header Card #231 Nikon fttensnreecop* Model II Toolmakers microscope with unusually (ugh image resolution and field brightness. Modular system provides choice of seven cross travelling micrometer stages. Interchangeable optical beads, objective*, focusing mounts and illuminators. Also, photomicrography capability, Reader Card #232 Nikon Epl Dark Field Microscope For examination of fine structural details on electronic circuits, on papec textile or wood. The Epi dark field illuminator has a sliding p *n which allows instant v .angeability from dark field to bright field observation. Reader Card #233 ptm r- ucc 034022 In the case of "Styrofoam" brand polysty rene foam and similar products, the rich poly styrene spectrum masks much of the usable fingerprint region of spectrum. With the ex ception of carbonyl-containing compounds, identification of minor additives would be difficult. Paper is another example of a poor medium for the detection of low level additives. The spectrum of paper is dominated by intense cellulose absorption bands and a characteris tically high level of scattered radiation. How ever, some additives that have absorption bands in regions where cellulose is trans parent have been identified directly by apply ing Fourier Transform techniques. In many of these cases, the additive could not be isolated by conventional extraction and separation methods. 0 9S0 700 EMISSION SPECTRA ,perchloroethylene vapors above a beaker placed at 23 and 66 m from the instrument. Top and bottom spectra show 20 coadded scans computer ratioed ipainst black body emitter. Center spectrum is single ' -vrr jean taken at 23 m. Eliminating the cold look Special low temperature techniques and conventional dispersive spectrometers have been used extensively to measure oxygen as an impurity in silicon down to the detector limit of about 50 parts per billion atomic (ppba). Since the oxygen content of many gas float-zoned crystals ranges from less than 50 ppba to 300 ppba, an improvement in the lower limit of detection for oxygen would be most desirable. The measurement itself is not straightfor ward since the absorption band due to the Si-O stretching vibration lies beneath a strong Si-Si lattice band. One must, therefore, first compensate or ratio out the lattice band to observe and measure the Si-O band. Through Fourier Transform techniques, it has been recently shown that a lower limit of detec tion for oxygen of about 8 ppba can be achieved with the sample at ambient tem perature. This eliminates the need for cryo genic equipment and special handling of the samples, all of which contribute substantially to the overall time of analysis. During the last ten years, many of the so- called inherent limitations shrouding various spectroscopic techniques have fallen by the wayside. This is particularly true of infrared emission techniques in which most chemical spectroscopists showed little interest. Sensi tivity limitations on existing dispersive in strumentation partially justified this atttitude, but there was also a mistaken belief on the part of many that good emission spectra could not be obtained at temperatures much below a few hundred degrees centigrade. This, of course, spells instant disaster for the organic chemist, who visualizes months of synthesis work going up in smoke. In fact, this is not the case at all, for emis sion spectra have been obtained using disper sive instrumentation at temperatures well be low 100 C, but the techniques are exceedingly tedious and time consuming and certainly not amenable to practical applications work. But the infrared emission picture has changed dramatically during the past few. years since the introduction of Fourier Trans-- form systems. Usable emission spectra cani be obtained in a reasonable length of timej (minutes) with the sample at near-ambient temperatures using Fourier Transform tech-f niques. 1 In the emission experiment, the source unit of the Fourier Transform spectrometer is re placed by an attachment designed to pass external radiation through the system to the detector. The sample, mounted on a specular mirror that is a poor emitter, is heated in an external "bird house" like enclosure. One of the major shortcomings of infrared emission spectroscopy is the effect of sample thickness on the spectrum. The spectrum of thin, uniform samples, 10 mp. to 0.1 mm in thickness, clearly shows discrete emission bands. As the sample thickness is increased beyond 0.5 mm, the band structure is no longer discernible and the recorded spectrum resembles that of a gray body. This is indeed a limitation but it applies only to emission from thick samples. Thin coatings of such things as plastic, grease, and paint, on metal surfaces can sometimes be more easily characterized by emission spectra than by multiple internal reflection (MIR) i i i Ir , I ! i LV ! Mil. v, n i i k i r ri~ - t 1 L i i { if INDUSTRIAL RESEARCH--OCT 1W ucc 034023 spectra, which is more or less the quick con ventional technique applied in these cases. This is particularly true if the metal substrate is rigid and the coated surface rough so that good optical contact between the sample and MIR plate cannot be achieved. Measuring Into the wind A most intriguing application of the Fourier Transform emission system is its use in de tecting radiation emitted from a hot sample that is located some distance away from the instrument. In a simple laboratory set-up, emission spectra were recorded of hot vapors above a beaker of boiling perchloroethylene placed over 65 m away from the instrument. Similar spectra were recorded in field tests with the sample 33 m away from the instru- Black body emission fronts an 20-cm-dla. hot plate placed 6.5 m away from the Interferometer, (B) Black body emission curve of the vapors above a beaker containing boil ing perchloroethylene an the hot plate. (C) Obtained by com- . puter raticing B to A, .if Is the emission spec-, rju (rum of 1 perchloroethylene SarlM of emission spectra obtained by uccaMivaly (tacking layers of 10-nv. thick polyethylene sheets. The spectra, covering . the range 2000-400 cm4, were obtained with the sairiplc at SO C and are ratload against a black body emitter. The discrete emission ; bands due to polyethylene merge Into a broad band centered around 1300 cm4 an the sample thlcknesa la bteroasad.. ment but exposed to a 48 km/hr (30 mph) wind. Systems using cooled detectors and a Cassegrain telescope collector have been used to monitor hot gas emission from smoke stacks. The advantages of Fourier Transform sys tems are real and significant, and the system offers problem solving capabilities previously not available to the spectroscoptist The ex amples cited only point to the great potential of this extremely sensitive instrument. There are obviously many analogous situations where the sensitivity and speed of a Fourier Transform spectrometer, as well as its flexi bility to adapt to unusual experimental con figurations, will prove this to be a valuable laboratory tool. REFERENCES: 1. Mittledorf, A. J,, The Spex Speaker, XIV (1969). 2. Lowenstein, E. V., Applied Optics, 5, 845 (1966). 3. Conner, J., Rev. Opt., 40, 45, 116, 171, 231 (1961). For abstracts of related articles from ARAC, NASA Regional Dissemination Center, circle number 678 on the reader service card. THE AUTHOR5 Dr. Ronald O. Kagel, right, below, it a senior analytical specialist in the instrument group of the Analytical Laboratories of Dow Chemical Co, Midland, Mich, His PhD in i physical chemistry , is from the Unia. i j He presently is i of the Fourier . Transform Users Group. Dr. Stanley S. T. Xing is an analytical specialist in the polymer group at Dou> Midland. After undergraduate work in Taiwan, he also earned a PhD in physical chemistry at Unia. of Minnesota. In this photo they are using the Digilab FTS-14 system. r'"w The band structure no longer Is discern ible for a single sheet 0.5 mm thick. 034024 |VC|.> JUN 6 1977 R N WHEELER JR. ucc 034025