**Meet the editor**

Mark T. Stauffer was born in 1957. He graduated with a Bachelor of Science degree in Chemistry from the University of Pittsburgh in 1979, worked in the industry for 12 years, and then returned to Pittsburgh, receiving a PhD degree in Chemistry in 1998. He joined the Faculty of Chemistry at the University of Pittsburgh at Greensburg in 2001, receiving tenure in 2007. Since 2001, he has

been collaborating with projects in archeology, foods, test kit evaluation, mine drainage, and data analysis. He and coauthors presented over 100 papers and posters at technical conferences and published 13 papers in peer-reviewed journals. He presented a short course on analytical data treatment at the annual Pittcon Analytical Chemistry Conference. He conveys his enthusiasm for research and teaching via mentoring undergraduate research and through his courses in analytical chemistry.

Contents

**Preface VII**

Mark T. Stauffer

**Section 2 Calibration Approaches 15**

**Deliberations 17** Marwa S. El-Azazy

**LC-MS/MS 61**

**Spectroscopy 85**

Dong

**Spectroscopic Methods 35**

Chapter 1 **Introductory Chapter: The Many Faces of Calibration and**

Chapter 2 **Analytical Calibrations: Schemes, Manuals, and Metrological**

Chapter 3 **Multivariate Calibration for the Development of Vibrational**

Chapter 4 **Internal Standards for Absolute Quantification of Large Molecules (Proteins) from Biological Matrices by**

Morse Faria and Matthew S. Halquist

Chapter 5 **Calibration Methods of Laser-Induced Breakdown**

Chapter 6 **Linearity of Calibration Curves for Analytical Methods: A**

Seyed Mojtaba Moosavi and Sussan Ghassabian

Ioan Tomuta, Alina Porfire, Tibor Casian and Alexandru Gavan

Hongbo Fu, Junwei Jia, Huadong Wang, Zhibo Ni and Fengzhong

**Review of Criteria for Assessment of Method Reliability 109**

**Validation in Analytical Methodology in the Present Day 3**

**Section 1 Introduction 1**

## Contents



Chapter 6 **Linearity of Calibration Curves for Analytical Methods: A Review of Criteria for Assessment of Method Reliability 109** Seyed Mojtaba Moosavi and Sussan Ghassabian

### **Section 3 Validation Approaches 129**


Preface

referencing process is known as *calibration*.

at, e.g., particular wavelengths/frequencies or response times.

Ask any analytical chemist what constitutes the most important aspect of any quantitative ana‐ lytical determination, and the most frequent response one might receive is "calibration" and per‐ haps even "calibration and validation." One might ask: "Why would this be so?" If one mulls over this matter for a moment, the answer might become apparent. Consider any analytical in‐ strument, e.g., a UV-visible spectrophotometer, and the absorbance response produced as radia‐ tion from the spectrophotometer's source lamp passes through the solution of absorbing analyte and is absorbed to a particular degree. At face value, that absorbance has no meaning at all *un‐ less* that signal is referenced to a *known* amount of the absorbing analyte present in the measured solution. Now, the absorbance does have meaning, as it can be expressed as a number of absorb‐ ance units per unit amount of the sought-after analyte and may potentially be used for determi‐ nation of analyte concentration in an unknown solution or other types of sample. This

The calibration process establishes a relationship between analyte signal and analyte concentra‐ tion that is useful for quantitation of a sought-after analyte in a sample, using a given analytical method. A typical approach is to measure the signals produced by a series of calibration *stand‐ ards* of known analyte concentrations that cover a particular range of concentration and then re‐ gress the blank-corrected signals (response variable) on the standard concentrations (predictor variable) to obtain the equation (i.e., calibration curve) that best fits the experimental calibration data and yields predicted values of the response variable. This type of approach to calibration is termed *univariate* (one response variable, one or more predictor variables) calibration and can also hold for the opposite arrangement—a response variable of concentration regressed on a pre‐ dictor variable of signal. Also, included under the umbrella of univariate calibration methods is multiple linear regression (MLR), comprised of one response variable (usually concentration) re‐ gressed into two or more predictor variables (usually signal). Another broad class of calibration model, known as *multivariate* (two or more response and predictor variables each) calibration, involves the use of data arrays for the response and predictor variables, with the response ma‐ trix regressed on the predictor matrix, using the principles of linear algebra and statistics. Most multivariate calibrations are designed so that the response matrix is the array of concentration vectors for the calibration standards containing the sought-after analytes and the predictor ma‐ trix is the array of signal vectors for, e.g., spectra or chromatograms of the calibration standards

The *validation* process is equally important, as it verifies the signal-versus-concentration relation‐ ship acquired from the calibration samples via analysis of another, separate set of samples with known concentrations of analyte. A set of predicted analyte concentrations for the validation samples are then obtained using the calibration relationship for the method employed. The known and predicted concentrations of the validation samples are compared, and the *residuals* (errors or differences between known and predicted concentration) are analyzed statistically, e.g., by standard deviation and bias, to evaluate the accuracy and precision (i.e., the reliability), respectively, of the calibration system and, ultimately, the analytical method used for analyte quantitation. Other validation parameters assess the goodness of fit of the experimental calibra‐
