r. w: 1111 - 1.11.... can... 17w. 1--.—.-::.. 19-.» "3' Va NS GPMEM, ‘3 E5. DEV OFA -' ‘ s was: V. . V V V . V V V . . i V .VV V . V : . I . . 4 : V 1..... 1uV- V .V..VV.......VV...VV V . .VVV...V.V.V..VV .V......V.....__V.. V: ..V._V.. _..V....1 V V VV V...... VV .1 V V . 11 111.12.. V .. . 1 . .11..q....y: V V V V. V. V V V V .../...‘..V11_.,V.V ./ 1.1.1123: . .33.... (#1.: .V 1 . ...1..1.:..1.1..L..V11 V :1...11c: .:V.: V V. : V 1.1.1501... V V r. V . ..1. .V.. V .. . . V 171.11.111.1r11s .r:..:.1p:n.1 . V 1:. '11.». 1 11 V.V..V:.V . :1... 1.V:1u,.1:..VV.:V..i..t:V11/( 11:: :11".1...1:..V1V..:r. . V V V . .:.4: 1.11:. .V...1.._.;. 111' V.V,.1..1:V.1;1.V..n. 4..1v..:VV:m 1 . V Hr: . 11V: . . I. .:¢-o-.-l‘ . V ..~;1_.s1V../:1¢O:.:1.V1:1111V1 11:111.... .1 21.1.. . V. V v 21. ..../ V11.¥.:V1.:.ar:,11_11¢ .1.¢.1.. V V V V .V .121? 11.111. 2.111.111 12111.1...1) #11: V .11..... w. 1......1..rf {:4 . V V :31;.111.11.1:V.Tr.o:1:.91.......1:.1:1... . V 11 ....VVV1 V.:.V11. 115:}. 1 ..1 1..1.V..11,:. au- ‘N. 1K ’- V11 .11.. . . V1: 7...: 1..! . V 11.1.1. 1.11.1.1... V V V V 2.7:. V: .1 1 :..o...:...: .4V.:.11.>1.V.'.1.V 1 1.. SPECTRO‘ V. V 1 . 20-57. V» 1111 .01 .1111... .111 1... Vin/v1. 114111.. 11V". .! Vrrwl NTROLLED 1::r.,».1 1.11:9:.V:.. 171:1».V 1:4: anoxr: :11... . . U71V1VV .1Vr:.:..V1 .11....1..::—1.:V3.V . vi. :1VV..:v:or1.v. ammo "CG ‘— {COIN fi » 7 AM) APP OMPUTER l 1-1.3.11. ...:.1 . .l Vr:.:V C 1 .. x V V . V. .V V . .V ..V. :_ V.. . . V V . . V .VV . . . . .: 1 V: V I . . 1 . : n. V t r' l : I l: o . .. . V. . .2 . .. VVV . . . .. 1. . V V V . - . 1’ :. 1 V . u 1:- Vc ' r I: :1 1 : IV. : : . . V V V . . : 1 . V V -4 V V.... n ..t .V. .....1 . .: :V1: 1 V V V 1 .V V. V : . . I . V — v . t '1 r 1 . .V . V . : : .: V 1:! 1 :. :.V1v V V V. P . . .. .: I : u l I 1 '1 n n! 1 t 1: u A ..h D .1 A. r: ' c I: . : :1 0 I AI 1 1 r q . V V .V . .V . n 1 A v V . . 1. V. V .V. . 1 V:V.LV 1:. . V .1.: 1 . v: V 1 V .V1V_:._1..... . 11:11.: ”.VV .V 1. V 1 . . V V 1.. V .V 1 .V. . VVVV . V. I... .1... .1 .V.—.V....V V.. .V . Cr».r15.3.1:V.V. . . V “"4“; .V : LIFER 41;i Y Micl igan Stiff? U dvczsity , ~ This is to certify that the thesis entitled DEVELOPMENT, CHARACTERIZATION, AND APPLICATIONS E? OF A COMPUTER-CONTROLLED SILICON VIDICON SPECTROMET presented by Timothy Alan Nieman has been accepted towards fulfillment of the requirements for PhD deg“. in Chemistry [9/9 2;; Major professor 5' alumna av ‘ MUM; & SIMS ABSTRACT :- DEVELOPMENT, CHARACTERIZATION, AND APPLICATIONS OF A COMPUTER-CONTROLLED SILICON VIDICON SPECTROMETER By Timothy Alan Nieman A computer-controlled rapid scan spectrometer has been develOped which uses a silicon vidicon tube as a multichannel detector. The spectrometer can monitor a 230 nm window in the range of 380-900 nm with 4.2 nm resolution and wavelength linearity better than 0.3%. Spectrum scan times as fast as 2 milliseconds are possible. Under computer control the wavelength window can be divided into between 32 and 4096 wavelength channels. The tube readout beam can be deflected to any channel at random or made to scan them sequentially. The readout beam can be inhibited from scanning any channel, to increase the target's integration time and enhance weak signals. The system has a signal-to-noise ratio (S/N) of 220. Signal averaging increases the S/N in proportion to the square root of the number of spectra averaged. A S/N of 104 has been reached by averaging 2048 spectra. The S/N increases linearly with the target integration time up to at least a twenty-fold S/N enhancement. The detector responds linearly to the incident light level over at least 3 l/Z orders of magnitude, with response becoming nonlinear above 60% of target saturation. To provide guidelines for application of least squares smoothing in processing the spectra from the vidicon spectrometer, a systematic study was made of the compromises between noise reduction and signal Timothy Alan Nieman distortion due to smoothing. The standard deviation of normally distributed noise was seen to be reduced in proportion to the square root of the width of the smooth and approximately in proportion to the eighth root of the number of passes of the smooth. Distortion of peak height, width. and area is a function of the smoothing ratio, which is the ratio of the width of the smooth and the full width at half maximum (FNHM) of the peak. Distortion is minimal for a single pass of a smooth- ing function with a smoothing ratio less than one. Combining observa- tions on signal distortion and noise reduction indicates that for a peak of given FNHM there is a maximum S/N enhancement factor attainable through smoothing. The rapid scan capabilities of the vidicon spectrometer were used to study the kinetics of the complexation reaction between cyanamide and sodium pentacyanoammineferrate (SPF) in buffered, basic solutions. The initial rate equation at constant pH was found to be first order in SPF and first order in cyanamide. At pH 10.5 the second order rate constant is l6.4 l mole”1 sec". The pH was seen to affect the initial rate by controlling the fraction of cyanamide present as the monoanion and the fraction of SPF in the aquated form.= By monitoring the absorbance of the SPF-cyanamide complex at 530 nm, it was shown that the initial rate can be used to determine trace amounts of cyanamide over a concentra- tion range of 3 l/2 orders of magnitude with a limit of detection of 0.24 ppm. The kinetics of the chemiluminescent oxidation of luminol in DMSO and in lzl DMSO-EtOH was studied by monitoring the absorption spectrum and by simultaneously measuring the solution conductance, temperature, Timothy Alan Nieman and intensity of chemiluminescent emission. The decomposition of a proposed peroxy intermediate to form excited state 3-aminophthalate was determined to have a pseudo first order rate constant of 1.3 sec'] in the DMSO-EtOH solvent. DEVELOPMENT, CHARACTERIZATION, AND APPLICATIONS OF A COMPUTER-CONTROLLED SILICON VIDICON SPECTROMETER By Timothy Alan Nieman A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1975 To Sandy (my sweet Sport) and Mom and Dad ii ACKNOWLEDGMENTS I wish to offer my sincere appreciation to Professor Chris Enke for his guidance and friendship during the course of my research. Of particular value were many helpful and thought provoking discussions we had in which he shared his experiences and insights into the career of an analytical chemist. Thanks go to the faculty and staff of the MSU Chemistry Department for the excellent support they offered. I wish to express gratitude to the members of my guidance committee, especially to Professor J.L. Dye for serving as second reader for this thesis, and to Professor Stan Crouch for many useful discussions. Special thanks are extended to Kathy Shippy who typed the final copy of this thesis under the usual pressure of a deadline. I am very appreciative of financial support received from the MSU Chemistry Department in the form of an L.L. Quill Fellowship, from a fellowship provided by the Eastman Kodak Co., and from the ACS Analytical Division Fellowship sponsored by the Proctor and Gamble Co. The Heath Co. is also acknowledged for supplying the funds for the silicon vidicon tube and the deflection assembly used in the project. I wish to thank the fellow members of the C.G. Enke & Co. research group for their beneficial interactions and valuable friendships, especially Keith Caserta, Brian Hahn. Ed Darland, and Jim Holler. iii To my parents I am very grateful for the support and encouragement they have always given me. Their pride in me has made the success of my graduate school career all the more rewarding. Most important thanks and acknowledgment go to my wife Sandy. Besides being a loving wife she has been a valuable consultant and my best friend. She has encouraged me throughout my graduate education, taught me all of the probability and statistics used in my research, read and critically commented on this entire thesis, typed the initial draft, and helped me proofread the final c0py. iv TABLE OF CONTENTS LIST OF TABLES ......................... LIST OF FIGURES ......................... CHAPTER l - Introduction .................... CHAPTER 2 - Perspective on Imaging Devices in Spectroscopy . . . A. Introduction ...................... 8 Silicon Vidicon Tube Operation ............. C. Survey of Imaging Devices ............... 0 Historical Applications of Imaging Devices in Spectroscopy ...................... CHAPTER 3 - Routine Operation of the Instrument ......... CHAPTER 4 - Performance Characteristics of the Silicon Vidicon Spectrometer ................ A. Introduction ...................... B. Multichannel Spectral Response ............. C. The Analytical Signal ................. 0. Dark Current ...................... E. Random Noise ...................... F. Wavelength Linearity and Resolution .......... 6. Software ........................ CHAPTER 5 - Evaluation of Least Squares Smoothing of Digitally Recorded Data ............... A. Introduction to the Smoothing Process ......... B. Noise Reduction .................... C. Signal Distortion ................... D. Signal-to-Noise Ratio Enhancement ........... Page ix x l 7 7 15 35 42 55 55 56 59 71 74 80 84 86 86 89 102 112 Chapter Page E. Frequency Analysis ................... 120 F. Software ........................ 125 6. Conclusions ...................... 125 CHAPTER 6 - Studies on the Reaction of Cyanamide and Pentacyanoammineferrate ............... 132 A Background ....................... 132 B Role of the Vidicon Spectrometer ............ 133 C. Observations on the Reaction .............. 140 0 Initial Rate Studies .................. 142 E Conclusions ...................... 149 F. Software ........................ 151 CHAPTER 7 - Studies on the Chemiluminescent Oxidation of Luminol in DMSO ................. 156 A. Background ....................... 156 B. Conductance Studies .................. 158 C. Spectroscopic Studies ................. 161 CHAPTER 8 - Documentation of System Design ........... 166 A. Introduction ...................... 166 B. The Computer System and the Computer-Interface Buffer . 166 C. Analog Circuits .................... 168 D. The Vidicon Interface ................. 173 E. Programming Notes ................... 180 F. Optical Setup ..................... 182 CHAPTER 9 - Conclusions ..................... 185 APPENDIX A - Probability and Statistics ............. 187 A. Least Squares Estimates of Parameters of the Straight Line ..................... 187 vi Chapter 8. t-Test for Hypothesis Testing ............. C. Congruential Method to Generate Uniformly Distributed Random Numbers ............... 0. Central Limit Theorem Approximation to Generate Normally Distributed Random Numbers .......... E. Chi-Square Goodness of Fit Test ............ APPENDIX B - Supplementary Data Tables for Chapter 5 ...... APPENDIX C - Routine Operation Program Listings ......... Program VCIDCO.FT ..................... Program VSIGAV.FT ..................... Subroutine DBL2.FT ..................... Subroutine DBL3.FT ..................... APPENDIX D - Instrument Characterization Program Listings. . . . Program VDARK5.FT ..................... Program VDRDC.FT ...................... APPENDIX E - Smoothing Evaluation Program Listings ....... Subroutine IRANU.FT .................... Subroutine IRANG.FT .................... Subroutine RAND.SB ..................... Subroutine SMOTH.SB .................... Subroutine SMOTH.FT .................... Program CHISQ.FT ...................... Program RANDT4.FT ..................... Program RANDT3.FT ..................... Program LSTSSl.FTN ..................... Program GAUSSB.FT ..................... vii Page 188 188 189 190 192 198 198 206 213 214 215 215 218 222 222 222 223 227 233 235 239 240 242 243 Chapter Page Program GAUST2.FT ..................... 245 Program FOUR4.FT ...................... 247 APPENDIX E - Kinetics Program Listings ............. 250 Program VNTAMO.FT ..................... 250 Program VNTMIN.FT ..................... 250 Subroutine DSCRIB.FT .................... 251 Subroutine DRCTRY.FT .................... 252 Subroutine NINIT.FT .................... 254 Subroutine VNTDA.FT .................... 257 Program VNTADS.FT ..................... 266 Subroutine PTPLT.FT .................... 271 Subroutine LLSQ.FT ..................... 271 REFERENCES ........................... 272 viii LIST OF TABLES Table Page 2-1 Vidicon Target Comparison ................. 22 2-2 Comparison of Silicon Vidicon, SIT, and SEC Tubes ..... 24 4-1 Linearity and Range .................... 61 4-2 Signal Enhancement with Charge Integration ......... 68 4-3 wavelength Linearity .................... 81 5-1 Smoothing Uniform and Normal Noise ............. 94 5-2 Noise Reduction with Multiple Pass Smoothing ........ 97 5-3 Noise Reduction with Mixed Length Smoothing ........ 101 5-4 Height Response After Smoothing Gaussian Peaks ....... 192 5-5 Hidth Response After Smoothing Gaussian Peaks ....... 193 5-6 Area Response After Smoothing Gaussian Peaks ........ 194 5-7 Height Response After Smoothing Lorentzian Peaks ...... 195 5-8 Hidth Response After Smoothing Lorentzian Peaks ...... 196 5-9 Area Response After Smoothing Lorentzian Peaks ....... 197 5-10 S/N Enhancement Factors .................. 114 5-11 Maximum S/N Enhancement for a 16 Point Hide Peak ...... 118 5-12 Software for Evaluation of Least Squares Smoothing ..... 126 6-1 Initial Rate Studies .................... 143 8-1 Power Supply Requirements ................. 169 8-2 Codes for Wavelength Points per Spectrum .......... 177 8-3 DS-IOP Codes for the Vidicon Spectrometer ......... 181 ix LIST OF FIGURES Figure Page 1-1 Single Channel Versus Multichannel Detection ...... 2 2-1 Imaging Device Block Diagram .............. 8 2-2 Silicon Vidicon Target Construction ........... 10 2-3 Silicon Vidicon Signal Production ............ 10 2-4 Vidicon Tube Geometry .................. 13 2-5 Beam Scan Pattern .................... 13 2-6 Family Tree of Imaging Devices ............. 16 2-7 Photoemissive Signal Production ............. 18 2-8 Photoconductive Signal Production ............ 18 2-9 SEC Target and Signal Production ............ 21 2-10 Image Dissector ..................... 26 2-11 Linear Photodiode Array ................. 29 2-12 Discrete Component Bucket Brigade Device ........ 31 2-13 Bucket Brigade Device .................. 31 2-14 Charge Coupled Device .................. 33 2-15 Charge Transfer Scanning ................ 34 2-16 Publications Concerning Imaging Devices in Spectroscopy ...................... 37 3-1 System Diagram ..................... 43 3-2 Target Scan Pattern, Incident Spectral Lines, and Resulting Signal .................... 44 3-3 Program VCIDCD Teletype Interactions .......... 46 3-4 Mercury Emission Spectrum ................ 50 3-5 Absorption Spectra for Mixtures of SPF and Cyanamide . . 51 3-6 Transmission Spectrum and Beer's Law Plot for PrC13. . . 52 X Figure 3-7 4-1 4-2 4-3 4-4 4-5 4-6 4-8 4-9 4-10 4-11 4-12 4-13 4-14 4-15 4-16 4-17 4-18 5-1 5-2 5-3 5-4 5-5 Single Precision Versus Double Precision Averaging . . Gross Spectral Response Curve .............. Channel Sensitivity Curve ................ Overall Detector Response Curves ............ Components of the Vidicon Spectrum ........... Nonlinearity Near Detector Saturation .......... Signal Transfer Function ................ Signal Enhancement with Charge Integration ....... Observed Signal Versus Exposure Time .......... Baseline Glitches from Selective Integration ...... Fixed Pattern Noise due to Background Dark Current . . . S/N Enhancement due to Signal Averaging without Dark Current Subtraction ................ S/N Enhancement with Signal Averaging (Using the Same Number of Scans of the Signal and the Dark Current). . . S/N Enhancement with Signal Averaging (Using a Stored Estimate of the Dark Current) .............. RMS Noise with Charge Integration ............ S/N Enhancement with Signal Averaging and Charge Integration ....................... Dynamic Range of the Vidicon Spectrometer ........ ‘Response Across Light-Dark Edge ............. Response Versus Number of Wavelength Channels ...... Noise Output from Random Number Generators ....... Histograms of Noise Generator Outputs .......... Smoothing Normal and Uniform Noise ........... Noise Reduction with Multiple Pass Smoothing ...... Multiple Pass 5 Point Smoothing on Noise Arrays with Different Standard Deviations .............. xi Page 54 57 57 58 60 63 64 66 67 70 72 73 75 75 77 78 79 83 83 91 92 95 98 100 Figure 5-6 5-7 5-9 5-10 5-11 5-12 5-13 5-14 5-15 5-16 5-17 5-18 5-19 6-2 6-3 6-5 6-6 6-8 6-9 Gaussian and Lorentzian Peaks Before and After Smoothing ........................ Wing Distortion due to Smoothing Gaussian Peaks ..... Single Pass Distortion for Gaussian Peaks ........ Single Pass Distortion for Lorentzian Peaks ....... Height Response Factor for Multiple Pass Smoothing . . . Width Response Factor for Multiple Pass Smoothing. . . . Area Response Factor for Multiple Pass Smoothing . . . . S/N Enhancement for an 8 Point Wide Peak ........ S/N Enhancement for a 16 Point Wide Peak ........ S/N Enhancement for a 32 Point Wide Peak ........ Example of Smoothing to Maximize S/N Enhancement . . . . Frequency Relations of Signals, Noise, and Smoothing Functions ........................ Frequency Analysis of Minimum (0.5%) Peak Height Distortion due to Smoothing ............... Frequency Analysis for Maximum S/N Enhancement due to Smoothing ...................... Absorption Spectra for Mixtures of SPF and Cyanamide . . Absorption Spectra During Reaction of SPF and Cyanamide. Absorbance Versus Time for Figure 6-2 .......... Absorption Spectra During Reaction of SPF and Cyanamide. Absorbance Versus Time for Figure 6-4 .......... Absorbance of Reactant and Product During SPF-Cyanamide Reaction ........................ Initial Rate Versus Cyanamide Concentration ....... Initial Rate Versus SPF Concentration .......... Effect of pH on Initial Rate Compared to the Monoanion Fraction ................... Page 104 105 106 107 109 110 111 115 116 117 119 122 123 124 134 137 137 138 138 139 144 145 147 Figure Page 6-10 Kinetics Software Package ................ 152 6-11 Parameter Input for Timed Data Acquisition ....... 153 6-12 Data Run Descriptive Text ................ 154 7-1 Conductance, Chemiluminescence, and Integrated Chemiluminescence During Reaction ............ 162 7-2 Absorbance and Integrated Chemiluminescence During Reaction ........................ 164 8-1 Block Diagram of Instrument Circuits .......... 167 8-2 Schematic of Line Deflection Circuit .......... 171 8-3 Schematic of Preamplifier Circuit ............ 172 8-4 Signal Conditioning, Sampling, and Conversion Circuit. . 174 8-5 Wavelength Deflection Circuit .............. 176 8-6 Logic for Wavelength Counter .............. 178 8-7 Control Signals for the Wavelength Deflection Circuit. . 179 8-8 Optical Setup of the Vidicon Spectrometer ........ 183 xiii CHAPTER 1 INTRODUCTION The modern analytical chemist is often concerned with developing new methods and instrumental tools for analysis. We seek to examine increasingly more complex problems whose solutions require ever greater amounts and different types of information at ever faster rates. In our quest to understand the nature of the chemical world we must often reach out to other disciplines for tools to extend or complement those tools we currently possess. Particularly in the area of spectroscopy, chemists often discover they can benefit from a new method of excitation, isolation, or detec- tion, developed in another field of science or engineering, but modified for chemical analysis. Conventional (non-multiplexed) spectroscopy has used either single channel electrical detection (with photodiodes or photomultiplier tubes) or multichannel parallel detection with a photo- graphic plate (Figure l-l). With single channel detection, after the spectrum is dispersed an exit slit is used to isolate all but a narrow wavelength region which strikes the detector; at any time, the majority of the available spectral information is wasted. Use of a photographic plate for multichannel detection makes simultaneous use of all the available spectral information but requires several additional steps to obtain an easily used electrical signal. Recently, considerable interest has been aroused by the prospect of using TV-type multichannel detectors or "imaging devices" to combine the advantages of the single DISPERSING ,A‘ ELEMENT ,- _\ 1 “ A ‘ A 4A 1 SOURCE EXIT SLIT FOCAL PLANE ENTRANC/E SLIT Figure l-l. Single Channel (top) Versus Multichannel (bottom) Detection. 3 channel electrical detectors with those of the multichannel photographic plate. The subject of this thesis is the development, characterization, and applications of a spectrometer based on such a detector. An imaging device responds to an incident light pattern to produce a position-encoded electrical signal whose magnitude at a particular time is proportional to the incident light intensity at a corresponding location. The imaging device can be thought of as an array of semi- independent detectors. When the device is placed in the focal plane of the dispersed spectrum, each detector element monitors the light intensity at the wavelength region it intercepts. All of the detectors are active simultaneously, but the information they record is read out sequentially. Most imaging devices are integrating detectors, which means that weak signals can be enhanced by lengthening the exposure time between read- outs. Also, the device can record the total light output of flashes whose duration is shorter than the detector read out time. Just as with single channel detectors, imaging devices have greater sensitivity than photographic plates and produce signals in the electrical domain. A unique property of imaging devices is that they can scan a spectrum without any mechanical motion. A discussion of the types of imaging devices applicable to spectroscopy and of the operation of the silicon vidicon tube used in this research is contained in Chapter 2. Chapters 3 and 8 discuss the routine operation and the construc- tion details, respectively, of the rapid scan vidicon spectrometer. The instrument can monitor a 230 nm window in the visible and near IR regions with scan times down to 2 milliseconds. A dedicated mini- computer and associated interfacing electronics have been integrated 4 into the system to add power and flexibility through the computer's ability to handle and analyze the large amount of data produced by the instrument, make decisions, and provide real-time control of instrument operation. Through the computer and the operating software, the user can select the wavelength resolution and the spectrum signal-to-noise ratio (S/N) using signal averaging, background subtraction, control of detector integration time, and least squares smoothing of the acquired data. Before an instrument or method can be used with confidence its operation must be well enough characterized to know whether the measurement in question has been distorted, and if it has been distorted, to enable interpretation of the altered form. The necessity for complete characterization was made graphically clear, when during the course of using the vidicon spectrometer to study the kinetics of a reaction. an undiagnosed imperfection was incorrectly interpreted as originating from the chemical system under study. Chapter 4 describes the work done to study the performance characteristics of the vidicon spectrometer. Items examined include multichannel spectral response, wavelength linearity and resolution, and signal dependence on the incident light level and on the integration time. Factors limiting S/N and computer-interactive methods for S/N enhancement are investigated. Least squares smoothing is often used during data analysis and display to remove random noise. This technique was found to be of use in the software set of the vidicon spectrometer system. Since complete and easily interpreted guidelines were lacking, a systematic study was undertaken, as described in Chapter 5, to learn the compromises between 5 noise reduction and signal distortion that result with smoothing. It was seen that for a given spectrum. there is a maximum S/N enhancement attainable through smoothing. Chapter 6 describes the application of the rapid scan and multi- channel capabilities of the vidicon spectrometer to study the kinetics of a complexation reaction of cyanamide. The instrument software was extended to handle acquisition, storage, display, and analysis of kinetics data. The instrument was able to monitor over a wavelength region that covered the absorbances of both reactant and product. Besides providing insight into the reaction the investigation showed that an initial rate method could be successfully employed to determine cyanamide at sub part-per-million levels. A brief but multifaceted study of the chemiluminescent oxidation of luminol was undertaken and is described in Chapter 7. After stopped flow mixing, the reaction was followed by simultaneously monitoring the bulk solution conductance, the solution temperature, and the intensity of light emission. In separate experiments the absorbance spectrum was monitored during the reaction. Each chapter of this thesis has been written to be as independent of the others as possible to enable the reader to examine only those parts of particular interest to him. In a few places it is felt that familiarity with the material in one section is a prerequisite to the understanding of another section. The studies on performance character- istics in Chapter 4 can be more fully appreciated by first reading the section concerning the theory of operation and signal production with the silicon vidicon tube as described in Chapter 2 and the description 6 of routine instrument operation contained in Chapter 3. Familiarity with the routine operation chapter will also aid analysis of the instrument design details outlined in Chapter 8. CHAPTER 2 PERSPECTIVE 0N IMAGING DEVICES IN SPECTROSCOPY A. Introduction The purpose of this chapter is to provide operational and historical perspective on the use of imaging devices in spectroscopy. The first section places emphasis on the theory of operation and signal produc- tion of the silicon vidicon tube. It should be read and understood before proceeding to the chapters concerning operation (Chapter 3), characterization (Chapter 4), and design (Chapter 8) of the instrument developed in the course of this research. The second section briefly discusses the entire family of imaging devices. This survey will enable the reader to evaluate each type of device for potential spectroscopic applications. The last section traces the history of imaging device applications in spectroscopy from the origin of the idea in 1938 to the present flurry of interest. The contents of the last two sections are helpful but not essential to an understanding of the subsequent parts of this thesis. 8. Silicon Vidicon Tube Operation At the heart of the silicon vidicon spectrometer is the silicon vidicon camera tube. Its Operation can best be understood by first noting the functional units common to all imaging devices of the signal-generating charge-storage type (Figure 2-1). First there is .Emgmmwc xuopm mup>mo mcwmmsm ._-~ meaowm 40”:on 25m l ._.Dn:.DO 65.05 m <3: m m< 0220:5on /0 0 :0 S— VIDEO OUTPUT Figure 2-4. Vidicon Tube Geometry. START Figure 2-5. Beam Scan Pattern. 14 Tr=readout time to scan area A, sec Te=detector exposure time between readouts, sec (5). If the photon flux is constant over the exposure time, this reduces to I=kAN(Te/Tr) where k equals the constant QeG. Typical signals are on the order of a few hundred nanoamperes. Considering these relations and the material presented earlier we can see the following points concerning use of the silicon vidicon as a spectroscopic detector. 1) All areas (wavelengths) of the detector are active simultaneously and are divided into discrete channels of information by the scanning pattern of the electron beam. 2) One of these channels at a time is sequentially converted into an electrical signal. 3) There is a linear relationship between the intensity of the incident light and the magnitude of the resulting photocurrent. 4) The spacing of the diodes places an upper limit on the spatial resolution of the detector. 5) The scanning time of the electron beam places an upper limit on the temporal resolution of the detector. 6) The detector has storage capability and holds a signal until read off by the electron beam. 7) Even though the detector can not time-resolve signals occurring faster than the scan time, it will be able to record the time-integrated spectral output Of a short flash (provided the intensity is great enough). 8) A large portion of the sensitivity Of the device is due to combination of the advantages of a multichannel detector and an 15 integrating detector. For an illumination constant over the time Te, the ratio of Te/Tr determines the factor Of signal enhancement through integration. If we consider the target to contain 500 areas of information on each axis, and the beam spends an equal amount of time on each area, then (Te/Tr)=(At/A)=250’OOO’ where At is the total target area. 9) The velocity of the beam across the target must be held constant in order that Tr remain constant. This requires the beam deflection circuits to be very linear. 10) The ability to control the time and sequence of the electron beam readout provides the user with a great deal of flexibility in the operation of the device. The remainder of this chapter contains material which is helpful but not essential to the understanding of the remainder of this thesis. C. Survey of Imaging Devices From the outline of the family tree of imaging devices (Figure 2-6) one can see that not only must a chemist decide if his problem is compatable with the general idea of imaging device detection, but he must also choose which device will be the Optimum choice for his application. Except for the charge transfer devices, all of the varieties listed have been applied as spectroscopic detectors. The following pages offer brief discussions Of these devices to facilitate evaluation of a specific device for a specific application. It is assumed that the reader is familiar with both the material in the first section of this chapter concerning the operation of the silicon vidicon and with the basic ideas in Chapter 1 concerning the use of .mmow>mo mcwmmsu mo mosh zpwsmd .o-~ ugamwu memwfim‘ zouom_ nmm mmmmm iouEEo —mmmmzixg m>~PUDQZOUOHOImfi m>Hmw~2m0hozm A—mOHuwmwuo— m T mkmn oz~w PHOTOCONDUCTOR SIGNAL PLATE Figure 2-8. Photoconductive Signal Production. 19 field and neutralize a portion of the surface charge deposited by the electron beam. The current required for the beam to restore the surface charge creates the photo-signal. These targets provide storage but no gain (Figure 2-8). The Silicon Electron Bombardment Induced Response tube will be seen tO employ a photoemissive transducer and a photoconductive target. Image Orthicon. The image orthicon tube was used in some early work (7-10) but has been largely superseded by the silicon vidicon and related tubes. The image orthicon uses a target Of glass or magnesium oxide with a target gain of 3-12. Rather than monitoring the current passing through the target to restore it to its initial state (as in vidicon tubes), the return beam (beam current in excess of that required to recharge the target) is monitored. As a result the signal-tO-noise ratio (S/N) Of the tube is quite sensitive to fluctuations in the beam current. The return beam passes into an electron multiplier with a gain of about 1000 (13). The image orthicon offers moderately high resolution, but very poor S/N, a limited dynamic range Of 10-60, and the lack of a unique gamma. The image orthicon is no longer used in spectroscopy (5,9-13). Image Isocon. The image isocon is a modification of the image orthicon in which only a portion of the return beam passes into the electron multiplier. The return beam consists of two parts: 1) electrons that are energetically unable to land and are reflected from the target, and 2) electrons that are scattered elastically from the target in numbers that are a function of the target potential at that point (14). The image isocon separates the two fractions and 20 amplifies only the scattered electrons. Since a major portion Of the beam noise is in the reflected part, the image isocon has a several fold S/N improvement over an image orthicon with the same photocathode and target (15,16). Only limited spectroscopic application has been made Of the image isocon (l8). Secondany_Electron Conduction (SEC) Tube. In the SEC tube (17,19) photoelectrons are accelerated into a target constructed of fibrous potassium chloride. Each incident electron causes release Of up to 100 secondary electrons which drain Off to ground through a conducting layer (Figure 2-9). The high target gain allows the video signal to be produced by monitoring the recharge current passing through the target rather than amplifying the return beam. The SEC tube has a very low dark current, allowing integration times of several hours. Vidicon. The term vidicon is Often used as a generic name for camera tubes with a photoconductive transducer. There are three types of vidicons in use today, differing only in the material used to form the photosensor and target. The antimony trisulfide vidicon (often simply called a vidicon) (20) has been used for spectroscopy (11,21-23) but has some serious shortcomings that preclude its future use in view Of the availability of the silicon vidicon and better tubes. The tube is limited to use between 400 and 600 nm, has low sensitivity and high dark current. The transducer has a gamma Of about 0.6 so it exhibits a quite nonlinear response to incident photon flux. The target can be damaged by 21 SUPPORTING LAYER (A1203) CONDUCTING LAYER (Al) .1 r\/;7C:"\_ FIBROUS kCT——L U25. ./3 ELL GB’T'JF . e- ELECTROI: BEAM > [/71 ACCELERATING GRID PHOTOEMITTER F / C) I 92/ g 9173: I L—aw T (D 4;: Figure 2-9. SEC Target and Signal Production. 22 viewing images which are too intense or by viewing the same image for an extended period (image persistence). Especially serious is the problem of lag which limits the faithful representation Of rapidly changing images. Lag is defined as the percentage of the signal remaining on the target three readout frames after the optical signal has been removed. It is as high as 30% in some szs3 vidicons (24). The problem Of lag becomes worse at low light levels. Cooling the tube reduces dark current but increases lag. Vidicons with a lead oxide target (25) have lower dark current, less lag, increased sensitivity and a gamma closer to unity. Lead oxide vidicons are marketed under the names Leddicon, Oxycon, Plumbicon, and Vistacon (26). The silicon vidicon discussed Table 2-1. Vidicon Target Comparison Target Sb253 PbO Si A region (nm) 400-600 400-600 350-1000 Sensitivity* (relative) 1 2 20 Dark current (relative) 7 l 2 Lag (z) 30 5 7 Gamma 0.6 0.95 l *On this scale a good photomultiplier would have a sensitivity of 10,000 to 100,000. 23 extensively in the first part Of this chapter is quite popular in analytical spectroscopy at present (5,6,27-35). It responds over a wide spectral range (generally 350-1,000 nm, but tubes with limited response down to 200 nm are available) with high sensitivity, low lag, low dark current, unity gamma, high resolution, freedom from image persistence or damage from intense sources, and a linear dynamic range of at least 104 (29,36). Some brand names include Epicon, ST-Vidicon, Sivicon and Tivicon (26). A comparison of some of the characteristics of the three vidicons is given in Table 2-1 (4). Silicon Electron Bombardment Induced Response (SiEBIR) Tube. The SiEBIR tube (37,38) uses the same silicon diode array target structure as the silicon vidicon, but precedes the target with a photoemissive transducer. Emitted photoelectrons are focused and accelerated by 10 kV into the silicon target. While an incident photon produces only one electron-hole pair in the silicon target, each 10 kV electron produces about 2000 holes. This high target gain has proved useful in low light applications (38). Due to the silicon target, this tube exhibits the same dynamic range, dark current, resolution, and lag as the silicon vidicon. The spectral response curve depends on the photocathode. As in the SEC and vidicon tubes, the output is taken from the target charging current. Trade names are Silicon Intensifier Target (SIT), Electron Bombardment Silicon (E85), and Secondary Electron Multiplication (SEM) (26). The name SIT is sometimes used as a generic term instead of SiEBIR. 24 Table 2-2. Comparison of Silicon Vidicon, SIT, and SEC Tubes. Tube Silicon SIT SEC Vidicon Dynamic Range 103-104 103-104 102 Internal Gain 1 2000 100 Relative Response (assuming a S-20 photocathode for the SEC and SIT) 250 nm 2 240 120 450 nm 8 400 200 650 nm 7 100 50 850 nm 4 2 1 Dark Current _ _ electrons s cm 20°C 101° 101° 103 Maximum Integration 4 Time, sec. 3 3 10 Relative Cost 2 10 7 25 The silicon vidicon, SIT and SEC tubes show great promise in analytical spectroscopy. Mitchell (5) has recently compared these tubes. Table 2-2 summarizes the important characteristics. In absorption work, the higher light levels permit use Of the silicon vidicon with a great reduction in cost. However, for atomic or molecular emission the higher gain Of the SIT or SEC would be desirable, though, at very long wavelengths the silicon vidicon has the sensitivity advantage. Since the tubes employ the same readout and external amplification, in low light situations where preamp noise is dominant, the SEC and SIT tubes always have a S/N advantage over the silicon vidicon (5). The extremely low dark current Of the SEC permits long integration times for enhancement of weak signals. The SEC however is not as robust as the SIT and more likely to be damaged by over exposure. The extreme high gain of the SIT target makes it comparable in sensitivity to a PMT (12). An SIT tube divided into 500 electronic channels can be used in the range of 101-105 photons/channel/sec (39). Image Dissector. The image dissector (40), a non-storage tube used in earlier studies (41-43), is recently receiving revived interest (44,45). The optical image is incident on a photocathode, and the emitted photoelectrons are accelerated toward an anode containing a small aperture. The two dimensional image is maintained by use Of an external magnetic field to focus the drifting electrons. Only those electrons that pass through the aperture are amplified in the multiplier chain. By suitable deflection Of the electrons in 26 [ FOCUS ] ANODE WITH APERTURE [ DEFLECTION I PHOTOCATHODE —-—~~ ~ — ~ ———~‘ ~.—-—- ‘0- ~_. -—_- VIDEO OUT l l DRIFT REGION 1 1 Figure 2-10. Image Dissector. 27 the drift section Of the tube, the signal from different areas Of the target can be scanned into the multiplier Chain (Figure 2-10). The instantaneous current through the aperture is proportional to the instantaneous irradiation on the corresponding area Of the photocathode. The tube Operates as a multichannel photomultiplier tube with similar spectral response, sensitivity, dark current and dynamic range. The dissector tube can count single photons (44,45). Since there is no storage target, the tube has no lag but cannot use integration to enhance weak signals or to detect pulsed events. The resolution is higher than most storage tubes and depends on the size Of the aperture. The aperture of the dissector functions exactly as the slits Of a scanning monochromator, isolating the spectral unit to be monitored, and determining resolution and sensitivity. A magnetic field scans the spectrum past the aperture just as a rotating grating or mirror causes a spectrum to scan past a conventional exit slit. Some feel that the dissector has greater scan versatility than a storage tube (42,44). Since the signal does not depend on the time between readouts, it can scan different areas at different rates and Operate at stationary deflection indefinitely. Image Intensifier An image intensifier is a device whose output is an Optical image which is brighter than the incident image (46,47). Photons incident on a photocathode release photoelectrons which are accelerated into a phosphor screen to produce a bright image. If the input and output images are in different spectral regions, the 28 device is called an image converter. The intensifier is not an imaging device but can be used to amplify a weak signal which is then detected with an imaging device (18,22). Solid State Devices The solid state devices employ a one or two dimensional array Of discrete silicon photodiode sensors contained on an integrated circuit chip. Due to the silicon diode detector used, they will have the same method and characteristics of signal production (spectral response, sensitivity, linear range, gamma, and dark current) as the silicon vidicon tube. The electron beam is replaced by some other method Of accessing the individual sensors (12,48). Rather than a continuous video signal, the output is a series of pulses of varying amplitude which occur as each sensor is sampled. 'The two basic types of solid state devices are the X-Y addressed array (Often referred to as a self-scanning photodiode array or simply a photo- diode array) and the charge transfer devices. In the photodiode array, the signal is monitored at the point it was created by sequentially switching the amplifier from one sensor to the next. In the charge transfer devices the charge pattern is clocked through the array in an analog shift register to the edge Of the sensor where the video signal is formed. The two types Of charge transfer devices (the Bucket Brigade Device or 880 and the Charge Coupled Device cn-CCO) differ in the method Of shifting the analog signal. Photodiode Array. A linear photodiode array may be visualized as in Figure 2-11. Photons striking the reverse biased photodiodes cause generation Of a charge pattern on the storage capacitors. 29 f—D— VIDEO OUTPUT q “w. “a 14""‘2. 1 .J Li J I l I 7 Give» F + 1P __ +V l l SCAN GENERATOR 0 <7 Figure 2-11. Linear Photodiode Array. 30 Sequentially, each diode is pulsed. When the diode becomes foreward biased, the capacitor discharges through the common video line to produce an output pulse (48). Horlick and Codding have recently done extensive work with the photodiode arrays (49-56). Arrays with spacings of 1-2 mils between diodes and containing up to 1024 diodes in a linear array or up to 50 x 50 diodes in a two dimensional array are available (57). Developmental models of at least 360 x 360 photodiodes have been constructed (48). The clock rate is typically 1 KHz - 10 MHz, but the maximum rate is only useful for very intense Signals (49,57). Bucket Brigade Device (BBD). The 880 (58-61) can be simulated out of discrete components. Each sensor element consists of two capacitors, two transistors and two photodiodes (Figure 2-12). While the device is sensing, both clocks are at the same potential, all Of the transistor switches are closed, and a photoinduced charge pattern develops on the capacitors. To read out this information, the clock lines are alternately pulsed. When clock line 1 is pulsed the charge on C1 flows into C2, C3 into C4, and C5 into C6, etc. Subsequently, pulsing clock line 2 causes flow from C2 into C3, C4 into C5, etc., and the net effect is to shift the charge pattern by one element. In solid state form, a 880 would appear more as in Figure 2-13. It can be seen that information is stored as majority carriers in the p regions and transferred as minority carriers in the n region. Bucket Brigade Devices are not yet commercially available. Charge Coupled Device (CCD). The CCD (59,60,62-65) has no discrete component analog. It is a junctionless device except for 31 SENSOR 1" ELEMENT ‘1 CLOCK l 1 T - {—— CLOCKZ hCI-t—‘cz EC he 1,—1. “~a‘v ‘“B- ' '“nt" ‘”V\ ' ‘PVE ' “‘~. ' +V -Iv~nt er Figure 2-12. Discrete Component Bucket Brigade Device. SENSOR l" ELEMENT ’1 j T 7CLOCks GATE ELECTRODE ’/_..Sio2 INSULATOR L"-’-’-‘-p REGION “‘-n REGION Figure 2-13. Bucket Brigade Device. 6 32 a small p-n junction at the output. The photo-produced charges are stored in potential wells at the surface of a doped semiconductor, and moved across the surface by translating the potential wells. The potential wells are created by applying a voltage to an electrode on the surface Of an insulator covering the semiconductor and driving the surface of the semiconductor into depletion. Figure 2-14 illustrates the construction and operation Of a CCD with a three phase clock. It can be seen that in the CCD, information is both stored and transferred in the form of minority carriers. Commercially available CCD's have sensor spacing of 1-2 mils and clock rates up to 6 MHz. They are available with up to 500 elements in a linear array or up to 320 x 512 elements in a two dimensional array (66,67). To date, no spectroscopic applications have been reported with the CCD. Device Comparisons Neither the self-scanning (X-Y) photodiode arrays nor the charge transfer devices are clearly superior at this point. Due to switching transients in addressing the separate sensors, the self-scanning arrays have a significant level of fixed pattern dark current. The charge transfer device has a lower, more uniform dark current. In the photodiode array, a bad sensor causes loss of information only at that one point, but on a charge transfer device a bad sensor will also effect any signal shifted through it from other sensors. Since the photosensors are always active, continued excitation during charge transfer scanning will result in image 33 SENSOR 1" ELEMENT ’( I I I f_>CLOCKS ELECTRODE 1...“, u ”4,. (INSULATOR ‘ 1 ‘7“ ' --—-SEMICONDUCTOR 19”“, 1g (POTENTIAL WELL "—1.9;- "‘""""l' e- """m‘lwg: 1 —---~1 Lee! Lee. I I’" "l .___:-_‘ rm- LL: ’ {Ci—9.1 19.3.} 129-91 Figure 2-14. Charge Coupled Device. 34 COL—£3215 4:: 1:1 :3 Figure 2-15. Charge Transfer Scanni ng. 35 smear. TO overcome this problem, charge transfer devices have a second charge transfer readout area (shielded from light) connected to the charge transfer image area (exposed to light). Periodically the signals in the image area are gated into the readout area and the signal is then shifted out (Figure 2-15). For analytical spectrOSCOpy there is no unanimous favorite among the various solid state and tube devices. The solid state devices Offer long life, small size, durability, low power consumption, and simplified circuitry since there is no requirement to create or deflect an electron beam. However, they are limited by a fixed sequence of scanning. The ability to control the target scan sequence with the tube's electron beam provides great flexibility by permitting read out of only selected areas (wavelengths) or to sample different areas at different rates. The sensitive SEC, SIT, and dissector tubes eclipse solid state sensitivity. Until technology can produce a device combining the simplicity Of the solid state device with the sensitivity and scanning flexibility Of tube devices, the choice of an imaging device for spectroscopic application will involve some compromise. Very late in the proof stages of completion Of this dissertation, a pair of excellent articles appeared which surveyed the field of imaging devices applicable to spectroscopy (149,150). 0. Historical Applications of Imaging Devices in Spectroscopy The method of rapid-scan multichannel spectroscopic detection using imaging device detectors has recently aroused considerable 36 interest. However, the origin Of the technique can be traced back over thirty years to a patent granted in 1941 to H.A. Snow for a device he called a color analyzer (68). The device was capable of recording 16 spectra per second (using an early TV tube called an iconoscope), and displaying the spectra on an oscilloscope. Snow's invention failed due to poor image tube technology (11) and probably because the idea was a bit before its time. Figure 2-16 is a cumulative plot Of the number Of papers concerning application Of imaging devices in spectroscopy since Snow's patent. The idea remained fairly dormant until the sudden explosion Of interest in the early 70's. Work prior to the 70's was mainly from physicists, engineers, and astronomers (non-spectroscopic astronomical applications are not included in this discussion). In the early 70's several commercial spectroscopic systems employing imaging device array detectors became available. Since 1972, work in the area has been dominated by analytical chemists who have evaluated and accepted imaging device array detectors as useful new tools for multichannel and rapid-scan spectroscopic analysis. Some early work included the study Of emission spectra from heptane-oxygen explosions (7,8), reflections from xenon shock waves (69), laser Raman spectra Of azobenzene, toluene and bromine (74), hyper-Raman spectra Of ammonium chloride and water vapor (75), and the excited-state absorption spectrum Of chromium ions (10). Lightner (11) developed a general purpose visible spectrophotometer employing complex compensating circuitry to correct for nonlinear response. Harber and Sonnek (41) and Baker and Steed (42) employed image dissectors for preliminary studies and commented on their scan 37 60-4 W 35 o. < n. o E on 2 3 2 ‘$’ :2 20- < .J :3 Z :3 u _r 1940 1950 1960 1970 YEAR Figure 2-16. Publications Concerning Imaging Devices in Spectroscopy. 38 flexibility and sensitivity over storage tubes. An application Of image dissector tubes to analytical emission spectrometry was made by Fassel (43), but the tube was not used as a multichannel detector. Instead each exit slit-photomultiplier tube combination was replaced by a separate dissector tube. This reduced the necessity for critical slit alignment and facilitated rapid reprogramming Of monitored wavelengths. In order to be able to cover a wide spectral region and maintain satisfactory resolution, Anderson (9) divided a spectrum into three overlapping sections (380-520, 480-660, 630-820 nm) and stacked them in strips on the detector face. This idea has been developed further by Margoshes (70-72), Danielsson (44,45,73), and Wood et al. (89). Each Of these later systems uses an echelle to separate the region between approximately 200 and 850 nm into up to 120 orders. The systems have computer controlled readout of intensity at selected wavelengths accurate to a few hundredths of an Angstrom. This technique has been extended by Grotch et al. for use as a detector for mass spectrometry (90). Ions impinge on a microchannel electron multiplier coupled to a phosphor screen. Fiber Optics take the signal from the long, narrow phosphor, divide it into four sections stacked on top Of each other, and transport it to the vidicon detector. Beaver and McIlwain (76), and Boumans et al. (77,78) studied small arrays of photodiodes. They noted that the dark current could differ by as much as an order Of magnitude between sensors and that the S/N was 100-500 times poorer than for a photomultiplier tube. 39 The recent surge Of interest in array detectors was precipitated by the availability Of commercial systems based on the silicon vidicon (79-85) and by Pardue's preliminary studies (23,86) to evaluate a spectrophotometer containing an antimony trisulfide vidicon. Since then, analytical applications have been in the areas of multichannel atomic absorption, atomic emission, molecular absorption, and rapid-scan monitoring of kinetic studies. Horlick used flame emission studies on K, Rb, and Na in early characterization of photodiode arrays (49,53). The arrays were not sensitive enough to detect Ca, Cr, or Ni emission. By multiplying the array output signal by an appropriate gate function he Obtained smoothed, sharpened, or differentiated spectra. Pardue (32) and Morrison (38) have done simultaneous determinations of electrolytes in serum. Morrison combined first and second order diffraction to Obtain analytical lines within the detector's wavelength window, and used filters to partially absorb certain lines to equalize intensities. Simultaneous emission analysis Of up to eight metals has been performed (5,28,29,3l,52), but the S/N for the array detector is an order Of magnitude lower than for single channel detection (5). Winefordner compared single channel and multichannel detection limits for eleven metals (31). Morrison has used spectral stripping (33) to eliminate interferences due to overlap Of analytical lines with flame bands or bands and lines from other constituents. Horlick has identified DC arc spectra from 35 metals using logic Operations on reduced spectra (51) and in a recent study has used the photodiode array to 4O perform time studies on the DC arc spectra Of mixtures Of Cu and Ge in various matrices (55). Simultaneous AA analysis on up to four elements (27,30,56,87,88) has shown that multichannel array detectors are suitable for AA. Detection limits are 7-10 fold higher than for single channel detection (27,30). Multichannel detection allows summation of the signal from several lines of the same element to increase sensitivity (56). For AA, use of a non-storage image dissector may be advantageous since it would allow signal modulation to eliminate interference from flame emission (30). Array detection has been used to determine phenols in the .1-17 ppm range by using difference spectra of the acid and basic forms (34). Using a photodiode array to study laser intra-cavity enhanced absorption Of rare earths allowed collection of a complete spectrum from one to four laser pulses (54). Reflectance and emission power spectra from colored surfaces and phosphors have been measured by using a silicon vidicon (91). Pardue et al. have used a computer interfaced vidicon spectro- photometer tO study the kinetics Of several reactions (23,29,34), collecting spectra as Often as every four milliseconds. The power of the multichannel rapid-scan technique enabled them to detect intermediates in the reactions. They have developed a simultaneous kinetic method for the determination of two enzymes, based on absorption changes at two wavelengths(35). By tracing its history we have seen the image device array detector develop into a useful, versatile spectroscopic detector. 41 Its future is probably limited only by the ingenuity of chemists to develop new applications subject to performance limitations such as described in Chapter 4 for the silicon vidicon. CHAPTER 3 ROUTINE OPERATION OF THE INSTRUMENT The purpose of this chapter is to describe the operation of the computer-controlled vidicon spectrometer as a general purpose, single- beam scanning spectrophotometer. The software and teletype inter- actions for this general purpose Operation will be discussed at length as they form the framework for all other data acquisition, analysis, and instrument characterization software. Unless otherwise noted, all software discussed in this thesis was written in FORTRAN II and SABR (assembly language) to be run on a PDP-8/I with 12K words of memory and DECtape. All programs relevant to this chapter are listed in Appendix C. The basic instrument outline is shown in Figure 3-1. A brief discussion Of only those points necessary for routine operation will be presented here. More extensive discussion is reserved for Chapter 8. The vidicon detector sits in the focal plane Of a modified monochromator such that the beam slow-scan axis is parallel to the wavelength axis Of the dispersed spectrum. An image focused at the entrance slit is dispersed and focused on the target. With the grating and Optics used in these studies, the vidicon can view a wavelength window Of 230 nm in the region Of 380-900 nm with about 4.2 nm resolution. The top half Of Figure 3-2 illustrates the image on the vidicon detector from a line source at the entrance slit. The two dimensional detector is converted into a one dimensional detector 42 43 .Emcmmpo Emamam .~-m mcamwa uz~4a2 1 mmDnhm amaz: thw>m 4_ E -" QUANTUM EFFICIENCY t: 23 ‘ > H H U. —I F— u. H LIJ (I) Z z _ O D a. F- U‘) Z “J < O: 3 CT 10 I 1 l I 400 600 800 1000 WAVELENGTH - NANOMETERS Figure 4-1. Gross Spectral Response Curve. 2000-4 6? 1: 5 1500.. SE < Ci :3 g 1000- :5 E: 5; 500.4 53 *— 55 C l r l l 100 200 300 400 CHANNEL Figure 4-2. Channel Sensitivity Curve. 58 AmAHza >mm< chmwm og mzu pcmsmucmscm 2\m g>ND and NT==NS. If the number of sample scans averaged equals the number of dark current scans averaged, then ND==NS and NT==NS/2. Signal enhancement through charge integration can also be used to increase the S/N Of a spectrum. In this case the signal increases lihearly with the amount of integration time (until nonlinearity occurs near saturation), while the rms background noise remains constant, increasing only at very long exposure times (Figure 4-14). So, the S/N increases linearly with the amount of time spent Observing the signal rather than with the square root Of the time as with averaging. Figure 4-15 illustrates three spectra in which the same total amount Of exposure time was spent in data acquisition, but the relative times spent on signal averaging and charge integration were varied. Going from the top spectrum to the bottom a l6-fold larger fraction Of the time was 77 .aoweacmaccH amcacu EAL: amLoz mzm .a_-a seamen uz LIJ U) E 5- O. «.0 Lu c: 4. T I I T T r I I 32 64 128 256 512 1024 2048 4096 NUMBER OF CHANNELS Figure 4-18. Response Versus Number Of Wavelength Channels. 84 the center of the beam travels down the center of one channel, the edges Of the beam do not scan any of the adjacent channels. Along the plateau, as the beam center travels down a channel center, the beam edge overlaps a portion Of the next adjacent channel. When that next channel is scanned, the effective area is reduced since some of the channel has already been sampled. Across this plateau region the reduction in effective area and the increase in exposure time balance each other. In the descending region Of the graph, the wavelength channels are so close together that the beam completely overlaps more than one channel. As a result the signal drops rapidly. The intersection Of the rising and plateau sections gives the effective width of the beam as about 1/192 Of the target width. Since this value is less than that determined with the response across the tape assuming a rectangular beam profile, it is obvious that the beam profile is shaped more like a peak than a rectangle. Most of the "reading ability" is concentrated about the center, but the beam wings extend to the sides. G. Software Programs VCIDCO and VSIGAV (described in Chapter 3 and listed in Appendix C) were used in performance evaluation. In addition, two other programs, VDARK5 and VDRDC, were developed and are listed in Appendix 0. Large parts of the programs are identical to VCIDCO, so only the sections which differ have been listed with notations to indicate deletion of common material. VDARK5 was used to generate data concerning the S/N as a function of signal averaging for the experiments discussed in the sections on 85 dark current and on noise. VDRDC was written to expand the dynamic range of the vidicon spectrometer by using selective integration. The first half Of this program (normal scanning of the spectrum) is identical to VCIDCO. In the second half the "scanning map" is constructed and the selective scanning begins. The output routine simply scales and displays the spectrum. CHAPTER 5 EVALUATION OF LEAST SQUARES SMOOTHING OF DIGITALLY RECORDED DATA A. Introduction to the Smoothing Process Any real signal consists of the desired signal which is related to the information of interest, and an undesired component called noise which hinders ready extraction Of the desired information. The signal-tO-noise ratio (S/N) is a useful way to describe the quality Of a measurement process. As shown in Chapter 4, performance Of the vidicon spectrometer suffers from a restricted S/N. Chapter 4 described studies applying digital signal averaging and analog signal integration to enhance the S/N. Signal averaging and integration can prove quite effective in reducing the noise component at the expense Of the time required to collect redundant data to be averaged into a single data point. In situations requiring the vidicon spectrometer to produce Spectra at its maximum rate, no redundant data can be taken and averaging or integration for S/N enhancement is not possible. If a particular experiment requires data acquisition at a rate which limits the extent to which signal averaging or integration may be used, then further S/N enhancement must be done using numerical procedures applied after data acquisition is complete. Least squares polynomial smoothing is widely used for this after-the-fact S/N enhancement. In this method a polynomial Of degree m is fitted through 2n+1 adjacent points (ngm) in the raw data using the least squares 86 87 method. The equation Of the polynomial is used to calculate a new value for the central point. This procedure is repeated for each group of 2n+1 points so that each point in the raw data (except for n points at each end of the data array) is replaced by a new value. Savitzky and Golay have described an equivalent method to obtain smoothed data which is easier to perform than the fitting procedure (93). They derived tables Of weighting coefficients which are convoluted with the raw data to yield smooth values. Their method assumes that the data points are evenly spaced on the abcissa axis, the data curves are continuous, and the uncertainty in the ordinate values is much greater than the uncer- tainty in the abcissa values. Least squares smoothing makes no assumption concerning the probability distribution Of the uncertainties, but requires the distribution to have a mean of zero (94). The least squares smoothing process uses local portions Of the data to determine a "better"representation of the central point Of each local portion. The magnitude of noise present on the signal at one point is independent of the magnitude of the noise at adjacent points. Consequently, smoothing of noise will yield a weighted average of random values (with mean zero) and, as a result, will reduce fluctuations and the effective bandwidth of the original signal. However, the value Of the true signal at a point is related to the signal at adjacent points. In this case, the smoothing function will yield new data values which are quite close to the original values. Information carried by the signal, within the remaining bandwidth, will be retained. Since the smoothing function will only be an approximate representation Of each 88 local section of data, the true signal will suffer some distortion which will depend on the details of the smoothing process and on the properties of the data being smoothed. Improvements in the S/N of a measurement occur at the expense of some other property Of the data. Signal averaging results in loss of time resolution but does not cause signal distortion. Loss due to averaging is decided at the time of data acquisition. Smoothing results in bandwidth reduction and signal distortion due to loss in resolution along the abcissa (wavelength axis of a spectrum). However the type and degree of compromise involved depend on parameters Of the smoothing process and may be decided after data acquisition is completed. The smoothing parameters which may be adjusted include the type Of smoothing function (quadratic-cubic, quartic-quintic, derivative, etc.), the width of the smoothing interval (the number of points in the smoothing function), and the number of repetitions Of the smoothing process. Decisions concerning these parameters are Often based on intuition rather than any firm guidelines. Since Savitzky and Golay published their tables of smoothing weights there have been several studies concerned with Optimum procedures in least squares smoothing (95-107). However, none Of these seemed to be sufficiently thorough nor to yfield guidelines in a conveniently useful form. Therefore, I decided to study the effects Of least squares smoothing of digitally recorded data in order to determine the resulting compromises between noise reduction and signal distortion. Peak-shaped signals (Gaussian and Lorentzian) were used in this study. This type of signal was of immediate interest to the vidicon project to provide 89 guidelines for smoothing Of spectra. These signal shapes were also of more general interest due to their widespread occurrence in signals from many types Of data sources (spectroscopy, chromatography, electrochemistry). The type of smoothing function used in all of the following studies was a quadratic-cubic smooth. The parameters which were varied were the width of the smoothing function (5 to 23 points) and the number of times the smoothing process was repeated (up to 1024 times) on the same data set. Higher order smoothing functions were not studied due to the limited word length (12 bits) of the POP 8/I minicomputer used in most Of these studies. 'A larger computer was not employed since the results from this study were to be used to decide Optimum smoothing techniques for data acquired by a minicomputer from on-line experiments. 8. Noise Reduction In any real situation both signal and noise would exist together and be smoothed together. However, to facilitate the study it was decided to separately generate noise free signals and random noise and to smooth each separately. Since the smoothing method studied is linear, this approach will yield the same results as if signal and noise were added together before smoothing. For the study performed it was assumed that the noise component of the system was white noise. White noise is a mixture of signals of all frequencies with random amplitudes and phase (109). The probability distribution of the ampli- tudes is normal (Gaussian) if the Observed noise is the sum of several independent noise sources (110). White noise can be simulated by a 90 random sequence Of numbers with a normal distribution Of magnitudes. Noise Of this type was generated by two methods. (See Appendix A.) Both began by using a congruential method to generate a sequence Of random numbers with a uniform distribution Of magnitudes. In one case the Central Limit Theorem was used to convert this sequence of numbers to a normal distribution. In the other case a look-up table derived from a standard normal table was used for the conversion. Figure 5-1 illustrates a sequence of 200 normally distributed numbers generated by each method. For comparison a sequence Of 200 uniformly distributed random numbers is included. All three sequences have the same standard deviation. To demonstrate that the calculated values do in fact approximate the desired distribution, a histogram was plotted for each white noise generator and compared to the theoretical frequency (Figure 5-2). In both cases the theoretical mean was 1000 and the theoretical standard deviation was 300. Observed values were 1001.5 and 301.3 using the Central Limit Theorem and 994.7 and 292.9 using the look-up table. The chi-square goodness Of fit test was used to test the distributions. (See Appendix A.) It indicated that each of the white noise generators produced values which can be described by a normal distribution with 95% confidence. In a simulation study by Rogers, et al. (108), synthetic noise with a uniform distribution was added to pure signals prior to numerical filtering. Since experimentally Observed noise has a normal, rather than uniform, distribution, I wanted to see if the result Of least squares smoothing depends on the form of the distribution of the noise .: ‘- °...° -° :’° ' NORMAL BY 0 0 . 'I .. . . . . . o . o . o . “.0 .. CENTRAL LIMIT 0. .0. . 0° 0 o .0 .0 O . .0 ° ' THEOREM . ... o . o .. 0.. g... . 0 . o . . .0 .0 . . . o O .. . . ’ . . . o . o. . . . o .. . ' ° ' NORMAL BY 0 . ..o. .’ o O. .0 . 1 LOOK-UP ..‘ °. '-'. '. , - ' TABLE 0 0.. ‘ . o .. . . . . . . coo . .0 I .. . co .0 2 6 . : ...00 :g . o o .0 ° C 0 O . .0 O. .. . 0.. ‘ o0. UNIFORM o . .0 . .0 ‘0 . o. . o ”.0 .0. .0 . . Figure 5-1. Noise Output from Random Number Generators. 92 loo-- {0 35 75- NORMAL BY g CENTRAL LIMIT THEOREM U 23 50- u. 0 85 g3 25- 3 z .3 , 400 1000 1600 MAGNITUDE 100- 0"} E? 75-4 a: a: 2 :3 c: 50-. LI... 0 £3 £5 25-1 2 0.J T l I l l 400 1000 1600 MAGNITUDE Figure 5-2. Histograms Of Noise Generator Outputs. 93 in the signal. The uniform noise generator and each of the white noise generators described earlier was used to generate an array of 1000 noise values. In each case the standard deviation was 300. The degree of noise reduction for the smoothing process was measured as the ratio of the standard deviation after (0) and before (00) a single pass of the smoothing function. For both normal and uniform noise, the ratio 0/00 was measured for smooths from 5 to 23 points wide. This study was performed using a SABR integer smoothing routine on the PDP 8/I and using FORTRAN real value smoothing routines on the PDP 8/I and PDP 11/40 minicomputers. The results appear in Table 5-1. On a log-log plot of 0/00 versus width of the smoothing function, the points fall along one of two straight lines, depending upon whether the noise had a normal or uniform distribution (Figure 5-3). Least squares fitting was used to estimate the slope, intercept, and standard deviation of the slope and intercept. The normal line has a slope of -0.496 1 0.009 and an intercept of 0.179 1 0.010 (antilog of the intercept is 1.51). The uniform line has a slope of -O.603 :_0.004 and an intercept of 0.227 t 0.005 (antilog of the intercept is 1.87). Savitzky and Golay indicated that the amount of noise decreases in proportion to the square root of the number of points in the smoothing function (93). A t-test was used to determine whether or not the slopes of the fitted lines were significantly different from -0.5 . (See Appendix A.) At the 95% confidence level the slope of the normal line is not significantly different from -0.5 while the slope of the uniform line is different. Therefore, for normally distributed noise, the amount of noise remaining after a single pass smooth is inversely proportional 94 NNo. mvoo. NNo. o_o. opo. mmo. Fmoo. meoo. cowpmw>wa vcmucmum poFQ mpm.- ~m¢.- Fom.1 Nn¢.1 em¢.- emo.u mmm.- moo.- mopimop mo maopm muom. mFFm. mppm. oemm. ommm. Fenm. mmmm. wmnw. mm mwmm. even. upmm. oxen. mmwm. ouom. Fm mme. «mom. nmmm. Pmom. Room. Numm. m_ mmkm. Pfimm. comm. omnm. muom. nmmm. opmm. _~mm. up Ream. mmpe. anm. vomm. mumm. mmmm. nmkm. «com. m_ mNFe. moee. o_o¢. meme. mm—e. ummm. ”upe. swam. mp mume. meme. mnmv. Name. mmve. Amme. emmc. m_ee. FF upom. mmmm. mmme. meom. upme. Pmme. mopm. name. a «com. comm. ommm. mvnm. mpom. omum. Pmmm. mmnm. A memo. mmon. muse. emmm. memo. poem. moo“. mpom. m oe\pp H\m o¢\pp H\m H\m oe\~F H\m H\m zmm vcmecmum Acacacc_o EAL; asaaae aaaaz ca e=_35665m Gesaa m amaa a_a_ppaz onhuzam czmzpoozm mo mmmmm2u >m mwmzzz “N 3 N 1m; m. a W 3 E m m ” UV 0 “a ulo.m .xama mum: acwoa m_ a to» “cosouemscm 2\m .¢_-m weaned onhuznm wznzhoo2m mo :hon 116 Pm mp up mF mp Pp m m m ,. p _ _ _ . . _ _ monpHcmawm do ammzzz mmem=u >m mumzaz F ..m F N .5 / N v 3 .uo.N mm V w W 3 N mp W... lute-N In. 4. NM WW II. 0 #0 H. l®.m 117 .xmma wu_3 “econ mm a Lac pemswucaccm Z\m .m~-m wcsm_u zoFFuzsu wzFIFOOZm no IFan mm FN mF FF mF mF FF _ F F F F _ _ I—nm t'rF -LD mzoFFFFmamm mo «unsaz mmFmzo >m xmmzsz UOlOVd 1N3N33NVHN3 N/S 118 is a different limit to the degree of S/N enhancement attainable through smoothing. The limit of S/N enhancement increases as the FWHM of the peak increases. For an 8 point wide peak the limit is 2.4; for a 16 point wide peak the limit is 3.1; while for a 32 point wide peak the limit is 4.0. For each width of peak there are several combinations of smoothing width and number of repetitions which yield the maximum S/N enhancement. For example, with a 16 point wide peak, the combinations listed in Table 5-11 yield nearly the maximum enhancement. Each combina- tion listed results in approximately the same degree of peak height and FWHM distortion, falling in the range of 10.4% to 12.9%. Table 5-11. Maximum S/N Enhancement for a 16 Point Wide Peak. Width of Smooth 11 13 15 17 23 Repetitions 64 32 16 8 3 (S/N)/(S/N)O 3.064 3.072 3.074 3.066 3.071 Height Reduction (%) 12.9 12.8 11.9 10.4 11.4 FWHM Increase (%) 11.8 11.8 11.8 11.8 11.8 When the smoothing parameters are chosen such that S/N is Inaximized, there is an obvious degree of distortion introduced into the peak. The user must decide if for his purposes the loss of accuracy is tolerable. If the purpose for smoothing is to "clean up" noisy data prior to display for human interpretation (but not measurement), then S/N enhancement more than compensates for signal distortion. Sinfilarly, if the analysis involves only relative measurements against standards that have been treated identically, then signal distortion 119 R AW DATA PEAK 8 POINTS WIDE prx/N/W\#fifV/WLA\VWPUMAMKJNJK: POINT SMOOTH U/J\JJ/\\/~rJ/\V\\~/\/p/\V\/:5 POINT SMOOTH //\Jj/p\L/~j/\\A\Jf«//’\\VI23 POINT SMOOTH Figure 5-16. Example of Smoothing to Maximize S/N Enhancement. 120 factors will cancel (provided the lines are equal in width) and the increased S/N will yield a higher precision measurement. If absolute measurements are required or if peak broadening causes resolution to decrease below an acceptable level, then it is not desirable to smooth to the maximum S/N. In this case a better approach would be to use peak distortion and not S/N as the criterion for judging the extent of smoothing. Use of S/N as a criterion to judge smoothing is illustrated in Figure 5-16. The original spectrum is shown at the top and is a pair of lines out of the spectrum of a mercury lamp as recorded by the vidicon spectrometer. The source was attenuated to produce a signal of low S/N. The lower three curves are the results after single passes of 5, 15, and 23 point smoothing functions. The original lines were sampled by eight points at their FWHM. The data after a 5 point smooth has suffered negligible distortion of the peaks, but still contains significant noise. A 23 point smooth produces a signal considerably free of noise but at the expense of significant peak distortion. The intermediate case, using a 15 point smooth, is seen to produce the desired compromise between noise reduction and peak distortion. Reference back to Figure 5-13 shows that for single pass smoothing of an 8 point wide peak, a 15 to 17 point smooth produces the maximum S/N enhancement. E. Frequency Analysis In order to broaden the perspective of conclusions reached concerning the least squares smoothing of peak-shaped signals, an analysis was made 121 of the frequency components of the smoothing function and the signal to be smoothed. The frequency spectra for a hypothetical signal and for band—limited white noise are shown in Figure 5-17. To reduce noise and yet pass the signal, a smoothing function should have its high frequency cutoff between those of the signal and noise. As the high frequency cutoff of the filter moves to lower frequencies, more noise is eliminated but greater signal distortion occurs. PrOper placement of the filter cutoff can be made in terms of the empirical guidelines developed concerning signal distortion and S/N enhancement. It is desirable to determine how close the filter response can approach the signal without harmful distortion. The proximity of the two frequency curves is related to the ratio in the time domain of the width of the smoothing function and the width of the peak (the smoothing ratio). The frequency spectra for several smoothing functions and Gaussian- shaped peaks were calculated using a Fast Fourier Transform (FFT) program. Figure 5-18 illustrates the frequency response for Gaussian peaks that are 8, l6, and 32 points wide. Each case is paired with the frequency response of the smoothing function that limits peak height error to 0.5%. In the time domain, the ratio of (smoothing width)/(peak width) falls in the range of 0.6 to 0.7. A similar observation can be made in the frequency domain by comparing the frequencies at which the response drops to half of maximum (f1/2). Where peak height error is limited to 0.5% the value of f1/2 for the smoothing function is 3 1/2 times the value of f”2 for the signal. By similar analysis it can be seen that when peak height error is limited to 1.0%, the values of f differ by 1/2 about 1.5. RESPONSE 122 ‘\\ \\ NOISE Q—Tfi- \ SIGNAL more less noise signal reduction distortion \ \ SMOOTHING FUNCTION FREQUENCY Figure 5-17. Frequency Relations of Signals, Noise, and Smoothing Functions. 123 PEAK WIDTH (——) 8 POINTS SMOOTH (- -) 5 POINTS 1 PASS 1 Response FREQUENCY PEAK WIDTH (-—) 16 POINTS SMOOTH (- -) 11 POINTS 1 PASS PEAK WIDTH (——) 32 POINTS SMOOTH (- -) 21 POINTS 1 PASS Figure 5-18. Frequency Analysis of Minimum (0.5%) Peak Height Distortion due to Smoothing. RESPONSE 124 PEAK WIDTH (——) 8 POINTS SMOOTH (- -) 11 POINTS 4 PASSES FREDUENCY PEAK WIDTH (——) 16 POINTS SMOOTH (- -) 11 POINTS 64 PASSES PEAK WIDTH (-—) 32 PASSES SMOOTH (- -) 23 POINTS 64 PASSES Figure 5-19. Frequency Analysis for Maximum S/N Enhancement due to Smoothing. 125 Results for maximizing the S/N can also be correlated with frequency response. Figure 5-19 illustrates the frequency response for Gaussian peaks of 8, 16, and 32 points wide along with the frequency response of a smoothing width and repetition combination which produces nearly maximal S/N enhancement. (These combinations are taken from data in Figures 5-13 to 5-15.) As can be seen, the least squares filter which provides maximum S/N enhancement is the filter whose frequency response most nearly matches the frequency characteristics of the signal to be smoothed. It is important to remember that the matched filter still distorts the signal since the output frequency spectrum is equal to the product of the frequency spectra of the input signal and the filter. F. Software The software which was developed for this study of least squares smoothing is outlined in Table 5-12. Listings of representative programs are given in Appendix E. Mathematical and statistical concepts behind some of the programs are explained in Appendix A. Unless other- wise specified, all of the programs were written in FORTRAN II and SABR (assembly language) to run on a PDP 8/1 (with EAE) under 05/8. G. Conclusions Having completed a study of least squares polynomial smoothing as a means to increase the S/N of spectra produced by the vidicon ‘spec- trometer, we can conclude that smoothing is indeed a useful technique (within the constraint of signal distortion) to be used in conjunction with, or in lieu of digital signal averaging or analog signal integration. 126 .mu\F new .FFV uFumFuaam emu» .quv socmmgy mo mmmcmmu muaauao .mumc on» soc» umuoswumm so uaFFFumam we age cowuanFEumFu Fugue: asp coF mgmuasmgua .cowuaawgumwu _asto= to stoc_== ”sesame “agapuua to HH z<¢p¢oa .mamp gonna .quumqu sogmv aunt Faucmswsonxm mpmoF ummu HFF Fe mmocuoom magnum-Fgu amaze .oe\FF new H\m an; 8:» cues co :3; mm: mcFuaog mng .z 6:» mo osmm mcwuaognam :Fguoosm :Fozm .ega: seven An capuwcz .macwoa mu 8:” .~_.m_.mF.P_.m.~.m mmnm ownauuqumgumaa msgomgma wcwuaognam mchuoosm IFOZm .cgmz cmmgm ;u_z ancpow cmuuFL: .gmnssc can .umm: owe mp wsFu :oFumgmcou .o mmFFFumam gum: .cmms och gsz Faggoc mmcou o» «Fang n=-xooF Fugue: ugmucmam AcowuznwgumFu Fasgocv a mom: can mg~g23= wouaapgumFu xFeLoFFca moungmcmu ocFuzognam goumgmcmm amass: soucum oz<¢ .Lmnssc gun .uoms me mF we.» :oFuogmcmu HF z<¢Fmom .8 new : mmFFFumqm com: AcoFuaanmeu Fmsgocv .emcom;F HFEFJ Fagucmu mom: can :z<¢~ mFqu mcpuzogasm goumgwcmm conga: soucmm uzm com ogmzumom .NF-m anmF 127 .mcwgaoosm 0» Loan mmaFn> Fncwmwco ssz moFunc can .nmgn ncn .223; .usmFm; xnmn no mnaFn> .Eastns xnmq mg» Fe :oFunqu mannpzo ncn waoum ananFn co mxnma mqun Engmogn .mgaoosm Fo Lassa: oz»-Fo-meoa gunm goum< FF z<¢Fmou .mmmmna we Lassa: anou ncn :uooEm mo spew: manFuonm cm»: mxnma no mcpguooEm mnFnapm NFm= Fnuwumgomgp n can mnoFm unanFumm cmmzumn pump-» n anon .mounsFumn mg» we mcopunw>nn ncnncnpm ncn unmugmucF .maoFm mouneFumm .ngnonanx soc» nonmucn mucwoa >H zg=u mmgnnam umnon mean .oo\o van 8 gmucFLa mch on mannuzo nan mnoom anaan co anggn mmFo: mqua Engmogn .mguooEm Fo amass: Nuyoigmzoa zunm gmuw< HF z<¢Fmom .mnmmnn Fo snags: Fnuop ncn gnoosm no cunpz mmFFFumnm gwms mmFo: mo mchpoosm mopnapm m»oz<¢ .mFoz<¢ op mcpnzu FF zFuumammg szguuonm xmgncn gowgzom any we moF mgp can Fmpcna agncanEF ncn ang Fa Sam many Ensuunam sagmcn nggaom as» mnFawFo mason ncn mason .annuuma so maxqunp song mucwon noan> ang mNF on a: yo snowmcnnu an» yo mgmzcn nan Fm; nmaFn> ang as» manann can mmuaasou FF z<¢F¢Ou Fmev ELOchnLF LmFgaom umnm canon .nozcwpcoo .mFum annF 129 All three of these techniques enhance S/N at the expense of some facet of the operation of the vidicon spectrometer. Analog charge integration increases the time required to acquire a spectrum. Until saturation, the S/N increases with the integration time. This method for S/N enhancement is unique to integrating detectors. Digital signal averaging requires collection of more than the theoretical minimum number of data points in order to define the measurement at each wavelength. This results in a corresponding increase in total elapsed time. Signal averaging increases S/N in proportion to the square root of the number of averages. Averaging filters out signals of all frequencies that are not in phase with the sampling, while smoothing removes only those signal components at higher frequencies than the cutoff of the smoothing function. In least squares smoothing, variations in adjacent data points (rather than variations in repeated measurements of the same point) are averaged together. As a result, the operation of smoothing improves as the spacing between adjacent points decreases, but conse- quently requires more points to complete a spectrum. For normally distrib- uted white noise, the rms value of the noise is reduced in proportion to the square root of the number of points in the smoothing function, and approximately in proportion to the eighth root of the number of times the smooth is repeated. Digital signal averaging and analog signal integration result in loss of time resolution to a degree which is determined at the time of data acquisition. Redundant data are lost but information carried in the wavelength axis is faithfully preserved. Least squares smoothing results in loss of wavelength resolution due to peak distortion. 130 Although this distortion is undesirable, the degree of the loss is not set at acquisition time, and can be altered as many times as the spectrum is processed. The factors which determine the degree of peak distortion include the number of passes of the smoothing function and the ratio of the width of the smooth to the width of the peak. If the degree of distortion is known, the smoothed data can be corrected. In theory, correction can be made for any degree of distor- tion. However, as spectrum complexity increases (especially if all peaks are not the same width), the corrections can become quite involved. In this case, the best alternative is to keep the degree of distortion within acceptable limits. Error in peak height can be kept to less than one percent if the smoothing function is narrower than 80% of the width of the peak, and only one smoothing pass is made. If the information contained in the data is to be extracted using least squares fitting by an automated procedure with no human interaction, then prior use of least squares smoothing does not result in improved precision. Smoothed data does not contain any additional information; in fact, some information has been lost due to bandwidth reduction. Smoothing makes the information in the data more easily accessible to human interpretation. For measurements involving human interaction with data interpretation, the optimum smooth is the one that produces the maximum S/N enhancement of the raw data. For a particular width peak, there is a maximum S/N enhancement factor which increases as the density of sampled points increases across the peak FWHM. There are several combinations of smoothing width and number of smoothing repetitions that will yield a particular factor of S/N enhancement. The filter functions which provide maximum S/N enhancement 131 are those whose frequency response approximately matches the frequency characteristics of the signal to be smoothed. As long as the user maintains perspective on his reasons for data collection and analysis, least squares polynomial smoothing can be a useful tool. Guidelines developed in this study can aid in the proper utilization of the technique. CHAPTER 6 STUDIES ON THE REACTION OF CYANAMIDE AND PENTACYANOAMMINEFERRATE A. Background Although cyanamide (HZNCN) has widespread uses in fertilizers (lll), textile fireproofing (112), weed control (113), defoliants (114), and in wood pulp processing (115), only limited attention has been directed toward development of fast and sensitive methods of analysis. Such a method would be of environmental importance since cyanamide is moderately toxic either by ingestion or inhalation (116, 117). Traditionally cyanamide has been determined by precipitation with silver followed by titration of the silver in the precipitate (117). This method is quite time consuming and sensitive only to macro amounts of cyanamide. Buyske and Downing (111) developed a spectrophotometric method for cyanamide based on the formation of a purple complex with sodium pentacyanoammineferrate(II) (Na3[Fe(CN)5NH3], from this point on called SPF) in a carbonate buffer at pH 10.5. The absorption of this complex was measured at 530 nm after reaction time of almost an hour. Their method had a limit of sensitivity of 1 pg cyanamide in 1.3 ml sample or 0.77 ppm, and exhibited linear response over a 53-fold range of concentrations. The cyanamide content of soil, blood, and urine was determined in this manner, but heavily colored samples interfered with measurement. Other workers (118) have analyzed for cyanamide using a similar reaction with sodium nitroprusside (Na3[Fe(CN)5N0]-2H20). 132 133 Holler (119) has conducted equilibrium studies on the SPF-cyana- mide reaction to determine the nature and composition of the purple complex. Spectra of mixtures of SPF and cyanamide showed a peak at 394 nm due to SPF, a peak at 530 nm due to the complex, and a single isosbestic point at 434 nm. Slope-ratio and mole-ratio studies indicated the stoichiometry was one mole of cyanamide and two moles of SPF; the molar absorptivity of the complex was determined to be 2860 at 530 nm. It was proposed that the cyanamide dianion replaced the NH3 ligand from each of two SPF molecules to form a carbodiimide bridge and yield a complex of the form [Fe(CN)5NCN(CN)5Fe]8'. The complex was not isolated for structural determination. 8. Role of the Vidicon Spectrometer It was decided to study the kinetics of the SPF-cyanamide reaction to gain insight into the mechanism and to see if cyanamide could be determined by an initial rate method with this reaction. Because the reaction was not too fast and the absorbance bands fell in the visible region, the reaction was suitable for study with the vidicon spectrometer. A 230 nm window containing both the 394 nm and 530 nm peaks could be monitored to observe both reactant and product simultaneously. Figure 6-1 illustrates the spectra for solutions (buffered at pH 10.5) having different ratios of SPF and cyanamide, with the SPF concentration held constant. The spectrum with the most intense absorbance due to the complex is of a stoichiometric ratio of cyanamide and SPF; since a significant fraction of the SPF still remains free, the SPF-cyanamide complex does not have a large apparent formation constant at this pH. In the region of the 394 nm peak the vidicon rapidly loses sensitivity. 134 1.01 SPF SPF-CYANAMIDE COMPLEX Amax = 394 nm Amax = 530 nm LIJ Q E a 0.5a c: (I) an <1 0.0 I ‘ 400 450 500 550 WAVELENGTH NM Figure 6-1. Absorption Spectra for Mixtures of SPF and Cyanamide. 135 As a result this peak was quite noisy and used mainly for qualitative and not quantitative observation. Both bands are seen to be quite broad and can be adequately defined using a small number of wavelength channels per spectrum. An estimate for the rate of the reaction was determined by simply mixing the solutions together on the benchtop and watching the rate of color change. The reaction half—life was observed to be on the order of a few seconds (with approximately millimolar solutions), and therefore was well within a suitable time scale for study using the vidicon spectrometer. 0n the basis of the rate of reaction and broadness of the peaks, it was decided to divide the wavelength window into 128 channels. With this degree of wavelength resolution the maximum data acquisition rate (without signal averaging or charge integration) was a spectrum every 8 msec. The fastest rates actually used were on the order of a spectrum every 50 to 100 msec, so some signal averaging was possible. Acquisition of a spectrum every 100 msec resulted in 1,280 data points per second. For this volume and rate of acquisition, interaction with the computer for timing, data manipulation, storage on a mass storage device, and subsequent analysis was essential. A special software set was developed to use the vidicon spectrometer in kinetic analysis, and is described later in this chapter. For studies of the kinetics of the reaction, solutions were mixed by using a stopped flow device built by the Hacker Machine Company. Pneumatically driven syringes injected a total of about 1.2 ml each time they were triggered. The observation cell was 2 cm long with quartz windows at each end. The light source used was a tungsten lamp 136 from a GCA McPherson EU-701-50 light source module. The dispersion system (described in Chapter 8) was placed between the stopped flow module and the vidicon detector. No modifications were made of the hardware of the vidicon spectrometer, however a flag was added to the interface to trigger the computer when the drive syringes stopped. (A relay in the stopped flow module closed at that instant.) Figures 6-2 and 6-4 illustrate the family of absorption spectra recorded during the course of a typical reaction. The reactant band at 394 nm decreases and the product band at 530 nm increases with time. These are the only bands observed; no bands for intermediates are seen. No peak shifts occur and only a single isosbestic point is seen, indi- cating a simple reaction. (The cause of the small dip at about 525 nm is not chemical but apparently due to an insensitive spot which developed on the vidicon target.) In Figure 6-2 data were acquired over most of the course of the reaction whereas in Figure 6-4 acquisition was mainly during the initial linear portion of the reaction. Figures 6-3 and 6-5 are plots of the absorbance at 530 nm versus time for these two runs. An interesting point to note is the nonlinear section near zero time. This stopped flow device was known to have a dead time problem which would cause such nonlinearity, but the dead time had been previously measured as 40 msec and not the 250-400 msec indicated here. The nonlinearity was initially thought to be due to some initial step in the reaction, but all attempts to study the nonlinear region by varying reactant concentrations failed to yield consistent results. It was finally determined (as described in Chapter 4) that the intensity of the source caused the vidicon detector to operate in the nonlinear region near saturation. Lowering the source intensity eliminated the 1.0 r 137 SPF=0.00075 M CYANAMIDE=0.00138 M pH 10.5 Carbonate Buffer ABSORBANCE 1 0.0 I l ’1 350 450 550 WAVELENGTH NM Figure 6-2. Absorption Spectra During Reaction of SPF and Cyanamide. 1.0 F ABSORBANCE (530 nm) 'r 0.0 I I l 0 10 20 TIME (SECONDS) Figure 6-3. Absorbance Versus Time for Figure 6-2. 138 1.0 ~— " SPF=0.00075 M ._ CYANAMIDE=0.00138M pH 10.5 Carbonate Buffer a , _ (I z \ c: __ t m ~- a: O 8 < l—- o'.o A 1 I T‘ 350 450 550 WAVELENGTH NM Figure 6-4. Absorption Spectra During Reaction of SPF and Cyanamide. ].OF .0 l— o. E __ .' c o O 3 F— 0 Lu "" . u z __ ‘ < o 3 . o r— m C on < F . r'a ' . 0.0 ' l l l I 1 ~1 1 4L; 1 O 1 2 4 8 16 TIME (SECONDS) Figure 6-5. Absorbance Versus Time for Figure 6-4. 139 0.50 o o o a Ant 0 o _ 394 nm Lu 0 z <: CD a: 8 530 nm co 4: 0.00 1 I I T I I *1 o 10 20 30 4o 50 60 70 TIME (SECONDS) Figure 6-6. Absorbance of Reactant and Product During SPF- Cyanamide Reaction. 140 nonlinearity as seen in Figure 6-6 which illustrates the increase in product absorbance and the decrease in reactant absorbance during a typical stopped flow run. C. Observations on the Reaction SPF (Fisher) for the study was analyzed for carbon, hydrogen, and nitrogen by Chemalytics (Tempe, Arizona). It was found to contain 16.71% carbon, 2.34% hydrogen, and 23.02% nitrogen which is consistent with the formula Na3[Fe(CN)5NH3]'5H20 giving a formula weight of 362. Only enough solution to be used in one day was prepared and this was stored in the dark, since over several days the solution begins decomposing, especially upon exposure to light. Cyanamide (Eastman) was used without further purification. Stock solutions were standardized by precipitating cyanamide with ammoniacal AgN03, then dissolving the yellow precipitate in dilute nitric acid and finally, determining silver by the Volhard titration (117, 120). Buffers were prepared at 0.2M and added to the SPF solution prior to mixing with cyanamide. The final buffer concentration was 0.05M. Early in the course of study there was some indication that the measured initial rate of the reaction depended upon the age of the SPF solution, with the rate seeming to increase over the first one to three hours (rather than decrease as might be explained by solution decomposition). More careful study failed to confirm this early idea. However before the notion was refuted, several experiments were performed which provided interesting results. Fearon (121) had observed that the NH3 ligand on SPF was easily replaced. With this in mind, it was thought that when the SPF was placed in a basic aqueous solution the 141 NH3 ligand was slowly replaced by H20 or 0H" and that both the resulting aquated species and the original ammoniated species reacted with cyanamide but at different rates. This would explain the slow drift in initial rate and also the initial nonlinear portion of the reaction curve. (The detector nonlinearity had not been discovered at this point.) The reaction was run in a pH 10.5 carbonate buffer since that was the buffer used by Buyske and Downing. (They made no mention of the reason for choosing a carbonate buffer.) To suppress the displacement of the ammonia ligand, the reaction was tried with a pH 10.5 ammonia buffer. In this buffer the reaction did not proceed at all; no purple complex was formed. Thinking that perhaps the reaction was specific for carbonate buffer, a pH 10.5 borate buffer was used and the reaction proceeded as in the carbonate buffer. At this point the reaction was attempted in carbonate buffer to which solid NH4C1 had been added; the purple complex did not form. Also if the purple complex was allowed to form in carbonate buffer and then NH4Cl was added, within a few min- utes the purple color disappeared leaving a yellow solution which dis- played a single absorption band at 394 nm just as the original SPF did. It now appeared that the complexation reaction would not proceed unless the ammonia ligand was replaced, probably by water. To examine this poSsibility, sodium pentacyanoaquoferrate (Na3[Fe(CN)5H20].H20) was prepared from sodium nitroprusside using the method of Hofmann (122). In carbonate buffer the aquo complex displayed an absorption band at 394 nm and reacted with cyanamide to produce a purple complex absorbing at 530 nm. For a given concentration of cyanamide, the absorb- ance at 530 nm per mole of iron complex added was the same, regardless of whether the aquoferrate or ammineferrate complex was used. In each 142 case the purple complex could be destroyed by addition of NH4C1. This work with the aquoferrate complex was felt to be strong evidence that prior to reaction with cyanamide, SPF exchanges its ammonia ligand for a water ligand. However, the initial rates observed with the two iron complexes did not agree. With pH 10.5 carbonate buffer and 1.372x10-3M cyanamide (after mixing) the pseudo zero order rate constant observed with the ammineferrate complex was 7.6 times greater than that for the aquo complex. The reason for this discrepancy was not resolved and attention was turned to initial rate studies of the SPF-cyanamide reaction. 0. Initial Rate Studies The vidicon spectrometer was used to gather absorbance data during the initial portion of the SPF-cyanamide reaction in order to determine the reaction order with respect to cyanamide, SPF, and hydroxide ion. All work described in this section was done with the ammineferrate complex in carbonate buffers. Concentrations mentioned are the initial concentration after mixing in the stopped flow cell. At pH 10.5, with an initial SPF concentration of 2.530x10'3M, the initial rate was measured (monitoring absorbance at 530 nm versus 3M to time) as the cyanamide concentration was varied from 2.754x10- 1.102x10'5M (Table 6-1). At lower concentrations the total absorbance change was too small for the rate to be measured. A plot of log(initial rate) versus log([cyanamide]o) (Figure 6-7) is linear over this range with a slope of 1.08 indicating that the reaction is first order in cyanamide. The intercept of the log-log plot is 2.429 (antilog = 269 A/sec). With the molar absorptivity of the complex at 530 nm equal to 2860 and with a 2 cm path length this yields a pseudo first order 143 Table 6-1. Initial Rate Studies. pH Initial Concentrations Initial Rate (after mixing) Absorbance/sec Cyanamide SPF 530 nm 10.5 2.754x10‘3 2.530x10'3 0.4469 1 2.203 4 0.3489 1.652 0.2505 1.102 0.1582 5.508004 0.0996 2.754 0.0443 2.203 0.0322 1.652 0.0223 1.102 0.0142 5.508x10‘5 0.0070 2.754 0.0028 1.652 0.0017 1.102 I 0.0013 1.377x10‘3 2.533x10'3 0.2388 4 1.013 0.1193 5.065x10‘4 0.0504 2.532 0.0181 V 1.520 0.0118 9.23 3.000003 0.1122 9.34 9 0.1185 9.78 0.1126 9.90 0.0875 10.23 0.0600 10.34 0.0373 10.87 0.0170 11.00 0.0136 11.09 0.0094 11.15 I I 0.0074 144 0.10-q o 1 E C O _ C m m 3: 'U \ < .1 3 Lu '— 3‘ 0.01—J _II :5 i '— E ‘1 q .1 .00] T F I I l T I I I j 10'5 10‘4 10'3 Figure 6-7. INITIAL CONCENTRATION OF CYANAMIDE (M) Initial Rate Versus Cyanamide Concentration. 145 o s 4 c o m m 3 o E 0.10-4 Ii .— 13 a E —J < i t: E - 0 0°01 I I I I I I 10'4 10'3 INITIAL CONCENTRATION OF SPF (M) Figure 6-8. Initial Rate Versus SPF Concentration. 146 rate constant of 4.69x10"2 sec-1. Next, the order in SPF was determined. At pH 10.5, the cyanamide concentration was held at 2.754x10'3M while the SPF concentration was 3M to 1.520x10'4M (Table 6-1). A plot of varied from 2.533x10' 1og(initial rate) versus log([SPF]o) (Figure 6-8) was linear over this range with a slope of 1.12. A t-test indicated that at the 95% confidence level the estimate of the slope was not significantly different from one, so the reaction is also first order in SPF. The intercept of the log-log plot is 2.349 yielding a pseudo first order rate constant of 3.90x10"2 sec-]. Combining the results for the reaction order with respect to cyanamide and SPF we can say that at pH 10.5 the initial rate of appearance of the complex is given by d[Complex]o/dt = k[SPF]o[H2NCN]O where k = 1.64 1 mole-1 sec". The effect of solution pH on the rate was determined by preparing a series of ten carbonate buffers spanning the range from pH 9.2 to 3M SPF and pH 11.1. The initial rate for the reaction of 3.000x10- 1.372x10'3M cyanamide was measured in each buffer (Table 6-1). On a plot of log(initial rate) versus pH (Figure 6-9) the data do not exhibit a good linear region but show the rate to decrease as pH increases. At high pH the slope is almost -1, indicating inverse first order in hydroxide ion. The relationship between pH and rate can be understood by considering the forms of cyanamide present in solution. K1 K2 HZNCN + 014' 22:: WW“ + 011' :__—- NCN'Z Cyanamide has a pK] of 1.1 and a pK2 of 10.27 (123). In the pH region studied, all of the cyanamide will be divided between the 147 INITIAL RATE (dA/dt)530 nm 0.01 I I l I 9.0 9.5 10.0 10.5 11.0 pH Figure 6-9. Effect of pH on Initial Rate Compared to the Monoanion Fraction. Experimental Points --Calculated Monoanion Fraction FRACTION AS MONOANION . d 148 monoanion and dianion. If a log concentration plot is constructed for the monoanion and compared with the experimentally observed rate it can be seen that the initial rate is proportional to the concentration of the cyanamide monoanion. (To obtain this agreement a pK2 of 9.96 instead of 10.27 was used. This difference can easily be explained by activity corrections at the high buffer concentration used in this study.) Varying the pH causes a change in the reaction rate by changing the fraction of cyanamide present as the monoanion. Since the initial rate was determined to be first order and not second order in SPF (as anticipated from consideration of probable mechanisms based on a two to one stoichiometry) and since no evidence for more than one complex was seen from the absorption spectra either during the reaction or at equilibrium, there was some doubt as to the validity of the two to one stoichiometry determined by Holler. Holler indicated (124) that an attempt to use Job's method (125) failed to yield consistent results and that he had then used the slope-ratio and mole-ratio methods instead. Furthermore his confidence in the result from the mole-ratio method was low. It was decided to check Holler's stoichiometry by reattempting Job's method. At pH 10.5 poor results were obtained indicating stoichiometries of three to one or four to one or higher. In light of the pH dependence of the initial rate, the problem was easily determined to result from the equilibrium between the cyanamide monoanion and dianion. Repeating Job's method at pH 9.0 (where all of the cyanamide was present as the monoanion) resulted in excellent data yielding an unambiguous result of two to one stoichiometry and supporting Holler's conclusion. 149 E. Conclusions It has been seen that the reaction between pentacyanoammineferrate and cyanamide in aqueous base proceeds by the loss of the ammonia ligand from the SPF and complexation with the cyanamide monoanion. (Recent work by Toma and Malin indicates that'hiunbuffered aqueous solutions the ammonia ligand is lost from pentacyanoammineferrate with a half- life of about one minute (126).) The initial rate equation at constant pH is first order in SPF and first order in cyanamide. At pH 10.5 the second order initial rate constant is 16.4 1 mole-1 sec']. Since the rate equation is first order in SPF, then the rate limiting step occurs before attachment of the second SPF ligand and is probably the attachment of the first ligand. A kinetic method for cyanamide is quite practical using the initial rate as measured at 530 nm. In a carbonate buffer at pH 10.5 with an SPF solution of 5.5x10'3M the initial rate is linearly proportional to the cyanamide concentration over almost three orders of magnitude with a limit of detection of 0.24 ppm. The kinetic method for cyanamide has several advantages over the equilibrium method reported by Buyske and Downing. The linear concentration range is five times greater and the time for determination is an order of magnitude shorter. The detection limit of 0.24 ppm for the kinetic method is lower than that for the equilibrium method. The detection limit for the kinetic method was measured using an SPF solution only one-fourth as concentrated as that used in the equilibrium method. Had the same concentrations been used, the anticipated limit of detec- tion for the kinetic method would be proportionately lower, or 60 ppb. 150 The kinetic method will be less susceptible to error due to analysis in colored solution since only relative measurements are made. Based on the study concerning the effect of pH on the reaction rate, the kinetic method would be more sensitive at a pH of about 9 rather than 10.5 as proposed for the equilibrium method. At pH 10.5 only 30%-40% of the cyanamide exists as the monoanion, while at pH 9.0 greater than 95% is the monoanion. Finally, since ammonia inhibits the reaction, determinations of cyanamide in samples containing ammonia (such as urine) should be compared to standards run in the same matrix. The vidicon spectrometer proved quite useful for investigating this reaction. By monitoring an entire wavelength region, it was easily seen that there were no intermediates (at least on the time scale and over the wavelength region studied) and that the reaction proceeded through a simple mechanism. For routine analysis of cyanamide, however, based on information from this research, one would be better off to use a photomultiplier tube and monitor only at 530 nm. With the higher S/N of the photomultiplier, one would expect that the detection limit could be further lowered by perhaps an order of magnitude. Most of this work was performed at pH 10.5 because that was the pH that Buyske and Downing felt to be optimum. It was not until late in the course of this research that the reaction was discovered to be faster and less complicated by other equilibria at lower pH (around 9). Had this pH effect been noted earlier, all work would have been performed at pH 9. Besides the equilibrium involving the cyanamide monoanion and dianion, the hydrolysis of SPF is anticipated to be affected by the pH due to the acid-base equilibrium of NH3 and NH: 151 in solution. Continued work is planned to study the hydrolysis of SPF and the stoichiometry of the complex as a function of pH. A kinetic method for cyanamide (at around pH 9) will be developed. F. Software An extensive vidicon software package was developed for kinetic analysis (Figure 6-10). The package is built around an input monitor (VNTMIN), an output monitor (VNTAMD), and a display program (VNTADS), each of which calls several subroutines. (Listings are in Appendix F.) Only one of these programs can fit into memory at a time but it can call the other two. The data acquisition routine was written to permit use by an individual with only minimal knowledge of the instrument operation and characterization (Figure 6-11. All input by the user is underlined.). The program asks for parameters concerning wavelength resolution, time resolution, duration of data acquisition, and S/N enhancement. The program then times aquisition according to the desired parameters to determine if it is possible in the allotted time. The user is instructed to alter his parameters if they are incompatible. If the program decides that there is sufficient time to acquire the data, it then examines its options to store the data. If acquisition is sufficiently slow (at intervals longer than half a second) the program will write each spectrum on DECtape as it is received. If acquisition is too rapid, then the program attempts to store all of the data in its buffer area (3072 words). Failing in this attempt, the program instructs the user that data acquisition will be interrupted whenever 152 VNTMIN VNTADS VNTAMO SMOTH '-‘{1east squares smoothing output , monitor J NINIT . 1 PTPLT parameter . input and ~ * paint plot check DSCRIB DRCTRY 7 file directory descriptions and file line plot 4 . manipulations ~ - VNTDA LLSQ timed data slope acqu151t10n calculation analysis routines Figure 6-10. Kinetics Software Package. 153 I=EXIT32=ACQUIREa3=DESCRIBE34=DISPLAY3S=DIRECTORY‘§_ POINTS/SPECTRUM = 512 SEC. BETWEEN SPECTRA - :1. O SPECTRA a.gl a SAMPLE AVERAGES =.Lg I SAMPLE EXPOSURES - I PARAMETERS REQUIRE “ 0.554740 SEC BETVEEN SPECTRA CHANGE PARAMETERS OR TIME POINTS/SPECTRUM = 51g SEC.BETVEEN SPECTRA a ;I_ l SPECTRA =.gl I SAMPLE AVERAGES . ;_ I SAMPLE EXPOSURES = ;_ PARAMETERS REQUIRE INTERMITTENT ACQUISITION 2 SECOND GAPS EVERY 6 SPECTRA I-OK. PROCEED. 2-CMANGE PARAMETERS g_ POINTS/SPECTRUM . 128 SEC. BETVEEN SPECTRA - EL. l SPECTRA -.gl ' I SAMPLE AVERAGES = _8_ a SAMPLE ExPOSUREs ”.L CONTINUOUS ACQUISITION o DARK AVERAGES . gg_ t REFERENCE AVERAGES . 6i 0 DARK EXPOSURES a l_ o REFERENCE EXPOSURES ‘.L ENTER IDENTIFYING TEXT TYPE CONTROL 6 TO STOP HELLO THERE THIS IS A TEST (CTRL G) DATA CODE 2 g; DARK SIGNAL To OBTAIN TRIGGER ENTER A I (ONE)___ DATA CODE . i REFERENCE TO OBTAIN TRIGGER ENTER A I (ONE)___ DATA CODE '.1 SAMPLE TO OBTAIN TRIGGER ENTER A 1 (ONE) 1. DATA CODE - §_ Figure 6-11. Parameter Input for Timed Data Acquisition. 154 .R VNTAMO IBEXITJZSACQUIRE:3=DESCRIBEIA=DISPLAY35=DIRECTORY‘§_ ItLIST: 2=DELETEa 3'COPY3 482E303 S=RETURN £_ ARE YOU SURE? I'YESO__ I'LISTJ 23DEI.ETEJ 3=COPY1 119-ZERO: 5=RETURN 5 I-EXITaZIACQUIREaO'DESCRIBEaAIDISPLAY:SSDIRECTORY 2_ DESCRIBE RUN '22 21 SPECTRA 128 POINTS PER SPECTRA 0.2500 SECONDS BETWEEN SPECTRA CONTINUOUS SIGNAL AVERAGES: DARK - 64 REF - 64 SAMPLE s 16 EXPOSURESI DARK I I REF 8 I SAMPLE a I 10/22/74-2 SAME AS LAST RUN BUT TWICE AS FAST LUMINOLIoOI M K-T-BUOISATD BOTH IN DMSO DESCRIBE RUN I 1 Figure 6-12. Data Run Descriptive Text. 155 the buffer fills to allow sufficient time (not over two seconds) to transfer the data to tape. Time lost in data transfer is measured and stored with the data so that analysis is still possible. The user can either accept this intermittent acquisition or alter the acquisition parameters. (For the studies described in this chapter, intermittent data acquisition was not allowed.) Next, identifying text may be entered to identify the data run; the text is stored with the data. Finally the Spectra corresponding to dark signal, reference, and timed acquisition of the sample (after a trigger from the stOpped flow) are taken and stored. The output monitor (VNTAMO) permits directory Operations on the data tape (listing contents, deleting runs, transferring data between tapes, and zeroing the tape) and description of data runs. If the "describe" option is chosen (Figure 6-12) all of the parameters and identifying text concerning a specified data run are output. The display program (VNTADS) allows display of the family of spectra recorded over the course of the reaction or the time development at any wavelength point in the spectrum. Also contained in this program is a least squares fit routine to calculate the initial rate of reaction. CHAPTER 7 STUDIES ON THE CHEMILUMINESCENT OXIDATION OF LUMINOL IN DMSO A. Background Recent reviews by Seitz and Neary (127) and Isacsson and Wetter- mark (128) have indicated the wide applicability of chemiluminescence (CL) for trace analysis. The most pOpular system at present is the oxidation of luminol which can be used to determine some 20 transition metals (down to 10'11M) and numerous organic substances. The luminol reaction was discovered by Albrecht (129) in 1928 and has received continued interest due to its high quantum yield (about 5%). Since Albrecht's discovery, many investigators have studied the reaction but the full details of the mechanism have yet to be elucidated. The stoichiometry of the reaction has been Shown by White (130, 131) to be C‘N 'H '0- ' ~ oxidant (::) N-H + 20H Fa's—e—" . + 2H20 + N2 + hV NH2 NH Luminol (LH 2) 2) 3-AminOphthalate (3-AP' The reaction may be carried out in water or aprotic solvents such as DMSO. In basic aqueous solutions, the oxidant is hydrogen peroxide plus a trace of a metal catalyst. In DMSO, using potassium tertiary butoxide (K-t-BuO) as the base, dissolved molecular oxygen acts as the 156 157 oxidant. The mechanism proposed by White for the reaction in DMSO is shown below: .. K] _ LH2 + 0H ;;;::3: LH + H20 K LH‘ + 0H“ ::;§::: L”2 + H20 -2 k3 -2 L + 02 -————> [L02 ] k [LO-2] ___5L. [3—AP'2]* + N 2 2 -2 k5 -2 [3-AP ]* ———> 3-AP + h\) where [L052] is a bridged, cyclic peroxide: R\\. NO firm evidence has been reported concerning 0}: I:- detection of this or any other intermediate in 9)- the reaction. White Observed the reaction to be NHZ' g first order in luminol, base and 02. Several studies (132-134) of the Spectral characteristics of the CL emission have supported the belief that an excited state Of 3-AP'2 is the emitter. Gorsuch and Hercules (133) have performed limited stopped flow studies on the reaction in DMSO and DMSO-H20 mixtures. They concluded that the rate limiting step is decomposition of [L022] to form [3-AP-2]*, and proposed a rate constant of 0.12 sec.1 for that step in pure DMSO. To expand the knowledge on the CL oxidation of luminol we proposed to study the reaction by using stopped flow kinetics and monitoring with bipolar conductance and Spectroscopic detection. 158 B. Conductance Studies All work which used conductance to monitor the CL reaction and some of the work which involved monitoring the intensity of the emitted light was performed jointly with Keith J. Caserta. Conductance measure- ments were made by using a computer-interactive bipolar conductance instrument developed and characterized by Caserta. The instrument was capable of making measurements every 30 usec and so was very appropriate for use in studies of reaction kinetics. Discussion of the instrument and experimental details concerning study of the luminol reaction may be found in Caserta's PhD dissertation (135). The present section will discuss only the results of that research. The system initially chosen consisted of K-t-BuO as the base and DMSO as solvent for both luminol and base. When the solutions were mixed in a stopped flow device at room temperature a smooth increase in conductance (G) was observed which asymptotically approached a maximum with a half-life of about one second. Extensive studies were performed on the temperature changes involved in mixing and in reaction and on the conductance-temperature behavior of the system. It was concluded that most of the observed conductance increase was due to the reaction and not merely to temperature effects. By simultaneously measuring the solution conductance and temperature during the reaction, it was possible to have the computer correct for the small portion Of the observed conduc- tance change which was due to changes in the temperature. After preliminary investigation, it was decided to replace K-t-BuO by some other base. Difficulties were encountered in producing accurate solutions of K-t-BuO in DMSO due to incomplete dissolution and also due 159 to decomposition which became significant after one to two hours. A decision was made to use KOH as the base. Since KOH is insoluble in DMSO a l/l mixture of DMSO and EtOH was tried but the reaction did not proceed in this solvent. Instead of using a premixed 1/1 solvent, luminol was dissolved in pure DMSO and KOH was dissolved in pure EtOH. When these solutions were mixed, intense light (significantly more than with K-t-BuO in DMSO) was emitted and a smooth conductance increase was observed. After correcting for temperature changes (these solutions mix endothermically) there still remained a large portion Of the conductance increase that was not due to the reaction but rather due to solvent mixing effects which influenced the conductance of KOH. Subtraction of a reference obtained without luminol corrected for this phenomenon, leaving a smooth conductance increase due only to the reaction. The reason for the conductance increase was sought in order to determine what step(s) Of the reaction was (were) being followed. Since the curve was monotonically increasing, it was not caused by intermediates. From WhiteS mechanism, this then left either the initial proton transfers or the formation of 3-AP'2. If the base (KB) were weak, then the proton transfer steps would create K+ ions which were not previously present. 2 KB + _ KB + - 2——-K+HB+LH—5K+HB+L The strength of KOH in 1/1 DMSO-EtOH was not known but was measured 8 LH by conductance. Over the concentration range of 4x10'2M to 8x10" M a 1/2 graph of G/C versus C was linear, as expected for a strong electrolyte. Furthermore, measurements were made Of the conductance of solutions of KOH in DMSO-EtOH before and after complexing K+ with dicyclohexyl-lB- 160 crown-6. Since the K+-crown complex has only a slightly lower conductance than K+, if the KOH were not dissociated, addition of sufficient crown to complex all of the potassium would create many OH' and K+-crown ions and cause a sharp conductance increase. Instead, a 5% decrease in conductance was Observed after addition of crown indicating that the KOH had been dissociated. In addition it was felt that the proton transfer steps would be much faster (even in an aprotic solvent) than the time (tl/Z 2 2 sec) indicated by the conductance versus time curve. This belief was confirmed by doing stopped flow studies on phthalic acid and phthalamide (also in EtOH-DMSO) which are structuraly similar to luminol: 2| 8 o C ‘9 O O 010“” -H c" 3’0” ’\\0 NH2 6 0 Luminol Phthalic Acid Phthalamide The conductance increase in these two cases occurred virtually instantaneously with mixing, indicating that, in fact, the proton transfers were very fast. Since the acid-base reaction of phthalamide and KOH in DMSO-EtOH resulted in a conductance increase, the phthalamide anion must have a higher conductance than hydroxide ion. This can be explained if OH- is more extensively solvated by bulky solvent molecules than is the phthalamide anion. The higher charge/mass ratio and greater degree of charge localization on 0H“ could cause such behavior. Since 3-AP'2 is structurally similar to the phthalamide anion, it also could be 161 expected to have a higher conductance than OH'. The conductance increase during the reaction would then be due to creation of 3-AP'2. To support this idea, the intensity of light emission (I) was monitored simultaneously with the conductance (G) during a reaction. Light emission increased to a maximum and then decreased to zero. The CL has been Shown to originate from the conversion of [3-AP'2]* to 3-AP'2 so the integral of the light intensity curve is then proportional to the concentration of 3-AP"2 product. Figure 7-1 illustrates I, II, and G as measured during reaction of luminol and KOH and shows that the [I and 6 curves are quite similar, strenghthening the belief that the conductance instrument was monitoring creation of 3-AP'2. Further work was not pursued due to extensive corrections (temper- ature and solvation) required when using KOH in EtOH. C. Spectroscopic Studies At the beginning of the studies concerning the luminol reaction, it was intended to complement the conductance work by measurements Of the emission spectrum during the reaction by using the vidicon spectrom- eter. It had been reported that luminol exhibited more than one emission band and there was a possibility that the relative intensities of the bands shifted with time (136). The plan to use the vidicon spectrometer was made before its strengths and weaknesses were thoroughly understood, however. It was discovered that although the CL emission (495 nm) was well within the spectral range of the vidicon, the sensi- tivity of the vidicon was not great enough to permit detection after the emission had been dispersed through a polychromator. 162 .conunmm mchaa mucmumm:_E=FFEmgu nounemmpcm can .mucmummcFE=F_eo;u .wucnaosucoo .F-~ mesmpe mm (sawv) NOISSIW3 mum.~ L Amazonnmv nzne ow mF oF m c .11 11 p E b E p E - VIMOQM . zomeFZm hzmumszz=4F2mzu 3 0 N 1 mm 3 I. v N 3 3 muz I I f g g ——'( A/D CONVERSIONJ l 1 ° "’ 0 . I 1 —rl GATED DRIVER J 1 L v - - - - - _ - — J [B & IOP AC in INTERFACE BUFFER PDP 8/1 COMPUTER Figure 8-1. Block Diagram of Instrument Circuits. 168 card punch (138), display controllers for a Varian F-80 x-Y recorder, a Heath EU-205-11 strip chart recorder, and a Tektronix 535A display scope (139), a Character generator for the display scope (140), and a versatile mainframe real-time clock (141,142). The Heath EU-801E Interfacing System (143) was used for all of the in-house interfacing mentioned above and formed the backbone of the interface for the vidicon spectrometer. This interfacing system provided, in parallel, 12 accumulator input (AC) lines, 12 buffered accumulator output (BAG) lines, and 12 buffered memory buffer (8MB) lines. Control of functions in the vidicon spectrometer was accomplished by using the device select (05) lines, input output transfer pulses (IOP), and the skip (SKP) line. C. Analog Circuits The power supply and the tube and deflection assembly support circuits are based on conventional television circuitry (144). The silicon vidicon tube (RCA 4532), deflection assembly (Cleveland Electronics VYLFA-959), and Circuits for magnetic focus and alignment, line deflection (fast scan axis), and signal preamplification are contained in a small aluminum box which constitutes the detector. The latter circuits are in close proximity to the tube and deflection assembly because of their high speed and low level signals. Power for the circuits and for the tube is obtained through shielded cables from the power supply located in a separate box. The destination, and voltage and current requirements of the power supply lines are given in Table 8-1. The magnetic focus and alignment control circuits are adjustable current sources. The focus circuit provides between 40 and 45 mA through 169 Table 8-1. Power Supply Requirements. Destination Voltage Current Tube pin 1 Heater 6.3 VAC 0.1 A 3 Grid 4, decelerate 340 5 Grid 2, accelerate 300 6 Grid 3, electronic focus 290 8 Heater 6.3 VAC 0.1 A Tube target flange Target bias 8 Line deflection -15 45 mA Preamplifier -6 Magnetic Focus -30 45 mA Alignment -15 :90 mA +15 :90 MA 170 the 360 9 focus coil (purple and white/purple wires). There are two pairs of alignment coils (One pair is the brown and white/brown wires, the other is the orange and white/orange wires.) at 140 0 per pair. Each pair of coils is driven by a current source adjustable between -40 mA and +401mA. Pots to control the focus and alignment currents are located in the detector housing. Once they have been properly adjusted there is no reason to change them. The line deflection circuit (Figure 8-2) creates a 185 mA current sawtooth in the deflection coil (red and black wires). Transistors 01 and 02 and associated components generate a square wave at about 15 KHz which is buffered by Q3. Transistor 04 is normally held in saturation. The differentiated square wave from 03 brings 04 out of saturation. The 22 mH inductor increases this pulse to about 40 volts which is applied through a 10 pF capacitor to the deflection coil. Because of the inductive reactance of the coil at 15 KHz (0.94 mH, 2.69) a volt- age pulse is required to generate a current ramp through the coil. Since information is contained only in the wavelength axis of the vidicon target, the line deflection waveform need not be perfectly linear. It must however be absolutely consistent from line to line. The video preamplifier (Figure 8-3) is capacitively coupled to the vidicon target. The amplifier is located within one inch of the target and connected by a short piece of shielded cable. Target signals of a few hundred nanoamps are raised to about 100 mV and transmitted on shielded cable to the interface for further conditioning and sampling. The preamplifier has a frequency response of about 4.5 mHz. 171 .neeecwu eeenea_can ae_S Lo eveneeeam .N-n ee=m_a >m+ g, x, ¥~.~ uz>m zm I emec u mo.mo voczm u .co.mo.~o mNFczm u Fo 202200 UL? Foo No ——_— Fa Foo. mw um .n< u:n=.n u 621 l X28 F....,\1....g X8'9 .uFauLFu LOFFFFnsnan mo uFunsmgum .m-m «Laure mmmmzm u oFo.mo.wo.mo 5.28 , >9. “:5. comm “38:... SN aamwooir 352 new 3.2: .41 I E r a 2: Ew. :11} mo 2 ...... 8 n 856% So “and 3.8 0 an: 8 I .._ .8. r 3. r e E. hr v 88 3.2: 3.3 35: can :2 V enn¢mn 173 D. The Vidicon Interface The interface portion of the instrument contains circuitry for controlling beam deflection in the wavelength axis, and for condition- ing, sampling, and conversion Of the video output signal. In addition the unit contains a data latch, a gated driver, and flags and logic required for transfer of data to and from the computer through the computer interface buffer. The signal conditioning, sampling, and conversion circuit (Figure 8-4) receives the video signal from the preamp. The signal is capaci- tively coupled and adjusted to just positive of zero volts (the preamp output is biased at -6 volts) by adding in plus and minus five volt supplies. The +5 volt supply is from the ADD module; the -5 volt . supply is Obtained from the -15 volt ADD module supply by using a voltage regulator (National Semiconductor LM304). The signal is then buffered by 0A1 and amplified by 0A2 (both are 741's). The gain of 0A2 is normally -10 but can be varied by changing the feedback resistor. 0A3 (Analog Devices 1428) integrates the video signal as the line deflection circuit samples the target at one wavelength position. Switches 51 and S2 (FET switches on 3 Heath analog switch card) control the timing of the integrator for integration and discharge. The signals controlling S1 and S2 come from a series of four monostables which are triggered by the line scan sync signal. This signal occurs every time the line deflection circuit has completed sampling the target at one wavelength and starts to sample at another wavelength. The waveforms for the line scan sync and the outputs of the four monostables are shown at the bottom of Figure 8-4. Integrating the DS 44 IOP 1 DS 44 IOP1 AC IN LINE SCAN SYNC SEQUENCING MONOSTABLES 680K FROM 47°PF PREAM LINE SCAN SYNC DELAY 21:39—77. 174 FLAG TEST SK P r-T J;K DS44 IOP2 GATE R STATUS P“, HOLD Gnd I I v A/D S/H E E5 ES ‘7 1o 31 + V 220-F 10K CA] as» 22K 0A2 S2 680K —L— FOLLOWER AMPLIFIER INTEGRATOR ' I 9 us gr- 15 s [- DISCHARGE INTEGRATOR ($2) | Is us __F 40 .51— INTEGRATE (5]) TRIGGER A/D CONVERSION Figure 8-4. 231$ ll Signal Conditioning, Sampling, and Conversion Circuit. 175 signal produced along a line Of diodes results in noise reduction (due to averaging out variations in diode sensitivity and dark current) and in signal enhancement. With the specified components and integration time 0A3 has a gain of about -8 but the gain may be adjusted by chang- ing the input resistance. After 40 us integration time the A/D (Analog Devices ADC-120U) is triggered to convert. The A/D status line causes the Sample and Hold (Intronics F5201) to hold during the course of the conversion (15 us or less) and sets a flag when conversion is complete. When the computer senses the flag it gates the 12 bit data word into its accumulator and clears the flag. The basic wavelength deflection (Slow scan axis) circuit is diagramed in Figure 8-5. This circuit produces a current in the deflec- tion coil (white and green wires) to drive the beam to the wavelength of interest. At the heart of the circuit is a D/A converter (Analogic MP1812-2-C high speed 0 - +10 volt converter, slews at 10 V/us) and a unique presettable counter with programmable increment (described in detail later). The 12 bit number in the counter determines the wave- length tO which the beam is deflected. The converter's output is buffered by 0A4 (Analog Devices 1428) before being applied to the wavelength deflection coil. Since this coil is driven at fairly low frequencies, the waveform of the applied voltage produces an identical current waveform in the coil. Pots are available to adjust the amplitude and centering of the deflection waveform. FET switches S3 and 54 are driven by complementary signals and determine whether the target is scanned or allowed to integrate. Closing S3 allows the D/A to drive the deflection coil. Closing S4 applies an offset to the coil which 176 ACCUMULATOR 12 BIT LATCH PRESETTABLE COUNTER WITH OVERFLOIN. PROGRAMMABLE INCREMENT l 12 BIT D/A I - 10 , 10K [3/ WAVELENGTH SCAN COIL SCAN ll) CONTROL / l K S +5V 4 4.7K OFFSET 10K +15V -15V Figure 8-5. Wavelength Deflection Circuit. 177 drives the beam off of the active portion of the target. The wavelength counter can either be set to a preselected wave- length from the 12 bit data latch or incremented by the line scan sync signal. If the counter is incrementing and overflows back to zero, a flag is raised which the computer recognizes as Signaling the end Of a spectrum. When the counter is clocked by the line scan sync, the word in the data latch controls the magnitude by which the counter incre- ments and so determines the number of points on the wavelength axis. The counter is a synchronous counter composed of eight flip flops (for the least significant bits) and a four bit binary counter (for the most significant bits). By suitable logic controlling the J and K inputs, the lower bits in the counter can be bypassed depending upon which bit is set in the latch. A portion of the counter is diagramed in Figure 8-6. A flip flop can toggle only when its J and K inputs are at logical “one". This occurs when either 1) the flip flOp enable bit is set in the data latch or 2) the 0 output and J-K inputs of the previous flip flop are at logical "one". The data latch code words and the resulting number Of points in the wavelength axis are given in Table 8-2. Table 8-2. Codes for Wavelength Points per Spectrum. r 0 Code (octal) # Points (decimal) Code (octal) # Points (dec1ma1) 1 4096 16 256 2 2048 32 128 4 1024 64 64 8 512 128 32 178 DATA LATCH LOADS I B I MSB -{;J * J ' I4t3 T T . . *T K K K c c CLEAR CLOCK I Jr Edi. T0 D/A Figure 8-6. Logic for Wavelength Counter. 179 [NS 45 OVERFLOW T K 11$ 45 1013 2 A/D TRIGGER D8 415 1013 2 Q c 1 s Q SCAN DS 47 --— CONTROL 1 C 1 IOP 2 SCAN OFF Figure 8-7. Control Signals for the Wavelength Deflection Circuit- 180 Figure 8—7 illustrates the logic required to generate the control signals for the wavelength deflection circuit. Notice that although the computer can issue an instruction to clear and set the counter, the timing of these instructions is determined by signals from the sequencing monostables mentioned earlier. Whether the counter is allowed to increment or is preset to a desired value, the wavelength change occurs at the same time relative to the line deflection. E. Programming Notes All of the software described in this dissertation was developed and run under the 08/8 Operating System (145). Programs were written in FORTRAN II with in-line coding of SABR (assembly language) as required for interaction with the experiment (146). In the programs listed in Appendices C through F an "S" in column one indicates a SABR statement. Large arrays of data which were to be accessed by both SABR and FORTRAN were placed in FORTRAN COMMON (locations in field one above address 0200). To access data in field one using SABR indirect address- ing it was necessary to fool the assembler because the SABR assembler generates a CDFD instruction (change to data field zero) whenever it encounters an indirect addressing statement (e.g. TAD I DATA). Genera- tion of a CDFO instruction was eliminated by defining a new mnemonic containing the indirect bit (e.g. 0PDEF TADI 1400) (147). Although the assembler issued an error message every time it assembled such an instruction, the proper code was generated. Several programs generated data that were stored on DECtape to be analyzed later by another program. When timing was not critical (as 181 Table 8-3. DS-IOP Codes for the Vidicon Spectrometer. DS IOP Function 44 1 Clear flags for A/D, external trigger, and line scan* 44 2 Check A/D flag 44 4 Gate driver 45 1 Check wavelength scan f1ag* 45 2 Clear wavelength scan flag 45 4 Check external trigger flag 46 l Latch wavelength data latch 46 2 Clear wavelength counter 46 4 Set wavelength counter 47 1 Turn on wavelength scan 47 2 Turn off wavelength scan 47 4 Check line scan flag *Note-In some of the program listings an earlier nomenclature for these scan flags was used. Line scan was called "horizontal", and wavelength scan was called "vertical" after conventional television nomenclature. 182 in VCIDCO and CHISQ) advantage was taken of the 05/8 directory, and data was output into files using device independent unit four output (wHNIOOPEN, OCLOSE, and IOPEN statements). In the kinetics data acquisi- tion routine where timing was critical it was found to be up to ten times faster to use the RTAPE and WTAPE statements to output to absolute block address on the tape without directory consultation. However, these statements work only with the T008 tape controller on the PDP 8/I. They will nOt work with a TDBE controller. In addition, data written in such a manner cannot be accessed by any standard 05/8 utility program. In all vidicon data acquisition routines, it was necessary to eliminate acquisition of the last two points in each spectrum. This step was required to allow sufficient time for checking and clearing of flags and for resetting of software counters and pointers before it was time to acquire the next spectrum. All interactions between the vidicon interface and the computer program are through input-output instructions of the form 6nnx, where nn is the device address code and x is the IOP pulse during which the operation is to be performed. In Table 8-3 the DS-IOP combinations used in the vidicon spectrometer are listed along with the functions that they initiate. F. Optical Setup The dispersion system used in the vidicon Spectrometer (Figure 8-8) was a modified Heath EU-700 Czerny-Turner Monochromator (148). The exit slit was covered, the second folding mirror removed, and the focus— ing mirror (35 cm) moved forward about one inch to cause the focal plane 183 SAMPLE CELL MODIFIED HEATH CZERNY'TURNER MONOCHROMATOR ’1 0R L \ 1 STOPPED now 11 1 \ I’m 1\\1 \ / 1 ’1 1 \1 \ I 1 H ‘ \ / / l ‘ (‘1 \ I, X 1 1 1 1 \ / 1 / l \ / 1 1 I \ \ / 1 1 1 \ )l / 1 1 1 SOURCE “1 \ / 1 7/ 1 1 .* ______ _. _ __= 7’ \ / \/ 11 \/ 11 H \U— VIDICON DETECTOR Figure 8-8. Optical Setup Of the Vidicon Spectrometer. 184 to fall just outside the front of the housing where the vidicon detector was placed. The original grating with 1180 grooves/mm was replaced by a Bausch and Lamb certified precision grating with 133.6 grooves/mm and blazed at 5461A (2°05'). These dispersion optics resulted in a reciprocal linear dispersion of 18 nm/mm in the focal plane. The wavelength axis Of the vidicon deflection coils must be exactly perpendicular to the slit image focused on the target. A simple test using a line source and the charge integration option in VCIDCO will determine if the Slit and deflection coils are properly aligned. Using charge integration to enhance a line signal should not change the location of the maximum of the peak. However, if the coils and slit are not aligned the peak maximum will shift with integration; the degree Of shift increases as the integration time increases. CHAPTER 9 CONCLUSIONS The experience with the vidicon spectrometer develOped in the course of this research has resulted in increased appreciation of the strengths and weaknesses of imaging devices as detectors in analytical spectroscopy. This type Of detector forms a useful complement to more conventional detectors and is uniquely suited to multicomponent analysis and preliminary investigation of the kinetics of new chemical reactions. It is particularly gratifying to a researcher when the fruits of his labor open the doors to other avenues of investigation. Currently planned in our laboratory is the use of the vidicon spectrometer for analysis of turbid solutions using dual wavelength or derivative techniques. Also being considered is application of the vidicon spec- trometer as the detector for a centrifugal analyzer. Work is presently in progress to study the kinetics and equilibrium of the aquation of SPF by using potentiometric methods and to examine the stoichiometry of the SPF-cyanamide complex over a wider pH range. In the very near future, these results will be combined with those reported in this thesis to develop a firm basis for a sensitive determination of cyanamide using an initial rate method. The flexibility and modularity Of the software and hardware of the vidicon spectrometer make it an easily modifiable system. One improve- ment that this author would hope to see evolve is the successful implementation of the selective integration technique discussed in Chapter 4. 185 186 The author sincerely hopes that through the medium of this thesis, he has successfully transmitted some of the knowledge and insight which he acquired during the course of performing, analyzing, and reporting this research. APPENDICES APPENDIX A PROBABILITY AND STATISTICS A. Least Squares Estimates Of Parameters of the Straight Line (151) Least squares estimates of the slope, intercept, and standard deviation of the slope and intercept for the best line (y = a+bx) through a set of x,y data points can be calculated using the following equations: Slope = b = "XX% ' ZXZYZ an - (2x) Intercept = 3 =.§X - be n n 52 Slope standard deviation = 5b = 2 2 - (2X) /n Intercept standard deviation = Vfl/ZS Van - (Xx)2 where n the number of x,y data points estimate of the standard deviation of the y measurements U1 ll associated with a Single x value and is calculated from (n-2)S2 = 2(yi-yi)2 where §i = bxi+3 or equivalently (n-2)s2 = Zyz - 1%213. Ligy - ExEy/n)2 - (2X)2 /n 187 188 B. t-Test for Hypothesis Testing (152) The t-test was used to test the hypothesis that the least squares estimate of the slope of a line through a set of x,y data points was equal to a theoretical value for the slope. The t statistic is computed by t = b-b* \( $2 2(X1'i)2 where S2 = 2(y1-;112/(n-2) y]. = 3.0.. 6 and g are the least squares estimates Of the slope (b) and intercept (a) of the line y = a+bx b* = the theoretical value for the slope n = the number of data points. A t table is used to find the critical value Of t for the proper number of degrees of freedom (n-2) and the desired confidence level (a). (The value found in the table is t(]_a/2)(n_2).) If the calculated value of t exceeds the critical value of t, then we reject the hypothesis that the Slope is equal to the theoretical value; the error in rejecting the hypothesis is no greater than a. C. Congruential Method to Generate Uniformly Distributed Random Numbers (153) With a computer, a series of pseudo random numbers is often generated using the following equation: 189 3 I 1+] - nim(mod d) where m multiplier (see below) d modulus (usually the computer word size) 111. = a number in the random series. On a binary computer, a sequence of maximum length (before repeating) can be obtained if the starting number is odd, and if the multiplier leaves a remainder of 3 or 5 after being divided by 8. A common choice is the largest odd power Of 5 that can be stored in a computer word. (For a 12 bit machine this is 55 or 3125.) If the word length is s, the sequence length is 25'2. In subroutine IRANU, 110 = 501 and "i = (ni_])(3125) (mod 4096). The random values are scaled from 0-1000 prior to output, but the "seed" used for the next number is the value before scaling. 0. Central Limit Theorem Approximation to Generate Normally DistribUted Random Numbers (154) If X is a sequence of independent random numbers with a uniform distribution (Although the method works with any distribution, for my case I used a uniform distribution.), then by the Central Limit Theorem a standard normal random variable (W) is approximated by W = Zx-E(Zx) MVIZx) where Ex is the sum of a series of n independent, uniformly distributed random numbers (over the interval 0-1) 190 E(£x) is the expected value of the sum E(2x) nE(x) = nfl;xdx = n(l/2) V(Zx) is the variance of the sum (For independent measurements the variance of a sum is equal to the sum of the variances.) nV(x) = n[ExZ-(Ex)2] n[{lxzdx - ({lxdx)2] n[l/3 - (1/2)2] = n(1/12) (x]+x2+x3+...+xn)-(n/2) /(n/12) V(Zx) 50, W = A standard normal random variable, W(0,1), has a mean (p) of zero and a standard deviation (0) of one. A normal variable with specified parameters, N(p,o), may be obtained using N(p,o) = OW(0,1)+p. E. Chi-Square Goodness of Fit Test (155) The chi-square test can be used to check a set of experimental data to see how well it approximates a given distribution. The test divides the data range into intervals and calculates how many data values would be expected to fall in each interval, given a particular distribution and the total number of data points. Intervals should be constructed such that expected frequencies less than 5 in any interval are avoided. The test statistic (T) is computed from _ k (oi-E1.)2 d1=1 Ei where i is the interval number k is the total number of intervals 191 Oi is the observed frequency in interval i E1 is the expected frequency in interval i. The hypothesis that the data follows the assumed distribution is rejected if the calculated value of T is greater than Ca, where Cu is the critical value Obtained from a chi-square table and is dependent upon the number Of degrees of freedom (df), and on the significance level (a). For testing a uniform distribution df=k-l. For testing a normal distribution df=k-l if the mean and standard deviation are known, and df=k-3 if they are estimated from the data. APPENDIX B SUPPLEMENTARY DATA TABLES FOR CHAPTER 5 192 enee. e_nn. enme. nnee. neen. Penn. menu. nen.m n\en enee. emee. nnen. Anmn. neee. nnne. nnee. nNF.~ n\e_ emne. nnnn. neon. ennn. eene. eenn. nann. nee.n n\nn n__n. ween. nenn. eeee. N_~n. m_nn. meme. nnn.n n\en nene. nenn. neen. _e_e. neee. nenn. _Ppe. nnee. nee._ n_\em Anne. nnnn. enee. ne_n. enee. nene. enne. nee.” n\_P n_nn. Anne. nnnn. Knee. nene. eene. onee. nNF.F n\e nnPn. nnen. e_eu. Aeen. enen. name. nmne. none. ene.P n_\e_ ween. none. “_Nn. none. Nnne. eene. Fnee. mane. eee. nn\np neee. nenn. F__e. NAee. Aeee. eene. A_ee. nan. n\e meme. ee_n. n_An. en_e. _Nne. nnee. nnne. Knee. nee. n_\n_ new“. eenn. nnee. e_ee. nnne. enne. Neee. Pnee. e_e. nn\n~ neon. e_ue. en_e. n_ne. _eme. enne. meee. _nee. enn. np\.F nPNe. nnne. neee. enne. “Nee. nnee. enee. nun. n\n e_nn. nene. nnne. nnee. nnne. meee. _nee. enee. enn. nF\e _een. n_me. eene. nnee. mnee. _eee. nnee. enee. _en. NM\A_ n_~e. Pene. neee. enee. “Nee. _nee. enee. eeee. ene. ~n\n_ nmee. nene. eene. N_ee. Knee. neee. nnee. neee. nee. n_\e Knee. neee. emne. A_ee. nnee. neee. enee. eeee. nee. ~m\ep neee. eene. NNee. nnee. neee. nnee. eeee. neee. nee. me\PP neee. neee. neee. neee. neee. neee. neee. neee. n_n. n_\n nnee. nnee. nnee. nnee. nnee. neee. neee. neee. new. me\e eeee._ eeee._ eeee._ eeee.F eeee._ eeee.P eeee.F eeee.P new. NM\A eeee.P neee.. eeee._ eeee._ eeee.P none._ eeee._ eeee.F nn_. NM\n en. en me n_ n e N _ seen annex eneeEn eeneeEn Le Lensez eenenaeEn seen: .mxnme :nanano mchuooEm LmuF< Onconnnm usmFo: .cnm annF 193 eeee.m nANR.~ nene.~ ____.~ Annn._ eene.F eeee.~ nun.~ n\n~ nnnn.N __PF.N eenn.F Annn._ neee.F NNNN._ --.F n~_.~ n\s_ meem.~ ennn._ Annn._ eeee._ --.F -N~.P neee.. nun._ n\np ennn._ Annn.F eeee._ NNN~._ -N~.P eeee.F eeee._ nun._ n\np nne_.n Neee._ Nnnn.n neee.F nnnm._ nA__._ nA_F._ eeee._ nee._ n_\n~ Annn._ neee.F N-~.F meme.” eeee.P neon._ neon.F nem.n n\_P eeee._ NNNN._ --._ eeee.F eeee.~ eeee._ eeee.P nNP.F n\e eneA._ neee._ ennm._ nnn~.F nm__._ eeee._ eeee._ none._ nne. n_\AP Nnnn.F emne._ mneN.F nA__._ eeee._ eeee.~ eeee._ eeee._ nee. n_\np NNNN._ eeee._ eeee._ eeee._ eeee._ eeee.. neee.. nun. n\A ennn._ enmm.. nAPP.F nRFP._ eeee.F eeee.~ neon.” eeee.F e_n. n_\ep nnnn.n enen._ NFNP._ none.F eeee._ eeee.F eeen._ eeee.. ens. ~e\e~ enn~.F nA_P.P nm_..F eeee.p oeee.~ eeee.F neoe._ eeee.~ nnn. n_\.P eeee.p eeee.n eeee._ eeee._ none._ eeee._ eeee.. eeee._ nen. n\n nA_F.F eeee.P eeee._ eeee._ eeee._ eeee.~ eeee.F none._ Nnn. n_\e NFNP.F nene.n neon.” enee._ none._ eeee.P eeee._ neon.P _en. me\np NPNF._ none._ eeee._ neee.~ eeee._ eeee._ eeee.F eeee.. ene. mn\n. eeee.. eeee._ eeee., eeee.~ enee._ eeee.~ eeee._ none._ nee. n_\e none._ eeee.n eeee._ eeee._ eeee._ eeee.F eeee.~ eeee._ nee. ~n\m. eeee._ eeee._ eeee.~ eeee._ neon.P eeee._ eenn.. eeee._ nee. ~e\P. none._ eeee.F eeee.F eeen._ neon.” neon._ eeee._ none.P nee. n~\n neee._ eeee.. eeee._ eeee.F eeee._ oeee.~ eeee.. eene._ _nn. me\e eeee._ none.P eeee.. eeee.F eeee.n neon.F neee.. none._ new. ~e\e eeee._ none._ eeee._ oeee.. eeee.. enee.P eeee._ none.P nnp. ~e\n n~_ en me n_ n e N _ .Tnnepn crane sneeze meanesn Le tense: eeeeneeen gene: .nxnmn :nmesno mchuooem Lmuw< unconnmm spew: .mum «Fan» 194 mFoF.F ano.F nmwo.F mnmo.F mmno.F mmmo.F ammo.F mum.~ m\m~ ammo.F Ammo.F quo.F Ammo.F emno.F memo.F NnFo.F .mNF.~ m\~F nmmo.F oeuo.F Fcno.F oomo.F ammo.F oFmo.F mmoo.F mum.F m\mF quo.F Nuno.F nmmo.F ammo.F om~0.F NNFo.F weco.F mwn.F «\mF nomo.F mFNo.F memo.F mneo.F memo.F nnFo.F mnoo.F nmoo.F mmn. mF\m~ mnno.F mmmo.F mono.F Ammo.F mmFo.F nmoo.F nFoo.F mum.F m\FF meo.F nwmo.F nemo.F mmFo.F meoo.F mFoo.F mooo.F mNF.F m\m NFno.F mmno.F memo.F memo.F Nooo.F smoo.F oFoo.F Nooo.F moo.F nF\~F FFmo.F memo.F ammo.F mNFo.F omoo.F oFoo.F coco.F Fooo.F mma. nF\mF omNo.F FmFo.F choo.F mmoo.F sooo.F Nooo.F Fooo.F mum. m\~ ammo.F nn~0.F snFo.F mnoo.F mFoo.F nooo.F Fooo.F Fooo.F mFm. nF\mF oFmo.F mmFo.F coco.F «moo.F FFoc.F mooo.F Fooo.F Nooo.F an. NM\m~ memo.F ano.F nnoo.F mFoo.F mooo.F Fooo.F Fooo.F coco.F man. nF\FF woos. omoo.F sooo.F Nooc.F Nooo.F coco.F coco.F mun. m\m ocFo.F «voo.F omoo.F mooo.F coco.F coco.F oooo.F coco.F Nnm. nF\m AFFo.F mnoo.F nFoo.F nooo.F Nooo.F Nooo.F Fooo.F coco.F me. Nm\NF mnoo.F omoo.F mFoo.F mooo.F Nooo.F oooo.F coco.F coco.F men. ~m\mF meoo.F mmoo.F mFoo.F oFoo.F mooo.F mooo.F Fooo.F Fooo.F man. nF\~ mmoo.F omoo. mooo.F Nooo.F Fooo.F coco.F coco.F coco.F mac. ~m\mF coco.F coco.F oooo.F coco.F coco.F coco.F coco.F coco.F men. NM\FF coco.F coco.F coco.F coco.F coco.F coco.F coco.F coco.F mFm. nF\m coco.F coco.F coco.F coco.F oooo.F coco.F coco.F coco.F me. ~m\m coco.F coco.F coco.F coco.F coco.F oooo.F coco.F coco.F mFN. NM\~ coco.F coco.F coco.F coco.F coco.F coco.F coco.F coco.F mmF. NM\m mNF no Nm mF m e N F 2:3“ oFunm spoosm unpoosm we Lmasaz mchpoosm squz .mxnme :nansnu mchuoosm Loum< mneonmom nmL< .nIm annF 195 ”one. nnee nenn. neon. nenn. neoe nun.n n\nn nonn. nnnn nnen. neon. noun. eo_n nnp.n n\ep nnnn. nmnn noen. n_nn. nnon. neon. nun.n n\np _n_n. _nen. none. nnee. _nen. nonn. nnn.. n\n_ _nnn. nnee. nnee. nenn. enen. nnoe. nne._ n_\nn nn_n. nnnn. menu. nnee. nnen. Penn. ee_e. nun._ n\PF Lnnn. oeee. _non. ennn. nnen. Anne. Pnne. nnp.. n\e _een. nnnn. on_n. n_nn. nnoe. oeee. nnne. eno. _ n_\n_ n_oA. nee“. neon. __en. nnne. npne. nene. nee. np\np none. Anne. nonn. enAn. nnee. nnee. nnne. neee. nan. n\R n_nn. onnn. nnnn. nnnn. enpe. nnee. nnne. oone. n_n. np\nn onnn. npnn. nenn. Anoe. nnee. none. _nue. nnne. e_e. ~n\nn nnnn. nunn. nan. enpe. nnee. nnne. oone. nnne. nnn. nn\pp neon. Rnnn. none. nnee. none. none. nnne. nnee. nnn. n\n nnon. _enn. nnne. None. Anne. e_ne. nene. _oee. nnn. np\e ennn. onen. nene. _one. Rene. onne. Noee. neee. Pen. ~n\mp nnnn. none. Anne. mane. none. nnne. _eee. nnee ene. nn\np nnee. oeee. eone. nnee. nnne. nnee. nnee. nnee. nno. n.\A onpe. none. nnne. nnee. mane. nnee. Fnee. onee. non. nn\np nnee. enne. nnee. nnne. nnee. .nee. onee. oeee. non. ~n\.~ nnee. neee. nnne. nnee. nnee. onee. oeee. neee. n_n. n_\n oene. none. nnne. nnee. nnee. onee. oeee. neee. an. ~n\e oooo. — oooo._ oooo._ oooo.~ oooo.. oooo._ oooo._ oooo._ n_~. ~n\e oooo. _ oooo._ oooo.n oooo.~ oooo.P oooo.n oooo._ oooo._ nn_. nn\n on. en me n_ n e n _ zzzn onone nooOEn neoooEn no tense: eennoern none: .nxnmn ene~ucmeon mceguoosm Lmum< oncoqnmm uszmz .mum annF 196 nnnn.n nnnn.n nnnn.n ennn.P Annn._ Annn._ nan.n n\nn nnnn.n FP_P.N ennn._ Annn.P oeee.p nnnn.n nnp.n n\A_ .P.P.~ ennn.. Annn.— eneo.F nnnn._ nnnn._ nun._ n\n_ ennn._ Annn.. oeee.. nnnn.p nnnn._ oeee.. nnn.p n\n_ nnnn.n nnnn._ none._ oeee., nnnn._ nA._._ nno.p np\nn ennn., Annn.. nono._ nnnn._ nnnn.p oooo.P oooo.p nan.P n\F_ ennn.P oeee.P nnnn.F nnnn.P oooo.~ oeee.. oooo._ nnp._ n\e enoA.P nouo.P ennn._ nnnn.P nR._._ oooo._ oooo.. nno._ nP\AP nnnn._ ennn.. nnnn._ ne_.._ nep_._ oooo._ oooo.. nee. np\n_ oeee._ oeee.F nnnn._ nnnn.n oooo._ oooo.P oooo.n oooo._ nan. n\R nnnn._ oeee., nnnn._ nA_P.P nepp._ oooo._ oooo._ oooo.n n_n. n.\n_ enon._ nnnn.~ enon.. npnp., nono., nono.P oooo.. oooo._ e.A. nn\nn nose._ nnnn.F nnnn.p nepp._ oooo._ oooo.. oooo._ oooo.P nnn. nP\PF nnnn._ oooo._ oooo.o oooo._ oooo.~ oooo.F oooo._ oooo._ nnn. n\n ennn.. nR_F._ nA_F.P oooo.F oooo.p oooo.P oooo._ oooo.F nnn. nP\e onon._ n_n_.P n_n_._ nono.P oooo._ oooo.F oooo._ oooo._ Pen. nn\e_ npn_._ npn_.F nono.n oooo._ oooo._ oooo.P oooo._ oooo._ ene. nn\n. nAPP.. oooo._ oooo., oooo.F oooo.. oooo.~ oooo._ oooo._ nee. n_\A n_~_.. nono.. oooo.n oooo._ oooo._ oooo._ oeee.. oooo._ non. ne\n_ nono.F oooo., oooo._ oooo.P oooo.P oooo.F oooo._ oooo.p non. nn\__ oooo.P oooo., oooo.p oooo._ oooo.. oooo._ oooo._ oooo.P n_n. n_\n oooo.~ oooo.P oooo.F oooo._ oooo._ oooo.p oooo._ oooo._ Fen. nn\e oooo._ oooo.. oooo._ oooo.n oooo._ oooo.. oooo.F oooo._ npn. nn\e oooo._ oooo._ oooo._ oooo.P oooo._ oooo.P oooo._ oooo._ nnp. nn\n en. on me n, n e n p zzzn canoe snooze nneeoEn Lo tense: econoeoEn none: .nxnnn anNucoLon acnsuooem Lmum< uncommon canF: .mam annF 197 vao.F oeoo.F coco.F mooo.F Fooo.F Fooo.F mmw.~ m\m~ «Foo.F coco.F maaa. maaa. waaa. aaaa. mNF.~ w\FF nooo.F oFoo.F mooo.F mooo.F Nooo.F oooo.F muw.F m\mF eaaa. Nooo.F Nooo.F Nooo.F naaa. oooo.F mam.F m\m~ Faaa. maaa. Faaa. Faaa. oaaa. oaaa. wmv.F mF\mm wnaa. oaaa. coco.F coco.F mooo.F mooo.F oooo.F mum.F m\FF omaa. oooo.F mooo.F nooo.F coco.F Nooo.F oooo.F mNF.F m\a Fwaa. amaa. vaaa. oaaa. maaa. naaa. maaa. moo.F oF\nF waaa. waaa. maaa. maaa. caaa. maaa. maaa. mma. oF\mF aNoo.F wFoo.F aooo.F nooo.F coco.F Nooo.F Fcoo.F aaaa. mum. w\n coco.F aooo.F nooo.F coco.F naaa. maaa. maaa. maaa. mFm. oF\mF mwaa. mmaa. mwaa. mmaa. Fwaa. Fwaa. omaa. Fmaa. aFn. NM\m~ nFoo.F mooo.F Nooo.F Fooc.F waaa. Faaa. maaa. maaa. wwo. oF\FF coco.F eFoo.F mFoo.F oFoo.F mooo.F mooo.F Fooo.F oooo.F mum. w\m aeoo.F «Foo.F mooo.F mooo.F aaaa. waaa. Faaa. naaa. mom. oF\a mmoo.F oooo.F maaa. Naaa. amaa. wmaa. smaa. mwaa. me. NM\FF aFoo.F Nooo.F maaa. eaaa. oaaa. wwaa. nmaa. Fwaa. amc. NM\mF mmoo.F omoo.F mooo.F Nooo.F coco.F aaaa. waaa. maaa. mac. mF\F FFoo.F mooo.F maaa. naaa. Naaa. oaaa. oaaa. amaa. mow. NM\mF cooc.F maaa. nmaa. maaa. Naaa. Faaa. Faaa. Faaa. men. ~m\FF maaa. msaa. mmaa. vaaa. naaa. aaaa. maaa. maaa. mFm. mF\m mmaa. onaa. mmaa. swaa. amaa. Faaa. Naaa. Naaa. me. NM\a vaaa. vaaa. eaaa. vaaa. eaaa. eaaa. caaa. caaa. mFN. Nm\n oaaa. oaaa. maaa. oaaa. oaaa. oaaa. maaa. waaa. omF. Nm\m wNF cm mm oF w v N F 2:3; onnm spoosm mcuoosm mo Lmnszz achuooEm :uqu .mxnma :nFNucmLoq a:F:HOOEm Lmuw< mm:OQmmm nos< .aum annF APPENDIX C ROUTINE OPERATION PROGRAM LISTINGS Program VCIDCO. FT OGOOOOOOOOOOOOOOOOO GUIUIU‘U‘DU‘IWUIJI(n01MLIWCTUIMMMMMMMMMMMMWWMMO 999 FILENAME I VCIDCOoFT VIDICON DATA COLLECTION AND OUTPUT PROGRAM. 198 ALLOWS CHARGE INTEGRATION DY STOPPING THE READOUT BEAM. TIMOTHY A. NIEMAN JANUARY 24.1975 THIS PROGRAM VAS VRITTEN FOR THE ENKE GROUP PDP B/I PROGRAMMABLE REAL-TIME CLOCK. EAE. H-Y DISPLAY SYSTEM. DECTAPE. AND LINEPRINTER. VITH I2K MEMORY. CALLS SUBROUTINE XYSYS.AXIS.SMOTH. DATA COLLECTION CODES! I-DATA GOES IN ISAMP ARRAY 2-DATA GOES IN IDARK ARRAY A-DATA GOES IN IREF ARRAY GINO MORE DATA.GO TO OUTPUT SECTION. COMMON IVORK.ISAMP.IREF.IDARK.NPT' DIMENSION IVORKCSIR).ISAMPISIR).IDARK(S12).IREFCSIZ) OPDEF TADI IAGO [INDIRECT TAD TO FOOL SABR 0PDEF DCAI GAO. [INDIRECT TAD TO FOOL SABR 0PDEF STORE 6.51 [PUT DISPLAY SCOPE IN STORAGE MODE. 0PDEF KLDAD 6'6! [LOAD X COORDINATE INTO SCOPE. 0PDEF YLP 6'66 [LOAD Y COORDINATE AND PLOT POINT. 0PDEF LPSET 6|2| [LOAD CLOCK PRESET REG FROM AC 0PDEF CLCIC 6IRA [CLEAR CLOCK AND INITIALIZE 0PDEF LCTRL 6I3I [LOAD CLOCK CONTROL REG FROM AC SKPDF SKPOF 6|32 [SKIP ON CLOCK OVERFLOU SKPDF SKOFE 6133 [SKIP ON OVERFLOV ERROR FLAG 0PDEF CLOFE 613A [CLEAR OVERFLOV AND ERROR FLAGS 0PDEF CDFI 620! [CHANGE TO DATA FIELD I 0PDEF CDFI 62!! [CHANGE TO DATA FIELD I 0PDEF CLCF 6AAI [CLEAR AID CONVERTER FLAG 6 TRIGGER FLAG SKPDF CHCF 6442 [CHECK CONVERTER FLAG 0PDEF GATE 6444 [GATE DRIVER 0PDEF GDCC 6AAS [GATE DRIVER.CLEAR CONVERTER FLAG SKPDF CHVF 6AS| [CHECK VERTICAL SCAN FLAG 0PDEF CLVF 6652 [CLEAR VERTICAL SCAN FLAG SKPDF CHTF 6ASA [CHECK EXTERNAL TRIGGER FLAG 0PDEF VLTCH 6C6! [LATCH WAVELENGTH 0PDEF VCNTC 6‘62 [VAVELENGTH COUNTER CLEAR OPDEF UCNTS 646A [VAVELENGTH COUNTER SET 0PDEF VCNTL 6466 [VAVELENGTH COUNTER LOADCCLEAR THEN SET) 0PDEF USON 6ATI [TURN ON WAVELENGTH SCAN. 0PDEF VSOFF 6472 [TURN OFF HAVELENGTH SCAN. 0PDEF DVI 1497 [DIVIDE ACAMO BY NEXT LOCATION CONTENTS 0PDEF MQL 7A2! [CLEAR MO. LOAD FROM AC. CLEAR AC 0PDEF MOA TSOI [PUT OUOTIENT IN AC 0PDEF LAS 7604 [LOAD AC FROM SR 0PDEF CMDA 1761 [CLEAR AC. LOAD AC FROM MO VRITE(1.12I) FORMATCI'DATA INPUT CODES!'I'IISAMPLE'l'zflDARK SIGNAL'I I'A-REFERENCE‘I'S'NO MORE DATA. GO TO OUTPUT'J VG!TE(I.ICI) FORMAT([’OUTPUT CODES!'I'SINOTHING'I'I-SAMPLE'I I'2-REFERENCE'I'J'ST‘I'A-ADSORDANCE‘I'S-VRITE FILE'I 2'6'STANDARD DEV.'I’T-LOCATE’I'S'EVALUATE'I'9-RETURN'I) CONTINUE S NPTL.CLA CLL READ(I.III)NPT FORMATI‘POINTS PER SPECTRUM 0 '.IS) Cimmmfimmmmm‘nmwmmInb‘mI-AC‘IIAMMUIUIMMUIM mmmmmmmmmwmmmmcnmmmmmmmm U! OK. ROT. NPT. MNPT. TmP. CODE. OUT. 199 CLA CLL TAD \NPT [GET POINTS/SPECTRUM DCA up? [STORE IT AVAY. TAD up? mo ("II was: our WACCEPTAGLE umaEns. SZA ' [VALID NUMBER or POINTS? JMP on [YES d on GENERATE conE. CLA CLL IND. run (201 TLS . [RING TTY BELL CLA CLL JMP NPTL [TRY AGAIN. CLA CLL TAD up? CIA I-NPT IAC [IN ORDER TO HAVE ENOUGH TIME TO RESET POINTERS IAC [AND COUNTERS :1 ts NECESSARY TO 5x1? 2 POINTS. DCA INPT [STORE INVERSE or NPT ron USE as COUNTER. CLA CLL ran up? RAR onvon 37 THO. DCA HNPT [STORE IT ran USE DY cooE GENERATOR. CLA CLL NPT-NPT-Z I [THE FOLLOVING SETS THE PROPER CONTROL [GIT FOR THE VAVELENGTN COUNTER SO IT HILL GENERATE NPT [UAVELENGTH POINTS PER SPECTRUM. [ CLA CLL TAD HNPT DCA TEMP [INITIATE TEMP CLA CLL STL [INITIATE CODE RAL [SHIFT THE CODE BIT. DCA CODE TAD TEMP [GET HNPT RAL SZL [IS MSGOIT JMP OUT [YES. CODE IS GENERATEO DCA TEMP [NO CLA CLL TAD CODE [GET CODE READY TO SHIFT AGAIN JMP ROT [GO AND SHIFT AGAIN - GGGG [CONTAINS THE NUMBER OF POINTS/SPECTRUM GOGG [CONTAINS I[2 OF THE POINTS[SPECTRUM GGGG [CONTAINS SHIFTED VERSION OF HNPT GGGG [CONTAINS CONTROL VORO FOR WAVELENGTH COUNTER CLA CLL [ALL DONE. CODE IS CORRECT. TAD CODE VLTCH [LATCH THE CONTROL VORD TO THE COUNTER VCNTC [CLEAR THE VAVELENGTH COUNTER. VSON [TURN ON VAVELENGHT SCAN. I DENT. CLA CLL 99G UIUIUIUIMUOUIUIUIUDUIMUIUI DO 998 I-I.SI2 IVORKIIIOO CONTINUE READII.II2)IVHIC FORMATI'DATA CODE I '.II) CLA CLL TAD \IVHIC [GET DATA CODE DCA UHICH TAD VHICH [GET DATA CODE. AND IOGOI SEA [IS THE CODE A I FOR SAMPLE? JMP SAMPL [YES - GO SET UP SAMPLE STORAGE POINTER. CLA CLL [NO. CHECK AGAIN. TAO VHICH [GET CODE AGAIN. AND (IGOR SEA [IS THE CODE A 2 FOR DARK SIGNAL? JMP DARK [YES . GO SET UP DARK SIGNAL STORAGE POINTER. CLA CLL IND. CHECK AGAIN. TAD VHICH [GET CODE AGAIN. AND (GGO4 SEA JMP‘REF TAD VHICH AND (GOIG SZA JMP EXTND CLA CLL TAD (2'1 TLS JMP IDENT S'AMPL. CLA CLL 200 IS THE CODE A 4 FOR REFERENCE? [YES - GO SET UP REFERENCE STORAGE POINTER. [GET CODE AGAIN [IS THE CODE A G FOR EXTENDED OUTPUT? [YES IND. ILLEGAL DATA CODE. [RING TTY BELL. [TRY TO INPUT PROPER DATA CODE THIS TIME. VRITEII.II3) FORMATI'SAMPLE') DO 997 I-IoSIR ISAMPII’IO S CLA CLL S TAD (IZOD S DCA VHICH S S IIG [SAMPLE STORAGE STARTS AT IIZGGo JMP SCANS CLA CLL VRITEII.II4) II4 FORMATI'DARK SIGNAL') . DO 996 I-I.SI2 IDARKIIIIG CLA CLL TAD (3200 DCA VHICH JMP SCANS CLA CLL VRITEII.IIS) IIS FORMATI‘REFERENCE') DO 995 I-IoSIZ IREFII)OG CLA CLL TAD (2200 DCA UHICH JMP SCANS SCANS.CLA CLL READII.II6IN II6 FORMATI'SPECTRA TO AVERAGE O CLA CLL TAD \N SNA JMP IDENT _DCA N TAD N CIA DCA M CLA CLL (READCIollTININT IIT FORMAT('INTEGRATION PERIODS 0 IFCNINT’GG.GG.39 GB NINT-G 39 CONTINUE [DARK STORAGE STARTS AT I32... “MIRIAM 99S [REFERENCE STORAGE STARTS AT I22GG. .0 mcnoam '.IS) [ARE ANY AVERAGES UANTED? [YES‘ SET UP TO AVERAGE N SPECTRA. [-N , [STORE THE NEGATIVE OF N TO SAVE TIME. MMMMUIUDUIUIUI '.ID) 5 CLA CLL S TAD \NINT [GET THE NUMBER OF INTEGRATION PERIODS. S CMA S DCA NINT S JMP INIT [GO TO IT. 5 [ S [HERE UE FINALLY GET TO TAKE SOME DATA. S [ 5 PAGE S NINT. GOGG [COUNTER FOR INTEGRATION PERIODS. S VHICHoGGGG [STARTING ADDRESS OF DATA STORAGE. S INPT. OOOG [NEGATIVE OF POINTS PER SPECTRUM S N. GOOO [NUMBER OF SIGNAL AVERAGES S HIGH. DOGS .[ADDRESS OF MOST SIGNIFICANT HALF OF DATA S LOU. GOG. [ADDRESS OF LEAST SIGNIFICANT HALF OF DATA S COUNT.OGOG [COUNTER FOR POINTS PER SPECTRUM S M. GIGS I-N INIT. CLA CLL TAD UHICH DCA HIGH TAD (GOG DCA LOU TAD INPT DCA COUNT TAD \NINT SZA CLA JMP INTGR CLVF START.CHVF JMP START CLVF CLEARoCLCF INPUT.CHCF JMP INPUT CLA CLL GDCC CDFI TADI LOU DCAI LOU RAL TADI HIGH DCAI HIGH CDFG ISZ LOU ISZ HIGH ISZ COUNT JMP INPUT ISE M JMP INIT CLA CLL TAD INPT DCA COUNT TAD UHICH DCA HIGH TAD (2CD DCA LOU CDFI CLA CLL TAD N CDFI DCA DIVIS TADI LOU MOL TADI HIGH DVI DIVIS.GGGG CLA CLL MOA TAD (AGGO CLL ' DCAI HIGH ISZ LOU ISZ HIGH ISZ COUNT JMP NEXT CDFG JMP IDENT INTGR.CLA CLL TAD \NINT CIA DCA NINT CLVF CHVF JMP INTI CLVF USOFF ISE NINT JMP INTI CHVF JMP'INTR CLVF USON JMP CLEAR 201 [SET UP HIGH ADDRESS [SET UP LOU ADDRESS [- I POINTS/SPECTRUM [ARE ANY INTEGRATIONS UANTED? [YES - GO AND INTEGRATE [CLEAR VERTICAL SCAN FLAG [CHECK VERTICAL SCAN FLAG [CLEAR VERTICAL SCAN FLAG [CLEAR A/D FLAG [CHECK AID FLAG [GATE DRIVER. AND CLEAR CONVERTER FLAG. [DATA FIELD I [ADD TO SUM OF PREVIOUS DATA [PUT CARRY INTO HIGH UORD [DATA FIELD G [INCREMENT THE STORAGE ADDRESSES. [HAVE ENOUGH POINTS BEEN TAKEN? [NO - GO TAKE ANOTHER POINT. [YES. HAVE ENOUGH SPECTRA BEEN AVERAGED? [NO. TAKE NEXT SCAN TO AVERAGE [YES. DIVIDE SPECTRA SUM BY N [INITIALIZE COUNTER AGAIN. [INITIALIEE DATA ADDRESSES AGAIN. [DATA FIELD I [GET NUMBER OF AVERAGES TAKEN. [DATA FIELD I [DEPOSIT IT FOR USE AS DIVISOR. [LOAD MG [DIVIDE [PUT GUOTIENT IN AC [CONVERT TO 2'5 COMPLEMENT [INCREMENT ADDRESS COUNTERS. [HAVE ALL POINTS BEEN DIVIDED? [NO - GO DIVIDE NEXT ONE. [YES. CHANGE DATA FIELDS. [SEE UHAT THE NEXT THING IS TO DO. [INITIALIZE COUNTER [CLEAR VERTICAL SCAN FLAG. [CHECK FOR VERTICAL SCAN FLAG [INTEGRATE [ENOWH I NTmRATI ONS DONE? [NO - UAIT LONGER. [YES - TAKE DATA UHEN NEXT SCAN STARTS. [TURN ON SCAN. [RETURN TO DATA ACOUISITION. OOOMOOO 4GB 4GI 402 4'3 4G4 4OS 496 4'? 4GB 4IG GOI 200 I02 ISO 000 00000 203 IIG 303 3I2 394 GGS 2I3 202 OUTPUT SECTION EXTND.CLA CLL SUBTRACT DARK SIGNAL DRUG... . DO 4GB III.NPT DAVGIDAVGOFLOAT(IDARKCII) DAVGIDAVG/FLOAT‘NPT) FSPO$.2547. FSNEG-‘EGAT. DO AIB I-I.NPT S'ISAMP‘I) 'R-IREFII) D-IDARK(I) SIS-DODAVG R-R-DODAVG IFIS‘ESNEG)4GI.4G2.4GE S'FSNEG. IFIS-FSPOS)4G4.4G4.4GG SIFSPOS IF‘R-FSNEG’4G5.4G6.4G6 R-FSNEG IFIR-FSPOS)4GB.4GG.4GT R'FSPOS ISAMPCIIOS IREFIIIUR READIIalGIIIA FORMATI'LIST '.IS) ICODE'I IFCIA)2.2.GBI ' GO TO (2‘I.202.BG3.2G4.50G.6GG.TGG.BBG.999).IA URITEIG.IGR)‘IUORK(I).I-I.NPT) FORMATI/(IX.IG(IS.SKI)I GO TO I READCI.IGG)IA FORMATI'PLOT '.IS) ICODE'E IFIIA)I.I.3BI SAMPLE DO 2II III.NPT IUORKII)IISAMP(I) GO TO IZOD.GGG)ICODE REFERENCE ST READII.IIO)ITR FORMATC'THRESHOLO I ‘.IS) THR-ITH DO 2I3 III.NPT R-IRBFCI) SIISAMPII) IF!(R-S)-THR)304.3I2.3I2 T-lOOO.t¢S-DAVO)[(R-DAVG) IVORK(I)-T IF2.2.2o -MI(ISIZE+I)[2 NINPT-M CALL SMOTHIIUORK.NPT.ISIZE) GO TO 3 GOO MUD ~94 mm "H ~— ~. 500 120 IIS [23 501 II9 mmmmm UIUIUIUIUIU'D 502 503 GOO 000 700 124 70! 702 000 80! 802 I25 205 ROUTINE TO OUTPUT A DATA FILE TO STORAGE DEVICE READII.I20)FNAME FORMATI'FILE NAME '.A6) CALL OOPEN('SYS'.FNAME) VRITEI4.II8)NPT' FORMATIIS) URITEII.123)DAVG FORMATIFIO.4) DO 50I III.9 URITEI4.II8)NINT URITEII.II9) FORMATI/‘ENTER IDENTIFYING TEXT'I'CONTROL G TO END'[) CLA CLL KSF - [KEYBOARD STRUCK YET? JMP TEI2 [NO TRY AGAIN KRB [READ CHARACTER TLS " [ECHO DCA \IA CLA CLL URITEI4.II8)IA CLA CLL TAD \IA TAD (TSTI [UAS THE CHARACTER A CONTROL G? SZA JMP TEII IND. READ ANOTHER CHARACTER. CLA CLL [YES. STOP TAKING TEXT. DO 502 I=|.NPT VRITEI4.II8)ISAMP(I) DO 503 I=l.NPT VRITEI4.IIB)IREF(I) CALL OCLOSE GO TO I LOCATE READ(|.|24)LOC FORMATI‘POINT '.xs> xrtLoc>1aa.1ao.7a| IFINPT-LOC)?OO.702.702 IXIXSCALEIFLOAT(LOCI- xv-o CALL xvsr¢xx.xv) IYI(FLOAT(IVORK(LOC))-SMIN)IYSCALE CALL XYPLT¢IX.IY) CALL XYEND GO TO I EVALUATE READIIII243LOC IFCLOCIGUGIDGB.CGI IF‘NPT-LOC)BGD.GG2.802 S'ISAHPILOC) M-IREFCLOC) SIS'DAVG RIR'DAVG T‘S/R RI-OGDOGOALOGIT) URITE‘I.I25)S.R.T.A FORMATC'SI '.FI004.T RI '.FI004.' TI '.F805.' GO TO I END A- :.ra.5) 206 Program VSIGAV.FT OOOOOOOOOOOOOOOOOOOOO nmmmmmmuzmmmmmmmmmwmmmmmmmmmmmmmmo FILENAME I VSIGAV.FT VIDICON DATA COLLECTION AND OUTPUT PROGRAM. ALLOUS CHARGE INTEGRATION BY STOPPING THE READOUT BEAM. EXTRA BITS ACCUMULATED BY SIGNAL AVERAGING ARE RETAINEDIEVEN IF NOT SIGNIFICANT). TIMOTHY A. NIEMAN DECEMBER I2. I974 THIS PROGRAM UAS URITTEN FOR THE ENKE GROUP PDP 0/I UITH 12K MEMORY. PROGRAMMABLE REAL-TIME CLOCK. EAE. X-Y DISPLAY SYSTEM. DECTAPE. AND LINEPRINTER. CALLS SUBROUTINE XYSYS.AXIS.DBL2.PTPLT. DATA COLLECTION CODESI IIDATA GOES IN ISAMP ARRAY 2IDATA GOES IN IDARK ARRAY 4IDATA GOES IN IREF ARRAY GINO MORE DATA.GO TO OUTPUT SECTION. COMMON JDARK.IDARK.JSAMP.ISAMP.JREF.IREF.NPT.NEXP.N.DAVG.NORMD I.NORMS DIMENSION JDARKI5I2).IDARK(SI2).JSAMP(SI2).ISAMP(SI2) DIMENSION JREFISI2).IREF(SI2) OPDEF TADI I400 [INDIRECT TAD TO FOOL SABR OPDEF DCAI 3400 [INDIRECT TAD TO FOOL SABR OPDEF STORE 6057 [PUT DISPLAY SCOPE IN STORAGE MODE. OPDEF XLOAD 606I [LOAD X COORDINATE INTO SCOPE. OPDEF YLP 6066 [LOAD Y COORDINATE AND PLOT POINT. OPDEF LPSET 6I2| [LOAD CLOCK PRESET REG FROM AC OPDEF CLCIC 6I24 [CLEAR CLOCK AND INITIALIZE OPDEF LCTRL 6I3I [LOAD CLOCK CONTROL REG FROM AC SKPDF SKPOF 6I32 [SKIP ON CLOCK OVERFLOU SKPDF SKOFE 6IGG [SKIP ON OVERFLOU ERROR FLAG OPDEF CLOFE 6I34 [CLEAR OVERFLOU AND ERROR FLAGS OPDEF CDFG 620I [CHANGE TO DATA FIELD 0 OPDEF CDFI 62II [CHANGE TO DATA FIELD I OPDEF CLCF 644I [CLEAR A/D CONVERTER FLAG A TRIGGER FLAG SKPDF CHCF 6442 [CHECK CONVERTER FLAG OPDEF GATE 6444 [GATE DRIVER OPDEF GDCC 6445 [GATE DRIVER.CLEAR CONVERTER FLAG SKPDF CHVF 64SI [CHECK VERTICAL SCAN FLAG OPDEF CLVF 6452 [CLEAR VERTICAL SCAN FLAG SKPDF CHTF 6454 [CHECK EXTERNAL TRIGGER FLAG OPDEF ULTCH 646I [LATCH UAVELENGTH OPDEF UCNTC 6462 [UAVELENGTH COUNTER CLEAR OPDEF VCNTS 6464 [WAVELENGTH COUNTER SET OPDEF UCNTL 6466 [UAVELENGTH COUNTER LOADICLEAR THEN SET) ‘ OPDEF VSON 647I [TURN ON UAVELENGTH SCAN. OPDEF VSOFF 6472 [TURN OFF UAVELENGTH SCAN. OPDEF DVI 7407 [DIVIDE ACAMO BY NEXT LOCATION CONTENTS OPDEF MOL 742I [CLEAR MO. LOAD FROM AC. CLEAR AC OPDEF MOA 750I [PUT OUOTIENT IN AC OPDEF LAS 7604 [LOAD AC FROM SR OPDEF CMOA 770I [CLEAR AC. LOAD AC FROM MO 207 READII.800)SATD 800 FORMATI'SATURATION I '.FI2.2) 999 CONTINUE S NPTL.CLA CLL READII.III)NPT III FORMAT('POINTS PER SPECTRUM I '.IS) S CLA CLL S TAD \NPT [GET POINTS/SPECTRUM S DCA NPT [STORE IT AUAY. S TAD NPT 5 AND (I700 [MASK OUT UNACCEPTABLE NUMBERS. S SZA [VALID NUMBER OF POINTS? 5 JMP OK [YES - GO GENERATE CODE. 5 CLA CLL [NO. S TAD (207 ' S TLS [RING TTY BELL S CLA CLL ‘ S JMP NPTL [TRY AGAIN. 5 OK. CLA CLL . S TAD NPT 5 CIA l-NPT S IAC [IN ORDER TO HAVE ENOUGH TIME TO RESET POINTERS S IAC [AND COUNTERS IT IS NECESSARY TO SKIP 2 POINTS. S DCA INPT [STORE INVERSE OF NPT FOR USE AS COUNTER. S CLA CLL , S TAD NPT 5 RAR [DIVIDE BY TUO. S DCA HNPT [STORE IT FOR USE BY CODE GENERATOR. S CLA CLL NPTINPT-2 S [ S [THE FOLLOUING SETS THE PROPER CONTROL 5 [BIT FOR THE UAVELENGTH COUNTER 50 IT UILL GENERATE NPT S [UAVELENGTH POINTS PER SPECTRUM. S [ S CLA CLL S TAD HNPT S DCA TEMP [INITIATE TDIP S CLA CLL 5 STL [INITIATE CODE S ROT. RAL [SHIFT THE CODE BIT. S DCA CODE S TAD TEMP [GET HNPT S RAL S SZL [IS MSBIIT S JMP OUT [YES. CODE IS GENERATED S DCA TEMP [NO 5 CLA CLL S TAD CODE [GET CODE READY TO SHIFT AGAIN S JMP ROT [GO AND SHIFT AGAIN S NPT. 0000 [CONTAINS THE NUMBER OF POINTS/SPECTRUM S HNPT. 0000 [CONTAINS I/2 OF THE POINTS/SPECTRUM S TEMP. 0000 [CONTAINS SHIFTED VERSION OF HNPT 5 CODE. 0000 [CONTAINS CONTROL UORD FOR UAVEENGTH COUNTER S OUT. CLA CLL [ALL DONE. CODE IS CORRECT. S TAD CODE S ULTCH [LATCH THE CONTROL UORD TO THE COUNTER S UCNTC [CLEAR THE UAVELENGTH COUNTER. S USON [TURN ON UAVELENGHT SCAN. S IDENT.CLA CLL 208 'READ(I.II2)IVHIC II2 FORMAT('DATA CODE I '.II) S CLA CLL S TAD \IVHIC [GET DATA CODE S DCA VHICH ' S TAD VHICH [GET DATA CODE. 5 AND (000I 5 '52A [IS THE CODE A I FOR SAMPLE? S JMP SAMPL [YES - GO SET UP SAMPLE STORAGE POINTER. S CLA CLL IND. CHECK AGAIN. S TAD VHICH [GET CODE AGAIN. S AND (0002 S SZA [IS THE CODE A 2 FOR DARK SIGNAL? S JMP DARK [YES - GO SET UP DARK SIGNAL STORAGE POINTER. S CLA CLL [NO. CHECK AGAIN. S TAD VHICH [GET CODE AGAIN. S AND (0004 S SZA [IS THE CODE A A FOR REFERENCE? 5 JMP REF [YES - GO SET UP REFERENCE STORAGE POINTER. S TAD VHICH [GET CODE AGAIN 5 AND (00I0 S SZA [IS THE CODE A 8 FOR EXTENDED OUTPUT? S JMP EXTND [YES . S CLA CLL [NO. ILLEGAL DATA CODE. S TAD (207 ‘ S TLS [RING TTY BELL. S JMP IDENT [TRY TO INPUT PROPER DATA CODE THIS TIME. 5 SAMPL.CLA CLL ’ VRITE(I.II3) II3 997 VOUIU'OUIUI IIA 996 MMUIUIUI II6 WML1MUIUIUIMUI II? DARK. ronuarc'snane'z no 991 I-I.512 annp¢x)-o xsnnP-o CLA CLL TAD (2200 DCA uuxcn an? SCANS CLA CLL unxrscx.lna> FORMAT(‘DARK SIGNAL') no 996 t-n.512 ' JDARK(I)-0 onnxcx)-a CLA CLL [SAMPLE STORAGE STARTS AT I2200. .TAD (020a DCA VHICH [DARK STORAGE STARTS AT [0200. JMP SCANS CLA CLL VRITE(I.II5) FORMAT('REFERENCE') DO 995 III.5|2 ' JREF(I)'0 IREF(I)-0 CLA CLL TAD (4200 DCA VHICH JMP SCANS [REFERENCE STORAGE STARTS AT I4200. SCANS.CLA CLL READ(I.II6)N FORMAT('SPECTRA TO AVERAGE ' '.I5) CLA CLL TAD \N SNA [ARE ANY AVERAGES VANTED? JMP IDENT DCA N [YES- SET UP TO AVERAGE N SPECTRA. TAD N CIA l-N DCA M [STORE THE NEGATIVE OF N TO SAVE TIME. CLA CLL READ(I.II7)NINT FORMAT('INTEGRATION PERIODS - '.I5) Uim‘nUIMtnUImtnuzmcnuzmcnuamcnuzmcnuzmtnuzmcnuwmcnuzmcnuIMIn(nu:mcnu:mtnuamcnuamcnuxmtnUImtnurm(nuzmInuamtnuamcnuwm 38 39 NINT. IF(NINT)38.39.39 NINT-0 CONTINUE CLA CLL TAD \NINT _CMA DCA NINT JMP INIT [ 209 [GET THE NUMBER OF INTEGRATION PERIODS. [GO TO IT. [HERE VE FINALLY GET TO TAKE SOME DATA. [ PAGE 0000 VHICH.0000 INPT. N. HIGH. LOW. 0000 0000 0000 0000 COUNT.0000 M. INIT. 0000 CLA CLL TAD VHICH DCA HIGH TAD HIGH TAD (I000 DCA LOV TAD INPT DCA COUNT TAD \NINT SZA CLA JMP INTGR CLVF START.CHVF JMP START CLVF CLEAR.CLCF INPUT.CHCF JMP INPUT CLA CLL GDCC CDFI TADI LOV DCAI LOU RAL TADI HIGH DCAI HIGH ‘CDFO ISZ LOU 152 HIGH ISZ COUNT JMP INPUT ISZ M JMP INIT CLA CLL CDFO JMP IDENT INTGR.CLA CLL INTI. INT2. TAD \NINT CIA DCA NINT CLVF CHVF \ JMP INTI CLVF VSOFF ISZ NINT JMP INTI CHVF JMP INT2 CLVF VSON JMP CLEAR [COUNTER FOR INTEGRATION PERIODS. [STARTING ADDRESS OF DATA STORAGE. [NEGATIVE OF POINTS PER SPECTRUM [NUMBER OF SIGNAL AVERAGES [ADDRESS OF MOST SIGNIFICANT HALF OF DATA [ADDRESS OF LEAST SIGNIFICANT HALF OF DATA [COUNTER FOR POINTS PER SPECTRUM I-N [SET UP HIGH noonass [SET UP LOV ADDRESS [- I POINTS/SPECTRUM [ARE ANY INTEGRATIONS VANTED? [YES - GO AND INTEGRATE [CLEAR VERTICAL SCAN FLAG [CHECK VERTICAL SCAN FLAG [CLEAR vsarxan SCAN FLAG' [CLEAR AID FLAG [CHECK AID FLAG [GATE DRIVER. AND CLEAR CONVERTER FLAG. [DATA FIELD I [ADD TO SUM OF PREVIOUS DATA [PUT CARRY INTO HIGH UORD [DATA FIELD 0 [INCRDIDIT THE STORAGE ADDRESSES. [HAVE ENOUGH POINTS BEEN TAKEN? [NO - GO TAKE ANOTHER POINT. [YES. HAVE ENOUGH SPECTRA BEEN AVERAGED? [NO. TAKE NEXT SCAN TO AVERAGE [YES. DIVIDE SPECTRA SUM BY N [DATA FIELD 0 [SEE UHAT THE NEXT THING IS TO DO. [INITIALIZE COUNTER [CLEAR VERTICAL SCAN FLAG. [CHECK FOR VERTICAL SCAN FLAG [INTEGRATE [ENOUGH INTEGRATIONS DONE? [NO - UAIT LONGER. [YES - TAKE DATA UHEN NEXT SCAN STARTS. [TURN ON scan. [RETURN TO DATA Acouxsxrxou. (100 210 OUTPUT SECTION S EXTND.CLA CLL 300 I2I 30A 30] I22 802 303 (00101000 (100 @400 I04 I05 I06 I07 000 I5 I4 I08 READ(I.I2I)IA FORMAT('I-PLOT.2-DEV.3'EVALUATE.A'RETURN '.I5) Human. 390.304 ‘ GO TO(30I.30S.B02.999).IA READ( I. I22)NORMD. NORMS FORMAT('NORMALIZE DARK SIGNAL BY '.I5[ I’NORMALIZE SAMPLE BY ’.IS) - . IF(NORMD)30I.302.302 - IF(NORMS)30I.303.303 CONTINUE DAVO-'o PUT SCOPE IN STORAGE MODE CLA CLL 6051 CLA CLL CALL DBL2(I.PMIN.I) FIND MAX AND MIN AND SCALE FOR PLOTTING PMAx-PNIN JMIN-I JMAXII DO 0 IIIuNPT CALL oaLé(x.t.|) IF(T-PMINI5.0.6 PMIN-T JHIN-I IF(PMAX-T)1.B.B PMAXIT JMAX-I courxuus wax::rmxu.anxu VRITE(I.105)PMAX.JMAX FORMAT(/‘MIN - '.r|o.a.' AT poxur '.xs> FORMAT(‘HAX - ';rn...." AT POINT 'zzsl) READCI.I06)SHIN£SMAX ‘ ' ' roanArc'scOPs sorrow - '.ra..o/'scops rap - '.t|a.o> vscALs-20a7./’ ' XN-FLOAT(NPT-I)[I0. xscALa-zaav./rL0AT xrch)nA.|..:5 PLOT AXES CALL AXIS(I0.I0) READ(I.IOB)IA FORMAT('I'FAST. RETURN-SLOV '.IS) IF(IA)999.I6.I1 211 C LINE PLOT I6 IXIXSCALEOFLOAT(I) CALL DBL2(I.T.I) IY-(T-SMIN)OYSCALE CALL XYST(IX.IY) DO I8 I32.NPT IX'XSCALEOFLOAT(I) CALL DBL2(I.T.I) YI-(T-SMINIIYSCALE IF(2047.-YI)2I.22.22 2| YI-2047. 22 IF(YI)23.24.2‘ 23 YI'G. 24 IYIYI IB CALL XYPLTCIX.IY) CALL XYEND GO TO 300 POINT PLOT 000 I7 CONTINUE DO 9 I'I.NPT IXiXSCALEiFLOAT(I) CALL DBLRCI.T.I) YI'(T'SMIN)‘YSCALE . IF(20‘7.-YI)I0.II.II I0 YI'EBAT. II IFIYI)I2.I3.I3 I2 YI'B. I3 IYIYI CALL PTPLTCIX.IYI 9 CONTINUE GO To 300 STANDARD DEVIATION 000 305 URITE(I.I23) I23 FORMAT('ENDP POINT -_') XSCALE-20AT.[FLOAT(NPT) S CLA CLL S 6032 S DCA DELAY DO 306 IENDI-I.NPT IXIXSCALE:FLOAT(IENDI) IY-O DELAI.ISZ DELAY JMP DELAI CALL PTPLT(IX.IY) KSF JMP \306 JMP \307 S DELAY.0000 306 CONTINUE 307 URITE(I.IZA)IENDI IBA FORMAT(IS) VRITE(I.I23) S CLA CLL S 6032 S DCA DELAY DO 308 I-I.NPT IEND2-NPT-IOI IX-XSCALE‘FLOAT(IEND2) IY-I S DELA2.ISZ DELAY S JMP DELA2 CALL PTPLT(IX.IY) S KSF S S (JIM MMUI JMP \308 JMP \309 308 CONTINUE 309 VRITE(I.I24)IEND2 212 CALCULATE DEVIATION 0‘30 SUN'CO sunso-a. ‘ DO 310 x-xs~0:.15~02 CALL DBL2(I.X.I) SUH=SUM+X _ 31a sunso-sunsaox.x XMEAanun/FLOArcstoz-xsunlox) Dav-(sunso/rLoArcnzuoa-xauolo1>)- DSVISQRT(DEV) VRITE(I.I25)XMEAN.DEV :25 FORMAT('MEAN - ‘.F|0.2.' ST.DEV. - '.FI2.A) XMEANIXMEAN‘(I0000.[4096.) n:v-osv:<:CCeC./AC96.) URITE(I.I25)XMEAN.DEV DEVISATD/DEV vnxtsct.aot)nsv an: FORMAT('S[N - '.rlo.2> co to 309 EVALUATE A POINT 002 CONTINUE CLA CLL 6032 J-I BII DO 003 I-J.NPT IXIXSCALEtFLOAT(I) CALL OBL2(I.Y.I) YI.(Y-SMIN)OYSCALE IF(2047.-YI)00A.005.005 800 YI'ZUCT. 00S IF(YI)000.007.007 B06 YI'0. 007 IY-YI S DELA3.ISZ DELAY S JMP DELA3 CALL PTPLTCIX.IY) S KSF 5 JMP \003 S JMP \000 003 CONTINUE GO TO 300 000 SUM-0. XMAx-Y S 6030. DO CIR J'I.NPT IX'XSCALE‘FLOAT(J) CALL DBL2(J.Y.I) SUM'SUM’Y IF(Y-XMAX)CIA.0IA.BI3 BIO KHAN-Y CIA CONTINUE S DELAA.ISZ DELAY JMP DELAA CALL PTPLT(IX.IY) KSF JMP \BI2 JMP \BIS BI2 CONTINUE BIS URITE(I.BIO)XMAX.SUM BIO FORMAT('MAX I '.FI600.' AREA 0 '.EI606) S 6032 - GO TO BII GO TO 300 END cam "(30 UIMIA M 213 Subroutine DBL2.FT FILE NAME IS DBL2.FT SUBROUTINE DBL2(I.DBL.ID) O TIMOTHY A. NIEMAN NOVEMBER 6. I974 SUBROUTINE TO TAKE DOUBLE PRECISION INTEGER VORDS FOR SAMPLE SIGNAL AND DARK SIGNAL. CORRECT FOR THE AVERAGE DARK SIGNAL . CREATE A REAL UORD EQUAL TO THE DIFFERANCE OF_THE SAMPLE AND DARK SIGNALS. AND NORMALIZE AS DESIRED. 000000000 COMMON JDARK.IDARK.JSAMP.ISAMP.JREF.IREF.NPT.NEXP.N.DAVG.NORMD I.NORHS DIMENSION JDARKISIZI.IDARKCSIZI.JSAMP(5I2).ISAMP(5I2) DIMENSION JREF(5I2).IREF(5I2) ILO'IDARK(I) IHIIJDARK(I) NORM-NORMD'I ICODE'I IF(NORM)B.7.7 B DLO'ILO DHI-IHI IF(DLO)3.A.A 3 DHI'DHI’I. A DARK'A096-tDHIODLO DBL-DARK GO TO(S.6).ID 5 ILOIISAMP(I) IHI'JSAMP(I) NORMONORMS‘I ICODE'Z IF(NORM)9.7.7 9 SLO-ILO SHI-IHI IF(SLO)I.2.2 I SHI'SHI0I. 2 SIG'A096..SHI*5LO DBL-SIG‘DARKODAVG 6 RETURN 7 CONTINUE CLA CLL TAD \NORM DCA NORM TAD \ILO 742I [LOAD MULTIPLIER OUOTIENT TAD \IHI TAIS [SHIFT RIGHT NORM. 0000 [BITS TO SHIFT DCA \IIII 750I [LOAD MO INTO AC DCA \ILO CLA CLL GO TO(8.9).ICODE END ([IU'IU'IUIU‘IU'IUIMUIUIUIUI 214 Subroutine DBL3.FT GOO QDUG FILE NAME IS DBL3.FT SUBROUTINE DBL2(I.DBL.ID) TIMOTHY A. NIEMAN FEBRUARY 22. I975 COMMON JDARK.IDARK.JSAMP.ISAMP.JREF.IREF.NPT.NEXP.N.DAVG.NORMD I.NORMS DIMENSION JDARKISIZ’.IDARKC5I20.JSAHP(5I2).ISAMPCSIZ) DIMENSION JREFISIE).IREF(5I2) DARK". DLO‘IDARK‘I) DHI'JDARKII) IFCNORMD)7.7.B IFCDLO)3.A.A DHI-DI‘IIYI . DARK-(4096.#DHI§DLO)/FLOAT(NORMD) DBL'DARK GO TO(5.6).ID ’SLOIISAMP(I) SHI-JSAMP(I) IF(SLO)I.2.2 SHI-SHIOI. SIG-(4096.*SHI+SLO)[FLOAT(NORMS) DBL-SIG-DARK+DAVG RETURN END APPENDIX D INSTRUMENT CHARACTERIZATION PROGRAM LISTINGS 215 Program VDARK5.FT FILENAME I VDARK5.FT VIDICON DATA COLLECTION AND OUTPUT PROGRAM. ALLOWS CHARGE INTEGRATION BY STOPPING THE READOUT BEAM. EXTRA BITS ACCUMULATED BY SIGNAL AVERAGING ARE RETAINED(EVEN IF NOT SIGNIFICANT). TIMOTHY A. NIEMAN FEBRUARY 22. I975 THIS PROGRAM VAS VRITTEN FOR THE ENKE GROUP PDP B/I VITH I2K MEMORY. PROGRAMMABLE REAL-TIME CLOCK. EAE. X-Y DISPLAY SYSTEM. DECTAPE. AND LINEPRINTER. CALLS SUBROUTINE DBL2. DATA COLLECTION CODES! I-DATA GOES IN ISAMP ARRAY 2-DATA GOES IN IDARK ARRAY A'DATA GOES IN IREF ARRAY BCNO MORE DATA.GO TO OUTPUT SECTION. OOOOOOOOOOOOOOOOOOOOO COMMON JDARK.IDARK.JSAMP.ISAMP.JREF.IREF.NPT.NEXP.N.DAVG.NORMD I.NORMS DIMENSION JDARK(SI2).IDARK(5I2).JSAMP(SI2).ISAMP(SI2) DIMENSION JREF(SI2).IREF(SI2) NGNNNMNEMONIC DEFINITIONS READ(I.B00)5ATD B00 FORMAT('SATURATION O '.FI2.2) 999 CONTINUE S NPTL.CLA CLL READII.III)NPT III FORMAT('POINTS PER SPECTRUM I '.IS) OAVG'OO INHICIE READ(I.I22)NORMD.NORMS READ(I.I205)IV2.IVI I205 FORMAT('CODE AFTER OUTPUT -".I5['TIMES NORMD O '.IS) READ(I.II7)NINT *iNNNSET UP CODE FOR NUMBER OF POINTS PER SPECTRUM 216 I203 SUM2-0. DO I20I ICRUDII.I0 S IDENT.CLA CLL S CLA CLL S TAD \IVHIC [GET DATA CODE 5 DCA WHICH S TAD VHICH [GET DATA CODE. 3 AND (000! S SZA [IS THE CODE A I FOR SAMPLE? S JMP SAMPL [YES - GO SET UP SAMPLE STORAGE POINTER. S CLA CLL [NO. CHECK AGAIN. S TAD VHICH [GET CODE AGAIN. S AND (0002 S SZA [IS THE CODE A 2 FOR DARK SIGNAL? 5 JMP DARK [YES - GO SET UP DARK SIGNAL STORAGE POINTER. S CLA CLL IND. CHECK AGAIN. S TAD VHICH [GET CODE AGAIN. S AND (000A 5 SZA [IS THE CODE A A FOR REFERENCE? 5 JMP REF [YES - GO SET UP REFERENCE STORAGE POINTER. S TAD WHICH [GET CODE AGAIN 5 AND (00I0 S SZA [IS THE CODE A 0 FOR EXTENDED OUTPUT? S JMP EXTND [YES 5 CLA CLL IND. ILLEGAL DATA CODE. 5 TAD (207 S TLS [RING TTY BELL. S JMP IDENT [TRY TO INPUT PROPER DATA CODE THIS TIME. 5 SAMPL.CLA CLL N-NORMS INHIC-B DO 997 I'I.512 JSAMP(I)¢0 997 ISAMP(I)=0 S CLA CLL S TAD (2200 ' S DCA WHICH [SAMPLE STORAGE STARTS AT I2200. S J”? SCANS S DARK. CLA CLL NSNORMD IVHICII DO 996 III.SI2 JDARK(I)'0 996 IDARK(I)=0 CLA CLL TAD (0200 DCA WHICH [DARK STORAGE STARTS AT I0200. JMP SCANS REF. CLA CLL DO 995 I‘I.SI2 JREF(I)-Z 99$ I?EF(I)$3 I CLA CLL ; TAU (“200 J DCA JHICH [REFERENCE STORAGE STARTS AT 14200. C JIH’ SCANS mmmmm {OOQNSET UP SOFTWARE POINTERS AND COUNTERS NOONQACQUIRE DATA 000 217 OUTPUT SECTION S EXTND.CLA CLL 300 I22 305 000 3I0 BOI I20I I202 I204 802 I800 IA-2 :wuxc-xuz FORMAT('NORMALIZE DARK SIGNAL BY '.15/ I‘NORMALIZE SAMPLE BY '.I5) ' ' IENDI-(NPT[2)-50 ' IENDZIIENDI+I00 CALCULATE DEVIATION SUM'B. SUMSQ-0. DO 3I0 I-IENDI.IEND2 CALL DBL2(I.X.I) SUM-SUM+X SUMSOSSUMSO+X*X XMEAN=SUMIFLOAT(IEND2-IENDI+I) DEVI(SUMSO[FLOAT(IEND2-IENDIOIII-(XMEAN4XMEAN) DEVISORT(DEV) XMEAN-XMEAN#(I0000.[4096.) DEV‘DEVI(I0000.[4096.) SNISATD/DEV VRITE(I.80I)DEV.SN FORMAT(FIO.4.2X.FI2.4) SUM2ISUM29SN SUM2'SUM2/I0. WRITE(I.I202)NORMD.NORMS.SUM2 FORMAT(IS.SX.IS.5X.FI2.4) IF(NORMS-I024)l204.I800.I800 NORMD-NORMDtIVI NORMSINORMS#2 GO TO I203 CONTINUE CALL EXIT END 218 Program VDRDC.FT 00000000000006 FILENAME I VDRDC.FT INCREASES THE VIDICON'S DYNAMIC RANGE BY STOPPING THE BEAM FROM READING LOU INTENSITY REGIONS UNTIL A SIGNIFICANT SIGNAL HAS ACCUMULATED. TIMOTHY A. NIEMAN MARCH 5. I975 THIS PROGRAM UAS VRITTEN FOR THE ENKE GROUP PDP 8/I VITH I2K MEMORY. PROGRAMMABLE REAL-TIME CLOCK. EAE. X-Y DISPLAY SYSTEM. DECTAPE. CALLS SUBROUTINES OUTPUT.DOUBL.XYSYS.AXIS.PTPLT.SMOTH. COMMON JDARK.IDARK.JSAMP.ISAMP.IUINT.IMINT.NPT.NEXP.N DIMENSION JDARK(SI2).IDARK(SI2).JSAMP(5I2).ISAMP(5I2) DIMENSION IVINT(SI2).IMINT(SI2) NNNNNMNEMONIC DEFINITIONS 999 WRITE(I.II9) . FORMAT(/['DATA CODES ARE.‘/5X.'I-SAMPLE'I ISX.'2-DARK SIGNAL‘ISX.'4-REFERENCE'ISX.'BIOUTPUT'II) CONTINUE {SSNDSET UP CODE FOR NUMBER OF POINTS PER SPECTRUM .....IDENTIFY INPUT 5 SCANS.CLA CLL [NORMAL DATA ACQUISITION ROUTINE. |I6 II7 38 39 READ( I. I I6)N FORMAT( 'SPECTRA TO AVERAGE ' '. I S) REAOII.II7)MNEXP FORMAT( 'MAX. NUMBER OF UPOSURES O '. I S) NTEXPIO NEXPII NINT-NEXP-I CDCNNSET UP SOFTWARE POINTERS AND COUNTERS ANGNNACQUIRE DATA 219 S DNRNG.CLA CLL [ROUTINE TO DECIDE THE RELATIVE INTENSITIES. MIAUIM¢AUIMIAUIM 524 SII SI2 5I0 5I3 52. 52] 514 I20 SIS saé 533 5:8 5I7 519 S23 DAVG'C. DO 524 III.NPT DAVGIDAVGOFLOAT(IDARK(III DAVGIDAVGIFLOAT(NPTI FSPOSI2047. FSNEGI-2047. THIFSPOS+FSNEG PMAKIFLOAT(ISAMP(I))IFLOAT(IDARK(IIIIDAVG XSCALEI2047.[FLOAT(NPTI DO 5I0 III.NPT SIISAMP(I) DIIDARK(I) IMINT(IIIMNEKP IF(PMAX-(S-D*DAVG))5II.5I2.5IR PMAXIS-DIDAVG CONTINUE IXIXSCALEIFLOAT(I) IYISID+DAVG CLA CLL CML RTR TAD \IX RAL CMA 606I CLA CLL TAD \IY CMA 6066 CLA CLL CONTINUE IF(PMAX-TH)SI3.5I3.5I4 NEXPINEXPI2 IFINEXP-MNEXP)S20.520.5I4 DO 52I III.NPT JSAMP(I)I0 ISAMP(I)I0 GO TO 39 JII READ(I.I20)TH FORMAT('THRESHOLD(X.) I '.FB.0) DO SIT III.NPT SIISAMP(I) DIIDARK(I) IF(S‘D)532.S33.533 SID IF((SIDI'TH)5I7.5I7.5IO ISAMP(I)IIDARK(I) IMINT(I)IJ CONTINUE THITH/2. JIJ*J IFIJ'MNEXP)SI5.5I5.5I6 CONTINUE DO 5I9 III.NPT IYI2047.[FLOAT(IMINTIIII IXIXSCALEIFLOAT(I) CALL PTPLT‘IX.IY) CONTINUE DO 523 III.NPT IMINTCI’I'IMINT(I) UBUIUIW 525 I18 526 527 528 U'IU'IUIUIUI'JI $29 530 \SJI. A6. A3. A7. U701mmLTMMU’UIUIUIUIUIMUTUIUIUImme‘IL‘ImmmmmmADI/7b1mmtnmbqugmmmm 220 [ [HERE STARTS THE SECTION VHICH INTEGRATES WEAK POINTS [LONGER THAN STRONG POINTS. [ READ(I.II8)IA FORMAT('DATA CODE (DR) I '.IS) IF(IA)525.525.526 . GO T0(S27.S29.525.S2S.S2S.SES.SES.ACI).IA DO 528 III.NPT IUINT(I)IIMINT(I) ISAMP(I)I0 JSAMP(I)I0 CLA CLL [SET UP SAMPLE SIGNAL STORAGE POINTERS. TAD (2I77 DCA HI TAD (3I77 DCA LO JMP \S3I D0 530 III.NPT IUINT(I)'IMINT(I) IDARK‘IIID JDARKIII=0 . CLA CLL [SET UP DARK SIGNAL STORAGE POINTERS. TAD (OI77 DCA HI TAD (II77 DCA LO JMP \53I CLA CLL TAD N CIA CLL DCA M USON [TURN ON SCAN BEAM TO START OUT ERASED. CLA CLL TAD (7776 CHVF [DELAY TO MAKE SURE SCREEN IS ERASED. PAGE [INITIALIZE POINTERS AND COUNTERS. DCA LOI [SET POINTER FOR LEAST SIGNIFICANT PART. DCA HII [SET POINTER FOR MOST SIGNIFICANT PART. TAD INPT TAD (4200 DCA VINTP TAD (S200 DCA MINTP TAD D2 DCA JUMP mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmwmmmmmmmmmmmmmmmmm A5. GO. JUMP. A4. L0. HI. CDFI CLVF CHVF JMP A5 ISZI UINTP JMP NO USON TAD DI MOL TADI MINTP DCAI UINTP HLT $006 0600 GOOD MNEXP.OGGG MINTP.GGGG UINTP.GGGO CNT. DI. DZ. N0. 8. C. DFZ. GOOD JMP A JMP D USOFF TAD 02 MOL JMP JUMP INC LOI INC L02 INC HII INC HI2 CLCF JMP C CHCF JMP A GDCC TADI DCAI RAL TADI DCAI CMOA DCA JUMP INC UINTP INC MINTP ISZ JMP GO ISZ JMP A1 USON CLA TAD DF DCA DF2 HLT ISZ M JMP A6 JMP \525 EXTND.CLA CLL AOI CALL OUTPUT GO TO 999 END 22] [CHANGE TO FIELD I FOR DATA STORAGE [CLEAR VERTICAL SCAN FLAG [CHECK FLAG TO START AT BEGINNING OF SCAN [SHOULD THIS POINT BE ACQUIRED? [NO. INTEGRATE LONGER. [YES. TURN ON THE SCAN BEAM TO TAKE THE POINT. [SET UP CODE TO TAKE POINT [TEMPORARY STORAGE IN EAE REGISTER [RESTORE WORKING VALUE FOR INTEGRATIONS. [REPLACED BY JMP A OR JMP B [COUNTER FOR NUMBER OF TIMES TO REPEAT CYCLE. [MASTER INTEGRATION POINTER. [WORKING INTEGRATION POINTER [COUNTER FOR POINTS PER SPECTRUM [TURN OFF SCAN BEAM AND LET POINT INTEGRATE. [SET UP CODE TO SHIP TAKING POINT. [TEMPORARY STORAGE IN EAE REGISTER. [SKIP DATA POINT [PUT CARRY IN HIGH ORDER UORD [TAKE JMP INSTRUCTION OUT OF EAE. [STORE IT [INCRDGEVT PO INTERS [AND COUNTERS. [HAVE VE CYCLED THROUGH INTEGRATIONS ENOUGH? [NO DO ANOTHER. [YES-DATA IS COLLECTED. ERASE SCREEN. [TIME TO CHANGE DATA FIELDS AGAIN [REPLACED BY A CDF [HAVE ENOUGH AVERAGES BEEN DONE? [NO DO ANOTHER. [SEE WHAT TO DO NEXT APPENDIX E SMOOTHING EVALUATION PROGRAM LISTINGS 222 Subroutine IRANU.FT OMMMOOOOOOOOO MMUIUIUIUIU'IUIUIUIUIUIUOUI SUBROUTINE IRANU(INIT.IRAN) GENERATES UNIFORMLY DISTRIBUTED RANDOM INTEGERS IN THE RANGE G-IGOO. BY MULTIPLICATIVE CONGRUENCE. TIMOTHY A. NIEMAN JUNE 23. I974 INIT'G INITIALIZES THE GENERATOR IRAN CONTAINS THE RANDOM NUMBER OPDEF MUY TAGS [MULTIPLY OPDEF MOL 742I [LOAD AC INTO MO OPDEF MOA TTGI [LOAD MO INTO AC IF(INIT)2.I.2 CONTINUE JRAN-SOI INIT-I CLA CLL TAD \JRAN [GET PREVIOUS NUMBER MOL MUY [MULTIPLY BY SPPSACGIES) 606$ [5“5 MOA [MODULUS 4096 ANSVER IAC [ADD A CONSTANT (I) DCA \JRAN TAD \JRAN MOL [SCALE TO RANGE OF I-IOCI MUY [MULTIPLY BY IOOO I150 [AND DIVIDE BY A096 DCA \HRAN CLA CLL IRAN-KHAN RETURN END Subroutine IRANG.FT OOOOOOOOOOOOOOOOOOOO SUBROUTINE IRANGIINIT.IDEV.IMEAN.INUM.IRAN) GENERATES NORMALLY DISTRIBUTED (GAUSSIAN) RANDOM INTEGERS USING THE CENTRAL LIMIT THEOREM APPROXIMATION. TIMOTHY A. NIEMAN JUNE 20.I9TA INIT-G INITIALIZES THE GENERATOR IDEV-STANDARD DEVIATION IMEAN-MEAN INUMINUMBER OF APPROXIMATION TERMS TO USE IRAN-RANDOM NUMBER METHOD! XIII-UNIFORM RANDOM VARIABLE V(J)-STANDARD NORMAL VARIABLLE VCJIOICSUM OF N VALUES OF X)-E)[SORT(V) EIEXPECTED VALUE OF SUM OF X - NCI/2) VIVARIANCE OF SUM OF X I NCI/IR) IRANIIMEAN.IDEV)IU(JI¢G.II‘IDEVOIMEAN' SUM-O DO I I-I.INUM CALL IRANUCINIToIRAN) SUM-SUMOFLOAT(IRAN) SUM-SUMIIOOG. RAN-(SUM-FLOATIINUM)[2.)[SORT‘FLOATIINUM)[I2.) IRAN-RANOFLOAT(IDEV) ’ IRANIIRANOIMEAN RETURN END 223 Subroutine RAND.SB [ / SUBROUTINE RAND --- GENERATES RANDOM NOISE / z I - ARRAY To PLACE THE NUMBERS INTO / / N . NUMBER or VALUES To GENERATE [ / IDEV . STANDARD DEVIATION or THE NOISE w I _ / SUBROUTINE RAND(I.N.IDEV) i [ , OPDEF TADI IAOO : 090:. DCAI 34cc ; 0PDEF NOA TTOI [LOAD MULTIPLIER INTO AC P- OPDEF MOL 1421 [LOAD AC INTO NO opus. NOV 7405 [MULTIPLY OPDEF DVI 1.01 [DIVIDE ENTRY RAND RAND. BLOCK 2 [ / GET ARRAY LOCATION. NUMBER or POINTS. I DEVIATION , . TAD RAND DCA POINT pOINT. HLT [REPLACED BY cor. THEN DY ARRAY POINTER TADI RAND. DCA FIELD [SET FIELD or ARRAY INC RAND. TADI RAND. DCA POINT [STORE START or ARRAY IN POINT INC RAND. ‘ TADI RAND. DCA COUNT INC RAND. TADI RAND. DCA IDEV [ADDRESS or N INC RAND. COUNT. HLT [REPLACED BY CDFI THEN DY -N TADI IDEV [GET N CIA DCA COUNT [STORE -N As COUNT TAD RAND DCA SIGN SICN. HLT [REPLACED BY CDFITHIN BY SION or RANDON NOISE TADI RAND. DCA INDEX INC RAND. TADI RAND. [LOCATION or IDEV DCA IDEV INC RAND. INDEX. HLT [REPLACED BY cor. THEN USED As THE TABLE POINTER TADI IDEV DCA IDEV [STORE THE VALUE or IDEV TAD (620! [CHANGE DATA FIELD BACK To SAME As It 622. [READ IT DCA GEN [STORE CDF AT START OF NUMBER GENERATOR Q\\\ IDEV. FIELD. OUT. START. NUM. 224 RANDOM NUMBER GENERATOR AND SCALING SECTIONS HLT [REPLACED BY A CDF TO CURRENT FIELD TAD NUM [GET PREVIOUS RANDOM NUMBER MOL MUY [MULTIPLY OLD VALUE BY s..s (3125) 6065 MOA [MODULUS 4096 ANSVER IAC [ADD A CONSTANT (I) DCA NUM [SAVE NEV VALUE TAD NUM MOL [CONVERT VALUE TO RANGE OF 0 - 999. MUY [BY MULTIPLYING BY I000 I750 [AND DIVIDING BY 4096. VALUE NOV IN AC TAD START [CONVERT NUMBER INTO A POSITION IN THE TABLE DCA INDEX TADI INDEX [GET THE VALUE CLL SPA [IS VALUE POSITIVE? CMA CML IAC [IF NEGATIVE. MAKE POSITIVE MOL RAL DCA SIGN [SAVE SIGN OF THE VALUE MUY [MULTIPLY BY IDEV 0 DCA INDEX [SAVE HIGH ORDER UORD MOA TAD (62 [ADD ROUND - OFF FACTOR MOL RAL [MOVE CARRY INTO POSITION TAD INDEX DVI [DIVIDE BY I00 I44 ' CLA CLL TAD .SIGN [GET SIGN RAR [AND MOVE IT INTO THE LINK MOA [GET THE FINAL ANSUER SZL [IS IT NEGATIVE? CIA [VELL. THEN MAKE IT NEGATIVEI HLT [REPLACED BY A CDF TO THE ARRAY FIELD DCAI POINT [STORE RESULT IN ARRAY INC POINT A ISZ COUNT [MORE POINTS NEEDED? JMP GEN [YES TAD GEN [NO. CHANGE DATA FIELD TO CURRENT FIELD DCA OUT HLT [REPLACED BY CDF RETRN RAND TABLE TABLE! PAGE LAP TABLE. 225 DECIM “3263-2973-2801-2703-260I'2541-2401°2433-2381-234 -2303'2271-2233-22II-BIGI~2ISI-2I31-2I01-2001-206 '2041-202!-2001'I9BI-I961-I951-I933'I9II-I901-I00 TIGTI-I051-ISAI-IBGI-IBII-IGBI-ITOI-I701-I76I-I75 -I743-I73)-I723-I7I1-I70I-169I-I683-I671-I661-I65 “Ibdi-I6JI-I621-I6I3-I603“I593-I581'I57I-I56I-I55 'ISSI-ISAI-I53)-I528-ISII-ISII-I501'I49I-I481-IQ7 -I473-I46I~I45)-IASI-IAAI-IAOI'IABI-IA2!-IAI)-I60 -I408-I39I-I381-I381-I371-I361-I36I-I353'I341-I34 -I331-I833-132)-I8II'I3I3-I303-I301-I29I-I201-I28 -I273-I271-I26I-I26I-1253'I251'I241-I2AI'I2GI-I22 -I22I-I211-I2II-I20I-I20I-II9l-II9I-IIBI-IIBI-II7 'IITI-II6I-lI63-II53-IISI-IIAI-IIAI°IIOI-II31-II2 -II2I-IIII-IIII-IIOI'II03-IIBI-IB9I-I09I-I083-I08 -I073-I013-I061-I061-I063-IOSI-I053-I041-IOAJ'IBG 'IGGI-I03)-I02I-I023-IOI3-IOII-IO0I-I003-I001-99 '993-983-988'903-971-973'963-963-931-95 '953-941-943-941'933-933-921-92I-923'9I '9II-908'901-903-093-09I-89I'881-001'07 ‘873-871-868-061'063-053-851'053-041-04 “033-031-033-82I-02I-021-8II-BII-BII-00 '80)'001-791-793-793-783-781~701-771-77 '771-761-763~763-733'751'751-7AI-7AI-74 -731-731-731~721-721-721-TI)-7II-7II-70 '70!-703-69]-69I-69I-68I-6BI-681-673-67 '67)~673-66I-66I-661-651-6SI-6SI-643-64 ‘643-631-631-633°623-62)-623-621-6II-6I -6|)-60I-603-60)~59I-S9I-59I-59I-503-50 -58)'57)-571-573-563-561'563-561-551-55 '551-548-543~543-543‘531-53I’533-521-52 '52}-S2I-SII-SII'SII‘SBI-$03'501-A91-A9 -49)-49I'481-483-403'483-473-471-471-46 '46!-A6!'46!'43!'451-451-443-443-443-44 '433-431-433-43I-423'42)-A23-AII-AII-4I -AI)-401-401-403-403-391'391'391-301'30 '381-383'371-37I-371-371-361-363'363-35 “353“353-353-343-GAI'GAI-34I-333-333-33 '323-32I-GZI-32I-3II-3II-3II-8II-303'30 '303-303-291-291-293'29)'203-281-203-20 '271-27fi-2TI-273'261'26I-261-251'258-ES '25)-24I-2AI-243-2AI-233'231-231-231'22 '221-221‘221-2II'2II'2II-2II'203-201-20 “ROI'IOI-I9I-I9I-I9I-IBI-IGI'IBI-IGI-IT -ITI-ITI'I61-I6I-I63'I6I-ISI'ISI-ISI'I5 -I41'IAI-IAI-IAI-IOI-IGI-IGI-IGI-IBI-IB “IZI-I2I-III-III'III-III-IBI-IBI-IBI'IB -9I-9I-9I-91-8)'08-01-01-73-7 '7)-71-6)-6I-6)'61-SI-$I°SI-S -AI-AI-AI-AI-33-33-31-31-23-2 '23-2I-II-II-I)-I30303010 226 03010301I1I1I1I1212 2121333131314141414 5151515163616161717 7171010181819191939 I01 I03 I01 I01 II1 II1II1 II1 I21 I2 I21I21I31I31I31I33IA1I43I43IA I51I51I53I53I63I61I61I61I71I7 I71IB1101I03I81I91I91I91I9120 2012032012I12I12I12I122122122 22123123123123126124124124125 25125125126126326127127127327 28128128128329329129329130130 8013013I13I13I331332132332333 33133133134136134334135335135 35136136136137137337137338338 3833813913933914016016034034I Al34|14l142142142343143148143 44344344144145145145146146146 461471473473481481AG1AG1A9149 491A915015035035I15I1SI352352 52352353153353154154154354155 55355156156356156357157157158 5835815935935915916016036016! 6I16I162362162362163363363164 64364365165165166366166167167 67167168168160169169169170370 7017137I17I372172172173173173 7A174374375375175176176176177 77177378170178179179179380180 8030I18I10I182182102383183383 84184385185185106106166387387 8718818838918910919019039019I 9I39239219219319319A194196195 95395196196197397198198390199 9911001I001I001IOI1IOI11021I021I031I03 I031I041IOA3I051I053I061I061I061I071I07 I083I083I093I093II01II03II01III1III1II2 II23II31II31II41II41I151II53II61II61II1 IIT1II01IIB3II91II93I201I201I2I1I2I1I22 I223I231I241[2&1I251I251I261I261I271I27 I281I283I2931301I301I3I1I3I1I321I331I33 1341I341I351I361I361I371I381I301I391IAO IA01IAI1I421I421IAJIIAAIIA51I451I461IA7 I471I401I493I501I5I1I5I3I521I531I541I55 I553I561I571I5831591I601I6I1I621I631I66 I651I661I671I601I691I701I7I3I723I731I74 I751I761I781I7911801I8I1I031I841I051I07 I881I901I9I1I933I953I961I98120012021204 206120012I012l312I512I0122I122332273230 234128012433248325‘12603270128012973326 END .-—:.—.—‘-__ ‘1. .‘T' ‘ 227 Subroutine SMOTH.SB \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ SUBROUTINE SMOTH -- GENERAL SAVITZKY AND GOLAY SMOOTH (OUADRATIC-CUBIC) BRIAN HAHN BI2A/73 LENGTH: 3 PAGES OF CORE DESCRIPTION! SMOTH HILL PERFORM AN N-POINTS SAVITZKY AND GOLAY SMOOTH ON ANY INTEGER ARRAY. THE INPUT ARRAY MAY RESIDE ANYVHERE IN CORE. THE INPUT VALUES MUST BE INTEGERS. BUT THEY ARE UNRESTRICTED IN VALUE. I.E. ALL POSITIVE AND NBEATIVE INTEGERS ARE PERMISSIBLE. THE FOLLOWING SMOOTHS ARE CONTAINED IN THES ROUTINE! 5-POINT . T-POINT 9-POINT II-POINT I3-POINT I5-POINT IT-POINT 23-POINT ANY ATTEMPT TO USE A I9-POINT OR A 2I-POINT SMOOTH VILL PRODUCE A I7-POINT SMOOTH ATTEMPTS TO USE SMOOTHS LARGER THAN 23-POINTS VILL GIVE THE COMPUTER A MIGRAINE HEADACHE AND IT HILL VIPE OUT YOUR SYSTEM TAPE IN REVENGEIIII (HOVEVER. IF YOU ARE VERY LUCKY. IT UILL ONLY GIVE YOU GARBAGE FOR THE SMOOTHED VALUES.) INPUT PARAMETERS! IRAY I NAME OF ARRAY TO BE SMOOTHED NRAY I NUMBER OF POINTS IN ARRAY TO BE SMOOTHED ISIZE I SIZE OF THE SMOOTH TO BE USED (I.E. 5. 7. 9. II. I3. I5. I7. OR 23) SUBROUTINE SMOTH(IRAY.NRAY.ISIEE) DEFINITIONS OF NON-STANDARD CODES TADI I400 [INDIRECT TAD TO FOOL SABR DCAI 3400 [INDIRECT DCA TO FOOL SABR JMPI 5400 [INDIRECT JMP TO FOOL SABR MOA 77Gl [LOAD MO INTO AC MOL 7Q2l [LOAD AC INTO MO MUY 7405 [MULTIPLY DVI 7407 [DIVIDE SMOTH. POINT. COUNT. NMI. \\\ FROM THE MAIN LAP ENTRY SMOTH BLOCK 2 GET ARGS CLA CLL TAD SMOTH DCA POINT HLT TADI SMOTH! DCA SHIFTA INC SMOTH! TADI SMOTH! DCA POINT INC SMOTH! TADI SMOTH! DCA COUNT INC SMOTH! TADI SMOTH! DCA INITI INC SMOTH! TADI SMOTH! DCA NMI INC SMOTH! TADI SMOTH! DCA LTIA INC SMOTH! HLT TADI INITI DCA COUNT HLT TADI LTIA 228 PROGRAM [CLEAR EVERYTHING -- JUST IN CASE [REPLACED BY CDF1 THEN USED AS A POINTER [SET FIELD OF ARRAY INTO SHIFT [STORE STARTING ADDRESS OF IRAY AS POINT [SET UP RETRIEVAL OF NRAY [SET UP RETRIEVAL OF ISIZE [REPLACED BY CDF3 THEN BY NRAY [GET NRAY [AND STORE IT [REPLACED BY CDFI THEN BY ISIZE'I [GET ISIZE SET UP SMOOTH PARAMETERS AND POINTERS RAR TAD DCA TADI RAR CLL RAL DCA TAD CLL DCA TAD TAD CIA DCA TAD DCA TAD TAD DCA CIA POINT PLACE LTIA NMI NMI MNMI COUNT MNMI COUNT SHIFTA DATAF LTI NMI LPOINT [DIVIDE ISIZE BY 2 [FIND LOCATION OF FIRST POINT TO BE SMOOTHED [AND INITIALIZE PLACE [GET ISIZE AGAIN [DIVIDE IT BY 2 AGAIN [DROP THE LOVEST BIT IN ISIZE [THIS SHOULD NOV BE ISIZE-I [FORM -NMI [SUBRTACT NUMBER OF POINTS NOT SMOOTHED [STORE COMPLIMENT OF NUMBER OF POINTS AS COUNT [SET ARRAY FIELD INTO OUTPUT SECTION [SET POINTER FOR END OF TD‘IP STORAGE BUFFER \\\\ \\\ INIT. I‘\\\ COP. INITI. In\\\ HIFT. SHIFTI. SHIFTQ. SHIFT2. 229 SET UP POINTER TO THE PROPER SET OF SMOOTHING CONSTANTS AND SET THE NORMALIZATION TAD TAD DCA TADI DCA INC TADI DCA INITIALIZE THE TEMPORARY TAD DCA LPARAM NMI INITI INITI NORM INITI INITI MATHA MNMI INITI JMS SHIFT ISZ JMP INIT INITI FACTOR [FIND THE PROPER SET OF POINTERS [GET THE NORMALIZATION FACTOR [AND PUT IT IN THE MATH ROUTINE STARTING ADDRESS OF THE CONSTANTS IT IN THE MATH ROUTINE [GET THE [AND PUT STORAGE BUFFER [INITIALIZE THE COUNTER [SHIFT TEMP BUFFER DOUN AND GET ANOTHER POINT [TEMP BUFFER INITIALIZEDT [NOD MAIN PROGRAM LOOP JMS SHIFT JMS MATH INC PLACE ISZ COUNT JMP LOOP TAD (6203 6224 DCA INITI 0 RETRN SMOTH [SHIFT TEMP BUFFER DOVN AND GET A NEV VALUE [CALCULATE THE SMOOTHED VALUE [INCRmmT THE SMOOTHED VALUE POINTER [DONE? [NO. [YES. CHANGE OF TO SAME AS IF [READ IF [GENERAL PURPOSE ADDRESS - SHIFTS TH‘IP BUFFER ONE PLACE AND GETS NEV POINT SHIFT - 0 TAD LTI DCA LTIA TAD LTI IAC DCA LTIB TAD MNMI DCA SHIFT2 TAD I LTIB DCA I LTIA INC LTIA INC LTIB ISZ SHIFT2 JMP SHIFTI HLT TADI POINT DCA I LPOINT INC POINT JMPI SHIFT 0 [SET START OF BUFFER POINTER [AND THE SECOND POINT IN BUFFER [INITIALIZE SHIFT LOOP COUNTER [GET A POINT [AND MOVE IT OVER ONE SPACE [INCREMENT ADDRESS POINTERS [DONET [NO- [REPLACED DY ARRAY FIELD [YES. GET NEXT POINT [AND PLACE AT END OF BUFFER [RETURN [LOOP COUNTER HNHI: LPARAM: LT I 0 PLACE: LTIA: LTIB: LPOINT: \\\\ \\\\ 230 HERE ARE ALL THOSE VEIRD NAMES I'VE BEEN USING AND WHAT THEY ARE 0 PLACE ITI G D G C l-NMI ( -(ISIZE-I) ) [START OF PARAMETER LIST: [START OF TEMP BUFFER AREA [POINTER TO CURRENT ARRAY VALUE BEING SMOOTHED [POINTER TD TEMP BUFFER FOR SHIFT ROUTINE [SAME AS LTIA: BUT OFFSET BY I [LAST POINT IN TEMP BUFFER OFFSET BY A LIST OF THE POINTERS AND VALUES USED TO SET UP MATH VITH THE PROPER SET OF CONSTANTS. DECIM [43 CIGPT IISS CISPT 323 CITPT 323 CI7PT 323 CITPT 865 C23PT (VALUES ARE DECIMAL) [NORMALIZATION FOR S-PT [START OF S‘PT SMOOTHING CONSTANTS [NORM FOR T-PT [START OF T-PT CONSTANTS [ETC. [£16. II9-PT GETS A IT-PT SMOOTH [2I-PT ALSO GETS A IT-PT SMOOTH [HOWEVER A 23-PT SMOOTH IS KOSHER ATTEMPTS TO USE OVER A 23°PT SMOOTH VILL PROBABLY BLDV THE PROGRAM: PAGE SO DON'T TRY IT! 3\\\ ATM: MATHI: MLTPLR: MATHZ: 231 MATH -- CALCULATES SMOOTHED VALUE USING DOUBLE PRECISION MATH G CLA TAD DCA DCA DCA CMA TAD DCA TAD DCA TAD DCA CLL CLL TAD DCA RAL TAD TAD DCA INC INC ISZ CLL PLACE PLAC2 SUMB SUMA MNMI COUNTR MATHA MATHB ADITI MATHI IAC MATHB IAC MLTPLR SIGN HIGH SIGN MATH2 CMA IAC LOU - HIGH IAC HIGH LOU SUMB SUMB HIGH SUMA SUMA MATHI MATHB COUNTR JMP MATHI [SET RESULT POINTER [INITIALIZE THE DOUBLE PRECISION RESULT [INITIALIZE COUNTER FOR THE MATH LOOP [INITIALIZE CONSTANT STORAGE POINTER [INITIALIZE TAD ITI STATEMENT [TAD VALUES OF POINTS USED IN THE SMOOTH [HULTIPLICAND POSITIVE? xuo. roan 2fs COMPLIMENT [GET MULTIPLIER FROM CONSTANTS ARRAY [MULTIPLIER POSITIVE? [N00 FORM 2:5 COMPLIMENT [SAVE LINK AS SIGN INDICATOR [MULTIPLY [MULTIPLIER [SAVE HIGH ORDER RESULT [RETURN SIGN T0 LINK [IS PRODUCT NEGATIVE? [NO [YES. COMPLIMENT PRODUCT [LINK IS USED TO COUPLE CARRY FROM LOV TO HIGH [ORDER NORDS [START DOUBLE PRECISION SUM [LOU ORDER SUM [MOVE CARRY FORM SUMB INTO BIT II [HIGH ORDER SUM [ INCREMENT TAD STATEMBJTS [LAST POINT? [NOo \\\ oxvi. NORM: DATAF: SUMA: SUMB: COUNTR; MATHA: MATHI: ADITI: SIGN: HIGH: LOU: PLACE: ITI: CSPT: CTPT: COPTO CIIPT: CIBPTO CISPTO CITPTO CROPT: 232 NORMALIEE THE SUM TO GET THE SMOOTHED VALUE TAD SUMA [CHECK THE SIGN OF SUM RAL [MOVE SIGN INTO LINE CLA SEL [IS LINK I I? CMA [YESI SUM IS NEGATIVE DCA SIGN [STORE SIGN INDICATOR SNL ‘ JMP DIVI TAD SUMB [SUM NEGATIVE CLL CMA IAC [COMPLIMENT LOV ORDER UORD DCA SUMB ' TAD SUMA CMA [COMPLIMENT HIGH ORDER UORD SZL [LINK USED TO COUPLE CARRY BETVEEN VORDS CLL IAC ' DCA SUMA TAD NORM [ADD ROUND-OFF FACTOR AND LOAD DIVIDEND RAR ' CLL TAD SUMB MOL RAL [MOVE ROUND-OFF CARRY INTO POSITION TAD SUMA ' DVI [DIVIDE BY NORM I [NORMALIEATION VALUE MOA [GET GUOTIENT ISE SIGN [GUOTIENT NEGATIVE? SKP [NO CIA [YES HLT [REPLACED BY THE DATA FIELD OF THE ARRAY DCAI PLACR [DEPOSIT SMOOTHED VALUE JMPI MATH [RETURN I [HIGH ORDER SUM G [LOU ORDER SUM I [COUNTER FOR LOOP O [ADDRESS OF START OF CONSTANT ARRAY I [TEMP STORAGE POINTER TAD ITI [TAD FIRST VALUE IN TEMP BUFFER I [SIGN REGISTER FOR MULT AND DIVIDE O [HIGH ORDER PRODUCT I [LOV'ORDER PRODUCT I [COINTER FOR RESULT ADDRESS 3:2:R R3 [TEMPORARY STORAGE BUFFER FOR UNSMOOTHED POINTS THESE ARE THE TABLES OF SMOOTHING COEFFICIENTS 'JIIEIITIIZI-G 'RIGIGITIOIGI-B 'RIIIAISOISAISOISAIGOIIAI-2I “3689344369384189IGAI69;:A:::;?:’ " - IGIOII612IIRAIESIZAI I - -TSI-I3IAZISTII223IA?II62IIGTII62)IATIIREIGTIAzl-IGI-TS 'EII-OITII812?)OABGOIAEIAGIAZIOOIGAIRTIIGITI-OI-EI ‘421-2II-23ISIGOIAGISAIOGITOITSITG TOITGITSITOISOISAIAOIGGIISI-EI-ZII-Az 233 Subroutine SMOTH.FT SUBROUTINE SMOTH C TIMOTHY A. NIEMAN AUGUST IA; I974 DIMENSION XNORMCIO):IVEIGH(I30IC):IRAY(IOZA);IBUF(23) KNORM(I)-35o XNORM(G)I2I- XNORH(3,.23IO KNORM(A)-429. XNORMCSIIIAGo XNORM(6’.II050 XNORMI130323o XNORM(8)-226Io XNORMC9DUGOS9o XNORM(IO)-805o IVEIGHCIoI)I-3 IVEIGH(2:I)II2 IUEIGH‘GoIIIIT IVEIGH(I:2)--2 IVEIGHC2o2)-3 IUEIGHC3a2)-6 IUEIGH(A:2)87 IVEIGH(I:3)--2I IUEIGHC2a3)8IA IVEIGH(3:3)-39 IVEIGH(A:3)I54 IVEIGH(503)-59 IVEIGHCIoADI-36 IVEIGH‘204)-9 IVEIGH¢304)IAA IUEIGHCGJG’IO9 IVEIGHCSoA)-84 IVEIGHI630)-G9 IVEIGHCIoS)I-II IVEIGH¢205)'O IVEIGHCJ:S)O9 IVEIGH¢Ao5)-I6 IVEIGHCSaS)-2I IVEIGHI6:S)I24 IVEIGH(7:5)I2$ IUEIGH(I:6)I-78 IVEIGH<2¢6)--I3 IVEIGH(3:6)-42 IVEIGH(4¢6)887 IVEIGHI329 IVEIGHCIoIOJO'AZ IVEIGHCZoIGIO'ZI IUEIGH(3;IG)I-2 IVEIGH(AoIO)DIS IUEIGHCSoIO)-30 IVEIGHI6:IG)'AG IVEIGHI?:IO)'54 IVEIGH(8:IG)-63 IVEIGH(9;IG)-70 IVEIGHCIG:IO)-?S IVEIGH(II0IG)=78 IUEIGH(I2:IC)I79 235 Program CHISQ.FT 00000000000 93 IIS 999 ISO III 281 000 IIA I03 I a 000 000 IGA 30 SI 32 CHISGoFT PROGRAM TO DO CHI- SQUARE GOODNESS OF FIT TEST TIMOTHY A. NIEMAN JULY I6aI97A INPUT MAY COME FROM TTYoPTR: OR DECTAPE FILE. FIRST UORD INPUT IS THE TOTAL NUMBER OF WORDS TO INPUT. DISTRIBUTIONS COMPARED ARE UNIFORM AND NORMALT COMMON IDATA:IENDP:EXPT:OBS:FX DIMENSION OBS(I28):EXPT(I28):IDATA(IO2AI:IENDP(I28):FX(200) DIMENSION TRVL(A)oNUINT(A) READ(2:IG3)INPUT DO 93 I-Io2GG READ(2:IIG)FX(I) FORMAT(FToA) URITE(I0IOI) FORMAT(/I URITE(I5IGI) FORMAT(‘CHI-SGUARE GOODNESS OF FIT TEST’Il) READ(I0287)LPT FORMAT(‘ENTER A I (ONE) TO DELETE LINEPRINTER OUTPUT ‘aISI READ(I:I02IINPUT FORMAT(‘INPUT FROM? n-rrv.z-pra.3-DTA 'aIS) GO TO(I32:3):INPUT TTY xuévr READ(|:IIA)INPUT FORMAT(‘TOTAL NUMBER OF POINTS I 'oIS) FORMAT(IS) VRITE(I:IOO) IF(IO2A-INPUT)II:II:II INPUT-IDEA I DO I2 I-IoINPUT READ(IoIO3)IDATA(II GO TO A PTR INPUT READ(20I93)INPUT IFCIBZA-INPUT)2D02I02I INPUT-IDEA . D0 22 I'IOINPUT READ(20IB3’IDATA(I) GO TO A DECTAPE INPUT READ(I:IOA)FNAME FORMAT(‘FILENAME(A6) IS 'oAb) CALL IOPEN('DTAI’oFNAME) READ(A:I03)INPUT IF(I62A-INPUT)30a3I03I INPUT-IO2A DO 32 I-IoINPUT READ(A:I03)IDATA(I) GO TO A 000 000 000 000 II9 27A I05 87 II? 86 SO 524 53 89 SA 56 I66 I67 6| 236 CHOOSE DISTRIBUTION URITE(I0IOO) READ(IOII9)IDIST FORMAT(‘I'LIST DATA: O'SUPPRESS LIST 'OIS) IF(IDIST-I)27A:276027A URITE(IOIOD) READ(I:IOS)IDIST FORMAT('I-UNIFORHoE-NORMAL 'aIS) GO TO(5;6)OIDIST UNIFORM MAX'IDATA(I) MIN-MAX IDIST-O . READ‘I:IDO)IPARM GO TO (87088’0IPARH READCIIII7)MIN0MAX FORMAT(‘MIN ' 'oIS/‘HAX I 'oIS) GO TO 39 DO 53 I-ZOINPUT IF(FLOAT(IDATA(I))‘FLOAT(MIN))5005I15I MIN'IDATA(I) IF(FL0AT(IDATA(I)I‘FLOAT(HAXJ)53052052 MAXIIDATA(I) CONTINUE URITECIJIITIMINJMAX UNIFORM - SET UP INTERVALS AND EXPECTED FREQUENCIES. K-FLOAT(INPUT)/S. IF(K-I27)SS:SSoSA K-I27 ITRVL-(FLOAT(MAX-MIN)[FLOAT(K))+.5 IENDP(I)-MIN ‘ XPECT-FLOAT(INPUT)[FLOAT(K) DO 56 I-IoK IENDP(I+I)IIENDP(II+ITRVL EXPT(IIIXPECT GO TO 7 NORMAL READ‘I0I06)IPARH FORMAT('I'INPUT PARAMETERS: Z'ESTIHATE THEM 'OIS) GO T0(839):IPARM READ‘IJIBTIXMEANIDEV IDIST-O FORMAT(‘MEAN I 'oFIO-Z/‘STANDARD DEVIATION 3 'OFIDoE) GO TO 60 SUN". SUNSQ'D- DO 6I I'IalNPUT X3IDATA( I ) SUM'SUM*X SUMSQ'SUMSQ’XAX XHEANISUM/FLOAT(INPUT) DEV'( SUNSQ/FLDAT‘ INPUT) )‘ (XMEAN‘XHEAN) DEV'SQRT ( DEV) URITE(I:I37)XMEANJDEV GO TO 69 (TCIO 60 62 63 65 6A 66 67 68 69 .70 TI 72 9? I09 9A 95 96 I10 73 276 I2I I20 237 NORMAL - SET UP INTERVALS AND EXPECTED FREQUENCIES. XK'FLOAT(INPUT)I50 IF(KK-IRG-)63063062 XK-I20. K'O EXPT(I)I.I9IS EXPT(2)00I500 EXPT(3)I.09IS EXPT(A)I.0AA0 D0 64 I'IIA NUINT(I)'XKOEXPT(I)¢.S IF(NUINT(I))6S:6506A NUINT‘IIOI K'K‘NUINT(I) K-KOI DO 66 I'IoA TRVL(I)-(.SADEV)/FLOAT(NUINT(I)) URITE(IOI00) FORMAT(‘VALUES FROM A STANDARD NORMAL TABLE'/ I' X (I F(X)'/) ' IENDPCI)-'20A7 IENDP(EAK+I)-20A7 I-I J-A OFSET'(FLOAT(J)[2.)ADEV IENDP(IOI)IXMEAN-OFSET IENDP(B.HOI'I)IXMEANOOFSET I'I’I IF(K-I)72a72069 IF(NUINT(J)'I)7007007I J'J'I GO TO 67 NUINT(J)-NUINT(J)-I IENDP(IOII-FLOAT(IENDP(I))0TRVL(J) IENDP(2AKOI'I)IFLOAT(IENDP(2tK+2-I))-TRVL(J) GO TO 68. IENDP(KOI)OXMEAN PI... KIN-INPUT DO 73 I-IoK . XI(FLOAT(IENDP(IOIII-XMEANIIDEV va:r;¢:.1¢9)x. ' FORMAT(FSoE) IXC-XGIOOO IF(IX)9A59A095 P230500. 00 TO 96 P2'FX(IX) VRITE‘IOIIOIPE FORMAT(' 'oFToA) EXPT(I)5(P2-PI)AXIN EXPT(E¢K¢I‘I)OEXPT(I) PIIPE KOKOK GO TO 7 URITE(30I2I) FORMAT(IHI) VHITE‘O:I20)(IDATA(I)oI-IOINPUT) FORMAT(I0(5XJI5)) GO TO 27A 000 000 000 77 70 76 79 GA 05 03 62 _9I I16 9i was nus so 238 cuzcx ran zxéncrzn VALUESII31I31IA AREAIAREA‘FLOATCIUORKIK)) IF(IWORK(K)IMAX)37037:38 MAXIIUORKIK) MEAN-K CONTINUE CONTINUE INN-HAX/z IFVHNIO DO 39 K‘KIIKZ IF‘IUORKIK)‘IHH)AOO4IJQI IFVHMIIFVHMOI CONTINUE CONTINUE PCNTAIIAREA/SUH)II000 PCNTHIIFLOATIHAX)/FLOAT(IHAX))IIDC. PCNTVI(FLOAT(IFVHM)/FLOAT(IVIDE))II00. DO IO K-KIIKZ IXIQIIK‘I) IYIIWORKIK) CLA CLL CML RTR TAD \IX RAL CMA 606I CLA CLL CML RTR TAD \IY RAL CMA 6066 CLA CLL CONTINUE 606I CLA CLL 6666 HRITEC3:I021I.MEAN:MAX:PCNTH0IFUHMoPCNTUoAREAaPCNTA GO TO 5 END ‘p-“urm .. xm‘m . . CIOCHCIO¢1C30 247 Program FOUR4.FT I05 999 I07 25 I08 I09 997 998 '0. I IIO 27 26 I03 200 PROGRAM FOUR4.FT FAST FOURIER TRANSFORM - COMPUTES REAL TRANSFORM TIMOTHY A. NIEMAN MARCH 2Ia I975 ACCEPTS INPUT FROM TELETYPE OR DATA FILES 0N DTAI COMMON A18 DIHENSION A(256).8(I28) 6051 UPITEII.IO$> FORMAT(/I'FAST FOURIER TRANSFORM PROGRAH'II) CONTINUE ' READ¢IoI07)IA . FORMATI'INPUTT I-TTY.O-OECTAP£\ '.I5I ITIIA>25.25.26 REAOII.IOUITNANE TORNATI'TILENANE 15 '.A6) CALL IOPENC'DTAI'oFNAME) REAOIO.I99)NOATA‘ FORMAT(IS) IF(NDATA-128)998.998.991 NOATA-Iea CONTINUE URITSII.IOOINOATA FORMAT(‘NO. OF INPUT POINTS - '.Is> READ(I.III)NTVO FORNATI'NO. 0! POINTS To TRANSFORNIPOUER or 2) - '.I5) 00 21 IIIINDATA ‘ REAOIO.IIOIIA TORNATIIS) A<201)I0. A(2II-I)IIA GO TO 28 READ(I.100)NDATA R:AONTUO URITBII.I02) FORMAT(/'INPUT ON: POINT PER LINE'I) DO I I-Iizso A‘I)"o DO 2 I-I.NOATA READ(I:I03)A(2II-I) FORMAT(FI6.5) FORMAT(/(IOIZXoFIO.4))) (TCIO I04 OIB~J I3 I4 IS I6 248 TRANSFORM READ¢IoI0¢>ISIGN FORMAT('IIFORBVAflbo-IIINVERSE TRANSFORM 'oIS) IPa-z. . ' IPJIIPIONTVO IBREVII DO I0 IBII.IP3.IP0 IF(IB-IBR£V)6.1.1 TEHPR-A(IB) ‘ TEHPIIA(IBOI) AIIB)IA(IBR£V) ACIB+I)IA(IBREV¢I) ACIBREV)ITEHPR A‘IBREV+I)ITEMPI IPIIIP3/2 . IF(IBRBV-IPI)I0.IO.9 IBREVIIBREV-IPI IPIIIPI/Z IFCIPI-IPO)IO.8.8 IBREVIIBREV+IPI IPIIIPO IF(IPI-IP3)12.I5.I5 IP2IIPII2 THETAIZoIJ.IAIOIFLOATtlSIGNtIPZIIPl) SINTHISIN XMINIA‘I) DO 20 IIlaNTVO IF(A(2II-I)-XMIN)I70I00I8 XMINIA<2II-I) IFCXMAX-A(ZII-I))I9.20.20 XMAXIAI2II-I) CONTINUE URITECI.II2)XHAX.XHIN FORMAT(‘MAXI 'oEI2vo' MINI 'oEI2oA) READCI.II3)XMAX.XHIN ” ’ FORMAT('SCOPE MAXI 'aEIOoO/‘SCOPE MINI 'aEIO-B) FSCALIXMAX-XMIN ' " ' 6051 READ(I.III)IA FORMAT(‘IIFASTo0-SLOU 'aIS) IFtIA)29.29.30 ° SLOV LINE PLOT CALL XYSTIOoI02A) CALL XYPLT(20A7:I02A) CALL XYEND 6057 IYIIAII)II024-IXMAX)+I0230 CALL XYST(0:IY) YSCALII02A./XMAX IXSCI(2047./FLOAT(NTUO)) DO 2I 1.20NTUO IYIAIZII-I)IYSCALOI023. IXICI-I)¢IXSC CALL XYPLTIIXoIY) CONTINUE CALL XYEND GO TO 3I FAST POINT PLOT YSCALIIO2Ao/XMAX IXSCII2047.IFLOAT(NTUO)) IYII024 DO 32 III05I2 IXICI-I)IA CALL FTPLT(IXIEY’ CONTINUE DO 33 IIIINTUO IYIAI2II-IIIYSCALOI023. IXI(I-I)IIXSC CALL PTPLTCIXoIY) CONTINUE GO TO 3I COMPUTE POWERS OF TRANSFORM READII:I06)N FORMAT(‘POUER I 'oIS) IFIN)999:999:22 ' DO 23 IIIoNTUO A(2II-I)IB(I)IIN CONTINUE GO T0 20 END APPENDIX F KINETICS PROGRAM LISTINGS 00000 250 Program VNTAMO.FT III III I0 20 30 40 VNTAMO.FT VIDICON. NON-FILE. TIMED-ACOUISITION MONITOR. TIMOTHY A. NIEMAN JULY 26; I97A COMMON IVORKoIIaIZaI3 DIMENSION IUORHISI2IoIIII032>oI2CI032)oI3(I032) URITEIIOICCI FORMAT(/I READCIOIOIIIA FORMAT('IIEXITIZIACOUIREo3IDESCRIBEoAIDISPLAYoSIDIRECTORY 'oIS) GO TO (I002003CIACaSCIIIA CALL EXIT CALL CHAINI'VNTMIN') CALL DSCRIB ‘ GO TO I CALL CHAINI'VNTADS') GO TO I CALL DRCTRY GO TO I END Program VNTMIN.FT 00000 III 998 999 VNTMINoFT VIDICON NON-FILEb TIMED-ACOUISITION MONITER FOR DATA INPUT. ATIMOTNY A. NIEMAN JULY 26. I914 COMNDN IUORKoIIoIEJIO DIMENSION INORKCSIEIOIIIIC3EIJIECI032III3IIC32) CALL NINIT 'RITE‘IOIDC) FDRNAT‘II A 2:33;:z121;;IT02.ACOUIREO3-DESCRIBEO‘-DISPLAYOS'DIRECTORY :oIS) CD TOC9990I199CJ99BJ99OIIIA CALL CHAINI'VNTAMOT) CALL EXIT END 251 Subroutine DSCRIB.FT SUBROUTINE DSCRIB‘ SUBROUTINE TO DESCRIBE DATA SETS GENERATED ON DECTAPE BY THE VIDICON TIMED DATA ACQUISITION SERIES. NON-FILE STRUCTURED FORM. TIMOTHY A. NIEMAN JULY 2: I974 IDIRIVIDICON DATA DIRECTORY IDENTIIDENTIFIER PARAMETERS AND TEXT IRUNIRUN BEING DESCRIBED 00000000000 COMMON IDIRoIDENT DIMENSION IDIRII28I0IDENT(260) CALL RTAPE(I.O.I28.IDIR) I VRITEIIoIOZ) IOU FORMAT(/I R:ADII.IOI)IRUN IOI FORMATI'DESCRIBE RUN I '.I5) IFIIRUN)§99.999.2 2 IFIIRUN-IDIRIIIIOoao9 9 HRITEII.I01)IRUN la? FORMATI'RUN '.IS.' NOT FOUND') GO TO I 8 IFBRIIDIRIJIIRUN-I) CALL RTAPEtlaIFBnoassoIDENT) IBASEIIDENTIOI ITIHEIIDENTIIO) TIME-ITIHE IIASEIIBASE-b TIMEITIMEICIOoIIIBASE) VRITE¢IaIOO> URITSII.I02)IDENT(II.IDENTIR).TIME I02 FORMAT(IS.’ SPECTRA'IIS.' POINTS PIA SPECTRA'I IFII.A.' SECONDS BETWEEN SPECTRA') IFIIDENTII2))4.4.3 3 URITECIolfla) 103 FORMAT('INTERMITTENT') GO T0 5 a URITEII.IOA) I04 FORMAT(‘CONTINUOUS'I 5 VlIT:(I.IOSIIDENT(3).IDENT(A).IDENTIS) IOS FORMATC'SIONAL AVIAACES‘I'DARK I '.I$.' RE! I 'aISo' SAMPLE I ' IIISI l A URITICIoIIOIIDINTCO).IDINTC1).IDINT¢C) I06 PORMAT('IXPOSURES‘I‘DAAK I ‘.IS.' It! I ‘.I5.' SAMPLE I '.15) PRINT IDENTIFIER TEXT CLA CLL TLS DO 6 IIJBIBSO LETTERIIDENTCI) CLA CLL TAD \LETTEW ‘ TAD (TSTI l-207 FOR CONTROL G SNA IIS IT A CONTROL 6? JMP \7 ’YESo STOP PRINTING TEII: TSF INOo IS PRINTER READY? JMP TEII INO. CHECK AOAIN CLA CLL IYES: PRINT CHARACTER TAD \LETTER TLS CLA CLL CONTINUE CLA CLL GO TO I 999 RETURN END £00000 UIO‘I'JIUIUIUIUIMMUIUI U! I ~10 s 252 Subroutine DRCTRY.FT 0000000000000000000000 000 000 I02 29 I09 II 24 I07 23 I03 IA I04 28 SUBROUTINE DRCTRY‘ SUBROUTINE TO PERFORM DIRECTORY OPERATIONS ON DATA TAPES FROM VIDICON TIMED DATA ACQUISITION SERIES. NON-FILE STRUCTURED FORM. TIMOTHY A. NIEMAN JULY 2. I974 IFBDIFIRST BLOCK TO DELETE ILBDILAST BLOCK TO DELETE IFBRIFIRST BLOCK TO READ ILBRILAST BLOCK TO READ IFBUIFIRST BLOCK TO WRITE ILBUILAST BLOCK TO WRITE ILBCILAST BLOCK TO COPY IBPTIBLOCKS PER TRANSFER IUPTIUORDS PER TRANSFER IDECTIINPUT DECTAPE UNIT NUMBER JDECTIOUTPUT DECTAPE UNIT NUMBER IDIRIVIDICON DATA DIRECTORY NRUNTITOTAL NUMBER OF RUNS ON TAPE IRUNIRUN TO BE OPERATED UPON COMMON IDIRalTEMP DIMENSION IOIRII28I.IT:NP<2¢OOI URITE(I.I00) TORNATIII READII.I02)IA FORMAT(‘IILIST. 2IDELETE. 3ICOPY. .azsao. SIRETURN '.15) CALL RTAPE(I.0.I28.IDIR) ' NRUNTIIDIR(I) OO TOIII0I20I3529099OIOIA READCI.I09)IA FORMAT(‘ARE YOU SURE? IIYES. 'IIS) ITIIA-III.2S.I ' LIST DIRECTORY IFINRUNTI2Ao2Ao23 URITECIoI07) FORMATI’EMPTY TAPE') GO TO I I URITECIoI03) FORMAT(/'RUN NO. STARTING BLOCK TOTAL BLOCKS NSPEC'I) DO I4 IIIoNRUNT .. URITECI.I04)IoIDIR(3IIII).IDIR(3II)0IDIR(3II‘I) FORMAT(IS.9X.ISoIOXIISoSXoIS) GO TO I SET UP PARAMETERS TO DELETE IDECTII READII.IOI)IRUN FORMAT(‘DELETE RUN 0 '.ISI IF‘IRUNIIJIIZ - IFBDIIDIRIOIIRUN-I) . ITIIRUN-IOIRII)>2S.¢.21 ILBDIIDIRIOIIRUNIIIFBD-I IBPTIILBD-IFBDII IFTIME . FORMATI'SEC. DETUEEN SPECTRA - 'oFI6.6) ITITINE-IOOOC.>3.A.A ITITINC-.CI)S.6.S URITCII.ID3I FORMAT(‘TIME TOO LONC') 00 To 2 . URITCII.IOA) FORMAT('TIME Too SHORT') DO TO 2 READII:IOSINSPECINSAVEJNSEXP FORMAT(" SPECTRA I '1I5/‘I SAMPLE AVERAGES I 'IIS/ I'l SAMPLE EXPOSURES I 'CIS) TNEEDIFLOATINPNTS/I28IIVTINE‘FLOATINSAVE)IFLOATINSEXPIO.OOOS IFITIME'TNEEDITICIC URITECIOI06)TNEED FORMAT(‘PARAMETERS REQUIRE '0FI6o60' SEC BETWEEN SPECTRA'I I'CMANCE PARAMETERS OR TIME') I GO TO I TURITEI.AOFLOAT(NPNTS/IZCIIIOZS IFCTNEEDOTURITE'TIHE)9090IO URITEIIJIOT) FORMATI'CONTINUOUS ACOUISITION') IDENTIIZIIO GO TO I2 IAIOIISIZINPNTS) IFCNSPEC'IAIZZJ 2202I URITE‘IIIOT) IDENTIIZII'I GO TO I2 URITEIIIIOSIIA FORMATC'PARAMETERS REOUIRE INTERMITTENT ACOUISITION'I I'2 SECOND CAPS EVERY 'OISI ' SPECTRA') READIIOIO9IIA FORMAT(‘I'OKO PROCEED: EICMANCE PARANETERS 'OISI CO TOIIIOIIJIA IDENTIIEIII 'READII.IIO)IDENTIJ):IDENT(4).IDENTC6).IDENT(7) FORMAT(‘O DARK AVERAGES I 'aIS/ I'I REFERENCE AVERAGES I 'oISl‘O DARK EXPOSURES I 'oIS/ 2'0 REFERENCE EXPOSURES I 'oIS) IDENTII)INSPEC IDENTI2IINPNTS IDENTCSIINSAVE IDENT(8)INSEXP TIMEIITIMEII000000.) IBASEIO IF IORc-IVORKIIAOI SET CLOCK BASE AND COUNTER NBLKSINPNTS/I28 CLA CLL CML TAD \IBASE RTR RTR DCA IBASE ISAVE FOR FUTURE USE TAD \ITIME CIA DCA ITIME IDITTO CLA CLL TAD \LDASE RTR RTR DCA LBASE IDITTO SET WAVELENGTH CONTROL CLA CLL TAD \NPNTS DCA NPT ISTORE IT AUAY. TAD NPT CIA l-NPT IAC IIN ORDER TO HAVE ENOUGH TIME TO RESET POINTERS IAC IAND COUNTERS IT IS NECESSARY TO SKIP 2 POINTS. DCA INPT ISTORE INVERSE OF NPT FOR USE AS COUNTER- CLA CLL TAD NPT RAR IDIVIDE DY TUO. DCA HNPT ISTORE IT FOR USE BY CODE GENERATOR. CLA CLL MMUIUDUICAML'IUIMMMUIMUIUIUIVIMMMUIIAUIMUIC‘IUIMMUIMUIUI MMLIMMMMMUIUIMMMMMMMUIUIUIUIUIM .- I0 [ 259 [THE FOLLOHING SETS THE PROPER CONTROL [BIT FOR THE UAVELENGTH COUNTER 50 IT HILL GENERATE NPT [VAVELENGTH POINTS PER SPECTRUM. I CLA TAD DCA CLA STL ROT: RAL DCA TAD RAL SZL JMP DCA CLA TAD JMP CLL HNPT TEMP CLL CODE TEMP OUT TEMP CLL CODE ROT NPT. 0000 HNPT: 0000 TEMP: 0000 CODE: 0000 OUT. CLA TAD CLL CODE ULTCH UCNTC USON IDENT.CLA DCA DCA TAD DCA CLL CLCDE TRCDE (I20I WHICH [INITIATE TEMP [INITIATE CODE [SHIFT THE CODE BIT. [GET HNPT [IS MSG-I? [YES. CODE IS GENERATED [NO [GET CODE READY TO SHIFT AGAIN [GO AND SHIFT AGAIN [CONTAINS THE NUMBER OF POINTS/SPECTRUM [CONTAINS I/2 OF THE POINTS/SPECTRUM [CONTAINS SHIFTED VERSION OF HNPT [CONTAINS CONTROL UORD FOR VAVELENGTH COUNTER [ALL DONE. CODE IS CORRECT. [LATCH THE CONTROL UORD TO THE COUNTER [CLEAR THE WAVELENGTH COUNTER. [TURN ON VAVELENGHT SCAN. [SET CODE FOR NO CLOCK UAIT [SET CODE FOR NO TRIGGER UAIT [SET BEGINNING OF DATA STORAGE READII:II2)IVHICH CLA TAD AND SZA JHP CLA TAD 'AND SZA JMP CLA TAD AND SZA JMP TAD AND SZA JMP CLA TAD TLS JMP CLL \IUHICH (000I SAMPL CLL \IVHICH (0002 DARK CLL \IUHICH (0004 REF \IVHICH (00I0 EXTND CLL (207 IDENT FORMAT(‘DATA CODE I 'oII) [GET DATA CODE [IS THE CODE A I FOR SAMPLE? [YES - GO SET UP SAMPLE STORAGE POINTER. INC. CHECK AGAIN. [GET CODE AGAIN. [IS THE CODE A 2 FOR DARK SIGNAL? [YES - GO SET UP DARK SIGNAL STORAGE POINTER. INC. CHECK AGAIN. [GET CODE AGAIN. [IS THE CODE A 4 FOR REFERENCE? [YES - GO SET UP REFERENCE STORAGE POINTER. [GET CODE AGAIN [IS THE CODE A 8 FOR EXTENDED OUTPUT? [YES [NO. ILLEGAL DATA CODE. [RING TTY BELL. [TRY TO INPUT PROPER DATA CODE THIS TIME. 26C! 5 SAMPL.CLA CLL VRITE(I.II3) IIJ FORMAT(‘SAMPLE'I [BLOCK-IsnosznLKSoz N-NSAVE , lGOHF-6¢(512INPNTS) ICONSO NDONE-l NINT-NSEXP-I READCI.IIG)1A II6 FORMAT(‘TO OBTAIN TRIGGER ENTER A I (ONE) 'oIS) IFIIA-I)998:997.998 S \997. CLA CLL IAC S DCA TRCDE 998 CONTINUE S JMP SCANS S DARK. CLA CLL UR!TE(I.I|4) Ila FORMAT(‘DARK SIGNAL‘) IBLOCKnISB+2 NINDAVE NINT-NDEXP-l READtl.Il6)IA lF(IA-l)998.997.998 S JHP SCANS S REF. CLA CLL VRITE(I.IIS) IIS FORMAT(‘REFERENCE'> IBLOCK-ISB+2+NBLKS N-NRAVE NINT'NREXP-I READ(I.I16)IA . lFtlA-l)998.991.998 JNP SCANS SCANS.CLA CLL CLA CLL TAD \N SNA IARE ANY AVERAGES WANTED? JMP IDENT DCA N IYES- SET UP To AVERAGE N SPECTRA. TAD N CIA l-N DCA N ISTORE THE NEGATIVE OF N To SAVE TIME. CLA CLL xrcnnnr)3a.39.39 38 NINT-B U‘IUIMUIMUIU'IUIUICIUI 39 CONTINUE s CLA CLL . s TAD \NINT IGET rue NUMBER or INTEGRATION PERIODS. s can 5 DCA NINT mmmuuummmmmmmmmmmmmmmmmmmmmmmwmmmmmmmmummmmmmmmmmmmummmumumm CLA CLL TAD (0200 DCA ZLOC DCA LOU TAD (0600 DCA TEMP CLA CLL TAD (620I RIF DCA DFI CPAGE II ZEROC.CLA CLL ZLOC: DFIo TRIG. CLCH: CDFI DCAI ZLOC SKP 0000 152 ZLOC ISZ TEMP JMP ZEROC TAD (I200 DCA ZLOC TAD LOU DCAI ZLOC CLA CLL TAD ITIME LPSET CLA CLL TAD TRCDE SNA JMP CLCH CLA CLL DCA TRCDE CLCF CHTF JMP TRIG CLA CLL TAD CLCDE SNA JMP CINIT CLCIC CLOFE CLOCK.SKPOF JMP CLOCK CLOFE JMP INITI CINIT.CLCIC CLOFE JMP INITI INITI.CLA CLL TAD N CIA DCA M , TAD \NINT CMA DCA NINT JMP INIT 261 [FOLLOWING ROUTINE ZEROES COMMON STORAGE [DEPOSIT CHANGE DATA FIELD INSTRUCTION [CONTAINS ADDRESS TO EERO [STORE LOST TIME COUNTS [REPLACED BY A CDF INSTRUCTION [SET CLOCK BASE AND MODE [LOAD CLOCK COUNTER [CHECK TRIGGER UAIT CODE [CLEAR TRIGGER UAIT CODE [CLEAR FLAG [HAS FLAG BEEN RAISED? [CHECK CLOCK UAIT CODE [INITIALIZE CLOCK [CLEAR CLOCK FLAGS [UAIT FOR CLOCK [INITIALIZE CLOCK [CLEAR CLOCK FLAGS mrmmmmmmummmmmmmmummmmmmmmmmmmmummmmmmmmmmmmmmm [ 262 JHERE VE FINALLY GET TO TAKE SOME DATA. / PAGE IBASE.0000 LBASE.0000 ITIME.0000 TRCDE.0000 CLCDE.0000 NINT. 000a uutcu.ooaa INPT: 0009 N. 0000 HIGH. aaea Low. 0900 COUNT.0000 M. 9090 INIT. CLA TAD DCA TAD DCA TAD DCA TAD SZA JMP CLVF START.CHVF JMP CLVF CLEAR.CLCF CDFI INPUT.CHCF an CLA GDcc TADI DCAI RAL TADI DCAI CLL ‘UHICH HIGH (20! LOU INPT COUNT \NINT CLA INTGR START INPUT CLL LOU LOU HIGH HIGH ISZ LOU I52 I52 JMP HIGH COUNT INPUT [CLOCK TIME BASE AND MODE. [TIME BASE AND MODE FOR LOST TIME. [CLOCK COUNTS [CODE TO UAIT FOR TRIGGER [CODE TO UAIT FOR CLOCK [COUNTER FOR INTEGRATION PERIODS. [STARTING ADDRESS OF DATA STORAGE. [NEGATIVE OF POINTS PER SPECTRUM [NUMBER OF SIGNAL AVERAGES [ADDRESS OF MOST SIGNIFICANT HALF OF DATA [ADDRESS OF LEAST SIGNIFICANT HALF OF DATA [COUNTER FOR POINTS PER SPECTRUM l-N [SET UP HIGH ADDRESS [SET UP LOU ADDRESS [- I POINTS/SPECTRUM [ARE ANY INTEGRATIONS UANTE07 [YES - GO AND INTEGRATE [CLEAR VERTICAL SCAN FLAG [CHECK VERTICAL SCAN FLAG [CLEAR VERTICAL SCAN FLAG [CLEAR AID FLAG [CHECK A/D FLAG [GATE DRIVER. AND CLEAR CONVERTER FLAG. [ADD TO SUM OF PREVIOUS DATA [PUT CARRY INTO HIGH UORD [INCREMENT THE STORAGE ADDRESSES. [HAVE ENOUGH POINTS BEEN TAKEN? [NO - G0 TAKE ANOTHER POINT. MCTML1L1LIUTMUIDTMMUIUIC‘IU‘IU‘DLTDTC‘ILOLIUIDTLIU'ILTLTUICIUIL‘IU'IL1‘JIMU‘UIUIUIOTL’ILIOTUIMUILDU‘ILTLTL‘O U! DONE. ISZ JMP CLA TAD DCA TAD DCA TAD DCA CLA TAD DCA TADI MOL TADI DVI DIVISo0000 CLA MQA TAD CLL DCAI I52 152 ISZ JMP TAD DCA HLT JMP INTGR.CLA TAD CIA DCA CLVF NEXT) DFZ: INTI: JMP M INIT CLL INPT COUNT UHICH HIGH (20I LOU CLL N DIVIS LOU HIGH CLL (4000 HIGH LOU HIGH COUNT NEXT DFI DF2 TURKY CLL \NINT NINT CHVF INTI CLVF USOF ISZ JMP INTEJ JMP F NINT INTI CHVF INT2 CLVF USON JMP TURKY.CLA TAD AND SZA JMP CLA CLEAR CLL \IUHICH (0001 USAMP CLL 263 [YES. HAVE ENOUGH SPECTRA BEEN AVERAGED? [NO. TAKE NEXT SCAN TO AVERAGE [YES. DIVIDE SPECTRA SUM BY N [INITIALIZE COUNTER AGAIN. [INITIALIZE DATA ADDRESSES AGAIN. [GET NUMBER OF AVERAGES TAKEN. [DEPOSIT IT FOR USE AS DIVISOR. [LOAD MO [DIVIDE [PUT OUOTIENT IN AC [CONVERT TO 2'5 COMPLEMENT [INCREMENT ADDRESS COUNTERS. [HAVE ALL POINTS BEEN DIVIDED? [NO - GO DIVIDE NEXT ONE. [YES. CHANGE DATA FIELDS [REPLACED BY CDF [SEE UHAT TO DO UITH THE DATA. [INITIALIZE COUNTER [CLEAR VERTICAL SCAN FLAG. [CHECK ron VERTICAL SCAN'FLAG [INTEGRATE [ENOUGH INTEGRATIONS DONE? [NO - UAIT LONGER. [YES - TAKE DATA UHEN NEXT SCAN STARTS. [TURN ON SCAN. [RETURN TO DATA ACQUISITION. [FIND OUT UHAT TO DO UITH THE DATA [REF AND DARK URITTEN THE SAME CALL UTAPE(I.IBLOCK.NPNTS.II) JMP IDENT [FIND OUT NEXT THING TO DO m 6.1 U! (.1 L1 mtnUlMInUthnOTW(DUTMIAUTMZDCTMIA mmmmm ML‘IL‘ILTL‘I UTWU‘IL‘IU‘I 264 USAMP.CLA CLL [SAMPLE TAD \IORC [CHECK FOR CONTINUOUS OR INTERMITTENT SKA JMP NTRMIT [INTERMITTENT CLA CLL [CONTINUOUS CALL UTAPE(I.IBLOCK.NPNTS.II) NDONE=NDONETI IBLOCK-IBLOCKONBLKS CLA CLL TAD \NSPEC CIA TAD \NDONE SMA [IS NDONE’ONSPEC? JMP IDENT [YES. FIND OUT NEXT THING TO DO. SKOFE [CHECK FOR TIMING ERROR SKP JMP ERROR CLOFE [CLEAR CLOCK FLAGS INC CLCDE [UAIT FOR CLOCK NEXT TIME CLA CLL TAD (200 » DCA ZLOC [SET UP TO ZERO UHAT USED IN COMMON TAD \NPNTS CIA DCA TEMP [ STORAGE OF STOP ADDRESS DCA LOU ' ' JMP ZEROC [ZERO COMMON AND RECYCLE NTRMITaCLA CLL [INTERMITTENT \2: NDONE-NDONEOI ICOM-ICOM+I IF(ICOMF-ICOM)2.2.I IF(NSPEC-NDONE)2.2.3 CLA CLL [OUTPUT ALL OF COMMON THAT IS FULL TAD LEASE [RECORD HOU MUCH TIME IS LOST LCTRL CLA CLL CLCIC INARY-IOEAINPNTS IFCICOM'INARYISaAoA MPNTSIINARYtflpNTS CALL UTAPECIJIBLOCKOMPNTSIII) ICOM'ICOM'INARY IBLOCK'IBLOCK’INARYCNBLKS CLA CLL TAD \ICOM SNA JMP IDENT [SAMPLE ACQUISITION IS COMPLETE CLA CLL . IF(ICOM‘INARY)7:606 MPNTS'INARY‘NPNTS CALL UTAPEIIJIBLOCKJMPNTSOIE) ICOH'ICOM'INARY IBLOCK'IDLOCK’INARY‘NBLKS CLA CLL TAD \ICOM SNA JMP IDENT [SAMPLE ACQUISITION IS COMPLETE CLA CLL IF(ICOM-INARY)9.8.8 MMMMMMMMMMMMMCTMMM LOU) (DUI 265 NPNTSaINARY.NPNTs CALL VTAPE(1.IBLOCK.MPNTS.13) Icon-O CLA CLL TAD \NSPEC CIA TAD \NDONE SMA [IS NDONE>-NSPEC JMP IDENT [YES. FIND OUT NEXT THING To DO CLA CLL /No. TAKE RORE DATA RCNTR [READ ELAPSED LOST TIME DCA LOU [TEMPORARY STORAGE DCA CLCDE [DONT UAIT TOR CLOCK NExT TIME TAD (0200 ' DCA zLOC TAD (0600 DCA TEMP TAD»<|20| DCA wHICR JMP zEROC NPNTs-ICON+NPNTs CALL VTAPE(I.IBLOCK.MPNTS.II) CLA CLL JHP IDENT [SAMPLE ACOUISITION Is COMPLETE NPNTs-ICOR.NPNTS CALL UTAPE(I.IDLOCK.NPNTS.12) CLA CLL JMP IDENT ISARPLE ACOUISITION Is CONPLETE RPNTs-ICON.NPNTS CALL UTAPEII.IDLOCK.RPNTS.I3I CLA CLL ' JMP IDENT [SAMPLE ACOUISITION Is COMPLETE IOFSTINPNTS/IZB CLA CLL TAD VHICH [SET UP STARTING DATA ADDRESS TAD \NPNTS TAD \IOFST DCA UHICH SKOFE [CHECK FOR TININC ERROR SKP JNP ERROR CLOTE INC CLCDE CLA CLL TAD c0200 DCA zLOC [SET UP To zERO IUORR TAD (Taco DCA TEMP [STORE STOP ADDRESS DCA.LOU JMP zEROC ERRORoCLA CLL [TIMING ERROR URITE(I.II7) FORMAT(/'TIMING ERROR'I) RETURN ' EXTND.CLA CLL [DONE UITH DATA ACQUISITION RETURN END 000000 nonm GOO 000 266 Program VNTADS.FT 103 I02 VNTADS.FT DISPLAY PROGRAM FOR THE NON-FILE STRUCTURED VIDICON TIMED DATA ACQUISITION SERIES. TIMOTHY A. NIEMAN FEBRUARY 26. I975 COMMON IDIR.IDENT.IDARK.IREF.ISAMPoITIMD.ITIML‘ DIMENSION IDIR(I28).IDENT(32).IDARK($I2).IREF(5I2).ISAMP(SI2) DIMENSION ITIMD(I00).ITIML(I00) CALL-RTAPE(I.0.I28.IDIR) NRUNTOIDIR(I) 6057 GIVE OPTIONS VRITE(I.I08) ' ~ , FORMAT(/['DISPLAY OPTIONS:'l'G-NOTHING'I'I-SAMPLE'I'2-REFERENCE' ll'3-IT‘I'4-ABSORBANCE'I'SITIME DEVELOP'I'G-AXES'I'T-RETURN'II GET RUN TO DISPLAY URITE(I.I00) FORMAT(/J READ(I.IOI)IRUN FORMAT(‘DISPLAY RUN I '.IS) IF(NRUNT-IRUN)3.A.A URITE(ILI02)IRUN FORMAT(‘RUN I '.I5.' NOT FOUND') GO TO I ISBOIDIR(3fiIRUN-I) CALL RTAPE(I.ISB.32.IDENT) NPNTSIIDENT(2) NSPECIIDENT(I) NBLKS'NPNTS/I28 IBLOCK-ISBOZ CALL RTAPE(I.IBLOCK.NPNTS.IDARK) IBLOCK-ISBOZONBLKS CALL RTAPE(I.IBLOCK.NPNTS.IREF) AVERAGE DARK SIGNAL DAVGI'D J-NPNTS-I DO 5 Iizod DAVGIDAVGOFLOAT(IDARK(I)) DAVGIDAVGIFLOAT(J-I) FSPOSIZOAT. FSNm-'2047 o 000 I3 I04 I2 I05 I06 I01 I09 000 600 60I 602 603 604 6I0 605 606 GOO 000 700 UIUIUIU'UIUIUIUIUIMUI 70! U'IUIUIUIUI 267 INPUT DISPLAY OPTION. AND DISPLAY PARAMETERS READIIJIOAIIOPT J'NPNTS'I IF‘IOPT)IOI.I2 FORMAT(‘DISPLAY OPTION 0 'OIS’ GO TOIOOOJIA.I40I5110999)0IOPT READII.I05)ITH FORMAT(‘THRESHOLD 3 '.I5) THRIITM REAOIIOI06IISIZE FORMAT(‘LENGTM OF SMOOTH ' 'OISI READ(I.I07)SMIN.SMAX FORMAT(‘SCOPE BOTTOM I '0FI004/‘SCOPE TOP - '.FI004) Y5CALE'2047o/(SMAX‘SMIN) KSCALE'ZOATo/FLOATCJ‘I) READII.I09)IPLOT FORMAT(‘I'FAST. E'SLOU. G'RESTART '0I5) CO TOIOOGOIIIPLOT IBLOCK'ISB‘ZONBLKS¢2 DO 2 I'IONSPEC CALL RTAPE(I.IBLOCK.NPNTS.ISAMP) IBLOCK'IBLOCK’NBLKS DO 6|. K'ZoJ S'ISAMPIK) D-IDARKIK) RIIREF(K) ICODE'I GO TO(200030004000500).IOPT SCALE AND SMOOTH SIYSCALE‘(S-SMIN) IF(S)60I.602.602 5'0. IF(S-FSPOS)60A.604.603 S-FSPOS ISAMP(K)-S CONTINUE IF(ISIZE)606.606.605 ISAMP(I)IISAMP(2) CALL SMOTH(ISAMP.J.ISIZE) GO TO(700.000).IPLOT PLOT AXES CALL AXIS(I0.I0) GO TO I3 FAST PLOT DO 10! KI2.J IX-XSCALEtFLOAT(K-2) IY-ISAMP(K) CLA CLL CML RTR TAD \IK RAL CMA 606I CLA CLL CML RTR TAD \IY RAL CMA 6066 CLA CLL CONTINUE CLA CLL KSF SKP JMP \I3 HLT GO TO 2 GOO UIU‘IUIUI GOI S \002. 000 200 GOO 300 006 400 A02 A03 A01 404 (100 500 502 503 504 SOI 505 268 SLOU PLOT IKIO IYflISAMP(2) CALL XYST(IX.IY) DO 00I K33.J IXIXSCALEOFLOAT(K-2) IYOISAMP(K) CLA CLL KSF SKP JMP \802 CALL XYPLT(IX.IY) HLT GO TO 2 CONTINUE GO TO I3 SAMPLE SOS-DODAVG GO TO (600.2I).ICODE REFERENCE SIR°D+DAVG GO TO (600.2I).ICODE 1T S's-D R-R-D IF(R)AOI.AOI.A02 IF((R‘S)-THR)AOI.A03.403 T-I000.0(S[R) IF(T-I000.)AOA.AOA.AOI T-I000. SCT GO TO (600.2I).ICODE ABSORBANCE SIS‘D R-R'D IF(R)50I.50I.502 IF((R-S)'THR)SOI0503.$03 TIISIR) IF(T'I.)5001504.50I IFIT350I050I0505 TI]. . SO-ALOG(T).AGA.3 GO TO (OOOOEIIIICODE 000 IS 30 23 II2 2A IIA 22 71 I9 I6 113 I7 25 IIS 34 33 269 TIME DEVELOPMENT VRITE(I.I00) READ(I.|II)IOPT FORMAT(‘TIME DEVELOP USING OUTPUT OPTION . '.I5) IF(IOPT)30.I3.30 ' ICODE-a READ(I.II2)IPNT FORMAT(‘TIME DEVELOP POINT . '.IS) ITIIPNT-NPNTS>22.22.2. URITEII.IIAINPNTS FORMAT(‘ONLY '.IS.' POINTS/SPECTRA UERE TAKEN') GO To 23 READ(I.IO6)ISIZE Do I9 I-I.NSPEC IDLOCR-Isa.2.NaLRs.¢I.II CALL RTAPE(I.IBLOCK.NPNTS.ISAMP) ITIML(I)-ISAMP(I) IFIISIzE)I9.I9.11 CALL SMOTHCISAMP.NPNTS.ISIZE) ITIMD(I)-ISAMP(IPNT) TSNEG--2OAT. rspos-zoav. READCI.I13)J|.J2 FORMATC'FROM SPECTRA . '.I$['T0 SPECTRA . '.I5) IF(JI-NSPEC)I7.I?.25 IF(J2-NSPEC)I8.IB.2S VRITE(I.IIS)NSPEC FORMAT(‘ONLY '.I5.' sPECTRA HERE TAKEN') Go To ID TLOST-G. READ(I.IB6)ISIZE CALCULATE TIME SCALE LTIM-0 DO 33 I'JIIJE IFCITIMLII))33.33.34 LTIM'LTIM’I TLOSTITLOST2FLOATCITIMLIIII CONTINUE LBASE'IDENT‘II)‘6 IBASE'IDENT(9)'6 TIME'IDENTCIO) DASET'IOOOCIBASE DASEL'IOO'9LBASE TIME-TIME‘DASET TLOSTITLOSTfiaASEL READIIOIOTISMINDSMAX YSCALE'20470[(5MAK'SMIN) XSCALC'ZOGTo/(TLOST’TIME‘FLOAT‘JE’JI'LTIM)I 0.2047.[XSCALE URITEIIOIIOID FORMAT(/FIODA.’ SECONDS FULL SCALE'[) DO 29 I'JIPJE - D’IDARKCIPNT) RSInEFCIPNT) SIITIMDII) GO TOCZOOOOOOPAOOOSOO’aIOPT S’YSCALEOIS'SMIN) IF(S)26.27.27 5'00 IFCS’FSPOS)29.29.28 S'FSPOS ITIMDIII'S IF(ISIZE)32.32.3I CALL SMOTMCITIMDoNSPEC0ISIZE) 000 32 36 35 20 37 999 270 POINT PLOT IYIITIHD(JI) TLOST'CO TLOS'ITIML(JI) TLOST-TLOS‘BASELOTLOST IX'(TLOST)OXSCALE CALL XYSTCIXOIY) CALL XYEND I'JI‘I DO 20 J'IIJZ IY'ITIMD‘J) TLOS'ITIHL‘J) TLOST'TLOS'BASELOTLOST LTIH'G IF(ITIHL(J))35035036 LTIH'I . IX'CTLOST‘FLOAT‘J‘JI‘LTIH)‘TIHE)‘XSCALE CALL XYST(IXJIY) CALL XYEND URITE‘IOIIT) FORMATC'SLOPE') READ(IJII3)JIOJ2 CALL LLSQ(XIY:DELX:DELYII) DO 37 J'JIOJQ X'FLOATCJ)*TIHE YflITIMD(J) Y'CY/YSCAL5305MIN YaY/IZGOO CALL LLSQ‘XOYIDELXIDELYJZ) CALL LLSQ(XIYIDELXIDELY03) URITE‘IPII8)XODELX:YIDELY FORMAT('SLOPE . 'OEIOOCO' 0' 'OEI6080' ISEC'/ I'INTERCEPT - ',£16.8.' +- ‘oEI6-B) ' “ GO To I5 CALL EXIT END 271 Subroutine PTPLT.FT 00000 U'IU'IUIUIUIUIUIUIUIUIUI SUBROUTINE PTPLTIJXPJY) TIMOTHY A. NIEMAN DECEMBER II: I914 POINT PLOTS ON THE DISPLAY SCOPE IXIJX IYUJY CLA CLL CML RTR TAD \IX RAL CMA 606I ILOAD X CLA CLL CML RTR ' TAD \IY RAL CMA 6066 ILOAD Y AND INTDJSIFY CLA CLL RETURN END Subroutine LLSQ.FT OCOOOOOOCCCGOGO SUBROUTINE LLSQ(XOYIXDEVIYDEVIHODE’ LINEAR LEAST SQUARES FOR Y'AX‘B TIMOTHY A. NIEMAN FEBRUARY 27: I975 MODES I=ZERO SUMS 2-ACCUNULATE SUMS 3'CALCULATE VALUES RETURNED VALUES X-SLOPE Y-INTERCEPT XDEVSSLOPE STANDARD DEVIATION DYDEV-INTERCEPT STANDARD DEVIATION GO TO (I:2:3)OMODE SUI‘IKBG o SUMY‘Q o SUMXSQIO. SUMYSQ-z. SUMKYIB. XNUH‘G o RETURN SUMK-SUMX+X SUMY-SUMY+Y SUMNSQISUMXSQ+CX¢X) SUMYSQ-SUMYSQ§(YtY) SUHXY'SUMXY+(XtY) XNUM=XNUM+I. .IETUWN X-(XNUM‘SUHYY)-(SUMXISUMY) XIX/I(XNUMtSUMXSQ)-(SUMX'SUMX)) Y-(SUMY/XNUM)-Xt(SUMX/XNUM) XDEV8(SUMYY-(SUMX‘SUMY)/XNUH).¢2 XDEV=XDEV/(SUNXSO-(SUMXtSUMX)/XNUM) XDEVISUMYSQ-ISUMYOSUMY/XNUMI-XDEV SSXDEV/(XNUMTZo) . XOEV35/(SUHXSQ-(SUMXtSUMX/XNUMI) XDEV'SORT(XDEV)' ' YDEVI(StSUMXSQ)l(XNUMtSUMXSQ*SUHX¢SU€() YDEVISQRTCYDZV) RETURN END REFERENCES 10. 11. 12. 13. 14. 15. 16. 17. 18. REFERENCES L.M. Biberman and S. Nudelman, eds., Photoelectronic Imaging Devices, Vol. 2, Plenum Press, New York, 197l, pp. 5-ll. Ibid., PP. 253-57. Ibid., PP. 303-6. R.E. Johnson, Applications of RCA Silicon Diode Array Target Vidicons, Camera Tube Application Note AN-4623, RCA Electronics Components, Harrison, N.J., 197l. D.G. Mitchell, K.N. Jackson, K.M. Aldous, Anal. Chem., §§(l4), 1215A (1973). K.W. Busch and G.H. Morrison, Anal. Chem., g§(8), 712A (l973). J.T. Agnew, R.G. Franklin, R.E. Benn, and A. Bazarian, J. Opt. Soc. Am., .§g(5), 409 (1949). R.E. Benn, N.S. Foote, and C.T. Chase, J. Opt. Soc. Am., 32(7), 529 (1949). J.E. Anderson, Rev. Sci. Inst., §Z(9), 12l4 (1966). S.A. Johnson, H.M. Fairbank, and A.L. Schawlow, Appl. Opt., 19(10), 2259 (1971). G.E. Lightner, A291. Opt., §(3), 439 (1966). R.E. Santini, M.J. Milano, and H.L. Pardue, Anal. Chem., fl§(ll), 915A (1973). L.M. Biberman and S. Nudelman, eds., Photoelectronic ImagjngADevices, Vol. 2. PP. 193-202. Ibid., PP. 23-24. Ibid., pp. 203-15. Ibid., p. 30. Ibid., pp. 2l7-5l. C.M. Savage and P.D. Maker, Appl. Opt., 19(4), 965 (l97l). 272 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 273 M. Margoshes, paper presented at Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy. 1970. L.M. Biberman and S. Nudelman, eds., Photoelectronic Imaging Devices, Vol. 2, pp. 263-73. P. Gloerson, J. Opt. Soc. Am., 4800’), 712 (1958). M. Delhaye, Appl. Opt.,.1(1l), 2195 (1968). R.E. Santini, M.J. Milano, H.L. Pardue, and D.w. Margerum, Anal. Chem., 44(4), 826 (1972). Data sheet for 7735A, 7735 and 4478 vidicons, RCA Electronic Components, Harrison, M.J. L.M. Biberman and S. Nudelman, eds., Photoelectronic Imaging Devices, Vol. 2. pp. 275-98. Camera Tube Names, Camera Tube Application Note AN-4795, RCA E1ectronic‘COmponents, Harrison, N.J., 1971. K.H. Jackson, K.M. Aldous, and 0.6. Mitchell, Spect. Let., 6(6), 315 (1973). K.M. Busch, N.G. Howell, and G.H. Morrison, Anal. Chem., 46(4), 575 (1974). M.J. Milano, H.L. Pardue, T.E. Cook, R.E. Santini, D.H. Margerum, and J.M.T. Raycheba, Anal. Chem., 46(3), 374 (1974). K.M. Jackson, K.M. Aldous, and 0.6. Mitchell, Appl. Spect., 28(6), 569 (1974). D.O. Knapp. N. Omenotto, L.P. Hart, F.H. Plankey, and J.D. Winefordner, Ana. Chim. Acta, 62(2), 455 (1974). T.E. Cook, M.J. Milano, and H.L. Pardue, Clin. Chem., gQ(11), 1422 (1974). K.M. Busch, N.G. Howell, and G.H. Morrison, Anal. Chem., 46(14), 2074 (1974). M.J. Mi1ano and H.L. Pardue, Anal. Chem., 41(1), 25 (1975). M.J. Milano and H.L. Pardue, Clin. Chem., 21(2), 211 (1975). Data sheet for 4532A and 4532 vidicons, RCA Electronic Components, Harrison, N.J. L.M. Biberman and S. Nudelman, eds., Photoelectronic Imagjng Devices, V01. 2. PP. 253-62. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 274 K.M. Busch, M.G. Howell, and G.H. Morrison, Anal. Chem., 4§(9), 1231 (1974). Data sheet on model 1205 Optical Multichannel Analyzer, SSR Instruments, Santa Monica, Ca. L.M. Biberman and S. Nudelman, eds., Photoelectronic Imggjng_0evices, V01. 2, pp. 179-181, 493-8. R.A. Harber and G.E. Sonnek, A221. 02t., 2(6), 1039 (1966). D.J. Baker and A.J. Steed, A221. 02t., 2(11), 2190 (1968). 0.”. Golightly, R.N. Kniseley, and V.A. Fassel, Spect. Acta , 222(9), 451 (1970). A. Danielsson, P. Lindblom, and E. Soderman, Chem. Scr., 2(1), 5 (1974). A. Danielsson, paper no. 311 presented at Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy. 1975. L.M. Biberman and S. Nudelman, eds., Photoelectronic Imaging Devices, Vol. 2, pp. 119-124. M.G. Hyzer, 222, 22(12), 34 (1974). L.M. Biberman and S. Nudelman, eds., Photoelectronic Imaging Devices, Vol. 2, pp. 453-79. G. Horlick and E.G. Codding, Anal. Chem., 52(8), 1490 (1973). G. Horlick and E.G. Codding, Anal. Chem., 52(9), 1749 (1973). E.G. Codding and G. Horlick, Appl. Spect., 22(5), 366 (1973). E.G. Codding and G. Horlick, Spect. Let., 2(1), 33 (1974). G. Horlick and E.G. Codding, paper no. 25 presented at Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, 1973. G. Horlick and E.G. Codding, Anal. Chem., 32(1), 133 (1974). G. Horlick, E.G. Codding, and S.T. Leung, App1.,§pect., 22(1), 48 (1975). G. Horlick and E.G. Codding, Appl. §pect., 22(2), 167 (1975). Data sheets on self-scanning photodiode arrays, Reticon, Mountain View, Ca. P.K. Weimer, M.G. Kovac, F.V. Shallcross, and w.s. Pike, IEEE Trans., ED-18(1l), 996 (1971). 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 275 .H. GiIder, E1ect. Des., 21(1), 36 (1973). J L. Altman, Elect., 55(4), 62 (1972). L. Boonstra and F.L.J. Sangster, Elect., 55(4), 64 (1972). G .F. Amelio, M.J. Bertram, and M.F. Tompsett, IEEE Trans., ED-18(11), 986 (1971). M.F. Tompsett, M.J. Bertram, D.A. Sea1er and C.H. Sequin, Elect., 55(2), 162 (1973). G.F. Ame1io, Sc. Am., 230(2), 23 (1974). R. Me1on, E1ect., 55(6), 106 (1973). Data sheets on CCDlOl, CCDllO, and CCDZOl, Fairchild Semiconductor, Mountain View, Ca. Data sheet on 51051232, RCA, Lancaster, Pa. H.A. Snow, U.S. Pat. 2,240,722, May 6, 1971. P. Gloersen, J. Opt. Soc. Am., 52(10), 712 (1958). M. Margoshes, Spect. Acta , 252(3), 113 (1970). M. Margoshes, paper presented at Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, 1970. M. Margoshes, Opt. §pect., 5(11), 26 (1970). A. Danielsson and P. Lindblom, Chem. Scr., 5, 227 (1972). M. Delhaye, A221. 02t., 2(11), 2195 (1968). C.M. Savage and P.D. Maker, A221. 02t., 52(4), 965 (1971). E.A. Beaver and C.E. McIlwain, Rev. Sci. Inst., 52(9), 1321 (1971). P.H.J.M. Boumans and G. Brouwer, 5pect. Acta, 278(6), 247 (1972). P.N.J.M. Boumans, R.F. Runphorts, L. Hillemsen, and F.J. deBoer, Spect. Acta, 288(7), 227 (1973). F.w. Karasek, Opt. Spect., 25(1), 47 (1972). G.G. Olson, Opt. Spect., 5(1), 38 (1972). G.G. Olson, Amer. Lab,, 5(2), 57 (1972). 276 82. P. Schierer, paper no. 51, presented at the Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, 1972. 83. J.M. Marrs, paper no. 48, ibid. 84. J.M. Marrs, Amer. Lab., 5(11), 57 (1972). 85. P. Burke, Res.[0ev., 25(4), 24 (1973). 86. R.E. Santini, M.J. Milano, H.L. Pardue, and D.N. Margerum, paper no. 227, presented at the Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy. 1972. 87. G. Horlick, H.K. Yuen, and K.R. Betty, paper no. 392, ibid., 1975. 88. K.M. Aldous and 0.6. Mitchell, paper no. 83, ibid., 1973. 89. D.L. Hood, A.B. Dargis, and D.L. Nash, paper no. 310, ibid., 1975. 90. M.J. Dreyer, A. Kuppermann, H.G. Boettger, C.E. Giffin, D.D. Norris, S.L. Grotch, and L.P. Theard, Clin. Chem., 22(8), 988 (1974). 91. Programs written by B.K. Hahn. 92. H.V. Malmstadt, C.G. Enke, S.R. Crouch, and G. Horlick, Optimization of Electronic Measurements, Module 4, H.A. Benjamin, Inc., Menlo Park, Ca.,*l974. pp. 144-148. 93. A. Savitzky and M.J.E. Golay, Anal. Chem., 55(8), 1627 (1964). 94. H.C. Hamilton, Statistics in Physical Sciences, Ronald Press Co., New York, 1964, p. 130. 95. R.R. Ernst, "Sensitivity Enhancement in Magnetic Resonance", in Advances in Magnetic Resonance, J.S. Waugh, Ed., Academic Press, Now York,-1966, Vol.2, pp. 34-50. 96. H.P. Yule, Anal. Chem., 55(1), 103 (1966). 97. H.P. Yule, Trans. Am. Nucl. Soc.,_15, 38 (1967). 98. H.P. Yule, Nucl. Inst. & Meth., 55, 61 (1967). 99. R.N. Jones, R. Venkataraghavan, and J.M. Hopkins, 5pect. Acta , .225, 941 (1967). 100. K. Yamashita and S. Minami, Jgp, J. Appl. Phys., 5(12), 1505 (1969). 101. J.P. Porchet and Hs. H. Gunthard, J. Phys., E, 2, 261 (1970). 102. M. Caprini, S. COhn-Sfetcu, and A. Manof, IEEE Trans., AU-18(4), 389 (1970). 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 277 K. Yamashita and S. Minami, Jap. J. Appl. Phy§,, 55(8), 1097 (1971). H.P. Yule, Anal. Chem., 55(7), 1245 (1972). M.J.E. 601ay, IEEE Trans., 521, 299 (1972). H. Tominaga, M. Dojyo, and M. Tanaka, Nucl. Inst. & Meth., 25, 69 (1972). T.H. Edwards and P.D. Hillson, Appl. Spect., 25(6), 541 (1974). T.A. Maldacker, J.E. Davis, and L.B. Rogers, Anal. Chem., 55(6), 637 (1974). H.V. Malmstadt, C.G. Enke, S.R. Crouch, and G. Hor1ick, Optimization of Electronic Measurements, Module 4, pp. 8-9. H.C. Hamilton, Statistics in Physical Sciences, p. 67. D.A. Buyske and V. Downing, Anal. Chem., 52(13), 1798 (1960). D.J. Donaldson, F.J. Normand, G.L. Drake, and H.A. Reeves, U.S. De t. A r. A r. Res. Serv. Re . , ARS 72-98, 68-9 (1972). origina not avai a e - Chem. Abstr., 55, 28245 (1974).) V.I. Lukin, Tr., Sakhalin, Obl. Sta. Zashch. Rast., 1970 (1), 19. (original not available - Chem. Abstr., 15,1019 (1973).) V. Tupalov, Rastenievud. Nauki, 2(5), 145 (1972). (original not available - Chem. Abstr., Z5, 12560 (1973). S. Fukuda and H. Kunisho, Ja an Kokai, 66, 590 (l974).(original not available - Chem. Abstr., 52, 32715 (1975).) ' N. Irving Sax, Dangerous Properties of Industrial Materials, 2d ed., Reinhold, New York, 1963. pp. 648-49. Data sheet on Cyanamide-50, American Cyanamid Co., Wayne, N.J. Y.I. Mushkin, Zavod. Lab., 55(3), 296 (l967).(original not available - Chem. Abstr., 57_, 39959 (1967).) F.J. Holler, "A Spectrochemical Investigation of Cyanamide and Sodium Pentacyanoammineferrate (II) in Aqueous Solution", (M.S. Thesis, Ball State University, 1972). H.A. Laitinen, Chemical Analysis, McGraw-Hi11, New York, 1960, pp. 214-216. H.R. Fearon, Analyst, 21, 562 (1946). 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 278 K.A. Hofmann, Justus Liebigs Ann. Chem., 312, 1 (1900). 0.0. Perrin, Dissociation Constants of Organic Bases in Aqueous Solution, Butterworths, London, 1965, p. 442. F.J. Hol1er, private communication. F.J.C. Rossotti and H. Rossotti, The Determination of Stability Constants, McGraw-Hill, New York, 1961, pp. 45-51. H.E. Toma and J.M. Malin, Inorg, Chem., 55(7), 1772 (1974). W.R. Seitz and M.P. Neary, Anal. Chem., 55(2), 188A (1974). U. Isacsson and G. Wettermark, Anal. Chim. Acta , 55(2), 339 (1974). H.0. A1brecht, Z. Phys. Chem., 155, 321 (1928). E.H. White, 0. Zafiriou, H.H. Kagi, and J.M. Hill, J. Am. Chem. Soc., 55, 940 (1964). E.H. White and M.M. Bursey, ibid., p. 941. M.P. Beck and F. J06, Photochem. & Photobiol., 15(6), 491 (1972). J.D. Gorsuch and D.M. Hercules, Photochem. & Photobiol., 15(6), 567 (1972). J. Lee and H.H. Seliger, Photochem. & Photobiol.,55(2), 227 (1972). K.J. Caserta. "A Computer-Interactive Bipolar Pulse Conductance System", (PhD Dissertation, Michigan State University, 1974). R.E. Santini and H.L. Pardue, Anal. Chem., 52(7), 706 (1970). Controller design by B.K. Hahn; Interface design by T.G. Kelly. Interface design by M. Rabb and B.K. Hahn. Design by R. Davis and B.K. Hahn. Design by T.A. Last. B.K. Hahn and C.G. Enke, Anal. Chem., 55(7), 651A (1973). Redesigned and rebuilt by E.J. Darland. T.A..Nieman, and F.J. Holler. Product bulletin 595—1422, Computer Interface ADD, Heath-Schlumberger, Benton Harbor, Michigan, 1972. G. Davis, Radio Electronics, 23 (July, 1969). 279 145. Digital Equipment Corp., Introduction to Programming, Digital Equipment Corp., Maynard, Mass., 1972, Chap. 9. 146. Digital Equipment Corp., ProgrammingLanguages, Digital Equipment Corp., Maynard, Mass., 1972, Chap.C13 and 15. 147. Ibid., pp. 15-73 to 15-76. 148. Product bulletin 595—1208, Instrumentation for Spectroscopy EU 700 Series, Heath-Schlumberger, Benton Harbor, Michigan, 1969. 149. Y. Ta1mi, Ana1. Chem., 51(7), 658A (1975). 150. v. Ta1mi, Anal. Chem., 4_7_(7), 697A (1975). 151. W.J. Youden, Statistical Methods for Chemists, John Wiley, New York, 1951. pp. 40-45. 152. H.J. Larson, Introduction to Probability Theory, John Wiley, New York, 1969. pp. 332-39. 153. T.E. Hull and D.D.F. Day, Computers and Problem Solving, Addison-Wesley, Ontario, 1970, pp. 137-39. 154. M. Dwass, Probability and Statistics, W.A. Benjamin, New York, 1970. pp. 297, 318, 329-30. 155. W.J. Conover, Practical Nonparametric Statistics, John Wiley, New York, 1971, pp. 186-95. .i #9! n H P ..I1ldu 5 IF 7. #25... . .LC sort... ‘4'. (v $101.4. 1). ‘IIL ..