Basics of Analytical Chemistry and Chemical Equilibria A Quantitative Approach 2Nd Edition Brian M Tissue Online Ebook Texxtbook Full Chapter PDF
Basics of Analytical Chemistry and Chemical Equilibria A Quantitative Approach 2Nd Edition Brian M Tissue Online Ebook Texxtbook Full Chapter PDF
Basics of Analytical Chemistry and Chemical Equilibria A Quantitative Approach 2Nd Edition Brian M Tissue Online Ebook Texxtbook Full Chapter PDF
https://ebookmeta.com/product/chemistry-in-quantitative-language-
fundamentals-of-general-chemistry-calculations-2nd-edition-
oriakhi/
https://ebookmeta.com/product/archaeology-the-basics-4th-edition-
brian-m-fagan/
https://ebookmeta.com/product/world-prehistory-the-basics-1st-
edition-brian-m-fagan/
https://ebookmeta.com/product/practical-soft-tissue-pathology-a-
diagnostic-approach-2nd-edition-jason-l-hornick/
Statistics for Chemical and Process Engineers: A Modern
Approach, 2nd Edition Shardt
https://ebookmeta.com/product/statistics-for-chemical-and-
process-engineers-a-modern-approach-2nd-edition-shardt/
https://ebookmeta.com/product/the-thermodynamics-of-phase-and-
reaction-equilibria-2nd-edition-ismail-tosun/
https://ebookmeta.com/product/atlas-of-wound-healing-a-tissue-
regeneration-approach-soheila-s-kordestani/
https://ebookmeta.com/product/chemical-reaction-engineering-a-
computer-aided-approach-2nd-edition-tapio-salmi/
https://ebookmeta.com/product/quantitative-chemical-analysis-
tenth-edition-daniel-c-harris-charles-a-lucy/
BASICS OF ANALYTICAL
CHEMISTRY AND
CHEMICAL EQUILIBRIA
BASICS OF ANALYTICAL
CHEMISTRY AND
CHEMICAL EQUILIBRIA: A
QUANTITATIVE APPROACH
By
BRIAN M. TISSUE
Virginia Tech
Department of Chemistry
Blacksburg, VA
Second Edition
Copyright © 2023 by John Wiley & Sons, Inc. All rights reserved.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form
or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as
permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior
written permission of the Publisher, or authorization through payment of the appropriate per-copy fee
to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750–8400, fax
(978) 750–4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be
addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030,
(201) 748–6011, fax (201) 748–6008, or online at http://www.wiley.com/go/permission.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts
in preparing this book, they make no representations or warranties with respect to the accuracy
or completeness of the contents of this book and specifically disclaim any implied warranties of
merchantability or fitness for a particular purpose. No warranty may be created or extended by sales
representatives or written sales materials. The advice and strategies contained herein may not be suitable
for your situation. You should consult with a professional where appropriate. Neither the publisher nor
author shall be liable for any loss of profit or any other commercial damages, including but not limited to
special, incidental, consequential, or other damages. Further, readers should be aware that websites listed
in this work may have changed or disappeared between when this work was written and when it is read.
Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages,
including but not limited to special, incidental, consequential, or other damages.
For general information on our other products and services or for technical support, please contact our
Customer Care Department within the United States at (800) 762–2974, outside the United States at (317)
572–3993 or fax (317) 572–4002.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may
not be available in electronic formats. For more information about Wiley products, visit our web site at
www.wiley.com.
A catalogue record for this book is available from the Library of Congress
Set in 11/13pt TimesNewRomanPSMT by Integra Software Services Pvt. Ltd, Pondicherry, India
CONTENTS
PREFACE ix
1 MAKING MEASUREMENTS 3
1.1 Introduction / 3
1.2 Glp and a Few Other Important Acronyms / 10
1.3 Precision and Random Error / 14
1.4 Discarding a Suspected Outlier / 25
1.5 Calibration / 28
1.6 Maintaining Accurate Results / 44
Practice Exercises / 49
2 SAMPLE PREPARATION, EXTRACTIONS, AND
CHROMATOGRAPHY 53
2.1 Sampling and Control Samples / 53
2.2 Sample Preparation / 59
2.3 Solvents and Solutions / 65
2.4 Introduction to Solubility / 69
2.5 Extraction and Partitioning Theory / 72
v
vi contents
INDEX 469
PREFACE
This text will introduce you to analytical chemistry: the science of making
quantitative measurements. Quantifying the individual components in a com-
plex sample is an exercise in problem solving. An effective and efficient analyst
will have expertise in
ix
x preface
The beginning of each chapter lists learning outcomes that serve as a brief
outline to help categorize new material. After completing a chapter, make a
concept map to help yourself see the big picture and underlying concepts.
You will often encounter a repeat of concepts in the text. Making connec-
tions with prior material makes learning analytical concepts much easier.
Treating every topic as something new becomes overwhelming. Each chapter
preface xi
contains sample calculations and practice exercises. I assume that your goal
is success. Achieving success requires skills, and acquiring skills takes
practice.
Variables and constants are italicized to not be confused with other text. As
much as possible, I use the conventions and terminology in the IUPAC Gold
Book.1 You will find other symbols in other books and resources, so use the
context to decipher the differences. This second edition adds substantial
information on instrumental methods and manufacturers will often use differ-
ent terms for similar instruments. Relevant spreadsheets and links to useful
resources are available at https://www.achem.org.
1
See IUPAC Compendium of Chemical Terminology, “The Gold Book,” https://goldbook.iupac.
org; Accessed August 2022.
ABOUT THE COMPANION WEBSITE
www.wiley.com/go/tissue/analyticalchemistry2e
xiii
PART I
MAKING MEASUREMENTS
Learning Outcomes
1.1 INTRODUCTION
There are few areas in our modern life in which the quantity of substances is
not important. Industries and government agencies spend substantial
resources to determine and monitor the safe levels of chemicals in foods, phar-
maceuticals, and the environment. Setting permissible levels of contaminants
is based on quantitative results from toxicological studies and raising or low-
ering a level has significant costs and consequences. Similarly, companies
compete for sales by providing high-quality goods at the lowest price.
Optimizing industrial processes depends on making decisions based on ana-
lytical measurements. Poor measurements or incorrect data interpretation will
lead to poor decisions.
Basics of Analytical Chemistry and Chemical Equilibria: A Quantitative Approach, Second Edition. Brian M. Tissue.
© 2023 John Wiley & Sons, Inc. Published 2023 by John Wiley & Sons, Inc.
Companion Website: www.wiley.com/go/tissue/analyticalchemistry2e
3
4 making measurements
You might not make many measurements yourself, but you probably rely on
data and quantitative results to make decisions. You’ve probably read the
ingredients or nutritional information on a product label to choose one prod-
uct over another. I certainly want manufacturers to perform quality checks on
the contents of the products that I buy. I’m also expecting an independent
agency, say the FDA or USDA,1 to check that there is not too much of a min-
eral, or contaminants such as Pb or rat poison, that could make the food
unhealthy. Think about the last time that you had a medical checkup. Did the
doctor determine your health by just looking at you? At the least you had a
quantitative measure of pulse rate and blood pressure. Modern medicine relies
on a variety of technological tools and clinical analyses. At some time, you
might need to make a significant decision such as beginning daily doses of a
cholesterol-lowering drug. We all hope that doctors and clinical technicians
analyzing our samples were paying attention when they took an analytical
chemistry course!
When you make a measurement, or you need to make a decision based on
someone else’s measurement, do you trust the value? This chapter introduces
the terminology and statistical tools to describe and assess quantitative results.
Some of the details will be new to you, but they all fit into a framework for
collecting and reporting quantitative measurements. Table 1.1 begins building
our vocabulary of measurement science and data-handling concepts by
defining general terms. Many of these terms are used rather loosely in the
scientific and manufacturer literature. You might need to dig into the details to
know exactly what is meant when a procedure refers to the sample, the signal,
etc. It is also common that a given term will have a different definition for dif-
ferent instruments or techniques. Resolving power is a measure of selectivity
in mass spectrometry. However, resolving power and the related term resolu-
tion have different meanings when discussing a spectrum, a microscope image,
or the separation of components in a mixture.
In quantitative analysis, we want a measurement or detector signal that we
can relate to an analyte concentration. Doing so can be quite involved, and
Chapter 2 discusses various sample preparation methods to isolate an analyte
from interferences so that it can be measured. What mechanisms are available
to detect an analyte? I can think of only three general strategies to detect and
quantitate an analyte:
1
US Food and Drug Administration and US Department of Agriculture. See “FDA
Fundamentals,” https://www.fda.gov/about-fda/fda-basics/fda-fundamentals; accessed August
2022. Other countries and trade blocs have similar agencies.
introduction 5
Table 1.2 lists some examples in each of these categories. We will discuss
most of these methods, so do not worry if they are unfamiliar. Chapter 3 dis-
cusses classical methods that rely on physical measurements and Part III of
the text introduces instrumental methods based on electrochemistry, spectros-
copy, and mass spectrometry.
Although I list only three general detection strategies, each of these general
categories encompass a multitude of specific analytical techniques. For example,
spectroscopic methods have been developed to use most of the electromagnetic
spectrum, including X-ray, ultraviolet (UV), visible (Vis), infrared, and radio
waves. The different regions of the electromagnetic spectrum interact with
matter differently and provide different types of information. This text concen-
trates on quantitative methods for analytes in aqueous solution. There are
6 making measurements
2
EAS (Eastern Analytical Symposium), https://eas.org; FACSS (Federation of Analytical
Chemistry and Spectroscopy Societies), https://facss.org; or Pittcon (Pittsburgh Conference on
Analytical Chemistry and Applied Spectroscopy), https://pittcon.org; URLs accessed August 2022.
3
SI is the abbreviation for “The International System of Units.” For more information see
https://physics.nist.gov/cuu/Units/index.html; accessed August 2022.
introduction 7
4
Note that parts per thousand is sometimes abbreviated as ppt in older literature. It should not
be confused with parts per trillion.
5
US Environmental Protection Agency, https://www.epa.gov; accessed August 2022.
6
Solid and liquid particles less than 2.5 μm in diameter suspended in air. US Envi
ronmental Protection Agency, National Ambient Air Quality Standards, https://www.epa.gov/
environmental-topics/air-topics; accessed August 2022.
introduction 9
Example 1.1 Unit Conversion. The EPA maximum contaminant limit (MCL)
for benzene, C6H6, in drinking water is 0.005 mg/l. What is this limit written in
units of ppm, ppb, M, and wt%?
The question does not specify the density of the water. Given that drinking
water is purified, we will take it to be 1.0 g/ml or 1.0 kg/l. As we have only one
significant figure in 0.005 mg/l, this assumption does not affect this calcula-
tion.7 Given a density of 1.0 kg/l, the units mg/l and ppm are the same:
0.005 mg C 6 H6 1.0 l
= 0.005 ppm C 6 H6
1.0 l water 1.0 kg
1000 ppb
0.005 ppm C 6H6 = 5 ppbC 6H6
1ppm
5 ×10−6 g C 6 H6
= 6.4 ×10−8 molC 6 H6
78.11g/mol
To find the weight percent, wt%, we need grams of benzene per 100 g of solu-
tion. The MCL of 0.005 mg/l for a water density of 1.0 g/ml is 0.005 mg/1000 g.
I multiply both numerator and denominator by 0.1 to get the units that I want
in the denominator:
As you can see, 0.005 mg/l or 5 ppb is simply more convenient for describing
this limit than other concentration units.
You-Try-It 1.A
The conversions worksheet in you-try-it-01.xlsx contains two tables
of measurement results. Convert these results to other common units.
The different regulations are similar in their overall structure and purpose,
but they are tailored to specific types of chemicals and laboratories. The details
of the Code of Federal Regulations are not important to us, but they are use-
ful to recognize the origin of regulations that we will discuss. Reference to
sections of the federal code use an abbreviation based on title, part, and sec-
tion number. The following paragraph illustrates the coding with a small
excerpt of the Code of Federal Regulations, 21CFR58.83, where the .83 in the
title refers to this specific section.
8
Available from US Government Printing Office: https://www.ecfr.gov; accessed August 2022.
glp and a few other important acronyms 11
Besides codifying good science and common sense as federal law, the GLP
regulations stipulate a framework for personnel responsibilities, analytical
methods, and record keeping. These tasks include
With the formalized GLP regulations, quality assurance (QA) now indicates
an auditing role, while quality control (QC) refers to instrument calibration and
method validation (discussed in Section 1.5). My point in this discussion is not
that we must memorize government regulations but that all workers in an analyt-
ical laboratory have certain roles. Laboratory workers must follow documented
and reproducible procedures and create an audit trail of all work. Following
GLP ensures the reliability of analytical measurements and maintains the
credibility of reported results. An analyst should develop “GLP” habits-of-mind
to become adept at measurement science and to work safely in the laboratory.
1. Purpose
To ensure correct and safe usage of the FL-400 Flame-Resistant
Fiber Probe when recording flame emission spectra.
2. Scope
This SOP provides operating procedures for the Ocean Optics FL-400
Flame-Resistant Fiber Probe for recording flame emission spectra.
Consult other SOPs or your instructor for sample handling and use
of the flame source and spectrometer.
3. Precautions
● Do not touch the tip when hot.
● Handle the probe with care. It contains a glass fiber and should
not be dropped or bent.
● Let tip cool for at least 10 seconds before placing in sample or
cleaning solutions. Do not place a hot probe tip in a cool solution
as the glass fiber could crack.
4. Specific Procedures
4.A. Set-up: The flame-resistant fiber probe should be connected to a
1-m optical patch cable using the stainless steel SMA splice bushing
stored with the probe. Do not attempt to use the probe connected
directly to the USB2000® spectrometer.
4.B. Use: To load a test portion onto the flame loop (1) dip into a
solution and allow to dry or (2) wet the loop with dilute HCl and dip
into solid sample. Insert only the loop into a flame and not the whole
tip. A separate sample loop or splint may also be used.
4.C. Cleaning: The flame loop and fiber tip should be cleaned with
distilled water. A mild detergent and ultrasound is acceptable if
necessary.
The point of this list is not that you should memorize acronyms. The
purpose is that you can recognize the purpose and origin of regulations when
you see them in the context of your work or study. Many of these regulatory
structures are available online, and a search will often find all that you need to
know about the details of the regulations. The chances are high that you will
deal with scientific regulations sometime during your work life, possibly in
more than one laboratory role as analyst, supervisor, or auditor.
9
See for example the Toxicology Procedures Manual available at: https://www.dfs.virginia.gov/
documentation-publications/manuals; accessed August 2022.
10
GCP and GMP are both administered by the Food and Drug Administration in the United
States.
11
“OECD Principles on Good Laboratory Practice” and “Guidelines for Testing of Chemicals”
can be found by searching at https://www.oecd.org; accessed August 2022.
14 making measurements
equipment (PPE), and work practices to work with a chemical safely. A lab
worker should read the SDS information before using a substance, that is,
before spilling, inhaling, or igniting anything. It is very difficult to read an
SDS when you are dizzy, burned, or unconscious. Some of the more useful
sections of an SDS include
SDS sheets must be readily available to lab workers. Medical personnel can
refuse to treat a case of chemical exposure without an SDS for the substance
involved. If you do not have an SDS, most chemical suppliers provide them
online. The following excerpt illustrates the type of information that is avail-
able for hydrochloric acid.
measurements can identify gross errors such as omitting one step in a procedure,
one-time instrument glitches, or writing a value incorrectly. Even when making
individual measurements, an analytical method will specify remeasuring a stan-
dard or sample at some frequency. The frequency could be twice daily, one
duplicate measurement for each batch of samples, or once every 10 measure-
ments. The frequency depends on the stability that is expected for a given method
or instrument. Making duplicate measurements is especially useful to identify
drift. Drift is the gradual change in instrument response over time. It introduces
a bias in measurements that gets worse with time. Observing drift in a duplicate
measurement can indicate that a method or instrument requires recalibration.
The repeatability in making replicate measurements is called precision.12 We
use the term repeatability when we are talking about replicate measurements
made on the same sample and performed under identical conditions. When we
compare measurements of the same sample by different analysts or different
methods, we use the term reproducibility. The calculation of precision is the
same for repeatability and reproducibility, the difference is the source of the
measurements and the purpose of the result. Quantitative measures of preci-
sion include standard deviation, standard error, and confidence limits (CL).
These measures quantitate the variation or spread in the individual measure-
ments due to random fluctuations, which we call random errors. The distribu-
tion of random fluctuations follows a Gaussian-shape “bell curve,” and
precision is a measure of the width of this distribution. Graphically, we display
the precision by placing error bars about a data point.
The accuracy of a measurement is how close an experimental result comes
to the true value. The difference between a measurement and the true value is
called the systematic error or bias. We determine the bias in a measurement by
measuring samples of known composition, which we call standards. We will
discuss accuracy and systematic errors fully in Section 1.5 on calibration,
which is the process of measuring standards to be able to obtain accurate
results for unknowns.
We can never know with 100% certainty if we have determined the true
concentration of an analyte in a real sample. The accuracy of a measurement
depends on the care in validating sample processing procedures, calibrating
the measurement, and making sure that standards match the samples. The best
that we can do, by following GLP, is to
13
The EPA action level is the concentration of a contaminant that requires a response such as
public notification, exposure monitoring, or remediation.
precision and random error 17
the same. The same goes for analytical measurements. If you want repeatable
results, keep all conditions the same. The archer on the right just needs more
practice to get steadier. As in any human endeavor, achieving high accuracy
and precision requires practice and care.
s=
∑( xi − x )2 (1.3)
n −1
where n is the number of data points and n − 1 is called the degrees of free-
dom.14 By squaring each deviation before summing them, we eliminate the
14
The degrees of freedom is the number of measurements minus the number of results obtained
from the measurements. When calculating the mean of a data set, it is n − 1. On obtaining the
slope and y-intercept of a line, it is n − 2. Taylor, J. R. An Introduction to Error Analysis: The
Study of Uncertainties in Physical Measurements, 2nd ed.; University Science Books: Sausalito,
CA, 1996.
18 making measurements
problem of an average deviation going to zero. Taking the square root of the
summed squares gets the standard deviation back to the scale of the mean.
The standard deviation provides a “typical” deviation for a series of measure-
ments. We will see later that this typical deviation encompasses approximately
68% of all measurements in a set of data.
The equation for s is used to describe the precision of a relatively small number
of replicate measurements. For a large number of measurements, say 20 or more
for a reliable procedure, we can use the true or population standard deviation, σ:
σ=
∑( xi − µ )2 (1.4)
n
Concentration (M)
0.88
0.86
0.84
0.82
the true value. Students B and C were both accurate, their means were very
close to the true value. Compared to Student B, Student C’s measurements
were less precise. The error bars on her results were larger than for the other
two students, leading to greater uncertainty in her reported result. Given a
choice, I will take a large error bar on an accurate result in place of a very
repeatable wrong answer.
15
The relative standard deviation is also called the “coefficient of variation” in some disciplines.
20 making measurements
Example 1.2 Percentage Relative Standard Deviation. The table lists the
results from four students who each measured a different sample using the
identical analytical procedure. Let us see which measurement result was most
precise.
A 1.03 0.01
B 10.03 0.01
C 3.58 0.01
D 0.11 0.01
Each measurement has the same standard deviation, but each measurement
was not made with equal precision. Calculating the standard deviation as
%-RSD for Student A gives
0.01 ppm
%-RSD A = ×100% = 1%
1.03 ppm
Repeating this calculation for the other students and adding the results to the
table shows clearly which measurements are most and which are least precise.
The result from Student B was measured to 0.1%, much better than the ones
the other students achieved. We cannot say anything about the accuracy of the
results based only on this data, but Student B certainly had a more repeatable
technique in performing the measurement.
measurement. The 68% range of the standard deviation is not always the best
measure to state the degree of confidence that we have in the repeatability of a
measurement or in the spread in results from multiple samples. Several other
quantities—variance, standard error, and CL—are described here for com-
pleteness. Each of these measures of precision derive from the standard
deviation and are useful for different purposes.
1.3.3.1 Variance The variance is simply the square of the standard deviation:
n
∑( xi − x )2
v = s2 = i =1
(1.7)
n −1
The advantage of working with variance is that variances from independent
sources of variation may be summed to obtain a total variance for a
measurement. The topic of the propagation of uncertainty is beyond the needs
of this text, but it is treated in several other sources on practical statistics.16
1.3.3.2 Standard Error The standard error, sm, also called the standard
deviation of the mean (SDOM), is the standard deviation divided by the square
root of n:
s
sm = (1.8)
n
By dividing by n, the sm provides a better reflection of the amount of data
collected than other measures of precision. Taylor considers it the “best” mea-
sure of the uncertainty in a set of multiple measurements.17 A significant
difference between the standard error and the standard deviation is that the
standard error will decrease with increasing number of measurements,
although slowly for large n.
16
Taylor, J. R. An Introduction to Error Analysis: The Study of Uncertainties in Physical
Measurements, 2nd ed.; University Science Books: Sausalito, CA, 1996; also useful are Bevington
P.; Keith Robinson, D. K. Data Reduction and Error Analysis for the Physical Sciences, 3rd ed.;
McGraw-Hill: New York, 2002 and Garland, C. W.; Nibler, J. W.; Shoemaker, D. P. Experiments
in Physical Chemistry, 7th ed.; McGraw-Hill: New York, 2003.
17
Taylor, J. R. An Introduction to Error Analysis: The Study of Uncertainties in Physical
Measurements, 2nd ed.; University Science Books: Sausalito, CA, 1996.
22 making measurements
To answer these questions, we use a more definitive description for the impre-
cision of repetitive measurements.
Consider again our example of lead in drinking water, for which the EPA
has set an action level of 15 ppb Pb. Suppose that measurements from a
random sampling of water in homes show a mean and standard deviation of
14 ± 1 ppb. For this municipal water supply, 68% of the measurements fall bet-
ween 13 and 15 ppb. 16% of the water samples contain less than 13 ppb Pb,
which is not an issue for the quality and safety of the drinking water. However,
16% of the water samples contain Pb greater than the 15 ppb EPA action level.
In this case, reporting the standard deviation might not be the best measure of
precision on which to base water treatment decisions. If I lived in this locality,
18
The histogram data was generated in Excel using Random Number Generation in the Data
Analysis Toolpak and the curve was generated with the NORMDIST function.
precision and random error 23
I might not want to take the risk that my tap water is in the 16% of cases with
high Pb levels. So the question is: To what level should the average lead level be
reduced?
In such cases, we introduce the “level of confidence” to specify the proba-
bility of a value being within or outside of a given distribution of measure-
ments. CL is a statistical measure of the precision for replicate measurements
that can be calculated for different levels of confidence. It is calculated from
the standard deviation, s, using
t×s
CL = (1.9)
n
19
NIST/SEMATECH e-Handbook of Statistical Methods; https://www.itl.nist.gov/div898/
handbook/eda/section3/eda3672.htm; accessed August 2022.
24 making measurements
where CI, confidence interval, is the span of values that encompasses the true
value with the stated level of confidence (see Example 1.3).
Table 1.8 also introduces the α notation, the values in parentheses, which is
called the significance level. Referring to a significance of α = 0.05 or a
confidence level of 0.95 (or 95%) is the same for our purposes, and the symbol
for confidence level is 1 − α.
Returning to the lead in the drinking water example, if the 14.0 ppb mean
was the result of six measurements, the 99% CL is
2.78×1.0 ppb
CL = = 1.2 ppb (1.11)
5
2.09×1.0 ppb
CL = = 0.5 ppb (1.12)
20
In these two cases, the number of measurements affects our level of confidence.
With only six water samples the range of 14.0 ± 1.2 ppb encompasses the EPA
action level, but for 21 measurements the result of 14.0 ± 0.5 ppb is safely
below the action level.
Example 1.3 CI Calculation. What is the 95-% CI for the following four tri-
als of an iron analysis of a soil?
Weight Percent of Fe in Soil
Test Portion Result
1 3.44% Fe
2 3.11% Fe
3 2.98% Fe
4 3.27% Fe
The sum of the four data points is 12.80, the mean is 3.20, and the standard
deviation is 0.20 (determined from a spreadsheet).
Using the expression for CL and a t value from Table 1.8 for n − 1 of 3:
(3.182)(0.20)
CL = = 0.32
4
The CI at 95% confidence is (3.2 ± 0.3)%.
You-Try-It 1.B
Replicate measurement data is in the precision worksheet in you-try-
it-01.xlsx. Open this worksheet and type formulas to find the mean,
discarding a suspected outlier 25
If one value in a set of data appears very different from the rest, it might be
possible to discard that value. The basis to discard a value is that some unrec-
ognized error causes that value to not be representative of the sample. There
are a number of criteria for removing an apparent outlier from a set of mea-
surements. This section will illustrate Dixon’s Q test and Peirce’s Criterion.
Note that performing a test is not necessary to discard data that you know is
erroneous. For example, if you were making a UV/Vis absorption measurement
and one solution was turbid due to a precipitate being disturbed from the bot-
tom of a reaction tube, you may make a note of your observation in your note-
book and discard the data value for that one experimental trial. Similarly, it is
not necessary to discard a suspected outlier, and the conservative approach is
to report all data with the larger standard deviation. If you do remove a point
from a set of data, you must acknowledge the change to your data and specify
your criterion for discarding the data.
If Q is larger than the critical value, Qc, for n repetitive measurements (see
Table 1.9), then the outlier may be discarded. This expression and the table of
Qc is only valid for discarding one data point from a series of measurements
(see Example 1.4). A more complete list at different levels of confidence can be
found in the literature.20
20
Adapted with permission from Rorabacher, D. B. “Statistical treatment for rejection of deviant
values: critical values of Dixon’s ‘Q’ parameter and related subrange ratios at the 95% confidence
level,” Anal. Chem. 1991, 63, 139–146. Copyright 1991 American Chemical Society.
26 making measurements
The value of 36.5 mV appears lower than the rest of the values in the set.
Can this value be discarded?
To answer this question, we find Q and compare it to Qc:
36.5 − 39.2
Q= = 0.794
39.9 − 36.5
At the 95% confidence level, which is usually suitable for working with analyti-
cal data sets, Qc = 0.710 for five data points. Our calculated Q is larger than Qc,
so the suspected outlier may be discarded. Recalculating the mean after dis-
carding the outlier gives a much smaller standard deviation 39.6 ± 0.3 mV,
compared to an original value of 39.0 ± 1.4 mV.
If the probability is very low, then the suspected outlier(s) may be discarded.
Using the data in the previous example, the mean and standard deviation of
the data are 39.0 ± 1.4 mV. The suspected outlier, 36.5 mV, is nearly two stan-
dard deviations from the mean. A value this distance from the mean should
arise in approximately 5% of measurements. We would expect a value such as
36.5 mV to appear in a set of more than 20 measurements, but not when we
have only 5 measurements.
Two similar methods that use this approach are Chauvenet’s criterion and
Peirce’s criterion. I will describe Peirce’s criterion because it is easier to imple-
ment using a published table of critical values (Table 1.10).21 This method
compares the deviation of the suspected outlier, that is, the difference from the
mean value, to the probability of a data point being that different from the
mean assuming a normal distribution of n measurements. Peirce’s criterion
can test multiple outliers, but we will use it here to test only one suspect value.
The procedure to use Peirce’s criterion is to first calculate the absolute value
of the deviation from the mean for a suspected outlier:
d i = xi − x
where xi is the suspected outlier and x is the mean. This deviation is then com-
pared to the maximum deviation, dmax, that is expected for a random distribu-
tion given the number of data values. dmax is found from
d max = sR
21
Adapted with permission from Ross, S. M. “Peirce’s criterion for the elimination of suspect
experimental data,” J. Eng. Technol. Fall 2003.
28 making measurements
where s is the sample standard deviation and R is taken from a table of values.
If the deviation of the suspected outlier is larger than dmax, it is improbable
that the outlier occurred owing to a random fluctuation. The suspected outlier
may be discarded if this condition is met:
| d i| > d max
You-Try-It 1.C
The above data is replicated in the outlier worksheet in you-try-it-01.
xlsx. Use this worksheet to test for the suspected outlier using Dixon’s
Q test at the 99% CL and using Peirce’s criterion. Use the = MIN(range)
and = MAX(range) functions to find the potential outliers in the data
set. Check that you get the same result as the Q-test calculation above.
1.5 CALIBRATION
The preceding section described the statistical tools to describe the repeat-
ability of replicate measurements. Calibration is the process of measuring a
known quantity to determine the relationship between the measurement signal
and the analyte amount or concentration. Calibration allows the analyst to
estimate the accuracy of a measurement, procedure, or instrument.
Calibration is critical to GLP and method development and validation.
Method development is the process to determine the experimental conditions
for sample collection, preparation, and measurement that produce accurate
and repeatable results. Very often this work has been done by someone else
and you are using one or more standard methods. When a standard method is
not available, analysts usually start with a method for an analyte in a similar
type of sample and modify it to obtain good results. Method validation is the
process to ensure that you are obtaining accurate and repeatable results. To
quote 21 CFR 211.165 (e) 22:
22
Part 211—Current Good Manufacturing Practice for Finished Pharmaceuticals.
calibration 29
Most of the analytical methods that we will discuss assume that the measured
signal is directly proportional to an analyte amount or concentration. The
generic expression of a direct or linear proportionality is
y = a1x (1.14)
100 g
50 g
5g
5g
Calibration weights
This calculation is a one-point calibration. You expect 0.0 mS/cm for the
blank, but there is some residual salt in the tap water. To correct the stan-
dard and unknown measurements, subtract the nonzero blank from each
value. The corrected measurements are 9.21 mS/cm for the standard and
5.45 mS/cm for the soil sample. There are two approaches to work with a
proportionality: determine the sensitivity, a1, or set up a simple ratio. The
two calculations are equivalent, so use whichever one is most intuitive for
you. I will show the calculation first by finding the sensitivity and then using
23
The potentiometric methods discussed in Chapter 8 are one notable exception.
Another random document with
no related content on Scribd:
XIII
Then, at the drooping of the twilight hour,
We wander in the ancient plaza where
We breathe the attar of the jasmine flower
Like incense on the altar of the air;
And list, as music swells
Down drifting from the old cathedral tower,
The arpeggio of the bells.
XIV
We linger by the sea-wall while the tide
Below us murmurs like a sad refrain,
Bearing from outer ocean reaches wide
The lore and legend of the Spanish main,
Nor leave that spot serene
Till Sleep, as with the mantle of the bride,
Wraps fair Saint Augustine.
XV
Days dedicate to rapturous things were these;
It was as though Youth came again, and brought
Past aims, past ardors and past ecstasies,
And toward the shrine of Beauty turned our thought.
And there were after times
Of exultation, prismic harmonies,
When hours ran by in rhymes.
XVI
Once, ’mid cathedral Carolinian pines,
We saw the Springtide, at its radiant birth,
Kindle to fragrant gold the coiling vines,
And make a garden of the wakened earth;
And every morning heard
Within the treetops, melody linked with mirth,
The hidden mocking-bird.
XVII
And while the cardinal through the waving bredes
Of pendulous moss swift flitted like a flame,
Back flooded to our minds the illustrious deeds,
Emblazoned on the honor-scroll of Fame,
When Liberty was won,
Hearkening the Ashley whisper to its reeds
The name of Marion.
XVIII
From Gloucester cliffs and brown Nantucket dunes
The mountains lured you, and the mountain star;
For us the Woodland sang its lyric runes
Where’er we followed it, or near or far,
In sun or shadow cool,
Or loitered through long languorous afternoons
By Dian’s darkling pool.
XIX
Far up the valley Wittenberg’s vast form,
Its summit beckoning, with you I view,
And above sweeping slopes where wild bees swarm
Glimpse timid deer at dawn and fall of dew;
Through Panther Kill we roam,
And mark the purple streamers of the storm
Ascend behind the Dome.
XX
And, too, in bookmen’s mines of dusty ore
Ever shall I remember how we delved,
Plucking from out the musty treasure-store
Rich rarities within the darkness shelved,
Elated if we found
Leaves that some name we long had honored bore
In frayed morocco bound.
XXI
Thus, step by step, we trod adown the years,
Thus, side by side, with ne’er a break between;
We shared our laughter and we shared our tears,
Nor deemed inexorable Fate might intervene
To sever the strong cord
That bound us, Fate with its “abhorrèd shears,”
That is man’s over-lord.
XXII
You that in Autumn came, in Autumn went;
How vain to say the mourning word! how vain
To beat the bars of that arbitrament
That metes to mortals pleasurement or pain!
How vain!—how vain!—and yet
We beat upon them, and we only gain
The poignance of regret!
XXIII
Autumn again with all its loveliness;
Autumn again that brought an end to joy,
Despite the sight of earth in amber dress,
And airs that bear the blitheness of a boy!
Autumn, and leaves that toss
In bright brief triumphing, while they express
The brooding sense of loss.
XXIV
Autumn again down every winding way
That, in the days gone by, our footsteps pressed!—
Instead of woven amaranth would I lay
Above your dust—you gone by paths unguessed—
Love’s deathless asphodel;
Until some happier hour,—when, who shall say?—
Brother in song, farewell!
ONE HUNDRED AND FIFTY COPIES
PRIVATELY PRINTED ON ITALIAN HAND-MADE
PAPER DURING OCTOBER MCMXVII
TRANSCRIBER’S NOTE:
New original cover art included with this eBook is
granted to the public domain.
*** END OF THE PROJECT GUTENBERG EBOOK ELEGY IN
AUTUMN ***
1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside
the United States, check the laws of your country in addition to
the terms of this agreement before downloading, copying,
displaying, performing, distributing or creating derivative works
based on this work or any other Project Gutenberg™ work. The
Foundation makes no representations concerning the copyright
status of any work in any country other than the United States.
1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if
you provide access to or distribute copies of a Project
Gutenberg™ work in a format other than “Plain Vanilla ASCII” or
other format used in the official version posted on the official
Project Gutenberg™ website (www.gutenberg.org), you must, at
no additional cost, fee or expense to the user, provide a copy, a
means of exporting a copy, or a means of obtaining a copy upon
request, of the work in its original “Plain Vanilla ASCII” or other
form. Any alternate format must include the full Project
Gutenberg™ License as specified in paragraph 1.E.1.
• You pay a royalty fee of 20% of the gross profits you derive from
the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information