Optimization in HPLC E-Book

• Herausgeber: Wiley-VCH GmbH
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch

Seitenzahl: 601

Veröffentlichungsjahr: 2021

Beschreibung

Learn to maximize the performance of your HPLC or UHPLC system with this resource from leading experts in the field

Optimization in HPLC: Concepts and Strategies delivers tried-and-tested strategies for optimizing the performance of HPLC and UHPLC systems for a wide variety of analytical tasks. The book explains how to optimize the different HPLC operation modes for a range of analyses, including small molecules, chiral substances, and biomolecules. It also shows readers when and how computational tools may be used to optimize performance.

The practice-oriented text describes common challenges faced by users and developers of HPLC and UHPLC systems, as well as how those challenges can be overcome. Written for first-time and experienced users of HPLC technology and keeping pace with recent developments in HPLC instrumentation and operation modes, this comprehensive guide leaves few questions unanswered.

Readers will also benefit from the inclusion of:

• A thorough introduction to optimization strategies for different modes and uses of HPLC, including working under regulatory constraints
• An exploration of computer aided HPLC optimization, including ChromSwordAuto and Fusion QbD
• A treatment of current challenges for HPLC users in industry as well as large and small analytical service providers
• Discussions of current challenges for HPLC equipment suppliers

Tailor-made for analytical chemists, chromatographers, pharmacologists, toxicologists, and lab technicians, Optimization in HPLC: Concepts and Strategies will also earn a place on the shelves of analytical laboratories in academia and industry who seek a one-stop reference for optimizing the performance of HPLC systems.

Leseprobe

Table of Contents

Cover

Title Page

Copyright

Preface

About the Book

Part I: Optimization Strategies for Different Modes and Uses of HPLC

1.1 2D‐HPLC – Method Development for Successful Separations

1.1.1 Motivations for Two‐Dimensional Separation

1.1.2 Choosing a Two‐Dimensional Separation Mode

1.1.3 Choosing Separation Types/Mechanisms

1.1.4 Choosing Separation Conditions

1.1.5 Method Development Examples

1.1.6 Outlook for the Future

Acknowledgment

References

1.2 Do you HILIC? With Mass Spectrometry? Then do it Systematically

1.2.1 Initial Situation and Optimal Use of Stationary HILIC Phases

1.2.2 Initial Situation and Optimal Use of the “Mobile” HILIC Phase

1.2.3 Further Settings and Conditions Specific to Mass Spectrometric Detection

1.2.4 Short Summary on Method Optimization in HILIC

References

1.3 Optimization Strategies in LC–MS Method Development

1.3.1 Introduction

13.3.2 Developing New Methods for HPLC–MS Separations

13.3.3 Transferring Established HPLC Methods to Mass spectrometry

Abbreviations

References

1.4 Chromatographic Strategies for the Successful Characterization of Protein Biopharmaceuticals

1.4.1 Introduction to Protein Biopharmaceuticals

1.4.2 From Standard to High‐Performance Chromatography of Protein Biopharmaceuticals

1.4.3 Online Coupling of Nondenaturing LC Modes with MS

1.4.4 Multidimensional LC Approaches for Protein Biopharmaceuticals

1.4.5 Conclusion and Future Trends in Protein Biopharmaceuticals Analysis

References

1.5 Optimization Strategies in HPLC for the Separation of Biomolecules

1.5.1 Optimizing a Chromatographic Separation

1.5.2 Optimizing the Speed of an HPLC Method

1.5.3 Optimizing the Sensitivity of an HPLC Method

1.5.4 Multidimensional Separations (See also Chapter 1.1)

1.5.5 Considerations for MS Detection (See also Chapter 1.3)

1.5.6 Conclusions and Future Prospects

References

1.6 Optimization Strategies in Packed‐Column Supercritical Fluid Chromatography (SFC)

1.6.1. Selecting a Stationary Phase Allowing for Adequate Retention and Desired Selectivity

1.6.2. Optimizing Mobile Phase to Elute all Analytes

1.6.3. Optimizing Temperature, Pressure, and Flow Rate

1.6.4. Considerations on SFC–MS Coupling

1.6.5. Summary of Method Optimization

1.6.6. SFC as a Second Dimension in Two‐Dimensional Chromatography

1.6.7. Further Reading

References

1.7 Strategies for Enantioselective (Chiral) Separations

1.7.1 How to Start?

1.7.2 Particle Size

1.7.3 Chiral Polysaccharide Stationary Phases as First Choice

1.7.4 Screening Coated and Immobilized Polysaccharide CSPs in Normal‐Phase and Polar Organic Mode

1.7.5 Screening Coated and Immobilized Polysaccharide CSPs in Reversed‐Phase Mode

1.7.6 Screening Immobilized Polysaccharide CSPs in Medium‐Polarity Mode

1.7.7 Screening Coated and Immobilized Polysaccharide CSPs under Polar Organic Supercritical Fluid Chromatography Conditions

1.7.8 Screening Immobilized Polysaccharide CSPs in Medium‐Polarity Supercritical Fluid Chromatography Conditions

1.7.9 SFC First?

1.7.10 Are There Rules for Predicting Which CSP Is Suited for My Separation Problem?

1.7.11 Which Are the Most Promising Polysaccharide CSPs?

1.7.12 Are some CSPs Comparable?

1.7.13 “No‐Go's,” Pitfalls, and Peculiarities in Chiral HPLC and SFC

1.7.14 Gradients in Chiral Chromatography

1.7.15 Alternative Strategies to Chiral HPLC and SFC on Polysaccharide CSPs

1.7.16 How Can I Solve Enantiomer Separation Problems Without Going to the Laboratory?

1.7.17 The Future of Chiral Separations – Fast Chiral Separations (cUHPLC and cSFC)?

References

Notes

1.8 Optimization Strategies Based on the Structure of the Analytes

1.8.1 Introduction

1.8.2 The Impact of Functional Moieties

1.8.3 Hydrogen Bonds

1.8.4 Influence of Water Solubility by Hydrate Formation of Aldehydes and Ketones

1.8.5 Does “Polar” Equal “Hydrophilic”?

1.8.6 Peroxide Formation of Ethers

1.8.7 The pH Value in HPLC

1.8.8 General Assessment and Estimation of Solubility of Complex Molecules

1.8.9 Octanol–Water Coefficient

1.8.10 Hansen Solubility Parameters

1.8.11 Conclusion and Outlook

Acknowledgments

References

1.9 Optimization Opportunities in a Regulated Environment

1.9.1 Introduction

1.9.2. Preliminary Remark

1.9.3. Resolution

1.9.4. Peak‐to‐Noise Ratio

1.9.5. Coefficient of Variation, VC (Relative Standard Deviation, RSD)

References

Part II: Computer‐aided Optimization

2.1 Strategy for Automated Development of Reversed‐Phase HPLC Methods for Domain‐Specific Characterization of Monoclonal Antibodies

2.1.1 Introduction

2.1.2 Interaction with Instruments

2.1.3 Columns

2.1.4 Sample Preparation and HPLC Analysis

2.1.5 Automated Method Development

2.1.6 Robustness Tests

2.1.7 Conclusions

References

2.2 Fusion QbD

®

Software Implementation of APLM Best Practices for Analytical Method Development, Validation, and Transfer

2.2.1 Introduction

2.2.2 Overview – Experimental Design and Data Modeling in Fusion QbD

2.2.3 Analytical Target Profile

2.2.4 APLM Stage 1 – Procedure Design and Development

2.2.5 Chemistry System Screening

2.2.6 Method Optimization

2.2.7 APLM Stage 2 – Procedure Performance Verification

2.2.8 The USP <1210> Tolerance Interval in Support of Method Transfer

2.2.9 What is Coming – Expectations for 2021 and Beyond

References

Part III: Current Challenges for HPLC Users in Industry

3.1 Modern HPLC Method Development

3.1.1 Robust Approaches to Practice

3.1.2 The Classic Reverse‐phase System

3.1.3 A System that Primarily Separates According to π–π Interactions

3.1.4 A system that Primarily Separates According to Cation Exchange and Hydrogen Bridge Bonding Selectivity

3.1.5 System for Nonpolar Analytes

3.1.6 System for Polar Analytes

3.1.7 Conclusion

3.1.8 The Maximum Peak Capacity

3.1.9 Outlook

References

3.2 Optimization Strategies in HPLC from the Perspective of an Industrial Service Provider

3.2.1 Introduction

3.2.2 Research and Development

3.2.3 Quality Control

3.2.4 Process Control Analytics

3.2.5 Decision Tree for the Optimization Strategy Depending on the Final Application Field

3.3 Optimization Strategies in HPLC from the Perspective of a Service Provider – The UNTIE

®

Process of the CUP Laboratories

3.3.1 Common Challenges for a Service Provider

3.3.2 A Typical, Lengthy Project – How it Usually Goes and How it Should not be Done!

3.3.3 How Do We Make It Better? ‐ The UNTIE

®

Process of the CUP Laboratories

3.3.4 Understanding Customer Needs

3.3.5 The Test of an Existing Method

3.3.6 Method Development and Optimization

3.3.7 Execution of the Validation

3.3.8 Summary

Acknowledgments

References

3.4 Optimization Strategies in HPLC

3.4.1 Definition of the Task

3.4.2 Relevant Data for the HPLC Analysis of a Substance (see also Chapter 1.8)

3.4.3 Generic Methods

3.4.4 General Tips for Optimizing HPLC Methods

Reference

Part IV: Current Challenges for HPLC Equipment Suppliers

4.1 Optimization Strategies with your HPLC – Agilent Technologies

4.1.1 Increase the Absolute Separation Performance: Zero Dead‐Volume Fittings

4.1.2 Separation Performance: Minimizing the Dispersion

4.1.3 Increasing the Throughput – Different Ways to Lower the Turnaround Time

4.1.4 Minimum Carryover for Trace Analysis: Multiwash

4.1.5 Increase the Performance of What you have got – Modular or Stepwise Upgrade of Existing Systems

4.1.6 Increase Automation, Ease of Use, and Reproducibility with the Features of a High‐End Quaternary UHPLC Pump

4.1.7 Increase Automation: Let your Autosampler do the Job

4.1.8 Use Your System for Multiple Purposes: Multimethod and Method Development Systems

4.1.9 Combine Sample Preparation with LC Analysis: Online SPE

4.1.10 Boost Performance with a Second Chromatographic Dimension: 2D‐LC (see also Chapter 1.1)

4.1.11 Think Different, Work with Supercritical CO2 as Eluent: SFC – Supercritical Fluid Chromatography (see also Chapter 1.6)

4.1.12 Determine Different Concentration Ranges in One System: High‐Definition Range (HDR) HPLC

4.1.13 Automize Even Your Method Transfer from other LC Systems: Intelligent System Emulation Technology (ISET)

4.1.14 Conclusion

References

4.2 To Empower the Customer – Optimization Through Individualization

4.2.1 Introduction

4.2.2 Define Your Own Requirements

4.2.3 An Assistant Opens Up Many New Possibilities

4.2.4 The Used Materials in the Focus of the Optimization

4.2.5 Software Optimization Requires Open‐Mindedness

4.2.6 Outlook

4.3 (U)HPLC Basics and Beyond

4.3.1 An Evaluation of (U)HPLC‐operating Parameters and their Effect on Chromatographic Performance

4.3.2 “Analytical Intelligence” – AI, M2M, IoT – How Modern Technology can Simplify the Lab Routine

References

4.4 Addressing Analytical Challenges in a Modern HPLC Laboratory

4.4.1 Vanquish Core, Flex, and Horizon – Three Different Tiers, all Dedicated to Specific Requirements

4.4.2 Intelligent and Self‐Contained HPLC Devices

4.4.3 2D‐LC for Analyzing Complex Samples and Further Automation Capabilities (see also Chapter 1.1)

4.4.4 Software‐Assisted Automated Method Development

Abbreviations

References

4.5 Systematic Method Development with an Analytical Quality‐by‐Design Approach Supported by Fusion QbD and UPLC–MS

References

Index

End User License Agreement

List of Tables

Chapter 1-1

Table 1.1.1 Summary of conditions for the IPRP–IPRP separation of dye‐labeled...

Table 1.1.2 Summary of conditions for the HILIC × RP separation of PS20.

Chapter 1-2

Table 1.2.1 Classification of functional HILIC phases and interaction possibi...

Chapter 1-5

Table 1.5.1 Commonly combined LC–separation methods and their orthogonality a...

Chapter 1-7

Table 1.7.1 Comparison of coated and immobilized polysaccharide CSPs

Table 1.7.2 Recommendation for CSPs to be included in a screening.

Chapter 1-8

Table 1.8.1 Interactive forces between molecules.

Table 1.8.2 Impact of functional moieties.

Table 1.8.3 Overview over selected functional moieties and their properties.

Table 1.8.4 Comparison of ethyl acetate and 1,4‐dioxane.

Table 1.8.5 p

K

a

‐Values of selected carboxylic acids.

Table 1.8.6 p

K

a

‐Values of selected amines (p

K

a

‐values of the corresponding ac...

Table 1.8.7 Log 

P

‐values of selected (amino‐)alcohols.

Table 1.8.8 Hansen parameters of paracetamol and selected solvents as well as...

Chapter 2-1

Table 2.1.1 Reversed‐phase columns.

Chapter 2-2

Table 2.2.1 Gradient time ranges for sample workup and chemistry screening.

Table 2.2.2 Trend response goals for the chemistry screening study.

Table 2.2.3 Mean performance goals for the optimization study.

Table 2.2.4 Maximum expected variations in study parameters used in the simul...

Table 2.2.5 Coordinated robustness goals for the optimization study.

Table 2.2.6 Tolerance interval analysis results – receiving lab.

Chapter 3-4

Table 3.4.1 Important points for the definition of the task.

Table 3.4.2 Data of a substance relevant for HPLC analysis.

Table 3.4.3 Plate number and USP tailing of two bases of different strengths.

Table 3.4.4 Acid constants of typical organic acids.

Table 3.4.5 Acidity constants of typical organic bases.

Table 3.4.6 Octanol–water partition coefficient of some compounds and acetoni...

Table 3.4.7 Typical decomposition reactions of dissolved substances.

Table 3.4.8 Typical decomposition reactions of dissolved substances.

Table 3.4.9 General method parameters and rationales.

Table 3.4.10 Enhancements to the generic method.

Table 3.4.11 Basic data for the analysis of butamirate in cough syrup.

Table 3.4.12 Possible difficulties with the HPLC analysis of butamirate in co...

Table 3.4.13 HPLC parameters for the analysis of butamirate in cough syrup.

Table 3.4.14 Structural formula and UV spectrum of pendimethalin.

Table 3.4.15 UV cut‐off of some solvents.

Table 3.4.16 HPLC parameters for gradient method with trifluoroacetic acid.

Table 3.4.17 Step‐by‐step procedure to reduce peak tailing.

Table 3.4.18 Results plate number and USP tailing with different mobile phase...

Table 3.4.19 Conditions cannabinoid analysis.

Chapter 4-1

Table 4.1.1

Chapter 4-2

Table 4.2.1 Exemplary specification sheet for optimization measures on the HP...

Chapter 4-3

Table 4.3.1 Compressibility of common (U)HPLC solvents.

Table 4.3.2 Analytical conditions used for analysis shown in Figure 4.3.3.

Table 4.3.3 Signal, noise, and S/N ratio obtained from analysis of pyrene.

Chapter 4-5

Table 4.5.1 General method development considerations.

List of Illustrations

Chapter 1-1

Figure 1.1.1 Illustration of an instrument configuration typically used for ...

Figure 1.1.2 Illustration of four different modes of 2D‐LC separation.

Figure 1.1.3 Matrix illustrating the compatibility of different separation m...

Figure 1.1.4 Illustration of a commonly used approach to estimate the fracti...

Figure 1.1.5 Identification of an impurity in a synthetic dye–labeled oligon...

Figure 1.1.6 Separation of the constituents of commercial PS20 using LC × LC...

Chapter 1-2

Figure 1.2.1 Scheme of the polarity ranges of separation techniques. Reverse...

Figure 1.2.2 Toluene and 4‐hydroxybenzoic acid as a prominent example for th...

Figure 1.2.3 Classification of stationary HILIC phases (analogous to [2])....

Figure 1.2.4 Typical HILIC gradient with course and composition of the mobil...

Figure 1.2.5 Chromatograms of the separation of 1‐hydroxy‐ and 4‐hydroxybenz...

Figure 1.2.6 Chromatograms of the separation of 1‐hydroxy‐ and 4‐hydroxybenz...

Chapter 1-3

Figure 1.3.1 Peak capacity vs. gradient volume, expressed as a multiple of t...

Figure 1.3.2 Automatically generated sequence table for the automated determ...

Figure 1.3.3 Left: single‐heartcut setup for a dedicated isolation of an unk...

Chapter 1-4

Figure 1.4.1 Common modifications observed in therapeutic proteins and chrom...

Figure 1.4.2 Analysis of 10 representative mAbs using generic salt gradient ...

Figure 1.4.3 Glycan analysis of digested/reduced trastuzumab subunits under ...

Figure 1.4.4 IEX‐MS analysis of different mAb samples with pI values ranging...

Figure 1.4.5 Possible combinations of chromatographic approaches in the firs...

Figure 1.4.6 CEX × RPLC–MS and HILIC × RPLC–MS profiles of cetuximab after r...

Figure 1.4.7 Multidimensional LC–MS/MS setup consisting in IEX in the first ...

Chapter 1-5

Figure 1.5.1 Separation of released and labeled

N

‐glycans from a monoclonal ...

Figure 1.5.2 Separation of intact NIST antibody via pH‐gradient‐based CEX ch...

Figure 1.5.3 Separation of intact infliximab via pH‐gradient based CEX chrom...

Figure 1.5.4 Ion pair (IP)‐RP‐HPLC separation of trastuzumab and infliximab ...

Figure 1.5.5 HILIC separation of released and labeled

N

‐glycans of a monoclo...

Figure 1.5.6 Separation of trastuzumab charge variants within only three min...

Figure 1.5.7 SEC separation of a monoclonal antibody‐based protein with two ...

Figure 1.5.8 Important steps in setting up an HPLC method. Shown are the thr...

Chapter 1-6

Figure 1.6.1 Packed‐column SFC instrument with the different features to opt...

Figure 1.6.2 Enantioselective stationary phases principally used in chiral S...

Figure 1.6.3 Typical stationary phases employed in achiral SFC ranked accord...

Figure 1.6.4 Features of the mobile‐phases employed in SFC: composition in a...

Figure 1.6.5 The effects of back‐pressure, oven temperature and mobile‐phase...

Figure 1.6.6 Proposed method development process in achiral and chiral SFC....

Chapter 1-7

Figure 1.7.1 Application areas of enantioselective chromatography.

Figure 1.7.2 Structure of rac‐Norketotifen.

Figure 1.7.3 Enantiomer separation of AHC 2102224 and AHC 2082728 on Chiralp...

Figure 1.7.4 Systematic screening for polysaccharide CSPs in normal‐phase mo...

Figure 1.7.5 Enantiomer separation of a nicotinoyl derivative on Chiralcel O...

Figure 1.7.6 Systematic screening for polysaccharide CSPs in reversed‐phase ...

Figure 1.7.7 Enantiomer separationof cyproconazole on CHIRAL ART Cellulose‐S...

Figure 1.7.8 Systematic screening immobilized polysaccharide CSPs in the med...

Figure 1.7.9 Baseline resolution of HCQ enantiomers can be achieved under fo...

Figure 1.7.10 Enantiomer separation of

D

‐ and

L

‐ABGA on Chiralcel OJ‐3 Rt

D‐A

...

Figure 1.7.11 Systematic screening polysaccharide CSPs in the polar organic ...

Figure 1.7.12 Systematic screening of immobilized polysaccharide CSPs in the...

Figure 1.7.13 Enantiomer separation of OTL38 on Chiralpak ZWIX(+): 150 × 3.0...

Figure 1.7.14 Finding separations for rac fenoxaprop‐

p

‐ethyl.

Figure 1.7.15 Separation of spiked plasma (0.5 mg/ml of racemic ketorolac an...

Chapter 1-8

Scheme 1.8.1 Examples for hydrogen bonds. (a) Between water and ethanol. (b)...

Scheme 1.8.2 Hydrate formation of carbonyl compounds.

Scheme 1.8.3 Radical chlorination of acetyl pyridine.

Scheme 1.8.4 Dimethyl formamide.

Scheme 1.8.5 Formation of ether hydroperoxides.

Figure 1.8.1 pH dependency of RP retention for acids and bases.

Scheme 1.8.6 Lincomycin.

Scheme 1.8.7 Streptomycin.

Scheme 1.8.8 (Amino)alcohols.

Scheme 1.8.9 Paracetamol.

Figure 1.8.2 3‐D‐illustration of Hansen solubility parameters of paracetamol...

Chapter 1-9

Figure 1.9.1 Separation can become worse or better when using different air ...

Figure 1.9.2 By turning the column upside down, double peaks are prevented, ...

Figure 1.9.3 By reducing the time constant from 1 seconds to 50 ms, lower ch...

Figure 1.9.4 Coefficients of variation in the evaluation over the peak heigh...

Figure 1.9.5 Coefficient of variation of peak height and peak area as a func...

Figure 1.9.6 Coefficient of variation of peak height and peak area as a func...

Figure 1.9.7 Coefficient of variation of peak height and peak area as a func...

Figure 1.9.8 VC depending on the used data rate recording (“Sample Rate”); f...

Chapter 2-1

Figure 2.1.1 Schematic overview of the HPLC method development system.

Figure 2.1.2 Limited proteolysis of IgG1 by IdeS.

Figure 2.1.3 A method development workflow for early stages of drug developm...

Figure 2.1.4 A method development workflow for late stages of drug developme...

Figure 2.1.5 (a) Chromatogram after screening. Column Agilent Advanced Bio R...

Figure 2.1.6 Results of the rapid optimization.

Figure 2.1.7 Chromatogram after the fine optimization. Column temperature = ...

Figure 2.1.8 Effect of temperature on the resolution of the critical pair.

Figure 2.1.9 Effect of flow rate on the resolution of the critical pair.

Figure 2.1.10 Temperature – flow rate effect of the resolution of the critic...

Figure 2.1.11 Effect of temperature and gradient time on the resolution of t...

Figure 2.1.12 Effect of the flow rate and temperature at increased gradient ...

Figure 2.1.13 Chromatogram of the final method. Temperature = 68 °C, flow ra...

Chapter 2-2

Figure 2.2.1 Example chromatogram at start of chemistry screening study.

Figure 2.2.2 Screening Trellis graph series – three stationary phases.

Figure 2.2.3 2D and 3D resolution map graphs.

Figure 2.2.4 2D overlay graph.

Figure 2.2.5 Example of method with mean response = 2.00 and

C

pk

 = 1.33.

Figure 2.2.6 (a) Example of MODR established for mean performance and robust...

Figure 2.2.7 (a) and (b) Example replication strategies and associated T.I. ...

Chapter 3-1

Figure 3.1.1 POPLC Scheme

Figure 3.1.2 Resolution equation:

RS

, resolution,

N

, theoretical plate count...

Figure 3.1.3 Overview of the selectivity of the individual separation system...

Chapter 3-2

Figure 3.2.1 Pathway for the development of an optimized method.

Chapter 3-3

Figure 3.3.1 The UNTIE pyramid.

Figure 3.3.2 Schematic procedure of an automated “Method Scouting.”

Figure 3.3.3 Schematic structure of the Shimadzu Nexera Method Scouting Syst...

Figure 3.3.4 Easy creation of an analysis plan with the “Shimadzu Scouting S...

Figure 3.3.5 Automated evaluation of the “Scouting” runs and graphic display...

Figure 3.3.6 The UNTIE process of the CUP laboratories.

Chapter 3-4

Figure 3.4.1 Protonation equilibrium of aniline in water.

Figure 3.4.2 Proportion of the two aniline species as a function of pH value...

Figure 3.4.3 Chromatograms of aniline and phenol at different pH values.

Figure 3.4.4 UV spectra of aniline at different pH values.

Figure 3.4.5 Chromatograms of aniline and benzylamine at pH 2.3.

Figure 3.4.6 Acidity constants of cetirizine calculated with MarvinSketch.

Figure 3.4.7 Proportion of acetonitrile in the mobile phase to obtain

k

′ of ...

Figure 3.4.8 Retention times as a function of the log 

P

OW

(column: Nucleosil...

Figure 3.4.9 UV spectra of formic and acetic acid.

Figure 3.4.10 Structural formula of prednisolone.

Figure 3.4.11 Chromatogram of a sample solution of butamirate cough syrup.

Figure 3.4.12 Chromatogram of pendimethalin at 238 nm (a) and 430 nm (b).

Figure 3.4.13 UV spectra of some solvents (measured against water).

Figure 3.4.14 UV spectra of phosphoric acid.

Figure 3.4.15 UV spectra of formic acid and ammonium formate.

Figure 3.4.16 UV‐spectra of acetic acid and ammonium acetate.

Figure 3.4.17 UV spectrum trifluoroacetic acid.

Figure 3.4.18 Chromatogram of a blank injection at different wavelengths.

Figure 3.4.19 Chromatograms of benzylamine with different mobile phases.

Figure 3.4.20 Chromatograms of benzylamine with alkaline mobile phases.

Figure 3.4.21 Chromatographic profile HPLC‐DAD.

Figure 3.4.22 Chromatographic profile UHPLC–HRMS (BPI, ESI positive).

Chapter 4-2

Figure 4.2.1 The mindset of the user determines which results are obtained f...

Figure 4.2.2 Optimization levels that are taken into account.

Figure 4.2.3 The assistant AZURA ASM 2.2L from KNAUER Wissenschaftliche Gerä...

Figure 4.2.4 AZURA HPLC system with assistant with two‐column switching valv...

Figure 4.2.5 Schematics of the fractionation valve (a) and continuous chroma...

Figure 4.2.6 A continuous, simulated moving‐bed (SMB) system from KNAUER Wis...

Figure 4.2.7 Shown are the possible materials for the rotor seal (a) and the...

Chapter 4-3

Figure 4.3.1 Effect of compressibility setting on pump pressure pulsation. (...

Figure 4.3.2 Recommended compressibility setting for a premixed mobile phase...

Figure 4.3.3 Effect of sample solvent and injection volume on peak shape.

Figure 4.3.4 “Co-injection” settings in the LabSolutions software.

Figure 4.3.5 Effect of co-injection of water on peak shape of caffeine in me...

Figure 4.3.6 Visualization of the effect of slit width on the UV spectrum of...

Figure 4.3.7 (a) Air bubble trapped in flow path. (b) Illustration of auto‐d...

Figure 4.3.8 Real‐time mobile‐phase monitoring, using weight sensors and a s...

Figure 4.3.9 Integration example of overlapping peaks using (a) a valley‐to‐...

Figure 4.3.10 Data matrix (D) showing spectral data (S) and peak profiles (C...

Figure 4.3.11 Illustration of the i‐PDeA function for the individual quantif...

Figure 4.3.12 Use of IoT for resource optimization in an analytical laborato...

Chapter 4-4

Figure 4.4.1 The Vanquish Duo‐based Dual LC concept, (a) Fluidic configurati...

Figure 4.4.2 Vanquish Core features for seamless adjustment of system GDV, (...

Figure 4.4.3 Loop‐based single‐heart‐cut 2D‐LC setup. The switching valve in...

Figure 4.4.4 Loop‐based multi‐heart‐cut 2D‐LC setup with the Vanquish platfo...

Figure 4.4.5 Trap‐based single‐heart‐cut 2D‐LC setup for eluent strength red...

Figure 4.4.6 Trap‐based single‐heart‐cut 2D‐LC setup for eluent strength red...

Figure 4.4.7 Example for use of setup from Figure 4.4.6 in two different mod...

Figure 4.4.8 Fluidic setups for automated method scouting, with (a) single‐c...

Figure 4.4.9 ChromSwordAuto

®

‐based method development workflow for cate...

Chapter 4-5

Figure 4.5.1 Chemical structures of budesonide (a) and formoterol fumarate (...

Figure 4.5.2 The plot demonstrates the differences in selectivity of LC colu...

Figure 4.5.3 The “best‐looking” chromatogram from the chemistry screening ex...

Figure 4.5.4 Influence on peak retention of ionizable compounds by the pH of...

Figure 4.5.5 Fusion QbD graph of the design space and the APR obtained from ...

Figure 4.5.6 Four replicate injections of formoterol, budesonide, and its re...

Guide

Cover Page

Title Page

Copyright

Preface

About the Book

Table of Contents

Begin Reading

Index

WILEY END USER LICENSE AGREEMENT

Pages

iv

xv

xvii

1

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

73

74

75

76

77

78

79

80

81

82

83

84

85

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

239

240

241

242

243

244

246

247

248

249

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

311

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

387

388

389

390

391

392

393

394

395

Optimization in HPLC

Concepts and Strategies

Edited by

Stavros Kromidas

Editor

Dr. Stavros Kromidas

Consultant

Breslauer Str. 3

66440 Blieskastel

Germany

All books published by WILEY‐VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.

Library of Congress Card No.:

applied for

British Library Cataloguing‐in‐Publication Data

A catalogue record for this book is available from the British Library.

Bibliographic information published by

the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at <http://dnb.d-nb.de>.

© 2021 WILEY‐VCH GmbH, Boschstr. 12, 69469 Weinheim, Germany

All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law.

Print ISBN: 978‐3‐527‐34789‐6

ePDF ISBN: 978‐3‐527‐82850‐0

ePub ISBN: 978‐3‐527‐82851‐7

Obook ISBN: 978-3-527-83748-9

Cover Design Formgeber, Mannheim, Germany

Preface

The “HPLC world” is a diverse one – a lucky chance and challenging at the same time. Successful strategies for a “good” result can therefore look completely different.

The aim of this book is to provide interested colleagues successful strategies and proven ways for method development and optimization for all important areas in the field of HPLC and UHPLC. With this goal in mind, experts were invited to present their knowledge and experience in a practical and compact manner.

It was important to take both into account: Different challenges of a chromatographic nature, but also different framework conditions in everyday life. Only this enables a differentiated perspective and consequently a target‐oriented approach: Hence, the authors are researchers or employees of well‐known manufacturers, are service providers in industrial companies or private laboratories, or they have developed tools themselves.

Readers may find inspiration in the book for developing their individual optimization strategy.

I would like to thank my fellow authors for their time and commitment as well as WILEY-VCH, who made the realization of this project possible.

Blieskastel, June 2021

Stavros Kromidas

About the Book

The book is designed as a guide and does not have to be read in a linear fashion. The individual chapters represent self‐contained modules; it is possible to “jump” at any time. In this way, we have tried to do justice to the book's character as a reference and hope that readers may benefit from this.

The book consists of four parts:

Part I: Optimization Strategies for Individual Problems

In the first part, optimization strategies for different analytes are discussed, from small molecules and chiral substances to biomolecules. Different modes of operation are also covered: LC–MS, 2D‐HPLC, HILIC, SFC. Finally, optimization strategies based on structural info of the analytes are presented, and optimization possibilities in a regulated environment are discussed.

Part II: Computer‐Aided Strategies (In silico Applications)

In Part II, concepts for computational method development for small molecules and biomolecules are presented, based on specific problems.

Part III: Users' Report

Service providers from two industrial companies and two private laboratories present their concepts for method development in Part III, based on the specifications and requests of internal and/or external customers.

Part IV: Manufacturers' Report

Employees of 5 well‐known HPLC manufacturers show how the design of HPLC instruments, different tools, and the underlying philosophy support HPLC users in establishing the most efficient HPLC method possible, adapted to the problem at hand.

Part IOptimization Strategies for Different Modes and Uses of HPLC

1.12D‐HPLC – Method Development for Successful Separations

Dwight R. Stoll Ph.D.

Gustavus Adolphus College, Department of Chemistry, 800 West College Avenue, St. Peter, MN 56082, USA

1.1.1 Motivations for Two‐Dimensional Separation

Historically, much of the research devoted to multidimensional separations and their application to real analytical problems has been focused on dealing with complex samples. These have traditionally been described as containing hundreds or thousands of compounds and are often derived from natural sources such as plant extracts or body fluids (e.g. blood or urine). Increasingly, however, we observe that multidimensional separation can be exquisitely effective for dealing with samples containing analytes that are difficult to separate but are not complex by the traditional definition. Since this distinction can have a big impact on how one approaches method development, we start here by explicitly differentiating the two cases.

1.1.1.1 Difficult‐to‐Separate Samples

The difficulty associated with a separating a particular sample may originate from its sheer complexity (i.e. thousands of compounds). In this case relying on chromatographic separation alone will not be enough to fully separate the mixture, and some other source of selectivity will be needed (e.g. sample preparation, and/or selective detection such as mass spectrometry). However, it is now common to encounter samples that contain only a few compounds but are difficult to separate simply due to the high degree of similarity of the compounds in the mixture. For example, a mixture may only contain six compounds, but if two of those six compounds are enantiomers (1a and 1b), then fully separating the mixture using a single column may be difficult even if the separation of compounds 2–5 from 1a/1b is straightforward. Such situations are encountered more frequently now compared to the past, in part due to the development of small‐molecule drugs with multiple chiral centers [1], and the increasing recognition of the importance of both the D‐ and L‐ enantiomers of amino acids ([2], see Chapter 1.7), for example.

1.1.1.2 Complex Samples

As stated above, traditionally complex samples have been thought of as containing hundreds or thousands of different compounds. These samples often come from nature, but not always. For example, surfactants and polymers produced by chemical synthesis can result in highly heterogeneous mixtures of thousands of different compounds. Historically, the analysis of such samples by multidimensional chromatography has been mainly focused on so‐called comprehensive methods of separation that yield a kind of global profile or “fingerprint” of the contents of the sample. However, in cases where only one or a few particular molecules in the sample are of importance to the analysis, simpler multidimensional separation methods such as heartcutting can be adequate, and even preferred.

1.1.1.3 Separation Goals

As is often discussed in the multidimensional separation literature, and below, the process of developing a multidimensional separation method is one full of compromises. For example, conditions that favor shorter analysis times do not lead to the best detection sensitivity, and vice versa. Therefore, it is important for the analyst to identify – at the very beginning of method development – what are the characteristics or performance metrics for the method that are most important to him/her. For example, if achieving baseline resolution of six critical pairs of analytes is critically important for the method to be successfully applied, then method development decisions should support this objective, even if it comes at the cost of increased analysis time, and/or lower detection sensitivity.

1.1.2 Choosing a Two‐Dimensional Separation Mode

All two‐dimensional separations can be executed either “offline” or “online.” In the offline mode, one or more fractions of 1D effluent are collected in some kind of storage device such as a set of vials or a wellplate. These fractions are then injected at some later time (minutes to years) into another LC system (i.e. the same LC system running different conditions from the 1D separation, or a different LC system altogether), either with or without intermediate processing of these fractions. For example, in proteomics applications of 2D‐LC, it is common to desalt the fractions, or dry them down by evaporation to remove organic solvent, before analysis by the 2D separation [3]. In the online mode, fractions collected from the 1D column are either processed immediately by direct injection into the 2D column, or stored for a short time (seconds to hours) in some kind of device (typically capillary loops or sorbent‐based traps) that is internal to the instrument. An example of an instrument configuration commonly used for this purpose is shown in Figure 1.1.1. In this case, the interface valve situated between the 1D and 2D columns has two positions. Switching between them changes the roles of loops 1 and 2 between collecting 1D effluent and introducing the fraction of the 1D effluent into the 2D flow stream, effectively injecting that material into the 2D column.

Figure 1.1.1 Illustration of an instrument configuration typically used for 2D‐LC.

Source: Dr. Gabriel Leme.

As commercially available equipment for 2D‐LC separation has become more sophisticated and reliable, the trend in the industry has been to move away from offline separations because of challenges associated with implementation of offline separations for large numbers of samples, and with degradation and contamination of 1D effluent fractions when they are handled external to the instrument [4]. Given this trend, I have chosen to focus entirely on online 2D‐LC for the rest of this chapter. Readers interested in learning more about offline 2D‐LC are referred to review articles dedicated to this topic [5, 6].

1.1.2.1 Analytical Goals Dictate Choice of Mode

Starting in the late 1970s, different groups began developing the modes of 2D‐LC separation we have come to know as “heartcutting” and “comprehensive” [4, 7]. In the most recent decade, two additional modes have been developed, which are now known as “multiple heartcutting” and “selective comprehensive” 2D separations. Each of these four modes will be discussed in some detail in Section 1.1.2.2. At this point, though, I want to emphasize that choosing which separation mode you will use should be driven by the overall goals of the analysis. For example, if you have a complex sample and you want to learn as much as you can about that sample (i.e. identify hundreds of compounds), then the comprehensive mode of 2D separation will almost always be the best choice. However, if you are only interested in a few target compounds in the sample – even if the sample matrix is highly complex – then a more targeted mode of 2D separation such as heartcutting or multiple heartcutting will likely be the best approach. In practice, time spent on each 2D separation is one of the most precious resources of the 2D‐LC instrument, and allocating effort to 2D separations that are not necessary to achieve the overall analytical goals of the analysis is costly (in terms of both time and supplies), wasteful, and adds unnecessary complexity to the method.

1.1.2.2 Survey of Four 2D Separation Modes

The vast majority of 2D‐LC applications being developed today fit into one of the four modes of 2D separation illustrated in Figure 1.1.2. In the single‐heartcut mode (A; LC–LC), a single fraction of 1D effluent containing analytes of interest is captured at the outlet of the 1D column and transferred to the 2D column where this submixture of the original sample will be further separated if the separation mechanisms employed in the first and second dimensions are complementary. Perhaps the biggest advantage of the LC–LC mode is that the time that can be dedicated to separation of the 1D effluent fraction in the second dimension is not strictly limited. This provides tremendous flexibility in terms of choosing parameters for the 2D separation, including flow rate, column dimensions, and injection volume. The biggest disadvantage of LC–LC, however, is that the scope of the analysis is limited. We are restricted to the analysis of compounds that can be captured in a single fraction of 1D effluent. Nevertheless, the LC–LC approach has been used to great effect in application areas ranging from identification of small‐molecule pharmaceutical impurities [8] to the detection of drug metabolites in plasma [9].

Figure 1.1.2 Illustration of four different modes of 2D‐LC separation.

The extreme opposite of LC–LC in terms of analytical scope is the comprehensive mode of 2D separation (D; LC × LC). As the illustration shows, in this case, fractions of 1D effluent are collected and transferred – one at a time, in a regular, serial fashion – to the 2D separation. Typically, this results in a long string of many (tens to hundreds) 2D chromatograms collected in a single detector datafile. This long data string can then be parsed into pieces that correspond to individual 2D separations and reformatted to produce a two‐dimensional data array, which can then be viewed either as a contour map or a 3D surface rendering of the data. The advantages and disadvantages of the LC × LC approach are effectively the converse of those for the LC–LC approach. The main advantage is that the scope of the 2D separation is as wide as the scope of the 1D separation; the main disadvantage is that the time that can be dedicated to each 2D separation is severely restricted because of the sheer number of fractions of 1D effluent that must processed by the second dimension.

The two other modes illustrated in Figure 1.1.2 are hybrids of the LC–LC and LC × LC modes. In the case of multiple heartcutting (B; mLC–LC), one fraction of 1D effluent is collected per region of the 1D separation targeted for further separation, just like LC–LC, but this is repeated two or more times over the course of the 2D separation. Finally, in selective comprehensive separations (C; sLC × LC), multiple fractions of 1D effluent are collected across a zone of interest in the 1D separation, stored in loops or traps associated with the interface, and then injected one at a time into the 2D column as in LC × LC separations. These hybrid modes are attractive in many situations because they capitalize on the strengths of LC–LC and LC × LC while mitigating their weaknesses. Specifically, mLC–LC and sLC × LC provide the analyst with a lot of flexibility in development and implementation of a 2D‐LC method because they provide a means to decouple the process of collecting 1D effluent fractions from the process of further separating those fractions in the second dimension [10].

1.1.2.3 Hybrid Modes Provide Flexibility

There are multiple ways that the added flexibility provided by mLC–LC and sLC × LC is practically useful, but I provide two examples for consideration here. First, sLC × LC is helpful for avoiding the so‐called undersampling problem in 2D separations. Undersampling refers to the negative effect of collecting 1D effluent fractions that are wider than about one‐half of a 1D peak width, whereby analytes eluting closely from the 1D column are mixed back together in the sampling process. This effectively diminishes the performance of the first dimension of a 2D separation [11–13]. Overcoming this problem in the LC × LC mode is especially difficult when 1D peaks are narrow (e.g. less than five seconds wide), but the sLC × LC mode can be used to manage this challenge by collecting several narrow (as low as one second or less) fractions over a particular region of interest in the 1D separation. Second, sLC × LC can also be used to manage the volume of 1D effluent that is injected into the 2D column for each region of interest in the 1D separation. A concrete example will make this benefit more clear. Suppose we have an existing 1D‐LC separation running at 1 mL/min and we want to transfer a particular peak of interest to a 2D column for further separation, and/or characterization by mass spectrometry. If the 1D peak is 15 seconds wide, then the volume of the peak that has to be transferred is 250 μl. While it is certainly possible to transfer this volume in a single fraction, there are many cases where injecting such a large volume into the 2D column will compromise the performance of the 2D separation, especially when there is a mismatch between the mobile phases used in the 1D and 2D separations [14]. With sLC × LC, however, one could collect four fractions of the 1D peak of interest instead of one, with each of the fractions being about 60 μl, and then these four fractions would be injected into the 2D column one at a time [15]. This, of course, is likely to add time to the overall analysis and requires a more complex interface, but this kind of flexibility can be very valuable during method development.

1.1.3 Choosing Separation Types/Mechanisms

Once one has chosen which mode of 2D separation to use, the next most important decision involves choosing which two separation types will be used in the first and second dimensions of the 2D system.

1.1.3.1 Complementarity as a Guiding Principle

There has been much discussion in the 2D separations literature about the principle of “orthogonality” as it is related to choosing two separation types to use in a 2D separation. The reason for invoking orthogonality is that – from a purely theoretical standpoint – it is best if the retention patterns obtained from the 1D and 2D separations are not at all correlated [16]. However, I think it may be more practically relevant to think about the relationship between two separation types used in the 2D separation in terms of complementarity. To what extent does the separation type used in the second dimension complement the separation already used in the first dimension? A concrete chemical example will help make this point. Suppose we are separating a mixture of peptides that vary in both the total number of amino acids and the number of lysine residues such that the degree of positive charge on these peptides in solution varies as well (low pH). If we set up a 2D‐LC system with reversed‐phase C18 columns and low pH mobile phases in both dimensions, this will not yield an effective 2D separation because the 2D separation does not add anything new to the separation in terms of selectivity. On the other hand, suppose we change the 1D separation to cation‐exchange (CEX) where peptides will elute mainly according to their degree of positive charge (low charge elutes first, high charge elutes last). Now, if we add a 2D separation using a reversed-phase (RP) C18 column, this will nicely complement the 1D separation because it will separate mainly according to the water solubility of the peptides (most soluble elutes first, least soluble elutes last). In this case, we can have two peptides that carry the same charge – and thus coelute from the CEX separation – but have very different water solubilities due to differences in the number and/or type of amino acids, and can be easily separated by the 2D RP column.

Historically, a lot of effort has been dedicated to learning which separation types are most complementary for different sample types and applications. New users can use this prior research as a foundation for their own work. For some application areas, there are specific papers that illustrate the complementarity of different separation types for specific types of molecules such as peptides [17]. I encourage readers to consult databases of 2D‐LC applications to quickly learn about which two separation types might be useful for their application (http://www.multidlc.org/literature/2DLC-Applications).

1.1.3.2 Pirok Compatibility Table

Unfortunately, we need to consider more than just the complementarity of the selectivities of two separation types used in a 2D‐LC separation. Other factors such as the compatibility of the mobile phases used with each separation type are often important, and in fact can render useless a pairing of separation types that looks quite attractive from the point of view of selectivity. For example, pairing a normal‐phase (NP) separation (i.e. bare silica stationary phase; hexane mobile phase) with an RP separation is attractive for some applications because the NP separation is dominated by adsorptive analyte–stationary phase interactions, whereas the RP separation is dominated by the partitioning of analytes into a bonded stationary phase. This difference in retention mechanisms can lead to highly complementary selectivities. However, we encounter a major practical difficulty in this case because the nonpolar organic solvent–rich mobile phases used for NP separations are not miscible with the water‐rich mobile phases used for RP separations – at least not across a wide range of compositions. This difficulty has limited the use of some combinations of separation types such as NP–RP, although even in this case the miscibility problem can be managed by injecting very small volumes of 1D effluent into large 2D columns [18]. Pirok and Schoenmakers have summarized a lot of the knowledge in the 2D‐LC field about which separation types work well together using the table shown in Figure 1.1.3. Combinations shaded with green colors are likely to work well, whereas combinations shaded with red colors present at least one major difficulty that will have to be managed if they are chosen for a 2D separation. Readers interested in a more detailed explanation of all of the information in this table are referred to the original paper of Pirok and Schoenmakers [19]. The table is also being updated as 2D‐LC technology evolves; a current version can be found at our website (www.multidlc.org/megatable).

1.1.3.3 Measuring the Complementarity of Separation Types

Once we have made initial decisions about which two separation types to use in our 2D‐LC separation, we need to assess the quality of the resulting separations. For more targeted separations, usually we are most interested in resolving one or more target compounds from the sample matrix, or from themselves. In this case, it is a straightforward matter to evaluate the extent to which the 2D separation has resolved compounds that coeluted from the 1D column, and thus the complementarity of the two separation types. For more comprehensive separations, we are usually interested in the extent to which the 1D and 2D separations – working together – spread the constituents of the sample out across the entire separation space. The need to assess this has led many groups to develop a variety of metrics, which have been critically discussed and compared in recent articles [20, 21]. In our own work, we have used the approach illustrated in Figure 1.1.4, which amounts to estimating the fraction of the available 2D separation space that is occupied by peaks by counting the number of bins that are occupied by peaks and dividing by the total number of available bins in the space. During method development, we adjust elution conditions in both dimensions to spread the peaks out as much as possible with the goal of reaching 100% coverage of the space. This is rarely achieved in practice, but some applications are able to achieve greater than 90% coverage [23].

Figure 1.1.3 Matrix illustrating the compatibility of different separation modes when used in the first or second dimension of 2D‐LC systems.

Source: Reprinted with permission from ref. Pirok et al. [19]. Licensed under CCBY 4.0.

Figure 1.1.4 Illustration of a commonly used approach to estimate the fraction of 2D separation space that is occupied by peaks when analyzing real samples. The dark perimeter captures the bins where peaks can reasonably appear (outside of this perimeter lie the dead times and re‐equilibration times associated with the first and second dimensions).

Source: Adapted from Davis et al. [22].

1.1.4 Choosing Separation Conditions

Newcomers to multidimensional separations often describe being overwhelmed by the number of variables to consider when developing a 2D method, and the apparent complexity of 2D separations in general. While it certainly is true that there are more variables to consider in a 2D method than a 1D one, I also think it is possible to approach method development in a methodical way that should prevent being overwhelmed. Here, we consider method development from two distinct starting points: (i) cases where a 1D method already exists and we want to turn that into the 1D separation of a 2D method; and (ii) cases where we want to start building a 2D method from scratch, such that we are not constrained by the parameters of an existing 1D method.

1.1.4.1 Starting with Fixed First‐Dimension Conditions

Suppose we aim to develop a 2D method by adding a 2D separation to an existing 1D method. This is common in cases where a 1D method is used to separate and determine the concentrations of several impurities in an active ingredient. If this method relies on UV detection, and suddenly a new peak appears, the first question will be – what is that new peak? This question can often be addressed quickly using a single‐heartcut approach (LC–LC) and adding a 2D separation that is compatible with MS detection. This enables capture of the new, unknown peak, transfer of that peak to the 2D column, separation of the constituents of the fraction from themselves and/or the 1D mobile phase buffer, and eventually elution and detection by mass spectrometry. Development of this type of method is straightforward in principle, but as soon as we start looking at the variables we quickly run into a problem. Suppose the existing 1D method runs at a flow rate of 1 ml/min, and the target unknown peak is five seconds wide at half‐height. If we want to capture all of that peak, and add a little margin on the front and back sides of the peak to account for slight retention shifts from one analysis to the next, then we should collect a fraction of 1D effluent that is about 20 seconds wide (8σ peak width, plus 10% on each side). Given the flow rate of 1000 μl/min, this corresponds to a fraction volume of about 333 μl. The volume of this fraction, which becomes the injection volume for the 2D separation, is large compared to even the largest analytical columns (e.g. 150 mm × 4.6 mm i.d.) in common use today, and can have a serious negative impact on the quality of the 2D separation. This is the first point where an important decision must be made. We have three main options; readers interested in more detail on each approach are referred to the recent literature [14, 24, 25]:

(1) Inject the entire fraction as it is into a large

2

D column, and hope for the best. This will work well in cases where the target analyte elutes from the

1

D column in effluent that is a weak solvent for the

2

D column (e.g.

1

D effluent that is 10/90 acetonitrile (ACN)/water injected into a

2

D mobile phase that is 20/80 ACN/water), but this is not always possible or convenient.

(2) Inject a smaller fraction of the

1

D peak to make the fraction volume more manageable (e.g. 40 μl). This approach will address the injection volume problem, but potentially introduces new problems: (i) one runs the risk of missing analytes that elute in the front or the tail of the target

1

D peak; and (ii) the repeatability of this approach will not be good because it will be sensitive to small shifts in

1

D retention time.

(3) Finally, the option that gives the analyst the most control is to use an active modulation technique to manage impact of the large fraction volume on the

2

D separation. The most common approaches are: (i) dilution of the

1

D effluent with weak solvent using an additional pump (e.g. diluting with water in the case of RP separation in the second dimension); (ii) Active solvent modulation using a valve to adjust the properties of the

1

D effluent fraction; and (iii) using a sorbent‐based trap to separate the analytes of interest from the

1

D effluent matrix prior to injection into the

2

D column.

Once the 1D effluent fraction volume and the approach to transfer this fraction to the second dimension are chosen, the dimensions and conditions for the 2D column can be established. In the case of a LC–LC separation like the one described in this scenario, the analyst has a lot of flexibility in the second dimension, and most of the guidelines we use for choosing columns and conditions in 1D‐LC apply here too. The column length, particle size, and flow rate are strongly dictated by the analysis time for the 2D separation, and the theory to guide these choices is readily available and easily applied [26]. Second‐dimension separations that are long and require high resolution will benefit from long columns (>100 mm). On the other hand, separations that are short, or in cases where only a crude 2D separation is needed (e.g. desalting applications), short columns (<100 mm) will suffice. When using UV detection in the second dimension, larger diameter columns (>3.0 mm i.d.) and higher flow rates (>1 ml/min.) are desirable because these conditions will mitigate the effect of the large volume of 1D effluent that is injected. On the other hand, when MS detection is used in the second‐dimension, smaller‐diameter columns (<2.1 mm i.d.) and lower flow rates (<1 ml/min) are desirable to avoid flooding the MS inlet with solvent.

1.1.4.2 Starting from Scratch – Flexible First‐Dimension Conditions

When developing a new 2D‐LC method from scratch, we don't have as many constraints to deal with as we do when adding a 2D separation to an existing 1D‐LC separation. In most cases, users leverage this freedom by reducing the flow rate of the 1D separation so that the volume of 1D effluent transferred to the 2D column is not so large as in the scenario described above. This is especially helpful in the case of LC × LC separations, where 2D columns tend to be small to facilitate fast 2D separations. In this case, a typical 1D flow rate is 50 μl/min. However, even for LC–LC and hybrid 2D‐LC separations, dealing with a 1D flow rate of 200 μl/min is a lot easier than dealing with 1000 μl/min.

1.1.4.3 Special Considerations for Comprehensive 2D‐LC Methods

As illustrated in Figure 1.1.2d, in comprehensive 2D‐LC separations, many fractions of 1D effluent are collected and transferred to the 2D column over the course of a single 2D‐LC analysis. This means that all of the steps associated with a single 2D separation happen tens or hundreds of times over the course of an LC × LC analysis, including one or two valve switches per fraction, and typically a change in mobile‐phase composition to elute the injected components. Most useful LC × LC separations are on the order of 30 minutes to 3 hours in length, and the timescale of each 2D separation is usually 15–120 seconds. This, in turn, requires that the second dimension is operated with short (<50 mm), narrow (2.1 mm i.d.) columns, and high (>1 ml/min) flow rates. Of course there are exceptions to these trends, but these conditions are typical. In our work, we have observed that both the design of the valve used to transfer 1D effluent fractions to the second dimension and the pressure at which the 2D column is operated can have a significant impact on the lifetime of 2D columns, and therefore these parameters are worthy of serious consideration when developing LC × LC methods [23, 27].

1.1.4.4 Rules of Thumb

(1) When using RP separations in both dimensions, try to use the less retentive column in the first dimension. This will help mitigate the effect of the volume of

1

D effluent transferred to the

2

D column.

(2) Estimate the solvent strength of the

1

D effluent compared to the starting point in each

2

D separation. For RP separations, ensuring that the

1

D effluent contains about 10% less organic solvent will lead to good results more often than not. If this is not the case, then the organic content can be adjusted using active modulation approaches

[24]

.

(3) A good starting point is to use columns of the same diameter for the

1

D and

2

D separations.

1.1.5 Method Development Examples

The scope of this chapter is such that there is not enough space to cover many aspects of method development in detail. I think the next best option is to talk through examples of 2D‐LC methods and explain some of the key method development decisions. This will enable readers to use a similar thought process when developing their own methods.

1.1.5.1 Example 1 – Use of LC–LC to Identify an Impurity in a Synthetic Oligonucleotide

This first application example, which is from the work of Koshel et al., uses a single‐heartcut 2D‐LC method to identify an impurity in a synthetic oligonucleotide by mass spectrometry [28]. Historically, the identification of such impurities by MS has been difficult because the high concentrations of long‐chain amine‐based ion‐pairing agents used for RP separations of oligonucleotides are not very compatible with MS detection. However, this 2D‐LC method enables separation of the impurity from the target oligonucleotide by the 1D column as shown in Figure 1.1.5a, and then separation of the impurity from the long‐chain ion‐pairing agent by the 2D column prior to detection. In this case, the same RP column chemistry is used in both dimensions (i.e., ion-pairing reversed-phase (IPRP)). Although this is unusual in 2D‐LC separations generally, it is effective here because the ion‐pairing agents used in the first and second dimensions are chemically quite different (hexylamine vs. triethylamine plus hexafluoroisopropanol) and give rise to different selectivities. The important parameters for the separation are shown in Table 1.1.1. In this case the 1D and 2D flow rates are quite similar; this is possible because this is a single‐heartcut method, and the speed of the 2D separation is not critical. A single 50‐μl fraction of 1D effluent (0.5 minutes × 0.10 ml/min) is transferred to the 2D column after dilution with aqueous eluent using the 2D pump. This dilution step is important as it reduces the organic solvent content of the sample injected into the 2D column, resulting in excellent peak shape in the 2D chromatogram as shown in Figure 1.1.5b.

1.1.5.2 Example 2 – Comprehensive 2D‐LC Separation of Surfactants

In this second example, we examine the use of LC × LC to separate Tween 20 (also known as polysorbate 20, or PS20), which is a complex mixture of fatty acid esters of ethoxylated sorbitan. The comprehensive mode of 2D‐LC is clearly the best one to characterize a complex mixture like this. The separation shown in Figure 1.1.6, which is from the work of Vanhoenacker et al. [29]