12222:" 'f" {:22 -¢. "2 22W“ flaw-Q22 '0'" "”2220 .{. (_ 2' \_I 220222.?" ‘_'|2r “.:?‘}:‘%:.J.Q\;E\\ 1w :vh:,+b:2 \ i?" llll‘fl 2". l ‘ ' 2:22“ fi‘lvl“ 5“ 222.32 ”3"": f: 2' I 2,2'I:'2:2:.|‘ ‘3‘}: :l‘ £222.42! "'.-"':I22- -I. .‘ ‘2 "'2: -22.\ 21‘9“: 5:: Ltd 22:22; fi/gjifi’yuvfiygvhil.mfi 2 ': [222122 JI'n'II'v" ”‘8'!“ figfii #2 ' 2:335 ‘22 .. 2: 2 I 22222.2 2 ' ”222 '2 '22 22'th 21“" 2 22" «fltu~n"’2flm‘2« W . f- 2 2 22.222 I2" '2‘ J -:I 2- l I" 222 .ll 2 '2 2 “-2.2“: :2 2|“: 2'. 422221.222... I I22“ fling]; 2‘. 22:; II :22 2|: 2‘ ’2 ”2222‘” 2224. 222 'fl'HIJ .42 222 '--2212|-I2'2'2.‘|21| 922.22 : - :20: I“. :IIII, 2 2 22b». \.:Y:::\::¢$:?£°£;'4>-" 7:22“ W- . 2: 22» 5,22, 5‘12““.3.‘ M: ._-,-: . '12:?4-"212 '2'; . ":2 *- 2a 22‘: ’22 ‘z'w‘fr ~::: '. OH: I “‘222.'l%: 2&5“! JG A.’£ 2'. 2‘4 .i’ifiilxz; l'fb‘ 3'» I. i: . I' ‘ IV :2” 2 2 :2. ‘:7'2)2' 26233222ch 2:2 £2,222 [is law/222': ‘Ml. >212.“ J I‘- . 22E‘ 22.32 ' *6 ‘22‘ 2‘ ‘ ' 2 2:"'2:222[.r2».2 II 9'." -2 «'2' 22 2 “ 222:2'22' ‘22:.” 22.2,. WU 2.‘ ‘22}! 2' ? "':’{:2r"".‘|"2: 2‘21"“ I? ' 2'{";\"‘nl-\J"' .3} T".- ;1 2~ ' :2 < 'r 2 . 2 ; ..' .. _' _ ,5.) ‘2 g , 2 - ’13 .,I'*_-'::\ 1.. . . 4,2. 2. . ‘9‘: :_-. “9:... rdlb‘xl.‘o ”.21.:24 '“I't: 2.1.:221“: J,' 222‘,- ‘22? 2' 2'1”“- ):N'ixa' 32:22.22??? 212252c/222fgfijfi‘; . 2 "2 2 "Yo .I W 4‘22), .m “2. 2,"ij '52- I'U‘H‘YOII 2 2' "W‘- 2‘2‘1‘.".2I," ‘2" ”i“: . 2t . II ‘2' 2:. ., 2 | "1 ,1’2'“ .0 "2' ,2 h) ,J .'."‘ ‘IJ .. v/ “V- . 2'”: I .222 TfiI I:T2~I-.:.:'li: 2 'IQ: wh‘bnrzziozh's‘él :{fbfi‘h 3: 2222:? 2:K:.‘?rtm{£$%:|:2{2%u ; . 2, .. , . .. _ . £32212. -; 2.322.322: w-IIJ : f Lil”. ”:6 ( 2:02:21, r7” 2 2’4" -.>{'I:2“';2:‘;IL:"'.>‘,‘ $3.. :10. g. ”:52'124 2 ' 1-2: 2:22‘ " :::::‘ 2 :21: 222'! 2,‘ 52:22.32" 2‘.‘ 2222'”:r *2--- 2’“: 22- 'I’"‘:{L'-'-&-' ‘ “2 .4222 2 2. .“HI ‘.:2 2:52 22"1‘2fi; 223:0 52 I" .22 I I - 2| {2.2.2122-22'“ ‘C' 22M .‘4‘. 2' 2' IK‘AL" I2l'"2\' ‘ ‘ :.:::l M ”21.2.2222. 2’ I 2_ 2":;22~' “3.22.11: 2692112212 :Hkuyw §q33|£::1;‘3.:\'1fl5;‘| 22 ':.«~~»w222 :aw2xw2m¢%wwaww a“ 2: 2 $I2I2' 0:32...“ , , 2.2 . lLthtfl I. l U" 2‘ I “in... ‘2’ M|‘(~ | I”: I~\ |, WI I 2.2 ~ 2.2:: 2 «I .2 . I -. . I ‘C H. " o . .2 2222?. : ‘ w '2 “AI: q. 1‘ 2 1, M2: $5132va r:'2I.I¢' _ I" 2. “213%?" 2: 22h :22 22‘3’21 2 31111.2“ 22 222', 2222 I.‘ 222222: $222,222 2222' 2‘22 2 .2222 ' :2: :|. 12125::222'2’“ ,,‘:\2'2::'2:2'22a2'.‘1222')2:;:" . .:I r 22 ~222| .22 ”E2222”. 222'“ ”2' ‘2 22‘2" ' 2%.". 222-3“2 2 THESlS \ \ A “:1. " ‘ ‘ § "PI”? 1’, f" . am..." a»... I m. ‘ ,. 3-2-5g.~.a:5£ mu Mathew .»~«' «I.» '1‘, _ N?- W “‘9’“ : I ‘ ' ' ' ..'-'."'..." .3 w my. . ‘..-, I 1 ‘t "“1 Q r .. . of . .‘ I - u ' J .w_" w, This is to certify that the dissertation entitled A Microcomputer Controlled Automatic Recording Spectropolarimeter Using a Linear Diode Array Detector presented by Peter Joseph Aiello has been accepted towards fulfillment of the requirements for Ph.D. degree in Chemistrl £%ZL\ Major professor Date 8/ 25/83 MS U is an Affirmative A ction/Eq ual Opportunity Institution 0-12771 MSU RETURNING MATERIALS: Place in book drop to remove this checkout from LIBRARIES . “ your record. FINES Wlll be charged if book is returned after the date stamped below. L';} C r E 35' {fiéfi ? rm: m - ‘. u.'\l;v A MICROCOMPUTER CONTROLLED AUTOMATIC RECORDING SPECTROPOLARIMETER USING A LINEAR DIODE ARRAY DETECTOR By Peter Joseph Aiello A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1983 ABSTRACT A MICROCOMPUTER CONTROLLED AUTOHATIC RECORDING SPECTROPOLARIMETER USING A LINEAR DIODE ARRAY DETECTOR By Peter Joseph Aiello A microcomputer controlled system for the simultaneous measurement of absorbance and Optical rotatory dispersion has been deve10ped. The system also incorporates an automatic optimization of integration time method to improve the dynamic range of the detector. The instrument uses a linear diode array detector for simultaneous multiwavelength detection. The array used has 512 individual light sensitive diodes each of which is sensitive to electromagnetic radiation with wavelengths from 200-1000 nm. Photodiode detectors have a limited dynamic range relative to a photomultiplier tube. Sensitivity (in terms of intensity) can be adjusted by changing the integration time, but the integration time for all diodes in the array detector is the same for any given scan, and all portions of a spectrum are acquired within the same limited dynamic range. However. the results of scans taken at varying integration times can be combined to produce a data set with improved dynamic range. A microcomputer controlled system for the automatic sequencing Of detector integration times and the storage Of only the optimum readings is demonstrated. An improvement in dynamic range Of 215 times that for a single integration time is theoretically possible, but in most systems. stray light and dark current will limit the practical dynamic range attainable. The inlet Optics allow the simultaneous measurement Of absorbance and Optical rotatory dispersion (0RD) and consist Of a light source, 2 polarizers (one of which is rotated under computer control). a sample cell and a focusing lens. By collecting successive spectra at different angular positions Of the polarizers the absorbance and 0RD can be calculated. The systems studied include the inversion of sucrose and Vitamin 31,. ACKNOWLEDGMENTS My sincere appreciation goes to Professor Chris Enke for the help and encouragement that he has given me during the course Of my research here at Michigan State University. I would also like to thank Professor Stan Crouch for many invaluable discussions I would like to thank Michigan State University, Abbott Laboratories and the Office Of Naval Research for the financial support they provided. I would like to thank the members Of Professor Enke's research group for their help and friendship, especially Hugh Gregg, Bruce Newcome and Phil Hoffman. Finally. I Offer my thanks and appreciation to my wife, Pat, for her continuous love and support. ii Chapter LIST OF LIST OF CHAPTER A. B. C. CHAPTER A. CHAPTER TABLES. FIGURES . . . TABLE OF CONTENTS I. INTRODUCTION. The History Of Imaging Type Detectors. Optical Rotatory Dispersion and Circular Thesis Overview. II. AN Introduction The Optical Design The Data Acquisition System. Data Acquisition, III. CONTINUED. . . . . . A. CHAPTER A. Introduction The Effect Of Cooling the Linear Diode Array Saturation Effects The Effect Of Arc Lamp Stabilization Iv. Introduction Why Vary the Integration Time? Programmed Sequence Control Of Integration Total Dynamic Range for the Instrument Data Storage OVERVIEW OF THE INSTRUMENT. Reduction and Analysis CHARACTERIZATION iii AUTOMATIC OPTIMIZATION OF INTEGRATION Dichroi OF THE LDA SPECTROMETER, TIME 12 14 14 14 20 25 28 28 28 31 35 37 37 38 4O 43 44 9 Chapter F. Performance. . . . CHAPTER V. A. Introduction . . . B. Hardware . . . . . C. Software . . . . . CHAPTER VI. 0RD O O O O O O O O O O O O A. Introduction . . . B. The Experimental Technique . C. The Hydrolysis Of Sucrose. . D. Vitamin B12 . . . E. Conclusion . . . . APPENDIX A: APPENDIX B: A. FORTH Documentation. B. PDP 11/40 Software REFERENCES 0 O O O O O O O 0 iv THE DATA ACQUISITION SYSTEM SELECTED PROGRAM LISTINGS. THE SIMULTANEOUS MEASUREMENT OF ABSORBANCE INTERFACE SCHEMATICS AND TIMING DIAGRAMS Page AND . 78 .101 .101 .114 .139 Figure TABLE 3.1 TABLE 5.1 TABLE 5.2 TABLE 6.1 LIST OF TABLES Percent Lag for 4 Consecutive Scans after Saturation Condition. . . . . . . Commonly used FORTH Words for Instrument Operation. . . . . . . . . . Data Reduction and Analysis Routines. . Observed Precision Of the Phase Information produced by the Curve Fitting Procedure . . . . . . . . . . . Page 71 Figure FIG. 2.1 FIG. 2.2 FIG. 2.3 FIG. 2.4 FIG. 3.1 FIG. 3.2 FIG. 4.1 FIG. 4.2 FIG. 4.3 FIG. 4.4 FIG. 4.5 FIG. 4.6 FIG. 4.7 FIG. 4.8 FIG. 4.9 FIG. 4.10 FIG. 5.1 LIST OF FIGURES A Block Diagram Of the Optical System. . An Approximate drawing Of the Spectrograph . . . . . . . . . . . . . . A Block Diagram Of the Microcomputer . . A Block Diagram Of the Data Acquisition Circuitry. . C O O O C O I O O O O O O 0 Dark Current versus Temperature at a 1 s Integration Time . . . . . . . . . . . . Dark Current versus Integration Time at Temperatures Of 25 and -2 degrees C. . . 0.1. Lamp Spectra at varying integration times. . . . . . . . . . .‘. Integration Time Control Sequence. . . . Data and Integration Time Code Storage. In this illustration, a minimum integration time Of 3 ms is used . . . . Hg Lamp Spectrum at a 25 ms integration time 0 O O I I O O I I O O O O C I O O O Enlargement Of the Hg Lamp Spectrum at a 25 ms integration time . . . . . . . . . Hg Lamp Spectrum at a 1 s integration ting O O O O O O O O O O O O O O O O O O Hg Lamp Spectrum acquired with an initial integration time Of 15 ms with 6 integration time doublings (7 readings). Enlargement Of FIG. 4.7 . . . . . . . . Dynamic Range Measurement using Neutral Density Filters. . . . . . . . . . . . . Dynamic Range Measurement using known KMnO4 solutions. . . . . . . . . . . . . A Block Diagram Of the Microcomputer System and Linear Diode Array Interface. Page . 32 Figure FIG. FIG. FIG. FIG. FIG. FIG. FIG. FIG. FIG. FIG. FIG. FIG. FIG. FIG. FIG. 5.2 A Block Diagram Of the Integration Time Control Circuit. . . . . . . . . . . . . A Block Diagram Of the ADC Board . . . . The Absorbance and Rotational Spectra of sucrose. 0 O O O O O O I O O O 0 O O O O The Hydrolysis of Sucrose. acid concentration 3.0 M. . . . . . . . . . . The Hydrolysis Of Sucrose, acid concentration 1.5 M. . . . . . . . . . . The Rate of Hydrolysis Of Sucrose at acid concentrations Of 3.0 and 1.5 M versus wavelength. . . . . . . . . . . . The Absorbance Spectra Of Vitamin B12 at concentrations Of 16.4, 11.5, 8.2, and 4.9 malmlt 0 O O O O O O O O O O O O O O The Rotational Spectra Of Vitamin B12 at concentrations Of 16.4. 11.5. 8.2, and 4.9 uglmlo O O O O O O O O O O O O O O O Absorbance and Rotational Working Curves Of Vitamin B12 at 355 nm . . . . . . . . ADC Board Schematic Diagram. . . . . . . Schematic Diagram Of the Address Control Circuit. 0 O O O O O O O O O O O I O O 0 Schematic Diagram of the Integration Time Control Circuit . . . . . . . . . . Schematic Diagram Of the Data Acquisition Control Circuit. . . . . . . Single Scan Timing Diagram . . . . . . . Multiple Scan Timing Diagram . . . . . . vii Page 61 65 83 85 86 87 90 91 92 95 96 97 98 99 .100 CHAPTER I INTRODUCTION A. The History of Imaging Type Detectors Over the past decade, there has been considerable development in imaging type detectors for the measurement Of ultra-violet (UV) and visible light. These new detectors have attracted the interest Of a number Of analytical spectroscopists. Traditionally. the analytical chemist has used such instruments as the photometer. which uses a narrow-band light source (for example. the 254 nm emission line from a low pressure Hg lamp or a continuous source and a selective filter). a sample cell and a photomultiplier tube (PMT) detector. While useful for many specific applications, the photometer cannot determine multiple sample components simultaneously or provide general absorbance characterization Of the system. When information at multiple wavelengths is desired. a continuous source (such as a tungsten lamp) and a dispersive element (a prism or grating) is used. The dispersive element is mechanically rotated tO vary the wavelength Of light passed through a fixed exit slit. This selected monochromatic light beam then passes through the sample cell and is detected by a PMT. Spectrometers Of this type maintain the single photometer's characteristics Of wide dynamic range Of 2 absorbance, gOOd sensitivity and rapid. linear response. In addition, they provide a continuous variation in the wavelength sampled with a relatively high degree of resolution in wavelength selection. Despite the great convenience and analytical power Of the scanning spectrometer. it has two characteristics which are limiting in a number Of areas Of application. First. by sampling over only a narrow range Of the dispersed light at any given time, it makes inefficient use Of the Optical information available and thus it prolongs the necessary measurement time. Second, because the wavelength must be physically scanned to provide measurements at multiple wavelengths. this data is acquired at different times. This limits its applicability in situations where the sample's Optical properties are changing, particularly if the rate of such changes is an Object Of the measurement. An Obvious improvement in both these limitations could be achieved with simultaneous multiple wavelength detection. Of course, one would like to have such capability without sacrificing high resolution Of wavelength selection, wide dynamic range and gOOd sensitivity at all wavelengths, linear response, gOOd geometric stability, no stray light, rapid response and rapid electronic readout. Unfortunately. no currently available multi-wavelength spectrometer has all these desirable properties. Several spectrometers have been developed which use multiple detectors arranged in the focal plane Of the dispersing element, thus achieving multi-wavelength detection. Multiple detectors are available in a variety Of types and spatial. geometries. These types Of detectors include: the photographic plate, multiple PMT's (direct reader), the image dissector, the silicon vidicon, the charge-coupled device, the charge-injection device, and the photodiode array. The number of photosensitive elements in these detectors can vary from just a few to many thousands, and these elements can be arranged linearly or in a two-dimensional array. As the number Of elements increases. greater wavelength resolution is possible. When a linear array is placed in the focal plane. each element detects a different wavelength region. This is also true for two-dimensional arrays as the spectrometer entrance slit image is focused on a separate row Of detectors for each wavelength region, and the signal from all elements in each row are averaged together. The two-dimensional array, however, has another advantage in that it can be used for two-dimensional imaging applications such as Eschelle grating and streak camera images. Both linear and two-dimensional arrays are available with photosensitive elements Of various dimensions. In general, the larger the element, the greater the dynamic range; the smaller the element, the greater the resolution. The ideal element dimensions would match those of the spectrometer slit, i.e. 4 narrow and tall for the best resolution and dynamic range. The following is a discussion Of a few types Of multi-channel detectors which includes a summary of their advantages and disadvantages. The first multi-wavelength detector used was the photographic plate. It has several advantages: relatively easy channel identification, simple Operation, low-cost, photometric integration and physical dimensions to suit a wide variety Of spectrometer geometries. Its disadvantages include limited dynamic range, non-linear response, difficult calibration and a very slow data retrieval procedure. Another type Of multichannel detector in use is called the direct reader. This instrument uses multiple PMT's arranged across the focal plane Of a polychromator (1). Many desirable characteristics are maintained through the use of the PMT detectors; however the number Of channels for a reasonably sized instrument is severly limited (usually < 10) and the PMT's must be arranged at wavelengths appropriate for predetermined specific applications. An early electronic scanning spectrometer utilized an image Adissector as the detector (2). An image dissector is nothing more than an electron multiplier which is sensitive to a small, electronically selectable region Of a relatively large photocathode. The image dissector tube Offers advantages such as high resolution, the absence Of 5 broadening effects Of intense spectral lines and the capability Of single photon counting (3). However, even though electronic wavelength scanning can be much faster than mechanical scanning, the image dissector does not provide simultaneous wavelength detection and light impinging on the unsampled area Of the photocathode is lost. The single largest application Of image dissector tubes is the measurement Of low—light levels in astronomy (4). The silicon vidicon (SV) has been widely used as a spectrometric detector. The development and characterization Of vidicon spectrometers has been described in many recent papers (5-11). The light-sensitive surface Of a vidicon tube is a twO dimensional array Of up tO 1000 by 1000 pixels (picture elements). Absorbed photons reduce the charge stored in the pixel. The charge is restored by an electron beam focused on and scanned across the reverse side Of the diode surface. The measured charging current at each pixel is proportional to the charge lost which is in turn related tO the light intensity integrated Over the interval between scans. Advantages Of the SV include high scanning speed and large number Of pixels. Disadvantages are many, however. Dynamic range is limited ((1000). blooming (adjacent pixel cross-talk) causes spectral line broadening, UV sensitivity is low and incomplete readout or lag is nearly 10% Of the charge. For low light level spectroscopy. a silicon intensified target tube (SIT) is 6 used. In the visible region of the spectrum, spectral response and sensitivity are comparable to those Of a PMT; however, its sensitivity is 1 to 2 orders of magnitude lower in the UV region (12). The SIT has been compared to the PMT for the measurement Of transient flourescent signals and has been proven to provide comparable signal to noise ratios (SIN) (13,14). A charge-coupled device (CCD) is a solid state imaging device. CCD's contain up to 2000 pixels in a two-dimensional array which provides relatively high resolution. CCD's are also less expensive than other solid state imaging devices (15,16). High scan speeds are possible, the response is linear, and the dynamic range is about 1000. One significant feature Of this detector is that all elements Of the array integrate intensity during exactly the same time frame. The integration period is terminated by simultaneously transfering the' integrated intensity signals to an analog shift register from which they are subsequently read out serially. Unfortunately, blooming and lag do Occur if saturation is attained. Sensitivity is below that Of a typical silicon device in the UV region Of the spectrum. The charge-injection device (CID), also a solid state imaging device, has several unique features (17). One Of these is its capability of nondestructive readout. With this capability, blooming (still a problem if saturation is 7 attained) can be avoided by periodically scanning the array in a nondestructive readout mode. This mode can determine which pixels are near saturation. These pixels can then be selectively sampled in the normal destructive mode so that saturation is avoided while adjacent pixels with low incident light intensity can continue integrating. Another unique capability is that the integrated circuit is fabricated tO allow random addressing Of pixels. Therefore, even faster readout speeds can be achieved when only a subset Of the pixel information is required for a particular application. The characteristics Of silicon photodiode arrays (SPD) have been discussed in many recent papers (18-25). Over the last few years, self-scanned linear SPD arrays have been used for a number Of analytical spectrochemical measurements (26-34). These arrays are currently available with up to 4096 photodiodes. It has been shown that SPD arrays have superior blooming and lag performance when compared to most other imaging type detectors (20.23). This allows the signal integrating capability Of the array to become a very powerful asset. Since blooming does not occur, one can allow several photodiodes to saturate while adjacent photodiodes can integrate low light level signals. E.G. and G. Reticon (35) has made available photodiode arrays specifically designed for spectroscopy. These arrays are self-scanned and provide real-time electronic readout with up to 1024 diodes. Each photodiode is a slit-shaped 25 um wide by 2.5 mm high. This relatively large active area gives a dynamic range Of up to 10,000 and the narrow width allows for excellent wavelength resolution. Although the arrays are much less sensitive than the PMT, useful measurements can be made over a region from 200 to 1000 nm. When improved sensitivity is needed, an electron multiplier type image intensifier can be used. These image intensifiers use an array Of electron multipliers in a structure called a microchannel plate (MCP). Each channel in the MCP provides a gain Of up to six orders of magnitude (36) and the density Of these channels Often determines the spatial resolution Of the detector. Complete MCP's can be stacked tO provide even higher gains. For response in the vacuum ultra-violet spectral region (50-200 nm) a SSANACON, self-scanned anode array with microchannel plate electron multiplier, has been used (37). This involves photoelectron multiplication through two MCP's, collection Of the electrons directly on aluminum anodes and readout with standard diode array circuitry. In cases where analyte concentrations are well above conventional detection limits, multi-element analysis with multi-channel detectors by hatomic emission has been demonstrated tO be quite feasible (12,38). Spectral source profiling has also been done with photodiode arrays (28,30,32). In molecular spectrometry, imaging type 9 detectors have been used in spectrophotometry, spectroflourometry and chemiluminescence (24,25,27,34). These detectors are Often employed to monitor the output from an HPLC or GC (14.39.40.41). B. Optical Rotatory Dispersion and Circular Dichroism Many chemical substances rotate the plane Of polarized light and are therefore described as being Optically active. Optical rotatory power has its origin in any structural asymmetry in the substance. This may be inherent in the structure Of the molecule (for example, the presence of a chiral carbon atom) or it may be a property Of the crystalline form Of the substance (as in quartz). Normally, light waves are oriented in all possible angles normal to the light path. Linearly polarized light is produced by passing light through a polarizer (such as a prism or film) which passes only a single orientation Of the light waves. This plane-polarized light can be thought Of as the sum Of left and right circularly polarized light. Left circularly polarized light has its electric vector rotating counter-clockwise along the direction Of the wave (right circularly polarized light rotates clockwise). Optically active substances affect left and right circularly polarized light in two distinct ways. First, it is known that these substances have different 10 refractive indices for left and right circularly polarized light. When this occurs, the speed Of the left circularly polarized light through the material is different than that of the right circularly polarized light. The result is the rotation Of the plane Of the incident linearly polarized light. Since refractive index is a function Of wavelength, the rotation Of the linear polarized light is also a function Of wavelength. The curve which results from plotting the number Of degrees Of rotation versus wavelength is called the Optical rotatory dispersion (0RD) spectrum. This characteristic curve gives both qualitative and quantitative information about the substance. Second, Optically active substances have different molar absorptivity coefficients for left and right circularly polarized light. This differential absorbance causes the light tO become elliptically polarized. A Circular Dicroism (CD) spectrum is a plot Of the degree of ellipticity versus wavelength. Theoretical relationships exist between 0RD and CD which permit calculating one curve from the other. These relationships are called the Kronig-Kramers transformations (52,53). This means that in principle no additional information can be Obtained about a molecule by measuring its 0RD spectrum if its CD spectrum is known or vice versa. However, there are practical considerations which sometimes make it advantageous to measure one or the other. For 11 example, the 0RD Of a molecule .can be measured at wavelengths quite different from the Optically active absorption band causing it. Such measurements in the visible region can be used to determine concentrations, purity, rates Of reaction, etc., but basically what is being studied is the CD Of an absorption band at a shorter wavelength. CD has Obvious advantages over ORD in that relatively weak transitions can be identified by CD, but may Often be masked by 0RD. The higher resolution Of CD bands makes it superior tO 0RD, particularly in the structural studies Of complex molecules. These can possess many asymmetric elements which give rise to many Optically active electronic transitions (58-61). The photoelectric polarimeter was the first instrument developed for the measurement Of Optical rotation (54,55). This instrument uses a sodium lamp and a polarizer tO produce a linearly polarized monochromatic light beam. This beam then passes through the sample cell and a second polarizing plate called the analyzer. The analyzer is rotated until the transmitted light intensity is at a minimum. The number Of degrees Of rotation is then read from a calibrated scale. In order to measure 0RD, the sodium lamp is replaced with a continuous source and a monochromator. This is called a spectropolarimeter. Several commercial instruments are available and some Of the more recent Of these have been put under microcomputer 12 control (62,63). If the first polarizer is replaced by a device which alternately generates left and right circularly polarized light (such as a Pockels cell). CD can be measured. By measuring the differential absorbance Of the left and right circularly polarized light, the molar ellipticity or CD can be calculated. Several commercial circular dichrometers are currently available and are capable of high-precision measurements over a wide region Of the spectrum. However, there is still uncertainty concerning the calibration Of these instruments because Of the lack Of a universally accepted standard. Recently compounds such as 10-camphorsulfonic acid and tris(ethylenediamine)- cObalt(III)iOdate have been suggested (64-72). C. Thesis Overview An instrument was developed that uses a linear photodiode array detector (LDA) for the simultaneous measurement of absorbance and Optical rotatory dispersion (0RD). Through the use Of the array detector, both spectra can be acquired at high speed, which allows the monitoring Of kinetic systems. Chapter II presents a general overview Of the instrument. Several important features Of the spectrometer are described including the inlet Optics. polychromator, 13 array detector, microcomputer and software. Chapter III is a continuation Of the diode array characterization begun in my M.S. thesis (22). The effect Of temperature on the detector dark current and Of the linearity Of the dark current versus integration time is studied as the detector can now be cooled to a minimum temperature Of -50 C. Chapter IV discusses a means by which the detector integration time is used to maximize dynamic range. When using any type Of array detector, dynamic range is a problem. Since no single integration time is Optimum for every wavelength, a programmed sequence Of increasing integration times is implemented to provide intensity information for each photodiode at its Optimum integration time. This greatly increases the dynamic range and only doubles the total data acquisition time for a given sensitivity. Chapter V presents a detailed description Of the microcomputer system, LDA interface, implementation of. integration time Optimization, data acquisition software and data analysis software. This chapter also includes normal instrument Operating procedures. Chapter VI presents the experimental procedure for the simultaneous acquisition Of absorbance and 0RD spectra. Also included are the chemical systems studied and the sensitivity limits achieved. CHAPTER II AN OVERVIEW OF THE INSTRUMENT A. Introduction In this chapter, a general overview Of the instrument which was developed is presented. The goal Of the instrument is to provide simultaneous absorbance and Optical rotatory dispersion measurements in the ultra-violet (UV) and visible spectral regions and to repeat this at very high speeds. An instrument capable Of measuring simultaneously more than one physical property Of a chemical system has great potential utility. Also such an instrument can assure that the properties measured actually refer to the same material under the same conditions. All the major parts Of the spectrometer are discussed: the Optical design, the microcomputer, data acquisition system and software, and the minicomputer data processing/analysis which produces the absorbance and 0RD curves. B. The Optical Design Figure 2.1 is a block diagram Of the Optical setup. Descriptions Of each block follow. The light source used is a high intensity xenon arc lamp (150 watt). The Osram HBO-150W lamp powered by a PRA Model 303x power supply 14 15 seesaw sausage one no acumen: soosm < H.~ .eHm w ozEmo - EMN>4910: .2255» Ewmmwhw J
a I DATA ‘ . MEMORY INTEGRATION ' TM . CI! CONTROL (pATA ACQUISITION eus DIODE , VIDE AD DA'm ARRAY C MULTIPLEXER INTEGRATION TIME CODE GENERATOR f FIG. 2.4 A Block Diagram Of the Data Acquisition Circuitry 24 which, under CPU control, are exchanged between the CPU bus and the data acquisition bus. Since the CPU does not need tO be inactive during acquisition, it can unload one RAM bank Of data as well as prepare for the next acquisition while the other RAM bank is being filled with new data on the data acquisition bus. Another feature Of the LDA interface is its ability to collect up to 16 successive spectra, each one at twice the integration time Of the previous one. The automatic ranging Of integration time and the data collection from the Optimum range are explained in Chapter IV. In a research environment, a software system must allow ease Of instrument Operation, ease Of data storage and the ability to reprogram quickly for new experiments. The program polyFORTH, sold by FORTH Inc. (44), has all these capabilities. FORTH is a language which contains a list Of 'dictionary' entries called 'words'. Each word is a small program, or subroutine, that performs a particular function. A word can be at as low a level as a single assembly language instruction or a small program written in assembly language. However, a word can also use any other word currently in the dictionary as an instruction in its program. Therefore, the building of a vocabulary Of increasingly higher function words can result in powerful set Of high level commands which are, at the same time, specific to the tasks at hand. Over 100 words have been 25 written to control the LDA spectrometer, collect the data. store both data and software on local floppy disk or the PDP 11/40 and plot the data on the local graphics monitor. Use Of the software can be found in Chapter V. Documentation Of all the current dictionary words for this system can be found in Appendix B. D. Data Acquisition, Reduction and Analysis In order tO acquire data from the LDA spectrometer, the use Of the microcomputer system is essential. Upon power up, the FORTH software present in the microcomputer PROM's becomes active. This then allows all the applications programming to be loaded from the local disk or the PDP 11/40. There are many FORTH words which perform the following functions: setting up scan parameters, collecting data. storing data and plotting data on the local graphics monitor. The scan parameters consist Of a sequence Of detector scans. The parameters for each scan in the sequence can have several Options. These include data acquisition [Y,N]. autoranging [Y,N], number of doublings [0,15]. and RAM bank number [0,1]. The sequence Of scans is then executed using a single word, GO. The data can then be stored either on the local flOppy disk or the PDP 11/40 minicomputer. The data is stored in a coded, unformatted binary form in order 26 to minimize storage space requirements. The data can be plotted in many different ways on the local graphics monitor very easily with the highest level FORTH graphics words. The microcomputer can also be made to appear transparent thereby allow direct access to the PDP 11/40 as a remote terminal. There are many pOlyFORTH words all Of which are described in Appendix B part 1. The normal commands for instrument Operation are described in more detail in Chapter V. All the data analysis software (which runs on the PDP 11/40 minicomputer) consists Of FORTRAN and MACRO-11 (PDP 11 assembly code) routines. All this software acts directly on the coded, unformatted, binary data which were acquired by the microcomputer system, and transferred to the 11/40 via direct serial line transmission or floppy disk. A single core routine, CRUNCH, is used to call all other subroutines and provides data transmission from one subroutine to another. This allows any combination Of the following data manipulation subroutines to be performed upon any data set: ENTER decodes and enters a data file, LOG takes the base 10 logarithm, ABSORB calculates the absorbance using reference, sample and dark current data files, SUBTR subtracts any 2 data files, NRED smooths the data using a modified Savitsky-Golay algorithm (56,57), FIT curvefita the data to a cosine squared wave, fitting the phase, amplitude, and Offset parameters knowing only the 27 frequency (i. e. x-axis spacing) and STORE stores the data in an ASCII file. At this point, MULPLT, a multiple data set plotting routine written by Dr. T. V. Atkinson (45), is used for data display. CHAPTER III CHARACTERIZATION OF THE LDA SPECTROMETER, CONTINUED A. Introduction In this chapter, several characteristics Of the linear diode array spectrometer are discussed. Previous work done in this area was presented in my M.S. thesis (22). Improvements made tO the spectrometer such as detector cooling and a new computer interface provided the basis for further experimentation and discussion on several detector characteristics. B. The Effect Of Cooling the Linear Diode Array One Of the major Operational characteristics Of the linear photodiode array is its electronic background noise or dark current. Since the dark current is very reproducible, it can be removed from any spectrum by subtraction. However, the presence Of dark current can severely limit the use Of the integrating capability Of the array. This is because the charge from the dark current increases as the integration time increases and can, at integration times Of longer than 2 s, completely saturate the array. At this extreme there is no dynamic range left for signal measurement. Fortunately, the dark current 28 29 (which is due to charge leakage) is temperature dependent. By cooling the array, the dark current level can be dramatically reduced (20). In this instrument. the photodiode array is cooled using two single stage thermoelectric coolers (Melcor CP-1.4-3-06L). These coolers are placed between the detector integrated circuit and a copper heat sink. Each cooler is beneath one half Of the detector. By passing a large current (>6A) through these thermoelectric devices. heat is pumped from one side to the other. With these coolers, the detector can be cOOled to a minimum temperature Of --5. C. If lower temperatures were desired, multistage thermoelectric coolers could be used. In order to prevent frost from building up on the face of the detector, the entire spectrograph housing is kept under a dry nitrogen atmosphere at a slightly higher than ambient pressure. This simple antifrosting technique is very effective at these temperatures. Figure 3.1 shows the dependence Of the detector dark current on temperature. The integration time was held constant at 1 s. Over a 30 degree range, the dark current varies by an order Of magnitude. Thus cooling makes significantly longer integration times available for the measurment Of low light levels. At a temperature Of -20 C, the detector dark current reaches saturation at an integration time Of 17 s or more. It has been shown that at room temperature the dark current response versus the Dork Current ( % of saturation) 50.0 —1 40.0 '- 30.0 — 30 X 20.0 - d x - x — X ,x 10.0 '- _ _,xf -x ' X- 0.0 l l l I I l l I l l l l l l I l l l l l l l l I i ”5 0 5 10 15 20 0 Temperature ( C) FIG. 3.1 Dark Current versus Temperature at a 1 5 Integration Time 31 integration time is non-linear (22). It was determined that this response was due to a charge leakage from the video recharge line on the LDA integrated circuit and that cooling should alleviate this problem as well. It can be seen in Figure 3.2 that at a temperature Of -20 C, the detector dark current responds linearly over a wide range Of integration times. C. Saturation Effects Blooming, an important characteristic of electron imaging sensors, refers to the situation where a strong signal induced in one region spreads to adjacent sensor elements. In the case Of the silicon vidicon (SV), a tube device, blooming is a serious problem. With CCD's and CID's, both Of which are solid state devices, blooming is not a problem as long as saturation is not attained. Blooming is Observed with these devices, however, if saturation occurs. With SPD arrays, it has been reported that blooming is negligible (22,23), even under saturation conditions. Lag, or the incomplete readout Of the signal stored, is also a problem with electronic imaging devices. In the case Of the SV, lag is as high as 10% (46). Lag is undesirable if the detector is being used for time resolution studies. The lag Of the LDA was measured in the following manner: Dark Current ( % of saturation) 32 100.0 90.0 80.0 700 600 500 40.0 lllllllllllllllllllllllllllllllllll 30.0 200 10.0 lllllllllllll X x x"xx X 0-0 'lflllllillllllllllTlllllllTllllllllllll]:I ..fi! 0 l 2 C5 4 5 6 7 8 (L) - lntegrotkma tune (sec) FIG. 3.2 Dark Current versus Integration Time at Temperatures Of 25 and -2 degrees C 33 The detector dark current was allowed to saturate the array. At this time, five sequential readout scans were recorded. each at a 50 ms integration time. Any {psidual charge not read in the first scan would show in the following four scans. This experiment was done at room temperature and at -2. C. The $ residual charge remaining in each scan is given in Table 3.1. The $ lag is calculated as shown below: sLAG = [ (51-Sd) / (Si-8d) l x 100 (1) where Si is the initial scan which reads out most of the charge. S1 is the next readout and Sd is the dark current at the 50 ms integration time. In the LDA, lag barely exceeds 1% at -2° C. This compares well with a previous reported value Of 1% (47). This phenomenon, however, has been attributed to the 1 ms time constant Of the preamplifier feedback circuit by Y. Talmi (23) as it was not Observed with an E.G. and G. PARC voltage mode preamplifier. With this preamplifier, lag was determined to be less than 0.1%. However, the temperature dependence Of the lag indicates that the problem is not with the preamplifier, but with the LDA integrated circuit itself. 34 % LAG AT FOLLOWING TEMPERATURES SCAN # 25 °c -2 °c 1 3.6% 1.3% 2 0.9% 0.2% 3 0.2% 0.0% 4 0.0% 0.0% TABLE 3.1 Percent Lag for 4 Consecutive Scans after Saturation Condition 35 D. The Effect Of Arc Lamp Stabilization In applications Of the Optical system, it quickly became apparent that the measurement precision decreased as the length Of the experiment increased. This was attributed to an intensity variance of the xenon arc lamp. It was determined that Over a 1 minute period. the intensity at 315 nm varied as much as 20%. In order to determine the frequency components of the intensity variation, 2000 successive intensity values were acquired at 315 nm using an integration time of 3 ms with 4 doublings. A Fourier transform was performed on this data. This determined conclusively that most Of the drift occurs at frequencies lower than 0.7 Hz. The length Of the experiment is approximately 1.5 minutes when 50 spectra are collected and transmitted to the PDP 11/40. Because Of this time limit. two Options were considered. The first Option involved the modulation Of reference and sample cells at a rate of 1 Hz. This proved to be ineffective as the cell position was not reproducible. The second Option involved the stabilization Of the lamp drift. This was done using the previously mentioned Optical feedback circuit. This circuit monitors the lamp intensity (using a beam-splitter and a photodiode) and modulates the output current Of the arc lamp power supply thereby keeping the light intensity constant. This circuit is damped so that it only responds to low frequency 36 ((50 Hz) intensity fluctuations. TO test its effectiveness, 2000 successive intensity values were again acquired in the same manner as before and a Fourier transform performed. With the corrective feedback, the lamp intensity now varies less than 2% at frequencies below 0.016 Hz. All simultaneously acquired spectra (absorbance and 0RD) presented in this work were Obtained with this Optical feedback circuit in Operation. CHAPTER IV AUTOMATIC OPTIMIZATION OF INTEGRATION TIME A. Introduction Analytical chemists have long sought a multichannel imaging device with excellent resolution and linear response over a wide dynamic range. One Of the most severe limitations Of current imaging detectors is dynamic range. The Reticon RL512S photodiode array has been shown tO have a dynamic range Of about 10,000 (23.35) for a single integration time. The conventional single channel detector, the photomultiplier tube (PMT) exceeds this by more than 2 orders Of magnitude. In order to extend the dynamic range, the integration time (time between successive scans) Of the LDA can be varied. In most analytical spectroscopy applications, however, no single integration time is Optimum for every element (wavelength) in the array. This chapter describes a solution to this problem which involves acquiring successive spectra at increasing integration times so that data taken at the Optimum integration time for each photodiode is available. 37 38 B. Why Vary the Integration Time ? The integrating capability Of the linear diode array is achieved by charge integration, a technique which enhances the signal and averages the noise. This is very useful as the signal-to-noise ratio (SIN) under many conditions increases linearly with the integration time, whereas S/N enhancement by averaging replicate measurements increases only with the square root Of the number Of scans averaged (46,48). Therefore. the maximum resolution ind S/N are Obtained if each photodiode is allowed to integrate charge until it nears saturation. Figure 4.1 shows multiple spectra Of a quartz halogen lamp at integration times Of 5. 10, 20, 50, 200 and 500 ms. The detector dark current has been subtracted out in each Of these spectra. Because Of this. the saturation level appears tO decrease. Actually, the total dynamic range is decreasing because Of the increase in the integrated dark current. It should also be noted that the profile Of these spectra is a product Of the polychromator transfer coefficient, the lamp spectral output and the detector sensitivity at each wavelength. It can be immediately Observed that at the shorter integration times, the photodiodes detecting the wavelengths above 500 nm are nearing saturation. At longer integration times the wavelengths near 300 nm are Optimized. However, this is achieved at the expense Of the information at the longer INTENSITY (relative) FIG. 39 ' .. . , _ _ _ 1 ....... w‘wt-W' ' ' " ‘ J~.’s‘« . w “go-2‘- ‘1‘“ ‘ 1 “Nndfip. Karma ‘:O ‘l 1.. . . 7.“- - Rum-33’ I 35"”‘4 web's}! . 'fl" ‘" ' . wart .m. .. .. 2\J._u-..J‘s-WFVMWH"~' ' ‘7' u-‘qG'n ’ 1," w?- F‘QT .1!" -‘R .- -'.' " ‘A' .- d' ' . " . . afi a. - ’ . a ,.2 r. , a. . a - g c a V I W ' » ’73.L""-'V "o-h‘ mWWM~-f (,7 ‘ I’lf'f‘." ‘ . *’ ' f ’u .- j , n“ " I . q ._ I ‘ \':'«’.' - .. ' ' a e‘ . . . ‘ q . ,4 . . .13..- : ° .- -.-':".' -' 1.." a. '. . ‘ . -..-' I f . . '1‘! ' 0 f . ‘1 ' . J q' gy \- V I. 0 w . o I I O .' e . ' ‘ . . _ , .. . . - 3 " ‘ - ' ' IO 5' 500- 200 50 20’ r ’ : ' I’ .m ’ .' - .' - .I I . l . . . . i ,I . _l . I ‘ J . ' 4' \ . 4' : .. . . u “- J I . v' n' . l " : . 4" ’ ‘. If -' / . .0 . .73!“ - -‘ I v v I , ' 4' . O I ,' .' l i D- . I. l . .' I '. I ' s V I . a v 6 v a .- t I «I ' I f I f v e l b I. .. I M f. V w I M :0 3 v I ‘: J ’ s' V' V' - a _ I .f ' I- ~ I O ’ v f l " vv / / .‘ " £4? ,~ ~\“\. 35...”- «va's’mm U \\"‘; Tj—llIll[TTIIIIITTITTIIIIITITITTITTTTTIIIIIII 300 350 400 450 500 550 600 550 700 WAVELENGTH (nm) 4.1 0.1. Lamp Spectra at varying integration times 40 wavelengths which is now lost due to saturation Of those photodiodes. Normally, spectra are acquired at various trial integration times and then data are taken at a single integration time which is determined to be the best compromise. This usually results in a severe degradation Of the photometric accuracy at the wavelengths for which the signal is weakest. The next section describes a multiple integration time technique which alleviates this problem. C. Programmed Sequence Control Of Integration Time If different integration times are used for successive scans, the longest integration time scan will provide the best reading for the wavelengths where the signal is weakest, while the shorter integration time scans will give the best reading for those diodes that were saturated during the longer times. Figure 4.2 shows that in this system. the integration time is successively doubled. The initial integration time, t, should be chosen such that no photodiode has neared saturation. The number Of doublings should be chosen to provide the maximum output for the lowest intensity wavelengths that is possible within the time limits Of the experiment. If the integration time is doubled n times ( in Fig. 4.2, n=4 ). then the total experiment time, T, will be: 41 Uo< Zn mums nn< mum: mzmu zonuomazm muzo mwzmnmzoo : mod mzmu ACH>HCHmzmm zaznzsz moi mznu zonu=._. Integration Time Control Sequence 4.2 FIG. 42 1‘ = 2(2nt) - t (1) This means that data are acquired at the optimum integration time for each photodiode in a total time of approximately twice the longest integration time in the sequence. For each element, the optimum integration time is that which results in a readout between 50% and 95% of saturation. The dynamic range improvement, D, over that of a single integration time experiment. will be: 1) = 2‘3““ (2) assuming the dark current is negligible at the longest integration time used. The number of integration time doublings. n, can be increased until either a time limitation imposed by the experiment is reached. or the integrated dark current exceeds 50% of saturation. The dark current can be reduced by cooling the LDA. The thermoelectric coolers used in this instrument cool the detector to a temperature of about -5° C. Further cooling continues to diminish the dark current thereby allowing integration times as long as 24 hours at liquid nitrogen temperatures (20). It should be noted that as the integration time is increased, many of the photodiodes may reach saturation while adjacent photodiodes have yet to attain their optimum 43 integration time. This would be a significant problem in other imaging detectors such as CCD's, CID's and vidicons. In the case of these detectors, when saturation is reached, charge will leak from the saturated pixels (picture elements) to adjacent pixels. This 'blooming' effect has been shown to be minimal in linear photodiode arrays (22,23). Thus integration times that cause many photodiodes to saturate can be used while useful information is still obtained from the non-saturated photodiodes. This information will contain only the integrated light intensity for each photodiode, free of bleed-through from adjacent saturated photodiodes. D. Total Dynamic Range for the Instrument The total dynamic range is a function of the detector dynamic range and the range over which the integration time is varied. A 12-bit analog to digital converter (ADC) is used which provides nearly enough resolution to use all the dynamic range information available for each photodiode (about 10‘). When a 4-bit code is used to identify the doubling number (this code will be described in Chapter V). up to 16 doublings can be performed. With the LDA at a temperature of 00 C, the dark current nearly saturates the array in about 17 s. In this case, the maximum useful integration time is about 9 s. Again, the maximum 44 integration time could be increased by further cooling. The minimum integration time is 2 ms with a 512 element array clocked at 1 MHz. By doubling this 13 times the final integration time is about 16.4 s, longer than the maximum usable integration time at 00 C. For the general case, disregarding the dark current, the total dynamic range D is D a 2(number of ADC bits) x 2(n-l) (3) For a 12-bit ADC and a limit of 16 for n, the maximum dynamic range allowed by the electronics is 1.34 x 10', neglecting the loss of ADC resolution due to dark current. If the LDA is cooled so that dark current is negligible, the total dynamic range available using this technique is thus greater than that of a conventional PMT detector with an autoranging amplifier or converter. B. Data Storage In an experiment in which spectra are to be recorded frequently and dynamic range enhancement by the variable integration time method is desired. the storage of every reading can require a huge amount of memory. This memory requirement can be reduced, however, by only storing the measured intensity for each photodiode at its optimum 45 integration time. This is advantageous for two reasons: first. the memory requirement is reduced by a factor of n; and second, the data are in a readily retrievable form. That is. it would not be necessary to search several data sets for the Optimum integration time of each photodiode. In order to do this. an indication of the integration time must be part of the intensity value word stored in memory for each photodiode. An integration time code. which is simply the number of readings or 1 more than the number of integration time doublings, is generated and combined with the ADC conversion value. Since the computer has a record of the initial integration time. the actual integration time for each photodiode can be calculated from the code. Figure 4.3 shows how the complete sequence of up to 16 integration time doublings is stored in 1 [byte of memory. Two bytes of space are allocated to each photodiode. 12 bits of which are the ADC output. The remaining 4 bits comprise the integration time code. The ADC conversions from the initial scan along with an integration time code of 1 (first reading) are written in the 1 Kbyte block of memory. 0n the next scan. for which the integration time has been doubled. the ADC conversions and the new integration time code (1 greater than the previous code) are written over the same block of memory with the exception of those channels for which the photodiodes which are within 95% of saturation. In the case of 1i in Figure 4.3. the data at each of the 46 m6 :0 :10 .10 SE8 02 :m m. .25. . mtfm v. _ 4. .. _.. I 800 rm... fl — Poo— ppp—FPFFPFFP mom * ooo— oopoppo—Fopp vmm PF—o oooooooopppo Nap OFFO PooooooFFPoo mm Pope Fppoppoppooo we oo—o Ppppooppoooo em ppoo pppppopooooo NP opoo —_poo~oooooo 9 FOOD FP—ooooooooo m moou >m83 w u o n 255 r I c .1 y \ I I M a I e I]! r e u. m L y u I e 25 c2:m_:ao< Sec v .2200 .M f .1 IN mmeauu< o D - t m u :5. , _ x a a e e; 1 mm . 35:3 - e r mu...“u .233 m E< mam 25%. .u m m :23. 2:. 5. H .m m B s A m. L v L Y L 1 gm . ago , I v :8 s D h Q h D 1 1 a . 3:03:00 .322 32:02 .35 2< F . x mm 33.25 533m —mIcEnSG >mwn320u v. «a v.9 9v: 3583... 60 The first mother board contains the modules comprising the $30 as described above. The second mother board, an extension Of the CPU bus only, contains an additional 8 Kbytes Of RAM for program space, a transceiver module and a stepper motor driver module. The transceiver module is used tO transmit a subset Of the CPU bus to a remote box which houses a similar transceiver module (reconstructing the CPU bus subset) and interface modules to floppy disk and graphics systems. The floppy disk system consists Of a dual 8-inch single-sided drive and an intelligent controller, Persci model numbers 760 and 1070 respectively. This provides 500 Kbytes Of mass storage for program and data space. The graphics system includes a 9-inch Ball monitor and a Matrox MTX-256“2 graphics display board. This system provides a 256 by 256 dot matrix display, each individually addressed. The stepper motor driver module (49) is used to rotate a polarizer in the Optical setup. The stepper motor is a RAPID-SYN Model 23D-6102, which has 200 steps/revolution. This is geared down using a worm screw gear set so that each complete revolution results in a 2 degree rotation of the polarizer. Therefore, the CPU can control rotation in 0.01 degree increments as well as determine the direction Of rotation. The third mother board, also an extension Of the CPU bus, contains another 8 Kbyte RAM module for program space, and the integration time control circuitry. Figure 5.2 is a 61 START SCAN [INT TIME COUNTERSI<——%6MI-Iz CLOCK [SHIFT REGISTERS J<— START OF SCAN L MICROCONPUTER] FIG. 5.2 A Block Diagram Of the Integration Time Control Circuit 62 block diagram Of the integration time control circuitry. A schematic diagram can be found in Appendix A. The integration time can be controlled directly by the microcomputer or it can be automatically doubled by the data acquisition control circuitry. When direct computer control is desired, the integration time is written into a set Of shift registers and transferred to the integration time counters. A 1 MHz clock is used for the detector scanning circuit and the integration time controller. The integration time counter is 28 bits long, which results in a possible range Of integration times from 2 ms to 268.4 s (2" us). When automatic integration time doubling is desired, the CPU only programs the initial integration time. Upon the start of each successive scan, the integration time shift registers then shift left, thereby exactly doubling the integration time for the next scan. The fourth mother board is divided into upper and lower halves. The lower half, an extension Of the CPU bus, contains the data acquisition control circuitry. The upper half, the data acquisition bus, contains hardware which generates the address, data and control lines for the direct memory access circuit. The data acquisition control circuitry is programmed by the CPU through an 8-bit control register. At the start Of each scan Of the detector, this register, as well as the integration time and address counters, must be loaded by the 63 CPU for control Of the next scan .(except when automatic doubling Of the integration time is in Operation). The low bit (D0) Of the control register enables the data acquisition cycle. Another bit (D1) enables autoranging Of the integration time. If enabled, the tap 4 bits (D4-D7) are loaded with the number Of integration time doublings desired (0-15). Another bit (D3) determines which memory bank the data is to be written into. The last bit (D2), currently unused, generates a pulse synchronous with the start of the scan pulse. This could be useful if some other equipment (for example, a shutter or stopped-flow) needs to be synchronized with the self-scanning detector. The data acquisition control circuitry uses the information at the control register, the LDA clock, start Of scan (generated by the integration time control circuitry) and the EOL (end Of line) signal from the LDA tO generate the following control signals necessary in the data acquisition cycle: Start Conversion, which signals the ADC to begin conversion Of the current photodiode; Latch Enable, which latches the 12-bit conversion and the 4-bit integration time code: GEILSB, which drives the low 8 bits Of the conversion onto the data bus: BE-MSB, which drives the high 4 bits Of the conversion and the 4-bit integration time code onto the data bus; ADDRESS COUNTER OLE, which increments the address for the storage Of the next data byte; -WR, which writes the data byte into memory and LOAD, 64 which loads the next set Of commands from the control register for the next scan. Schematics of the data acquisition control circuitry as well as a timing diagram showing the relationship of these signals to the detector output during a single data acquisition cycle can be found in Appendix A. Another set Of counters, the address control circuit, is simply loaded by the CPU with the initial address for data storage. These counters then drive the 16-bit address onto the data acquisition bus. The address is incremented after each byte Of the 1024 bytes/scan is written into RAM by the data acquisition control circuitry. The starting address for data collected from the next readout is written by the CPU during the current readout. When the autosequeneing of the integration time is enabled, the address counters are automatically reloaded with the same initial address for each successive scan. This is required for the efficient data packing described in Chapter IV. Schematics Of the address control circuitry can be found in Appendix A. The data lines on the data acquisition bus are generated by the ADC module. A block diagram of this circuitry can be found in Figure 5.3 (Schematic in Appendix A). The video signal (0.0 to 2.5 volts) output from the Reticon RC1024SA amplifier board is amplified by an Op-amp tO give the 0 tO 10 volt input signal required by the 65 Twooo.Iw2::H_z__ mmszaoo T zm x. zoo mmm> 1 _ D\<._..m NP IL.<<(I A Block Diagram of the ADC Board .3 5 FIG. 66 Computer Labs HAS-1202 2.2 as 12 bit analog-to-digital converter (ADC). This high speed ADC allows detector readout rates Of 4 as per diode. The resulting conversion is latched along with the 4-bit integration time code. This code is simply a 4-bit counter which, when automatic integration time doubling occurs, is incremented each time the integration time is doubled. The two bytes Of data are then written into RAM by the data acquisition control circuitry. Also, when autosequeneing Of integration time is used, a l2-input NAND gate is used to detect 95% Of photodiode saturation or more (all ADC bits are 1); its output inhibits the write pulse (generated by the data acquisition control circuitry) which results in data storage for only those photodiodes that have not yet reached 955 of saturation. The fifth mother board contains two half-populated 8 Kbyte RAM modules for data storage and also a bus switcher module (designed by Bruce Newcome) which can exchange these two memory banks between the data acquisition bus and the CPU bus. This arrangement allows the CPU to actively program the next data acquisition cycle and to unload data from one RAM bank while the ADC circuitry on the data acquisition bus fills the other with new data. 67 C. Software The software for the collection and subsequent analysis Of data is divided into two distinct categories. The microcomputer acquires and transmits data using programs written in the language pOlyFORTH (44). Data analysis and reduction occur in the PDP 11/40 by programs written in FORTRAN IV and MACRO-11 (PDP-11 assembly) which run under the RSI-11M Operating system. The multi-level language and Operating system pOlyFORTH is used exclusively on the microcomputer system (44.50.51). FORTH is characterized by four major elements: the dictionary, stack, interpreter and assembler. The dictionary occupies most of the memory space. It is a list Of 'words', each Of which is a program which can include any other dictionary entry. New words can easily be added to the dictionary which makes FORTH an extensible language and one which is versatile, almost to a fault. For example, already defined entries such as '+' can be redefined and thus drastically change the language. Also, several dictionaries can be maintained simultaneously allowing the user to create identically named entries with a different function in each dictionary. Two push-down stacks are maintained. The first, the parameter stack, provides an efficient mechanism for transferring numbers from one word to another. The second, the return stack, is used for 68 internal flow from word to word. FORTH is fundamentally an interpretive system. There are two interpreters, one which works in the conventional manner, simply passing text strings and looking up each word in the dictionary; the other interprets strings Of absolute memory addresses by executing words which contain the addresses Of the next word. This threading provides for a faster execution speed as each word does not have tO be located as in a conventional interpreter. FORTH also includes an assembler, which allows the user to define words on the assemble language level. Such words are Often necessary tO perform I/O Operations and highly time-critical functions. FORTH is used for many reasons. It is a very compact language using only about 8 [bytes of memory. FORTH is an extensible language, a great advantage when software modifications are required due to a change in the experimental procedure. FORTH is also highly portable. FORTH is most useful in applications where speed is Of the utmost importance. Since portions (or all) Of a FORTH program can be written in assembly language, FORTH programs can be written that compare favorably in speed with machine-language programs (about 90%). Since FORTH can grow to a high-level language, programming productivity is much higher than using machine-language, and yet speed is still much faster than other high-level languages such as FORTRAN. In addition to the words supplied in the basic FORTH 69 dictionary, over 100 words have been created which control the LDA spectrometer. As each word can call all others previously defined. a high level set of instructions has evolved tO perform various experiments. As an example, a listing Of six FORTH words which control rotation Of the polarizer follows: C1CO CONSTANT STPMTR C1C8 CONSTANT DIRMTR STEP 10 0 DO LOOP STPMTR C@ DROP 3 : CDIR 50 0 DO LOOP DIRMTR C@ DROP 3' : ROTATE 0 D0 STEP LOOP ; DEGREES 2DUP 0. D< IF CDIR DABS DROP 500 + ROTATE CDIR 500 ROTATE ELSE DROP ROTATE THEN 3 The first two words, STPMTR and DIRMTR, are CONSTANT definitions. When a CONSTANT definition is executed, the constant is placed on tOp of the stack. Therefore, either directly entering STPMTR from the terminal or calling STPMTR from another word will place the value ClCOH on tOp Of the parameter stack. The stepper motor interface hardware is designed so that reading from the addresses ClCOH and C1C8H causes the stepper motor to step once or change direction respectively. The third entry, STEP, executes a single read from STPMTR using the FORTH definition C@ and places the value read on tap Of the stack. The 10 0 DO LOOP is simply i wait loop inserted so that the maximum stepping frequency 70 Of the motor is not exceeded. This allows STEP to be called several times sequentially without loss of motor steps. The FORTH word DROP simply discards the top value from the parameter stack. Since a value was placed on top of the stack by C@ , it must be discarded when it is no longer useful. The fourth word, CDIR, Operates in a similar fashion to STEP and when called, changes the direction Of motor rotation. Calling the fifth word, ROTATE, steps the motor the number Of times found on tOp Of the parameter stack. The sixth word, DEGREES, is the only entry normally called by the Operator. Entering the command -18.75 DEGREES rotates the polarizer appropriately. DEGREES accomplishes this by first determining the direction Of rotation by the sign Of the number on top Of the stack. It then calculates the number Of steps required for the desired rotation. and rotates the polarizer accordingly. If a negative direction is entered, it rotates the polarizer 5 degrees past the final position and then rotates 5 degrees positively to eliminate any gear lash. This results in positive positioning Of the polarizer. This set Of words shows how easily and quickly FORTH can become a high level language for the needed experimental functions. Table 5.1 is a listing Of the most Often used and highest level words written for the Operation Of the instrument along with their functions. Since the CPU is actively programming cOntrOl of the FORTH word .—._.__ 5CD GO DEGREES MODIFY PARAM RESPTABLE SAVEPTABLE IVW AVW IVR IVT CLEAR DRES DSAVE IIDISK FLOPPY 71 FUNCTION Collects 50 sets of spectra, each differentiated by 3.6 degrees, and transmits this data to the 11/40 via direct serial line onto the FORTH pseudo disk LDA.FTH. Acquires a single set of spectra by executing the parameter table. Rotates the polarizer XXX.XX degrees. Modifies entries in the parameter table for the scan number on tOp of the parameter stack. Lists the parameter table on the terminal. Restores the parameter table from the FORTH disk block on tOp of the stack. Saves the current parameter table on the FORTH disk block. Plots intensity vs. wavelength spectra on the local graphics screen. Plots absorbance vs. wavelength spectra using 3 data files stored on disk. Plots intensity vs. rotation angle of the polarizer on the graphics monitor. Plots intensity vs. time on the graphics monitor. Clears the graphics monitor. Returns data to memory from FORTH disk block on tOp of the stack. Saves data currently in memory in the FORTH disk block on tap of the stack. Redirects mass storage device to the pseudo-disk on the 11/40. Redirects mass storage device to the local floppy disk. Words for FORTH used Instrument Operation Commonly TABLE 5.1 72 next detector readout during a current data acquisition cycle, it must already have access to the desired control parameters. In order tO do this, a 1 Kbyte parameter table is maintained in memory. This parameter table is a list Of detector control Options for up to 146 consecutive scans. For each scan, these Options include: data collection [Y,N]. doubling [Y,N], number of doublings [0-15]. external flag set [Y,N], memory bank [0,1], address of data storage, initial integration time, and a last scan flag in the parameter table [Y,N]. The command PARAM displays this parameter table on the terminal. Parameter table Options are modified by entering the scan number to be modified followed by the word MODIFY. MODIFY prompts the Operator for each possible Option and enters any changes for the appropriate scan in the parameter table. Because Of the potentially large size Of a particular set Of scan parameters, it is convenient to save this parameter table if similar experiments are likely to be performed. The words RESPTABLE and SAVEPTABLE restore and save the parameter table on the local floppy disk. Once the parameter table is established as desired for an experiment, the command GO executes the entire sequence Of scans (stopping at the scan marked with the last scan flag). After the data are acquired, they can be stored on the flOppy disk using the word DSAVE. If another look at any particular stored data set is desired, it can be returned to memory from the flOppy 73 disk using the word DRES. Since local mass storage is limited, it is necessary to Off-load the data quickly. Communications software utilizing a serial line connection between the microcomputer and our PDP 11/40 minicomputer is currently used. The program FORTHPIP was written by Phil Hoffman (73) for this purpose. FORTHPIP runs on the ll/40 and is capable Of transferring data to or from the serial line, pseudo disks or RSI-11M unformatted data files. Pseudo disks are files maintained on any 11/40 disk which emulate_ a local FORTH flOppy disk. By entering llDISR, all FORTH words requiring disk I/O now communicate with the pseudo disk (named LDA.FTH) instead Of the local floppy disk and remain completely transparent to the Operator. To return communication to the local floppy, the word FLOPPY is entered. Several graphics commands are available such as IVW which plots an intensity vs. wavelength spectrum on the local graphics monitor Of any data set currently in memory. The most powerful FORTH word written so far for this system is 5CD which collects 50 spectra, each taken after rotating the polarizer 3.6° and transmits all the data tO the 11/40. Once the data have been transmitted tO the minicomputer via direct serial line transmission or flOppy disk, the data analysis and reduction procedure can begin. These routines are written in FORTRAN and MACRO-ll and run on a PDP 11/40 minicomputer. A single core routine CRUNCH, makes available 74 several functions for the analysis of data from various types Of experiments. Table 5.2 lists these subroutines and thier functions. With the exception Of the EA subroutine, all the available subroutines read in the data from RSX data files that were generated by FORTHPIP. ENTER simply reads in a single data file. This data file normally contains the ADC conversions and integration time codes for 512 points. These points do not necessarily correspond to the 512 photodiodes. ENTER is used when direct intensity values are desired as output by the LDA. EA has the same function as ENTER with the exception that an ASCII formatted data file is entered. EA is useful when a data file output by the STORE subroutine requires further analysis. AVG simply averages the data files entered- a useful function for signal averaging. The subroutine LOG reads in a data file and returns the log Of each point. SUBTR subtracts any 2 data files, point by point. ABS calculates the absorbance Of the sample at each wavelength using the following equation: Ax = -log I Isl-n1) / (Rx-DA) I (1) where AA is the absorbance at wavelength 1, S; is the signal intensity, D1 is the dark current and RA is the reference intensity. Note that ABS calls both the LOG and SUBTR 75 SUBROUTINE FUNCTION ABS Calculates the absorbance at all wavelengths using Reference, Sample and Dark current files. AVG Averages together multiple data sets. DFIT Calculates absorbance and 0RD spectra from amplitude and phase information Obtained by fitting 50 points at each wavelength to a cos**2 function. EA Reads in any ASCII data file. ENTER Reads in any unformatted data file. FIT Calculates amplitude, phase and Offset information from a single data file; basically a single wavelength DFIT. LOG Takes the log of any data file. NRED Smooths any data file using a modified Savitsky-Golay algorithm. 0RD Calculates the observed rotation with polarizers at a fixed 45 degree angle from a reference, sample and dark current data files. STORE Stores any data file in an ASCII type compatible with MULPLT. SUBTR Subtracts any two data files. TABLE 5.2 Data Reduction and Analysis Routines 76 subroutines. NRED is a data smoothing routine which uses a modified Savitsky-Golay (S-G) algorithm (56.57). Linear, quadratic and derivative smoothing functions are available with an unlimited number Of passes and a 21 point window maximum. Normally, each pass through a S-G smoothing function results in the loss Of a single point at each end Of the data set. This limits the number of passes possible. The 8-6 algorithm used here has been modified to make estimates Of the end points and can therefore make an unlimited number of passes. However, as the number of passes increases, errors in the end-point estimates can cause misinterpretation Of the data. The 0RD subroutine calculates the Observed rotation, a. in degrees from 3 data sets with the polarizers at a fixed 45° angle by the following equation: aA = 0.5 cos.’1 [ ((sx-nx) / (Rx-01)) - 1 I — 45 (2) where SA is the signal intesity at wavelength A, D1 is the dark current and RA is the reference intensity. For the Observation of large rotations ()5') this works well and is very fast. DFIT is the routine which reduces the data generated by the 5CD command on the microcomputer. At each wavelength, there are 50 data points collected, each sampled at 3.6° intervals spanning a 180° sweep Of the polarizer. These 50 77 points complete a full cycle Of a cosa function. By curvefitting these points to this function, the phase, amplitude and Offset parameters are generated. This involves a least squares curve fitting procedure (74). By curvefitting both reference and sample spectra, absorbance and 0RD spectra can be calculated using the following equations: where SAMPX is the signal amplitude (from the curve fit) at wavelength A, RAMPA is the reference amplitude, SPHSA is the signal phase and RPHSl is the reference phase. The function Of FIT is very similar to DFIT in that a curve is fit, except that the 50 points are sequential in the data file and only a single curve fit is performed. This is useful for single wavelength studies as it takes only 1/512 of the total time. Finally, the subroutine STORE writes the resulting data file to a system device in an ASCII format compatible for input to a plotting routine MULPLT (written by Dr. T. V. Atkinson) (45). All data plots in this dissertation were generated by MULPLT. CHAPTER VI THE SIMULTANEOUS MEASUREMENT OF ABSORBANCE AND 0RD A. Introduction The instrument described in this dissertation can measure both absorbance and 0RD by 2 different techniques. The absorbance Of a sample can easily be calculated from 3 spectra (the reference, signal and background) using equation 1 in Chapter V, or in a similar manner, the ORD can be calculated from three spectra using equation 2 in Chapter V when the angle between the polarizer and the analyzer is known tO be 45 degrees. Using this procedure, however, the absorbance and 0RD cannot be measured simultaneously. An alternate technique involving the rotation of the analyzer does allow for the simultaneous measurement Of absorbance and 0RD. This technique will be described in the next section. The detection limits achieved and the studies Of the inversion of sucrose and Of Vitamin B13 are also described. B. The Experimental Technique It has long been known that the light intensity transmitted through two linear polarizers is a function Of the incident light intensity and the cos3 Of the angle at 78 79 which the polarizers are set. If the intensity is measured while the analyzer is rotated continuously, the signal detected will vary sinusoidally between 0 (when the polarizers are crossed) and some maximum value, Imax' with a frequency twice that of the rotation Of the analyzer. If the sample shows absorption but no Optical activity, the new maximum intensity value Observed, I, will be lower than the value Of Imax and the absorbance can be calculated using equation 3 in Chapter V. If the sample is Optically active but does not absorb, the amplitude Of the signal will not change (181 but a phase difference will be Observed max)’ between the sample and the reference signals. The Observed rotation, a, is exactly equal to this phase difference as shown in equation 4 in Chapter V. By recording the signal intensity at several analyzer positions, both the intensity and phase information are Obtained. If this signal intensity is measured at all wavelengths simultaneously, both the absorption spectrum and the 0RD spectrum are measured. To Obtain these cOSz waves with this instrument. several spectra are acquired at different angular positions Of the analyzer. The data at each wavelength are fit to a cosa function (with the subroutine DFIT) by a non-linear least squares algorithm (74). The resulting amplitude, phase and Offset parameters are used to calculate the absorbance and 0RD spectra. The only parameter not fit by 80 the routine DFIT is the spacing between positions Of the analyzer as this is known accurately. Several experiments were conducted in order to determine the relationship between the number Of data points taken along a 25 to 360 degree range Of rotation Of the analyzer and the precision Of the phase information produced by the curve fitting procedure. The results are shown in Table 6.1. It is easily seen that as the total number of points acquired increases, the error decreases. However, if 500 points are to be collected at every wavelength, 500 scans need to be acquired and transmitted directly tO the 11/40. At the minimum integration time, this process takes about 15 minutes and 500 [bytes Of mass storage are required. If only 25 points are collected, then a very much larger error results. As a compromise, 50 points are normally acquired. This results in a relatively low error and the data acquisition time and disk space requirements are reasonable (about 1.5 minutes at the minimum integration time and 50 [bytes respectively). It was also Observed that if the points are acquired over only a fraction of a complete cycle of the cosa function, relatively large errors result from the curve fit. For these reasons, the FORTH routine 5CD collects 50 spectra over a 1800 range Of analyzer rotation (each point is 3.60 apart). Degree Range Covered by Analyzer 81 Number of Points Acquired Evenly Spaced over the Degree Range Covered 25 50 100 500 90 +1.67 +0.23 180 +0.33 +0.11 +0.08 +0.05 360 +0.34 +0.10 TABLE 6.1 Observed Precision Of the Phase Information produced by the Curve Fitting Procedure 82 C. The Hydrolysis Of Sucrose Figure 6.1 shows the simultaneously Obtained absorption and 0RD spectra Of a 0.831% aqueous sucrose solution. These spectra were Obtained with a 100 um slit width and a 15 ms integration time followed by 4 doublings. The detector and sample were at 23 degrees C and the Xe arc lamp was used. The path length used was 1 dm. As explained previously, 50 sets Of spectra are acquired with each following a 3.60 rotation Of the analyzer. It is easily seen that the absorbance Of the sample is minimal, however, the Observed rotation, a, displays the expected behavior. When the Observed rotations are compared tO a value calculated from a 1 term Drude equation (75). the maximum absolute error over the entire wavelength region covered by the instrument is within $0.03 degrees. The hydrolysis Of sucrose to give equimolar amounts Of glucOSe and fructose is catalyzed by hydrogen ion. By monitoring the Observed angle or rotation, the rate at which this reaction Occurs is measured at various hydrogen ion concentrations. For this reaction, the integrated rate equation can be expressed in the following form: log(a-uo) = -kt/2.303 + log(ao-a°) (2) where a is the angle Of rotation at time t, a0 is the angle 83 OBSERVED ROTATION (2) Q o m o m o o l1111I41111L11I114+11111L 8 ‘ IN I t 1- P8 0 1'- ho? aI- -85 fi 1' 0% I" < 23 O -O I") O [ITTerflll’titrjrilrrrl—Tr a 9 '0. ca «2 O. "2 (I) BONvaaosav FIG. 6.1 The Absorbance and Rotational Spectra of Sucrose 84 Of rotation at the first reading (t=0), and a, is the rotation when the reaction has gone to completion. Since it is difficult to Obtain a reliable reading for a.co experimentally, and the characteristics Of the reaction are most sensitively determined by the data for the initial phases Of the reaction, this quantitiy need not be determined. Figures 6.2 and 6.3 show the Observed rotation spectra over a time span Of 4 to 32 minutes after the addition Of acid (each 4 minutes apart) at the HCl concentrations Of 3.0M and 1.5M respectively. Again all spectra were Obtained with a 15 ms integration time followed by 4 doublings, the sample and detector were at room temperature, the slit width was 100 um and the sucrose concentration was 25.03/100ml. The path length used was 1 dm. All the spectra were smoothed with the modified Savitsky-Golay algorithm (5 point window and two passes). Figure 6.4 shows the rate constant, k, at both acid concentrations versus wavelength. The rate constant was calculated with a non-linear least squares procedure. The program assumed a trial value for ac and then determined the best values Of the slope, intercept, and a” to fit the above rate equation. At all points, the correlation coefficient was greater than 0.997. OBSERVED ROTATION (degrees) 85 J ‘ ==:h~ . .\.‘ 1 O o I I I I I 1 I T T I I I I T r l I I —r I 200 300 400 500 600 700 WAVELENGTH (nm) FIG. 6.2 The Hydrolysis Of Sucrose, acid concentration 3.0 M OBSERVED ROTATION (degrees) 2&0 100 mo 86 FIG. 300 400 500 600 700 “MNELENGTH (nnfl 6.3 The Hydrolysis Of Sucrose, acid concentration 1.5 M RATE CONSTANT 87 orn5«: ammo-3 l I I OIHS-é QOHJ-i : PWW orms«€ 0.000:...rre.,...,...r...] 200 300 400 500 600 700 WAVELENGTH FIG. 6.4 The Rate Of Hydrolysis of Sucrose at acid concentrations Of 3.0 and 1.5 M versus wavelength 88 D. Vitamin B1, The discovery of vitamin B1, in 1947 culminated a worldwide scientific search that had lasted fOr over two decades. In 1926, Minot and Murphy had announced the successful use Of whole liver in the diet for the treatment Of pernicious anemia, which previously had almost always been fatal (76). This lifesaving discovery resulted in many studies all directed toward finding the anti-pernicious anemia factor in pure form. In early 1948, Merck Research Laboratories announced the discovery Of pure, crystalline vitamin B11 (77). Since the discovery of vitamin Btz, numerous laboratory and clinical studies have provided extensive information on its chemistry, its biological functions, its clinical uses, and the technology Of its production on a large scale. The first communication on the chemical nature Of vitamin B13 did little more than report that the molecule contained nitrogen and phosphorus, and that it appeared to be a coordination complex with the metal cobalt (78). Today, as the result Of research by many, the complete structure Of vitamin B13 is well known. In 1948, vitamin B was known only to possess high 13 activity as a microbial growth factor and to be active clinically in producing a hematologic response in pernicious anemia. Since then, it has been shown to be essential for 89 normal blood formation and neural function. Evidence indicates that it is also involved in other metabolic processes and that it is required for human and animal growth (79). Vitamin B1,, also known as cyanocobalamin, may be identified and assayed by its light absorption spectrum and its Optical rotatory spectrum in the visible and ultraviolet regions. In an aqueous solution, absorption maxima are Observed at 278, 355, and 550 nm. The solution also shows a strong Cotton effect at 355 nm (80). Figures 6.5 and 6.6 show the absorption and rotational spectra 0f vitamin 313 at the concentrations Of 16.4. 11.5. 8.2, and 4.9 mg/ml. Each pair Of spectra was acquired simultaneously using an integration time Of 20 ms followed by 3 doublings. The sample cell used was 1 dm long. Each spectrum was averaged 15 times. The Xe arc lamp was used and an infared cutoff filter was inserted into the light path in order to reduce stray light. The rotational spectra were smoothed by the modified Savitsky-Golay algorithm with a window size Of 5 points. Two smoothing passes were made through each data set. Figure 6.7 shows the linear response Of both the absorption and rotational spectra versus concentration at 355 nm. The deviation from Beer's law in the case Of the absorbance working curve is due to the presence Of stray light. This deviation Occurs at an absorbance value which is lower than the value Obtained from ABSORBANCE 90 25 20 I'Llill L5 LO 1 I LLLLJ I I l 05 L_I_JJ_lLll 000 U— T I I r I I I r I I l I I T I 200 300 400 500 600 700 WAVELENGTH (nm) FIG. 6.5 The Absorbance Spectra Of Vitamin B12 at concentrations of 16.4, 11.5. 8.2, and 4.9 mg/ml ROTATION (degrees) 91 1 .5 “j 1 1D - 05 - Ofl - "0-5 T I I I I l I I I I I I T I I I I I I I 200 300 400 500 600 700 WAVELENGTH (nm) FIG. 6.6 The Rotational Spectra of Vitamin 312 at concentrations of 4.9 mg/ml 16.4, 11.5, 8.2, and 92 OBSERVED ROTATION. degrees (2) "2 o «2 o "2 9 N N v- F O O LIL]llllljllljlllllllllll O N P X -."..’ L c L- L- 0 a. x 0" - é °z - F 8 E IJJ " 0 2 b (D O x ruq ”'0 I—rflTIT'FIIITTII'rI1I'II‘I O '0. ca '0. ca «2 Q N N v- v- C O (L) BOWBHOSBV FIG. 6.7 Absorbance and Rotational Working Curves of Vitamin 312 at 355 nm 93 the KMnO‘ experiments. This occurs because the percent stray light at the 355 nm photodiode is greater than that at the 545 nm photodiode. Even when the absorbance exceeds 2.0. the rotational spectrum obtained is still very accurate and shows a linear response. This demonstrates that valid rotational data can still be acquired even when the sample is highly absorbing. E. Conclusion Array detector spectrometry offers many advantages over scanning spectrometers, especially where near simultaneous information at multiple wavelengths is required. With suitable optics and electronics, the resolution and dynamic range of modern array detector spectrometers can match that of a medium resolution scanning monochromator and PMT. It has been shown that wide dynamic range can be achieved using silicon photodiode arrays with a variable integration time technique. In a total time of only twice the longest integration time, the optimum signal measurement is made at each photodiode while memory requirements remain at a minimum. An instrument capable of simultaneously measuring more than one physical property of a chemical system has great potential utility. Since the measurements are made simultaneously, it can be assured that the properties 94 measured actually refer to the same material under the same conditions. Also an instrument of this type can become cost-effective as the duplication of major pieces of equipment can be avoided. This LDA spectrometer has shown the ability to measure both absorbance and 0RD spectra over the near UV and visible regions of the spectrum simultaneously. Even when the absorbance of the sample is high. valid rotational data can still be acquired. Array detector spectrometers are somewhat more suceptible to stray light interferences than scanning monochromators because the baffling cannot be so exclusive and the use of switched in filters for various wavelength regions is precluded. Because of the many and proven advantages of this new technology, the years ahead should see great strides in both the technology and applications of array detector spectrometry. APPENDIX A INTERFACE SCHEMATICS AND TIMING DIAGRANS 95 From Control Circuitry 2% START CONVERSION Data Lines on Acqisition Bus HAS—1202 IN ANALOG 38.3K piece Jammy Aeuv epoga Luau FIG. A.1 ADC Board Schematic Diagram 96 From CPU From Control Circuitry hm > R 89> 4* L 755m 240 > DO 30 I=1.510 POINTS(I)=REALS(I.1) CONTINUE And call SMOOTH CALL SMOOTH(POINTS.NUM.COEFFS.TMP.NCOEFF.IWIN.IFUNC 5 ITER.IERR) IF(IERR .E0. 0) GO TO 40 WRITE(5.9IO)IERR 116 910 FORMAT(’ Subroutine SMOOTH error = ’.I7) GO TO 60 C C Now copy smoothed data back the array REALS(I.1) C 40 DO 50 I=1.510 REALS(I.1)=POINTS(I) 50 CONTINUE C C And 90 home C 60 RETURN END C LOG.FTN C C Subroutine which takes the 109(10) C oF input File which is requested or oF C data already in array REALS(512.1) iF parameter C N = 2. IF N = 1. asks For input Filespec. C C P. Aiello 5/14/82 C SUBROUTINE LOG(N) C C Variable declarations C COMMON IRDATA/ REALS REAL REALS<512.3) INTEGER IERR.N C C Determine iF Filespec is to be requested C GO TO (10.20)N C C Request input Filespec and put into array REALS C 10 CALL INPUT(3.1.IERR) IF (IERR .EG. 0)GO TO 20 GO TO 10 C C Calculate the log and store in array REALS(512.1) C 20 DO 30 I=1.512 REALS(I.1)=FULLOG(REALS(I.1)) 3O CONTINUE C C And go home C RETURN END C C C FULLOG 0000tfi0t‘10 0'" 0000 O 000 (")0 0 0 OO F-‘(TJOO 117 Function which takes the base 10 log oF parameter X. Returns 0 iF XC=O. FUNCTION FULLOGiX) IF(X .LE. 0.0)GO TO 10 FULLOG=ALOGIO RETURN FULLOG=0.0 RETURN END GSM.FTN Subroutine which questions user iF the data in the array REALS<512.1) is to be stored. smoothed stored again. STORE is called. then NRED is called and also STORE so that the smoothed data can be stored immediately SUBROUTINE QSM Variable declarations COMMON IPSTR/ STR COMMON /RDATA/ REALS REAL REALS(512.3) BYTE ANS(2).STR(82) INTEGER LEN Ask iF data is to be stored CALL PROMPT(15.STR.LEN) DECODE(LEN.900.STR)ANS FORMAT<2A1) IF(LEN .EQ. OlGO T0 10 IFiANS(1) .EQ. ’Y’)GO TO 20 Ask iF data is to be smoothed and then stored CALL PROMPT(16.STR.LEN) DECODE(LEN.900.STR)ANS IF(LEN .EQ. O)GO TO 40 IFiANS(l) .EO. ’Y’)GO TO 30 RETURN CALL STORE GO TO 10 CALL NRED(2) CALL STORE RETURN END INPUT.FTN Subroutine which requests For the File to be input. 00000 001‘) 00001000 -0 00 000 00000 00000 910 (‘1 0 (“l [t] 118 reads in the data into the array REALS(SET) and closes the input File. P. Aiello 5/14/82 SUBROUTINE INPUTUIUIUIUIUI 5 .AND. (AB8(BTEST-PHS))/BTEST .LE. 0.0001 5 .AND. (ABS(CTEST-OFF))/CTEST .LE. 0.0001 5 .OR. NITER .GT. 10)GO TO 50 GO TO 10 RETURN END DFIT.FTN 0 0017! [1.10005 0000() 0004] 0000000008 000 H 0 O 0 127 SUBROUTINE DFIT Delcare variables COMMON IRDATA/ REALS REAL REALS€512.3).FF(3.3).REFAMP(512).REFPHS<512) 5 .SAMAMP(512).SAMPHS(512) INTEGER INTS(100.15).IERR.LEN.NINTS(16) BYTE FILE(32).STR(B2) IERR=O DRCONV=3.14159/180.0 Read in the number oF data sets to be Fit CALL PROMPT(18.STR.LEN) DECODE(LEN.900.3TR)NDSETS FORMAT(1I4) Request For and read in First phase. oFFset. and XDIM. (the constant spacing between and NPTS. which is the number or in other words. the number IF NPDT = 0 then intermediate approximation oF ampl points in the X dimen oF points in each dat oF scans at each wave result printing is in iF it = 1 then they are printed. CALL PROMPT(19.STR.LEN) DECODE(LEN.910.5TR)AIN.BIN.CIN.XDIM.NPDT FORMAT(4GIS.7.1I4) Initialize loop For the reFerence and sample data 5 DO 210 NSET=1.NDSETS Now Fill in array REALS(512.1) with the X dimension parameters as this is clobbered later when the data is stored REALS(1.1)=0.0 DO 10 I=2.512 REALS(I.1)=REALS(I-l.l)+XDIM CONTINUE Ask For data set Filespec and CALL PROMPT(13.STR.LEN) DECODE(LEN.920.8TR)LENGTH.FILE FORMAT(G.32A1) FILE(LENGTH+1)=O CALL FOPEN(4.FILE.ISIZE.IERR) open File IF ((IERR .AND. 128) .LT. 0) GO TO 30 GO TO 40 _~ow mo 0 00008000() 000 50 940 60 90 000 0("10 128 WRITE(5.930) FORMAT(’ OPEN ERROR ’) GO TO 20 ' Calculate the number oF scans in the data set From the size oF the File returned by FOPEN NSCANS=(ISIZE-1)/2 Set up loop For 32 times. This is the number 0F 32 byte records there are For each scan NREC=1 DO 120 NRS=1.32 Set up loop to read in array INTS(NSCANS.16 words) DO 80 NCNT=1.NSCANS NCNTS=NCNT CALL FREAD(NCNTS.NREC.NINTS.IERR) IF ((IERR .AND. 128) .LT. 0) GO TO 50 GO TO 60 WRITE(5.940) FORMAT(’ READ ERROR. WILL TRY TO CLOSE THEN EXIT ’) GO TO 130 DO 70 NCOP=1.16 INTS(NCNT.NCOP)=NINTS(NCOP) CONTINUE CONTINUE NRECBNREC+1 Time to transFer INTS(NSCANS.16) to REALS(512.2) and Fit. store the data in REALS(512.3) or POINTS depending on the value oF NSET. DO 110 NFIT=1.16 DO 90 I=1.NSCANS REALS(I.2)=FLOAT(IBOT12(INTS(I.NFIT)))/ 5 2.*eFLOAT(ITOP4(INTS(I.NFIT))) CONTINUE AMP=AIN PHS=BIN OFF=CIN THIS IS WHERE FIT IS TO BE CALLED CALL SFIT(AMP.PHS.OFF.NSCANS) And type out result iF this option is selected IPT=(NRS-1)*16+NFIT IF (NPDT .EQ. 1)WRITE(5.950)IPT.AMP.PHS.OFF FORMAT(’ IPT.AMP.PHS.OFF= ’.lI4.3G15.7) 000 100 110 120 130 140 OOOHOOOO U‘ 0 160 170 970 000 180 980 210 000 129 Check which File is being Fitted and store IF(NSET .GT. l)GO TO 100 REFAMP(IPT)=AMP REFPHS(IPT)=PHS GO TO 110 SAMAMP(IPT)=AMP SAMPHS(IPT)=PHS CONTINUE CONTINUE And close the File CALL FCLOSE(IERR) IF ((IERR .AND. 128) .LT. 0) GO TO 140 GO TO 150 WRITE(5.960) FORMAT(’ CLOSE ERROR ’) RETURN IF we have Just Finished Fitting the reFerence File loop back IF (NSET .E0. 1) GO TO 210 Now calculate the absorbance and store the data DO 160 I=1.512 REALS(I.1)=SAMAMP(I)/REFAMP(I) CONTINUE CALL LOG(2) DO 170 I=1.512 REALS(I.1)=-REALS(I.1) CONTINUE WRITE(5.970) FORMAT(’$ABSORBANCE ’) CALL GSM Now calculate the ORD (phase diFFerence) and store DO 180 181.512 . REALS(I.1)=REFPHS(I)-SAMPHS(I) CONTINUE WRITE(5.9SO) FORMAT(’$ORD ’) CALL GSM CONTINUE RETURN END CRUNCH.FTN CRUNCH is the main program which is used to (J00 000 910 110 000000000 000 000 130 call all the number crunching subroutines via the First 2 letters oF the subroutine name. PROGRAM CRUNCH Variable declarations COMMON /PSTR/ STR COMMON IRDATA/ REALS REAL REALS(512.3) BYTE STR(82).TEST(SO) INTEGER IERR.LEN CALL ERRSET(72..TRUE...FALSE...FALSE...FALSE..15) CALL ERRSET(73..TRUE...FALSE...FALSE...FALSE..15) CALL ERRSET(74..TRUE...FALSE...FALSE...FALSE..15) Print out prompt CALL PROMPT(1.STR.LEN) Read in input command DECODE(LEN.900.STR)TEST FORMAT(SOAI) Call CMD. string parser to get the command and exec CALL CMD(TEST.IERR) IF no legal command was entered send error message. IF(IERR .NE. O)WRITE (5.910) FORMAT(’ Illegal command. type HELP iF you ned it’) GO TO 10 CALL EXIT And exit the program END EA.FTN Subroutine used to enter an ASCII data File into the array REALS(512.1). This is the same as ENTER.FTN but reads in ASCII data Files as opposed to FORTHPIP output RSX binary Files. P. Aiello 9/20/82 SUBROUTINE EA Variable declarations COMMON /RDATA/ REALS (300 900 910 10 ‘0 fie h. S. ~e ‘0 5e ‘s .s ‘0 ‘e IBOT1217 ITOP42: 131 REAL REALS(512.3).X BYTE TAG(2).FILE(32).STR(82) INTEGER IERR.LEN IERR=O Request For and read in the data File CALL PROMPT(17.STR.LEN) DECODE(LEN.900.STR)LENGTH.FILE FORMAT(Q.32A1) FILE(LENGTH+1)=O OPEN(UNIT=3.NAME=FILE.TYPE=’OLD’. 5 CARRIAGECONTROL=’LIST’.READONLY) DO 10 I=1.512 READ(3.910)TAG.X.REALS(I.1) FORMAT(2A1.2G15.7) CONTINUE and close the File CLOSE(UNIT=3) RETURN END .PSECT UNPACK .TITLE UNPACK .IDENT /V1.0/ .ENABL LC UNPACK.MAC Contains 2 routines. IBOT12 and ITOP4. which return the bottom 12 bits and top 4 bits respectively as a Function. IF the top 4 bits=0. a O is returned. iF > 0 then the number -1 is returned. P. Aiello 6/11/82 O.ARG=2 MOV @O.ARG(R5).RO BIC 17000O.R0 RETURN MOV @O.ARG(R5).RO BIC 7777.R0 SWAB R0 ASR R0 ASR R0 ASR R0 ASR R0 BEG 10$ So ”a 5. -0 ‘0 ‘0 ‘0 ‘I .- CSIBLK: FDB: DEC RETURN . END .PSECT .TITLE .IDENT FIO.MA This r Three reads P. HOF .MCALL .MCALL .MCALL OFFset Open c O.LUN O.NAM O.SIZ O.ER1 Read c O.SCN O.NDX O.DAT O.ER2 Close O.ER3 DeFine EFN = deFine CSIS .BLKB FDBDFS FDAT$A ' FDRCSA FDBK$A FDOP$A 132 FIO FIO /V1.0/ C outine does File I/O. open. read and close. separate entry points. Written For block From a FORTH emulator disk File. Fman. P. Aiello 7/12/82 CSI$.CSI$1.CSI$2 FDBDF5.FDATSA.FDRC$A.FDBK$A.FDOP$A OPEN$R.READ$.WAIT$.CLOSE$ s For the calling parameters alls 2 4 6 8. alls ween calls 2 event Flag For block I/O read 1 CSI and FDB For 1 block I/O C.SIZE R.FIX..512. FD.RWM BUF.512...EFN.STAT .CSIBLK+C.DSDS..FO.RD.FA.DLK BUF: STAT: BLKPTR: I i 5 i 3 i i 3 5 3 F OPEN:: CSIERR: ~. 5. -- so in s. 5. 1‘ ~. s. s. s. .. READ:: 133 DeFine buFFer size to 512 bytes STAT recieves I/O status word BLKPTR contains number oF virtual block to be read .BLKB 512. .BLKW 2 .WORD 0.0 File OPEN entry point For call sequence CALL FOPEN(lun.name.size.error) lun= logical unit number name= Filespec oF File to be opened size= returned. number oF RSX blocks + 1 error= returned. low byte is negative err IF high byte is 0 then FCS error IF high byte is neg then DSW error MOV CSIBLK.RO spoint to CSI block CSI$1 R0.0.NAM(R5). 32. sFilename syntactic BCC 10$ .iF no error.branch JMP CSIERR .set error Flag CSI$2 R0.0UTPUT iFilename semantic BCC 20$ iiF no error.branch JMP CSIERR iset error Flag MOV FDB.RO .point to FDB OPEN$R R0.QO.LUN(R5) sopen the File MOV F.ERR(RO).@O.ER1(R5) .move error code MOV (F.EFBK+2>=0 SUB R4.R2 ;subtract 16 INC (R1) sand increment 10$: READ$ R0... BLKPTR .read BCS RDERR iiF error WAIT$ RO await For and TSTB F.ERR(RO) ;? error read BMI RDERR sset error Flag DEC R2 scalculate oFFset .REPT 5 Sta desired record ASL R2 3(INDEX-1)*32 .ENDM sdone ADD BUF.R2 iadd base address MOV O.DAT(R5).R3 zpoints to data 20$: MOV (R2)+.(R3)+ sstore record SOB R4.20$ sloop until done RDERR: MOV F.ERR(RO).@O.ER2(R5) zset error Flag RETURN sand go home CLOSE entry point Fortran calling sequence CALL FCLOSE(error) error= error code fl ‘0 SI On 5. - CLOSE:: MOV FDB.RO spoint to FDB CLOSE$ R0 iclose the File MOV F.ERR(RO).@O.ER3(R5) sset error iF need RETURN sbye .END .PSECT PROMPT .TITLE PROMPT .IDENT /V2.0/ .ENABL LC PROMPT.MAC . Phil HoFFman and P. Aiello . Department oF Chemistry . Michigan State University . East Lansing. MI 48824 . 3-August-1982 i "-“'"S""""==_=—_—===—== ‘T‘T‘ —===r =—="-—-'”====== PROMPT is used by all CRUNCH subroutines to Fetch command input. This input may be obtained From (1)The MCR invocation string. (2)an indirect command File. and/or (3)in response to an explicit terminal prompt. The calling sequence is as Follows: Cs he .1 He ‘- 5. ‘- - 5. CALL PROMPT(IPMPT.STR.LEN) where the ar IPMPT =3 Ix; STR LEN II V PROMPT DOES WITHIN THE R be ‘a fie \ne ‘0 he 5' ‘0 .1 he ‘0 - -0 he ‘- ‘- .s fin. fie 5 Local DeFini 135 guments in the call are deFined as Follows: an INTEGERF2 variable whose value is the index oF the IPMPT-th string to be used as a terminal prompt. The address oF an byte array. 82. bytes in length to receive the command string in response to the selected prompt. an INTEGER*2 variable to receive the length oF STR. NOTE NOT CHECK THE VALUE OF IPMPT TO BE SURE THAT ANGE OF LEGALLY DEFINED PROMPT INDICES. tions. OFFsets. Storage. Macros. etc. .MCAL .MCAL 5 DeFine the o O.PMP O.STR O.LEN 3 SpeciFy the 5 prompt lengt P1: TEXT P2: TEXT P3: TEXT P4: TEXT P5: TEXT P6: TEXT P7: TEXT P8: TEXT P9: TEXT P10: TEXT P11: TEXT P12: TEXT P13: TEXT P14: TEXT P15: TEXT P16: TEXT P17: TEXT P18: TEXT P19: TEXT L GCMLB$.GCML$.EXIT$S L TEXT i From (MACROS.MLB) FFsets to the arguments in the call T 2 4 6 prompts. the prompt address table. and the h table. YCRUNCH} Y.P1L.1.0.0 {FILE TO ENTER : >.P2L.1.0.0 {FILE TO TAKE LOG OF : }.P3L.1.0.0 {REFERENCE FILE : }.P4L.1.0.0 {DARK CURRENT FILE : }.P5L.1.0.0 {ABSORBANCE FILE : }.P6L.1.0.0 {FILE TO BE SMOOTHED : }.P7L.1.0.0 {OUTPUT FILE : }.P9L.1.0.0 {FIRST FILE : }.P10L.1.0.0 {LESS SECOND FILE : }.P11L.1.0.0 {Input IWIN. IFUNC. ITER : >.P12L.1.0.0 {FILE TO BE FIT: >.P13L.1.0.0 (Input AMP. PHASE. OFF. XDIM. NPTS: }.P14L.1. {Store the data? [Y/NJ: }.P15L.1.0.0 {Smooth and store the data? [Y.NJ: }.P16L.1.0. {ASCII File to enter : >.P17L.1.0.0 (Number oF data sets to smooth: }.P18L.1.0.0 {Input AMP. PHASE. OFF. XDIM. NDPT: >.P19L.1.0 PADD: PLEN: 3' 136 .WORD P1.P2.P3.P4.P5.P6.P7.P8.P9.P10.P11.P12 .WORD P13.P14.P15.P16.P17.P18.P19 .WORD P1L.P2L.P3L.P4L.P5L.P6L.P7L.PSL.P9L.PIOL .WORD P11L.P12LP13L.P14L.P15L.P16L.P17L.P18L.P19L 3 SpeciFy the command block For the Get Command Line Macro. GETLIN: GCMLB$ 1...2 ’___ 3 Entry Point PROMPT2: 10$: 5‘ ~l ‘v V- - ‘O ‘1 5. ‘u ‘0 ~' MOV @O.PMPT(R5).R1 3 Get the prompt DEC R1 3 Convert to z-base ASL R1 3 word-oFFset addr MOV GETLIN.R0 3 Point GCMLS RO.PADD(R1).PLEN(R1) 3 Get the command TSTB G.ERR(RO) 3 Were there errors BGE 10$ 3 GE => No errors EXIT$S 3 Errors => quit MOV G.CMLD(RO).R2 3 Get the oF char MOV R2.@O.LEN(R5) 3 and send back MOV {G.CMLD+23(RO).RO 3 Get starting addr MOV O STR(R5).R1 3 Get starting addr 3 that string CALL SCVTUC 3 Convert to upper RETURN 3 Go home .END .PSECT CMD .TITLE CMD .IDENT /V1.0/ .ENABL LC CMD.MAC This routine uses the parser to determine input strings and execute the proper subroutines P. HoFFman. P. Aiello 6/14/82 Declare global reFerences. entry to state tables.symbol table and keyword table .GLOBL CRNCMD.CRNSTB.CRNKTB Call some oF HoFFman’s macros .MCALL PUSH.POP Macros necessary For parser Na 5' ‘0 ~e - 3' ARGLST: ARGl: 3' 137 .MCALL ISTAT$.STATE$.TRAN$ Create macro CLSUB. saves registers. points ARGLST necessary. calls the appropriate subroutine. and restores the registers .MACRO CLSUB.NAME.ARGS PUSH .IIF NB.ARGS. MOV ARGLST.R5 CALL NAME POP CR5.R4.R3.R2.R1.RO> .ENDM CLSUB Allocate space For the argument list .BLKW 1 .WORD ARGl .BLKW 1 OFFset to command string and error string in call O.STR=2 O.ERR=4 Entry point MOV MOV MOV MOV CALL O.@O.ERR(R5) O.STR(R5).RO 3clear the error 3pointer to cmd RO.R1 3put result 80 .R2 380 char string SCVTUC 3and convert to uc set up For parser MOV MOV MOV MOV PUSH MOV CALL POP BCC MOV RETURN Create ISTAT$ STATES TRANS TRAN$ TRANS .IS 1000.R1 32 char. abbrev. CRNKTB.R2 3point to keyword 80..R3 380 char string O.STR(R5).R4 3string pointer R5 3save R5 CRNCMD.R5 ‘state table entry .TPARS 3and parse it R5 3restore args 3check iF error 3set error Flag 3and go home -1.@O.ERR(R5) state table CRNSTB.CRNKTB 3initialize CRNCMD 3speciFy entry poin "LOGARITHM".$EXIT.LOGACT "SUBTRACT“.$EXIT.SUBACT "ABSORBANCE".$EXIT.ABSACT LOGACT: SUBACT: ABSACT: ORDACT: AVGACT: HELACT: STOACT: NREACT: ENTACT: FITACT: DFIACT: EAACT: EXIACT: TRAN$ TRAN$ TRANS TRAN$ TRANS TRAN$ TRAN$ TRANS TRAN$ TRANS TRAN$ STATES Parser .PSECT MOV MOV CLSUB RETURN MOV MOV CLSUB RETURN CLSUB RETURN CLSUB RETURN CLSUB RETURN CLSUB RETURN CLSUB RETURN MOV MOV CLSUB RETURN CLSUB RETURN CLSUB RETURN CLSUB RETURN CLSUB RETURN CLSUB RETURN . END 138 "ORD".$EXIT.ORDACT "AVERAGE“.$EXIT.AVGACT "HELP“.$EXIT.HELACT "STORE“.sEXIT.STOACT "NOISE".$EXIT.NREACT "NR".$EXIT.NREACT "ENTER".$EXIT.ENTACT "FIT".$EXIT.FITACT "DFIT".$EXIT.DFIACT "EA".$EXIT.EAACT "EXIT".$EXIT.EXIACT 3closes state table action routines SSTATE 1.ARGLST 3 ARGS=1 1.ARGl 3and it is =1 LOG.1 3call LOG 3and go home 1.ARGLST 1.ARG1 SUBTR.1 ABSORB ORD AVG HELP STORE 1.ARGLST 1.ARG1 NRED.1 ENTER FIT DFIT EA EXIT REFERENCES 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 139 REFERENCES Willard, Merritt and Dean. 'Instrunental Methods of Analysis'. Fifth Ed.. D. VanNostrand Co.. New York. 413 (1974) R. A. Barber, G. E. Sonner. Applied Optics. g. 1039 (1966) A. Danielsson. P. Lindblon. E. Soderman. 923513; Scripts. 6. 5 (1974) P. B. Boyce. Science. 198. 145 (1977) I. J. lilano, E. L. Pardue. T. E. Cook. R. E. Santini, D. W. Nargerun, J. Rachera, Anal Chen‘. 46‘, 374 (1974) J. A. deHaseth, I. S. Woodward. T. L. Isenhour, Anal Chem, 18. 1513 (1976) T. E. Cook, R. E. Santini, E. L. Purdue. Anal Chen. 12. 871 (1977) T. A. Nienan. C. G. Enke. Anal Chem, fig, 619 (1976) Y. Talmi. Anal Chem. 11. 658A (1975) Y. Talmi, Anal Chen, 11, 697A (1975) Y. Talmi, A; Lab. March, 79 (1978) N. G. Howell, G. M. Morrison, Anal Chem, 12, 106 (1977) R. P. Cooney. T. Vo-Dinh. G. Walden, I. D. 'inefordner. Anal Chem. 13. 939 (1977) R. P. Cooney. G. D. Boutilier, J. D. Winefordner, Anal Chem, 52, 1048 (1977) K. L. Ratzlaff, S. L. Paul. A221 figectroscogx, 33. 240 (1979) K. L. Ratzlaff. Anal Chem, 3;. 916 (1980) E. A. Lewis, N. B. Danton. l; 2; Anton, Chen, Preprint August (1980) G. Horlick. E. G. Codding. Anal Chem, 1;. 1490 (1973) D. A. Yates. T. Kuwana, Anal Chem, 48, 510 (1976) 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 140 G. Rorlick. A221 S2ectrosco2z. 12. 113 (1976) R. W. Simpson, Rev, Sci, Instrum. 29. 730 (1979) P. J. Aiello. N. S. Thesis. Michigan State Univ. (1980) Y. Talmi. R. W. Simpson. A221 02tics. ;2. 1401 (1980) D. F. Merino, J. D. Ingle Jr.. Anal Chem. 2;. 645 (1981) Y. Talmi. A221 §2ectrogco21. 12. 1 (1982) F. S. Chuang. D. F. S. Natusch. K. R. O'Keefe. Anal Chem. 29. 525 (1978) I. J. Iilano. K. Kim. Anal Chem. 12. 555 (1977) E. Franklin. C. Baber. S. R. Koirtyohann. S2ectrochinica Acts. 1;. 589 (1976) G. Rorlick. E. G. Codding. A221 §2ectrosco21. 29. 167 (1975) T. E. Ednunds. G. Horlick. A221 §2ectrogco21. 2;. 536 (1977) Y. Talmi, E. P. Sieper. L-Moenke-Bankenburg. Analytics Chinica Acts. ;21. 71 (1981) M. I. Blades. G. Rorlick. A221 §2ectr02co2x. 11. 696 (1980) B. R. Sandel. A. L. Broadfoot. A221 02tics. g;, 3111 (1976) N. A Ryan, R. J. Miller. J. D. Ingle Jr.. Anal Chem, 22, 1772 (1978) E. G. and G. Reticon. S-Series Solid State Linear Diode Array. Brochure (1978) E. H. Eberhardt. A221 02tics, 2Q. 1418 (1979) A. L. Broadfoot. B. R. Sandel. A221 02tics. $2. 1533 (1977) R. L. Felkcl. E. L. Purdue. 222; Chem, 50. 602 (1978) A. E. McDowell. H. L. Purdue. Anal Chem, 12, 1171 (1977) 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 141 R. E. Dessey. W. D. Reynolds. W. G. Nunn, C. A Titus. G. F. Moler. Clin Chem. 22. 1472 (1976) R. P. Cooney, T. Vo-Dinh. J. D. Winefordner. Anal Chim Acts. 22. 9 (1977) B. Newcome. submitted for publication Product Specifications on Model 1070 Intelligent Disk Controller, Persci. Marina del Rey. CA (1976) E. Rather. L. Brodie. 'Using FORTE'. Forth-79 Std. Ed.. Forth Inc. Third Ed.. C. Rosenberg. Editor (1981) T. V. Atkinson. Department of Chemistry. Michigan State Univ. (1982) E. M. Carlson. Ph. D. Thesis, Michigan State University (1978) S. S. Vogt, R. G. Tull. P. Kelton. A221 02tics. ;l. 574 (1978) T. B. McCord. M. J. Frankston, A221 02tics. lg. 1437 (1978) F. J. Holler. T. V. Atkinson. T. G. Kelly. C. G. Enke. 22 Lab. September. 9 (1976) C. E. Moore. the. August. 76 (1980) J. S. James, the. August. 100 (1980) I. i. 67 (1962) A. Moscowitz. Advan 1 Chem Phys Part C. Djerassi. 'Optical Rotatory Dispersion'. McGraw-Hill. New York (1960) T. B. Crumpler. W. R. Dyre, A. Spell. 2221 Chem, 21. 1645 (1955) Carroll. R. B. Tillem, E. S. Freeman, Anal Cheg. 1099 (1958) A. Savitsky. M. Golay. Anal Chem, 12. 1627 (1964) A. Proctor. Anal Chem. 2;. 2315 (1980) J. G. Foss. 1‘ Chem I3 12. 592 (1963) A573 (1974) '0 as u I“ [H K. Wong. 1‘ Chem 11 Chem A9 (1975) IN as 5 I‘ll IN 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 142 Wong. 1; Chem Ed. 5;. A83 (1975) K. V. C. Zadnik. J. L. Scott. R. Megargle. I. Kerkay. K. H. Pearson. 1y Autom Chem. 1. 206 (1979) D. V. Amato. G. I. Ewing. Analytical Letters. 1. 763 (1974) J. R. Krivacic. D. E. Wisnosky. D. W. Urry. Anal Biochem. 13. 547 (1971) W. M. Martx. S. Akipis. Anal Biochem. 39. 327 (1971) P. H. Schippers. E. Dekkers. Anal Chem. 5;. 778 (1981) J. Y. Cassim. I. T. Yang. Biochemistry. 8. 1947 (1969) D. F. DeTar. Anal Chem. 1;. 1406 (1969) G. C. Chem. J. T. Yang. Analytical Letters. ;9, 1195 (1977) V. C. Zadnik. K. H. Pearson. Analytical Letters. 1;. 1267 (1979) K. H. Pearson. V. C. Zadnik. J. L. Scott. Analytical Letters. 1; 1049 (1979) I. C. Krueger. L. M. Pschigoda. Anal Chem. 4;. 673 (1971) P. A. Hoffman. C. G. Enke. Computers and Chemistry. 1. 47 (1983) I. E. Deming. ’Statistical Adjustment of Data'. Wiley, New York (1943) T. M. Lowry. 'Optical Rotatory Power'. Dover Publications. New York (1964) G. R. Minot. W. P. Murphy. 1; A; M; A;. 81. 470 (1926) E. L. Rickes. N. S. Brink. F. R. Koniuszy. T. R. Wood. R. Folkers. Science. 101. 396 (1948) E. L. Rickes. N. S. Brink. F. R. Koniuszy. T. R. Wood. R. Folkers. Science. 198. 134 (1948) Merck Service Bulletin. 'Vitamin B 12' (1958) 143 80. E. L. Smith. 'Vitamin B 12'. Third Ed.. Wiley. New York (1965)