Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Chemical acoustic emission analysis of the Briggs-Rauscher iodine clock Brock, Ivan Heinz 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1995-059243.pdf [ 2.99MB ]
Metadata
JSON: 831-1.0059651.json
JSON-LD: 831-1.0059651-ld.json
RDF/XML (Pretty): 831-1.0059651-rdf.xml
RDF/JSON: 831-1.0059651-rdf.json
Turtle: 831-1.0059651-turtle.txt
N-Triples: 831-1.0059651-rdf-ntriples.txt
Original Record: 831-1.0059651-source.json
Full Text
831-1.0059651-fulltext.txt
Citation
831-1.0059651.ris

Full Text

Chemical Acoustic Emission Analysisof theBriggs-Rauscher Iodine ClockbyIvan Heinz BrockB. Sc. University of Toronto, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinThe Faculty of Graduate StudiesDepartment of ChemistryWe accept this thesis as conformingto the required standardThe University ofBritish ColumbiaSeptember 1995© Ivan Heinz Brock, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of______________The University of British ColuibiaVancouver, CanadaDate____(Signature)DE-6 (2/88)IIAB STRACTThe oscillations observed in the Briggs-Rauscher Iodine Clock were studied usingchemical acoustic emission (CAE) analysis. The experimental system evolved through threeseparate designs. The first early design employed a batch delivery system and utilized apiezoelectric transducer as the CAE sensor. The second design employed a flow deliverysystem and the oscillating reaction was monitored simultaneously by CAE, ultraviolet-visible spectrophotometry, and an ion-selective electrode. Sampling periods of 4.0 secondsmade this a “low resolution” system. The third design was an optimized version of thesecond, with shorter sampling periods (1.0 seconds), which led to higher resolution data.The concentrations of the reagents used in the first system were those recommendedby Briggs and Rauscher. Four unique phases of the oscillator were observed. The peakslopes during the positive slope phase increased according to second order kinetics, and thedecreasing peak slopes during the negative slope phase were found to be independent ofreagent concentration. Analysis of the averaged power spectra revealed two distinctfrequency regions. Chemometric analysis successfully identified signals due to noise.A series of experiments were conducted in which [iodate] varied over a range of Ca.0.004 to 0.037 M at three different temperatures: 25, 30 and 35°C. It was discovered that thenumber of oscillations (both CAE and iodine oscillations) was independent of temperature.The CAB peak rates (increase and decrease) were also found to be independent oftemperature while the iodine production and consumption rates were found to ca. triple witheach five degree rise in temperature. Integrated CAE d.c. rates followed second orderkinetics.Another series of experiments were conducted in which H20 was varied over therange of ca. 0.078 to 0.78 M at 35°C, The number of observed CAE oscillations was foundIIIto be independent of initial [H20] while the number of iodine oscillations increased asinitial [H20]increased. Integrated CAE d.c. rates followed no simple trend.Under the conditions examined, a minimum [iodate] of 0.0 19 M, and [H20] of0.16 M was required for oscillations to commence. Periodicity in the CAE data was notedin those experiments which did not strongly oscillate and attributed to a dissolved oxygenmodel.A refinement of the currently accepted mechanism is proposed and the utility ofCAE as a tool to investigate oscillatory kinetics is discussed.ivTable of ContentsAbstract ii1’J. 0P .,de#sList Of Tables xList Of Figures xiiGlossary xviChapter One - Introduction 11.1 Acoustic Emission 11.2 Chemical Acoustic Emission 21.3 Acoustic Emission Detection and Analysis 31.3.1 CAE Instrumentation 41.3.2 CAE Analysis 51.3.2.1 Data Analysis Protocol 61.3.2.2 Multivariate Statistics 91.4 Oscillating Chemical Reactions 121.4.1 History 121.4.2 General Theory 141.4.3 Briggs-Rauscher Oscillator 171.4.3.2 Heat Production 191.5. CAE and the Briggs-Rauscher Iodine Clock 201.5.1 Bubble Nucleation Theory 20VChapter Two - Preliminary Studies Using Batch Methods 232.1 Introduction 232.2 Experimental 242.2.1 Briggs-Rauscher Chemical System 242.2.2 Instrumental System 252.2.3 Timing and Methods 272.2.3.1 Stirring Effects 282.2.3.2 Sampling Frequency 282.2.4 Sensor Calibration 292.3 Discussion 302.3.1 Analysis of d.c. Signals 312.3.1.1 Induction phase 352.3.1.2 Negative Slope (neg) Phase 392.3.1.3 Positive Slope (pos) Phase 402.3.1.4 Shoulder Phase 422.3.2 Analysis of a.c. Signals 442.3.2.1 Signal Validation 442.3.2.2 Signal Analysis 442.3.2.3 Time Series Analysis 442.3.2.4 Descriptor Extraction and Analysis 522.3.2.5 Principal Components Analysis 55vi2.3.2.6 Clustering .552.4 Conclusions 58Chapter Three - lodate Studies Using Stopped Flow Methods 613.1 Introduction 613.2 Experimental 633.2.1 Briggs-Rauscher Chemical System 633.2.2 Instrumental System 643.2.2.1 Constant Temperature Bath 653.2.2.2 Flow Injection Development and Optimization System (FIDO) 663.2.2.3 Reaction Vessel 693.2.2.4 Ion Selective Electrode (ISE) 703.2.2.5 Primary Computer 703.2.2.6 Secondary Computer 703.2.3 Methods 713.2.3.1 Reaction Temperature 713.2.3.2 Reagents and Concentrations 713.3 Series PF-G, lodate varied from 0.005 to 0.050 M, 20°C 733.3.1 Data Work-up 743.3.2PF-G-1 and2 753.3.3 PF-G-3 and 4 763.3.4 PF-G-5 through 10 81VII3.3.4.1 Peak Count and Frequency .813.3.4.2 Peak Reaction Rates 843.3.4.3 Phase Relationships 853.3.4.4 Extent of Reaction 873.3.5 Series PF-G: Conclusions 923.4 Series PF-K, Jodate varied from 0.004 to 0.037 M, 25°C 943.4.1 Data Work-up 953.4.2 PF-K-1 and 2 953.4.3 PF-K-3 and 4 963.4.4 PF-K-5 through 10 993.4.4.1 Peak Count and Frequency 993.4.4.2 Peak Reaction Rates 1013.4.4.3 Extent of Reaction 1033.4.5 Series PF-K: Conclusions 1063.5 Series PF-O, lodate varied from 0.004 to 0.037 M, 30°C 1083.5.1 Data Work-up 1083.5.2PF-O-1 and2 1093.5.3 PF-O-3 and4 1093.5.3.1 Autocorrelation 1093.5.4 PF-O-5 through 10 1113.5.4.1 Peak Count and Frequency 111VIII3.5.4.2 Peak Reaction Rates 1123.5.4.3 Extent of Reaction 1143.5.5 Series PF-O: Conclusions 1203.6 Series PF-T, lodate varied from 0.004 to 0.037 M, 35°C 1213.6.1 Data Work-up 1223.6.2 PF-T-1 and 2 1223.6.3 PF-T-3 and 4 1223.6.3.1 Autocorrelation 1233.6.4 PF-T-5 through 10 1233.6.4.1 Peak Count and Frequency 1253.6.4.2 Peak Reaction Rates 1283.6.4.3 Extent of Reaction 1303.6.5 Series PF-T: Conclusions 134Chapter Four - Hydrogen Peroxide Study Using Stopped Flow Methods 1364.1 Introduction 1364.2 Series PF-S, Hydrogen Peroxide varied from 0.078 to 0.78 M, 35°C 1374.2.1 Data Work-up 1374.2.2 PF-S-1 and 2 1374.2.3 PF-S-3 and 4 1384.2.3.1 Autocorrelation 1384.2.4 PF-S-5 through 10 141ix4.2.4.1 Peak Count and Frequency.1414.2.4.2 Peak Reaction Rates 1434.2.4.3 Extent of Reaction 1444.2.5 Series PF-S: Conclusions 149Chapter Five - Conclusions 1515.1 Proposed Skeleton Mechanism of the BR Oscillator 1515.1.1 Iodine Production 1515.1.2 Iodine Consumption 1525.1.3 Reaction Scheme 1535.2 The Rate Limiting Step 1545.3 CAE as an Analytical Tool to Study Oscillating Reactions 155Chapter Six - Further Work 157References 160Appendix A - List ofManufacturers 163Appendix B - Software Developed 164Appendix C - Experimental Data 167xList of TablesTable 1-1 - Descriptors used to characterize individual CAE signals 11Table 2-1. Reagents used for batch studies 24Table 2-2. Apparatus list for batch studies 26Table 2-3. Phases present in Reagent Set A data, sorted by time 34Table 2-4. Phases present in Reagent Set A data, sorted by phase 35Table 2-5. Neg Phase Slopes for Run 1 40Table 2-8. Pos phase sopes for Run 1 41Table 2-7. Summary of peaks in reaction activity region of Run 1 49Table 2-8. Results ofWilcoxon Test (Reagent Sets A vs. B) 53Table 2-9. Results ofWilcoxon Test for Reagent Set A (run 1 vs 2) 54Table 3-1. UV-VIS absorption parameters for BR species 62Table 3-2. Reagents used for batch studies 63Table 3-3. Apparatus list for iodate studies 65Table 3-4. Flow system components 67Table 3-5. Experimental Series 72Table 3-6. Instrument parameters for series PF-G 74Table 3-7. Peak parameters for Series PF-G 82Table 3-8. Reaction rates for Pos and Neg events for Series PF-G 84Table 3-9. Curve fit results of integrated CAE d.c. data 88xiTable 3-10. Instrument parameters for series PF-K 95Table 3-11. Peak parameters for series PF-K 100Table 3-12. Reaction rates for Pos and Neg events 103Table 3-13. Instrument parameters for series PF-O 108Table 3-14. Peak parameters for series PF-O 111Table 3-15. Reaction rates for Pos and Neg events 112Table 3-16. Iodine consumptionlproduction ratios 113Table 3-17. Interpretation of transitions in integrated CAE d.c. data 118Table 3-18. Instrument parameters for series PF-T 121Table 3-19. Peak parameters for Series PF-T 126Table 3-20. Reaction rates for Pos and Neg events 126Table 3-2 1. Iodine consumption/production ratios 129Table 3-22. Interpretation of transitions in integrated CAE d.c. data 131Table 4-1. Instrument parameters for series PF-S 137Table 4-2. Peak parameters for Series PF-S 141Table 4-3. Reaction rates for Pos and Neg events 142Table 4-4. Second order curve fit results for integrated CAE d.c. data 147xiiList of FiguresFigure 1-1. Block diagram of basic experimental hardware 4Figure 1-2. CAE Data Analysis Road map 7Figure 1-3. Oscillating reaction publications 13Figure 2-1. Apparatus for preliminary batch studies 26Figure 2-2. Apparatus for calibration procedure developed for batch studies 30Figure 2-3. Colour changes in relation to CAE d.c 32Figure 2-4. Plots of Response vs Time(top) and LN(Responseadj) vs Time(bottom) 38Figure 2-5. Fit of l/(Pos phase rates) vs Event Number 41Figure 2-6. AVP of Reagent Set B 45Figure 2-7. AVP of Reagent Set B, with expanded Y axis 46Figure 2-8. AVP of Reagent Set A, Run 1 47Figure 2-9. Reaction activity region of Reagent Set A, Run 1 48Figure 2-10. AVP of Reagent Set A, Run 2 50Figure 2-11. Reaction activity region of Reagent Set A, Run 2 51Figure 2-12. Power spectra of K-Means clustering results for Reagent Set B 57Figure 3-1. Apparatus for stopped-flow studies. The acronyms are defined in Glossary. ... 64Figure 3-2. Schematic of the flow system 67Figure 3-3, The reaction vessel in detail 69XIIIFigure 3-4. Plots of LN([iodine]) and 1/[iodine] vs Time for runs PF-G-2, PF-G-3 andPF-G-4. In all cases the induction period data has been removed 78Figure 3-5. Autocorrelograms ofUV-Vis data for series PF-G-2, 3 and 4 80Figure 3-6. Reaction rates for different initial iodate concentrations 85Figure 3-7. Observed phase relationships:[iodine] (top), CAE d.c. (middle) and [iodide](bottom) 87Figure 3-8. Top:integrated CAE d.c. signal vs. reaction time. The numbers beside eachcurve identify the experimental run. Bottom: integrated CAE d.c. (at time=600 sec) vsinitial iodate concentration 90Figure 3-9. Plot of LN(integrated CAE d.c.) vs initial iodate concentration 91Figure 3-10. Autocorrelograms of CAE d.c. data for series PF-K-1 through 4. The timeconstant is six seconds 98Figure 3-11. Reaction rates for different initial iodate concentrations 102Figure 3-12. Top: integrated CAE d.c. signal vs. reaction time. The numbers beside eachcurve identify the experimental run. Bottom: integrated CAE d.c. (at time=480 see) vsinitial iodate concentration 105Figure 3-13. Autocorrelograms of CAE d.c. data for Series PF-O-l through 4. The timeconstant is four seconds 110Figure 3-14. Reaction rates for different initial iodate concentrations 114Figure 3-15. Integrated CAE d.c. data fort = 480 seconds, runs ito 6 116Figure 3-16. Integrated CAE d.c. data fort = 480 seconds, runs 7 to 10 117Figure 3-17. Integrated CAE d.c. data at time = 480 s 119xivFigure 3-18. Autocorrelograms of CAE d.c. data for series PF-T-1 through 4. The timeconstant is one second 124Figure 3-19. Reaction rates for different initial iodate concentrations 129Figure 3-20. Integrated CAE d.c. data for PF-T-6 through 9 130Figure 3-21. Integrated CAE d.c. data at time = 480 s 132Figure 4-1. Autocorrelograms of CAE d.c. data for series PF-S-i through 4. The timeconstant is two seconds 140Figure 4-2. Reaction rates for different initialH20concentrations 144Figure 4-3. Integrated CAE d.c. data for PF-S-1 through 6 146Figure 4-4. Integrated CAE d.c. data for PF-S-7 through 10 148Figure C-i. Chart recording of Reagent Set A, Run 1 - oscillations present 164Figure C-2. Chart recording of Reagent Set A, Run 2 - oscillations present 165Figure C-3. Chart recording of Reagent Set B - no oscillations present 166Figure C-4. Digitized chart recordings of Reagent Set A 167Figure C-5. Data from series PF-G-1 and 2: [iodine] (top), CAE d.c. (middle dotted) and[iodide] (bottom) 168Figure C-6. Data from series PF-G-3 and 4: [iodine] (top), CAE d.c. (middle dotted) and[iodide] (bottom) 169Figure C-7. Data from series PF-G-5 and 6: [iodine] (top), CAE d.c. (middle dotted) and[iodide] (bottom) 170Figure C-8. Data from series PF-G-7 and 8: [iodine] (top), CAE d.c. (middle dotted) and[iodide] (bottom) 171xvFigure C-9. Data from series PF-G-9 and 10: [iodine] (top), CAE d.c. (bottom) 172Figure C-l0. Data from series PF-K-l and 2: [iodine] (top), CAE d.c. (bottom) 173Figure C-il. Data from series PF-K-3 and 4: [iodine] (top), CAE d.c. (bottom) 174Figure C-12. Data from series PF-K-5 and 6: [iodine] (top), CAE d.c. (bottom) 175Figure C-13. Data from series PF-K-7 and 8: [iodine] (top), CAE d.c. (bottom) 176Figure C-l4. Data from series PF-K-9 and 10: [iodine] (top), CAE d.c. (bottom) 177Figure C-15. Data from series PF-O-l and 2: [iodine] (top), CAE d.c. (bottom) 178Figure C-16. Data from series PF-O-3 and 4: [iodine] (top), CAE d.c. (bottom) 179Figure C-17. Data from series PF-O-5 and 6: [iodine] (top), CAE d.c. (bottom) 180Figure C-18. Data from series PF-O-7 and 8: [iodine] (top), CAE d.c. (bottom) 181Figure C-19. Data from series PF-O-9 and 10: [iodine] (top), CAE d.c. (bottom) 182Figure C-19. Data from series PF-T-1 and 2: [iodine] (top), CAE d.c. (bottom) 183Figure C-20. Data from series PF-T-3 and 4: [iodine] (top), CAE d.c. (bottom) 184Figure C-21. Data from series PF-T-5 and 6: [iodine] (top), CAE d.c. (bottom) 185Figure C-22. Data from series PF-T-7 and 8: [iodine] (top), CAE d.c. (bottom) 186Figure C-23. Data from series PF-T-9 and 10: [iodine] (top), CAE d.c. (bottom) 187Figure C-24. Data from series PF-S-1 and 2: [iodine] (top), CAE d.c. (bottom) 188Figure C-25. Data from series PF-S-3 and 4: [iodine] (top), CAE d.c. (bottom) 189Figure C-26. Data from series PF-S-5 and 6: [iodine] (top), CAE d.c. (bottom) 190Figure C-27. Data from series PF-S-7 and 8: [iodine] (top), CAE d.c. (bottom) 191Figure C-28. Data from series PF-S-9 and 10: [iodine] (top), CAE d.c. (bottom) 192xviGLOSSARYa.c. (alternating current).aid analog to digital convertor, see DACA.CAE (chemical acoustic emission).CV % coefficients of variance, expressed as a percentage. This parameter is used togauge the accuracy of a fitted curve. It is computed as the normalizedversion of the standard error:CV % = 100% * std.err. in parameter valued/a digital to analog convertor, see DACA.DACA (data acquisition and control adapter). The primary interface used in thiswork. It has a max. 15 kHz sampling rate, 4 (12-bit) a/d devices, 2 (12-bit)d/a devices and 16 binary (TTL) i/o lines.dB (decibel). The logarithm of the ratio of the output signal to the input signal,multiplied by 20. Mathematically, dB = 20 log (I/Ia).digitize render a continuous signal into a set of [evenly spaced] discrete data points.FIDO (flow injection development and optimization). A flow injection analysissystem developed mostly in-house.GPIB (general purpose interface bus). A standard for transmitting data and controlcommands between electronic equipment. Based on the IEEE-488specification.HPIB (Hewlett-Packard interface bus). HP’s version of the GPIB (above).IEEE-488 see GPIB.i/o (input/output).null-modem a standard serial cable with two sets of control lines crossed:transmit data (ID) to receive data (RD)request-to-send (RTS) to clear-to-send (CTS)xviireal-time refers to the processing of data as it is acquired, as opposed to storing thedata for processing at a later time.RS-232 (serial port interface). The standard asynchronous communications adapterfound on most personal computers. In this work, the RS-232 was used inconjunction with a null-modem cable to allow two computers to exchangeinformation.signal has many meanings, depending on context. In this work, “acoustic emissionsignal” often refers to a collection of 1024 digitized points sampled from acontinuous data stream.Chapter 1 Introduction -1-Chapter One - Introduction“What do mean, you’ve got a little job for me?”- HerculesThis work was started with two goals. The first was to use chemical acousticemission to investigate the kinetics of a complex chemical reaction. The secondwas to develop the technology to a level where it could reliably deliver useful chemicalinformation in systems which are perhaps not readily studiable by more conventionalanalytical methods. These goals are not strictly separate; they are intimately intertwined.The complex chemical reaction studied is the Briggs-Rauscher oscillating iodine clock.1.1 Acoustic EmissionAcoustic emission is not a new phenomenon. It has been used in material sciencesince the early 193 Os’. The American Society for Testing of Materials adopted acousticemission as a recognized technique in 19722. The range of applications continues toincrease with acoustic emission being used to stress-test metals3, composite materials4and polymers5.Chapter 1 Introduction -2-An acoustic emission wave is a transient, elastic wave generated by a rapid massmotion of a collection of atoms. Unlike ultrasonics, where external sound waves areapplied, acoustic emission waves are generated from within the material under study.Physical processes which are known to emit acoustic emission waves include.6 7 •. 8bubble formation , crystal fracture , and phase transitions1.2 Chemical Acoustic EmissionUnlike photoacoustic spectroscopy, in which regular pressure fluctuations aremeasured in response to a chopped beam of radiation, chemical acoustic emission is apassive technique. No stimulation of the chemical reaction is required, reducing thepossibility of perturbation.The first published report of chemical acoustic emission was from van Ooijen etal. in 1978 who noticed audible cracking sounds accompanying the synthesis ofdichloro(pyrazine)zinc(II) A limited study of the phenomenon discovered that theintensity of the sound was found to be proportional to the concentrations of pyrazine andzinc chloride. The most intense emissions were found to be at 100 kllz, and the origin ofthe sound was thought to arise from either the rapid polymerization of short chainsegments or an abrupt change in metal coordination (further work by of thisresearch group, later proved van Ooijen’s hypothesis to be incorrect - acoustic emissionfrom the synthesis of dichloro(pyrazine)zinc(II) was found to be from crystal fracture).Since van Ooij en’s report, a number of researchers have investigated the potentialof chemical acoustic emission as a viable analytical tool. In 1981, Betteridge et al.conducted a survey of different chemical reactions and mixing processes to determine theextent of acoustic activity”. The research demonstrated that acoustic emission wasChapter 1 Introduction -3- Igenerated from chemical systems as diverse as acid-base reactions, ion-exchangereactions and gel formation. The Briggs-Rauscher iodine clock was also found to behighly acoustically emissive which formed the basis for this work.Sawada et al. took up acoustic emission research in the early 1 980s, with a reportof reproducible acoustic emission patterns from the gelation of sodium carbonate withcalcium chloride to form calcium carbonate’2.Plots of acoustic emission intensity as afunction of time were successfully used to distinguish three different reactionmechanisms: phase separation, precipitation, and gel formation. Encouraged by theseresults, Sawada investigated the phase transitions of sodium thiosulfate, p-cresol, water,and select liquid crystalline materials’3,to discover that acoustic emission accompaniesthe direction of the transition which results in volume contraction.In 1989, Lubetkin measured bubble nucleation rates by an acoustic method’4 andfound that for supersaturated solutions of CO2 the rate of bubble nucleation was closelycorrelated with the rate of bubble bursting at the free liquid surface.1.3 Acoustic Emission Detection and AnalysisAs noted above, acoustic emission as applied to chemical reactions is relativelynew and still in exploratory stages. This work is no exception. In the chapters that follow,the reader will witness the evolution of experimental design that led to the developmentof chemical acoustic emission as a useful analytical tool for solution phase reactions.Each successive experimental layout may differ in the instruments employed, yet there isa common thread. The purpose of this present section is to introduce the generalprinciples behind chemical acoustic emission detection and analysis.Chapter 1 Introduction -4-1.3.1 CAE InstrumentationThe simplest acoustic emission instrumentation consists of a sensor, amplifier,and an analog recorder. A system of this nature can be used for simple event counting oranalog recording of the continuous signal. It demonstrates the technique’s basicadvantages: i.e. it is non-invasive, non-destructive, and inexpensive. More elaboratesystems add a spectrum analyzer for real-time frequency analysis or a digitizer forcapturing a selected segment of the signal stream.Figure 1-1. Block diagram of basic experimental hardware.The system employed in this study is illustrated as a block diagram in Figure 1-1.A detailed description of the instruments involved is presented in the following chapters.Briefly, the sensor, a piezoelectric crystal, is placed in contact with the reaction vesselthrough a thin layer of acoustic couplant grease. The crystal transduces the acoustic waveinto an a.c. signal which is fed directly into an amplifier. The amplifier optionally appliesd.c.a.c.Chapter 1 Introduction -5-a bandpass filter to the amplified signal before passing it onto the digitizer. A d.c.representation of the signal (a peak-detect system with 200 ms time constant) is sent tothe analog recorder. The controller, typically a personal computer, supervises the digitizerand provides long-term storage and processing of the data. Shielded cable is used for allconnections.1.3.2 CAE AnalysisIn order to realize the potential of acoustic emission for chemical analysis,information about the characteristics of the emissions must be extracted. This informationis especially valuable when different mechanisms occur simultaneously’5.Theoretically, there is a large amount of information present within individualsignals. However, a signal is a complex transformation of the original emission due tomany transmission factors. Despite these distortions, emissions from the same source willoften show similar features’6.The variability in the emissions results in the characteristics(or features) being described by a population distribution of values rather than a singleunique value. Since hundreds or even thousands of emissions may be collected during asingle experiment, powerful data reduction and analysis tools are required in order touncover meaningful information. Unfortunately, commercially available tools of thisnature were not suitable for acoustic emission research. Specifically, most dataacquisition software was unable to acquire and store emission signals at an adequate rate,and data analysis software lacked the combined spectral analysis and pattern recognitiontools required for analyzing chemical acoustic emission data. To overcome thisdeficiency, a data analysis protocol was established in the laboratory and software writtenin-house to exploit it.Chapter 1 Introduction -6-1.3.2.1 Data Analysis ProtocolThe data analysis protocol, or road map, is illustrated in Figure 1-2. Signalacquisition is the obvious first stage, and programs have been written to acquire data froma variety of digitizers. See Appendix B for a complete list of the software written in thislaboratory. The acquisition cycle consists of arm/trigger/re-arm sequence in which thedigitizer captures a CAE signal when a preset trigger level is breached. This cycle and itslimitations are discussed in detail in the experimental sections of the following chapters.The second stage is signal validation. The purpose of this stage is to ensure thatonly valid CAE signals are captured and stored. Invalid signals include those which areover-ranged or under-triggered or those due to electrical spikes. An over-ranged signalarises when any part of the acoustic waveform exceeds the dynamic range of the digitizer.An under-triggered signal occurs when the digitizer captures a signal whose trigger levelis actually below the requested level. This may be due to trigger level instability orinstrument drift. Signals due to electrical spikes are detected through the use of the rootmean square (RMS) voltage of the signal. The RMS of each signal is compared to theRMS of known background signals and the signal is flagged for deletion if its RMS valueis comparable. This method suffers for experiments which have low signal to noise ratios.In this situation, signal validation based on noise modeling is a better alternative’6.Chapter 1 Introduction -7- ICAE Data Analysis Roadmap IOnce signals have been validated, information can be extracted from them usingtwo methodologies. In the first, the raw signal is analyzed directly either by hand or in anautomated fashion. Different physicochemical processes commonly produce emissionswhich have different characteristics in the frequency domain6. Though the powerspectrum for individual signals from a single experiment can vary appreciably, asdiscussed earlier, the averaged power spectrum is often highly reproducible6.Occasionally more than one acoustically active process will occur during the course of achemical reaction7.In this case, a time-resolved power spectrum may be of greater utility.Figure 1-2. CAE Data Analysis Road map.Chapter 1 Introduction -8-In order to produce a time-frequency surface of this nature, some prior knowledge of thetemporal behaviour of the chemical system is required.The acoustic emission of a chemical reaction is monitored as a function of time.Such a series of data measured in time is called a time series. An autocorrelogram is anexcellent means of determining the presence of periodicity in a time series data set. Tocreate an autocorrelogram, a new independent variable is created (t) which represents atime constant. The correlation coefficient, r(t), can be computed by17:[(x(t)— ) (x(t + t) -Eqn. 1-1. r(’t) =2(n — 1 — t) swhere x(t) is the value at time t, n is the number of data points, and SX is the standarddeviation of x(t).The second methodology is based on an indirect analysis of the raw signal.Numerical parameters called descriptors are computed for each signal. A complete list ofthe 26 descriptors used in this work is presented in Table 1-1. Detailed descriptions ofeach and their suitability for chemical acoustic emission are published elsewhere’8.Eachsignal is then an object which can be represented in a multi-dimensional space consistingof its descriptors as dimensions. Such a data space cannot be visualized and for more thenthree or four dimensions is inherently difficult to interpret. Fortunately there are a numberof chemometric tools which can be employed to assist in the analysis of such highlymultivariate data.Chapter 1 Introduction -- I1 .3.2.2 Mu Itivariate StatisticsMultivariate statistical techniques are necessary for analyzing complicated datasuch as obtained here. They provide a system for analysis under conditions in which theremay be several independent variables and many dependent variables all correlated tovarying degrees. Multivariate statistics represent a direct expansion of the more familiarunivariate statistics; i.e. multivariate statistics can be seen as the general case, withunivariate statistics a simplification of the more general multivariate model. In CAEanalysis, descriptive statistics are used to provide a best-guess description of thedescriptor populations. When descriptive statistics are used in this way, they are typicallycalled parameter estimates rather than sample statistics.17Two multivariate statistical tools are used in CAE analysis: rotated principalcomponents analysis (PCA) and non-hierarchical clustering. The most important use ofPCA is to represent n-dimensional data in a smaller number of dimensions, usually two orthree. This is accomplished by projecting a line through the cloud of points in multidimensional space such that there is maximum variation along the line and minimumvariation around it. Hence the line maintains maximum variation among the data. This isthe first principal component. The second component is a line projected orthogonal to thefirst component, positioned to maximize the variance along its axis. The third componentis orthogonal to the first and second, and so on. Each principal component is a newvariable made up of a linear combination of the original variables. The coefficients ofthese variables are the eigenvectors and the variance along the principal componentprojection is the eigenvalue.PCA results are likely to be uninterpretable unless rotated. Rotation is only usedto improve the interpretability of the solution, it does not change the mathematical fit.The factors may be rotated orthogonally or obliquely. Orthogonal rotation assumes thatChapter 1 Introduction -10-the underlying structures producing the results are uncorrelated. Oblique rotation does notmake that assumption but the results from an oblique rotation are typically more difficultto interpret.’7In non-hierarchical clustering, one wishes to classify the n-dimensional descriptordata into K clusters. The user selects the initial number of clusters, K, and the location ofthe cluster centers in the n-dimensional space (or the initial centers are chosen randomly).The algorithm assigns each object (in this case the object is a CAE signal) to one cluster(normally) based on the Euclidean distance between the object and the cluster’s center-of-mass. Once all objects have been assigned, the cluster’s center-of-mass is re-computedand each object is again assigned to the nearest cluster. The cycle is repeated until thesame clustering is found in two successive assigmnent steps.There are inherent limitations to clustering. The importance of choosing initialcluster centers is magnified if the data contain outliers, thus skewing a cluster’s center-ofmass computation. Then there is the problem of choosing the correct number of clusterswith limited or no a priori information about the system. Choosing too few clusters willnot reveal the structure of the data set and too many clusters may result in a large clusterbeing subdivided into two smaller ones. Fortunately, there are ways of detecting suchproblems and compensating for their effects which are discussed more fully in latersections.Chapter 1 Introduction -11-Table 1-1 - Descriptors used to characterize individual CAE signals.Descriptor Domain Description(abbreviation)CREST Time ratio of peak voltage to RMS voltageKURTOSIS Time measure of the deviation from a Gaussian distributionT@AREA!2 Time time to half-area (signal decay measurement)0-CROSS Time number of times signal crosses zero volts25-CROSS Time number of times signal crosses ±25% maximum voltage1/8T, 2/8T, etc. Time normalized time octiles of RMS voltageFRQMEAN Freq. frequency equal to the mean of the freq. intensitiesFRQMED Freq. frequency at mid-area of the integrated freq. intensitiesFRQCREST Freq. ratio ofmaximum power to RMS power• nd rd. .FQRTLBW Freq. bandwidth between the 2 and 3 intensity quartilesFBW>15% Freq. bandwidth of freq. whose intensities> 15% ofmax.DFB 1, DFB2, etc. Freq. eight user-defined frequency bandsChapter 1 Introduction -12-1.4 Oscillating Chemical Reactions1.4.1 HistoryThe history of the discovery of oscillating chemical reactions can be reliablytraced to the early 1 900s when A.J. Lotka published a series of papers demonstrating thetheory and practice of periodic reactions in a gas-phase system’9’201.Soon after, Braydiscovered the first liquid-phase example; the iodate-ion catalyzed decomposition ofhydrogen peroxide22. Liebhafsky joined Bray in an effort to map out the regions ofacidity where various reactions predominate, and then to isolate those reactions in orderto determine their rate constants23.The reaction became known as the Bray-Liebhafsky(BL) oscillator. Due to the instrumental limitations of the period, conclusive data couldnot be gleaned from this complex system, and it was decades later before interest inoscillating chemical systems was rekindled.It was in Russia, in 1951, that Belousov uncovered temporal oscillations in thecerium-ion-catalyzed oxidation of citric acid by bromate ion. This work was published in1958 in the proceedings of a minor radiation medicine conference24. Zhabotinskiicontinued Belousov’s work25 and the reaction became known as the BelousovZhabotinskii, or BZ reaction. Unlike previous oscillators, the BZ reaction demonstratedboth spatial and temporal periodicity. Supported by recent advances in instrumentationand mathematics26,many researchers began studying oscillating reactions.Chapter 1 Introduction -13-Oscillating Chemical ReactionsAnnual Number of Publications120 -.100 -°- 80-U,C04-’0-Q04-0ci) -0E •0- . — —1900 1910 1920 1930 1940 1950 1960 1970 1980 1990Year of PublicationOscillating Reaction Subtopics120 -1100-0 N- I Theory I\\N BL osc. BZ Osc. Other-20- I I I I1900 1910 1920 1930 1940 1950 1960 1970 1980 1990Year of PublicationFigure 1-3. Oscillating reaction publications.Chapter 1 Introduction -14-Burger researched the impact of homogeneous oscillators on the chemicalliterature27 and stated “A scientific revolution took place with the discovery ofperiodiccolour changes in the aqueous mixture of citric acid-bromate ion-cerium ion.” Burger’sraw data is expressed in graphical form in Figure 1-3.In a one-page article in the Journal of Chemical Education (1973) two high schoolteachers, Thomas Briggs and Warren Rauscher, reported an oscillating reaction which,, 28gives striking cyclic changes from colourless to gold to blue using simple reagents.The reacting solution incorporates the reagents from the Bray-Liebhafsky (BL) oscillatorwith the addition of an organic substrate (typically malonic acid) and a metal-ion catalyst(usually manganese). Unlike the BL oscillator which requires high temperatures (> 50°C)to oscillate appreciably, the Briggs-Raucher (BR) oscillator runs at a rate of severalcycles per minute at room temperature after a short induction period.In 1975, Cooke began the first systematic investigations of the BR oscillator inwhich each reagent was separately varied and the effect on the oscillation frequencynoted29. In 1979, the work was extended to cover the effects of the reagents on all theoscillator’s reaction phases: the induction period, waveshape, time periods, and the modeof conclusion30.Few details were provided concerning the mechanism, however, and itwas not until 1982 that a plausible skeleton mechanism was published simultaneously by31 32two research groups1.4.2 General TheoryContrary to conventional wisdom, oscillating reactions are quite common,especially in biological systems. In contrast to ‘normal’ or ‘ordinary’ chemical reactions,the concentrations of products in an oscillating reaction do not rise smoothly to amaximum. Instead, they exhibit a tendency to rise and fall cyclically. There is someChapter 1 Introduction -15-dispute in the literature as to whether the systems are exhibiting periodic or chaoticbehaviour33 . However, the kinetic and thermodynamic criteria that a system must meetfor oscillations to take place have been defined quite precisely:• oscillations only occur in systems far from equilibrium• the variables that determine the reaction rates are coupled kinetically• oscillating reaction mechanisms contain autocatalytic processesTo study a system far from equilibrium usually requires that the system be openso that energy and/or mass can be exchanged with the outside. Note that this is acharacteristic of a biological system, which constantly renews itself by interaction withthe outside world. Researchers have duplicated open systems through the use ofcontinuous-flow, stirred-tank reactors (CSTR)34.In a CSTR, reactants are pumped into arapidly stirred, fixed volume reactor. The partly reacted mixture is drained from the tankat the same rate as fresh reactants are pumped in, hence maintaining a constant volume.The second condition is the assertion that the concentration of one component willeffect the rate of change of concentration of a second component. These two componentsare then referred to as coupled species.The third condition is an empirical observation from studying oscillatingreactions. It has been found that an oscillating system requires some sort of feedbackmechanism in order to function. For example, consider a system in which an enzyme, EH,can undergo either reversible deprotonation:Eqn. 1-2 EH (active) E- (inactive) + H rate Kaor may catalytically deprotonate a substrate (SH):Eqn. 1-3 EH+SH—*EH+S+H ratek2Chapter 1 Introduction -16-If Ka is large, then Eqn. 1-2 initially predominates, producing large amounts ofinactive enzyme, leaving little active enzyme for Eqn. 1-3. However, as Eqn. 1-3proceeds, H is generated which drives Eqn. 1-2 to the left, raising the concentration ofactive enzyme. This is turn increases k2, producing even more H, and driving the firstreaction even further left.In a closed system, this sequence of steps would continue until the concentrationof SH is depleted, or there is sufficient acid to inhibit both processes. In an open system,however, H would begin to diffuse out of the reaction volume into regions of lower acidconcentration. This will have the effect of Eqn. 1-2’s equilibrium shifting to the right,once more inhibiting the second reaction. Over time, diffusion would replenish SHsufficiently to begin the autocatalytic cycle once more. The system will oscillate betweenperiods of rapid reaction when [Hj, [SH], and [EHJ are large, and periods of slowreaction when {EH] and [SH] are low.In practice, a system of this nature often has two relatively stable states. Anoscillating chemical reaction may be thought of as a system which fluctuates betweenthese two states. The feedback in the reaction mechanism is what provides the impetus tokeep moving between states rather than settle permanently in a single state.Chapter 1 Introduction -17-1.4.3 Briggs-Rauscher Oscillator1.4.3.1 Reaction MechanismFrom the historical perspective presented earlier, it is apparent that the study ofoscillating chemical reactions has enjoyed a rich and lengthy history. From the manypublications written pertaining to chemical oscillators based on iodate ion and hydrogenperoxide, there has arisen a pseudo-standard for the labeling of mechanistic steps. Eachstep is given one of the following labels:Iodine reactions where there is no net change in oxidation number of iodineUp reactions where there is a net oxidation of iodineDown reactions where there is a net reduction of iodineOxygen reactions involving oxygen species, with no net change in oxygenoxidation numberManganese reactions involving the Mn(II) catalystCarbon reactions involving the organic substrateFor the purposes of this work, previously published mechanistic steps will use thepreceding numbering system. All other mathematical and chemical equations will benumbered sequentially as illustrated by Eqn. 1-2 on page 15.In describing the BR oscillator, it is helpful to first characterize its predecessor,the BL oscillator. This reaction begins with three deceptively simple reagents: hydrogenperoxide, iodate, and water in mineral acid. Bray was studying this system22 when hehappened upon a narrow concentration region where the iodine concentration oscillatedslowly with a period of several days. Despite the limited equipment of the period, thebase processes were elucidated:Chapter 1 Introduction -18-iodine production:Eqn. 1-4 5H20 + 2IO + 2H 12 + °2 + 6H20iodine consumption:Eqn. 1-5 5H20 + 12 —> 2IO + 2H + 4H20The net result of Eqn. 1-4 and Eqn. 1-5 is the catalytic decomposition of hydrogenperoxide:Eqn. 1-6 10H2 —> 10H2 +These base processes were a starting point for those researchers investigating theBR oscillator. Eqn. 1-4 is the limiting step whose rate is greatly increased by the additionof Mn(II). Furrow and Noyes postulated a free-radical mechanism32 to explain the rateincrease:free radical initiation:03. H + I0 + H20 H0I0 + H20 +15. H + IO + H010 —> 210; + H20chain reaction.Ml. 10 + H20 + Mn2 H0I0 + Mn(QH)2M2. Mn (OH)2 + H20 —* H20 + Mn2 + H00D4. H + IO + H00 —> io; + H20 +chain termination:14. 2H010 —> I0 + HOl +Dl. HOl + H20 —> + °2 + H + H20Ii. HOl + I + H 12 + H20Chapter 1 Introduction -19-The stoichiometry of Eqn. 1-4 can be obtained with 4(M1) + 4(M2) + 4(D4) +2(14) + (Dl) + (Ii). The mechanism for iodine consumption may be explained withoutresorting to free radicals; unlike the BL system in which ‘2 consumption is viareoxidation to iodate, the BR system uses the organic substrate as a convenient halogensink. Again from Furrow and Noyes:enolization and halogen sinking.C3. RI-I <‘> enolC4. enol + 12 — RI + f + Hproduction ofHOl:13. f + iO + 2H HOIO + HOl12. HOlO + I + H 2H01The presented model is only a skeleton. Experimental rate constants are availablefor only seven steps, the remaining are estimated or adjusted so as to produce oscillationsin the computer models. There is no quantitative agreement between experiment andmodel. The magnitude of oxygen evolution is unknown. The presence of CO2 isdebatable.1.4.3.2 Heat ProductionCooke’s studies in 1979 included limited heat production data obtained using athermistor-Wheatstone bridge circuit30.He studied the BR reaction using highH20 andlow malonic acid concentrations in a 50 mL reaction volume. The resulting temperaturetime curves were low-resolution but definitely showed a three degree rise in temperatureover the nine minute life of the reaction.Chapter 1 Introduction -20-In 1987 Lamprecht and Schaarschmidt conducted microcalorimetricmeasurements on both the BZ and the BR reactions35.For the BR reaction, they reportedsome ten oscillations with a period of 56 s in the heat production. Unfortunately, all theirdata was presented in arbitrary units and phase relationships could not be directlyestablished due to the thermal inertia of the calorimeter. In the end they concluded thatfurther details concerning oscillatory heat production may be drawn after time-consumingdeconvolution of the calorimeter output.1.5 CAE and the Briggs-Raucher Iodine ClockThe pressing question is what physical and/or chemical processes in the iodineclock are giving rise to the acoustic activity reported by Betteridge. Since the iodine clockevolves copious amounts of 02 and bubble evolution is a known source of acousticemission6’14, it is a reasonable hypothesis that bubble evolution is the primary source ofacoustic emission in the iodine clock.1.5.1 Bubble Nucleation TheoryFor the purposes of this work, a bubble may be defined as gas surrounded entirelyby a liquid with a working upper diameter of about one millimeter36.A bubble may beformed in the bulk of the solution (homogeneous nucleation) and/or at a surface(heterogeneous nucleation). Studies of supersaturated gas/water systems foundheterogeneous nucleation to be the dominant mechanism of bubble formation’4’37.Thedegree of dominance depends mainly upon two parameters: contact angle and nucleationsite geometry.Chapter 1 Introduction -21-A bubble forms at a nucleation site at a given contact angle , characteristic of thegas/liquid system under study, of between 0 and 1200. A contact angle of 00 representsthe homogeneous nucleation case where bubble detachment is effectively instantaneousand the overall rate is determined solely by the nucleation kinetics. At contact anglesgreater than 00 the nucleation kinetics may be dominated by either the nucleation rate orthe detachment rate. The literature frequently reports the dimensionless factor L where Lis the ratio of the time required for detachment to the time required for nucleation. Hence,when L < 1, the nucleation rate dominates and when L> 1, the detachment ratedominates. Lubetkin’4reported that for an acoustic method to be useful L must be < 1,implying that the contact angle must be close to zero. Lubetkin’ s study of theC02/H0system on glass surfaces found contact angles of less than 5°.The effect of the nucleation site geometry has not been investigated for the case ofsupersaturated solutions. For details on the effects in the case of boiling refer to thereview ofCole38.Basic bubble nucleation mathematics theory remains much the same as when itwas first proposed in 1939 by Volmer39.The first equation relates the bubble pressure (Pr)to the external pressure (P) , the surface tension (a) and the bubble radius (r).2aEqn.1-7 = P +rEquation 1.7 expresses the concentration of the gas in solution as a product of Henry’sLaw constant (i) and the bubble pressure:Eqn. 1-8 C01 = KPrIf the bubble is not at equilibrium then the bubble will grow by diffusion of gas into it if1r < (C01/K).Chapter 1 Introduction -22-According to Volmer, density fluctuations in a solution create microscopic, short-lived cavities called embryos. An embryo must reach a critical radius before bubblenucleation can proceed:2aEqn. 1.8 ‘crit =CSS/1CPcDwith representing the supersaturated gas concentration. The rate of nucleation J inVolmer’s theory is proportional to exp(-W/kT) where W is the work required to expandthe embryo into a spherical cavity:Eqn.1.9 W=aA—(P—P)vwhere A is the area and v the volume. Equations 1.6 through 1.9 may now be combinedto form:16itc3Eqn. 1.10 W=3(C01 /K—P)where it is approximately true that,Eqn. 1.11 J=exp( 2)where a, 13 are constants. It can be seen that a large increase in J will occur when C01reaches a critical value.Chapter 2 Preliminaiy Studies Using Batch Methods -23-Chapter Two - Preliminary Studies Using Batch Methodswas it unlock the safe and then swim to the surface?”- Houdini2.1 IntroductionThis work represents research in an area of analytical chemistry which is still largelyunexplored. When the experiments were being planned, the challenge was to set upapparatus which would yield useful chemical information without requiring extensivetime or resources to assemble. This was accomplished by putting together instrumentsfrom a variety of disciplines and tying them together with software written in-house. Thisis described in more detail below.Chapter 2 Prelimina,y Studies Using Batch Methods -24-2.2 Experimental2.2.1 Briggs-Rauscher Chemical SystemTwo sets of reagents were made up from commercially available chemicals andwere used without further purification. Reagent Set A consisted of reagents which weremade up to the concentrations recommended by Briggs and Rauscher. Reagent Set B werethe same reagents, but the concentrations were reduced by two-thirds. The summary ofreagent concentrations can be found in Table 2-1.The hydrogen peroxide concentration was determined twice daily usingtitanium(IV) oxalate as described by Sellers40. The starch solution was prepared fresheach day by vigorously boiling the requisite amount of starch in distilled water until theoriginally clear solution turned translucent, as recommended by Vogel.4’Table 2-1. Reagents used for batch studies.Reagent Source and Grade Reagent Set A Reagent Set BConcentration ConcentrationH20 BDH, ACS Analytical 1.200 M 0.4000 MKb3 BDH, ACS Analytical 0.0529 M 0.0 177 MH2S04 BDH, ACS 0.0500 M 0.0 167 MAnalyticalCH2(COOH) Fisher, ACS 0.0670 M 0.0223 MMnSO4 MacArthur, ACS 0.0067 M 0.0022 Mstarch BDH, ACS Analytical 0.01% v/v 0.01% v/vChapter 2 Prelimina,y Studies Using Batch Methods -25-I2.2.2 Instrumental SystemThe instrumental apparatus is illustrated in Figure 2-1. A brief summary of eachcomponent’s make, model and settings is given in Table 2-2. Information about themanufacturers can be found in Appendix B. The chemical reaction took place in a large,shallow, flat-bottomed Pyrex® dish approximately 15 cm in diameter and 5 cm high. Thedish rests upon a large sheet of Styrofoam (not shown for clarity). The center of theStyrofoam has been cut out to accommodate a piezoelectric transducer which is in contactwith the dish through a thin layer of acoustic coupling grease. The piezoelectric crystaltransduces the chemical acoustic wave into an alternating current (a.c.) electrical signaland amplifies it with an internal 40 dB amplifier. This amplified a.c. signal is then fedinto a dual-stage conditioning amplifier. The first stage further amplifies the a.c. signalaccording to the amount specified by panel-mounted buttons, while the second stageconsists of a bandpass filter that rejects those frequencies which fall outside a user-defined range. In this experiment frequencies below 100 kHz and above 2 MHz wereexcluded. The lower limit was chosen to exclude any possible high-frequency noise (i.e.,audible frequencies), while the upper limit was chosen to exclude the range above thetransducer’s response limit. It should be noted that extremely intense signals (>10 volts)overload the amplifier and allow all frequencies to pass. Thus, the amplification settingwas carefully chosen to obtain the best response without the chance of signal overload.The output stage of the amplifier divided the signal into a conditioned a.c signal and ad.c. peak-detect signal with 200 ms time constant. The conditioned a.c. signal wassampled by a digital storage oscilloscope while the d.c. signal was fed directly into ananalog chart recorder.Chapter 2 Preliminaiy Studies Using Batch Methods -26-)fldltloflilamplifierFigure 2-1. Apparatus for preliminary batch studiesThe digital storage oscilloscope was operated in a triggered data acquisitionmode. Using this method, the oscilloscope continuously samples the analog signal fromthe conditioning amplifier, maintaining a small sampling buffer of 32 data points. Whenthe signal breaches a user-defined trigger voltage level, the oscilloscope is said to havebeen triggered, and it stores the buffered data points (pre-trigger points) as well as thenext 992 points into its memory and informs the control computer that a signal has beenTable 2-2. Apparatus list for batch studies.Instrument Manufacturer Model Settings/Otherpaddle stirrer Fisher 92a stainless steelpiezoelectric transducer Brüel & Kjer 8312 40 dB gainconditioning amplifier Brüel & Kjr 2638 40 dB gain100 - 2000 kHz filterdigital storage Tektronix T2430a 8 mV resolutionoscilloscope 2.50 MHz samp.freq.chart recorder BBC SE12O 6 cm/mm., 500 mV full scalepersonal computer Campus 3 86/20 GPIB interface used forComputers communication with scopepaddle stirrerdigital storageoscilloscopeVrMLIdigitizedac. signalconditioneac. signalanalogchart recorderconditionedignaltransducerChapter 2 Preliminaiy Studies Using Batch Methods -27-acquired and is awaiting transfer. It is now the responsibility of the control computer toinitiate signal transfer, store the signal (1024 points total) onto disk for post-processing,and re-arm the oscilloscope.The control computer was an IBM PC-AT compatible personal computer with anIntel 80386 processor. It communicated with the oscilloscope via the industry-standardGPIB interface. Software written in-house by this author was used to control theoscilloscope’s initial operating parameters. While the experiment was underway, thesoftware was responsible for arming the oscilloscope and then retrieving the signal itacquired once triggered. This arm-trigger-transfer sequence was repeated until theoperator deemed the experiment should be concluded. For more automated experiments,the software could be instructed to terminate signal acquisition after a user-defined periodof time had elapsed or a particular number of signals acquired.2.2.3 Timing and MethodsStrict timing control is necessary in order to reproducibly capture the oscillatingiodine clock’s unique temporal behaviour. A method was needed to deliver accurateamounts of reagent to the reaction vessel in a timely reproducible manner whilesimultaneously starting the recording equipment. To this end the reagents were premixedas three separate solutions. Solution #1 contained peroxide, solution #2 comprised theiodate and sulfuric acid, and solution #3 the malonic acid, starch and manganese. Withoutthe catalyst, the reaction proceeds slowly at room temperature and so 25 mL of eachsolution was pipetted into the reaction vessel in its numerical order.Once all three solutions had been delivered, the paddle stirrer was started alongwith the oscilloscope and chart recorder. Several trial runs were conducted for timingpurposes and it was noted that oscillations ceased after approximately 150 seconds. TheChapter 2 Preliminaiy Studies Using Batch Methods -28-progress of experiments was monitored and data collection halted after 300 seconds.Colour changes in the solution were observed visually and recorded manually. Thisexperimental sequence was repeated five times for each reagent set.2.2.3.1 Stirring EffectsBriggs and Rauscher reported in their 1973 paper that “Constant stirring improvesthe cycling, but is not essential.” While no detailed study on the effects of stirring in theBR oscillator has been published, Dutt and Menzinger42did report on the effect of stirringand temperature on a Belousov-Zhabotinskii-like reaction. They concluded that increasedstirring increased both the induction and oscillation periods and that the greatest effectswere observed at low temperatures of ca. 15°C.As mentioned in the previous section, a paddle stirrer was used in theseexperiments to provide mixing. Since the paddle stirrer operated at a fixed speed, and theexperiments were conducted at room temperature, we make the assumption that stirringeffects are constant between replicate runs.2.2.3.2 Sampling FreguencyIn this experiment there are two sampling frequencies to be considered. First,there is the rate at which the oscilloscope samples the analog a.c. signal which can be seteither manually or through computer control. The piezoelectric transducer has an upperlimit frequency response of 1 MHz, so a sampling frequency of 2.5 MHz was selectedto avoid aliasing. The second sampling rate requires greater explanation.In a triggered data acquisition mode, the oscilloscope, once triggered, is idle whileit waits for a re-arm command. The control computer does not issue the re-arminstruction until it has received and stored the digitized signal from the oscilloscope.Chapter 2 Preliminary Studies Using Batch Methods -29-Thus, the signal transfer process is the rate-limiting step in the data acquisition cycle. Theperiod of the limiting step is the longest of one of these three events: z) the oscilloscope’sdata transfer to the GPIB interface, ii) the GPIB ‘ s data transfer to the control computer oriii) the control computer’s data transfer to its internal hard disk. Measurements havedetermined the first event, data transfer to GPIB interface, is the limiting step, requiringapproximately 0.8 seconds.2.2.4 Sensor CalibrationAs mentioned in Chapter One, the chemical acoustic wave may undergo severaltransformations and attenuations before arriving at the piezoelectric sensor. Thus, thesensor must be secured to the reaction vessel in a manner both reproducible and allowingtransfer of maximum signal amplitude. To this end, a computer-controlled signalgenerator (model PCIP-SST, Metrabyte Corp.), with built-in calibration, was purchased.The device, a “signal generator on an adapter card”, was placed in one of the expansionslots of an IBM PC-AT compatible computer.Figure 2-2 illustrates the calibration procedure. The output of the signal generatorwas connected to a flat-response transducer (model FAC500, AET Corp.) to be used as asound source. The transducer was securely contacted to the reaction vessel and coupledthrough a thin layer of acoustic grease. Instead of the reaction mixture in the vessel, anequivalent volume of distilled water was used to ensure a nearly identical mass loadingon the transducer. This is possible since the sulfuric acid in the reaction mixture is presentat less than 1% by weight, having a density of 1.0051 g/mL43, which is comparable todistilled water. Calibrations were performed with and without the paddle stirrer in placeto determine if the mixing gave rise to an acoustic signature.Chapter 2 Preliminaiy Studies Using Batch Methods -30-)The signal generator was programmed to step through a range of frequencies from100 kHz to 2 MHz at a known output voltage. The test signals were captured andprocessed as if during an actual experiment. The resulting power spectra were comparedto those captured during earlier runs to see if the transducer response had changed.Stirring effects were not observed in the averaged power spectra, and thetransducer reponse remained unchanged over the time period the experiments wereconducted, in agreement with the work ofWentzell44.OD—signal generatorconditionedac. signalconditionedd.c. rms signalanaloga.c. signalpiezoelectrictransducerFigure 2-2.An experiment of this nature yields a large volume of data. First there are the rawa.c. signals which have been pre-processed by the conditioning amplifier beforedigitization. Consider that one digitized signal is 1024 data points of information, eachdigitizedac. signalHconditioningamplifierApparatus for calibration procedure developed for batch studies.2.3 DiscussionChapter 2 Preliminary Studies Using Batch Methods -31-one byte in length, and that more than 300 signals are captured during the course of a fiveminute experiment. An average run thus produces more than 307,000 bytes of raw data; aconsiderable amount of information to be processed and analyzed. There is also the d.c.signal output from the conditioning amplifier and recorded by the chart recorder. At achart rate of 6 cm/mm., 30 cm of chart output is generated for every experiment which isa total of 300 cm for these batch studies.An adequate data reduction and analysis strategy is required. Over the years thisresearch has been conducted, we have attempted to devise efficient and robust dataanalysis protocols which yield useful chemical information. We begin by considering thed.c. signal data separately from the a.c. signal data.2.3.1 Analysis of D.C. SignalsFor the purposes of this discussion we randomly pick two of the five replicates forReagent Set A and one of the five replicates for Reagent Set B. Oscillations as evidencedby colour changes were observed for the former reagent set and recorded manually. Nooscillations were observed for the latter reagent set. Gas evolution was observed in allexperiments, albeit more vigorous during the Reagent Set A experiments.The relationship between the colour changes and the acoustic emissions isillustrated in Figure 2-3. When the CAE is at a minimum, the solution is colourless. Asthe solution slowly changes from colourless to gold, there is a corresponding rise in theCAE until a sharp transition to blue is observed at which time the CAE is at a maximum.The solution remains blue until the CAE has once again reached a minimum. The figureis drawn to emphasize all the features. From the experimental data, it can be seen that thepeaks are skewed to the right. This is discussed in greater detail in Chapter Three.Chapter 2 Preliminaiy Studies Using Batch Methods -32-q-ec-)C,)cusharp transitionto blueincreasing timeFigure 2-3. Colour changes in relation to CAE d.c.The figures C-i through C-3 are the scanned images of the original chartrecordings from the selected experiments. They show acoustic emission intensity as afunction of time. The first two figures (Figures C-i and C-2) have similar features, andwe will discuss these at length in the following paragraphs. Figure C-3 is nearly flatindicating that if the oscillating reaction did take place it was not detectable from the d.c.signal at the chosen amplification level.Figure C-i is the chart recording of the first run of Reagent Set A. Specialattention should be directed to the axes of the recording. Due to the nature of theexperimental setup, the horizontal axis is time (increasing left to right), while the verticalaxis is voltage which increases from top to bottom. That is, the 0,0 point is the upper leftcolourless colourlessChapter 2 Preliminaty Studies Using Batch Methods -33-of the image. The recording itself is straightforward. There is an initial induction phasewhich rises quickly to a maximum (ca. 22 seconds) before triggering a series ofoscillations. Each successive oscillation is attenuated and gains more fine structure untiloscillations can no longer be clearly distinguished (ca. 150 seconds). There is also ageneral increase in signal intensity until Ca. 270 seconds, after which a decreasing trendis noticeable.Figure C-2, the chart recording of the second run of Reagent Set A, exhibitssimilar features. A steep induction phase (ca. 22 seconds), increasingly attenuatedoscillations to ca. 140 seconds, and finally a decreasing trend after ca. 180 seconds.The fine structure of the chart recordings makes it difficult to isolate key phases.To improve the image quality, the chart recordings for Reagent Set A were digitized andthen smoothed with a 15% FFT algorithm (see Glossary). The results for the first 150seconds (when oscillations occur) are in Figure C-4. From this figure we will discuss thefollowing phases: induction, positive slope (pos), negative slope (neg), and shoulder.Induction represents the initial phase of the reaction before oscillations commence. A pusphase is when d(response)/dt is positive, neg when d(response)/dt is negative andshoulder represents a sudden decrease in the magnitude of d(response)/dt withoutchanging sign.Table 2-3 lists the time periods for each phase. The data indicate remarkablesynchronicity between the two runs until the third oscillation at which time synchronicityis lost on the neg phase. When the phases are sorted as in Table 2-4, other trends becomeapparent. These trends are discussed in detail below.Chapter 2 Prelimina,y Studies Using Batch Methods -34-Table 2-3. Phases present in Reagent Set A data, sorted by timePhase Run 1 Run 2 NotesStart Stop Dur Start Stop Dur (all times in seconds)Induction 0 22 22 0 22 22Neg 22 30 8 22 30 8Pos 30 35 5 30 35 5 osc.#1Neg 35 48 13 35 48 13Pos 48 54 6 48 54 6 osc. #2Neg 54 61 7 54 61 7Shoulder 61 66 5 61 66 5Pos 66 73 7 66 73 7 osc. #3Neg 73 81 8 73 80 7Shoulder 81 86 5 80 86 6Pos 86 92 6 86 92 6 osc. #4Neg 92 102 10 92 98 6Shoulder 102 108 6 98 106 8Pos 108 116 8 106 116 10 osc.#5Neg 116 125 9 116 125 9Shoulder 125 135 10 125 132 7Chapter 2 Preliminaiy Studies Using Batch Methods -35-Table 2-4. Phases present in Reagent Set A data, sorted by phase.Phase Run 1 Run 2 NotesStart Stop Dur Start Stop Dur (all times in seconds)Neg 22 30 8 22 30 8Neg 35 48 13 35 48 13Neg 54 61 7 54 61 7Neg 73 81 8 73 80 7Neg 92 102 10 92 98 6Neg 116 125 9 116 125 9Induction 0 22 22 0 22 22Shoulder 61 66 5 61 66 5Shoulder 81 86 5 80 86 6Shoulder 102 108 6 98 106 8Shoulder 125 135 10 125 132 7Pos 30 35 5 30 35 5 osc. #1Pos 48 54 6 48 54 6 osc. #2Pos 66 73 7 66 73 7 osc. #3Pos 86 92 6 86 92 6 osc. #4Pos 108 116 8 106 116 10 osc.#52.3.1.1 Induction phaseThe emphasis of previous research has been on the mechanism of the oscillationsor the effect of different reagents or physical factors. In the literature there is littlediscussion of the induction phase. Cooke29 briefly mentioned that “under suitableChapter 2 Preliminary Studies Using Batch Methods -36-conditions the oscillations are preceded by an induction period”. He also systematicallyvaried the concentrations of the various reagents and measured their effect on theinduction phase. He concluded that the rate of ‘2 production is first order in [H20j andindependent of the other species in the following reaction:Eqn2-l 2IO + 5H20 + 2H > + 12 + 6H20This provides a starting point from which we can investigate the kinetics of the inductionphase based on the chemical acoustic emission data. To this end a plot ofLN(Responseadj) versus Time was constructed and can be seen in Figure 2-4 (bottom).The responseadi was computed by subtracting the CAE response (from Figure C-2) fromthe value for inifinite-time CAE. This value was calculated by fitting the data from Figure2-4 (top) to a single exponential function. It is readily apparent that there are twoprocesses taking place in the induction period. There is a rapid, almost exponential risefor the first five seconds, followed by a less rapid nearly linear rise. These processes canbe explained by the following model:x k1 > 02 (soln) k2 > Nucleationwhere X is some process which evolves oxygen into solution at rate k1 (such as Eqn 2-1),followed by bubble nucleation at occurring at rate k2. The rate law would then be:Eqn2-2d[021=k1X—k2[0]dtthe concentration of oxygen in solution would quickly build to a steady state:d[02]Eqn2-3dt=k2[0]—kEqn 2-4d{02]=k2([0,] —[02])Chapter 2 Preliminaiy Studies Using Batch Methods -37-which upon integrationEqn 2-5 [021 = [021SS(l — exp(—k,t))Recall from Chapter One that the nucleation rate J is itself a function of [02]SOlfl.Eqn2-6 J=cexp( 2)[02 IsolnLubetkin pointed out in his work’4 that the rate of nucleation (J) is closelycorrelated with the rate of bubble bursting at the free liquid surface. Both Lubetkin’ sstudy and earlier work by Wade6 indicate that the bursting is the most likely acousticallyemissive process, which ties nucleation to CAE. But how is k2 related to J ? Asdemonstrated in Chapter One, J is a complex function of many variables and whichsuggests that no simple relationship between k2 and J may exist, which implies that thereis no simple relationship between k2 and CAE.Chapter 2 Preliminaiy Studies Using Batch Methods -38-Induction Period160 -_________________________140- 00000120- 000000000:- 100 -o 0>80-060- 040-20-0-I I I I I I I I I I I I I6-5--00C 0o 000O 00G) 000000Z 000002--1 — I I I I I I I I I I I I I I I I I0 5 10 15 20 25Time (s)Figure 2-4. Plots of Response vs Time (top) and LN(Responseadj) vs Time (bottom).Chapter 2 Prelimina,y Studies Using Batch Methods --I2.3.1.2 Negative Slope (neg) PhaseFor the purposes of this discussion, we define a neg phase as one in whichd(response)/dt has a negative slope, indicating the rate of acoustic emission is decreasing.Colour provides some indication of the dominant processes. Blue is observed wheniodine complexes with starch. The iodide ion gives a colourless solution. Roux andVidal52 noted that a gold colour is observed for the tn-iodide ion.During a neg phase, the colour is observed to change from deep blue to colourless,from which we can infer that:z) [‘21 is decreasing to a minimum.ii) [F] rapidly increases to a maximum. (iodide/iodine ratio not favourable to formsignificant amounts of tn-iodide ion).For ease of comparison, the neg phases are grouped together in Table 2-4 on page35. The phase durations are not informative. If the phase immediately following theinduction phase is discounted as a “transition phase”, then there is a general trend towardsdecreasing durations. More useful information can be gleaned from the slopes of the lines(Table 2-5). The rate of each neg phase is constant until the third oscillation implying thatsome unknown event occurred which disrupted the reaction at that point. The rate thendecreases slightly. The reasonable conclusion is that the rate of acoustic emission duringa neg phase is independent of the reagent concentrations.Chapter 2 Preliminaiy Studies Using Batch Methods -40-Table 2-5. Neg phase slopes for Run 1.Phase SlopeNeg (after induction) -10.0Neg(osc#1) -10.0Neg (osc #2) -10.0Neg (osc #3) -9.4Neg (osc #4) -9.4Neg (osc #5) -9.42.3.1.3 Positive Slope (pos) PhaseA pos phase is one in which d(response)/dt has a positive slope, indicating therate of acoustic emission is increasing. Once again, colour provides some indication ofthe dominant processes. During a pos phase, the colour is observed to change fromcolourless to gold to deep blue at the peak, from which we can infer that:i) [‘2] is increasing to a maximumii) [F] is low, then builds rapidlyFor ease of comparison, the pos phases are grouped together in Table 2-4. Theslopes of the lines can be found in Table 2-6. A plot of 1/(Pos rates) vs Event Number isdepicted in Figure 2-5. The resulting correlation coefficient was 0.945. The reaction steprepresenting apos phase might be:Dl. HOl + H20 + + H + H20Second order kinetics suggests that the rate of oxygen evolution is first order in bothhypo-iodous acid and hydrogen peroxide.Chapter 2 Preliminary Studies Using Batch Methods -41-ITable 2-6. Pos phase sopes for Run 1Pos Phase Results0 1 2 3 4Event NumberPhase SlopePos (after induction) 28.0Pos(osc#1) 17.5Pos(osc#2) 15.3Pos (osc #3) 14.0Pos(osc#4) 13.3Pos (osc #5) 1 1.40.09 —0.08 -0.07ci)0.06 -0.05 -0.04 -0.03= 094505 6 7Figure 2-5. Fit of 1/(Pos phase rates) vs. Event NumberChapter 2 Preliminaiy Studies Using Batch Methods -42-)2.3.1.4 Shoulder PhaseThe most interesting feature of the shoulder phase is that it is unreported in theliterature and hence may be unique to chemical acoustic emission. Published figuresshowing oscillations in iodine species generally consist of saw-tooth or sinusoidal tracesdepending upon the mineral and/or malonic acid concentrations29.The oscillations thenterminate abruptly or gradually damp away.The data collected in this work clearly shows formation of a shoulder beginningwith the second oscillation and increasing in size and duration until oscillations are nolonger distinguishable. Does this represent a process which can only be detected viachemical acoustic emission? If so, what is the process?To answer these questions, attention must be drawn to the reactions occurringduring the neg phase. It was suggested above that the following is happening:i) [‘2] is decreasing to a minimumiz) [F] is increasing to a maximumRecall from the introduction in Chapter One that iodine is consumed during halogensinking by the organic substrate:C4.enol +and subsequently produces iodous and hypo-iodous acid:13. P +IO -i-2F{ —* HOIO+HOIThese reactions, as well as further catalyzed production of HOIO, continue untilthe concentrations of HOTO and F reach a critical level. When reached, a “switch isthrown” and the oscillator begins consuming HOlO and 1 (Eqn 12, to form HOl) ratherChapter 2 Prelimina,y Studies Using Batch Methods -43-than producing them. On the chart recordings, the switch is represented as the suddentransition from a negative slope to a positive one. A shoulder appears to be aphenomenon whereby just as the reactants reach the critical level required to throw theswitch, there is a sudden reversal of one of the producing reactions, and productionbecomes slightly delayed. Another possibility is that the switch has been thrown in the“forward” direction, only to be thrown “backward” again only a few seconds later. Thelatter explanation appears to fit better when the shoulders of the terminal oscillations areexamined. Clearly, there is a tendency for the shoulder to rise as if it were a pos phase,only for it to come back down to its previous response level. Chemically, the processesbehind the shoulder phase are difficult to explain. One possibility may lie in theconcentration ofH20.When the “switch is thrown” there begins autocatalytic production of hypo-iodous acid:12. HOIC + I + H 2H01Dl. HOl + R20 -+ + °2 + H + H20As the concentration ofH20 decreases, reaction Dl may proceed too slowly to keep upwith reaction 12, and so HOIO and F will not be able to produce sufficient HOI beforefalling below their critical levels. The switch is then “reversed” and the oscillatorcontinues.Chapter 2 Preliminary Studies Using Batch Methods -44-2.3.2 Analysis of A.C. SignalsA data analysis strategy for processing CAE signals was outlined in Chapter 1,and will be followed here.2.3.2.1 Signal ValidationThe step immediately following signal acquisition is signal validation. Theprogram AESCHECK was used to remove those signals which were over-ranged orcaptured as a result of under-triggering.2.3.2.2 Signal AnalysisSignal classes (italics) were assigned based on the events observed and discussedabove:“indu” - if the signal was captured during the Induction phase“negs” - if the signal was captured during a Negative Slope (neg) event“poss” - if the signal was captured during a Positive Slope (pos) event“shou” - if the signal was captured during a Shoulder event“none” - if the signal was captured after events could be positively assignedClass assignments were made only for Reagent Set A experimental runs. ForReagent Set B runs, all signal classes were set to “none “.2.3.2.3 Time Series AnalysisThe program TRAPS was initially used to compute the averaged power spectra(AVP) of all the signals captured. Following that, TRAPS computed the AVP of thoseChapter 2 Preliminaiy Studies Using Batch Methods -45-Isignals assigned to each of the four classes mentioned above. The raw signal data weretransformed with a Welch window before the fast Fourier transform was applied. Theresulting power spectra were then smoothed with a 10% FFT smoothing algorithm.Reagent Set BAveraged Power Spectrum0.018 - I I0.015--C’)0.012 -C0.009 -->.0.006--I III I III0 200 400 600 800 1000 1200Frequency (kHz)Figure 2-6. AVP of Reagent Set B.The results for Reagent Set B are shown in Figure 2-6, on the same scale as theAVPs to follow (for ease of comparison). It is very difficult to make out any distinctpeaks. The most easily identified peaks are typically those due to the piezoelectrictransducer’s unique response signature. The transducer used in this study generates peaksat 786 and 1008 kHz. These peaks are difficult to see in Figure 2-6, but are visible if they-axis is adjusted as in Figure 2-7.The magnified view offered by Figure 2-7 reveals more detailed structure of theAVP. There appears to be some acoustic activity as evidenced by the region from 100 to600 kHz. Recall from the experimental section of this chapter that background noise ismostly eliminated by setting a trigger level voltage about the background noise level,leaving the conclusion that any significant intensity must be from chemical acousticChapter 2 Preliminaty Studies Using Batch Methods -46-activity. The key words here are “significant intensity”. It is too early in the analysis todetermine if relative intensities <0.002 are significant.Reagent Set BAveraged Power Spectrum0.0030>. 0.0025U)0.0020C0.0015>0.00100.00050.00001200Frequency (kHz)Figure 2-7. AVP of Reagent Set B, with expanded Y axis.The results for Reagent Set A, Run 1 are depicted in Figure 2-8. The top graph isthe AVP of all the CAE signals collected. The lower four graphs are the AVPs of thosesignals which fall during the labeled event periods. Those spectra may be tentativelydivided into two distinct regions. The first, from 100 to 700 kFIz, may be thought of asthe “reaction activity” region. The second, from 700 to 1200 kHz, is the “transducersignature” region. In the reaction activity region, each graph shows a prominent peak at500 kHz which corresponds to earlier studies conducted in this lab and publishedelsewhere. As it is unreported in other acoustically active chemical systems, it may be apeak characteristic of bubble evolution in the iodine clock reaction.The first half of the reaction activity region, Figure 2-9, contains those peaks andpeak intensities unique to a particular class. Of the four signal classes represented, theinduction period is the most unique, exhibiting two major peaks at 125 and 293 kHz. The0 200 400 600 800 1000Chapter 2 Preliminaiy Studies Using Batch Methods --I0.018 -0.015 ->,0.012 -- 0.009 -0.006 -0.003 -0.000 -0.015 -.‘ 0.012 —U)C. 0.009 —C0.006 -G) 0.003 —0.0000.0 150.012CoC® 0.009C0.006ci) 0.0030.0000 200 400 600 800 1000 1200 0Frequency (kHz)200 400 600 800 1000 1200Frequency (kHz)remaining classes have three peaks which are similar in frequency but quite different inintensity. They are summarized in Table 2-7. Given the experimental system’s frequencyresolution of approx. 2.5 kHz, there is significant correlation between the peaks in eachevent.Averaged Power Spectrum for Run II I0 200 400 600 800 1000 1200Frequency (kHz)InductionFigure 2-8. AVP of Reagent Set A, Run 1Chapter 2 Preiimina,y Studies Using Batch Methods -48-With the exception of the induction period, it is extremely difficult to separate theevents via frequency shifts in the reaction activity region of the averaged power spectra.The peak intensities offer some useful information. Peak #1 steadily decreases betweenevents, while peaks #2 and #3, increase then decrease again. Note that if the shoulderevent is not present (as in the early oscillations), then all three peaks follow the samedecreasing trend.0.0070.006 -(I)- 0.005 -0.004 ->0.003 -0.002 -0.0000.0070.0060.005x< 0.004>-0.0030.0020.000Figure 2-9. Reaction activity region of Reagent Set A, Run 1.The transducer’s signature region is readily identified from Figure 2-8 on page 47.There is little or no reaction activity in this region and the transducer’s signature peaks of786 and 1008 kHz are easily discernible.lull IllIlIlll— I I I IIi ci .Ict_ioI1- IINegative Slopi100 150 200 250 300 350 400 100 150 200 250 300 350 400Frequency (kHz) Frequency (kHz)Chapter 2 Preliminary Studies Using Batch Methods --IThe results for Reagent Set A, Run 2 are depicted in Figure 2-10. Similar toFigure 2-8 above, the top graph is the AVP of all the CAE signals collected and the lowerfour graphs are the AVPs of those signals which fall during the labeled event periods.Of the four signal classes represented in the reaction activity region (Figure 2-11),the induction period is no longer unique, exhibiting three major peaks at 125, 210 and310 kHz, similar to the remaining classes which also have three peaks located near thosefrequencies.Table 2-7. Summary of peaks in reaction activity region of Run 1.Event Peak 1 Peak 1 Peak 2 Peak 2 Peak 3 Peak 3Freq. Rel. Tnt. Freq. Rel. mt. Freq. Rel. Tnt.(kHz) (x105) (kHz) (x105) (kHz) (x105)Negative Slope 125 437 200 370 288 419Shoulder 125 415 195 458 288 453Positive Slope 127 300 198 401 291 386Chapter 2 Preliminaiy Studies Using Batch Methods -50-0.0 180.0 15>.,0.012a)- 0.0090.0060.0030.0000.0 15- 0.012U)a) 0.0090.006(Ua) 0.0030.000Averaged Power Spectrum for Run 20 200 400 600Frequency (kHz)800 1000 12000.015 -2:’ 0.012 -U)C‘) 0.009 -C0.006 -CUa) 0.003 -0.000 -Induction - Negative Slope.Shoulder0 200 400 600 800 1000 1200 0Frequency (kHz)Positive Slope..200 400 600 800 1000 1200Frequency (kHz)Figure 2-10. AVP of Reagent Set A, Run 2.Chapter 2 Preliminaiy Studies Using Batch Methods -51-0.007‘ 0.0060.0050.004>0.0031)c 0.0020.0000.0070.006C’)0.0050.004>0.0030.0020.000100 150 200 250 300 350 400 100 150 200 250 300 350 400Frequency (kHz) Frequency (kHz)Figure 2-11. Reaction activity region of Reagent Set A, Run 2.Chapter 2 Preliminary Studies Using Batch Methods -52-2.3.2.4 Descriptor Extraction and AnalysisReferring to the CAE data analysis road map which was detailed in Chapter 1, thenext step in the analysis of CAE signals is to extract descriptors from the raw signal data.The program FEATURES was used to accomplish this task. Default settings were usedwith the following exceptions:• a Welch window transform was applied to the signal data before Fouriertransformation• the bandwidth for computations was limited to the frequency response bandwidth ofthe transducer which is approximately 100 - 1200 kllzSince it is impossible to capture every CAE signal emitted by the chemicalsystem, each signal represents a sample from a much larger population. Thus a signal’sdescriptors are also a sample of a larger population. Before chemometric tests can beapplied, some knowledge of the distribution of values is important. For example, ifstatistical methods based on normal (Gaussian) distributions are applied to non-Gaussiandata, the methods may yield erroneous results leading to faulty conclusions. Reasonswhy non-Gaussian distributions are obtained may include heterogeneity of samples,rounding off error, and measurements near the detection limit.45 When a Gaussiandistribution cannot be obtained, one can use methods which do not require assumptionsabout the distribution.One method of determining the type of sample distribution involves dividing thevalues into a discrete number of buckets, which are then displayed as a histogram. Thenaked eye can then be used to judge the symmetry of the resulting graph. This method isreliable yet time-intensive. The program DESTREND implements this method.Chapter 2 Preliminaiy Studies Using Batch Methods -5-IDescriptor distributions may be viewed on the display screen or the data dumped to diskfor later plotting. Processing of the data from Reagent Sets A and B revealed that themajority of descriptors exhibited non-Gaussian distributions.One of the goals of this chapter is to correlate the visual colour changes with CAEdata. Since no colour changes were observed during the Reagent Set B experiments, it isTable 2-8. Results of Wilcoxon Test (Reagent Sets A vs. B)Set A, Run 1 vs. SetB SetA,Run2 vs. SetBDescriptor Z Score Descriptor Z Score3/8 T -0.91 3/8 T 0.354/8T 0.32 4/8T 1.525/8T -1.41reasonable to form the hypothesis that the Reagent Set B descriptors have differentdistributions than the Reagent Set A descriptors. A second likely hypothesis is that Runs1 and 2 of Reagent Set A should have similar descriptor distributions. Due to the nonGaussian nature of the data, the Wilcoxon two-tailed rank sum test was used to test ourhypotheses. This non-parametric test is described in Siegel46. The results are tabulatedbelow. Values of less than 1.96 indicate distributions which overlap. Only 4 of 26descriptors have overlapping distributions when comparing Reagent Set A with B, while16 of 26 descriptors have overlapping distributions when comparing Runs 1 and 2 ofReagent Set A.Chapter 2 Prelimina,y Studies Using Batch Methods -54-Table 2-9. Results of Wilcoxon Test for Reagent Set A. (run 1 vs. 2)Descriptor Z ScoreCrest 1.91Kurtosis 1.42tcareaJ2 -0.990-Cross 0.711/8T 1.322/8 T -0.983/8T 0.244/8T 0.615/8T 0.176/8 T 0.277/8T -0.178/8 T -0.85FrqMed 1.05FrqMean -0.95DFB4 -0.79DFB7 -1.31DFB8 1.32Chapter 2 Preliminary Studies Using Batch Methods -55-2.3.2.5 Principal Components AnalysisPrincipal components analysis (PCA), described briefly in Chapter One, was usedto reduce the multi-dimensional descriptor data into three dimensions. The data were firstautoscaled before used as input to the program PCA. The resulting eigenvectors wererotated with a normal varimax algorithm. The resulting plots showed a cluster with askewed centroid and no appreciable pattern.2.3.2.6 ClusteringThe primary purpose of clustering data is to determine if there is a statistical basisfor the classification of the data. For the purposes of CAE data, the presence of distinctclusters may indicate that the CAE is due to unique acoustic processes. In this study,there is the advantage that classes have already been assigned to signals based on thephases observed in the d.c. data. Hence, clustering may be used to provide statisticalweight to the classes already assigned.The program K-MEANS was used to execute the non-hierarchical K-meansalgorithm47.The descriptor files were first trimmed to remove any CAE signals whichcould not be classified according to the observed d.c. phases. In practice, this meantremoving all signals whose time stamp was greater than 150 seconds. The resultingdescriptors were autoscaled before clustering. Since four classes had been previouslyassigned (induction, negative slope, positive slope, and shoulder) the program K-MEANSwas instructed to separate the data into the best four clusters.One drawback to clustering experimental data is that the number of clusters mustbe chosen carefully, preferably with a priori knowledge of the system under study. WhenChapter 2 Preliminary Studies Using Batch Methods -56-Ilimited information is available, some method is needed to review the results to makesure they are reasonable. In the case of CAE data, we can exploit the frequency content ofthe signals to fulfill this role by computing the averaged power spectrum of each clusterand comparing it with the other clusters. If two clusters exhibit a power spectrum withsimilar peaks then it is reasonable to assume that the clustering algorithm has subdivideda large cluster into two smaller ones.In this work, the program TRAPS was used to compute the AVPs of each of thefour classes assigned by K-MEANS. The results for Reagent Set B are shown in Figure 2-12. The numbers located in the upper right of each graph indicate the number of signalsbelonging to that class and the percentage of signals compared to the total. The graphtitled “none” represents the power spectrum of those signals which occurred after visibleoscillations (i.e. after 150 seconds).It is evident for this reagent set that class 1 and class 4 are similar and class 2 andclass 3 are similar leading to the conclusion that there are only two distinct classespresent. The first class (from class 1 and 4) shows a typical power spectrum for noise orbackground signals and makes up 20% of the classified signals and 10% of the totalsignals. The second class (from class 2 and 3) is low intensity but dominant, whichexplains its similarity to the power spectrum for the unclassified signals.Chapter 2 Preliminaiy Studies Using Batch Methods -57-0.012—iiriiiirClass I Class 2-87(48.3%)Class 3 Class 4—- 56(31.1%)- - 7(3.9%) -EE_____________________0 200 400 600 800 1000 12000.012.11111 Frequency (kHz)None0.009 - - -ci)C0.006 —>ccici) 0.003-- -0.000 - IIIJ III 1111111 I II0 200 400 600 800 1000 1200Frequency (kHz)Figure 2-12. Power spectra of K-Means clustering results for Reagent Set B.Chapter 2 Preliminaiy Studies Using Batch Methods -58-2.4 ConclusionsAnalysis of the d.c. signal has allowed the oscillating iodine clock reaction to beclassified into four distinct phases: induction phase, negative slope phase, positiveslope phase and the shoulder phase.In the induction phase it is apparent that there are two processes taking placewhich may be attributed to the following model:x k > 02 (soin) k2 > Nucleationfollowing the rate law:d[O,]=k1X—k2[0jdtwhich upon integration yields[02] = [02 ],(l — exp(—k2t))Following the lead of Lubetkin’4,one would expect the acoustic emission intensity to beclosely correlated tok2[0](1— exp(—k2t))The negative slope phase, when 12 is decreasing and F increasing, appears to beindependent of reagent concentration. It is difficult to state this as fact on the basis of thelimited experimental data at this point. The following chapters will explore this further.The positive slope phase, when both ‘2 and F are increasing, could not be fitclosely to a standard rate law. The best fit was to a second order rate law in which it wassuggested that the rate of oxygen evolution was first order in both hypo-iodous acid andhydrogen peroxide.Chapter 2 Preliminaiy Studies Using Batch Methods --IThe shoulder phase was the most challenging to explain as it is unreported in theliterature and hence may be unique to chemical acoustic emission. A shoulder phaseoccurs when the oscillator experiences difficulty switching from one state to another.In the a.c. signal analysis section, time-series analysis of the reaction phasesprovided the basis for dividing a power spectrum into two distinct regions: the “reactionactivity” region and the transducer signature region.Multivariate statistics yielded mixed results. The majority of the descriptorsbelonged to non-Gaussian distribution populations which limited the analysis to non-parametric methods. The Wilcoxon two-tailed rank sum test validated the hypothesis to95% confidence that the descriptors from Reagent Set A and those from Reagent Set Bare drawn from statistically different populations. The test also validated the hypothesisthat the descriptors from replicate runs of Reagent Set A are drawn from statisticallysimilar populations. Thus, the statistical methods were able to distinguish between signalsdue to noise/background and those due to reaction activity.Principal components analysis of the multidimensional descriptor data did notuncover an underlying structure of the data.K-means clustering of the multidimensional descriptor data, followed by analysisof the averaged power spectrums was ineffective.-60-Chapter 3 lodate Studies - 61 -Chapter Three - lodate Studies“Attention to detail is essential for gleaning information from an unsuspecting source.”- Insp. Clouseau3.1 IntroductionThe preliminary studies documented in Chapter Two described a system whichmoved the reagents into the reaction cell manually via 25.00 mL pipettes andpipette bulbs. While practical, simple, and cost-effective, this type of delivery system isslow, lacks mixing repeatability and is prone to operator error. Since one of the goals ofthis work was to study the Briggs-Rauscher iodine clock over a range of chemical andphysical conditions, an improved delivery system was required.This chapter documents the next step in the evolution of the experimental systemfrom Chapter Two. Reagent delivery is automated using components of the FIDO system(see Glossary). The detection system is also improved by supplementing the CAEdetector with an ion-selective electrode (ISE) and an ultra-violet/visible diode-arrayspectrophotometer (UV-Vis).Obtaining the same performance from an ISE in a flow system as in a manualsystem poses some challenges. While some researchers have reported significant gains insensitivity44,this is usually at the expense of peak broadening and incomplete baselineChapter 3 lodate Studies - 62 -45 . . ‘‘recovery . A more serious problem is the appearance of overshoot peaks, firstobserved by Linder46 under steady-state conditions. These peaks have been attributed tosudden changes in activity at the electrode surface as a result of the desorption of primaryions due to the adsorption of excess interferent ions. In later work, Linder reported that noovershoots were observed when a chloride-selective electrode was used to monitor iodideions47. Based on this work, we decided to employ a chloride-selective electrode for ourstudies.The Briggs-Rauscher (BR) oscillator has been extensively studied via UV-Visspectrophotometry29.UV-Vis provides a rapid and repeatable analysis which is widelyused in flow systems. Many of the species in the BR oscillator absorb in the UV-Visregion giving the researcher a choice of wavelengths to monitor. Table 3-1 lists some ofthe species and their extinction coefficients where known. In this work, we are primarilyinterested in the aqueous iodine peak at 460nm.Table 3-1. UV-VIS absorption parameters for BR species.Wavelength Species Formula Reference(nm) (M1cm1)300 iodomalonate ICH(COOH)2 209 48310 hypoiodous acid HOl 200 49352 tn-iodide ion 13 26400 50460 aqueous iodine I(aq) 746 50Chapter 3 lodate Studies - 63-3.2 Experimental3.2.1 Briggs-Rauscher Chemical SystemOne set of stock reagents was used for these studies. The commercially availablechemicals were used without further purification. The maximum concentrations of eachstock reagent can be found in Table 3-2. The ionic strength adjuster (ISA) was preparedas 5 M NaNO3.The hydrogen peroxide concentration was determined twice daily usingtitanium(IV) oxalate and the starch solution was prepared fresh each day by vigorouslyboiling the requisite amount of starch in distilled water until the originally clear solutionturned translucent, both as mentioned in Chapter Two.Table 3-2. Reagents used for batch studies.Reagent Source and Grade MaximumReagentConcentrationH2O BDH, ACS Analytical 0.8500 MKb3 BDH, ACS Analytical 0.0375 MH2S04 BDH, ACS 0.0401 MAnalyticalCH2(COOH) Fisher, ACS 0.0370 MMnSO4 MacArthur, ACS 0.020 1 MISA Orion, ACS 0.2% v/vstarch BDH, ACS Analytical 0.13% v/vChapter 3 lodate Studies - 64 -3.2.2 Instrumental SystemThe instrumental apparatus is illustrated in Figure 3-1; single lines represent theflow of data and/or control signals in the direction of the arrows, while the larger lineswith hatchings represent the flow of reagents or the reaction mixture. A brief summary ofeach component’s make, model and settings are given in Table 3-3. Information about themanufacturers can be found in Appendix A. Selected components are described in moredetail in the following subsections.Figure 3-1. Apparatus for iodate studies. The acronyms are defined in Glossary.Chapter 3 lodate Studies - 65-Table 3-3. Apparatus list for iodate studies.Symbol Instrument Description Manufacturer Model Settings/OtherAMP conditioning amplifier Brilel & Kjer 2638 40 dB gain100 - 2000 kHzfilterCTB constant temp. bath Blue M Electric M-W varied byexperimentDACA aid and d/a converter IBM DACA varied byexperimentFIDO see Glossary in-house nia varied byexperimentHPIB see Glossary Hewlett-Packard n/a n/aISE ion selective electrode Orion 94-1 7B nlaP-PC primary computer Nora Systems 286/12 n/aPCIP pc adapter oscilloscope MetraByte PCIP 8 mV resolutioncard 2.50MHzsamp.freq.PT broad-band Brüel & Kjr 8312 40 dB gainpiezoelectric transducerRS-232 serial port interface see below n/a n/aS-PC secondary computer Campus 3 86/20 n/aComputersUV-VIS spectrophotometer Hewlett Packard 8452A varied byw/ 1 cm path length experimentquartz flow cell3.2.2.1 Constant Temperature BathThe stock reagents were poured into 250 mL Erlenmeyer flasks, covered withParafilm®, and placed inside the constant temperature bath. A small hole was punchedthrough each Parafilm® seal to allow the insertion of Teflon (PTFE) tubing. A mixture ofj Chapter 3 lodate Studies - 66-50% v/v MeOH/water was also placed in the bath in a similar manner, to be used as acleaning solution between experimental runs.An electronic temperature monitor was attached to the bath and its outputconnected to the primary computer via the DACA interface. The computer could thenmonitor and record the bath’s temperature as the experiments progressed.3.2.2.2 Flow Inlection Development and Optimization System (FIDO)This instrumentation housed the peristaltic pumps and gas-actuated valves whichwere used to deliver the reagents to the reaction vessel. The system has undergonesubstantial development and improvement since its first construction51.Details of theinterface circuitry can be found elsewhere52.New software, written by this author tooperate the system, is described in Appendix B. A schematic diagram of the flow systemis illustrated in Figure 3-2. A list of components is in Table 3-4. On the left of the figureis the constant temperature bath. The flow system is represented as the collection ofpumps (Pn) and valves (Vn). The funnel is the reaction vessel and UV-VIS is the UVVisible spectrophotometer. Unlike the previous figure, only reagent flow is indicated,there are no control or data lines shown.Chapter 3 lodate Studies - 67-Table 3-4. Flow system components.Symbol Description Manufacturer Model Other(P1-P5) Peristaltic pump Ismatec 2/6 0-76 rpm in 256 steps(P6) Peristaltic pump Alitea C4V 0-125 rpm in 4096 steps(V1-V2) Gas-actuated valve Rheodyne type 50 six portThe five vertical peristaltic pumps drew reagents from their containers in theconstant temperature bath to the reaction vessel at a rate of Ca. 5 mL/min. Tygon® pumptubing, 1.52 mm I.D., was used for all pumps except the pump interfaced to themethanol/water line. This pump required Viton® pump tubing (also 1.52 mm I.D.) whichFigure 3-2. Schematic of the flow systemChapter 3 lodate Studies - 68 -is solvent resistant. (aside: two forms of tubing are normally used in a flow analysisproject of this nature. The first, Teflon® (PTFE) tubing, is the tubing used to deliver thereagents to and from the pump. In this work, insulated 0.8 mm I.D. tubing was used. Thesecond form, poly(vinylchloride) (PVC) pump tubing, is used with the pumps themselves.The rollers on the pump alternately pinch the tubing, and roll to propel the desiredsolution. The PTFE tubing should fit snugly inside the pump tubing.) Note that while thetubing in the figure is drawn as different lengths, in fact all the tubing was carefully cut tothe same length (Ca. 200 mm), ensuring that the necessary quantities of reagent arrived atthe reaction vessel nearly simultaneously.Prior to an experimental run, the pumping rate of pumps P1 - P6 were calibratedutilizing a simple gravimetric procedure where a known amount of distilled water ispumped during a known period of time at a fixed RPM, and the weight differencemeasured.During an experiment the reagents were delivered simultaneously to the reactionvessel at a rate of Ca. mL/min via pumps P1 to P5. The total initial reaction volume was10.0 mL. Once all reagents had been delivered, the flow system activated pump P6 tobegin cycling the reaction mixture through the UV-Vis detector. With valve V2 in the“cycle” position, the reaction mixture was returned to the reaction vessel. This constantcycling at a rate of ca. 8.5 mL/min provided adequate mixing and a mean residence timeof ca. 120 seconds. With valve V2 in the “waste” position, the reaction mixture was sentto the waste container.Chapter 3 lodate Studies - 69 -3.2.2.3 Reaction VesselThe reaction vessel is illustrated in Figure 3-3. The compact design precludes theuse of a mechanical stirrer; all mixing was provided by the cycling mentioned above. Inthis way, any possible reactions which might have occured on the stirrer’s metal surfaceare avoided.The reaction vessel consisted of a short-stemmed Pyrex® funnel with an internaldiameter of 50 mm at its widest point, an internal diameter of 5 mm at the stem, a stemlength of 25 mm, and an overall length of 60 mm. The funnel’s stem was pushed througha rubber stopper which in turn was fastened to a retort stand via a standard clamp (notshown for clarity). The rubber stopper provided limited insulation against benchtopvibration.A piezoelectric transducer was affixed to one side of the funnel (through a thinlayer of acoustic couplant grease) with black electrical tape. The tape was wrappedaround both the transducer and the funnel in several layers to provide adequate thermalinsulation and protection from ambient light.Flow Linesfrom FIDOand UV-VisTo UV-VisTransducerFigure 3-3. The reaction vessel in detail.Chapter 3 lodate Studies - 70 -3.2.2.4 Ion Selective Electrode (ISE)The iodide potential in solution was measured via an ISE mounted such that theelectrode’s sensing surface was completely immersed in the reaction mixture. The ISEused was rated to be sensitive to iodide over a range of 1 M to 5x1 06 M, and boasted aninternal reference electrode which obviated the need for an external one. Twice each day,the electrode operation was tested against iodide standards and the resulting slope of thecalibration curve checked against the manufacturer’s specifications. Before eachexperiment, the sensing element required polishing to remove any build-up of hypohaloussalts on the membrane’s surface. Following that, an E° value was obtained by immersingthe electrode in a solution containing all reagents except the iodate.3.2.2.5 Primary ComjuterThe primary computer exerted control over and collected data from the variousinstruments via three different interfaces. The data acquisition and control adapter(DACA) was used to control the flow system, to collect data from the temperature sensorin the constant temperature bath, the ion selective electrode, and the transducer (via theamplifier’s d.c. rms output). The HPIB interface provided control over the UV-Visspectrophotometer. The RS-232 port, through a null-modem cable, allowed the primarycomputer to communicate with the secondary computer.3.2.2.6 Secondary ComputerThe secondary computer’s single task was to collect CAE signals from thetransducer via the amplifier’s a.c. output. The acquisition and storage of CAE signalsproved to be too demanding to be included with the other devices, which is why adedicated computer was necessary. As mentioned above, the RS-232 port providedChapter 3 lodate Studies - 71-communication and synchronization between the two computers. This work is the firsttime the flow analysis system had been used with two computers. Significant softwaredevelopment by this author was necessary to complete the system.3.23 MethodsThe immediate goal was to collect information from three separate detectors overa range of reagent concentrations and temperatures. The result was a three dimensionalgrid with the axes reaction temperature, reagent, and concentration, measured as afunction of time.3.2.3.1 Reaction TemperatureSince the BR oscillator operates well under room temperature conditions, weselected a series of temperatures close to room temperature: 20, 25, 30 and 35 degreesCelsius.3.2.3.2 Reagents and ConcentrationsIdeally, we would like to systematically vary all the reagents which make up theBR oscillator, but time and equipment constraints limit the scope of our investigations.Initially we decided to explore the role of iodate in the oscillator by varying the [iodate]from 0.0038 to 0.0375 M. A series of hydrogen peroxide experiments were alsoconducted, but only at one temperature: 35°C. The [H20j was varied from 0.0850 to0.8500 M. TheH20work is covered in Chapter Four.Each concentration region was explored in a series of ten steps, from low to highconcentration. While one reagent was being varied, all others remained unchanged.Chapter 3 lodate Studies - 72 -Constant reaction volume was maintained by the automated addition of distilled water.Table 3-5 outlines the experiments conducted. Since each series varies the concentrationof a single reagent in ten steps, the table represents a total of 50 experiments.Note that series PF-G is an exception. This series was intended as a trial run totest that the hardware and software was functioning correctly. However, the results fromthis run proved interesting and it was decided to include the findings in this discussion.Table 3-5. Experimental Series.Series Name Temp (°C) Reagent VariedPF-G 20 iodatePF-K 25 iodatePF-O 30 iodatePF-T 35 iodateChapter 3 lodate Studies Using Stopped Flow Methods - 73 -3.3 Series PF-G, lodate varied from 0.005 to 0.050 M, 20CThis series was designed to test that all the equipment and software wasfunctioning acceptably. The concentrations of the reagents used in this series are listedhere:H2S04 0.054MMn(II) 0.020 MCH2(COOH) 0.050 MH20 1.24 MKb3 0.005 to 0.050M (inten steps)The instrument settings are shown in Table 3-6. All instruments except theoscilloscope were operated in a continuous mode which means the instruments collecteda data point at the end of every sampling period. The scope, however, was operated in atriggered mode which means that the scope captured a CAE signal every time a presettrigger level was breached. See the section Timing and Methods in Chapter Two. In thiswork with the PCIP oscilloscope adapter card, a minimum of 1.5 seconds was required tocapture and store a CAE signal.Chapter 3 lodate Studies Using Stopped Flow Methods - 74 -3.3.1 Data Work-upThe raw data collected from the various instruments required some preprocessingbefore an analysis could be conducted.• CAE d.c. - the digitized signal was corrected for amplification and then thebackground level was subtracted (30 seconds of background level was capturedbefore every experimental run).• ISE - the data from the ion-selective electrode were initially corrected foramplification, the value of E° subtracted, and then the [iodidej assigned based oncalibration curves. The electrode was calibrated at the start of every experimentalseries.• UV-Vis - the data from the UV-Vis spectrophotometer were automatically blank-subtracted by the UV-Vis itself before transmission to the primary computer (a blankwas taken before every experiment). The concentration of aqueous iodine wascalculated from the absorbance at 460mmTable 3-6. Instrument parameters for series PF-G.Instrument Samp. Period Other settingsCAE d.c. 4.0 s 64 dB gainISE 4.Os 20x linear gainscope mm. 1.5 s 1600 mV input range, 2.5 MHz digitization rateUV-Vis 4.0 s 300 to 600 nm; 2 nm resolutionChapter 3 lodate Studies Using Stopped Flow Methods 75 -3.3.2 PF-G-1 and 2The data from PF-G-1 (initial {iodate] = 0.005 M) and 2 (initial [iodate] = 0.010M) are presented in figure C-5. The first run shows no UV-Vis or ISE activity. There issome acoustic activity, but this may be due to line noise or to the side reaction involvingthe oxidation ofmalonic acid by Mn(III):CH2(000R) + 2H0 + Mn(III) —* H000H + 2C0 + 6H + Mn(II)This particular reaction has not been observed by other workers, but similar work hasbeen done with Ce(IV)53. Mn(III) can also oxidize organic compounds, but suchexperiments are difficult to do quantitatively54.An attempt was made to determine if the malonic acid oxidation was the primaryacoustic source by directing any evolved gas into the sampling ioop of a gas partitioner(Fisher-Hamilton, Model 29). This instrument is a gas chromatograph specificallydesigned for the quantitative determination of substances which are gaseous at roomtemperature. It employs a dual-column, dual-detector chromatographic system to separateand measure carbon dioxide, oxygen, nitrogen, methane, and carbon monoxide.Unfortunately, reproducible results could not be achieved and so conclusions cannot bedrawn at this time.The second run shows a slight increase in both the UV-Vis and the ISE trace,indicating the build-up of aqueous iodine and iodide ion., but not enough to commenceoscillations. The CAE trace also shows a little more activity suggesting that anacoustically emissive chemical reaction is taking place with greater vigor.Chapter 3 lodate Studies Using Stopped Flow Methods - 76-3.3.3 PF-G-3 and 4In PF-G-3 (initial [iodate] = 0.015 M, figure C-6) the ISE trace shows fourdistinct regions: an induction period of ca. 20 seconds, a reagent “build-up” phase of ca.100 seconds, regular, small amplitude oscillations for ca. 220 seconds and finally a droptowards the baseline. Normally only three regions are expected: induction, oscillation andtermination (recall the experiments in Chapter Two). The presence of four regions may bedue to lack of rapid mixing. Cooke noted that in the absence of stirring, the behaviour ofthe BR oscillator is markedly dependent on the size of the reaction vessel. Specifically,Cooke observed that the amplitude of iodide ion oscillations decreased, then increasedagain as the reaction proceeded55. This may be due to variation in the amount ofdissolved oxygen. If so, then the behaviour of the BR oscillator would be dictated by theavailability of reaction vessel surface nucleation sites. The accompanying CAE would bedependent upon the available surface area of the gas/liquid interface to accomodatebubble bursting. It is with these considerations in mind that a funnel-shaped reactionvessel was chosen which affords large surface areas for both bubble nucleation andbubble bursting while simultaneously providing sufficient depth to allow the submersionof the ISE.The small size of the reaction vessel prohibits the use of stirring devices. Instead,mixing is provided by continuously drawing the reaction mixture from the bottom of thereaction vessel, through the UV-Vis, and back into the top of the reaction vessel via apump which circulates the reaction mixture at a rate of Ca. 8.5 mL/min. This system is notperfect, but provides better results than a solely diffusion-controlled system.The UV-Vis trace corroborates the ISE trace. An added feature is the steadypositive slope indicating a build-up of aqueous iodine. The data with the induction periodChapter 3 lodate Studies Using Stopped Flow Methods - 77- Iremoved, was plotted LN([iodine) and l/[iodine] vs time. and are depicted in Figure 3-4.A linear regression was performed on each plot. The correlation coefficients for the threeLN(iodine) plots are presented. It is apparent that the rate does not follow a strictly firstor second order trend.12.0 -— 11.0 -G)0- 10.0 --JFigure 3-4. Plots ofLN([iodinej) and 1/{iodine] vs. Time for runs PG-G-2, PF-G-3and PF-G-4. In all cases the induction period data has been removed.Chapter 3 lodate Studies - 78-C4C,Li011.010.00)0- 9.0-J8.012.011.0ci)•00- 10.0-j9.0Time (s)100 200 300 400 500 600 100 200 300 400 500 600Time (s)C)C,U03.Oe-425e-42.Oe-41.5e-41.Oe-45.Oe-50.Oe+01.Oe-47.5e-55.Oe-52.5e-50.Oe+01.Oe-47.5e-55.Oe-52.5e-50.Oe+0Time (s) Time (s)100 200 300 400 500 600 100 200 300 400 500 6000U00e00012=0.8879.0 I I100 200 300 400 500Time (s)600 100 200 300 400 500 600Time (s)Chapter 3 lodate Studies Using Stopped Flow Methods -- IThe CAE trace for PF-G-3 (Figure C-6) shows sporadic oscillations while that forPF-G-4 shows almost regular oscillations of varying amplitudes. Of interest here is theapparent lack of an induction period. Recall from Chapter Two that a significantinduction period was observed in all the experiments discussed. Also note that both theISE and UV-Vis traces suggest an induction period is occurring. This discrepancy meritsfurther investigation.The ISE data from the four runs discussed above indicate that under theseconditions, oscillations occur when the initial iodate concentration is greater than 0.010M (PF-G-3). The UV-Vis data are inconclusive on that point. The CAE d.c. does notshow significant activity until PF-G-4 ([iodate]=0.020 M) which is in keeping with thework of Roux and Vidal56 which found the minimum [iodate] required for oscillations tobe 0.0 19 M (under similar conditions).The plot of [iodine] vs. time for PF-G-2, and to some extent PF-G-3, appears toshow some periodicity, implying the presence of low frequency oscillations. Fouriertransforms of these data sets showed no low frequency peaks and the autocorrelograms(Figure 3-5) were also negative. The apparent oscillations cannot be explained at thistime.Chapter 3 lodate Studies Using Stopped Flow Methods - 80 -—1I I I I I I I I I I I I I I I I I40 60 80 100I I I I i i i i I i i i I-20 0 20 40 60I I I I I I IIII liii III80 100 120I I I II I I I I I I I I I I I I I I I I I I I I I I I I I I I I I-20 0 20 40 60 80 100 120Time delay (s)Figure 3-5. Autocorrelograms ofUV-Vis data for series PF-G-2, 3 and 42I0-—1-2-PF-G-2- 0000000000- 000000- 000000000I I I I I i i i i I i i i i I I I I-20 0 20-32I0--120PF-G-3I II0000000000Ooo,I I I I , I I I I I I , I I ,—1 —-2--3-2II I I I I I I I I I I I I I II I I I IPF-G-4I I I I I I I I I I I I I I I I I I I I II I I I I I I I I I I I I I I I-2 I IChapter 3 lodate Studies Using Stopped Flow Methods - 81 -3.3.4 PF-G-5 through 10The initial iodate concentrations studied are G-5 (0.025 M), G-6 (0.030 M), G-7(0.03 5 M), G-8 (0.040 M), G-9 (0.045 M), and G-10 (0.050 M).In these runs, the BR reaction is seen to oscillate strongly. This provides peak datafrom which reaction rates can be calculated and plausible mechanisms introduced. Thusthe data from these runs are analyzed a little more rigorously than for the previous runs.3.3.4.1 Peak Count and FreciuencyThe data from runs 5 through 10 are shown in the figures C-7 through C-9. TheUV-Vis trace (at 460 nm) takes on a skewed sawtooth appearance implying that theiodine production rates are faster than the iodine consumption rates. The ISE trace has asimilar sawtooth waveshape while the CAE traces have a variety of waveshapes. Severaltrends are more easily seen when peak data are compiled into a table, as in Table 3-7.Peak data were not included from the UV-Vis because reliable data could not be collectedbeyond PF-G-7 due to saturation of the detector. A shorter path length flow cell was notavailable to compensate for this.For the purposes of this work, a peak visible in the data trace is consideredsynonymous to an oscillation and is defined as having an initial rise (positive slope) of atleast 12 seconds duration, a peak of not more than four seconds duration, followed by adrop (negative slope) of at least 12 seconds duration. This definition is not entirelyarbitrary: the data were sampled at a rate of 0.25 Hz, thus each slope is guaranteed aminimum of three data points. Hence, each peak is defined by a minimum of ten datapoints.Chapter 3 lodate Studies Using Stopped Flow Methods - 82 -Table 3-7. Peak parameters for Series PF-G.Series initial CAE DC CAE DC ISE ISE[iodate] Number Number of Number Number ofof peaks peaks/mm. of peaks peaks/mm.PF-G-5 0.020 M 8 1.60 13 2.60PF-G-6 0.025 M 11 2.20 13 2.60PF-G-7 0.030 M 12 2.40 12 2.40PF-G-8 0.035 M 12 2.40 13 2.60PF-G-9 0.040 M 10 2.00 10 2.00PF-G-10 0.045 M 11 2.20 -- --The “number of peaks” column in the table refers to the number of peaks between100 and 400 seconds. This range was chosen to help eliminate interference from eitherthe induction or termination phases. The CAE d.c. data show that as the initial [iodate]increases, the number of peaks increases until a maximum is reached at PF-G-7. Then thenumber of peaks begins to decrease as [iodate] increases. The number of iodide peaks,however, remains generally constant with increasing [iodate] until PF-G-9. The exceptionto these observations is PF-G- 10, which had the strongest initial concentration of iodate.In this run, the iodide concentration remained low for Ca. eight minutes before increasingbeyond the capacity of the ISE to measure. Such behaviour typically occurs when theChapter 3 lodate Studies Using Stopped Flow Methods - 83-ISE’s membrane has become contaminated with ‘2 precipitate. There was also an adverseeffect on the CAE which posted low peak numbers and frequency.These differences may be explained by examining the kinetics of the peak slopes.Cooke5 reported that iodine production was enhanced by an increased iodateconcentration while iodine consumption remained nearly constant. To test thisobservation against our data, the slopes of all the peaks were measured and thenaveraged. The results are shown in Table 3-8. Recall the terminology introduced inChapter Two where a Neg phase referred to the part of the peak where d(response)/dt hasa negative slope and a Pos phase a positive slope. In the case of the ISE (monitoringiodide ion), a Pos phase represents production of iodide ion and a Neg phase isconsumption of iodide ion. For the UV-Vis data (monitoring iodine), a Pos phaserepresents production of iodine and Neg phase consumption of iodine. In Chapter Two itwas suggested that the primary source of CAE in the BR reaction is oxygen evolution, inwhich case the phases represent increasing and decreasing gas evolution, respectively.Chapter 3 lodate Studies Using Stopped Flow Methods - 84-Table 3-8. Reaction rates for Pos and Neg events for series PF-G.Run CAE CAE [iodide] [iodide] [iodine] [iodine]Pos Neg Pos Neg Pos Neg(x103 mV.s’) (x103mV.s1) (x103M.s1) (x103 M.s’) (x102Ms1) (x102M.s’)5 5.20±0.48 -5.20±0.51 9.98±0.20 -8.93±0.32 29.2±4.2 -24.1±2.76 6.10±0.40 -6.74±0.28 11.4+0.41 -8.12±0.39 34.3±2.3 -47.4±3.57 8.96±0.44 -8.19±0.54 14.5±0.51 -9.61±0.25 46.6+2.0 -57.5±2.58 8.94±0.40 -8.23±0.43 16.5±0.52 -14.9±0.44 75.0±3.8 -61.9±3.09 10.2±0.72 -9.43±0.61 14.1+0.49 -7.54±0.38 -- --10 11.6±0.75 -9.11±0.59 -- -- -- --3.3.4.2 Peak Reaction RatesThe data in the table for all three analytical methods show a steady increase inreaction rate with increasing run number (hence increasing initial [iodate]), for both Posand Neg phases, the only exception being the ISE data for PF-G-9 and the CAE data forPF-G-10 (already noted above). A graphical depiction (Figure 3-6) of the data providesinteresting results. In agreement with Cooke’s observations, the iodine production ratesshow the beginning of an almost exponential growth, while the iodine consumption ratesdecrease at an almost exponential rate. The CAE Pos rate is nearly linear (r2 0.9 14) andclosely parallels the iodide ion production rate (r2 = 0.982). The CAE Neg rate is alsolinear (r2 = 0.935) but the iodide ion consumption rate rapidly deviates from linearity forinitial [iodate] = 0.03 5 M.Chapter 3 lodate Studies Using Stopped Flow Methods - 85 -12- 20- -80UA15- o -608-A U10- A 0 -400 iiA5- - 200>E-° A [icdde} xs ° jcdrie] ,ws- 0 ‘• CAEnsg [icdcIeji • [icdr]i--cw •A .2IAa-8--15- • • --60-12- -20- I I I I --800.020 0.025 0.030 0.035 0.040 0.045 0.050irit [iocke] (M)Figure 3-6. Reaction rates for different initial iodate concentrations.3.3.4.3 Phase RelationshipsPhase relationships play a vital role in understanding the BR oscillator. They aredifficult to see, however, unless the reaction is oscillating strongly. Figure 3-7 shows thephase relationships for the strongly oscillating system present in PF-G-7. As iodine fallsto its minimum concentration, iodide ion increases to its maximum concentration. ThisChapter 3 lodate Studies Using Stopped Flow Methods - 86 -relationship has been noted by other workers57. The CAE appears to parallel theproduction and consumption of iodine. This was an unexpected result, however, asoxygen is evolved primarily during iodide ion formation:12. HOIO + I + H 2H01Dl. HOI + H20 > + °2 + H + H20Ii. HOl + I + H 12 + H20Reaction Dl is the rate-limiting step in this sequence, and hence it is expected thatiodine production would be significantly out of phase with oxygen evolution. The phaserelationship which was observed may be explained by drawing conclusions from thework of Wentzell et a158. In that work, it was suggested that oxygen produced fromdissolved hydrogen peroxide is depleted through a diffusion pathway to atmospherebefore bubble nucleation is energetically favoured. If the same model can be applied tothis work, then a time delay would be expected from the time reaction Dl commences tothe time at which bubble nucleation occurs (and hence acoustic activity is detected). Thisdelay would alter the phase of the CAE, and might make it appear that CAE is in phasewith iodine instead of with iodide ion.Rubin and Noyes studied homogeneous nucleation of bubbles resulting fromsupersaturated solutions59.They found there existed a threshold for nucleation beyondwhich it was impossible to push the level of supersaturation. Thresholds for diatomicgases such as 02 all fell between 0.012 and 0.07 M, while that for CO2 was 0.4 M. Theyalso noted that temperature had little effect on the threshold.Chapter 3 lodate Studies Using Stopped Flow Methods - 87-j 0.20-8240 244 248 252 256 260 264 268 272 276 280Time (s)Figure 3-7. Observed phase relationships: [iodine] (top), CAE d.c. (middle) and[iodide] (bottom).3.3.4.4 Extent of ReactionRecall from the Introduction in Chapter One, that an acoustic emission wave is atransient, elastic wave generated by a rapid mass motion of a collection of atoms. Thusevery CAE “burst” represents a physicochemical event. If those bursts are integrated overtime, the resulting curve represents the amount of chemical acoustic emission produced.Figure 3-8 shows the integrated CAE d.c. curves (top) and the integrated CAE d.c. at time= 600 seconds as a function of initial iodate concentration (bottom).The data for PF-G- 1 through PF-G-3 fit to a zero order function indicating that therate of CAE is independent of the concentrations of the species involved. PF-G-4 (initial[iodate] = 0.020 M) through PF-G-9 fit well to a first order rate law. Note the commonChapter 3 lodate Studies Using Stopped Flow Methods - 88-Table 3-9. Curve fit results of integrated CAE d.c. data.Series initial r2 see k. 3[iodate] (xlO s )PF-G-4 0.020 M 0.999 0.202 1.21PF-G-5 0.025 M 0.998 0.320 2.69PF-G-6 0.030 M 0.998 0.364 2.42PF-G-7 0.035 M 0.998 0.535 1.13PF-G-8 0.040 M 0.996 0.840 1.17PF-G-9 0.045 M 0.997 0.971 1.22induction period of Ca. 25 seconds, which was not apparent in the peak data, but is inevidence in this graph as a linear portion of the curve.The rates in the above table show an interesting trend. There is a steady increasewhich reaches a maximum at PF-G-5, suddenly plummets, then begins to rise again. Notefrom Table 3-8 (page 84) that all three analytical methods reported similar rates for theirpos and neg phases, suggesting that the oscillator is working at peak efficiency. Areasonable conclusion is that for the concentration range studied, and at theconcentrations of the other components used, an initial [iodate]=0.025 M is optimum foroscillations to occur.The data in the lower graph produces a straight line when plotted as in Figure 3-9.The total integrated CAE decreases with increasing initial [iodate] at a rate of -35.0 M’.This is in agreement with Cooke3°who reported that cessation of the oscillations leadingto iodine precipitation was favoured by an increase in iodate concentration. This isbecause high concentrations of iodate favour iodine production over iodine consumptionChapter 3 lodate Studies Using Stopped Flow Methods - 89 -(via iodination of malonic acid). These results also support the hypothesis that the sourceof chemical acoustic emission in the BR oscillator is bubble nucleation rather than iodineprecipitation, or else we would see increased integrated CAE as initial [iodate] increases.Chapter 3 lodate Studies Using Stopped Flow Methods - 90-Figure 3-8. Top: integrated CAE d.c. signal vs. reaction time. The numbers besideeach curve identify the experimental run. Bottom: Integrated CAE d.c. (at time600 sec) vs. initial iodate concentration.605040E30wC-)- 20a)3)wi06050 -CoCCDIL 40 -(13>C.)uJ20-c’) 10 -C0 100 200 300 400Time (s)500 600 7000970084006300521 000.00 0.01 0.02 0.03initial [lodate] (M)0.04 0.05Chapter 3 lodate Studies Using Stopped Flow Methods - 91--04-wC-)0G)-I-,Cu0)G)Cz-J3-5Integrated CAE Study02= 0.9962rate = -35.04 M1I I0.00 0.01 0.02 0.03 0.04 0.05[initial iodate] (M)Figure 3-9. Plot of LN(integrated CAE d.c.) vs. initial iodate concentration.Chapter 3 lodate Studies Using Stopped Flow Methods - 92 -3.3.5 Series PF-G: ConclusionsThe data shown in figure 3-9 illustrate the quantitative analytical potential ofchemical acoustic emission. The agreement was better than expected.The ISE’s membrane surface required careful polishing with emery cloth at thebeginning of every experimental run to remove build-up of AgI on the surface. The UVVis detector saturated before the end of the experiments. The next section will detail theimprovements to the experimental system to overcome these limitations.The experimental runs in which regular oscillations were not evident (runs PF-G1 through 3), provided kinetic information about the reduction of iodate to iodine in thepresence of hydrogen peroxide. The rate of iodine production was found to obey a firstorder rate law.The data from the ISE suggest that the minimum concentration of iodate requiredto commence oscillations is 0.015 M. The CAE data does not concur, and insteadsupports the claim of Roux and Vidal56 that a minimum of 0.019 M is required.The reaction rates of the Pos and Neg phases indicated that the rate of iodineproduction increases with increased iodate while the rate of consumption increases onlyslightly, which confirms Cooke’s observations30.The CAE’s Pos rate was found toclosely parallel the iodide ion’s production rate, but the CAE’s Neg rate paralleled theiodide ion’s consumption rate only at low iodate concentrations. At high iodateconcentrations, the iodide ion consumption rate increases dramatically.Even though reaction rates for the CAE parallel those for the iodide ion, the phaserelationships showed that CAE parallels that for iodine, not iodide ion.Chapter 3 lodate Studies Using Stopped Flow Methods - 93 -Finally, the case of the missing CAE induction period was solved. While noteasily visible in the raw CAE traces, an induction period of ca. 25 seconds is apparentfrom the CAE cumulative sum graph (Figure 3-8 on page 90). The experiments in whicha reaction was not seen to take place (PF-G-l through 3) resulted in a linear function ofthe integrated CAE d.c. The remaining series PF-G-4 through PF-G-9 fit well to a firstorder function. The integrated CAE d.c. at time 600 seconds as a function of initial[iodate] was found to decrease at a rate of 35.0 M’ indicating that CAE is in the BRoscillator is due to bubble nucleation rather than precipitation.Chapter 3 lodate Studies Using Stopped Flow Methods - 94 -3.4 Series PF-K, lodate varied from 0.004 to 0.037 M, 25CSeries PF-G provided useful information about the chemical system as well as theinstrumental system. It demonstrated that an ISE was not easily adapted to an automatedflow system and that the UV-Vis reached saturation conditions before the entire seriescould be completed.Series PF-K is the first of the “formal” iodate series, and hence the first series touse a slightly modified experimental system. The changes are as follows:1. The ISE was no longer in use. The need to polish the ISE’s membrane before everyexperimental run negated one of the advantages of an automated system: the ability tooperate unattended. Hence it was decided to take the ISE out of the loop and tocontinue using the CAE and the UV-Vis detectors. Another concern was the potentialpoisoning of the ISE. Sandifer exposed a chloride ISE to bromide and found thatprolonged exposure led to the formation of AgClBr crystals throughout themembrane, not just at the surface60.However, no gross deterioration of ISE responsehad been detected in the present work.2. The sampling period was decreased from 4.0 to 3.0 seconds to increase the resolutionof the spectra. Because limited data space was available, the total acquisition timewas decreased to 480 seconds from 600. We felt that this would still provide a safemargin as the PF-G experiments indicated that oscillations generally terminatedbefore 420 seconds.3. Initial concentrations of key species were reduced by approximately 25% to avoidIJV-Vis detector saturation. See Table 3-2 above for the complete list of reagents andtheir concentrations.Chapter 3 lodate Studies Using Stopped Flow Methods - 95 -3.4.1 Data Work-upThe raw data collected from the various instruments required some preprocessingbefore an analysis could be conducted. The work-up is identical to that described for theprevious series.Table 3-10. Instrument parameters for series PF-K.Instrument Samp. Period Other settingsCAE d.c. 3.0 s 64 dB gainscope mm. 1.5 s 1600 mV input range, 2.5 MHz digitization rateUV-Vis 3.0 s 300 to 600 nm; 2 nm resolution3.4.2 PF-K-1 and 2The data from PF-K-1 and 2 are presented in FigureC-lO. PF-K-1 (initial [iodate]= 0.004 M) shows no UV-Vis activity other than the occasional bubble resulting in asharp spike on the UV-Vis trace. There is some periodic acoustic activity.The data for PF-K- 1 shows an initial periodicity equal to 3t, or 18 seconds. PF-K2 shows the same periodicity. In the absence of visible oscillations, this long-periodperiodicity is likely due to the catalytic decomposition of hydrogen peroxide which has anet stoichiometry of:Eqn3-1. 2H0 —> 2H0 +Chapter 3 lodate Studies Using Stopped Flow Methods - 96-If the model proposed by Wentzell (discussed in the last section) is accurate, thenbubble formation occurs only when the dissolved oxygen process is energeticallyunfavourable, i.e. when the solvent has reached oxygen saturation. Hence, there will be aperiodic “pressure release” in the form of bubble evolution which appears on the CAEtrace as a low rise.The second run, PF-K-2 (initial [iodate] = 0.008 M), shows a slight increase in theUV-Vis after Ca. 360 seconds, indicating a build-up of aqueous iodine, but not enough tocommence oscillations. The CAE trace also shows a little more activity suggesting thatan acoustically emissive process (as described above) is taking place with greater vigor.3.4.3 PF-K-3 and 4In PF-K-3 (initial [iodate] 0.011 M), the UV-Vis trace shows a lot of activity,suggesting the possibility that the BR reaction is taking place. However, the irregularshape and distribution of the peaks indicates that the activity is due to extensive bubbleevolution, rather than iodine absorption. Also, looking ahead to PF-K-4 (initial [iodate] =0.015 M), we see a mostly quiet trace which is expected for this level of iodate(cf. PF-G-3).The regular oscillations in the previous CAE traces have become less regular inPF-K-3, but larger in amplitude suggesting that either the existing acoustically emissiveprocess is being replaced by another CAE process or it is being supplemented by one.The autocorrelogram for PF-K-3 shows that the periodicity has increased to 4T, or 24seconds, lending support to the hypothesis that a change is underway. Theautocorrelogram for PF-K-4 shows almost random fluctuations and the results for PF-K-5through 10 also yield no useful information and so are not included in this work. Thissuggests that as the [iodate] increases to a level where oscillations become possible, theChapter 3 lodate Studies Using Stopped Flow Methods - 97 -catalytic decomposition of hydrogen peroxide is no longer the dominant method ofoxygen evolution. This hypothesis is tested further in the sections to follow.Chapter 3 lodate Studies Using Stopped Flow Methods - 98-0.40.0-0.2-0.4P0.2P0 5 10 15 20-0.2 --0.40.40.2:•:— I I I I I I I I I I I I I I I I0 5 10 15 20— I I I I I I I I I I I I I IPF-K-3:— I I I I I I I I I I I- 1111111111111111\ PF-K-4- ---0.2-0.40.40.20 5 10 15 20-0.2-0.40 5 10 15 20tFigure 3-10. Autocorrelograms of CAE d.c. data for series PF-K-1 through 4. Thetime constant (t) is six seconds.Chapter 3 lodate Studies Using Stopped Flow Methods - 99 -3.4.4 PF-K-5 through 10The initial iodate concentrations studied are K-5 (0.019 M), K-6 (0.023 M), K-7(0.026 M), K-8 (0.030 M), K-9 (0.034 M), and K-iO (0.037 M).3.4.4.1 Peak Count and FrequencyThe data from runs 5 through 10 are shown in the following figures c-i 2 throughC-14. The UV-Vis trace for PF-K-5 shows irregular peaks, which are likely due to acombination of the initial stages of the oscillator and bubbles in the light path. Theremaining UV-Vis traces take on a skewed sawtooth appearance implying that the iodineproduction rates are faster than the iodine consumption rates. Indeed, the laterexperiments (cf. PF-K-8 and 9) show a pronounced change in the consumption rates.Several trends are more easily seen when peak data are compiled into a table, as in Table3-11.Chapter 3 lodate Studies Using Stopped Flow Methods - 100 -The data in Table 3-11 show some interesting trends, which are detailed in thefollowing paragraphs. The exception appears to be PF-K-10 (initial [iodate] = 0.037 M)which shows a sharp decrease in the number of CAE d.c. peaks and a sharp increase inthe number of UV-Vis peaks compared with PF-K-9. This cannot be adequatelyexplained at this time.The CAE d.c. data from PF-K-5 to PF-K-9 show that as the initial [iodate]increases, the number of peaks increases to a maximum at PF-K-6, then drops. The UVVis data for PF-K-5 to PF-K-9 shows a similar trend, with the exception that themaximum is reached at PF-K-7.Interpretation of the UV-Vis results can be explained by the difference in iodineproduction and consumption rates. As mentioned in the discussion for series PF-G,Table 3-1 1. Peak parameters for series PF-K.Series initial CAE DC CAE DC UV-Vis UV-Vis[iodate] Number Number of Number Number ofof peaks peaks/mm. of peaks peaks/mm.PF-K-5 0.019M 10 2.00 14 2.80PF-K-6 0.023 M 13 2.20 16 3.20PF-K-7 0.026 M 8 1.60 17 3.40PF-K-8 0.030 M 8 1.60 15 3.00PF-K-9 0.03 4 M 11 2.60 12 2.40PF-K-10 0.03 7 M 4 0.80 20 4.00j Chapter 3 lodate Studies Using Stopped Flow Methods - 101 -increasing the [iodate] tends to increase the iodine production rate greater than theconsumption rate. Thus, we would expect the following situations:Low fiodate]: Production rate slightly larger than consumption rateIn this situation, ‘2 would be produced at a slightly greater rate than could bescavenged by the organic substrate. The peak shape should consist of a shallowrise to a peak followed by a sharp drop. Moderate oscillation periods are expected.Moderate [iodate]: Production rate nearly equal to consumption rateHere, the increased [iodate] increases both the production and consumption ratesuntil they are nearly equal. Hence, we expect to have a large number ofoscillations over a large oscillation period. The reaction proceeds until one of thereagents is depleted (usually the malonic acid). Symmetrical, needle-shaped peaksare expected to dominate.High fiodate]: Production rate much greater than consumption rateAt this level of [iodate], the production rate has increased more quickly than theproduction rate, and highly skewed peaks occur. Short oscillation periods areexpected and the reaction terminates with extensive iodine precipitation.3.4.4.2 Peak Reaction RatesThe reaction rates for thepos and neg phases of the CAE d.c. and UV-Vis data arein Table 3-12. The UV-Vis data lends support to the above discussion where it can beseen that the iodine production and consumption rates are nearly equal for PF-K-7 (initial[iodate] = 0.026 M) which is also the experimental run with the most peaks. In theChapter 3 lodate Studies Using Stopped Flow Methods - 102-previous series, PF-G-5 (initial [iodate] 0.025 M) also posted nearly identical iodineproduction and consumption rates.The rates are plotted in Figure 3-11. The CAE d.c. pos and neg rates are nearlylinear and closely parallel the iodine rates. The exception in both cases is PF-K-7 (initial[iodate] = 0.026 M) which shows dramatically increased rates. This was also true forPF-G-5. This confirms the earlier suggestion that an [iodate] of at least 0.025 M isoptimum for maximum iodine production and consumption rates.12- -40o [iodine] pos• [iodine]nego C4Epos - 308- • C4Eneg0 0-204- o-10E 0 0c) CC 0: ::2O—12- I I I I I I I I I I I I I I —-400.015 0.020 0.025 0.030 0.035iriti [icxate] (M)Figure 3-11. Reaction rates for different initial iodate concentrations.Chapter 3 lodate Studies Using Stopped Flow Methods - 103 -Table 3-12. Reaction rates for Pos and Neg events.Series initial CAE CAE [iodine] [iodine][iodate] Pos Neg Pos Neg____________(x103mV.s1) (x103mV.s1) (x102Ms1) (x102MsjPF-K-5 0.019M 2.45±0.19 -2.51±0.19 4.11±0.35 -3.32±0.32PF-K-6 0.023 M 3.41±0.18 -3.60±0.15 8.13±0.47 -7.64±0.52PF-K-7 0.026 M 6.97±0.24 -7.04±0.34 12.7±0.7 -12.8±0.5PF-K-8 0.030 M 6.84±0.41 -7.05±0.40 20.3±1.3 -15.3±1.0PF-K-9 0.034M 7.57±0.17 -7.26±0.27 23.8±1.3 -15.4±1.2PF-K-10 0.037 M 2.64±0.29 -2.95±0.34 5.71±0.33 -5.92±0.313.4.4.3 Extent of ReactionFigure 3-12 shows the integrated CAE d.c. curves (top) and the integrated CAE attime = 480 seconds as a function of initial iodate concentration (bottom).The plots in the upper graph appear to show essentially linear (zero order)behaviour. In fact, the data in the plots 5 through 9 fit better to a first order function,although the scatter in the data prevents an unequivocal conclusion.Recall that for series PF-G, the rate peaked at [iodate] = 0.025 M (PF-G-5). Inthis series, the rate peaks dramatically at 0.026 M before decreasing. This furtherstrengthens the hypothesis that an initial [iodate] of ca. 0.025 M is optimum for oxygenevolution.Chapter 3 lodate Studies Using Stopped Flow Methods - 104-The lower graph in Figure 3-12 depicts the integrated CAE d.c. traces at time480 seconds vs. initial [iodate]. The first four points (PF-K-1 through 4) exhibit linearbehaviour. The next five points (PF-K-5 through 9) fit nicely to a second order functionof the general stoichiometry:Eqn3-2. A + 2B —+ ZThe differential form of the rate law is:Eqn 3-3. = 2k(a0 — x) (a0 — 2x)dtwhere k is the rate constant, a0 is the initial quantity and x the quantity at time t. In termsof CAE, a0 would be the background noise and x the cumulative CAE response. Note thatthe rate law would also be valid for:Eqn3-4. A+B—>Zprovided that the initial concentrations of the two species were the same.Recall from the discussion of series PF-G, that both of the main oxygen producingreactions in the BR oscillator are second order:02. 2H00 — H20 +Dl. HOT + H20 + 0 + H + H20In series PF-G, the integrated CAE d.c. decreased as initial [iodate] increased. Forthis run, we see the integrated CAE d.c. exhibit a linear behaviour until initial [iodate]increases beyond 0.015 M. At that point, the integrated CAE d.c. increases dramatically.Plotting the natural logarithm of the last five data points against time failed to produce aChapter 3 lodate Studies Using Stopped Flow Methods - 105 -straight line. A test for a second order function also failed, indicating there is no simplerate law to describe the data.54>LUC-)ci)(0a)C02Time (s)6005-——Cl)IDIL 4-C)w<2-()U)ci)C0--——0.000I I I0.005 0.010 0.015 0.020 0.025initial [iodate] (M)0.030 0.035Figure 3-12. Top: integrated CAE d.c. signal vs. reaction time. The numbers beside eachcurve identify the experimental run. Bottom: Integrated CAE d.c. (at time = 480 s.) vs.0 100 200 300 400 50000 98070601o20504initial iodate concentration.lodate Studies Using Stopped Flow Methodsj Chapter3 -106-I3.4.5 Series PF-K: ConclusionsThe CAE d.c. data for PF-K-1 through 4 exhibited some periodic activity eventhough the UV-Vis data indicated no reaction taking place. Autocorrelograms of the datadetermined that periodicity was present with a time constant of 18 seconds. This periodlengthened to 24 seconds as the [iodate] increased to 0.011 M. It was hypothesized theactivity was due to the slow decomposition of hydrogen peroxide, according to thedissolved oxygen model proposed by Wentzell. The long-period oscillations would be theresult of the gradual build-up of dissolved oxygen followed by its release as bubblenucleation. Consider the following postulated cycle:1. 02 produced by the catalytic decomposition ofH20 is dissolved into solution. Thesaturation ratio a continues to increase (a is defined as the ratio of the currentconcentration to the equilibrium concentration of dissolved gas).2. When the solution becomes supersaturated (a reaches a critical level), there is aspontaneous formation of bubbles accompanied by a decrease in free energy. Bubblesmay form in the bulk of the solution and/or on the surface of the reaction vessel. Thelatter is usually easier (ie. occurs at lower values of a) than the former mainly due totwo main parameters: contact angle and nucleation site geometry as stated in theIntroduction.3. The bubbles then rise to the liquid/gas surface and burst where they are detected byCAE as the rise to a peak.4. When enough oxygen has been released (a decreases past its critical level), theevolved 02 once again dissolves into solution and the CAE peak declines.Chapter 3 lodate Studies Using Stopped Flow Methods - 107 -Examination of the CAE d.c. peak data revealed that there was no direct, simplecorrelation with either iodine or iodide ion.The PF-K reaction rate data for pos and neg phases agreed with that observed forseries PF-G. The rates for the two phases were nearly identical when [iodatej approached0.025 M, for all the analytical methods employed and both series posted the largestoverall CAE d.c. rate for this concentration.The integrated CAE d.c. data demonstrated a shallow linear trend until the initial[iodatej increased beyond 0.015 M at which time the slope increased dramatically.There exists a relationship between initial [iodate] and integrated CAE whichappears to be quantitatively useful for [iodate] > 0.020 M.Chapter 3 lodate Studies Using Stopped Flow Methods - 108 -3.5 Series PF-O. lodate varied from 0.004 to 0.037 M. 302CThis series is the second iodate series. It uses the same experimental procedures asPF-K, the only difference being a five degree rise in temperature. As a reminder, theinstrumental parameters are present in the following table.Table 3-13. Instrument parameters for series PF-O.Instrument Samp. Period Other settingsCAE d.c. 3.0 s 64 dB gainscope mm. 1.5 s 1600 mV input range, 2.5 MHz digitization rateUV-Vis 3.0 s 300 to 600 nm; 2 nm resolution3.5.1 Data Work-upThe raw data collected from the various instruments required some preprocessingbefore an analysis could be conducted. The work-up is identical to that described for theprevious series.Chapter 3 lodate Studies Using Stopped Flow Methods - 109-3.5.2 PF-O-1 and 2The data from PF-O-1 and 2 are presented in Figure C-15. PF-O-l (initial [iodate]= 0.004 M) shows no UV-Vis or acoustic activity.The second run, PF-O-2 (initial [iodatej = 0.008 M), is mostly linear save for theoccasional peak likely due to bubbles in the light path. The CAE trace shows a little moreactivity than for PF-O- 1. Referring back to series PF-K, approximately the same numberof peaks are present, but the peaks are larger in amplitude suggesting that temperature hasa positive effect on the catalytic decomposition of hydrogen peroxide.3.5.3 PF-O-3 and 4In PF-O-3 (initial [iodate] 0.011 M) and PF-O-4 (initial [iodate] = 0.015 M),show peaks of an irregular shape and distribution indicating extensive bubble evolution,rather than iodine absorption.3.5.3.1 AutocorrelationIn series PF-K, there was evidence for some periodicity in the CAE d.c. data forthe first four experimental runs. The results were an initial periodicity of 18 seconds,which lengthened to 24 seconds before commencement of oscillations.The autocorrelograms for PF-O-1 through 4 are in Figure 3-13. Each graph showsan initial periodicity of ca. 5t, or 20 seconds. This compares well with series PF-K,Chapter 3 lodate Studies Using Stopped Flow Methods - 110-i!-0.4 - I I I I I I I I I I I I I I0.4 I\ccND_cx...,aA/2N.R\h-0.2----0.4 - I I I I I I I I I I I I I I I0 5 10 15 200.4 - I I I I I I I I I I I I I I I I--0.4 - I I I I I I I I I I I I I I0 5 10 15 200.2 - I I I I I I I I I I I IP F -0-4-0.2 - I I I I I I I I I I I I I I I0 5 10 15 20tFigure 3-13. Autocorrelograms of CAE d.c. data for series PF-O-1 through4. The time constant (T) is four seconds.which reported initial periodicities of Ca. 18 seconds which lengthened to 24 seconds. Thesame discussion as to the processes involved should apply here.Chapter 3 lodate Studies Using Stopped Flow Methods - 111 -3.5.4 PF-O-5 through 10The initial iodate concentrations studied are 0-5 (0.019 M), 0-6 (0.023 M), 0-7(0.026 M), 0-8 (0.030 M), 0-9 (0.034 M), and 0-10 (0.037 M).3.5.4.1 Peak Count and FrequencyThe data from runs 5 through 10 are shown in the C-17 through C-19. The UVVis trace for PF-0-5 shows irregular peaks, which are likely due to a combination of theinitial stages of the oscillator and bubbles in the light path. The remaining UV-Vis tracestake on a skewed sawtooth appearance implying that the iodine production rates are fasterthan the iodine consumption rates. Several trends are more easily seen when peak data arecompiled into Table 3-14.Table 3-14. Peak parameters for series PF-0.Series CAE DC CAE DC UV-Vis UV-VisNumber Number of Number Number ofof peaks peaks/mm. of peaks peaks/mm.PF-0-5 3 0.60 14 2.80PF-0-6 5 1.00 19 3.80PF-0-7 4 0.80 17 3.40PF-0-8 9 1.80 14 2.80PF-0-9 10 2.00 12 2.40PF-0-10 9 1.80 15 3.00Chapter 3 lodate Studies Using Stopped Flow Methods - 112-Table 3-15. Reaction rates for Pos and Neg events.Series initial CAE CAE [iodine] [iodine][iodate] Pos Neg Pos Neg(x103 mV.s’) (x103 mV.s’) (x102M.s’) (x102M.s’)PF-O-5 0.019 M 2.93±0.44 -2.58±0.34 26.6±1.4 -24.6±1.1PF-O-6 0.023 M 3.29±0.39 -2.85±0.35 27.6±0.9 -33.2±1.3PF-O-7 0.026 M 4.13±0.37 -4.09±0.28 39.6±2.9 -35.9±2.7PF-O-8 0.030 M 4.06±0.39 -3.95±0.35 51.4±3.6 -41.7±2.7PF-O-9 0.034 M 12.3±0.9 -14.7±0.2 56.6±3.9 -42.4±2.0PF-O-10 0.037 M 2.57±0.27 -2.67±0.24 102±4.7 -59.4±3.8The average number of CAE d.c. peaks, 7±3, has actually decreased somewhatover the series PF-K (10±2), although the large error present in the first value makesconclusions difficult. The average number of peaks for the UV-Vis data, 15±2, remainsthe same as for the previous series (15±2), implying that the temperature increase has noeffect on the number of peaks observed.3.5.4.2 Peak Reaction RatesThe reaction rates for thepos and neg phases of the CAE d.c. and UV-Vis data arein Table 3-15. To assess the impact of the five degree increase in temperature, the rates inPF-K were compared to those presented here. The CAE d.c. rates increased as often asChapter 3 lodate Studies Using Stopped Flow Methods - 113-they decreased. i.e. an effect of temperature on rates was not observed. This suggests thatunder these conditions the rate at which dissolved oxygen is released from solution isalready at its maximum.Table 3-16. Iodine consumption!production ratios.Run initial PF-K PF-O[iodate] (consumption! (consumption!production) production)5 0.019M 0.808 1.206 0.023 M 0.940 0.9077 0.026M 0.992 0.8118 0.030 M 0.754 0.7499 0.034 M 0.647 0.58210 0.037M 1.04For the UV-Vis data, the iodine production rates increased by an average of2.9±0.5 times, while the consumption rates increased an average of 3.2±0.8 times.Between the series, however, the iodine production rate quickly outstrips theconsumption rate. For example, in PF-O-9 (Table 3-16) the consumption rate has fallen toonly one-half the production rate, suggesting that temperature increases favour iodineproduction over consumption.The data clearly show that at the increased temperature, there is an initialemphasis on consumption of iodine in series PF-O. However, the production andChapter 3 lodate Studies Using Stopped Flow Methods - 114 -‘U)>Ec)0x4-,w()t3‘U)C’,0x4-ci)0t5consumption rates come closest together at run 7 in PF-K (0.992) and run 6 in PF-O(0.907). After that point, the iodine production rate for the higher temperature seriesnearly doubles the consumption rate by the end of the experiment.15 -[iodine] pos• [iodine] neg12- 0CAE pos• CAEneg9-06- 00o 03- 00--3- 1 •. I •• •-6 -.-9 --12 --15---÷-- ••‘‘‘ I , I0.015- 120- 100- 80- 60- 40- 20-0- -20- -40- -60- -80- -100-1—- -1200.0400.020 0.025 0.030 0.035initial [iodate] (M)Figure 3-14. Reaction rates for different initial iodate concentrations.3.5.4.3 Extent of ReactionThe integrated CAE d.c. data are plotted separately for each run in Figure 3-15(runs 1 to 6) and in Figure 3-16 (runs 7 to 10). The integrated CAE d.c. appears to bemore sensitive to transitions in this series than to the previous series. In the first four runswhere acoustic activity is due mostly to the catalytic decomposition of hydrogenChapter 3 lodate Studies Using Stopped Flow Methods - 115 -peroxide, the transitions appear to be random, with the exception of the transition at ca.200 seconds. This suggests an acoustic event has occurred which is independent of theconcentration of iodate. The differences in the slope of the line before and after thetransition point further suggests that either another acoustically active process has begun,in addition to the previous process, or replacing the previous process. The former appearsmore likely. Chemically, the initial process is the catalytic decomposition of hydrogenperoxide. The second process is the termination of the free radical stage of iodineproduction.Recall that iodine production in the BR oscillator is divided into stages: freeradical production of HOlO and non-radical consumption of HOlO to ultimately formiodine. The first stage produces oxygen and only switches to the second stage whensufficient HOlO has been generated. This is the proposed sequence of events:1. induction phase: iodate is consumed by reaction 13 to form intermediates HOlO andHOl. Oxygen evolution is primarily due to the catalytic decomposition of hydrogenperoxide.2. iodine production (free radical): when sufficient HOIO has been produced, it isconsumed by reaction 15 and a rapid chain reaction begins in the sequence 15, Ml,M2, 02. The chain reaction continues for ca. 20 seconds before terminating. In thattime, there is a large amount of oxygen evolved which appears on the integrated CAEd.c. trace as a sharp rise at ca. 200 seconds.3. return to normal: step 2 was unable to produce sufficient HOlO to enable the switchto iodine production (reactions 12, Dl and I 1). This is likely due to insufficient iodate.Chapter 3 lodate Studies Using Stopped Flow Methods - 116 -0 100 200 300 400 500 0 100 200 300 400 5000.80.60.40.2uJ1.21.00.8E0.6Ci)c.: 0.40.20.02.55 2.0E1.51.0LUO50.01.00.8 5E0.60.40-oO.2uj0.002.01.5E1.0U)C)0.5 -ow0.0 02.52.0 5E1.51.10.5 w0.0 00 100 200 300 400 500 0 100 200 300 400 5000 100 200 300 400 500 0 100 200 300 400 500Time (s) Time (s)Figure 3-15. Integrated CAE d.c. data fort = 480 seconds, runs ito 6.Chapter 3 lodate Studies Using Stopped Flow Methods - 117->E2DU)C.)0wC-)22DU)C.)I-I-I00 100 200 300 400 500 0 100 200 300 400 5000 100 200 300 400 500Time (s)2.5 3.02.0 2.511.5 2.01.5C.) 1.01.00.5 0.50.0 0.020 3.02.5152.010 1.5C1.0LI.150.50 0.00 100 200 300 400 500Time (s)Figure 3-16. Integrated CAE d.c. data for t = 480 seconds, runs 7 to 10.The fifth run, PF-O-5, is itself a transition experiment. Semi-regular peaks arepresent in the raw CAE d.c. data suggesting the BR oscillator is active. Sharp transitionsappear at ca. 400 seconds which the UV-Vis data indicates is the end of the oscillationperiod. The gas evolution at this time cannot be due to oscillations, but another process.Possibly the catalytic decomposition of hydrogen peroxide is again the main oxygengenerator, this time unhindered by competing reactions. Assuming the limiting reagent isiodate, we again have the situation where there is insufficient iodate to produce HOlOand hence oscillations cease. However, there is still sufficient hydrogen peroxide andaqueous iodine to evolve oxygen.Chapter 3 lodate Studies Using Stopped Flow Methods -- 118-IThe integrated CAE data for runs PF-O-6 through 10 generally all show atermination transition at Ca. 400 seconds. Many of the remaining transitions can beattributed to features in the UV-Vis traces of the individual runs. The following table liststhe transitions and interpretations for PF-O-6 through 8. Run PF-O-9 is more difficult tointerpret as the integrated trace is an order of magnitude larger than previous runs. Thishas the effect of masking out the key transitions.Table 3-17. Interpretation of transitions in integrated CAE d.c. data.Series Transition Interpretation based on correlation with UV-Vis data.Point_(s)PF-O-6 70-100 switch from induction phase to oscillationsPF-O-6 3 90-400 switch from oscillations to terminationPF-O-7 60-80 switch from induction phase to oscillationsPF-O-7 180-200 maximum amplitude oscillationsPF-O-7 270-280 iodine production rate accelerates more quickly thanconsumption ratePF-O-7 3 80-400 switch from oscillations to terminationPF-O-7 460-480 not assignedPF-O-8 40-50 switch from induction to irregular oscillationsPF-O-8 100-110 switch to regular oscillationsPF-O-8 180-200 maximum amplitude oscillationsPF-O-8 400-420 switch from oscillations to terminationChapter 3 lodate Studies Using Stopped Flow Methods - 119-25 -___________________________________________c20-°15- °cj 10 -LUC). 5) o 00-° ° 0I I I I I I0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040initial [iodate] (M)Figure 3-17. Integrated CAE d.c. data at time = 480 s.The above figure illustrates the integrated CAE d.c. data at time = 480 seconds.Similar to the previous series, there is a generally linear trend followed by a sharpincrease. In this series the linear trend extends to 0.030 M iodate, whereas in the previousseries, the linear trend extended to 0.0020 M iodate, the difference due to the five degreeincrease in temperature.Chapter 3 lodate Studies Using Stopped Flow Methods - 120 -3.5.5 Series PF-O: ConclusionsThere are significant temperature effects on all phases of the oscillator: peakamplitudes, pos and neg rates, as well as overall reaction rates.The integrated CAE d.c. data demonstrated distinct transition events which werenot initially seen in series PF-K. This may be due in part to both the increasedtemperature and a higher sampling frequency. In hindsight, some transitions can be seenin the previous series.The transitions observed often heralded events present in the UV-Vis data,indicating that the CAE d.c. data is almost, but not quite, in phase with the UV-Vis dataas we saw in the discussion of series PF-K. This has important consequences as we wereunable to establish a phase relationship in this data series. Even a brief inspection of thedata clearly suggests that no simple relationship is present. Yet this is a misleadingconclusion, as increased gas evolution may be creating “macro” events (i.e. many bubblesbursting nearly simultaneously) which are displayed as one single broad peak. There maybe no way of easily determining if several peaks are present where one is seen. Onepossibility is to increase the sampling rate, but the 200 ms time constant of the amplifier’sd.c. output limits our temporal resolution.Chapter 3 lodate Studies Using Stopped Flow Methods - 121-3.6 Series PF-T. lodate varied from 0.004 to 0.037 M. 352CThis is the third and last of the iodate series. It uses the same basic experimentalprocedures as PF-O with two changes:1. another five degree rise in temperature.2. sampling period decreased from 3.0 seconds to 1.0 secondsIt was apparent from the previous series that the resolution of both the UV-Visand CAE d.c. data needed improving. This would be even more true in an experimentwith elevated temperatures as increased reaction rates would give rise to events whichmay be missed by existing sampling rates.However, the control computers were already functioning near their limits andsignificant effort was required to re-write the control software to increase the system’soverall performance.Table 3-18. Instrument parameters for series PF-T.Instrument Sampling Other settingsPeriodCAE d.c. 1.0 s 64 dB gainscope mm. 1.5 s 1600 mV input range, 2.5 MHz digitization rateUV-Vis 1.0 s 300 to 600 nm; 2 nm resolutionChapter 3 lodate Studies Using Stopped Flow Methods - 122 -3.6.1 Data Work-upThe raw data collected from the various instruments required some preprocessingbefore an analysis could be conducted. The work-up is identical to that described for theprevious series.3.6.2 PF-T-1 and 2The data from PF-T-1 and 2 are presented in Figure C-20. PF-T-1(initial [iodate] = 0.004 M) show no UV-Vis or acoustic activity.PF-T-2 (initial [iodate] = 0.008 M), clearly shows the noise due to bubbles in thelight path. The CAE trace shows a little more activity than for PF-T- 1.3.6.3 PF-T-3 and 4In PF-T-3 (initial [iodate] = 0.011 M) and PF-T-4 (initial [iodatej = 0.0 15 M),show peaks of an irregular shape and distribution indicating extensive bubble evolution,rather than iodine absorption.Chapter 3 lodate Studies Using Stopped Flow Methods - 123 -3.6.3.1 AutocorrelationThe autocorrelograms of series PF-T- 1 through 4 (Figure 3-18) present nosurprises. The time constants of the periodicities generally increase as iodateconcentration increases. The periodicity of PF-T- 1 is unclear. PF-T-2 shows strongperiodicity with a time constant of ca. 16 seconds. PF-T-3 exhibits less forcefulperiodicity of ca. 18 seconds while PF-T-4 presents a tenuous periodicity of ca. 27seconds. Series PF-T-5 through 10 show no significant periodicity and so are notillustrated.3.6.4 PF-T-5 through 10The initial iodate concentrations studied are T-5 (0.0 19 M), T-6 (0.023 M), T-7(0.026 M), T-8 (0.030 M), T-9 (0.034 M), and T-10 (0.037 M).Chapter 3 lodate Studies Using Stopped Flow Methods - 124-PP0.40.20.0-0.2-0.40.40.20.0-0.2-0.40.20.10.0-0.1-0.20.40.20.0-0.2-0.40 10 20 30 40 50 6060600 10 20 30 40 50 60tFigure 3-18. Autocorrelograms of CAE d.c. data for series PF-T-1I I IPF-T-10 10 20 30 40 500 10 20 30 40 50I r I I r I I II I I IPF-T-4- IIthrough 4. The time constant (t) is one second.Chapter 3 lodate Studies Using Stopped Flow Methods - 125-3.6.4.1 Peak Count and FrequencyThe data from runs 5 through 10 are shown in the following figures. The UV-Vistrace for PF-T-5 confirms our earlier suspicions that run 5 is a transition run - a borderbetween non-osc. and osc. conditions - showing some signs of oscillations, but notforming the sawtooth peaks associated with regular oscillations. The remaining UV-Vistraces take on a skewed sawtooth appearance implying that the iodine production rates arefaster than the iodine consumption rates. Several trends are more easily seen inTable 3-19. The average number of CAE d.c. peaks, 11±1, is representative of theprevious series (10±2 and 7±3). The average number of UV-Vis peaks, 15±2, is alsosimilar to the previous series (15±2 and 15±2). These results provide strong evidence thatthe number of peaks observed in the time period from 100 to 400 seconds is independentof temperature, implying that the mechanistic steps responsible for switching theoscillator between its two steady states are also independent of temperature under theseconditions. These steps are discussed further in Chapter Five.Chapter 3 lodate Studies Using Stopped Flow Methods - 126-Table 3-19. Peak parameters for Series PF-T.Series initial CAE d.c. CAE d.c. UV-Vis UV-Vis[iodate] Number Number of Number Number ofof peaks peaks/mm. of peaks peaks/mm.PF-T-5 0.019M 10 2.00 -- --PF-T-6 0.023 M 12 3.20 16 3.20PF-T-7 0.026 M 11 2.20 15 3.00PF-T-8 0.030M 11 2.20 15 3.00PF-T-9 0.034 M 12 2.40 12 2.40PF-T-10 0.037M 7 1.40 11 2.20Table 3-20. Reaction rates for Pos and Neg events.Series initial CAE d.c. CAE d.c. [iodine] [iodine][iodate] Pos Neg Pos Neg(x103 mVs’) (x103mV.s1) (x102M.s’) (x102M.s1)PF-T-6 0.023 M 3.31±0.19 -3.15±0.20 86.4+6.5 -67.1±2.0PF-T-7 0.026 M 4.48±0.46 -4.60±0.38 114±5 -78.5±3.4PF-T-8 0.030 M 3.98±0.29 -4.37±0.22 154±2 -93.4±5.8PF-T-9 0.034 M 6.18±0.38 -5.60±0.36 194±8 -103±7PF-T-10 0.037M 4.57±0.36 -4.36±0.41 296±16 -181±9Chapter 3 lodate Studies Using Stopped Flow Methods - 127 -The trends observed in the above table’s data generally reflect the trends observedin the previous two series:1. CAE d.c. maximum number of peaks: the previous two series (PF-K and -0) showthe maximum number to be at [iodate] 0.034 M. In this series, the maximum occursat both [iodate] = 0.023 M and 0.034 M. This was the expected result, as increasingthe iodate concentration should increase the number of observed peaks. Except forthis one result, however, inspection of the data in the table shows an almost randomnumber of peaks occurring at any particular concentration. We attribute these“random” numbers to the presence of “macro” events which were discussed above.To briefly recap, macro events occur when a large number of gas bubbles burst withinthe sampling period. This results in a single broad peak instead of several distinctpeaks. Thus in any particular experimental run, it is difficult to estimate the truenumber of peaks.2. UV-Vis maximum number of peaks: the first series (PF-K) shows the maximumnumber to be at [iodate] 0.037 M. The next two series place the maximum at{iodate] = 0.023 M. This trend is misleading as we are only counting the number ofpeaks which occur between 100 and 400 seconds. Closer inspection of the raw datareveals the fact that the induction period decreases as iodate concentration increases.This observation has been made earlier in this work, but the data resolution had beeninsufficient to accurately determine when oscillations began. Taking into account allthe observed oscillations, the maximum number could be placed at[iodate] = 0.03 7 M.In the next table, we examine the reaction rates of the pos and neg events in theCAE d.c. and UV-Vis data.Chapter 3 lodate Studies Using Stopped Flow Methods - 128-3.6.4.2 Peak Reaction RatesThe reaction rates for thepos and neg phases of the CAE d.c. and UV-Vis data arein Table 3-20. There is a large temperature effect which will be discussed in latersections. We note here only the common trends between the three experimental series:1. Close correlation between pos and neg rates for CAE d.c. data in this series. Unlikethe iodine data, the CAE d.c. data have a high degree of correlation between the twoevent rates at all the concentrations studied. This observation implies that the posevent corresponds to the production of a species while the neg event corresponds tothe non-production of a species (i.e. 02), rather than consumption. Hence, a CAE d.c.peak should have the same general shape regardless of the exact reaction rates. Forthe UV-Vis, the peak shapes will vary as the neg event in this case corresponds toiodine consumption, and the rate of iodine consumption varies with iodateconcentration.2. CAE d.c. maximum rates occur at [iodate] = 0.034 M, regardless of temperature. Thisis the expected trend that rates should increase as concentration increases. Thisobservation further indicates that the “macro” events discussed above produce peakswith slopes which are still representative of the acoustic processes. As an aside, itshould be noted that a “local maxima” of reaction rates occurs at [iodate] = 0.026 Min each series at this temperature.3. The iodine consumption/production ratios generally decrease as iodate concentrationincreases (Table 3-21) and as temperature increases. Under maximum conditions(initial [iodate] = 0.037 M, temperature = 35°C) the consumption rate shrank to ca.one-half the production rate.Chapter 3 lodate Studies Using Stopped Flow Methods - 129-Table 3-21. Iodine consumption/production ratios.Run initial PF-K PF-O PF-T[iodate] (consumption! (consumption! (consumption!production) production) production)6 0.023 M 0.940 1.20 0.7777 0.026 M 0.992 0.907 0.6898 0.030M 0.754 0.811 0.6069 0.034 M 0.647 0.749 0.53 110 0.037M 1.04 0.582 0.61110 -_____- 2508- -2006- ci o -1504 - o- 100CI) 02- -50ci 0x 0- -0•-2- --50•-- : • - -ioo-6 - - -150-8- CAE pos D [iodine] pos- -200• CAE neg [iodine] neg—10— I I I I I I I I I I I I ——2500.015 0.020 0.025 0.030 0.035initial [iodate] (M)Figure 3-19. Reaction rates for different initial iodate concentrations.Chapter 3 lodate Studies Using Stopped Flow Methods - 130-EEU)C)w3.6.4.3 Extent of ReactionThe integrated CAE d.c. data are plotted separately for the strongly oscillatingruns (6 through 9) in Figure 3-20. The integrated CAE d.c. data generally show a14121086420016100 200 300 400125008EEDU)-oIJJC-) 400 100 200 300 400 500Time (s)Figure 3-20. Integrated CAE d.c. data for PF-T-6 through 9.Chapter 3 lodate Studies Using Stopped Flow Methods - 131 -Table 3-22. Interpretation of transitions in integrated CAE d.c. data.Series Transition Interpretation based on correlation with UV-Vis data.Point_(s)PF-T-6 140-150 switch from induction phase to oscillationsPF-T-6 350-3 60 switch from oscillations to terminationPF-T-7 140-150 switch from induction phase to oscillationsPF-T-7 190-200 maximum amplitude oscillationsPF-T-7 280-290 iodine production rate accelerates more quickly thanconsumption ratePF-T-7 370-3 80 switch from oscillations to terminationPF-T-8 80-90 switch from induction phase to oscillationsPF-T-8 280-290 iodine production rate accelerates more quickly thanconsumption ratePF-T-9 80-170 see discussionPF-T-9 320-330 switch from oscillations to terminationtermination transition at ca. 400 seconds. Many of the remaining transitions can beattributed to features in the UV-Vis traces of the individual runs. Table 3-22 lists thetransitions and interpretations for PF-T-6 through 9.The transition times given in the table generally precede the observed event by ca.20 seconds, which is the time constant seen in the autocorrelogram results. The PF-T-9transition at time = 80 seconds is difficult to interpret. Rather than a sharp transitionChapter 3 lodate Studies Using Stopped Flow Methods - 132 -point, a long, slight rise is observed which lasts ca. 90 seconds. One explanation may bethat the oxygen bubbles are bursting at their maximum rate during this time period. Thismay be a limit of the experimental system (i.e. cell geometry) rather than a chemicaleffect.188-6-42-—0.000Figure 3-21. Integrated CAE d.c. data at time = 480 s.The above figure illustrates the integrated CAE d.c. data at time = 480 seconds.The data presented is the culmination of a trend observed in previous series where thelinear portion of the curve was increasing at the expense of the non-linear portion. In thisseries, the linear portion (slope = 433 mV.s’, r2 = 0.974) dominates to the exclusion ofthe non-linear portion. That is, temperature is speeding up the overall oxygen evolution ofthe system to the extent that oxygen is forming bubbles at its maximum rate at the highest0cu 16-14-212-C’)>210 -w0a)Cua)C000PF-TI I I I0.005 0.010 0.015 0.020 0.025initial [iodate] (M)0.030 0.035Chapter 3 lodate Studies Using Stopped Flow Methods - 133 -temperature, 3 5°C. Note that this trend has also been observed for the rates of the pos andneg events.Chapter 3 lodate Studies Using Stopped Flow Methods - 1343.6.5 Series PF-T: ConclusionsThe increased sampling rate provided better resolution of the UV-Vis data, whichsubstantially aided the data analysis and interpretation. The single most importantdiscovery was that an initial [iodate] = 0.0 19 M was not sufficient to produce iodineoscillations in some cases. We conclude that this is a borderline concentration and anydata collected may reveal oscillatory information as often as not.Autocorrelograms of non-oscillatory data revealed a periodicity which increasedas iodate concentration increased. The source of oscillations is believed to be from thebuild-up and release of dissolved oxygen as discussed for the previous series. The timeconstant at [iodatej = 0.008 M was ca. 16 seconds, at 0.011 M was Ca. 18 seconds and at0.015 M was ca. 27 seconds. Autocorrelograms of oscillatory data showed no detectableperiodicity.The maximum number of CAE d.c. peaks observed occurred at {iodate]0.034 M, again reflecting the difficulty of obtaining reliable data for [iodate] = 0.03 7 M.This appears to be a valid observation as all three experimental series corroborate thisobservation.The CAE d.c. data showed close correlation of the absolute values of the pos andneg rates regardless of [iodate]. This is unique from the UV-Vis data which shows the pos(iodine production) rate increases while the neg (iodine consumption) rate decreases as[iodate] increases. The conclusion is that the CAE d.c. pos rate is related to production ofa species while the neg rate is related to non-production, not consumption of a species.This is further evidence that gas evolution is the dominant process responsible for CAE inthe BR oscillator.Chapter 3 lodate Studies Using Stopped Flow Methods - 135 -The integrated CAE d.c. data showed similar transitions as the previous series(PF-O). The integrated CAE d.c. at the end of the experiments (time = 480 s.) exhibited astrictly linear trend, unlike the previous series which demonstrated a linear trend followedby a sharp second order rise.Again, the quality of these integrated CAE d.c. measurements suggest analyticalpotential.Chapter 4 Hydrogen Peroxide Study - 136 -Chapter Four - Hydrogen Peroxide Study“How come I get all the hard questions?”-O.North4.1 IntroductionOne of the conclusions from the series in the previous chapter is that hydrogenperoxide plays an important role in the production of oxygen in the Briggs-Rauscheroscillator. The aim of this experimental series is to use the “high resolution” systempresented in the last chapter to monitor the effect of varying the concentration ofH20.The experimental conditions are the same as detailed in the previous series withthe exception thatH20 is varied instead of iodate.Chapter 4 Hydrogen Peroxide Study - 137 -4.2 Series PF-S, Hydrogen Peroxide varied from 0.078 to 0.78 M, 35C4.2.1 Data Work-upThe raw data collected from the various instruments required some preprocessingbefore an analysis could be conducted. The work-up is identical to that described for theprevious series.4.2.2 PF-S-1 and 2PF-S-1, initial [H20] 0.078 M, exhibits a data trace (Figure C-24) similar toPF-T-5 of the previous series. The most noticeable difference is that both the UV-Vis andthe CAE peaks begin the experiment with relatively large amplitudes and finish theexperiment with relatively small amplitudes. The data suggests that rapid catalyticdecomposition ofH20 is occurring, consuming theH20until it is depleted.The UV-Vis trace in PF-S-2 (initial [H20] = 0.16 M) shows large, broadlyskewed peaks, indicating high iodine production rates and low iodine consumption rates.Interestingly, the CAE peaks have all but disappeared, leading to the conclusion that theTable 4-1. Instrument parameters for series PF-S.Instrument Samp. Period Other settingsCAE d.c. 1.0 s 64 dB gainscope mm. 1.5 s 1600 mV input range, 2.5 MHz digitization rateUV-Vis 1.0 s 300 to 600 nm; 2 nm resolutionChapter 4 Hydrogen Peroxide Study - 138 -non-oscillatory process responsible for CAE in PF-S-1 has been replaced. This newprocess, part of the oscillator, favours iodine production over iodine consumption andalso over oxygen production.4.2.3 PF-S-3 and 4The UV-Vis data of PF-S-3 (initial [11202] = 0.24 M) and PF-T-4 (initial [H20]=0.31 M), show the beginnings of a trend. As initial [11202] increases, the broadly skewedpeaks are being replaced by the needle-shaped peaks seen in previous series. Moreimportantly, the needle-shaped peaks occur at the beginning of the experiment, when[H20] is greatest. The broadly skewed peaks occur towards the end of the experimentwhen [H201is lowest. By analogy, we can conclude that the broad peaks observed at theend of the experiments in the previous series are the result of low [H20]rather than low[iodate].4.2.3.1 AutocorrelationThe autocorrelograms for series PF-S-l through 4 are in Figure 4-1. PF-S-1 showsthe beginnings of periodicity, indicating this is a “transition” series similar to the PF-x-5series seen earlier. PF-S-2 has distinct periodicity with a time constant equal to 16t, or 32seconds. PF-S-3 also exhibits periodicity with a time constant of 40 seconds. PF-S-4 andthe following series, in which the broadly skewed peaks are replaced by needle-shapedpeaks, show no distinct periodicity. This suggests that the periodicity is tied to the non-Chapter 4 Hydrogen Peroxide Study - 139-oscillatory processes which generate CAE at low [11202], and these processes are notfavoured as [H201increases.Chapter 4 Hydrogen Peroxide Study - 140-P0.60.5 -0.4 - -0.3 -0.10.0-0.10.20.00.2t606060Figure 4-1. Autocorrelograms of CAE d.c. data for series PF-S- 1 through 4. The timeconstant (t) is two seconds.0.2 — I I I I I I I I I I I0 10 20 300.30.240 50 6000.60.410 20 30 40 50-0.20 100.60.420 30 40 500.00 10 20 30 40 50Chapter 4 Hydrogen Peroxide Study - 141 -4.2.4 PF-S-5 through 10The initial [H20] studied is S-5 (0.39 M), S-6 (0.47 M), S-7 (0.55 M), S-8(0.63 M), S-9 (0.71 M), and S-1O (0.78 M).4.2.4.1 Peak Count and FrequencyThe data from runs 5 through 10 are shown in the Figures C-26 through C-28. TheUV-Vis traces clearly indicate that as initial [H20] increases, the iodine production andconsumption rates increase producing the familiar “needle-shaped” peaks.Table 4-2. Peak parameters for Series PF-S.Series initial CAE d.c. UV-Vis[H20] Number Numberof peaks of peaksPF-S-2 0.16M 9 6PF-S-3 0.24 M 10 7PF-S-4 0.31 M 9 9PF-S-5 0.39 M 10 8PF-S-6 0.47 M 12 11PF-S-7 0.55M 12 11PF-S-8 0.63 M 9 12PF-S-9 0.71 M 10 12PF-S-10 0.78M 10 11Chapter 4 Hydrogen Peroxide Study - 142-The CAE d.c. peak data shows no correlation between the number of peaksobserved and initial [H20], with the average number of peaks between 100 and 400seconds to be 10±1. In contrast, the UV-Vis data shows the number of peaks increasingas initial [H20] increases. The average number of peaks is 11±2. Therefore, the UV-Visis not detecting just bubbles. The acoustically active process must be independent of[H20].Table 4-3. Reaction rates for Pos and Neg events.Series initial CAE d.c. CAE d.c. [iodine] [iodine][H20] Pos Neg Pos Neg(x103mV.s1) (x103mV.s’) (x102M.s’) (x102M.sjPF-S-2 0.16 M 0.93±0.04 -0.89±0.06 239±8 -52.1±6.2PF-S-3 0.24 M 3.30±0.21 -3.29±0.23 422± 16 -52.3±2.5PF-S-4 0.31 M 6.12±0.78 -5.88±0.76 330±10 -38.2±1.8PF-S-5 0.39 M 3.66±0.30 -3.28±0.32 295±15 -55.3±2.1PF-S-6 0.47 M 3.16±0.21 -3.26±0.26 244±5 -83.4±5.1PF-S-7 0.55 M 2.69±0.12 -2.60±0.16 413±9 -82.2±2.4PF-S-8 0.63 M 2.37±0.13 -2.28±0.12 303±15 -96.1±2.0PF-S-9 0.71 M 3.13±0.24 -3.19±0.22 286±14 -87.0±3.5PF-S-10 0.78M 3.27±0.19 -3.30±0.18 308±10 -60.1±1.7Chapter 4 Hydrogen Peroxide Study - 143 -4.2.4.2 Peak Reaction RatesThe reaction rates for thepos and neg phases of the CAE d.c. and UV-Vis data arein Table 4-3 and depicted graphically in Figure 4-2. The CAE d.c. data presents nosurprises as it shows an initial period of instability before settling down to a near lineartrend for both the pos and neg rates. As mentioned in previous discussions, there is aclose correlation between the pos and neg rates - the average neg/pos ratio is0.978±0.041.The UV-Vis data exhibits interesting behaviour. The iodine production ratesappear to oscillate with increasingH20.This trend is also seen in Cooke’s data3° , but toa lesser extent. The iodine consumption rates are nearly linear, again in agreement withCooke. Unlike the previous iodate series, the consumption / production ratios are muchlower - the average is 0.271±0.070.Chapter 4 Hydrogen Peroxide Study - 144-10- -5008-E D -400D6- o D -300o D. 4- -200..U) Cl)2- ° -100-0 :::-• I- 1 .jrn----.0 • • • • • o--200-6- • --300-8- --400o CAE pos [iodinej pos• CAE neg [iodine] neg-lO-iIiiiiiiiiiiiii iiiiiiiiiiiiIiiii--5000.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9initial [H20] (M)Figure 4-2. Reaction rates for different initialH20concentrations.4.2.4.3 Extent of ReactionThe integrated CAE d.c. data for series PF-S-1 through 6 are plotted in Figure 4-3and for series PF-S-7 through 10 in Figure 4-4. Numerical analysis of the data curvesrevealed the following facts:PF-S-1. This pre-oscillation series fit best to a first order function with k = 3.88x10 51r2 = 0.997 and see = 0.185, suggesting that catalytic decomposition of H20 is theprimary CAE process.Chapter 4 Hydrogen Peroxide Study - 145-PF-S-2. The data from this series provided a linear fit with slope 0.01, r = 0.998 andsee = 0.072. We suggest that at this [H20] (the minimum necessary for oscillations),there is competition for the H20 between the oscillatory and non-oscillatory processesresulting in a pseudo-stationary state where the rate of oxygen evolution appears constant.PF-S-3 shows the beginnings of a second order function of the form discussed earlier.The fit resulted in k = 4.91x10 mV’s’ , r = 0.998 and see = 0.136.PF-S-4 also fits well to the same second order function with k 3.97x10 mV’s1,= 0.98 8 and see = 0.23 7.PF-S-5 is the first of the series with multiple distinct transition events. It is extremelydifficult to curve fit data of this nature. We experienced similar difficulties with PF-S-6.PF-S-7 through PF-S-10 all fit best to the second order function described above. Theparameters are in Table 4-4.Chapter 4 Hydrogen Peroxide Study - 146-Figure 4-3. Integrated CAE d.c. data for PF-S-1 through 6.Beginning with PF-S-5, transition events can be seen which correspond to eventsobserved in the UV-Vis data. Unlike previous series, the acoustic events are in phase withthe iodine events - there is no time shift. For example, in PF-S-5 we observe transition0 100 200 300 400 500 0 100 200 300 400 50012108642012108642010864205’E2U)-DuJ5’22DU)C-)-ow5’22U)-Dw065432I0121086420151296300 100 200 300 400 500Time (s)5’22U)C-)0w4:05’EU)C)-Dw5’22U)C.)-ów4:00 100 200 300 400 500 0 100 200 300 400 500— I I I ii I III I I I ii II I Ii iiIIIIIIIIIIIIIIIIIIPF-S-50 100 200 300 400 500Time (s)Chapter 4 Hydrogen Peroxide Study - 147-events at 240 seconds (max. peak amplitude), 330 seconds (last regular oscillation) and400 seconds (termination of oscillations).We conclude that if [H20] is low (but high enough to provide regularoscillations), the oscillatory mechanism competes forH20more efficiently than the non-oscillatory mechanism. This results in a smooth, second order rise of integrated CAE d.c.data. Since in the oscillatory mechanism oxygen is evolved only during the iodineproduction stage, the CAE events will be in phase with the iodine events.Table 4-4. Second order curve fit results for integratedCAE d.c. data.Series initial k r2 see[H20] (mV4s’)PF-S-7 0.55 M 3.l5xl0 0.997 0.133PF-S-8 0.63 M 2.90x104 0.998 0.104PF-S-9 0.71 M 9.25x105 0.998 0.098PF-S-l0 0.78 M 2.40x10 0.996 0.157If [H2Oj is high, both mechanisms receive sufficient H20 to generate oxygengas. The steady oxygen evolved by the non-oscillatory mechanism inhibits the switchingability of the oscillatory mechanism resulting in a delay between moving from iodineconsumption to production. The switch cannot occur until sufficient dissolved oxygen isremoved from the reaction mixture.0Time (s) Time (s)Figure 4-4. Integrated CAE d.c. data for PF-S-7 through 10.Chapter 4 Hydrogen Peroxide Study - 148 ->22DU)C)LUC-)10864201080 100 200 300 400 500 0 100 200 300 400 5006>2EDIf)C)UiC-)410>822‘U)46LU2<C-)010>822C,)4 CLU2<C-)0200 100 200 300 400 500 0 100 200 300 400 500Chapter 4 Hydrogen Peroxide Study - 149-4.2.5 Series PF-S: ConclusionsThe data from PF-S-1 and 2 illustrate the different processes responsible for CAE.In PF-S-1, the non-oscillatory catalytic decomposition of hydrogen peroxide is the mainprocess generating the bubbles whose bursting is detected by CAE. In PF-S-2, the iodineproduction stage of the oscillator is the primary bubble nucleation process.The UV-Vis peaks (where oscillations are observed) begin the reaction as sharp,needle shaped peaks and end the reaction as broadly skewed peaks implying that low[H20Jfavours iodine production over consumption.Autocorrelation analysis of the CAE d.c. data uncovered periodicity in thoseseries where the system was oscillating weakly, or not at all. This is consistent with theobservations made in the iodate series. The time constant of the periodicity varied from32 to 40 seconds. In contrast, the iodate series (with [H20] fixed at ca. 0.78M) postedtime constants in the range 16 to 27 seconds. This suggests that periodicity is inverselyproportional to the [H20], which supports Wentzell’s dissolved oxygen model. That is,assuming that the rate of oxygen evolution is proportional to [H20], at lowconcentrations it would take a longer period of time for the dissolved oxygen to reach itssolubility limit and commence bubble formation.The number of CAE d.c. peaks observed (10± 1) was mostly independent of initial{H20]. The number of UV-Vis peaks (11±2) was directly proportional to initial [H20].The CAE d.c. data showed close correlation of the absolute values of the pos andneg rates independent of initial [H20]. This is unique from the UV-Vis data which showsthe pos (iodine production) rate oscillates while the neg (iodine consumption) rates showlittle variation.Chapter 4 Hydrogen Peroxide Study - 150-The integrated CAE d.c. data showed similar transitions as the previous iodateseries (PF-O and PF-T). In contrast, these transitions occurred mostly in phase with theiodine events rather than out of phase as observed for the iodate series. We conclude thatthe oscillatory mechanism competes for 11202 more effectively than the non-oscillatorymechanism.Chapter 5 Conclusions -151-Chapter Five- Conclusions“Life is full of little surprises.”- Pandora5.1 Proposed Skeleton Mechanism of the BR OscillatorBoth the CAE and the UV-Vis data show that the BR oscillator moves betweentwo steady states: iodine production and iodine consumption.5.1.1 Iodine ProductionStep 15 initiates the radical processes which then produces HOlOautocatalytically. The experimental rate constants are not available for 15 or Ml althoughthey have been estimated at 1 .5x1 M2s1 and 1 .Oxl O M’s’ by Furrow and Noyes32.Step M2’s rate65 is on the order of i04 at 25°C and that for step 02 is 7.5x10 M’s’ at25°C66 . This suggests that 15 will be the rate limiting step, and the oxygen evolved instep 02 will be representative of the consumption of iodate to form HOlO.The autocatalytic production of HOlO described above would carry onindefinitely if not for the presence of step 12 which becomes energetically favoured whenHOlO and F are at acceptably high levels. Both this step and step Ii are expected to berapid in comparison to Dl. Since step Dl evolves oxygen, the CAE produced should bein phase with the production of iodine. However, the phase relationships observed inChapter 5 Conclusions -152-series PF-G-3 in Chapter Three indicate that the CAE is slightly out of phase with iodineproduction. We postulate that the slight delay is due to Wentzell’s diffused oxygenmodel, discussed earlier.5.1.2 Iodine ConsumptionThe oscillator switches from iodine production to consumption when iodide ion is lowenough to make step C4 more favourable than step Ii. During consumption, no oxygen isevolved, and so the rate of decrease of CAE will parallel the decrease of ‘2•The nine elementary steps proposed for this mechanism were first introduced byFurrow and Noyes32 in 1982, who initially proposed a total of 30 steps. Numericalmodelling selected 10 steps as a necessary and sufficient set to account for the oscillatorydynamics. The other 19 steps steps may of course occur. The step we did not include inour mechanism is14. 2H010 —> IO + HOl + HThis step is a two-equivalent reduction of iodine, implying that it should lie withthe other two-equivalent processes in the non-radical iodine production stage. However,steps 12, Dl and Ii adequately explain the experimental observations without invokinganother step, and so 14 is discarded. This assertion will need to be tested via numericalmodelling to see if the model can be made to oscillate without 14.In 1992, Turyani67 applied principal components analysis to the rate sensitivitymatrix of the BR oscillator. A particular species was considered redundant if the omissionof all of its consuming reactions had no significant effect on the calculated concentrationsof important species. Using this system, Turyani reduced the number of mechanistic stepsChapter 5 Conclusions -153-to 8. This is close to our proposed mechanism with the exception that step 02 has beenomitted.5.1.3 Reaction SchemeFREE RADICAL PRODUCTION OF 11010:D3. H + IO + H20 H0I0 + R20 +15. H + IO + H0I0 —> 210 + H20Ml. 10 + H20 + Mn2 H0I0 + Mn(OH)2M2. Mn (QH)2 + F120 —> R20 + Mn2 + H0002. 2H00 H0 +CONSUMPTION OF H0I0 TO FORM 12:12. H0I0 + I + H 2H01Dl. HOl + H20 + + H + H20Ii. HOl + I + H 12 + H20CONSUMPTION OF 12:C4. enol + 12 — RI + f + H13. I + I0 + 2H HOTO + H0IChapter 5 Conclusions -154-5.2 The Rate Limiting StepThere has been some question in the literature as to which step in the BRoscillator is rate limiting. Reaction C4, the iodide abstraction by the organic substrate, hasa rate constant estimated at 40 M’s’ based on work with the methylmalonic acidvariation of the BR oscillator68.The currently accepted rate constant for reaction Dl, thereduction of HOl to F, is 37 Mrn’s’, determined by Liebhafsky62.More recently, Furrowstudied reaction Dl under low-light conditions69 with periodic sampling by UV-Visspectroscopy and published a rate constant of 3 ± 2 M’s’.Reaction Dl is the main oxygen evolving step of the BR oscillator and henceseemed a good candidate to study with CAE. In Chapter Two of this work, reaction Dlwas studied under pseudo first order conditions and a rate constant of 1 ± 5 s_i obtained.Direct comparison of rate constants is difficult when they are of different reactionorders. Estimations could be made based on time to half-peak-height, for example, butresearchers generally do not publish the raw data necessary to make those calculations.Another alternative would be to numerically integrate the proposed models, an option webriefly discuss in Chapter Six - Further Work.Despite the shortcomings mentioned above, we believe that it is safe to concludethat reaction Dl is the rate limiting step in the BR oscillator.Chapter 5 Conclusions -155-I5.3 CAE as an Analytical Tool to Study Oscillating ReactionsFrom the outset of this work, one of the goals was to develop an experimentalsystem utilizing CAE which could deliver useful chemical information.CAE has general attributes which make it useful for studying chemical systemswhich result in phase changes, gas evolution, crystal fracture, etc.. From an experimentalstandpoint CAE offers these benefits:non-invasive: the BR oscillator is extremely sensitive not only to other halides, but alsoto physical conditions such as light and stirring. The non-invasive nature of CAE made itideal to study this reaction.simple calibration: CAE transducers are calibrated without removing them from thereaction vessel by attaching a second transducer to the reaction vessel and delivering acontinuous signal from a signal generator.In this work, CAE helped our understanding of the BR oscillator. Some of theconclusions we were able to draw are detailedin the following paragraphs.In Chapter Two, the preliminary studies revealed that two processes were takingplace in the induction. The neg phases were found to be independent of reagentconcentration which suggested that a neg phase was in fact non-production of oxygen andthat the oxygen was leaving the system through physical processes rather than beingconsumed in other reactions. The pos phase exhibited second order function behaviour; aforeshadowing of the second order behaviour which will be uncovered in the detailedstudies and which confirmed step Dl as second order.An attempt was made to isolate the CAE processes involved through theapplication of multivariate statistics on the a.c. signal data. The computed averagedChapter 5 Conclusions -156-power spectra uncovered some structure in the frequency domain which led to thedivision of the spectra into two distinct regions: reaction activity and transducersignature.Descriptor analysis of the a.c. signal data correctly predicted which signalsbelonged to a valid experimental run, and which belonged to a noise run. Principalcomponents analysis failed to uncover any underlying structure of the data. Nonheirarchical clustering also failed to uncover further structure.In Chapter Three, the detailed studies provided supporting evidence for theconclusions drawn in Chapter Two. The studies also uncovered the influence oftemperature on oxygen evolution, correlated transition events in the integrated CAE datato events in the UV-Vis data, and discovered the presence of periodicity which has beenunreported in the literature.Despite the natural variability of CAE processes, several examples of usefulquantitative relationships between integrated CAE and chemical concentrations wereobtained. This bodes well for future analytical use of the technology. However, with allthe advantages of CAE, there have also been some difficulties. Within-run reproducibilityis poor, and significant errors accompany most calculations. The a.c. signals requireconsiderable resources to acquire and store. Multivariate analysis produced few tangibleresults because of the single CAE production mechanism (gas evolution) studied.Despite the shortcomings of this new technology, we believe that the utility ofCAE as a tool for quantitative chemical analysis has been amply demonstrated in thisexploratory work.Chapter6 Further Work -157-IChapter Six - Further Work“The best laid plans often go fowl.”- W.E. CoyoteThe difficulty with a first thesis in an area of experimental chemical exploration isthat it often raises more questions than it can answer. The work presented here is nodifferent. More experimenting at different temperatures and varying the other reagents inthe system would provide excellent groundwork for answering the following questions.What is a “shoulder event”?We came across shoulder events in Chapter Two. They were observed many timesin these experiments and appeared to be a phenomenon whereby the oscillator switchedfrom state A to state B, only to switch back to A again after a few seconds. Since thisevent is unreported in the literature, it can only be answered by further CAEexperimenting in conjunction with other in situ spectroscopic techniques.LuminescenceSince the BR oscillator gives rise to high-frequency acoustic waves detectable byCAE, there is the possibility that these waves are also producing light. An enclosedChapter6 Further Work -158-experimental apparatus equipped with a photodiode - or better still, a monochromator anda photomultiplier tube might answer this question.Are the Averaged Power Spectra regions valid?Again in Chapter Two, we divided the averaged power spectra into two distinctregions. It would be interesting to see if these regions are universal - i.e. that they can beapplied to all CAE spectra.PeriodicityIn Chapter Three periodicity was observed in those experiments which did notproduce regular oscillations. Are the oscillations indeed the result of dissolved oxygenrelease (as we suggest) or are they due to other physical or chemical processes?Is CO2 gas evolved?Whether or not CO2 gas is evolved is still argued in the literature. We attemptedto answer this question through the use of a gas partitioner, but the rapid oscillations ofthe BR oscillator prevented reproducible results.It is difficult to write possible CO2 evolving reactions - the only source of carbonin the reaction mixture is the organic substrate, and it spends considerable time flippinginto the enol form and abstracting 1 from iodine. Any side reaction which breaks downthe substrate to produce CO2 would have to compete with its iodide scavengingpreference.Chapter6 Further Work -159-IIn 1987, Rubin and Noyes published measurements of critical supersaturation forhomogeneous bubble nucleation63 . They discovered that the thresholds for the diatomicgases H2, N2, 02, CO and NO in aqueous solutions all lie between 0.012 aqnd 0.07 M,while that for CO2 was 0.4 M. Assuming that their measurements can be applied to theBR oscillator in at least a relative sense, then it may be that CO2 never reaches thethreshold required to give rise to significant bubble evolution.Numerical ModelingIn Chapter Five, we presented a modification to the BR oscillator’s skeletonmechanism and a rate for reaction Dl considerably lower than the currently acceptedvalue. Unfortunately, there wasn’t sufficient time to test the validity of our model bynumerical integration, and this is left for future researchers.I References -160-References1. Förster, F. and Scheil, E., Zeitshrftfür Metalkunde, 24, 245, 1936.2. Schofield, B.H.,ASTMSTP, 505, 11, 1972.3. Nozue, A. and Kishi, T., I Acoustic Emission, 1, 1, 1982.4. Williams, J.H. and Lee, S.S., Mat. Eval., 43, 561, 1985.5. Betteridge, D., Conners, P.A., Lilley, T., Shoko, N.R., Cudby, M.E., and Wood, D.G.,Polymer, 24, 1206, 1983.6. Wentzell, P.D., and Wade, A.P.,Anal. Chem., 61, 2638, 1989.7. Beichamber, R.M., Betteridge, D., Collins, M.P., Lilley, T., Marczewski, Z., andWade, A.P.,Anal. Chem., 61, 2638, 1989.8. Lee, 0., Koga, Y., and Wade, A.P., Talanta, 37, 861, 1990.9. van Ooijen, J.A., van Tooren, E., Reedijk, J., I Am. Chem. Soc., 100, 1978.10 Munro, B.F., MSc. Thesis, Univ. of British Columbia, 1991.11.Betteridge, D., Joslin, M.T., and Lilley, T.,Anal. Chem., 53, 1064, 1981.12. Sawada, T., Gohshi, Y., Abe, C., and Furuya, K., Anal. Chem., 57, 1743, 1985.13. Sawada, T., Gohshi, Y., Abe, C., and Furuya, K., Anal. Chem., 57, 366, 1985.14. Lubetkin, S.D.,I Appi. Electrochem.,, 19, 668, 1989.15. Beichamber, R.M., Betteridge, D., Collins, M.P., Lilley, T., Marczewski, C.Z., andWade, A.P., Anal. Chem., 58, 1874, 1986.16. Wade, A.P., Soulsbury, K.A., Chow, P.Y.T., and Brock, I.H., Anal. Chim. Acta, 246,23, 1991.17. Massart, D.L., ed., “Chemometrics: A Textbook”, pg. 202, Elsevier, New York, 1988.18. Wentzell, P.D., Lee, 0., and Wade, A.P., I Chemom., 5, 389, 1991.19. Lotka, A.J., I Phys. Chem., 14, 271, 1910.20. Lotka, A.J., Z. Phys. Chem., 80, 159, 1912.21. Lotka, A.J., I Am. Chem. Soc., 42, 1595, 1920.22. Bray, W.C.,I Am. Chem. Soc., 43, 1262, 1921.23. Bray, W.C., and Liebhafsky, H.A., I Am. Chem. Soc., 53, 38, 1931.24. Belousov, B.P., Sbornik Ref Rad. Med., p. 145, 1958.25. Zhabotinskii, A.M., Biofizika, 9, 306, 1964., References -161-I26. Turing, A.M., Phil. Trans. Roy. Soc. B., 237, 37, 1952.27. Burger, M., and Field, R.J., Homogeneous Oscillating Chemical Reactions. ABibliography, Dept. of Chem., Univ. ofMontana, 1983.28. Briggs, T.S. and Rauscher W.C., I Chem. Ed., 50, 496, 1973.29. Cooke, D.O., React. Kinet. Catal. Lett., 3, 377, 1975.30. Cooke, D.O., Inorg. Chim. Acta, 37, 259, 1979.31. De Kepper, P., Epstein, I.R.,I Am. Chem. Soc., 104,49, 1982.32. Furrow, S.D., and Noyes, R.M.,J Am. Chem. Soc., 104, 45, 1982.33. Noyes, R.M., and Ganapathisubramanian, N.J., I Chem. Phys., 76, 1770, 1982.34. De Kepper, P., Compt. Rend. SceancesAcad. Sci. Ser. C., 283, 25, 1976.35. Lamprecht, L. and Schaarschmidt, B., Thermo. Acta., 112, 95, 1987.36. Wedlock, D.J., ed., “Controlled Particle, Droplet, and Bubble Formation”, pg. 159,Butterworth-Heinemann Ltd., Oxford, 1994.37. Wilt, P.M., I Colloid. Inter. Sci., 112, 530, 1986.38. Cole, R., Adv. Heat Transfer, 10, 85, 1974.39. Volmer, M., “Kinetik der Phasenbildung”, Steinkopf, Leippzig, 1939.40. Sellers, R.M.,Analyst, 105, 950, 1980.41. Vogel, A.I., “Vogel’s Textbook of Quantitative Chemical Analysis”, LongmanScientific and Tech., Essex, England, 1989.42. Dutt, A.K. and Menzinger, M., I Phys. Chem., 96, 8447, 1992.43. Handbook ofChemistry and Physics, 64th edition, CRC Press.44. Wentzell, P.D., and Wade A.P., Anal. Chem. 61, 2638, 1989.45. Thompson, M., and Howarth, R.J., Analyst, 54, 454, 1970.46. Siegel, S., “Nonparametric Statistics for the Behavioral Sciences”, McGraw-Hill,New York, 1956.47. Hartigan, J.A., and Wong, M.A.,Applied Statistics, 28, 100, 1979.48. Trojanowicz, M. and Matuszewski, W., Anal. Chim. Acta, 138, 171, 1982.49. Davey, D.E., Mulcahy, D.E., O’Connell, G.R., Analyst, 117, 761, 1992.50. Linder, E., Toth, K., and Pungor, E., Anal. Chem., 54, 202, 1982.51. Linder, E., Toth, K., and Hrabeczy-Pall, A., “Dynamic Characteristics of IonSelective Electrodes”, CRC Press, Boca Raton, FL, 1988.52. Mahon, M.J., and Smith, A.L., I Phys. Chem., 89, 1215, 1985.References -162-53. Paquette, J., and Ford, B.L., Can. .J. Chem., 63, 2444, 1985.54. Awtrey, S. and Connick, B., I Am. Chem. Soc., 73, 219, 1951.55. Betteridge, D., Sly, T.J., Wade, A.P., and Porter, D.G., Anal. Chem., 58, 2258, 1986.56. Wentzell, P.D., Hatton, M.J., Shiundu, P.M., Ree, R.M., Wade, A.P., Betteridge, D.,and Sly, T.J., I Auto. Chem., 11(5), 227, 1989.57. Jwo, J.J., Noyes, R.M., I Am. Chem. Soc., 97, 5422, 1975.58. Furrow, S.D., and Noyes, R.M., I Am. Chem. Soc., 104, 38, 1982.59. Cooke, D.O., Intl. I Chem. Kinet., 12, 683, 1980.60. Roux, J.C. and Vidal, C., Nouv. I Chim., 3, 247, 1979.61. Field, R.J., and Burger, M. eds, “Oscillations and Travelling Waves in ChemicalSystems”, John Wiley & Sons, Toronto, 1985.62. Wentzell, P.D., VanSlyke S.J., and Bateman, K.P.,Anal. Chim. Acta., 246, 43, 1991.63. Rubin, M.B., and Noyes, R.M.,I Phys. Chem., 91, 4193, 1987.64. Sandifer, J.R., Anal. Chem., 53, 312, 1981.65. Davies, G., Kustin, K., Inorg. Chem., 7, 146, 1968.66. Watters, J.I., Kolthoffl.M., I Am. Chem. Soc., 70, 2455, 1948.67. Turanyi, T., React. Kinet. Catal. Lett., 45, 235, 1991.68. Furrow, S.F., I Phys. Chem., 85, 2026, 1981.69. Furrow, S.F., I Phys. Chem., 91, 2133, 1987.Appendix A List of Manufacturers -163-List of Manufacturers(in alphabetical order)AET Corporation Sacramento, California, USAAlitea USA Medina, Washington, USABDH Chemicals Toronto, OntarioBlue M Electric Co. Blue Island, Illinois, USABrUel & Kjr Naerum, DenmarkCampus Computers Vancouver, British ColumbiaCole Panner Chicago, Illinois, USAFisher Scientific Nepean, OntarioHewlett Packard Richmond, British ColumbiaIBM Corporation Boca Raton, Florida, USAIsmatec Zurich, SwitzerlandMacArthur Chemical Montreal, QuebecMetrabyte Corporation Taunton, Massachusetts, USANora Systems Vancouver, British ColumbiaOrion Research Inc. Boston, Massachusetts, USATektronix Beaverton, Oregon, USAAppendix B Proprietoty Software -164-Proprietory Software DevelopedThe software listed here was developed in-house. Those programs marked with anasterisk were written by this author.SIGNAL ACQUISITIONThese programs acquire CAE signals from various digitizers and store them to disk in aproprietory file format designed by members of this research group. This format, “AES”,is a binary format designed for fast access and compact storage.ACK*- supports Tektronix 2200 series digital storage oscilloscopesACQ243OA - supports Tektronix 2400 series digital storage oscilloscopesACQS2000 - supports Soltec SDA2000 digitizerNOFRILLS*- supports Metrabyte PCIP oscilloscope adapter cardSIGNAL VALIDATIONValidation modules were written to check the performance of the digitizer hardware andsoftware. They accept AES files as input and provide output to the screen or to flat ASCIIfiles.AESCHECK - flags under-triggered and over-ranged CAE signalsAESEDIT*- graphical interface for deleting noise signalsSIGNAL ANALYSISThe analysis programs accept AES files as input and provide interactive display andediting of CAE signals with the exception of TRAPS which is designed to run in anAppendix B Proprietoiy Software -165-automated mode. All programs generate flat ASCII files. TRAPS generates averagedpower spectra files: AVP (entire experiment) or TRA (time resolved).AESEDIT*- graphical interface for editing and manually classifying CAE signalsAEVIEW*- graphical interface for inspection of all CAE data filesTRAPS - generates time-resolved averaged power spectraDESCRIPTOR EXTRACTIONThe main program is FEATURES which generates binary format descriptor (DS1) filesfrom AES files. DS1EDIT accepts DS1 files for input and is especially useful forbreaking down large data sets into smaller, more manageable, data files.FEATURES - computes descriptors for CAE signalsDS 1 EDIT* - interactive editing of descriptor filesDESCRIPTOR ANALYSISA variety of software has been written to aid in the analysis of the descriptor data oncedescriptors have been computed.AETREND - calculates distributions and correlates with physical factorsCOMATRIX - computes various measures of descriptor correlationDS 1 STATS - generates univariate statistics for descriptorsHCLUSTER - performs hierarchical clusteringK-MEANS - performs non-hierarchical clusteringPCA - performs principle components analysisWILCOXON - computes the Wilcoxon two-tailed rank sum testAppendix B Proprietoy Software -166-IOTHER SOFTWAREPF*This is the software written to perform the experiments detailed in Chapters Three andFour. PF was designed to operate and oversee all the instruments in use. For maximumflexibility, PF used both an “immediate” and a “script-based” system. In the immediatemode, the user could directly control any/all instruments. In the script mode, the userwould write a simple, English language list of commands and store the list in a file. PFcould then be instructed to read the file and process the commands in the list. This maderepeated, unattended experiments possible.CD 0 C) t CD CD CD 0 CD CD 0)(s)amrj0CD 2 cm CDCu0Cu‘I-aC)tC)di—‘I_>0C-)VtC)C!)C)C’)CCeC)C’)IaS-0---+-tt14---C-)tC)bD::LiiZ: tz-:—±t&•ALtziTime(s)CD H C I CD C,) CD 0 C En C) C C C En CD CD(s)au.nj.CD C-) cp t CD CD CD CD Er C-, SCC-CD C-)-t ofl-f IAppendix C Experimental Data -1 70-3002501005005000Time (s)Time (s)Reagent Set A, Run ICo0200xU)U)C00U)ci)0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150Reagent Set A, Run 2U)0>1000 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150Figure C-4. Digitized chart recordings of Reagent Set AAppendix C Experimental Data -1 71-1.4- 2.0— i I I I I I I I I I I - 10.01.2 - PF-G-1 - 8.0- 6.01.0-- 4.00.8-- 2.00.6- 0.0ci)CLJ. -o o x0 rr —‘JJ (1)1.0o 00.8- =0.6 -0.4 -0.2—0.’1 — 0.0 — I I— —8.00 100 200 300 400 500 6001.4- 2.0- I I - 10.01.2- PF-G-2- 8.0- 6.01.0-- 4.00.8 - 2.00.6- 0.0xE- CD.0.2-o. .D nn . —(I)1.0 .5C) 00.8- =-0.4- 0.0- I I I I --8.00 100 200 300 400 500 600Time (s)Figure C-5. Data from series PF-G-1 and 2: [iodine] (top), CAE d.c. (middle dotted) and[iodide] (bottom).Appendix C Experimental Data -172-I2.0cx010.08.06.04.02.00.01.20.9Cx00.60.30.00 100 200PF-G-4300 400-8.0500 6001.4 -1.2 -1.0 -0.8 -0.6 -5 0.4 -E0.2 -QDuJC-)-0.41.4 -— 0.0 — I I I I I I I I I I I I —0 100 200 300 400 500 600Tine(s)Figure C-6. Data from series PF-G-3 and 4: [iodine] (top), CAE d.c. (middle dotted) and[iodide] (bottom).- 16.0- 12.0- 8.0- 4.0- 0.0Appendix C Experimental Data -1 73-10 -0xci)-D04.3-I I I I I I - 20.0PF-G-5- 15.0*0°0 — I I I I I I I I I I I—0 100 200 300 400 500 60010- I -0xci)C0IIIPF-G-6I Jill1.4 -1.2 -1.0 -0.8 -0.6 -5 0.4 -E0.2 -C)° 00-wC-)-0.4 -1.4 -___________________________________1.2 -1.0 -0.8-0.6- QxE ci) ..-o f..! 00.2- . .C.) ...0 rr —ci)LJJ . C4’-0 .23-2-1--0.4- 0- I I I --8.00 100 200 300 400 500 600Time (s)Figure C-7. Data from series PF-G-5 and 6: [iodine] (top), CAE d.c. (middle dotted) and[iodide] (bottom).20.0- 15.010.05.00.0I I I I I r I IAppendix C Experimental Data -174-I— II I —o ioo 200 300 400 500 600— I I I I I I I I I I I —0 100 200 300 400 500 600Time (s)axci)C0axci)C0Figure C-8, Data from series PF-G-7 and 8: [iodine] (top), CAE d.c. (middle dotted) and[iodide] (bottom).1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 -0.0 -,-0.4 -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 -0.0 --0.4 ->0t3wC-)>E6-dwC-)10 -axa)04,3-2-1—010a><ci)-o-o043-2-1-0-- 25.0- 20.0- 15.0- 10.0- 5.00.0-8.025.0- 20.0- 15.0- 10.0- 5.00.0-8.0I . . IAppendix C Experimental Data -175-10I1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 -0.0-0.4 -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 -0.0432I0>2U-ow0>20wC)- 25.0- 20.0- 15.0- 10.0- 5.00.0-8.0- 25.0- 20.0- 15.0- 10.0- 5.00.00 100 200 300 400 500 6000xJ)000xci)-o0100xa)-D043-21 PF-G-1 0I I I I I r I I I I i I i p p I r I P I I r I I I0- --8.00 100 200 300 400 500 600Time (s)Figure C-9. Data from series PF-G-9 and 10: [iodine] (top), CAE d.c. (middle dotted) and[iodide] (bottom).Appendix C Experimental Data -1 76-I I I I I I — 15.0PF-K-1, [iodatej = 0.004 M- 12.09.06.03.0-o.ob0.10oCI)0.05-0.00 - ---0.05-- I I I I I I I I I I I I0 100 200 300 400 500I I I I I I I I I I — 15.0PF-K-2. [iodate] = 0.008 M- 12.0- 9.06.0- 300.0 ‘b0.100 ><—U)W C0.05-n0.00- - - - - - - - -—0.05 — I I I I I I I I I I I I I I0 100 200 300 400 500Time (s)Figure C-1O. Data from series PF-K-1 and 2: [iodine] (top), CAE d.c. (bottom),Appendix C Experimental Data -177-I3.00.0><U)0I I I I I I I IPF-K-3, [lodate] = 0.011 MI I I I I —i i i i0xU)V0- 15.0- 12.0- 9.0- 6.0- 3.00.015.012.09.06.00 100 200 300 400 500E0.10 ‘V0.05-0.00-0.05E0.00-0.050 100 200 300 400lime(s)500Figure C-i i. Data from series PF-K-3 and 4: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -178-I- 15.0- 12.0- 9.0- 6.0-3.00.0x0V0- 15.0- 12.0- 9.0- 6.0-3.00.0xci)V0PF-K-5, [iodate] = 0.019 MCVUI0>VUI00 100 200 300 400 500PF-K-6, [iodate] = 0.023 M0.200.15 -0.10 -0.05 -0.00 --0.05 -0.200.15 -0.10 -0.05 -0.00 --0.05 -0 100 200 300 400 500Time (s)Figure C-12. Data from series PF-K-5 and 6: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -1 79->E0wC)>0-ouJC)PF-K-7, [iodate] = 0.026M0 100 200I I I I30015.012.09.06.03.00.0><ci)0020.015.010.05.00.0xci)000.200.15 -0.10 -0.05 -0.00 --0.05 -0.200.15 -0.10 -0.05 -0.00 --0.05 -400 500PF-K-8,[iodate] = 0.03DM0 100 200 300 400 500Time (s)Figure C-13. Data from series PF-K-7 and 8: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -180-I PF-K-9,[iodate]=0.034M20.015.010.05.00.0xU)-c020.015.010.0uJC)cSwC)0.200.15 -0.10 -0.05 -0.00 --0.050.20 -0.15 -0.10 -0.05 -0.00 --0.050 100 200 300 400 500I I I I —PF-K-10, [iodate] = 0.037 M— I I I I I I I I I I I I I I I I I100 200 300 400 500Time (s)Figure C-14. Data from series PF-K-9 and 10: [iodine] (top), CAE d.c. (bottom).0Appendix C Experimental Data -181-I I I — 15.0PF-O-1, [iodate] = 0.004 M- 12.0- 9.0- 6.0- -3.0>0.0 °0.05-0.00 - ““,....,..,..,: ‘.....,-. ,,, :-, ,.,-0.05- I V I0 100 200 300 400 500I •I— 15.0PF-O-2, [iodate] = 0.004 M- 12.0- 9.0- 6.00.05- .0.00 - : ‘ ‘V...-0.05- I I I I I I I I I I I I I I I I0 100 200 300 400 500Time (s)Figure C-15. Data from series PF-O-1 and 2: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -182-I I I I I I I I — 15.0PF-O-3 [iodate] = 0.011 M- 12.0- 9.0::‘o.io 0.0><—LU0.05-0.00- ... .. ...,:-0.05- I I I I I I I I0 100 200 300 400 500I I I I- 15.0PF-O-4, [iodate] = 0.015 M- 12.0- 9.00.10W0.05- .0.00 -.: ...... .-0.05- I I I I I I I I I I I I I I0 100 200 300 400 500Time (s)Figure C-16. Data from series PF-O-3 and 4: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -183-iJ I I I I — 15.0PF-O-5, [iodate] = 0.019 M- 12.0- 9.00.20’w 0.15-o 0.10- o0.05 -0.00 -:..-0.05- I I0 100 200 300 400 500I I- 15.0PF-O-6, [iodate] = 0.023 M- 12.0L/VTftUIA0.000.20 .Lii 0.15-o 0.10- .20.05 -...0.00 -.. ....:....... ..:-0.05- I I I I I I I I I I I I0 100 200 300 400 500Time (s)Figure C-17. Data from series PF-O-5 and 6: [iodine] (top), CAE d.c. (bottom).>E(.-ouJ0>EC-)uJ0Appendix C Experimental Data -184-I15.012.09.06.03.00.0><G)-o0PF-c-7, [iodate] =0.026 M100 200 300 400 5000.200i5 -0.10 -0.05 -0.00 --0.05 -00.200.150.100.050.00-0.0515012.090603.00.0xC-o0500Time (s)Figure C-18. Data from series PF-O-7 and 8: [iodine] (top), CAE d.c. (bottom).0 100 200 300 400Appendix C Experimental Data -185-I IPF-O-9,[iodate] 0.034MEC.)-oC-)00w(-)I I I0 100 200 300 400 50020.015.010.05.00.0xa)0020.0- 15.0- 10.0- 5.0xa)000.50.4 -0.3 -0.2 -0.1 -0.0 --0.1 -0.80.6 -0.4 -0.2 -0.0 --0.2 -PF-O-10, [iodate] = 0.037 MI I I I r II I I I0 100 200 300 400 500Time (s)Figure C-19. Data from series PF-O-9 and 10: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -186-II I - 15.0PF-T-1, [lodate] 0.004 M- 12.0- 9.0- 6.0—S -3.0>0.10 0.00.05-0.00 --0.05- I I I0 100 200 300 400 500I — 15.0PF-T-2, [iodatej = 0.004 M- 12.0- 9.0- 6.0-3.0>E :0.01:3ci)0.05- 10.00 --0.05- I I I0 100 200 300 400 500Time (s)Figure C-19. Data from series PF-T-1 and 2: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -187-ii — 15.0PF-T-3,[iodate]=0.011 M- 12.0- 9.0- 6.00.050.00 -—0.05 — I F F0 100 200 300 400 500I — 15.0PF-T-4, [iodate] = 0.015 M- 12.0- 9.0- 6.00.05:. H..0.00 - . .. . . .-0.05- F F I I I I I F I I I I I I F F F0 100 200 300 400 500Time (s)Figure C-20. Data from series PF-T-3 and 4: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -188-I0 100 200 300 400 500I . I I0 100 200 300 400 500PF-T-5, [iodatej 0.019 M-•••:•-•---- -- :------ --•-‘--- -..-...././-..•>-DwC-)>2C-DuJ0- 15.0- 12.0- 9.0- 6.0- 3.00.0- 15.0- 12.0- 9.0- 6.0- 3.00.0PF-T-6, [iodate] = 0.d23 M0.200.15 -0.10 -0.05 -0.00 --0.05 -0.200.15 -0.10 -0.05 -0.00 --0.05 -Qxa)00xa)C0Time (s)Figure C-21. Data from series PF-T-5 and 6: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -189-IPF-T-7, [iodate]= 0.026 M0 100 200E0.200.150.100.050.00 --0.05E‘ 0.200.15 -0.10 -0.05 -0.00 --0.0515.012.09.06.03.00.0xci)015.012.09.06.03.00.00xci)0300 400 500- ‘ IPF-T-8,[iodate]=0.030M0 100 200 300 400 500Time (s)Figure C-22. Data from series PF-T-7 and 8: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -190-IPF-T-9, [iodate] = 0.034 M>ELUC-)E-DLUC-)0 100 20020.015.010.05.0..0.00x0cD20.015.010.05.00.00xa)C-D0300 4000.200.15 -0.100.05 -0.00 --0.05 -0.200.15 -0.10 -0.05 -0.00 --0.05 -500I I I -I PF-T-10,[iodate]=0.037M0 100 200 300 400 500Time (s)Figure C-23. Data from series PF-T-9 and 10: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -191-PF-S-1, [H20]= 0.078 M0 100 200>20.100.08 -0.06 -0.04 -0.02 -0.00>20.100.080.060.040.020.0030020.015.010.05.00.0 0><ci)C20.015.010.05.00.0 0xa)-c0400 5000 100 200 300 400Time (s)500Figure C-24. Data from series PF-S-1 and 2: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -192-I0.0><U)C-o020.015.010.05.00.0 0xU)C-o020.015.010.05.00 100 200 300 400>S0.10008o 0.060.040.020.00>S0.100.08o 0.060.040.020.005000 100 200 300 400Time (s)500Figure C-25. Data from series PF-S-3 and 4: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -193-0 100 200 300 400>E0.100.080.060.040.020.00>20.100.08o 0.060.040.020.0020.015.010.05.00.0xC)C0020.015.010.05.00.0xU)C005000 100 200 300 400Time (s)500Figure C-26. Data from series PF-S-5 and 6: [iodine] (top), CAE d.c. (bottom).Appendix C Experimental Data -194-I>EC.)uJC-)>EwC-)0 100 200 30020.015.010.05.00.0ci)020.015.010.05.00.0xci)00.100.080.060.040.020.000.100.080.060.040.020.00400 5000 100 200 300 400Time (s)500Figure C-27. Data from series PF-S-7 and 8: [iodine] (top), CAE d.c. (bottom).>E‘ 0.100.080.060.0 axci)00.0 axU)00Appendix C Experimental Data -195-20.015.010.05.00 100 200 300 400 5000.040.020.00>E0.100.080.060.040.020.0020.015.010.05.00 100 200 300 400Time (s)500Figure C-28. Data from series PF-S-9 and 10: [iodine] (top), CAE d.c. (bottom).

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0059651/manifest

Comment

Related Items