《分析化学样品前处理手册》（英文版）Sample Preparation Techniques in Analytical Chemistry

THE MEASUREMENT PROCESS Homogenization, Size reduction Extraction Concentration Clean-up Analysis Figure 1.2. Possible steps within sample preparation. all aimed at producing those few microliters that represent what is in the fish. It is obvious that an error in the first three steps cannot be rectified by even the most sophisticated analytical instrument. So the importance of the prior steps, in particular the sample preparation, cannot be understressed l.1.1. Qualitative and Quantitative Analysis There is seldom a unique way to design a measurement process. Even an explicitly defined analysis can be approached in more than one ways. Dif- ferent studies have different purposes, different financial constraints, and are carried out by staff with different expertise and personal preferences. The most important step in a study design is the determination of the purpose, and at least a notion of the final results. It should yield data that provide useful information to solve the problem at hand The objective of an analytical measurement can be qualitative or quanti- tative. For example, the presence of pesticide in fish is a topic of concern. The questions may be: Are there pesticides in fish? If so, which ones? An analysis designed to address these questions is a qualitative analysis, where the analyst screens for the presence of certain pesticides. The next obvious question is: How much pesticide is there? This type of analysis, quantitative analysis, not only addresses the presence of the pesticide, but also its con- entration. The other important category is semiqualitative analysis. Here

all aimed at producing those few microliters that represent what is in the fish. It is obvious that an error in the first three steps cannot be rectified by even the most sophisticated analytical instrument. So the importance of the prior steps, in particular the sample preparation, cannot be understressed. 1.1.1. Qualitative and Quantitative Analysis There is seldom a unique way to design a measurement process. Even an explicitly defined analysis can be approached in more than one ways. Dif￾ferent studies have di¤erent purposes, di¤erent financial constraints, and are carried out by sta¤ with di¤erent expertise and personal preferences. The most important step in a study design is the determination of the purpose, and at least a notion of the final results. It should yield data that provide useful information to solve the problem at hand. The objective of an analytical measurement can be qualitative or quanti￾tative. For example, the presence of pesticide in fish is a topic of concern. The questions may be: Are there pesticides in fish? If so, which ones? An analysis designed to address these questions is a qualitative analysis, where the analyst screens for the presence of certain pesticides. The next obvious question is: How much pesticide is there? This type of analysis, quantitative analysis, not only addresses the presence of the pesticide, but also its con￾centration. The other important category is semiqualitative analysis. Here Homogenization, Size reduction Analysis Extraction Concentration Clean-up Figure 1.2. Possible steps within sample preparation. the measurement process 3

SAMPLE PREP ARATION: AN ANALYTICAL PERSPECTIVE Table 1.1. Common Instrumental Methods and the Necessary Sample Preparation teps Prior to Analy Analytes Sample Preparation Instrument Organics Extraction, concentration, GC, HPLC, GC/MS, LC/MS Volatile organics Transfer to vapor phase. GC GC-MS Metals Extraction, concentration, AA. GFAA. ICP ICP/MS tion Metals Extraction, derivatization, UV-VIS molecular absorp- concentration, specia- tion spectrophotometr tion ion chromatography Extraction, concentration, IC, UV-VIs derivatization DNA/RNA Cell lysis, extraction, PCR Electrophoresis, UV-VIS Amino acids, fats Extraction, cleanup GC, HPLC, electrophoresis carbohydrat Microstruc Etching, polishing, reac- Microscopy, surface spectros- tive ion techniques, ion copy GC, gas chromatography; HPLC, high-performance liquid chromatography; MS, mass spec AA, atomic absorption; GFAA, graphite furnace atomic absorption; ICP, inductively plasma: UV-VIS, ultraviolet-visible molecular absorption spectroscopy; IC, ion chro- the concern is not exactly how much is there but whether it is above or below a certain threshold level. The prostate specific antigen(PSa)test for the screening of prostate cancer is one such example. A PSA value of 4 ng/L (or higher) implies a higher risk of prostate cancer. The goal here is to determine if the PSa is higher or lower then 4 ng/L Once the goal of the analyses and target analytes have been identified, the methods available for doing the analysis have to be reviewed with an eye to accuracy, precision, cost, and other relevant constraints. The amount of labor, time required to perform the analysis, and degree of automation can also be important. I.1. 2. Methods of Quantitation Almost all measurement processes, including sample preparation and anal ysis, require calibration against chemical standards. The relationship be- tween a detector signal and the amount of analyte is obtained by recording

the concern is not exactly how much is there but whether it is above or below a certain threshold level. The prostate specific antigen (PSA) test for the screening of prostate cancer is one such example. A PSA value of 4 ng/L (or higher) implies a higher risk of prostate cancer. The goal here is to determine if the PSA is higher or lower then 4 ng/L. Once the goal of the analyses and target analytes have been identified, the methods available for doing the analysis have to be reviewed with an eye to accuracy, precision, cost, and other relevant constraints. The amount of labor, time required to perform the analysis, and degree of automation can also be important. 1.1.2. Methods of Quantitation Almost all measurement processes, including sample preparation and anal￾ysis, require calibration against chemical standards. The relationship be￾tween a detector signal and the amount of analyte is obtained by recording Table 1.1. Common Instrumental Methods and the Necessary Sample Preparation Steps Prior to Analysis Analytes Sample Preparation Instrumenta Organics Extraction, concentration, cleanup, derivatization GC, HPLC, GC/MS, LC/MS Volatile organics Transfer to vapor phase, concentration GC, GC-MS Metals Extraction, concentration, speciation AA, GFAA, ICP, ICP/MS Metals Extraction, derivatization, concentration, specia￾tion UV-VIS molecular absorp￾tion spectrophotometry, ion chromatography Ions Extraction, concentration, derivatization IC, UV-VIS DNA/RNA Cell lysis, extraction, PCR Electrophoresis, UV-VIS, florescence Amino acids, fats carbohydrates Extraction, cleanup GC, HPLC, electrophoresis Microstructures Etching, polishing, reac￾tive ion techniques, ion bombardments, etc. Microscopy, surface spectros￾copy aGC, gas chromatography; HPLC, high-performance liquid chromatography; MS, mass spec￾troscopy; AA, atomic absorption; GFAA, graphite furnace atomic absorption; ICP, inductively coupled plasma; UV-VIS, ultraviolet–visible molecular absorption spectroscopy; IC, ion chro￾matography. 4 sample preparation: an analytical perspective

THE MEASUREMENT PROCESS 5 the response from known quantities. Similarly, if an extraction step is in- volved, it is important to add a known amount of analyte to the matrix and measure its recovery. Such processes require standards, which may be pre- pared in the laboratory or obtained from a commercial source. An impor tant consideration in the choice of standards is the matrix. For some ana lytical instruments, such as x-ray fluorescence, the matrix is very important, but it may not be as critical for others. Sample preparation is usually matrix dependent. It may be easy to extract a polycyclic aromatic hydrocarbon from sand by supercritical extraction but not so from an aged soil with a Calibration curves The most common calibration method is to prepare standards of known concentrations, covering the concentration range expected in the sample The matrix of the standard should be as close to the samples as possible. For instance, if the sample is to be extracted into a certain organic solvent, the standards should be prepared in the same solvent. The calibration curve is a plot of detector response as a function of concentration. A typical calibra tion curve is shown in Figure 1.3. It is used to determine the amount of analyte in the unknown samples. The calibration can be done in two ways, best illustrated by an example. Let us say that the amount of lead in soil is being measured. The analytical method includes sample preparation by acid extraction followed by analysis using atomic absorption(AA). The stan- 3 吧 LoD(3×S/N) LOQ(10×S/N) Analyte concentration Figure 1.3. Typical calibration curve

the response from known quantities. Similarly, if an extraction step is in￾volved, it is important to add a known amount of analyte to the matrix and measure its recovery. Such processes require standards, which may be pre￾pared in the laboratory or obtained from a commercial source. An impor￾tant consideration in the choice of standards is the matrix. For some ana￾lytical instruments, such as x-ray fluorescence, the matrix is very important, but it may not be as critical for others. Sample preparation is usually matrix dependent. It may be easy to extract a polycyclic aromatic hydrocarbon from sand by supercritical extraction but not so from an aged soil with a high organic content. Calibration Curves The most common calibration method is to prepare standards of known concentrations, covering the concentration range expected in the sample. The matrix of the standard should be as close to the samples as possible. For instance, if the sample is to be extracted into a certain organic solvent, the standards should be prepared in the same solvent. The calibration curve is a plot of detector response as a function of concentration. A typical calibra￾tion curve is shown in Figure 1.3. It is used to determine the amount of analyte in the unknown samples. The calibration can be done in two ways, best illustrated by an example. Let us say that the amount of lead in soil is being measured. The analytical method includes sample preparation by acid extraction followed by analysis using atomic absorption (AA). The stan- 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 Analyte concentration Signal LOQ (10 × S/N) LOD (3 × S/N) Limit of linearity Figure 1.3. Typical calibration curve. the measurement process 5

6 SAMPLE PREPARATION: AN ANALYTICAL PERSPECTIVE dards can be made by spiking clean soil with known quantities of lead. Then he standards are taken through the entire process of extraction and analysis Finally, the instrument response is plotted as a function of concentration. The other option assumes quantitative extraction, and the standards are sed to calibrate only the aa. The first approach is more accurate; the latter simpler. A calibration method that takes the matrix effects into account is he method of standard addition, which is discussed briefly in Chapter 4 1.2. ERRORS IN QUANTTTATTVE ANALYSIS ACCURACY AND PRECISION All measurements are accompanied by a certain amount of error, and an estimate of its magnitude is necessary to validate results. The error cannot be eliminated completely, although its magnitude and nature can be char- acterized. It can also be reduced with improved techniques. In general errors can be classified as random and systematic. If the same experiment repeated several times, the individual measurements cluster around the mean value. The differences are due to unknown factors that are stochastic in nature and are termed random errors. they have a gaussian distribution and equal probability of being above or below the mean. On the other hand, systematic errors tend to bias the measurements in one direction. Systematic error is measured as the deviation from the true value 1. 2.1. Accuracy Accuracy, the deviation from the true value, is a measure of systematic error. It is often estimated as the deviation of the mean from the true value mean-true val accurac rue value The true value may not be known. For the purpose of comparison, mea surement by an established method or by an accredited institution is ac epted as the true value 1.2. 2. Precision Precision is a measure of reproducibility and is affected by random error. Since all measurements contain random error, the result from a single mea surement cannot be accepted as the true value. An estimate of this error is necessary to predict within what range the true value may lie, and this is done

dards can be made by spiking clean soil with known quantities of lead. Then the standards are taken through the entire process of extraction and analysis. Finally, the instrument response is plotted as a function of concentration. The other option assumes quantitative extraction, and the standards are used to calibrate only the AA. The first approach is more accurate; the latter is simpler. A calibration method that takes the matrix e¤ects into account is the method of standard addition, which is discussed briefly in Chapter 4. 1.2. ERRORS IN QUANTITATIVE ANALYSIS: ACCURACY AND PRECISION All measurements are accompanied by a certain amount of error, and an estimate of its magnitude is necessary to validate results. The error cannot be eliminated completely, although its magnitude and nature can be char￾acterized. It can also be reduced with improved techniques. In general, errors can be classified as random and systematic. If the same experiment is repeated several times, the individual measurements cluster around the mean value. The di¤erences are due to unknown factors that are stochastic in nature and are termed random errors. They have a Gaussian distribution and equal probability of being above or below the mean. On the other hand, systematic errors tend to bias the measurements in one direction. Systematic error is measured as the deviation from the true value. 1.2.1. Accuracy Accuracy, the deviation from the true value, is a measure of systematic error. It is often estimated as the deviation of the mean from the true value: accuracy ¼ mean
true value true value The true value may not be known. For the purpose of comparison, mea￾surement by an established method or by an accredited institution is ac￾cepted as the true value. 1.2.2. Precision Precision is a measure of reproducibility and is a¤ected by random error. Since all measurements contain random error, the result from a single mea￾surement cannot be accepted as the true value. An estimate of this error is necessary to predict within what range the true value may lie, and this is done 6 sample preparation: an analytical perspective

ERRORS IN QUANTITATTVE ANALYSIS: ACCURACY AND PRECISION 7 by repeating a measurement several times [1]. Two important parameters, the average value and the variability of the measurement, are obtained from this process. The most widely used measure of average value is the arithmetic mean di where 2 xi is the sum of the replicate measurements and n is the total number of measurements Since random errors are normally distributed, the common measure of variability (or precision) is the standard deviation,o This is calculated as When the data set is limited, the mean is often approximated as the true alue, and the standard deviation may be underestimated. In that case, the unbiased estimate of o, which is designated s, is computed as follows: ∑(x As the number of data points becomes larger, the value of s approaches that of o. When n becomes as large as 20, the equation for o may be used. Another term commonly used to measure variability is the coefficient of variation(Cv) or relative standard deviation(RSD), which may also be RSD or %o RSD 100 (1.3) Relative standard deviation is the parameter of choice for expressing preci- sion in analytical sciences. Precision is particularly important when sample preparation is involved The variability can also affect accuracy. It is well known that reproduci- bility of an analysis decreases disproportionately with decreasing concen- tration [2. A typical relationship is shown in Figure 1. 4, which shows that the uncertainty in trace analysis increases exponentially compared to the major and minor component analysis. Additional deviations to this curve are expected if sample preparation steps are added to the process. It may be prudent to assume that uncertainty from sample preparation would Iso increase with decrease in concentration rally speaking, analytic

by repeating a measurement several times [1]. Two important parameters, the average value and the variability of the measurement, are obtained from this process. The most widely used measure of average value is the arithmetic mean, x: x ¼ P xi n where P xi is the sum of the replicate measurements and n is the total number of measurements. Since random errors are normally distributed, the common measure of variability (or precision) is the standard deviation, s. This is calculated as s ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Pðxi
xÞ 2 n s ð1:1Þ When the data set is limited, the mean is often approximated as the true value, and the standard deviation may be underestimated. In that case, the unbiased estimate of s, which is designated s, is computed as follows: s ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Pðxi
xÞ 2 n
1 s ð1:2Þ As the number of data points becomes larger, the value of s approaches that of s. When n becomes as large as 20, the equation for s may be used. Another term commonly used to measure variability is the coe‰cient of variation (CV ) or relative standard deviation (RSD), which may also be expressed as a percentage: RSD ¼ s x or % RSD ¼ s x 100 ð1:3Þ Relative standard deviation is the parameter of choice for expressing preci￾sion in analytical sciences. Precision is particularly important when sample preparation is involved. The variability can also a¤ect accuracy. It is well known that reproduci￾bility of an analysis decreases disproportionately with decreasing concen￾tration [2]. A typical relationship is shown in Figure 1.4, which shows that the uncertainty in trace analysis increases exponentially compared to the major and minor component analysis. Additional deviations to this curve are expected if sample preparation steps are added to the process. It may be prudent to assume that uncertainty from sample preparation would also increase with decrease in concentration. Generally speaking, analytical errors in quantitative analysis: accuracy and precision 7