Trace Residue Analysis

Ann!no, Raymondf. 83. Caro, J . H., 25. C u r r i e , L l o y d A., 49. Dunn, W. J . , I l l , 195. Freeman, H. P., 25. Hogan, J . W . , 195. J o h a ...
0 downloads 0 Views 639KB Size
Download PDF
Author Index

R o s e n b e r g e r , James L , 133 S c h w a r t z , T. R., 195 S m i t , H. C , 101 S t a l l i n g , D. L . , 195 Tamayo, Gwen J . , 133 T a y l o r , A. W., 25 T s c h i l t s c h k e , Frauke, I T u r n e r , B. C., 25 W e g s c h e i d e r , W o l f h a r d , 167 Wold, S., 195 Z e r v o s , C., 235

Ann!no, Raymond 83 C a r o , J . H., 25 C u r r i e , L l o y d A., 49 Dunn, W. J . , I l l , 195 Freeman, H. P., 25 Hogan, J . W . , 195 J o h a n s s o n , E . , 195 K r a t o c h v i l , B., 5 K u r t z , D a v i d A., 133, 183 M i t c h e l l , D o u g l a s G., 115 M u h l b a u e r , Johann A., 37 P e t t y , J . D., 195

Trace Residue Analysis Downloaded from pubs.acs.org by 5.101.221.127 on 01/03/19. For personal use only.

f

Subject Index

A Accuracy model selection, 189 regression, 160-64 Amount intervals c a l i b r a t i o n graphs, 183-93 spline method, 192t transformed method, 191t,192t Amount transformation, 147-49 Analysts, use of chemometries, 261,264-65 Analyte detection a n a l y t i c a l model, 55-56 l i m i t of, hypothesis testing, 51-52 Analytical chemistry, hypothesis testing, 242-44 Analytical model, assumptions and p r a c t i c a l implications, 52-57 Analytical performance, correlation chromatography, 108 Analytical process, steps of, 7 Aroclors, isomer-specific analysis of, application of SIMCA, 195-232 Atomic absorption spectrometry, determination of iron i n water, 116 277

Autocorrelation function of a power signal, d e f i n i t i o n , 103 Automatic processing o f standard data, o u t l i e r processing, 38-43 B Bandwidth(s) comparison o f three research methods, 191-93 confidence l e v e l s , 153 data and estimated amount i n t e r v a l s , 189-91 d e f i n i t i o n , 185 Binary noise, correlation chromatography, 104 Binomial data, normalization of, 44 Bonferroni i n t e r v a l estimates, 138-39 C Calibration alternative models, 62-63 l i n e estimation, 139 representation by a matrix equation, 59

278

TRACE RESIDUE ANALYSIS

Calibration—Continued r i s k s to data quality, 116-18 spline function, testing of

accuracy, 170-76 Calibration curve(s) amount and estimated amount intervals, comparison, 183-93 confidence bands around the curve, 120f construction, 169 detection l i m i t s , 73-80 determination, 55-56,185 fenvalerate, 67f linear, 58-61 one-dimensional, pesticide detection, 57-67 Calibration-curve-based analysis method development, 130f use of multiple-curve and weighted least-squares procedures with confidence band s t a t i s t i c s , 115-31 Calibration data determination of iron i n water, 117f extrapolation caveat, 159 hypothetical, 1l8f modeling, 134 Calibration problems proposed solutions, 116,119 use of cubic spline functions for solving, 167-80 Chemical analysis, many dimensions of detection i n , 49-80 Chemometrics applications, 101-2 d e f i n i t i o n , 236 i n s t i t u t i o n of, f o r a n a l y t i c a l use, 253-67 resolution vs. physicochemical analysis, 68-70 Chlordecone applications, 135 power transformation, 146 Chloride, determination i n blood, 127f Chlorothalonil applications, 135 data set, 272 power transformation, 146 transformed-response variances, I44t Chlorpyrifos applications, 135 data set, 272 power transformation, 146 Chromatogram, description, 102 Chromatographic data applications of SIMCA, 210-18 matrix of, 204f Chromatographic detection, 134

Chromatography c a l i b r a t i o n problems, 133-34 correlation, theory of, 83-99 C l a s s i c a l assumptions for f i t t i n g regression l i n e s , 43 C l a s s i c a l sampling problem, description, 43 C l a s s i f i c a t i o n problems, use of SIMCA, 216,219-20 Clustering class of i d e n t i c a l samples, 205f three classes of samples, 207f Columbia National Fisheries Research Laboratory, studies of PCBs, 196 Composite samples, description, 8 Concentration calculations, general a n a l y t i c a l plan, 136-37 Confidence i n t e r v a l s ) analyses of t r i f l u r a l i n content of f i e l d s o i l , 29 calculation, 152-54 calibration-curve detection l i m i t s , 79-80 construction, 179-80 d e f i n i t i o n , 153 determination of fenvalerate, 179t inverse-transformed data, 152t regression on transformed data, 133-64 spline functions, 170 s t a t i s t i c s , use i n c a l i b r a t i o n curve-based analysis, 115-31 transformed data, 154t use i n reporting data, 255 Confidence l e v e l , choice of, i n scient i f i c studies, 247-48 Conjectures, role and contribution to s c i e n t i f i c research, 237-38 Constant variances, 145-46 Correlation d e f i n i t i o n , 83 weights, and standard errors, calibration-curve detection l i m i t s , 77-78 Correlation chromatograph schematic, 88f,105f setup, 106-7f,109f Correlation chromatography application i n trace analysis, 101-13 background, 83-87 description, 102 p r i n c i p a l s , 102-6 problems, 95-99 theory, 83-99 Correlation c o e f f i c i e n t ( s ) description, 85-86 vs. tau, 86, 87f

INDEX

279

Correlation c o e f f i c i e n t — C o n t i n u e d t r i f l u r a l i n disappearance rate, 32-33 Correlation HPLC system, setup, 109f Correlation noise, correlation chromatography, 96 Correlogram(s) construction, 89-93 description, 102 separation of phenols, 110f Covariance, t r i f l u r a l i n content of f i e l d s o i l , analysis of, 33 Cross-correlation function d e f i n i t i o n , 86 two power signals, 103 Cross validation, spline functions, 170 Cubic spline function(s) representation, 168-69 for solving c a l i b r a t i o n problems, 167-80 variance estimation, 185 D Data bases, problems with, 259-60 Data pretreatment, SIMCA, 208 Decision c r i t e r i o n , hypothesis testing, 51 Decision flow chart, examples, 45-57 Decision l e v e l ( s ) c a l i b r a t i o n curve, 73-80 pesticide detection, 58-59 Decision rules f o r s o c i e t a l l y important study systems, 246-50 Decision strategies, gating hypothesis, 244-45 Decision theory, i n metrics, 241 Design and measurement descriptions, use i n reporting data, 256-57 Detectable signals, reporting, regulations, and p r a c t i c a l implications, 56-57 Detection i n chemical analysis, many dimensions of, 49-80 Detection l i m i t ( s ) analysis o f phenol and dimethylphenol, 108 c a l i b r a t i o n curve, 73-80 defining of, 49-50 hypothesis testing, 51 pesticide detection, 59 Detector signal cross correlation with the sampling code, 96 cross correlation with the valveposition code, 89 Dieldrin, pathways and rate of loss from grass-meadow s o i l , 15-21

D i s j o i n t p r i n c i p a l components models, 206 Distributions found i n nature, 9-10

E Emulsifiable concentrate of herbicide, s o i l treatment with, 26 Environmental applications, SIMCA, 223-26 Error propagation, fenvalerate detection l i m i t s , 66 Error terms, a n a l y t i c a l model, 53,54t Estimated amount interval(s) and bandwidth data, 189-91 c a l i b r a t i o n graphs, comparison, 183-93 differences determined from three methods, 188-89 inverse-transformed data, 157-59 Estimated concentration, pesticide detection, 58 Estimated-response error bounds, inverse-transformed data, 155-56 Estimation minimum number and size of sample increments, 10-12 number and size of increments f o r a segregated population, 12-13 sample size when form of population d i s t r i b u t i o n i s unknown, 14

F False positive decisions, i n toxicology, 246 False positive and negative r i s k s , hypothesis testing, 50-51 Federal Working Group on Pest Management, 5-6 Fenvalerate applications, 135 c a l i b r a t i o n curve, 67f c a l i b r a t i o n data f o r GC measurement, 61-63 data sets, 184,269-71 detection l i m i t s , 63,65-67 determination by GC, 125f power transformation, 146 transformed-response variances, I43t use of spline functions i n determinations of, 174,177-79 F i e l d layout and management, t r i f l u r a l i n disappearance study, 26 Field s o i l , t r i f l u r a l i n disappearance, 25-35

280

TRACE RESIDUE ANALYSIS

First-order least-squares curves, inappropriate use of, 117f First-order regression algebraic equations for, 121t models, 45-47 Flow charts, o u t l i e r processing, 38,40-42,44 Fractional composition histograms PCB mixture, 212f,222f transformer f l u i d , 222f

G Gas chromatography (GC) analysis of PCB residues, 197-98 analysis of PCBs and transformer o i l s , 227-32 determination of fenvalerate, 125f Gating hypothesis, decision strategies, 244-45 Gaussian d i s t r i b u t i o n , 9-10 Geometrical constructs, evaluation of sample s i m i l a r i t y , 208 Granular formulation of herbicide, s o i l treatment with, 26

Intraclass correlation c o e f f i c i e n t , d e f i n i t i o n , 13 Inverse-transformed data confidence intervals, 152t estimated amount intervals, 157-59 response error bounds, 155-56 transformation to real values, 159-60 I r r e g u l a r i t y of d i s t r i b u t i o n of herbicides i n f i e l d s o i l , causes of, 34 Isomer-specific analysis o f PCBs, application of SIMCA, 195-232 Isomer structure assignment, response factors, and concentration, PCB mixture, 202-3t

K Kepone data set, 273 transformed-response variances, I43t Knots, spline functions, 168-69

H

L

Half-lives, t r i f l u r a l i n formulations, 33 Hand p l o t t i n g of data, 186-88 Hartley test, constancy of variance, 145 Herbicide(s) application methods, 34 distribution in field s o i l , e f f e c t on sampling, 34 High-performance l i q u i d chromatography (HPLC), separation of phenols, 109f Hypothesis testing a n a l y t i c a l chemistry, 242-44 metrics, 239 scalar signals, 50-52 toxicology, 241-42 Hypothetical mixture, changes i n fraction composition due to decreasing concentration, 215t

Laboratory analysis, systematic error i n , 257-58 Lead, determination i n blood, 123f,127f Least-squares procedures, weighted, use i n calibration-curve-based analysis, 115-31 Least-squares regression, requirements, 134 Legislators, education of, use of chemometrics, 263-64 Limit of detection, defining of, 49-50 Linear c a l i b r a t i o n curve(s) confidence bands from regression on transformed data, 133-64 decision and detection, 58-61 Linear regression models, 138,151f

I Ideal a n a l y t i c a l model, deviations from, 53 Interval estimate(s) concentration calculations, 142 unknown amounts, description, 156

M Management and f i e l d layout, t r i f l u r a l i n disappearance study, 26 Managers and manufacturers, education of, use of chemometrics, 261-63 Mathematical methods, o u t l i e r processing, 44t Mathematical models, c a l i b r a t i o n curve-based analysis, 119,122-24

INDEX

281

Matrix equation, representation of c a l i b r a t i o n , 59 Maximum reportable concentration iron i n water, 128f measurement of, 129 Mean values calculation of, 254-55 t r i f l u r a l i n content of f i e l d s o i l , 28-31 Measurement and design descriptions, use i n reporting data, 256-57 Metrics d e f i n i t i o n , 236-37 history, 239 hypothesis formulation and testing, 239 Minimum detectable bias, d e f i n i t i o n , 13 Minimum reportable concentration, 126-129 Model selection for accuracy, 189 regression, 160 Modeling power, d e f i n i t i o n , 206 Multidimensional data intercomparisons, 70-71 Multidimensional signal, chemical analysis, 69 Multiple-curve procedures, use i n calibration-curve-based analysis, 115-31 Multiple peak, correlation chromatography, 91-92,93f Multispectral sorting, example, 68 Multivariate environments, 256 Multivariate problem, description, 43 N Negative binomial d i s t r i b u t i o n , 9-10 Neymann-Pearson process of s t a t i s t i c a l hypothesis testing, 238 Noise addition, correlation chromatography, 92, 94-95 Nonconstant variance correction for, 122-26 treatment of, 144 Nonlinear c a l i b r a t i o n curves, 61 Nonlinearity, correlation chromatography, 96 Nonnegligible errors, c a l i b r a t i o n curve detection l i m i t s , 74-76 Nontransformed regression, comparison to transformed regression, 161 Normalization data i n t r i f l u r a l i n disappearance study, 32-33 SIMCA, 208-9 transformation equations, 44

Null hypothesis a n a l y t i c a l chemistry, 243-44 toxicology, 241-42 Number crunching, i n s c i e n t i f i c studies, 258-59 0 Observed response, a n a l y t i c a l model, 52 One-dimensional c a l i b r a t i o n curve detection i n chemical analysis, 49-80 detection o f pesticides, 57-67 One-sided normal standard percentiles, hypothesis testing, 51 Optimal amount transformation, convergence f o r the determination of data l i n e a r i t y , l48t Outlier processing automatic processing of standard data, 38-43 factors influencing choice of strategy, 38 flow charts, 38,40-42,44 strategies, 37-38,39f

P P a r t i a l least-squares method, prediction of composition of unknown samples, 220-23 P a r t i a l peak summary of replicate analyses, PCB mixture, 211 Pesticide analysis, application of sampling theory, 15-21 Pesticide detection, one-dimensional c a l i b r a t i o n curve, 57-67 Phenol(s) c a l i b r a t i o n graph, 111 f correlogram, 110f,112f separation by HPLC, 109f Physicochemical analysis vs. chemometric resolution, 68-70 Pictures, use i n reporting data, 256 Point estimate of unknown amounts, description, 156 Poisson data, normalization of, 44 Poisson d i s t r i b u t i o n , 9-10 Polychlorinated biphenyls. (PCBs) c a l c u l a t i o n of composition, 209-10 Columbia National Fisheries Research Laboratory studies, 196 data base, 198-200 description, 195 i n the environment, 195-96 f r a c t i o n a l composition histograms, 212f,222f

282

TRACE RESIDUE ANALYSIS

PCBs—Continued

f r a c t i o n a l composition i n transformer o i l s , 223 gas chromatogram, 201f gas chromatographic analysis, 197-98,227-32 isomer-specific analysis of, application of SIMCA, 195-232 isomer structure assignment, response factors, and

concentration, 202-3t p a r t i a l peak summary of r e p l i c a t e analyses, 211 p r i n c i p a l components plots, 213,217f,219f residues, 196 s t a t i t i s c a l summary for SIMCA analysis, 221 thermal conversion to polychlorinated dibenzofurans, 196 variable loadings, 2l4f Polychlorinated dibenzofurans, thermal conversion of PCBs to, 196 Power transformation chlordecone, 146 chlorpyrifos, 146 fenvalerate, 146 variance s t a b i l i z a t i o n , 185 Precision improvements, 126-30 measurements, 115-16 Prediction, composition of unknown samples, 156-57,220-23 P r i n c i p a l components method, application i n isomer-specific analysis of PCBs, 195-232 P r i n c i p a l components models, 204-6 P r i n c i p a l components plot(s) description, 207-8 PCB mixture, 213,217f,219f transformer f l u i d , 219f Protein binding assays, use of spline functions, 171 Pseudorandom binary sequences, correlation chromatography, 91-92,104-6

R Radioimmunoassay standard curves, of spline functions, 171-72 Random errors, i n analysis, description, 6 Random noise, c o r r e l a t i o n chromatography, 104 Random sampling, 7-8

use

Random sampling—Continued accuracy, 160-64 analysis of t r i f l u r a l i n disappearance rate, 32-33 c a l i b r a t i o n data modeling, 134 c l a s s i c a l assumptions for f i t t i n g l i n e s , 43 c o e f f i c i e n t s for transformed data, 150t estimated amount bandwidths at various responses, l87t,190t f i r s t - o r d e r calculations, 121t f i r s t - o r d e r models, 45-47 model selection, 160 residuals vs. transformed amount, examination of, 150,152 on transformed data, 133-64 Regularization, SIMCA, 209 Regulators, education of, use of chemometrics, 263-64 Relative confidence bandwidth d e f i n i t i o n , 126 determination of chloride and lead i n blood, 127f Relative standard deviation, determination of chloride and lead i n blood, 127f Representative sample, description, 8 Residues of PCBs, 196 Response error bounds c a l c u l a t i o n of, 155-56 inverse-transformed data, 155-56 Response transformation, 142-45 Risk assessment, problems of, 267 Routine chemical analysis, description, 115

S Sample analysis data quality, measurement of, 126 Sample increments, estimation of minimum number and size, 10-12 Sample loading terms, plots of, 207-8 Sample size when form of population d i s t r i b u t i o n i s unkown, estimation, 14 Sampling for chemical analysis of the environment, s t a t i s t i c a l considerations, 5-22 c o r r e l a t i o n chromatography, 95-96 d e f i n i t i o n of constant, 11 measurements of t r i f l u r a l i n disappearance from f i e l d s o i l , 25-35 for pesticides and pesticide residues, problems, 5-6

INDEX Sampling—Continued theory of, application to pesticide analysis, 15-21 Scalar signals, hypothesis testing, 50-52 S c i e n t i f i c method of inquiry overview, 237-40 value foundations, 240-44 S c i e n t i f i c thinking, role of statistics, 2 Second-order least-squares curves, inappropriate use of, 117f Segregated population, estimation of number and size of increments, 12-13 Signal detection, a n a l y t i c a l model, 55 Signal enhancement, trace analysis, 108 Signal-to-noise r a t i o s , correlation chromatography, 89,95 Single-impulse chromatography, description, 102 Single-peak c o r r e l a t i o n chromatography, 86,88-91 Single-pulse chromatogram, 89 Smoothing parameters, spline functions, 169,171-74 Soft independent method of class analogy (SIMCA) applications to chromatographic data, 210-18 applications i n isomer-specific analysis of PCBs, 195-232 environmental applications, 223-26 general discussion, 200,202,204-6 software a v a i l a b i l i t y , 226 use i n c l a s s i f i c a t i o n problems, 216,219-20 Spline function(s) amount bandwidths and ranges, 191-92t c a l c u l a t i o n with pesticide data, 174,177-79 cubic, f o r solving c a l i b r a t i o n problems, 167-80 determination, 168-71 Standard deviation(s) Bonferroni i n t e r v a l estimates, 140 correlation, and weights, calibration-curve detection l i m i t s , 77-78 transformed data, 150t t r i f l u r a l i n content of f i e l d s o i l , 28-31 Statistics h i s t o r i c a l development, 1-4 methodology, 142 use i n c a l i b r a t i o n , 138-42 use i n processing of o u t l i e r s , 34-47

283 Regression use i n sampling f o r chemical analysis of the environment, 5-22 Stochastic s i g n a l , correlation chromatography, 103-4 Straight-line c a l i b r a t i o n , equation, 58 Structure of data, problems with, 260-61 Subsamples, inconsistent, detection of, 43-44 Systematic errors description, 6 laboratory analysis, 257-58 Systematic sampling, 7-8

T Table of random numbers, 8 Target population, i d e n t i f i c a t i o n , 7 Theory testing procedures, 238-39 Toxicology false positive decisions, 246 hypothesis testing, 241-42 n u l l hypothesis, 241-42 Trace analysis, application of corr e l a t i o n chromatography, 97,101-13 Transformation, data into normal or exponential forms, 44 Transformation power of data sets, 146-47 Transformed data confidence i n t e r v a l s , 154t regression on, 133-64 Transformed regression amount bandwidths and ranges, 191-92t comparison to nontransformed regression, 161 Transformed-response variances chlorothalonil data, I44t fenvalerate data, l43t kepone data, l43t Transformer f l u i d f r a c t i o n a l composition histograms, 222f gas chromatographic analysis, 227-32 p r i n c i p a l components plot, 219f Trans-science, description, 240 T r i f l u r a l i n disappearance from f i e l d s o i l , 25-35

U Univariate Univariate

environments, 256 problem, description, 43

284

TRACE RESIDUE ANALYSIS

Unknowns, predictions of composition, 156-57,220-23 V V a r i a b i l i t y , changes with time, t r i f l u r a l i n content of f i e l d s o i l , 30-33 Variable loadings, PCB mixture, 214f Variance of concentration, pesticide detection, 58-61 measurements of t r i f l u r a l i n disappearance from f i e l d s o i l , 25-35 random errors, d e f i n i t i o n , 6 Visman equation, determination of d i e l d r i n content of s o i l , 18-20

W Weighted least-squares procedures, use in calibration-curve-based analysis, 115-31 Weights, correlation, and standard errors, calibration-curve detection l i m i t s , 77-78 Wholeness of thinking, loss of, 3-4 Working-Hotelling confidence band, f o r regression l i n e , 139,141,151f

Z Zero-dimensional case, hypothesis testing, 50