摘要:

111 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 New British Standards Dynamic Properties of Rubbers. Part C5. Nitrogen. Determination of Insulation Resistance. Fat Products (Reference Method). Liquid Milk and Cream. Part 12. of Milk. and Maintenance* Section 5.14. Determination of Thermal Expansion (Temperatures Up to 1500 "C). Method* BS 6063: Method for Evaluation of General BS 1941 : Specification for 4-Methylpentan- Purpose EthYlene-ProPYlene-Diene IWd~er 2-one (Isobutyl Methyl Ketone) for Industrial (EPDM). Use. Method for Determination Of BS 2842: Hygrometer. Measurement of Gross Alpha Activity in Method of Evaluation by Gas Adsorption. ... BS 4364: Specification for Industrial Flow Rate of Particulate Material in Gas- BS 684: Methods of Analysis of Fats and Isothiocyanate and BS 6069: Characterization of Air Quality. Content Fatty Oils.Part 2. Other Methods. Section Vinylthiooxazolidone Content. Part 9. Part 4. Stationary Source Emissions. Section 2.3. Determination of Insoluble Impurities Preparation of Test Samples. 4.3. Method for the Manual Gravimetric Content. Section 2.37. Determination of Determination of Concentration and Mass Residual Technical Hexane Content. Oxygen. carrying Ducts. BS 903: Physical Testing of Rubber. Part A24. Guide to the Determination Of BS 4366: Specification for Industrial BS 6783" Sampling and Analysis of Nickel, Ferronickel and Nickel Alloys. Part 12. Method for the Determination of Phosphorus Molecular Absorption Spectrometry. Part 13, Method for the Determination of Phosphorus BS 4899: Guide to User's Requirements for in Nickel Alloys by Molybdenum Blue BS 1679: Containers for Pharmaceutical Technical Manuals, Dispensing.Part 8. Specification for Glass and Plastics Containers for Solid Dosage BS 5086: ~ ~ ~ l ~ ~ i ~ of B ~ ~ ~ ~ ~ . part 6. in Nickel, Ferronickel and Nickel Alloys by Forms, Semi-solids and Powders. Method for Determination of Fat Acidity and Phosphovanadomolybdate Molecular BS 1741: Methods for Chemical Analysis of Free Fatty Acids in the Fat of Butter and Milk Absorption SpectrometrJ'. BS 7079: Preparation of Steel Before Determination of Total Phosphorus Content BS 5726: Microbiological Safety Cabinets. Application of Paints and Related Products. specification for ~ ~ ~ i ~ ~ , Group B.Methods for the Assessment of part 1. BS 1743: Methods for Analysis of Dried Construction and Performance Prior to Surface "eanliness. Part 132. Method for Milk and Dried Milk products. part 5 . Installation. Part 2. Recommendations for Determination of Chlorides on Cleaned Determination of Bulk Density. Information to be Exchanged Between Surfaces* Purchaser, Vendor and Installer and BS 1902.: Methods of Testing Refractory Recommendations for Installation. Part 3. BS 7164: Chemical Tests for Raw and Materials. part 5. Refractory and Thermal Specification for Performance After Vulcanized Rubber. Part 3- Methods for properties. section 5 ,o. ~ ~ ~ ~ ~ d ~ ~ ~ i ~ ~ . Installation. Part 4. Recommendations for Determination Of Extract.Part 23- Methods for Determination of Total Sulfur Content. Section 23.1. Oxygen Combustion BS 7455: Analysis of Nickel Alloys by Flame Atomic Absorption Spectrometry. Part BS 6068: Water Quality. part 2. physical, 7. Method for Determination of Aluminium. Specification for Whirling Chemical and Biochemical Methods. Section Part 2.4 1. Determination of Fluoride: Electrochemical Probe Method for Potable BS 7591 Porosity and Pore Size BS 3406: Methods for Determination of and Lightly Polluted Water. Part 3. Distribution of Materials. Part 1. Method of Particle Size Distribution. Part 4. Guide to Radiological Methods. Section 3.1. Evaluation by Mercury Porosimetry. Part 2. Microscope and Image Analysis Methods. Non-Saline Water: Thick Source Method. BS 4289: Methods for Analysis of Oilseeds. Section 3.2. Measurement of Gross Beta PD 6532: Reference Materials. Part 2. Part 10. Determination of Chlorophyll Activity in Non-saline Water. Part 6. Guide to the Contents of Certificates of Content if Refereed. Sampling. Section 6.9. Guidance on Reference Materials. Part 3. Guide to the Sampling From Marine Waters. Section Uses of Certified Reference Materials. Part BS 4325: Methods for Analysis of Oilseed 6.10. Guidance on Sampling of Waste 4. Guide to General and Statistical Principles Residues. Part 8. Determination of Total Waters. for the Certification of Reference Materials.

ISSN:0144-557X    

DOI:10.1039/AP993300P003

出版商:RSC    

年代:1993

数据来源: RSC

ISSN:0144-557X    

DOI:10.1039/AP99330FX025

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

July 1993 ANPRDI 30(7) 289-320 (1 993) Ana lyt ica I Proceedings Proceedings of the Analytical Division of The Royal Society of Chemistry CONTENTS 289 Editorial 290 VAM Viewpoint 'Reference Materials and Validated Methods for Evaluating Biocompatibility' 291 Correspondence 291 Ronald Belcher Memorial Lectureship 292 Analytical Viewpoint 'Charting to Success Within Analytical Laboratories' by Jill A. Gardner, Shirley Coleman and S. G. Farrow 296 Report by the Analytical Methods Committee 'Evaluation of Analytical Instrumentation. Part VIII. Instrumentation for Gas-Liquid Chromatography 'From Feedback to Chaos in Chemical Systems' by Alison S. Tomlin 307 Chemometrics Papers 307 310 'Chaos, Chemometrics and Chemistry: An Introduction to Chaotic Systems' by Steven R. Bishop 315 Equipment News 319 Conferences and Meetings 320 Course 320 Publications Received iii New British Standards 0144-557XC198317: 1-5

ISSN:0144-557X    

DOI:10.1039/AP99330BX027

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 289 Editorial In writing this Editorial my first concern is to thank all those Analytical Division members who responded to our question- naire last Autumn. Of the 7707 question- naires circulated a very pleasing 14% (1050) were returned. The replies were fairly favourable to the journal in its present form; for instance 96% of those who currently subscribed to the journal agreed that the aim of the journal should be to communicate Analytical Division activities. Naturally, some types of article proved more popular than others, sum- maries of papers coming out on top and Equipment News appearing to be the least popular. I was surprised by the latter verdict because some instrument manu- facturers have told me that Analytical Proceedings is a good source of reader enquiries; nevertheless, this column will be dropped for 1994, although we hope to retain apparatus-based articles in some form.Unfortunately, the response of non- subscribing members to what might be termed the ‘crunch’ question, i. e., ‘given all of the listed contents and a very reasonable members’ price (quoted), would you take out a member subscrip- tion?’ was less encouraging with only 143 saying ‘yes’ against 438 saying ‘no’. The three most commonly stated reasons for answering ‘no’ were: respondent fully or nearly retired; library or some other source of the journal felt to be adequate; and respondent had moved out of analyti- cal chemistry. The answers to this last question provided us with considerable food for thought.In particular, we felt that in angling the content of the journal towards 1 members we were only satisfying 38% of subscribers and that we should instead be looking to benefit the 62% of (non- member) subscribers who provide the journal with most of its revenue. The problem was placed before the Analytical Editorial Board of the RSC at its meeting held in January, 1993, which produced the following solution (agreed by the Journals Management Committee in May, 1993). 1. The journal should be re-launched in 1994 as a communications journal, acting as an analytical equivalent of Chemical Communications (see below for further details). The communica- tions would, in general, be published within 10-12 weeks from the date of receipt of the manuscript. 2. The journal should continue to carry material of a review nature.Much of this would, as now, result from lectures given at meetings organized by the Analytical Division or its Groups and Regions. This material would now be refereed, the criterion being signifi- cance (or usefulness) to the analytical community. Papers from the same meeting would only be collected together if they were submitted within 3 months of the date of the meeting; after that time they would still be considered for publication but they would not carry a footnote tying them to any particular meeting. Material not presented at Analytical Division meet- ings would now be considered. 3. Other current features of the journal that might be retained were Corre- spondence, Analytical Viewpoints, Chemometrics Articles, Historical Articles.the full text of sDecial lectures (AD Silver Medal, Theophilus Red- wood, etc.), Conferences and Meet- ings, Courses and details of new British Standards and IUPAC discus- sion documents. 4. New features would include commis- sioned articles and a wider range of items of interest to the European chemical community. 5. The content of Analytical Division news (except for news of international importance) will be reduced. Obviously, the most radical of these changes is the publication of communica- tions. The criteria for acceptance of a communication will be that the work is of current interest and significance, justify- ing a preliminary report. In most cases the communication would be followed up by a full paper in a primary journal such as The Analyst. It is expected that each analytical communication will occupy no more than two printed pages inclusive of illus- trations, tables of results and the refer- ence list (i. e., 5-6 A4 pages of typescript). Again, as with the review-type material, the work described in communications need not have been presented at Analyti- cal Division meetings and workers all over the world will be encouraged to submit them. We hope that with these enhancements Analytical Proceedings will increase its international appeal and will be of even greater interest to its current subscribers. It will, of course, continue to be abstracted by the major abstracting agen- cies and will be produced to the high standard expected of RSC journals. ROGER YOUNG

ISSN:0144-557X    

DOI:10.1039/AP9933000289

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

290 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 Reference Materials and Validated Methods for Evaluating Biocompatibility The application of biocompatible ma- terials in medical technology has been receiving increased interest over the last few years. However, the potential is currently being restricted by the fact that the manufacture of most biomaterials, and the subsequent evaluation of their biocompatibility , is subject to inconsist- encies, limitations and opportunities for misinterpretation. These result from the: batch-to-batch variation of bioma- terials; requirement for the use of processing agents during the manufacture, and processing, of the majority of bioma- terials, and their inadvertent incorpora- tion and associated biological risk; a lack of a database describing bioma- terialltissue compatibility, i.e. , the complex interrelation of the response of the biomaterial to the host environ- ment and of the host to the biomaterial; a lack of test methods giving objective, validated and comparable data; and a lack of reference materials, to assist in the development and application of such methods. If significant breakthroughs are to be made in the longer term, it is necessary to derive meaningful comparisons of the advantages and disadvantages, and pat- terns of behaviour, of existing bioma- terials. The current approach of reliance on the experienced and professional scientist needs to be augmented by stan- dardization of manufacturing protocols combined with a range of reproducible, quantifiable and quantitative, pre-clinical test methods which are supported by reference materials.In this way, confi- dence may be achieved in both the consistency of test data obtained within, and between, laboratories and in the conclusions drawn. In turn, this should lead to the development and calibration of new test methods. As part of the DTI Valid Analytical Measurement Initiative (VAM) , the Materials Group of the Laboratory of the Government Chemist has addressed these problems through the production of refer- ence materials and the independent vali- dation of procedures for evaluating biocompa tibili t y . Reference Material Production Three polymeric matrices, based on poly- ethylene, poly(viny1 chloride) and poly- urethane, containing precisely defined quantities of a variety of common addi- tives ( e .g . , antioxidants, stabilizers, ultra- violet absorbers, etc.) have been prepared under carefully controlled manufacturing conditions. These reference materials are presently undergoing complete character- ization as outlined below. Material Homogeneity and Stability Assessment of the homogeneity of the reference matrices is being undertaken in order to detect unexpected variability, e.g. , incomplete distribution of additives within the reference polymers or differen- tial contamination, etc. Furthermore, the stability of the reference matrices is being performed at periodic intervals to confirm that all values obtained for these bioma- terials are maintained from production until end-use. The effects of light, mois- ture, heat and time upon the reference matrices’ characteristics are being quanti- fied in order to provide advice on suitable storage location, packaging environment and lifespan (storage/expiry date).Biomaterial Characterization Surface composition is an important par- ameter in the consideration of the interac- tion between biomaterials and the host tissue. However, present methods do not allow characterization of the surface to a sufficiently high degree. Methodology, based on the measurement of the spread- ing of liquids on the biomaterial surface, has been developed within this project and provides a more sensitive interpre- tation of surface composition of the refer- ence matrices. The relative proportion of specific chemical groups within the sur- face of the biomaterials is determined more easily and with greater precision.Bulk composition is also important as many materials retain low relative mol- ecular mass manufacturing or processing agents, which might subsequently be released. Current standard methods, involving aqueous solutions or solvents, for extracting the additives from the reference matrices, frequently fail to achieve the desired criteria for an ideal extraction technique. Indeed, direct analysis using conventional spectroscopy or spectrometry is difficult and chromato- graphic analysis limited. However, research within this project has led to the more simple and rapid procedure of analytical-scale supercritical fluid extrac- tionkhromatography (SFEK) being employed very successfully both to ‘fingerprint’ the bulk reference matrices and to extract quantitatively any polymer additives.Biocompatibility Evaluation Biomaterials should exhibit negligible toxic responses to their cellular environ- ment and have minimal effects on several factors relevant to normal healing pro- cesses. Many methods exist at present that allow some degree of analysis of these effects. Two of the methods which have been investigated as potential standard test procedures are as follows. (In vitro) cell culture testing/screening, which has been utilized to determine the ability of the reference matrices to exhibit short-term toxicity to cells. The standard- ized tests have been based on the direct exposure of cells to bulk reference bioma- terial or to dilutions of extracts and/or degradation products of these bioma- terials.The cells were assessed using phase contrast microscopy or spectro- photometry. The results are currently being correlated with those obtained from in vivo studies. (In vivo) local tissue response to a biomaterial, which is recognized as being essential to biocompatibility evaluation because implantation results in alteration of tissue morphology. Traditional histolo- gical techniques have tended to be only subjective in nature. However, a number of novel immuno-enzymatic cell-specific staining techniques involving computer- aided image analysis have been developed by the University of Liverpool through an external research contract. These are enabling effective cell identification and counting at, or around, the site of implan- tation of the reference matrices. It is the intention to certify the reference biomaterials through an interlaboratory collaborative exercise and to establish several draft standardized procedures for evaluating biocompatibility . These are presently being disseminated widely to the UK national, and international, analytical and research community and presented to the relevant Standards Committees. Useful Contacts Reference materials and validated pro- cedures for evaluation of biocompatibi- lity: Dr. Julian Braybrook, LGC. (Tel. 081-943-7576.) For further details about the VAM Initiative, please contact Greta Carpenter, LGC. (Tel. 081-943-7393.)

ISSN:0144-557X    

DOI:10.1039/AP9933000290

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 291 Ronald Belcher Memorial Lectureship In recognition of the late Professor Ronald Belcher’s interest in education, the Council of the Analytical Division has instituted a Belcher Memorial Lecture to be given annually on an analytical topic by a graduate student. The award will be considered by the Honours Committee, acting on behalf of the Council of the AD. Students will be considered to be eligible if they are registered for an academic session in the year of the award. The decision concerning the award will be made in December. The aim of the award is to commemorate Professor Belcher as a teacher, by encouraging students to make a positive contribution to, and take an active part in, the profession of analytical chemistry. The paper, to be written by the student (see rule 2 below), should be in the form required for submission to ‘The Analyst’.Rules 1. Candidates must currently be regis- tered postgraduate students of a British University. 2. The merits of a particular candidate may be brought to the notice of the Honours Committee by any supervisor of postgraduate students registered with a British University or Polytech- nic who desires to recommend the candidate, by letter addressed to the president of the Division. The letter shall be accompanied by a paper (or manuscript ready for submission for publication) of which the student is a principal co-author. 3. The award shall be made annually in December and shall be based on an over-all assessment of the originality of the work described in the paper and the significance of its contribution to analytical chemistry. The winner of the award will be expected to present his or her work at the Research and Development Topics Meeting follow- ing the award.4. The award will take the form of a presentation scroll plus a sum of f300. The sum is to assist the candidate to attend a national or international con- ference. It will be given to the candi- date, up to two years after the granting of the award, on presentation of satis- factory evidence of the candidate’s intention to attend such a conference. 5 . An award shall not be made if it is considered by the Honours Committee that none of the papers submitted reaches the required standard. 6. The decision of the Council of the Analytical Division shall be final. 7. Any alteration to these Rules shall be subject to the approval of the Council of the Analytical Division. Submissions should be sent to the President, Analytical Division, The Royal Society of Chemistry, Burlington House, London WlVOBN. The closing date is Thursday, September 30th, 1993.

ISSN:0144-557X    

DOI:10.1039/AP993300291b

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

292 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 Analytical Viewpoint The following is a member of a continuing series of articles providing either a personal view of part of one discipline in analytical chemistry (its present state, where it may be leading, etc.), o r a philosophical look at a topic of relevance t o chemists in general or analytical chemists in particular. These contributions need not have been the subject of papers at Analytical Division Meetings. Persons wishing t o provide an article for publication in this series are invited to contact the editor of Analytical Proceedings, who will be pleased to receive manuscripts or to discuss outline ideas with prospective authors Charting to Success Within Analytical Laboratories Jill A. Gardner” Merck, Sharp and Dohme, Shotton Lane, Windmill industrial Estate, Cramlington, Northumberland NE23 9JU Shirley Coleman Industrial Statistics Research Unit, University of Newcastle upon Tyne, Newcastle upon Tyne NEI 7RU S.G. Farrow ICl Chemicals and Polymers Limited, Teesside Operations Analytical Services, ICl, Billingham, Teesside TS23 ILE This paper is the second in a series of papers, by J . A. Gardner, S. Y. Coleman and S. G. Farrow, discussing the need for and benefits resulting from using Statistical Process Control and allied statistical techniques within analytical laboratories. These papers are a direct result of the experience gained by J . A. Gardner working within ICI’s production support laboratories on Teesside. Introduction Teesside Operations Analytical Services (TOAS) are the analytical support laboratories for the ICI manufacturing plants on Teesside. Within TOAS’s 15 laboratories, analysis of raw materials, intermediates and manufacturing process end- products takes place on a 24-hour basis, 7 days per week, 52 weeks per year.It is estimated that from this analysis TOAS analysts produce over 7 million results per annurn. These results, for example stating the colour, concentration and impurity levels for a sample being tested, are transmitted to the manufacturing plant and in many instances are also given to the customer who purchases the end-product. From these figures adjustments are made both to the manufacturing plants and, in some cases, to the processes of the customers purchasing the product. Apart from passing these numbers to those requesting them, in many areas up until 2 years ago little was done by the analytical functions with the results.In fact, the precision of the results was not always known and little attention was paid to patterns or trends resulting from the data. Since January, 1991, much time and effort has been directed at encouraging and enabling the analysts within TOAS’s laboratories to make maximum use of the data that they produce. This paper discusses the many reasons for the use of charting and ‘control charts’ within analytical laboratories. In addition, it discusses the setting up of and rules for the charts, benefits resulting from having such charts and the propagation of charting within laboratories. * To whom correspondence should be addressed.Why Chart? There are many reasons why we should plot the data obtained from analytical methods. Just a simple plot of the data over time, for the results from a process end-product, can give useful information to both the laboratory and the plant regarding patterns and trends. Any such patterns or trends can indicate changes within the plant process or, in fact, with the analytical method. These will be due to ‘special assignable causes’ and investigations into such results should lead to the elimination of these causes and consequently to reduced variability and a more stable manufacturing process. Charts such as these, where data are merely plotted against time, as they are produced, and which are used mainly to indicate any patterns and trends, are called ‘Run Charts’.We can also produce ‘Control Charts’, which, as suggested by their name, are used to ‘control’ or monitor a process or analytical method. By using control charts we can monitor the variability of our analytical methods and in addition we have an objective means of deciding when to take action on an analytical process and when to leave it alone. Control charts can be used for many reasons, examples of which are as follows. To indicate when re-calibration is necessary, for example, with gas chromatographs. To indicate drifting of results, for example, as a result of standards decaying. To show problems such as instrument failure. To notify analysts of mistakes such as calculation errors. To indicate poor sampling. To give analysts confidence in analytical results when dealing with customer inquiries.To provide data to calculate the reproducibility and capability of analytical methods. Therefore, control charts can tell us when there are problems with the analytical method and, just as important, when there are not problems with the analytical method.ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 293 Setting up Control Charts In order to establish a control chart for an analytical method we firstly need to calculate the within laboratory reproducibility for that analytical method. This is done by running a stable check sample several times using ‘normal’ laboratory con- ditions; that is, using all the analysts within the laboratory who would normally perform the analysis, using any equipment that would ‘normally’ be used and doing the test at the frequency that it would ‘normally’ be done, e.g., once a shift.The standard sample used for this should be, as far as is possible, of a similar concentration to samples that are being received for analysis from the manufacturing process. Where there is a large variation in the concentration of the samples being sent for analysis it may be necessary to have more than one control chart in order to monitor the performance of the analytical method at the different concentration levels. By using the results from the check sample, the mean and standard deviation are then calculated and this .gives an indication of the variability of a randomly produced test result at this concentration. This standard deviation is the Within Laboratory Reproducibility Standard Deviation.Within TOAS we ask for a minimum of 29 degrees of freedom for reproducibility; that is, a minimum of thirty results when calculating these values. A line is then drawn on the chart at the mean value (this is sometimes referred to as the ‘target value’). ‘Warning Lines’ are then drawn onto the chart at f 2 a from the meadtarget line and ‘Action Lines’ are drawn onto the chart at +3a from the meadtarget line. These steps are summarized below. (i) Produce 30 data values for the method under ‘within laboratory reproducibility conditions’. (ii) Calculate the mean, 3, and the standard deviation, a, for these data. (iii) Draw a line at 3 (the mean or ‘target value’) on the graph. (iv) Draw ‘Warning Lines’ on the graph at R _t 2a.(v) Draw ‘Action Lines’ on the graph at R f 30. As an example: For an analytical method measuring the level of an impurity in an alcohol, the following results were obtained: Mean Impurity Level = 0.52 ppm; Standard Devi- ation = 0.08 ppm. The control chart shown in Fig. l was then drawn up. Using the Control Charts In order to use the chart we analyse the check sample, using the method, every so often and plot the points on the chart. Note that a change in mean level will not necessarily be detected immediately by a point going ‘out of control’. A small change will, on average, take longer to detect than a large change. Once the standard has been run and the value plotted, providing the response lies within k 3 0 from the target value, then the method is considered to be ‘in control’.However, if one value lies outside k30, two consecutive points are outside +2a or seven consecutive points are ascending or descending, then the process is deemed to be ‘out of control’. The latter rule is an additional rule as it had been found to be appropriate for the laboratories, where the nature of chemicals and solutions is such that they may deteriorate over time. P Pm 0.76 1 Upper Action Line 0.68 _ _ _ - __ __ __ - __ __ - _ - __ _ __ _ _ _ _____-__ Upper Warning Line 0.52 Target Value 0.36 _ _ - _ _ _ _ - _ _ _ _ _ - _____-_____ _ _ _ _ _ _ _ Lower Warning Line 0.28 Lower Action Line Fig. 1 Control chart for measurements of impurity in alcohol These rules are summarized in Table 1. _ _ _ _ _ _ _ _ _ ~ Table 1 Rules for control charts A method is ‘out of control’ if: (a) One point lies outside f3a from the target value, i.e., outside the action lines.(b) Two consecutive points lie outside +2a, the warning lines, or (c) Seven consecutive points are ascending or descending. N.B. Within TOAS owing to the nature of the substances with which they deal, i.e., chemicals, many instances are found where, although there are not seven consecutive points ascending or descending, there is a definite upward or downward trend made up of many points. In these circum- stances they rely on the experience and judgement of their analysts to decide whether or not the method is ‘out of control’. More rules could be applied to the control chart but this increases the complexity of using the charts and the occurrence of false alarms, i.e., cases where the method is ‘flagged up’ as being ‘out of control’ but the result is only a fluke result and the method is in fact working properly.How frequently the standard is run within TOAS depends on many factors. These include instrument availability, analyst time and the importance of the analysis, e.g., end product analysis for a product being sent to an external customer is highly important, as is analysis at the start up of process plants. Within TOAS, therefore, some check samples are run every time an analysis is carried out, whereas others are run on a daily, weekly or even monthly basis. Running the standard sample at a fixed time interval allows us to check the chart for ‘time effects’; however, this is not essential. When a process or method is ‘out of control’, action must be taken.The ‘action’ taken will depend very much on the analytical method being used. For a method using gas chromatography, recalibration of the instrument may be necessary; another action may be to check any calculations made during the experiment. There may be a fault with the instrument, an incorrect weighing or poor sample injection. A list of possible actions to be taken when a chart goes ‘out of control’ should be devised when a chart is being constructed. It is essential that the actions to be taken are established at the outset. Otherwise, on receiving an ‘out of control’ signal analysts may be inclined to ignore the signal or merely run the analysis through again and hope for a better outcome.It is also important that when an ‘out of control’ signal occurs and an action is taken, that the reason for it is noted on the chart. This ensures that in the future, when any action signals occur, analysts can look back at previous charts in an attempt to establish whether or not the same fault has occurred again. If a fault is constantly recurring, this may mean that a maintenance schedule for the machine can be established or that further training in the method is needed. Example. At one TOAS laboratory the recalibrations of gas chromatograph machines were changed from being on a fixed time basis to being determined by control charts. During the early use of the charts, on the occurrence of an action signal and after checking for machine faults and so on, the usual action was to recalibrate the machine.Gradually, as the analysts gained more knowledge and information about the technique and machines from the charts, it was realized that in many instances the method was in fact going ‘out of control’ because of deterioration of the GC liner. The chart shown in Fig. 2 was produced during August and September, 1992. During this time period, although the chart only indicated the method to be ‘out of control’ on one occasion, action was taken twice. On both occasions, after a downward trend in results had been noted, the GC liner was taken out of the machine, cleaned and replaced. This action was taken after points ten and twenty-two on the graph. The cleaning of the liner in this way enabled the machine to294 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 8.2 I 7.6 7.4 7.2 7.0 6.8 Q 6.6 - 6.4 I I l I 1 I I l I I L I I I 1 5 9 13 19 23 27 31 4 8 12 16 20 24 28 2 August September Date Fig.2 Control chart for gas chromatograph remain ‘in control’ for a greater length of time and meant that recalibration, which is lengthy, was not needed over this time period. The running of standards on a frequent basis has been found to be a much more effective and efficient way of determining the need for the recalibration of gas chromato- graphs, rather then recalibration at regular intervals. Why Charting Works The charts work for several reasons. The main one is that the rules for the chart are such that the probability of obtaining one value outside the action lines, two consecutive values outside the warning lines or seven consecutive points ascending or descending is very small.Therefore, should any of these events occur, the probability that the method is actually performing as it should do is also very small. Hence, there is a large probability that there is something amiss with the analytical method and it is therefore right to investigate and take action. Because the probability of a transgression of these rules is very small when the method is working as it should, the false alarm rate is also very small. Therefore, on most occasions when an alarm is raised it is possible to find a problem and hopefully to resolve it. This is very important as it makes the analysts realize the benefits of using the charts. Another reason for adopting just a few rules, against which action should be taken, is that the system of charting is made simple enough for people to remember and use quickly.It should be noted that 3.090 and 1.960 could be used instead of 30 and 20 and that these values are in fact more ‘statistically sound’. However, within TOAS we have found, once again in order to provide a ‘user friendly’ system, that it is better to use the latter values. The ‘one at a time chart’, which is mainly used within TOAS laboratories, is the simplest possible control chart to use. The use of a moving average chart has the advantage of faster detection of a change in mean level but requires calculation and the start-up values are difficult to use, as the chart is based on values calculated from the average of a certain number of points.There are several assumptions made regarding the data which are being analysed. These are that the data are normally distributed, data values are independent of each other and have constant variance. These conditions must be met for the charts to work effectively. However, even though the assump- tions are not always justified there is still considerable benefit to be derived from the charting. Continued Use of the Control Charts Once all of the initial check sample has been used the analyst is, of course, placed in a quandary. It should be noted that providing a check sample of a similar concentration to the initial one is used, the variability of the method for that sample can be assumed to be equal to the variability for the original.Hence a ‘new’ control chart should be drawn up using a ‘new’ target value but with the same estimate for standard deviation. To calculate the new target value, the new check sample should be run 5 times, the mean value calculated and the chart drawn up around this value. After the chart has been used for some time we may then decide to use all the values obtained to re- estimate the mean level and then once again draw up the chart based on this vahe. An alternative to basing a chart around the mean value is to use a standard of known concentration and set up the chart around this value. Once a new chart has been established the standard should be run as before. Previous charts should, if possible, be kept so that any ‘out of control’ occurrences can be benefited from and the data can be used for process capability studies.The standard deviation for an analytical method or process will not necessarily stay constant over time and eventually it may be necessary to recalculate this value also. For example, as an analytical instrument ages measurements made with it will not necessarily be as precise as when it was first introduced into the laboratories. On the other hand, the variability of the method in some instances may improve. This may be a result of improvements in the analytical technique, new apparatus or simply because of reduced variability brought about by using the control charts. Example. An autotitrator had been tested by a laboratory analyst with the intention of establishing that the method was robust enough to be used by process operators on the production plant.Initial tests to look at the variability of the over-all analytical method were followed by comparisons of analyses done in parallel by the analyst and the process operators. The above work showed that the variability of the method was reasonably small and that there was no statistically significant difference between the analysis when done by the operators or the analyst. A control chart was then drawn up based on the data collected for the comparison and this was used by the operators by running a standard on a daily basis. The control chart was as shown in Fig. 3. After a period of time it was noticed that there had been a reduction in the variability of the method, i.e., an improved standard deviation.This was believed to be due to the improved technique of the operators as they gained experience in using the instrument. A new control chart was then established using the recalculated standard deviation and the chart shown in Fig. 4 was produced over the next 30 days (see Fig. 4). The knowledge gained from using charts has been used not only within individual laboratories but has also been spread throughout the laboratories within TOAS. Another example where charting led to an improved situation is described below. Plant Trial, January-March 1992 38.0 37.5 37.0 8 36.5 2 36.0 35.5 35.0 34.5 I 1 t, - Fig. 3 Control chart for plant trial, January-March 1992ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 36.5 295 UAL . . . . . . . , . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . n UWL = _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ - _ - - - - _ - - - - - - - - - - - - _ - _ - - - Target 4- LWL 36.0 - $ 1 1 LAL 35.0 I ! 3 5 7 9 11 13 15 17 19 21 23 25 27 29 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Observation Fig. 4 Plant trial-new control chart Example. Within several laboratories on Teesside the same analytical method is used for determining total organic carbon content (TOC) in effluent. In one laboratory a chart had been established to monitor the variability of the method and to check that it was working correctly; charting had been taking place for several months. Another laboratory had just taken over the responsibility for TOC measurement in their area and it was suggested that they too have a chart to monitor the method.A chart was set up and after only a short time the points showed the method to be ‘out of control’. The action taken in this instance was to enlist the help of someone from the first laboratory, who, after looking at the chart, was able to establish immediately what the problem was and suggest the solution. Since then the chart has once more gone ‘out of control’ and again the same problem with the column had occurred. This time the analysts in the second laboratory were able to recognize and solve the problem themselves. This is a prime example of how the knowledge and benefit from using such charts can be capitalized upon. Charts can be incorporated into the over-all laboratory ‘Quality System’, as in the case where charts are used as a basis for recalibration of gas chromatographs.When customerlplant enquiries are received regarding the validity of an analytical measurement or the analytical method, charts can be used both to validate that the method is ‘in control’ at that moment and also to assess the capability of the analytical method. This gives an indication of how well the analytical process is performing over time when compared with the specification for the product being analysed. One can also use the data from the charts to calculate the over-all variability for the analytical method, which can then be compared with the variability of the over-all plant process. Whither the Charts? The benefits from charting are numerous, but how does a laboratory go about starting the charting process? The most important stage is, of course, the beginning.It is essential that in order to begin charting people are trained in the relevant statistics. This training, including how to calculate means and standard deviations, how to set up charts and the benefits from using them, need not take a large amount of time. Initially, charting should be done on one or two selected areas or methods and should be done on paper. The choice of these initial projects is also critical. Within TOAS, as there are several thousand analytical methods, we are prioritizing which methods to consider first. Initially, projects are being chosen which fit into one of the following categories. Firstly, analytical methods where customers are asking for statistical data, on either the plant process, the analytical method or both, to validate the product being purchased.Secondly, analytical methods performing analysis for plant processes where there are problems with the plant process owing, for example, to large variability, results being outside specification limits or results being unusual. Thirdly, analytical methods where there may be problems with the method, for example, due to large variability. By prioritizing our effort in this manner we are trying to ensure that maximum benefit is gained from the initial work and, as a result of this, we hope to encourage analysts both to carry on with the charting and to extend this work. Once the use of statistics is firmly established and understood we can then progress to moving the charting on to a computer.This we have found to be essential in some areas where, because of the large numbers of charts, it would take a substantial and unreasonable amount of time to draw up every chart and keep them maintained. Suitable computer packages can be found where it is possible to enter data and then ‘at the touch of a button’ produce all types of control charts, summary statistics, capability figures and so forth. It should be emphasized that laboratories should only proceed to using computers for the- purpose of charting and statistical analysis after analysts have ’been trained in the appropriate statistics and have got a ‘feel’ for the process using pencil and paper techniques. In the future we hope to have systems where the production of data from the analytical instruments is connected to a computer containing the relevant software such that, again, ‘at the press of a button’ the appropriate statistics and charts can be produced.This will save in both time and in the occurrence of errors when handling data. An Analytical Data Working Party has been formed, containing representatives from all laboratory areas within TOAS to promote further the work that is already being done and to encourage the uptake of charting throughout the fifteen laboratories. Members of the working party are being encour- aged to share their most ‘useful’ charts with other areas. Similarly, laboratory managers are being encouraged to take an active interest in the charts to promote further their uptake and to ensure that quality system requirements are being met. In January, 1991, when this work started, there were very few charts within TOAS laboratories: now there are over 400. The use of charting in laboratories here has proved to be an invaluable way of increasing analytical confidence and the knowledge of analytical methods. Only by studying the data being produced by analytical techniques can we begin to understand our analytical methods. Charting makes data available in such a way that it is much easier to understand, so that it is possible for the people who are producing the data to have confidence in and understanding of their analytical results. As well as the benefits already discussed, including greater knowledge and confidence in the analytical method, there are also substantial monetary gains to be made from using control charts and other statistical techniques to study analytical methods. These and other benefits will form the foundation of the next paper in the series.

ISSN:0144-557X    

DOI:10.1039/AP9933000292

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

296 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 Report by the Analytical Methods Committee Evaluation of Analytical Instrumentation. Part Vlll Instrumentation for Gas-Liquid Chromatography Analytical Methods Committee The Royal Society o f Chemistry, Burlington House, Piccadilly, London W I V OBN The Analytical Methods Committee has received and approved the following report from the Instrumental Criteria Sub-committee. Introduction This report was compiled by the above Sub-committee of the AMC which consisted of Professor S. Greenfield (Chairman), Professor E. Bishop, Dr. P. J. Potts, Mr. D. C. M. Squirrell, and Mr. P. Warren, with Mr. C. A. Watson as Honorary Secretary. The initial input of the features for consideration and the reasons for their consideration was undertaken by a working party Chaired by Dr.R. M. Smith with Mr. M. Humphrey and Dr. K. D. Bartle, to whom the committee express their thanks. The purchase of analytical instrumentation is an important function of many laboratory managers, who may be called upon to choose between a variety of competing systems which are not always easily comparable. The objectives of the Instrumental Criteria Sub-committee are to tabulate a number of features of analytical instruments which should be con- sidered when making a comparison between various systems. As explained below, it is possible to score these features in a rational manner, which allows a scientific comparison to be made between instruments. The overall-all object is to assist purchasers in obtaining the best instrument for their analytical requirements.It is hoped that, to a degree, it will also help manufacturers to supply the instrument best suited to their customers’ needs. It is, perhaps, pertinent to note that a number of teachers have found the reports of use as teaching aids. No attempt has been made to lay down a specification. In fact, the Committee considers that it would be invidious to do so: rather, it has tried to encourage the purchasers to make up their own minds as to the importance of the various features of the equipment that is on offer by the manufacturers. The eighth report of the Sub-committee deals with Instru- mentation for Gas-Liquid Chromatography. Notes on the use of this Document Column 1. The features of interest. Column 2. What the feature is and how it can be evaluated.Column 3. The Sub-committee has indicated the relative importance of each feature and expects the users to decide on a weighting factor according to their own needs. Column 4. Here the Sub-committee has given reasons for its opinion as to the importance of each feature. Column 5 onwards. It is suggested that scores are given for each feature of each instrument and that these scores are modified by the weighting factor and sub-totals obtained. The addition of the sub- totals will give the final score for each instrument. Notes of Scoring 1. (PS) Proportional scoring. It will be assumed, unless otherwise stated, that the scoring of features will be by proportion, e.g. ? Worst/O to Best/100. 2. (WF) Weighting factor. This will depend on individual requirements. An indication of the Sub-committee’s opinion of the relative importance of each feature will be indicated by the abbreviations VI (very important), I (important) and NVI (not very important).A scale is then chosen for the weighting factor which allows the user to discriminate according to needs, e.g., x 1 to x 3 or x 1 to x 10. The factor could amount to the total exclusion of the instrument. 3. (ST) Sub-total. This is obtained by multiplying PS by WF. Gas chromatography is a well established analytical technique with applications in many areas. An often bewildering range of instrumentation is available from well over a dozen different manufacturers. Systems range from simple instruments, with a single column and detector, to complex multi-channel systems with autosamplers and microcomputer-based controllers for continuous operation and sophisticated mass spectrometric or FT-infrared spectroscopic detectors.Selection of a suitable instrument for purchase is, therefore, not an easy task and the purpose of these notes is to provide some guidance to areas which should be considered so that the choice is based on a full consideration of the available options. However, the performance of any gas-chromatographic method depends primarily on the separation conditions and thus on the nature of the column material, stationary phase and carrier gas and whether a packed or open-tubular column is used. The type of detector will also influence the sensitivity and selectivity of the assay. A number of alternative instruments may thus be suitable. The first task in the selection of an instrument is to examine the range of analyses that it will be expected to perform.Care should be taken not to specify these requirements too closely as uses change with time. The analytical scientist should also not try to envisage every potential application or the selection criteria may become too detailed. The choice of the column and detector type are outside the scope of these guidance notes but any specific requirements should be noted, such as special detectors, injectors or accessories. With these requirements in mind, the user should then evaluate the instruments available on the market while bearing in mind the guidelines and any financial limitations. In many instances it will quickly become clear that a number of different instruments could be satisfactory and non-instrumental criteria may then be important.However, in some specialized cases only one or two instruments will have the ability or necessary features to carry out the assay. The guidelines are intended to be used as a check list of features to be considered, mostly of the instrument itself, but some also of its service requirements and of the relationship of the user with the manufacturer. Their relative importance will depend on the installation requirements of the instrument as well as the uses to which it will be put. Therefore, to some extent, the selection process will inevitably be subjective, but if all the points have been considered it should be an informed choice.In addition, because a separation depends so much on the column and operating conditions, it may sometimes be difficult to assess the actual operating performance of a particular feature from the manufacturer’s specifications. Few standardANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 297 tests have been developed for gas chromatography columns and for some applications it may be necessary to evaluate the performance of the column using the instrument under consideration. Gas chromatographs are often sold as complete units, so that compromises between features may have to be accepted, but it will still be important to distinguish between critical features and those which are optional. The Committee consider that, in general, gas-chroma- tography equipment is safe in normal use, but care should be taken to allow sufficient cooling time when changing columns and detectors and taking suitable precautions when handling inflammable gases such as hydrogen.If hydrogen is used as a carrier gas it is recommended that a suitable leak detector should be fitted in the oven. Finally, as many laboratories are now working to quality standards such as GLP/NAMAS/IS09000, some consideration should be given to third party recognition of the manufacturer to standards such as BS5750 and I S 0 9000. Such accreditation should extend to the service organization, which is particularly important when working to NAMAS or GLP criteria. INSTRUMENTAL CRITERIA SUB-COMMImEE INSTRUMENT EVALUATION FORM Type of Instrument: Gas chromatograph Manufacturer: Model No: Feature NON-INSTR U- MENTAL CRITERIA Selection of manufacturer ( a ) Previous in- struments (i) Innovation (ii) Reliability record (iii) Up-grad- ing com- patibility.Inter- change- ability of column and detec- tors (iv Similarity of layout and design to instru- ments existing in laboratory (v) Confidence in supplier ( b ) Servicing (i) Service contract (ii) Availabil- ity and de- livery of spares Definition and/or test procedures and guidance for assessment Laboratories in possession of other gas chromatographs should score highest for the manufacturer with the best past record based on the following sub-features: Company's record for developing instruments with innovative features. Company's record for instrument reliability. The ease with which columns, injectors, pneumatic modules and detectors can be changed between different instruments.Confidence gained from past personal experience. Score according to manufacturers claims and past record, judged by the sub- features (i)-(iv) below: The availability of a suitable service contract from the manufacturer or agent. Range of stock carried by the manufacturer and delivery time. [mportance I VI Reason The manufacturer should be alert to developments in technology and chromatography. Indicates history of sound desigdmanufacturing concepts. Shapes of packed columns often differ. Common column fitting methods allow columns to be transferred between instruments, giving greater flexibility. Open tubular columns are usually interchangeable. If detectors have the same base and electrical connections as existing instruments, the user can interchange detectors between instruments.Also common spares and components such as amplifiers can be interchanged. Similarity of layout means that operators can draw on in- house expertise, resulting in reduced training costs and time. It can also maximize the use of spares and fittings. Good working relationship already in place. Suggests long commitment to user. Often ensures preferential service and can guarantee a specific response time to call-outs. Rapid delivery of spares reduces downtime. Score PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST298 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 Feature ~~ (iii) Call-out time (iv) Effective- ness of service en- gineer ( v ) cost of call-out and spares (c) Technical Support (i) From Appli- cations Depart- ment (ii) Technical literature (iii) Telephone assistance INS TR UMEN TA A CRITERIA I .General feature. (a) Facilities re- quired for: (i) Location of connections and controls on instrument (ii) Snap-on coded fit- tings on gas lines (iii) Power and heat dissi- pation (iv) Dimen- sions 2. Carrier and detector gas supplies (a) Flow control (i) Flow con- trollers ~ _ _ _ _ _ _ _ _ Definition and/or test procedures and guidance for assessment The time for an engineer to reach the laboratory following a call. The ability of the service engineer as judged from previous experience and reports of others, including the carrying of adequate spares. It may be inappropriate to score this feature. As in ( b ) score in consideration of sub-features (i)-(iii) below.The advice and training available from the manufacturer’s applications department. The range and quality of technical literature including the operating manual. Willingness of the manufacturer to give effective advice on problems over the telephone. This can normally only be evaluated by reference to existing users. Score according to convenience, taking into account the proposed location for the instrument. Score for provision of snap on fittings and clarity of coding. The ability to remove heat particularly during oven cooling cycles. Score highest for instrument most suited to proposed location. Score according to compatibility of dimensions (width and depth) to space available.Score maximum for a flow control system which gives the most precise and reproducible gas flows. This may be tested using a bubble flow meter. Score maximum for the provision of the most precise mass-flow controllers. Score additionally if the controller is computer compatible. [mportancc VI for neu user VI VI Reason Keeps laboratory in operation by reducing down time [see also (i)]. Ability to repair on-site avoids return visit or removal of equipment to supplier and so reduces service time, costs and downtime. The proximity of service centre may be a factor in tra- vel costs. This helps in-house staff with new application problems. Guidance on optimum use of instrument suggests manufacturer’s awareness of applications. Rapidly available technical help reduces the number of call-outs. Depending on bench positioning and layout, these may limit accessibility for servicing and installation, particularly at rear of instrument.Coded supplies should be more secure and reduce risk of incorrect connections. Needed because the oven generates considerable heat during operation. Air conditioning or ventilation may be needed. Depending on the design of the oven exhaust vent an area of bench may need to be left free for heat dissipation. Availability of suitable bench space. This may be important in some circumstances. Control is needed to give a constant gas flow, upon which precision and reproducibility are dependent. Used for packed and wide- bore open tubular columns. Score PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF STANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 Feature (ii) Pressure controllers (iii) Con- trolled tempera- ture ( b ) Carrier gas (i) Ability to use hydro- gen or helium (ii) Carrier leak detec- tor supply (iii) ‘Make up gas’ (c) Connections ( i ) Gas supply lines (ii) Gas puri- fiers 3. Injection ports (a) General (i) Ease of cleaning (ii) Replace- able liners (iii) Septum replace- ment (iv) Septum purge ( b ) Heaters heater control (i) Injector Definition and/or test procedures and guidance for assessment Score maximum for the provision of the most precise pressure controllers.Score additionally if the flow controllers are contained within a controlled temperature environment. Score according to availability of sub-features (i), (ii) and (iii) if capillary columns are to be used.Score according to availability and ease of fitting of non- permeable gas lines. Score according to availability, stated efficiency, and ease of fitting of on-line traps, such as activated carbon or molecular sieves, into gas supply lines. Score according to ease by which units can be dismantled and reassembled for cleaning. Score for provision of replaceable liners for injection ports. Score according to ease of removal and replacement of septa. Score additionally for provision of bleed of carrier gas from just below septum. Score highest for most stable control of temperature of injector heater unit. [mportance VI I I VI VI I VI I VI VI I I Reason Used with open-tubular columns. For either gas control system, digital control is generally more precise and reproducible than manual control.This is particularly desirable if the settings will be changed frequently. A controlled temperature environment can increase the stability of the system. Needed for capillary columns as they give higher efficiency and faster separations than nitrogen. Needed if hydrogen is used as carrier gas to avoid build-up of gas in oven in the case of a leak. Needed to maintain a suitable gas flow through the detector when capillary columns are used. Nylon lines are porous to air and should not be used with an electron capture detector as air causes a high back- ground. They are also not recommended for the nitrogen-phosphorus detector (NPD). Nylon can age if exposed to sunlight.Metal tubing is more robust. Removal of oxygen and water from carrier gas is essential for ECD detectors and desirable for some sensitive stationary phases or for operation at high sensitivity. Oil may also need to be removed from air lines if a compressor is used. The need to remove involatile residues. Replacement of liners removes involatile residues and reduces contamination. Frequent changes of septa are necessary for satisfactory operation. Removes volatiles arising from degradation of the septum and reduces background peaks. Particularly needed for temperature programmed separation on open-tubular columns. Injector temperature can affect volatilization and sample stability on injection. Score PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST 299300 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 Feature (ii) Programm- able cool- ing/heating of injection zone (iii) Indepen- dent tem- perature control (v) Heater tempera- ture range (c) Types of in- jection port (i) Capillary or open- tubular column injectors (ii) Gas sam- pling valves 4.Column ovens Oven design Oven tem- perature ) Hysteresis (ii) Maximum tempera- ture (iii) Thermal fuse/elec- tronic cut- out (iv) Near-am- bient op- eration Definition and/or test procedures and guidance for assessment If capillary or open tubular columns are to be used, score for availability of a programmable unit for heating/ cooling the injection zone. Score additionally for provision of a separate oven for the sample injection ports.Score maximum for widest range normally provided. For capillary (open tubular) columns score for provision of both split/splitless and on- column injection facilities. Score additionally according to ease of changing ports and provision of manual or automatic operation. Fixed volume loops which can be switched into the carrier gas line. Score according to availability and ease of connection. General preference should be for a size of oven which will accommodate two columns with adequate access. Score accordingly. Score maximum for smallest temperature lag during heating and cooling cycles. Most ovens operate satisfactorily up to 350 "C. If higher temperatures required score additionally if oven will operate up to 450 "C. Score for provision of thermal fuse or electronic cut-out to turn off oven heater in case of controller failure.Most ovens operate satisfactorily down to 10 "C above ambient. For the examination of volatile samples score additionally for satisfactory control at near ambient temperature. [mportance I I I VI I :depending on appli- cation) I I VI VI I Reason Can be used in split/splitless injection (particularly important for capillary chiomatography) to focus sample and in on-column injection to raise rapidly the temperature after start of run. Temperature of the injection ports should not be altered by programming of column ovens if repeatable subsequent injections are to be obtained. Normal applications need up to 350 "C. Some high temperature applications may need up to 450 "C. Choice enables wider range of analytes to be examined.Split-Mainly used for samples with limited volatility range. Wide ranging samples may suffer some discrimination. Need ability to set split ratio for quantitative results. On-column-Needed for thermally sensitive samples and to avoid discrimination effects in wide boiling range samples. Needed for gaseous samples as syringe injections can give poor quantification. Usually only one column is used for open-tubular separations but with packed columns two columns are often used to provide column bleed compensation.Only one column is needed if electronic background compensation is available. Sufficient space is required to enable work in oven to install and replace columns. Slow response can limit programming and cooling rates and prolong re- equilibration time.Some high temperature separations using special columns may need up to 450 "C. Safety device to protect columns (and injector) from excessive heating. Enables repeatable analysis of volatile samples. Some ovens have minimum usable temperatures for reproducible control. Score PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF STANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 301 Feature (v) Sub-am- bient capa- bility (vi) Tempera- ture gradi- ents across oven (c) Oven pro- grammers (i) Tempera- ture/gradi- ent settings (ii) Number of steps avail- able (iii) Heating rate (iv) Extermal control pro- gramme (v) Reprodu- cibility of pro- grammed tempera- ture ( d ) Column installation (Column materials (i) Column fit- tings (ii) Inter- change- ability be- tween open tubular and packed columns (iii) Ability to use wide- bore col- umns 5.Detectors ( a ) Detector tY Pes ( b ) Availability Definition andor test procedures and guidance for assessment If required, score additionally for availability of an add-on cooling system. Score maximum for minimum temperature gradients within oven. Score highly for provision of digital control rather than analogue control for temperature gradient settings. Score according to the numbers of separate delay periods and temperature ramps that can be programmed. Score most for the highest maximum ramp rate that the oven can follow. If an external computer or data system is likely to be used, score for ability to control oven temperature by these means.Score highly for good reproducibility in temperature control on resetting programme. .mportance I I I I (for com- plex sam- ples) I VI v 1 Reason Needed for some gas samples and for highly volatile samples. Gradients due to poor air mixing in the oven can produce poor peak shapes with open-tubular columns because of their low thermal mass. Packed columns are less sensitive to the effects of thermal gradients. Digital controlled programmes are easier to reproduce. The more steps available, the greater the flexibility. Desirable for complex samples, particularly to flush off involatiles. Most samples will only need a limited number of steps in the programme. High rates are needed for cold on-column injections or with short columns.Also needed when chromatograph is linked to autosampler to co- ordinate injection with temperature programme and data collection. A computer can also vary conditions to match mixed samples in autosampler. Programme reproducibility is more important than accuracy. Needed to ensure consistency of results. nd stationary phases are outside the scope of this evaluation.) Score according to ease of changing columns. Score for the ability to interchange between open- tubular, packed and wide-bore columns. When narrow open- tubular columns are to be used, score additionally for minimum dead volume in detector. If the application may call for the use of wide-bore columns, score additionally for the provision of this feature. In most instances detector selection is circumscribed by the analytical method. Detector selectivity may be needed to distinguish between analyte and matrix. If appropriate, score maximum for highest number of detectors which can be fitted on standard instrument.I I NVI I Self evident. Gives maximum flexibility in use of system, bearing in mind that most instruments are dedicated to one mode. Presence of large dead volumes in detectors degrades separation efficiency and may cause discrimination effects. In some applications, wide- bore columns may produce better separations than packed columns. Enables a wide range of applications; increases versatility. Score PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST302 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 Feature ( c ) General detector and amplifier features (i) Linearity of response (ii) Sensitivity (iii) Dead volume (iv) Amplifier response time (v) Autorang- ing ampli- fier (vi) Dual detector channels ( d ) Kathar- ometer (hot wire) detec- tor (e) Flame ionis- ( i ) Ease of cleaning ation detector (ii) Ease of lighting (iii) Internal dead volume v> Electron cap- ture detector (i) Sensitivity to test sam- (ii) Resistance to contami- nation ple (iii) Carrier gas options Definition and/or test procedures and guidance for assessment Score maximum for widest linear dynamic range. For the major analytes of interest, score maximum for the best signal to noise ratio.Score maximum for lowest dead volume of detector cell and connections. Score maximum for the amplifier with the shortest time constant.These automatically switch gain according to signal, avoiding saturation of the data system. Modern instruments and data systems are normally compatible so it may be inappropriate to score this item. Score for provision to compare output signals from different columns if temperature programming is envisaged. Score maximum for the detector with the lowest internal dead volume. Score maximum for a design which best satisfies sub-features (i)-(iii). Score maximum for best sensitivity to test compound (usually lindane). Avoidance of cold or hot spots in detector which cause analyte deposition or degradation, respectively. It is difficult to score this feature except on reputation/experience . Score according to provision to use nitrogen and/or argon- methane.Importanct VI VI VI VI I I VI VI I I Reason Wide linear dynamic range facilitates an extended calibration range for quantitative studies. For ECDs this is usually greater in constant current mode than constant pulse rate mode. High signal to noise ratio permits increased sensitivity and enables smaller quantities of analytes to be detected. To minimize band broadening when used with capillary columns. Make-up may be needed to reduce effective volume. Peaks from capillary columns will be broadened or merged if response is not rapid. This is less important for packed columns. Excessive amplifier output can saturate A-D converter in some systems. Useful to compensate for changing background signals during temperature programming. Maximum performance will be found for detectors with the lowest ‘dead volume’.This is particularly important for use with capillary columns. Access and disassembly needed for routine maintenance. Electronic ignition reduces danger of contamination. Needs to be minimal to avoid band spreading from capillary columns at low carrier gas flows. Can be reduced by passing open-tubular column through jet to base of flame. This also reduces decomposition on metal surfaces of detector. Guide to over-all sensitivity of detector . Contamination gives increased background and reduces standing current. Because of radiation hazards ECDs are difficult to clean. Appropriate selection of carrier gas can alter sensitivity. Score PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF STANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 303 Feature (iv) Pulsing modes of controller (vi) Equili- bration time (8) Thermionic ionisat ion detector (nitrogen- phosphorous detector) (i) Ease of mode selec- tion (ii) Discrimi- natior.(ii) Sensitivity (iv) Lifetime of source ( h ) Flame photo- metric detec- tor (i) Linearizer built into amplifier (ii) Selectivity of filters (iii) Capability of simulta- neous detection of sulfur and phos- phorus 6. Auto Sampler ( a ) General (i) Compati- bility with chromato- graph Definition and/or test procedures and guidance for assessment Score maximum for ability to use both variable pulse rate and constant current modes. Score maximum for shortest equilibration time before stable standing current is obtained.Score according to ease of mode selection, usually by switching electronic configuration and changing hydrogen flow-rate. Score according to proven selectivity to different groups of compounds. Score maximum for highest sensitivity to test compound such as malathion. The time before there is a significant change in response. Scoring of this may be difficult, unless the manufacturers data is accepted or previous experience with the detector is available. Score for the provision of a linearizer built into the amplifier or data processing system and for evidence of satisfactory performance with test compounds. Score according to efficiency of filters to discriminate between emission signals due to sulfur and phosphorus and interference from the emission resulting from carbon compounds.Depending on proposed use, score additionally for a dual detector capability permitting simultaneous sulfur and phosphorus determination. Where the number of analyses justifies automated operation, score the capability of the recommended autosampler according to the availability of sub-features (i)-(iv) listed below. Ability of auto sampler to control integrator and chromatograph programmes. .mportance I I NVI I VI I VI (for sulfur :ompounds VI I Reason Can alter linear range and sensitivity. Extended equilibration times reduces availability of instrument and prolongs start- up time. Different modes can alter selectivity or specificity to analytes containing P, N or other elements. Can make system more specific or selective.Guide to sensitivity and selectivity. Some source materials may age over a relatively short time, which may be dependent on temperature and causes a reduction in response and discrimination. Sulfur compounds give a non- linear response with this detector. Determines the over-all selectivity of the detector. Increases analytical productivity, e . g . , for the analysis of oils and pesticides. Importance is application dependent. Needed for reliable automatic operation and data collection. The use of the chromatograph manufacturer’s auto-sampler should ensure compatibility with instrument. Score PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST304 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 Feature (ii) Inter- change- ability be- tween sam- ple injec- tion ports (iii) Carousel control (vi) Carousel tempera- ture (b) Injection system (i) Injection volumes (ii) Minimum sample size (iii) Sample carry over (iv) Injection modes available (v) Needle residence time (vi) Automatic one-shot injection (vii) Manual injection (viii) Multiple injections ( c ) Auto sample controller Definition and/or test procedures and guidance for assessment Ability of autosampler to inject into each of the available ports.Number of sample positions. Temperature control of sample awaiting injection. Score maximum for greatest range of injection volumes that can be programmed. Score maximum for system requiring minimum amount of sample in vial to flush needle and make injection.Score highest for most convenient mode of operation avoiding presence of residue which can contaminate next injection. Score additionally if needle wash is available. Score for provision of hot needle-cold needle injection and for on-column injection. Score maximum for shortest needle residence time in injection port. Score for facility to inject single sample automatically. Score for facility for manual injection without removing auto injection. Score for facility to make a number of injections from a single sample vial. Score according to ability to programme operation of auto sampler. Score additionally if control can be effected via an external computer and if different conditions can be used for specified samples. [mportancc ~ I VI VI I I VI VI VI I Reason Needed in dual-column instruments so that either column position can be used.More sample positions means that longer sets of runs can be carried out unattended. Permits pre-column derivatization or cooling for thermally labile or volatile samples. If different volumes can be programmed for each injection this increases versatility so that different levels of analyte concentration can be handled. Sample size may be limited. Amount can be dependent on position of needle tip in vial and hence shape of vial. Avoidance of cross contamination; alternatively, intermediate blank samples may be needed which will increase analysis time. Use of hot or cold needle and on-column injection can alter discrimination and reproducibility. To minimize sample degradation of sensitive samples within the needle during injection.Can be used to increase repeatability compared with manual operation, particularly by semi-skilled personnel or multiple users. Improves flexibility of apparatus for non-routine samples. Used to obtain repeatability measurements. Improves versatility of operation. Enables use of wash solutions on repeating standard reference solutions within run. Score PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF ST PS WF STANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 305 Feature 7. Data handling The selection of a data system is out- side the scope of this study as many soft- ware packages are available for data han- dling which are interchan- geable be- tween personal computers.Providing the instrument can output the data to a suitable computer, it should not affect the choice of the instrument, so scoring is therefore inappropriate. However, the following sub- features should be taken into consideration. ( a ) Dedicated system (6) Interfacing (i) Connec- tions requirements (ii) Control links 8. Additional features and accessories These features may be re- quired for specific ana- lytes or appli- cations, and enquiries should be made as to the availability of suitable acces- sories. These features should only be scored when appropriate. switching ( a ) Column Definition and/or test procedures and guidance for assessment Facility to carry out data collection, integration and evaluation supplied with the instrument. Standard connections/RS232/ parallel.Ability of data system to control chromatograph. Ability to transfer eluent gas flow to second column and to reverse flow through column. mportance Reason These systems are fully compatible with the chromatograph but may require an interface for data storage or more complex manipulation. Data system can be linked to external computer (or LIMS) using standard protocols. Required if there is a need for the external computer system to be able to control one or more chromatographs. Can be used to facilitate heart-cut and back flush methods for complex samples. Score PS WF ST PS WF ST PS WF ST PS WF ST306 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 Feature (b) Headspace sampling unit (c) Pyrolysis unit (d) Connections to mass spec- trometer of FTIR spec- trometer Thermal de- sorption unit ~ 9.Value for money (points per 2) Definition and/or test procedures and guidance for assessment Extraction and injection of headspace vapour from above solid, liquid or complex matrices. Ability to thermally degrade sample rapidly in inlet carrier gas flow. Ability to link the chromatograph to the spectrometer with heated transfer line. Ability to interface the chromatograph to thermal desorption unit. Sum of the previous sub-totals divided by the price of the instrument subject to proportional scoring and weighting factors, including ST in grand total. Importance I Reason Used for volatile analytes in comparatively less volatile matrix. Used to analyse involatile samples as characteristic volatile fragments. Used to give additional structural information about separated analytes. Thermal desorption units are often required for the analysis of volatile toxins in environmental air samples. ‘Simple’ instruments are often good value for money, whereas those with unnecessary refinements are more costly. Score PS WF ST PS WF ST PS WF ST PS WF ST Sum 01 totals sub- PS WF ST Grand Total A unique database of atomic spectrometry reference information for the practising analyst JAASbase is a new PC-based product from the Royal Society of Chemistry designed to meet every atomic spectroscopist’s need for a comprehensive, yet inexpensive source of current analytical atomic spectrometry information. It contains over 20,000 regularly updated references compiled from the atomic spectrometry literature. The database consists of listings of published atomic spectrometry papers and conference papers, and includes tabulated information relating to the application of relevant techniques. The references are easily searched with the database manager Idealist which also enables the addition of personal data to the database. Subscription Details JAASbase 1993 Updates f99.00/$218.00 JAASbase Backfile (1987-1992) f230.00/$506.00 Idealist Software &21O.O0/$462 .OO Six updates will be issued at regular intervals through 1993. JAASbase is an invaluable tool for all practising analysts - order your copy today! To order JAASbase and for further information please contact: Sales and Promotion Department, Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge CB4 4WF, United Kingdom. Tel: +44 (0)223 420066. Fax: +44 (0)223 423623. Telex: 818293 ROYAL. ROYAL SOCIETY OF CHEMISTRY Intormation ServlcPs

ISSN:0144-557X    

DOI:10.1039/AP9933000296

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 2.8 (b) 1.2 0.8 307 Chemometrics Papers The following papers were originally presented at a Meeting of the United Kingdom Chemometrics Discussion Group. From Feedback to Chaos in Chemical Systems Alison S. Tomlin" School of Chemistry, University of Leeds, Leeds LS2 9JT Chaos has undoubtedly been one of the most exciting topics in science over the past decade. It stretches across many disciplines and its visual aspects have captured the imagination of a large proportion of the general public. For physical scientists, however, some questions about chaos remain unanswered. Mostly there is uncertainty about whether the strange dynamics found under certain experimental conditions are a result of the internal mechanisms of the system or are simply an artefact of insensitive experimental control.Further, the type of chemical systems which can support chaos, and why, is still not fully understood. Chemists would like to know what ingredients, when put together, will lead to chaotic solutions. To aid this understanding it is important to model experimental systems which show chaotic type behaviour and to investigate the underlying mechanisms which produce chaotic dynamics. Three key conditions are said to be important: ( I ) far from equilibrium conditions: the thermody- namic equilibrium is stable and unique so once a solution has reached equilibrium it will remain there. Far from equilibrium conditions are achieved in flow or pool chemical systems; (2) non-linearities: a linear system cannot show chaos, but chemical systems are inherently non-linear ; (3) feedback mechanisms: the origins of feedback mechanisms are discussed below.Feedback Mechanisms Chemical feedback in the form of chain-branching or autoca- talysis can give rise to oscillatory behaviour in very simple chemical systems involving only two variables. The oscillations in such cases are always simple, period-1 solutions. After transients have decayed, each excursion has exactly the same period and amplitude as the next. There is, however, a growing body of experimental evidence in chemistry for more complex, higher-order periodic (bursting) patterns and even aperiodic (chaotic) states. Examples include the solution-phase Belousov-Zhabotinskii reaction, gas-phase combustion systems, heterogeneous catalysis and electrodissolution reac- tions.The need therefore arises to find simple models which can mimic this complex behaviour so that its underlying mechanisms can be easily analysed. It can be shown that the coupling of a second periodic forcing to an already oscillatory system can provide the key to understanding such complicated dynamics. This forcing can take very different forms. It could be applied externally via the periodic forcing of experimental parameters such as ambient temperature or species inflow rate. More importantly it can arise through internal mechanisms such as second feedback routes from autocatalysis or from the thermal characteristics of the reaction. Many chemical reactions are exothermic and this * Present address: Department of Chemistry, Princeton University, New Jersey 08544-1009, USA.chemical heat release can give rise to self-heating and hence to thermal feedback, where the temperature and the concen- trations vary. When chemical and thermal feedback are coupled, the range of responses which can be observed in a system can be increased dramatically. These effects can be studied using a simple prototype scheme involving the kind of feedback mechanisms which commonly occur in realistic chemical mechanisms. For example, the following mechanism contains the competition between auto- catalytic and first-order conversion of the species A to B. It contains highly non-linear terms. P - + A rate = koP (1) A + 2B -+ 3B rate = kiab2 (2) 1 .o I[ Time (T) Fig. 1 Isothermal model. ( a ) Bifurcation diagram; (6) time trace (p = 0.8); and ( c ) time trace (p = 0.7)308 t ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 0.3 - (" 0.3 m E - .f - N I, ( d ) - 200 500 200 500 200 Time (T) 500 Fig. 2 Mixed mode time traces from non-isothermal scheme A - + B rate = k3a ( 3 ) so the concentration does not effectively change with time. The system never reaches thermodynamic equilibrium in this 'pool B -+ C + heat chemical' case and so sustained oscillations are a possible type of response. Using these assumptions the set of dimensionless rate equations reduce to a two-dimensional set of ordinary rate = k2b (4) Simple Chemical Feedback differential equations: In the initial analysis it is assumed that the rate constants are dddx = ~1 - aUp2 - dp/dT = cup2 - %,a - p (5) (6) not temperature dependent and so the heat produced in the termination step is not fed back into the system.The only feedback route is therefore the autocatalytic conversion of A to B where the rate of reaction depends on the concentration of B. It is also assumed that the precursor P is in large excess and where p = ( k ~ ~ k l / k z ~ ) ~ ' ~ p o , a = (kl/k#a, p = (k,/kz)'b and xp = k3/k2. The term ~1 is chosen as the bifurcation parameter and the different types of qualitative behaviour as ~1 changes are 0 50 0 0 50 0 0.3 /ie) 10.3 50 0 50 0 Ti rn e/s Fig. 3 Modelling resultsANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 309 100 Y ‘OI 0 ‘ Ambient temperature 746 20 s H Time + Fig. 4 Experimental results: thermocouple trace investigated. This corresponds to eit$.er a change in rate constants or in the initial concentration of P .From a bifurcation analysis it is found that the reaction can be steady- state ( i . e . , after an initial transient period the concentrations of A and B do not change with time) or can show simple period-1 oscillations. The bifurcation diagram in Fig. 1 shows the changing steady-state concentration and maximum oscillatory peak in dimensionless B concentration as p changes. The dramatic change in the type of oscillatory response, from small- amplitude sinusoidal oscillations to large-amplitude oscilla- tions, which happens over a small parameter region, illustrates the extreme sensitivity of chemical systems to parametric conditions. The Coupling of Thermal Feedback Taking into account the temperature dependence of the rate parameters for the reaction, the heat produced in the final termination step is coupled back into the concentration equations providing another feedback route, The chemistry is effectively amplified by the temperature.A further dimension- less energy balance equation is added into the system and is of the form: d@ldt = p-y@ (7) where y represents the dimensionless newtonian cooling coefficient. As the concentration of B oscillates, the dimen- sionless temperature c$ follows. The bifurcation parameter p is no longer constant, since it depends on the reaction rate constants, which are now temperature dependent, and takes the form Foe@ in this non-isothermal case. For certain conditions, this provides a kind of internal forcing of the simple oscillatory response found in the isothermal case, and gives rise to some interesting complex behaviour. As the temperature oscillates, p effectively oscillates also, and for a single parameter condition, i.e., constant h, a mixture of sinusoidal and relaxation type oscillations is observed.Fig. 2 shows this ‘mixed mode’ response. In (f) there is no repeating unit and for long times the response is unpredictable. The solution is now chaotic. No matter how close two sets of initial conditions are chosen in this region, the two trajectories will never converge to the same solution. No long-term predictability exists for such chaotic solutions. Real Systems The above system is only a prototype chemical scheme and bears no direct relationship to a real system.It has, however, given insight into the kind of dynamics found in real systems by mimicking their behaviour, and gives clues as to the underlying mechanisms which cause such responses. For example, in the case of hydrogen oxidation in a continuous stirred tank reactor or CSTR, the same kind of mixed-mode responses as shown in Fig. 2 are observed experimentally. If the system is modelled isothermally using a realistic reaction mechanism it is found that this kind of oscillatory solution can change dramatically with the ambient temperature of the reaction, although it always remains simple period. Again both sinusoidal and relaxation peaks are seen arising due to the competition between chain-branching and termination steps in the system. If the temperature is allowed to couple to the chemistry and Arrhenius dependence for the rate parameters is assumed, a mixture of small and large oscillations appears as a second ‘internal’ forcing is produced.Fig. 3 shows an example of the kind of oscillations found in the model and Fig. 4 the experimental hydrogen oxidation system. Conclusions Chemical systems contain all the ingredients to produce chaotic dynamics. Their rate equations are highly non-linear even for isothermal systems, and feedback mechanisms can arise through either autocatalytic behaviour or through self-heating. Far from equilibrium conditions are inherent in flow sytems but can also be produced for long periods of time in closed systems if the decay of the initial reactant pool is slow enough. Transient chaos is therefore a possibility even for closed systems, Chaotic dynamics can be seen in both experimental and model systems, and the model systems show that the internal mechanisms of chemistry can be enough to produce such complexities. The coupling of chemical and thermal feedback can produce a wide variety of responses. However, the extreme sensitivity of chemical systems to parameter changes forces one to take account of the possibility that external factors (such as poor mixing or insensitive temperature control) can play a part. Such strong evidence for the existence of chaos in chemistry means that we should be aware of why it arises and how to recognize it via the well established routes to chaos. Although a system in a chaotic region is unpredictable, all the types of period response which surround it in parameter space are easily categorized. Armed with this information, chaotic systems need not cause a problem and a knowledge of the whereabouts of the chaotic region, coupled with an understanding of the extent of unpredictability found there, can often be enough.

ISSN:0144-557X    

DOI:10.1039/AP9933000307

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

310 ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 Chaos, Chemometrics and Chemistry: An Introduction to Chaotic Systems Steven R. Bishop Centre for Nonlinear Dynamics and its Applications, Civil Engineering Building, University College London, Gower Street, London WClE 6BT Historical Remarks Three hundred years ago Sir Isaac Newton in his ‘Principia’ proposed mechanical laws of motion which could be used to describe the dynamical behaviour of many systems such as objects falling to the ground and planetary motion. Newtonian dynamics have subsequently been applied on different scales ranging from molecule vibrations’ to galaxy formation2 and in the macroscopic world of engineering.3 Furthermore, these far-reaching laws of differential evolution have been extended to many other fields of study including chemical, biological, economical and other ‘softer’ scientific disciplines.It might have been expected that determinism was here to stay and that given information about the state of a system at a particular instant, it would be possible to predict the future for all time (in Newton’s time the main interest was in celestial mechanics and the stability of the solar system, but the same questions may today be asked of modern chemical systems). Since then many gifted mathematicians believed in this Newtonian paradigm and perhaps it was only Poincark who cast doubts upon this assumption in the late 1880s. Certainly he realized that analytical solutions might not exist and instead Poincark focused his attention on the qualitative dynamics4 of all motions within the solution space.In the 1920s, the work of Bohr in quantum mechanics placed restrictions on the use of Newton’s theories in basic physics while the research of Birkhoff’ opened up new channels of thought on dynamical systems. In the interim years up to 1963 many talented scientists added complexities and laid down some of the mathematical foundations of theories that were to follow. There were two major breakthroughs in 1963 which have shaped our ideas since then. Firstly, Smale6 provided a qualitative picture of how complicated behaviour could be explained by relatively simple rules of stretching and folding of ensembles in phase space (Smale’s Horseshoe). Secondly, though many experimental and early computational investi- gators may have noticed irregular behaviour of non-linear systems prior to this date, it is generally regarded that in the same year it was Lorenz7 who was the first to cement his thoughts and numerics and put them into print.The advent of the modern computer sparked off a revolution in the study of chaotic systems and Lorenz’s original research on convective flows for meteorological weather predictions has now been followed by a plethora of numerical studies. Stimulated by the results of May’ and others, who showed complex behaviour resulting from very simple discrete mappings, anyone equipped with a personal computer could explore the intricate, and seemingly endless, patterns of non-linear systems. Indeed, the beautiful and mystical that are springing from such systems have attracted much attention but this popular image of chaos should not detract from the significant advances to our understanding of scientific principles that have recently emerged. The results of Feigenbaumll (who proved a universal propert& called a period doubling cascade) and Ruelle and Takens (with their work on the nature of turbulence) are two such advances worthy of specific note.Mathematicians, always eager to prove theorems and propose conjectures, have found a new vehicle for their talents while chaotic motions have now been observed in a wide variety of physical experiments (it is not possible, nor correct, to quote here the applications of chaos or a list of where it has been observed: the interested reader might begin by consulting some of the general texts13-16).In chemistry, research into chaotic states appears to be divided mainly between studies of chemical reaction, particularly in the Belousov-Zhabotin~kii’~ system, and convection in Rayleigh-Bdnard systems. A number of references in the literature can be found which look at the complex behaviour of the non-linear systems typically found in chemistry through a mathematical treatment which follows the same pattern as the geometrical approach of dynamical systems theor Specific mention should be made of the research of Scott,‘. his co-worker Gray,” Rossler,*’ Epstein,21 and particularly Ottino,22 who made the first move towards using chaotic theories in chemical mixing. It has not been the intention here to give a complete historical breakdown of events, nor shall the underlying theory be fully considered.It is hoped rather to lay some of the foundations leading up to the present day ideas so that comparisons can be drawn within the reader’s own area of research. A readable, journalistic view of the birth of chaos can be found in G l e i ~ k ; ~ ~ for an interesting and humorous perspective see also Stewart,24 while those concerned with the more theoretical as ects can consult a positive cornucopia of texts in bookshops. 3,25-29 P Chaotic Attractors In order to evaluate the role of chaos in chemistry a clear understanding of the concepts involved has to be gained. For clarity some of the salient features of chaotic systems are mentioned here. Chaos is a technical expression for a specific type of irregular motion produced by a deterministic system. That is to say, ifwe know precisely the inputs to a particular system so that it can justifiably be termed deterministic and the output from that system is irregular (quasi periodic solutions apart) then we say the system is chaotic. The converse is not true.A non-regular response does not imply that a chaotic system is at work since the nature of the input has not been verified. Furthermore, it is no surprise that given a stochastic, random input then the response of a system is also random. However, provided levels of random elements are small (noise) we can still call upon many of the ideas developed to reveal information on the state of the system (see the comments on time series analysis). Chaos is a long-term phenomenon and therefore in the presence of energy dissipation requires a continuous input of energy.This poses severe restrictions within the macroscopic world of engineering or chemical systems since long-term determinism cannot often be guaranteed. What if the energy input decays, or the system evolves in some manner? Again, the ideas developed for chaotic systems can still be applied and the behaviour of chaotic transient solutions can be understood while an evolving system can be studied with particular regard to qualitative changes in behaviour as parameters vary. Chaos has been described as persistent instability but the qualitative structure of transient motions can also be addressed. It should be noted too that the discussions in this paper relate only to dissipative systems though most of the ideas and concepts carry over to their conservative.Hamiltonian counterparts.ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 311 3D phase portrait 1 I I x + f(k, x ) - Steady states ------m Transients 2D phase projection Poincare mapping: A+ B+ A n = 2 fixed point (sub-harmonic) Fsi ncli t Stroboscopic Poincare sections for forced oscillators Fig. 1 Sampling of a trajectory within a three-dimensional phase space by stroboscopically sampling the solution at multiples of the driving period. The diagram shows a sub-harmonic response of the system which can be represented by the points A and B on the sections (PoincarC sections). Also shown are transient motions which decay onto this attracting solution Chaos lies within a well ordered structure.Before under- standing this statement it is necessary to define some termin- ology. Today, the geometrical approach of Poincare is used to great effect through the study of the topological structure of an ensemble of trajectories which fill the governing phase space. For high-order systems this phase space has to be perceived conceptually, and it is thus no surprise that many of the studies of chaotic systems to date have been for low-dimensional systems, usually three. One of the most challenging problems now being posed is the question of whether the dynamics of higher dimensional systems can be viewed sensibly on a space of lower dimension (an inertial manifold). For ease of explanation the response of a periodically (sinusoidally) forced oscillator whose phase dimension is spanned by its displacement, its velocity and the time is initially considered here.A bundle of trajectories would spiral around the time axis as time itself advanced. In this space trajectories do not cross; if they did then we would instead have to consider a probabilistic system. Often, it is more convenient to view the two-dimensional projection of these trajectories, though on this displacement/velocity plane trajectories may now cross to form a phase portrait. A further technique to reduce the amount of information and yet still reveal all the significant dynamic behaviour, is the trick of stroboscopically sampling the trajectories at multiples of the period of the forcing (so called Poincare samples, or sections) as can be seen in Fig.1. This figure in fact shows a sub-harmonic solution which repeats after two periods of the forcing as well as transient motions that are attracted to this solution. A collection of these Poincard samples can also be viewed as a projection onto a single plane. Jn this manner the sub-harmonic solution in Fig. 1 would reveal itself as a sequence of two repeating points marked A and B. A harmonic solution, which has the same frequency as the excitation, would be seen as a single repeating point. A transient solution of the system would manifest itself as a sequence of points that decays onto a stable solution; that is, the stable solution attracts the transient motions and we say that its representative repeating point is a point attractor. If we knew an expression for the solution to the system then we could also write down a relationship between consecutive points on the Poincare sections.This relationship would be in discrete form and is known as the Poincare map. A fixed point of the Poincare map thus corresponds to a stable orbit of the continuous flow. If we are considering a non-linear system then we are not usually able to write down the solution and therefore the exact form of the Poincare map is not known. In this case we will have to resort to numerical solutions to the system from which approximations to the mapping can be obtained. There is much to be gained from the study of discrete maps and these are interesting in their own right, bearing in mind that there are many problems in which we are only able to view a system at intervals.When a chaotic trajectory is viewed within this sampled phase space we see not a sequence of repeating points (since the motion is irregular) but a pattern or collection of points which may have a complex shape. Chaotic motions may be stable in the same way as periodic solutions so that nearby trajectories are attracted. Similarly the pattern of points which corresponds to chaotic motion attracts points resulting from decaying transient solutions and hence we refer to the pattern as a chaotic attractor. Fig. 2 illustrates a chaotic response of a non-linear second- order ordinary differential equation of the form: i + 0.05i + x3 = 7.5 sint (1) Two trajectories are plotted, both of which decay to a stable chaotic motion as indicated by the chaotic attractor in the bottom right-hand diagram.This equation has been used by the author to study such problems as the behaviour of a buckled beam and the surge motions of a monohull offshore drilling vessel. The details are not important here but instead the reader should focus on the possible behaviour from this typical non-linear system. If a system were responding in a stochastic, random manner (normally distributed perhaps) then although sone points would be visited more than others, given sufficient time all points within the plane would be visited. Conversely in a chaotic system the points generated by sampling a single trajectory (though not simply stepping along the attractor but jumping about in a erratic fashion) always remain within the3 12 ANALYTICAL PROCEEDINGS7 JULY 1993, VOL 30 P C I x + 0.05 X t x3 - 7.5 cos f // Damped forced non-linear oscillator m& Beam at buckling load 5 Transient from x = 0, X 0 Steady-state chaos 2 OX t - 5 1 k = 0.05, 8 = 7.5 6 5 h frequency transient from x - 2 - 4 -6 -5 ' x Steady chaos Poincare 2ix sections (mapping points) defining the attractor Fig.2 persistent steady-state chaos. The long-term trajectory sampled at multiples of the periodic forcing reveals a collection of points: a chaotic attractor Chaotic response of a damped forced non-linear oscillator showing two separate trajectories that after transient decay settle onto the attractor. Thus we see that a chaotic solution has an in-built structure or order; it is not chaotic in the everyday sense of the word.We have already noted that random fluctuations, for instance of temperature and mixing, can produce an irregular response during a chemical reaction without chaos. A major topic of research is the study of chaotic systems whose response is masked by noise in some sense; the task is to reveal the underlying chaos by removing the noise (or indeed it may be easier the other way around). Having evaluated the structure of the chaotic behaviour it may then be possible to back-track and describe the form of the original system. Chaos can arise when there is non-linearity and phase dimension of at least three. This does not mean that all non- linear systems are chaotic, but what is true is that chaos is not found in continuous linear systems (the addition of a disconti- nuity, however, results in a non-linear system which can result in all the usual features3'). Any analysis based on classical linear stability, though perhaps easier to perform, will not be able to capture the complicated behaviour of chaotic systems. Therefore, it is vital that engineers and scientists are fully aware of the limitations of their model if they are to avoid unfavourable system responses. Chaos is a typical ingredient of chemical systems.Most chemical systems are inherently non-linear and there are no reasons why we should therefore expect them not to behave chaotically for some parameters. As we have already noted, chaos has been observed in various albeit well controlled experiments. Unpredictability No analytical solution can be written to describe chaotic motion quantitatively; we must resort to numerical solutions.However, given precise details of initial conditions the response of a chaotic system can be predicted at any time in the future and the experiment repeated exactly. I , Starts x = 3, x = 4 and x = 3.01, i = 4.01 3 2 1 g o -1 -2 -3 k = 0.05, B = 7.5 I I I I I I I 2x 4x 6x 8x 1011 12x 14x Fig. 3 Exponential divergence of adjacent starts. The two trajectories of a chaotic system are initiated very close together but soon lose all correlation so that effectively a limit is placed on the predictability of the system due to finite precision Chaos exhibits exponential divergence from adjacent starts. If the initial conditions of a chaotic system are known only with some margin of error then the final outcome becomes unpredictable. The associated motions of a chaotic system, though deterministic, are irregular and unpredictable, in the sense that infinite precision is required in order to repeat a particular solution exactly.Transient trajectories are still attracted to the chaotic solution but within the attractor there is an exponential stretching action that can cause two very near- by points to become separated so that their long-terms motions are no longer ~ o r r e l a t e d . ~ ~ Fig. 3 illustrates this divergence by plotting the trajectories from two close-by initial conditions. At first the trajectories remain together but after a few cycles they move rapidly apart. This sensitivity to initial conditions orANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 313 divergence is used as an indicator of chaos as well as a measure.In a chaotic system we cannot usually say exactly when the two trajectories will diverge from each other, all we know is that when they do they do so exponentially. It is a further folding action that allows this divergence within the attractor without dispersion of the trajectory. In order to appreciate the mixing and folding fully it is necessary to view an attractor dynamically as time advances. The simplest analogy for this behaviour is that of the action with which a baker kneads dough when making bread, but the best way to view this folding behaviour is to carry out dynamic simulations and visualize the response. The measure used for the divergence is called the Lyapunov exponent, with at least one positive value indicating chaos.However, since the system is bounded, the Lyapunov is a local measure found by averaging many points along a trajectory. The determination of a positive Lyapunov exponent is used in meteorology, for instance, to indicate when the system which describes the weather pattern is in a predictable non-chaotic state. Another measure often calculated in chaotic systems is the so called fractal dimension used to quantify the ‘strange- ness’ of the attractor (hence the term strange attractor which is often used synonymously with chaotic attractor). This measure gives an indication of the complexity of the dynamics within the attractor; a value of D > 2, which is also non-integer, indicates a chaotic attractor.Other measures can also be calculated for chaotic and non-chaotic systems though their usefulness in practice is open to question. Complexities of Non-linear Systems The complexities of non-linear systems seem almost limitless (the author often jokes that nothing surprises him and conjectures that almost anything is possible). Equilibrium positions, periodic states and chaotic responses can all co-exist at given system parameters. Under a change of parameter a path of solution can undergo fine detailed local bifurcations, or the particular solution might lose stability in a global bifurca- tion and the system will necessarily jump to a distant attracting solution. It appears that the only way to be certain of avoiding any non-linear behaviour is to ensure that your system behaves in a purely linear manner.A technique used in the study of global behaviour of non- linear systems is the examination of the ensemble of solutions leading towards a common attracting solution. The initial conditions (in terms of displacement and velocity) which result in the different solutions can be evaluated and plotted collectively to form basins of attraction. Basins can be smooth and simple, basins of co-existing attracting solutions can lead to a complicated global portrait, and basins may become fractal having finer and finer detail on increasing scales of magnifica- tion (fractal dimension). An example of a fractal basin is shown in Fig. 4 in which the different colours/shades represent points which lead to the various attracting solutions.The intricate patterns show that an initial condition chosen within the fractal region with only finite precision cannot, with any certainty, be guaranteed to decay onto a desired solution. In some cases it can be shown that a boundary point is in fact on the boundary of all the attractors. These fractal basins thus add an extra element of indeterminacy to non-linear systems. It should be noted that in this instance no chaos is present; however the existence of a fractal boundary implies the presence of chaotic transient motions. Chaotic transients have all the properties of steady-state chaotic motions except that eventually they converge to one of the existing periodic states. Their existence often indicates that a chaotic attractor exists in nearby parameter space.Fractal basins have important implications for engineering systems since it has been shown that the basin of attraction leading to safe solutions can be eroded by the fractal nature of the boundary in a dramatic way.32 Fig. 4 Fractal basin of attraction. Plot of displacement against velocity; any close-up near the boundary reveals stripes (tongues) of a basin resulting in one solution intermingled with areas which lead to another Conclusions Mathematics plays a vital role in evaluating the complex behaviour of chemical systems through an understanding of the qualitative response of non-linear systems. Some of the important concepts which have arisen from the new discovery of chaos, arguably the most exciting mathematical develop- ment of recent times, have been introduced.It is now known that chaos is a typical ingredient of dynamics which has been observed in systems governed by Newton’s laws of motion and experimentally visualized in mechanical experiments and fluid systems. The associated motions of a chaotic system, though deterministic, are irregular and unpredictable, in the sense that infinite precision is required in order to repeat exactly a particular solution. Chaotic motions can lead to unexpected failure of what was thought to otherwise be a safe and deterministic system. Steady-state chaos, chaotic transients and chaotic modes of instability all add a very real element of indeterminacy into the response and hence safety of chemical sys tems. An important area of current research is the analysis of time series data from which qualitative dynamical behaviour can be extracted from experimental results.33 Through this approach it is possible to quantify measures as already described and build up an idea of the model from which the data have arisen.To date, much of the research on chaotic systems focuses attention on an understanding of behaviour and most have been directed towards the avoidance of unwanted unpredic- table motions rather than the use of chaos. However, chaotic motions have very useful mixing properties and it is perhaps in the field of chemical mixing that chaos has its first usage.22 Fluids with high levels of viscosity can be made to behave in a chaotic manner leading to a thorough mixing without unduly disturbing the structure of the fluid (useful perhaps in the treatment of blood).The enhancement of flow transport for fluids with a high viscosity might also prove to be a useful area in which chaos can play a significant role. The author acknowledges the support of the Science and Engineering Research Council. Thanks are extended to colleagues at University College London for their help and advice. References 1 Debrunner, P. G., and Frauenfelder, H., Annu. Rev. Phys. Chem., 1982, 33, 283. 2 Aarseth, S. J., Gott, J. R., and Turner, E. L., Astrophys. J . , 1979, 228, 664. 3 Thompson, J. M. T., Notes Rec. R. SOC. London, 1988,42,97.ANALYTICAL PROCEEDINGS, JULY 1993, VOL 30 314 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Poincare, H., Les mkthodes nouvelles de la mkcanique ckleste, Gauthier-Villars, Paris, 1899, vols.1-3. Birkhoff, G. D., Dynamical Systems, American Mathematical Society Colloquium Publications, Providence, RI, 1927. Smale, S., in Dqferential and Combinatorial Topology, ed. Cairns, S. S., Princeton University Press, 1963, 63-80. Lorenz, E. N., J. Atmos. Sci., 1963, 20, 130. May, R. M., Nature, 1976, 261, 459. Peitgen, H.-O., and Richter, P. H., The Beauty of Fractals, Springer-Verlag, Berlin, 1986. Mandelbrot, B. B., The Fractal Geometry of Nature, Freeman, San Francisco, 1982. Feigenbaum, M. J., J. Stat. Phys., 1978, 19, 25. Ruelle, D., and Takens, F., Commun. Math. Phys., 1971, 20, 167; 23, 343. Thompson, J. M. T., and Stewart, H. B., Nonlinear Dynamics and Chaos, Wiley, Chichester, 1986. Thompson, J. M. T., and Gray, P., Philos.Trans. R. SOC. London, A , 1990,332,49. Moon, F. C., Chaotic Vibrations, Wiley, New York, 1986. The New Scientist Guide to Chaos, ed. Hall, N., Penguin Books, London, 1991. Roux, J. C., Rossi, A., Bachelart, S., and Vidal, C., Phys. Lett., 1980, 77A, 399. Scott, S. K . , Chemical Chaos, Oxford University Press, 1991. 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Gray, P., and Scott, S. K., Oscillations and Instabilities, Oxford University Press, 1990. Rossler, 0. E., J. Theor. Biol., 1972, 36, 413. Epstein, I. R., Physica, 1983, 7D, 47. Ottino, J. M., The Kinematics of Mixing: Stretching, Chaos, and Transport, Cambridge University Press, 1989. Gleick, J., Chaos, Heinemann, London, 1987. Stewart, I., Does God Play Dice?, Blackwell, Oxford, 1989. Guckenheimer, J., and Holmes, P., Nonlinear Oscillations, Dynamical Systems, and Bifurcations of Vector Fields, Springer- Verlag, New York, 1983.Wiggins, S., Introduction to Applied Nonlinear Dynamical Systems and Chaos, Springer-Verlag, New York, 1990. Lichtenberg, A. J., and Lieberman, M. A., Regular and Stochastic Motion, Springer-Verlag, New York, 1983. Devaney, R. L., An Introduction to Chaotic Dynamical Systems, Addison-Wesley, Redwood City, CA, 2nd edn. , 1989. Holden, A. V., Chaos, Princeton University Press, 1986. Foale, S., and Bishop, S. R., Phil. Trans. R. SOC. London, A , 1992, 338, 547. Lighthill, L., Proc. R. SOC. London, A , 1986, 407, 35. Soliman, M. S . , and Thompson, J. M. T., J. Sound Vib. , 1989, 135, 453. Broomhead, D. S., and King, G. P., Physica D (Amsterdam), 1987, 20, 217.MENDELEEV COMMUNICATIONS Mendeleev Communications i s a unique publication providing rapid access to the extensive chemical research activities of an important and fascinating world region -the Commonwealth of Independent States. It is not a translation journal; all material is published directly in English and is therefore as up-to-date as possible. Mendeleev Communications i s a joint publishing venture between The Royal Society of Chemistry and The Russian Academy of Sciences. It contains preliminary accounts of novel and significant results of wide general appeal or exceptional specialist interest on any branch of chemistry. Most papers are submitted from the CIS but some come from other parts of the world. Mendeleev Communications acts as both a means and a stimulus for international dialogue. I 1993 Subscription Details Published six times per annum plus annual author and subject index EC €140 .OO USA $280.00 Canada €147.00 (+ GST) Rest of World €140.00 ISSN 0959-9436 Back issues available on request. Mendeleev Com mu n icat ions: * Has proved so successful that it i s now * Is not a translation journal * Publishes rapidly- within 12 weeks of receipt of papers in the UK * Will accept high quality papers on all topics of chemistry * Offers a unique insight into the research activities of the CIS published six times per annum Mendeleev Communications is essential reading for everyone who is interested in keeping up-to-date with the latest chemical research. For further information on Mendeleev Communications please contact: ROYAL Sales and Promotion Department, SOCIETY OF Royal Society of Chemistry, CHEMISTRY Thomas Graham House, Science Park, Milton Road, Cambridge CB4 4WF, UK. Tel: +44 (0) 223 420066. Fax: +44 (0) 223 423623. Telex: 818293 ROYAL. @ && Information Services

ISSN:0144-557X    

DOI:10.1039/AP9933000310

出版商:RSC    

年代:1993

数据来源: RSC